(PDF) Consumer Finance Research Methods Toolkit

Consumer Finance Research Methods Toolkit Erin B. Taylor and Gawain Lynch For the Institute for Money, Technology & Financial Inclusion The Consumer Finance Research Methods Project is a collaborative research project funded by the Institute for Money, Technology, and Financial Inclusion and by the Canela Group. Erin B. Taylor and Gawain Lynch serve as project leads. Information can be found on the IMTFI’s website http://www.imtfi.uci.edu/cfrmp Please send comments and feedback to

Consumer Finance Research Methods Toolkit (BETA version) Authors Erin Taylor, Gawain Lynch with Ursula Dalinghaus Foreword by Bill Maurer April 2016 This work is licensed under the Creative Commons Attribution- ShareAlike 4.0 International License. To view a copy of this license, visit http://creativecommons.org/licenses/by-sa/4.0/ Consumer Finance Research Toolkit About the Institute for Money, Technology & Financial Inclusion The Institute for Money, Technology & Financial Inclusion (IMTFI), established in 2008 with funding from the Gates Foundation, is a research institute based out of the University of California, Irvine. Its core activity has been supporting original research in the developing world on the impact of mobile and digital financial services, undertaken by researchers from the global South. It has built an extensive transnational network of embedded scholars and researchers who focus on developing grounded, nuanced perspectives on people’s everyday financial practices and the impact of new technologies. IMTFI has established itself as the thought leader in this space and is well known in the academic, development, policy, and industry communities involved in emerging payments and their potential to change people's financial lives. To date, IMTFI has supported 147 projects in 47 countries involving 186 different researchers. Those researchers have produced 10 books and 100+ articles in scholarly and other venues, and have been mentioned in the media 170+ times, in venues ranging from Bloomberg Businessweek and the Guardian to Forbes, India. IMTFI has pioneered a complementary approach to researching digital financial inclusion: identifying talented researchers, telling compelling stories, translating local knowledge, and catching insights into the everyday lives of the poor that other research methods miss. You can learn more about their work at www.imtfi.uci.edu. Institute for Money, Technology & Financial Inclusion University of California, Irvine School of Social Sciences 3151 Social Sciences Plaza Irvine, CA 92697-5100 (949) 824-2284 •

Consumer Finance Research Toolkit Table of Contents Foreword 1 About the Toolkit 3 A Note on Ethics 6 Money Methods 16 Ethnography 17 What is it? (17) Examples of use (18) Strengths (18) Limitations (20) Case Study 1 Transactions in Ethiopia: Direct Observations of Use of Cash in Sales (Woldmariam Fikre Mesfin) 21 Method (21) Findings (22) Applications (22) Ethical Issues (22) Case Study 2 Debt in Guatemala: Using Long-Term Ethnography to Collect Data (David Stoll) 23 Method (23) Findings (24) Applications (25) Ethical Issues (25) Consumer Finance Research Toolkit More about the Method 26 Verbal Interviews 30 What is it? (30) Examples of use (30) Strengths (31) Limitations (32) Case Study 1 Learning “good” and “bad” financial behaviours: Semi-structured interviews with Hispanic American college students (US) (Kittichai Watchravesringkan) 32 Method (33) Findings (34) Applications (34) Ethical Issues (35) Case Study 2 Interviews for social network analysis: Mobile money in Kenya (Sibel Kusimba, Yang Yang and Natesh V. Chawla) 35 Method (36) Findings (38) Applications (39) Ethical Issues (40) More about the Method 40 Object-Centred Interviews 42 What is it? (42) Examples of use (42) Strengths (43) Limitations (44) Case Study 1 Combining multiple object-centred methods to understand financial management (US) (Joseph Kaye, Mary McCuistion, Rebecca Gulotta, and David A. Shamma) 44 Method (45) Consumer Finance Research Toolkit Findings (47) Applications (48) Ethical Issues (49) Case Study 2 “Portable kit” studies on poverty and personal finances in Haiti (Erin B. Taylor and Heather A. Horst) 49 Method (50) Findings (51) Applications (53) Ethical Issues (54) More about the Method 55 Financial Diaries 60 What is it? (60) Examples of use (61) Strengths (62) Limitations (63) Case Study 1 The “portfolios of the poor” in Bangladesh, India, and South Africa (Daryl Collins, Jonathan Morduch, Stuart Rutherford and Orlanda Ruthven) 64 Method (65) Findings (66) Applications (67) Ethical Issues (69) Case Study 2 Online financial diaries: A short-term commercial project (US) (Alexandra Mack) 70 Method (70) Findings (71) Applications (72) Ethical Issues (72) More about the Method 72 Consumer Finance Research Toolkit Digital Research 76 What is it? (77) Examples of use (78) Strengths (79) Limitations (80) Case Study 1 Studying Bitcoin using qualitative and quantitative methods (US) (Sarah Meiklejohn, Marjori Pomarole, Grant Jordan, Kirill Levenchko, Damon McCoy, Geoffrey M. Voelker, Stefan Savage) 80 Method (81) Findings (82) Applications (83) Ethical Issues (84) Case Study 2 Combining online and offline data collection on payments in Indonesia (Tom Boellstorff, Khanis Suvianita, Dédé Oetomo, and Nurul Ilmi Idrus) 85 Method (86) Findings (88) Applications (90) Ethical Issues (90) More about the Method 91 Experiments 96 What is it? (96) Examples of use (98) Strengths (99) Limitations (100) Case Study 1 Understanding Risk Preferences and Time Preferences (US) (James Andreoni and Charles Sprenger) 103 Method (104) Findings (105) Consumer Finance Research Toolkit Applications (106) Ethical Issues (106) Case Study 2 Microfinance games: Group lending versus individual lending in Peru (Xavier Giné, Pamela Jakiela, Dean Karlan, and Jonathan Morduch) 107 Method (108) Findings (108) Applications (109) Ethical Issues (110) Case Study 3 Randomized Controlled Trial: Innovative and Interactive Ways to improve the financial capability and savings of women in India (Deepti Kc and Mudita Tiwari) 110 Method (111) Findings (113) Applications (114) Ethical Issues (115) More about the Method 116 Challenges in Consumer Finance Research 119 Acknowledgements 125 Appendix I. Featured IMTFI Blog Posts 127 1. Where is the data? Analyzing customer footprints for better product design (Jacobo Menajovsky) (128) 2. One researcher's thoughts on money and metadata based on fieldwork in Ethiopia (W. F. Mesfin) (129) 3. Understanding the transformative value of Tongan women’s kau tou lālanga: mobile mats, mobile phones, and money transfer agents (Charmaine ‘Ilaiu Talei) (132) Consumer Finance Research Toolkit 4. The Economy of the Quota: The Financial Ecologies and Commercial Circuits of Retail Credit Cards in Santiago, Chile (José Ossandón) (135) 5. Grandmothers as Mobile Money Brokers in Kenya (Sibel Kusimba) (141) 6. Mobile Money Uptake, Savings and Social Networks in Tanzania: A Lesson in Methods (A Burlando, C. Kinnan, and S. Prina) (144) 7. Using the ATM Debit Card to Build Trust and Savings: A Study through Mexico's Oportunidades (Enrique Seira) (148) 8. Hand Held Wealth? A Case of Tigo Money in Bolivia (M. Balderrama and O. Rocabado) (151) II. Sample Research Instruments 155 1. Hand Held Wealth? A Case of Tigo Money in Bolivia (M. Balderrama and O. Rocabado) Footnotes 168 Images Figure 1 “Ediacaran/ Mid Cambrian Diorama” Credit D. W. Miller Figure 2 A spreadsheet with post-it notes for anonymization. Image courtesy of Jofish Kaye Figure 3 Summary pages for two participants, designed by Mary McQuistion. Image courtesy of Jofish Kaye Figure 4 IMTFI Methods Workshop (2014) Photo IMTFI Figure 5 Church Money Donations. Image courtesy of Woldmariam F. Mesfin Figure 6 Ethiopian Currency. Image Courtesy of Woldmariam F. Mesfin Figure 7 Sample document to record wedding gifts. Image courtesy of Woldmariam F. Mesfin Figure 8 Market Scenes, W.F. Mesfin Figure 9 Dorcas (Broker) Social Network Analysis. Image Courtesy of Sibel Kusimba Figure 10 Common items in a portable kit. Image courtesy of Erin B Taylor and Heather A Horst Figure 11 Object Centred Interview, Haiti. Photo by Erin B. Taylor) Figure 12 Financial event cards. Image by Jofish Kaye Figure 13 Financial event cards. Image by Jofish Kaye Figure 14 Financial event cards. Image by Jofish Kaye Figure 15 Bills on microwave—put there until you have to deal with them. Photo by Alexandra Mack Figure 16 Donation receipts for taxes. Photo by Alexandra Mack Figure 17 Bitcoin ATM, Irvine. Photo by Ursula Dalinghaus Figure 18 Screenshot of Bitcoin Wallet. Photo by Ursula Dalinghaus Figure 19 Images from a shopping mall in Surabaya. Photo by Tom Boellstorff Figure 20 Semantic fields for “gadgets” in contemporary Indonesia. Tom Boellstorff Figure 21 Financial Training. Photo courtesy of Deepti Kc Figure 22 Lock Box. Photo courtesy of Deepti Kc. Figure 23 Money Lei, photo by Bill Maurer Consumer Finance Research Toolkit Since 2008, the Institute for Money, Foreword Technology and Financial Inclusion has been supporting research into the lives of Why a Consumer Finance Research the poor around the world as they are, or Methods Toolkit? And why now? There are not, impacted by new mobile and has been an unprecedented expansion of digital payment and financial technologies. financial technology aimed at helping With a global footprint and researchers everyday people save, manage and from emerging markets sometimes transfer money. There are new insurance, working in remote regions, IMTFI has investment and money management provided an archive of data on the products, digital financial education tools intersection of money and technology, and alternative credit scoring and lending traditional and modern financial practices, services. What started with M-Pesa’s and ground-level insights into how people mobile phone-based money transfer really deal with their money and what they service in 2007 has blossomed into might do if they had access to a broader myriad new technologies and platforms for range of financial tools. electronic payment and other digital financial services accessed via phone, This how-to, hands-on guide spotlights the computer, tablet, kiosk, ATM, point-of-sale research methods our researchers and terminal or other device. others have been using successfully to understand people’s money worlds. It We have dubbed this a “Cambrian spotlights ethical and consumer protection explosion”: like the Cambrian period in issues and offers case studies and evolutionary history, when on planet Earth examples of successful research and biological organisms went through an dissemination to show how to turn insight incredible period of diversification into impact. From qualitative to occupying new environmental niches and quantitative, ethnographic to experimenting with new body forms and experimental, an expanded portfolio of types, financial technology is undergoing research methods can grow the a period of profound change. Yet what all conversation on financial inclusion—to this means for everyday people remains make it more broadly inclusive of diverse very much an open question. In some perspectives, peoples and paradigms. places, we see consumers rapidly adopting a new service; in others, the This toolkit could not have been same service simply does not find appeal. completed without the assistance of Investors and startups create new IMTFI’s dedicated staff and in-house products but it is not always clear what researchers, including IMTFI problem they are trying to solve. And cash Administrator Jenny Fan, postdoctoral and coin—one of the most ancient of scholar Dr. Ursula Dalinghaus, analyst human technologies—endure, especially among the world’s poor. 1 Consumer Finance Research Toolkit John Seaman, and graduate research assistants Nima L. Yolmo and Nandita Badami. Bill Maurer Director, Institute for Money, Technology and Financial Inclusion Professor of Anthropology and Law University of California, Irvine Figure 1 “Ediacaran/Mid Cambrian Diorama” Credit D. W. Miller 2 Consumer Finance Research Toolkit / or households using a range of financial About the Toolkit tools (for example, money, bank accounts, cards, transfer services, mobile money, microfinance) to achieve financial goals This toolkit has been produced as part of (such as saving, lending, borrowing, the IMTFI’s Consumer Finance Research investing). Methods Project.1 It demonstrates how different methods are being applied in Generally this definition does not include consumer finance research, across both money management within businesses. for-profit and not-for-profit sectors. It is However, we recognize that there are designed to help researchers and many households around the world that managers to: do not clearly demarcate their household  Learn about innovations taking expenses from their business expenses. We therefore maintain a flexible definition. place in consumer finance research  Broaden their knowledge of the Why a consumer finance strengths and limitations of different methods for toolkit? understanding financial behaviour  Facilitate connections between Consumer finance research across all sectors faces common issues: researchers and organizations who have complementary  Consumer finance products are expertise changing fast globally, and our methods need to be adaptable to Just as consumers have an ever- shifting consumer practices increasing choice of financial products,  Due to the global connectivity of researchers have an ever-increasing array financial systems, providers and of methods at their disposal. This toolkit researchers are finding that provides readers with inspiration for ways collaborations are becoming they can develop their research, either by increasingly helpful to understand themselves or in collaboration with others. consumers Readers can choose to learn about  Financial practices often include applications that are related to their own both qualitative and quantitative expertise, or discover entirely new elements, making it important for methods and the professionals who researchers to have a good basic develop them. knowledge of methods outside their own area of expertise What is consumer finance?  Getting good data while protecting research participants’ privacy is a We define consumer finance as money major challenge where issues of management practices by individuals and 3 Consumer Finance Research Toolkit money and technology are be applied on its own, or combined with concerned other methods, to find answers to some of  There are competing theories the trickiest problems in consumer finance about how people make financial research. decisions, and being familiar with a range of different research Methods methods can help us understand  Time-tested methods, such as why researchers’ explanations ethnography and experiments differ  Innovative methods, such as object-centred interviews and The right research design can help digital research produce rigorous and relevant data, while  Methods developed specifically for ensuring that the privacy of research money research, such as financial participants is protected. We hope that the diaries CFRM Toolkit will assist researchers in discovering new ways to tackle problems Case studies and forge fruitful collaborations.  Showcase all sectors of consumer finance research: commercial, government, academic, non-profit What’s in the CFRM Toolkit?  Show how individual methods are The CFRM Toolkit features six methods applied to particular problems or (qualitative and quantitative) used in can be combined with other consumer finance research. For each methods Resources  Further reading on different method, we describe what the method is, its strengths and limitations, and provide methods  Links to relevant institutions links to further information. We present two case studies for each method (the entry on Experiments showcases three) to conducting innovative consumer finance work  Links to useful data sources for show how each methodological approach can be used to understand aspects of consumer finance. consumer finance research The case studies are drawn from research Who is this toolkit for? carried out in different sectors of consumer finance around the globe: The toolkit is intended for use by anyone academic, commercial, and non-profit who needs to understand how people use research. The research we profile financial products and services: represents a wide range of practices, with data collected in labs, in the field, from  Managers data sets, and through experiments. The  Individual professionals case studies show how each method can  Research and design teams 4 Consumer Finance Research Toolkit  Organizations and companies Example of use II: Using the toolkit to  Instructors and students choose the right experts User insight specialists, designers, NGO Monica is a manager in a medium-sized workers, policy specialists, and academic NGO. Her team has been using researchers are among those who may household survey data to study benefit from the toolkit’s descriptions of conditional cash transfers in Ecuador, but how different methods are applied to a she is dissatisfied with the results. She wide range of problems around the world. suspects that the primary cash flow isn’t from government to household, but from Whether you work in the field, in a lab, or household to household. Looking through at home on your notebook, this toolkit this toolkit, she decides that her covers methods that are relevant to your organization needs to do a financial diary research context. study along with some ethnography to get the information they need. Monica isn’t an expert in these areas, so she contacts Example of use I: Using the toolkit to relevant professionals to form choose the right methods collaborations. David is an anthropologist working in a research institute. For ten years he has been studying informal financial practices How to use the toolkit among residents of the south Pacific, There are multiple ways that you can use primarily using a combination of surveys this toolkit: and third party data. Recently, the majority  Read up on individual methods to of his research participants have begun to learn the basics use online banking and, in some cases,  Get tips on how methods can be mobile money. He realises that he is used in combination going to have to adapt his methods in  Browse the case studies to learn order to understand how these new about innovations in consumer changes are affecting financial practices. finance  Explore the “examples of use” and David browses through this toolkit and “more about the method” sections decides to try a combination of interviews to find methods guides, data and digital research. Since he has some sources, and potential prior experience in these areas, he is able collaborators to continue to conduct most of the research by himself, occasionally calling upon other experts for advice. 5 Consumer Finance Research Toolkit applied, who applies them, and what the A Note on Ethics research will be used for. Designing research that is sensitive to Ethical guidelines and the review process ethical issues can be tricky. In consumer vary depending on whether the research finance research, ethical issues can be is done by a commercial, non-profit, particularly difficult to navigate. Most government, or academic institution, since people consider money to be a private each of these bodies will carry out and matter, and releasing one's personal apply research in different ways. financial information can be both embarrassing and dangerous. However, many considerations are common to all research, and across all People may be ashamed to reveal the sectors. Permission and procedures for state of their finances to a stranger, or informed consent, privacy, psychological they may worry that their financial comfort, data storage, and data sharing information will be leaked to neighbours, are just some of these common government officials, or companies. considerations. Millions of people depend upon accessing financial services such as credit and In this section we review the main ethical insurance, and the release of personal issues that affect consumer finance data can potentially affect their ability to research and suggest ways to begin access those services. addressing ethical considerations beginning with research design. To learn For these (and many other) reasons, it is more about the ethical implications of a important to consider ethical issues at the particular method, you can refer to the outset of research design. Importantly, ethics discussions attached to each case academic institutions and US Federal study in this toolkit. regulations require human subjects review and clearance, as well as training of Ethical codes and review project personnel in human subjects protection, before data collection can boards begin. Ethical codes are developed to protect the autonomy, safety, privacy, and welfare of Ethical issues vary depending upon who human subjects. Ethical issues arise will be included as research subjects in regardless of whether researchers interact the study. Are these adults who can with human subjects directly (e.g., in provide informed consent or do these interviews) or indirectly (e.g., statistical include subjects considered to be “at-risk analysis of a database). The methods of populations” such as children or the data access differ, but the general elderly? Ethical considerations inform, principles remain the same. which methods are used, how they are 6 Consumer Finance Research Toolkit Fortunately, these days there is plenty of The Belmont Report outlines three key advice available online as to how to principles in research with human design ethical research, receive training in subjects: human subjects protection, and be aware of the institutional, legal, and regulatory 1. Respect for persons: requirements for research with human Researchers must respect subjects. people's autonomy, treat them with courtesy, seek informed consent, Many researchers, especially within be truthful, and conduct no universities, can call on official bodies for deception. practical guidance. In most countries, 2. Beneficence: The research universities have their own review process should generate the maximum of that is carried out by a dedicated board, social benefits possible, while often called an Institutional Review Board minimizing harm to human (IRB) or an ethics committee. Each subjects. university will have a web page to tell you 3. Justice: Researchers must ensure how these boards work. These bodies are that research is carried out fairly required to adhere to national legislation and in a non-exploitative manner. and protocols. In companies and NGOs, ethics oversight tends to be specific to the In practice, these principles can be met organization. through meeting a number of general requirements: ensuring participants' safety Research with human subjects (physical and psychological), gaining informed consent from each participant, Probably the most influential ethical code protecting subjects' privacy and in human subjects ethics is the Belmont confidentiality, undertaking risk Report,2 a United States Government management, and gaining approval from guide that has become a reference for an authorized oversight body (such as an institutional review boards (IRBs) in the Institutional Review Board, IRB). United States. Codes for the protection of human subjects became imperative after Rules and procedures of local IRBs or biomedical experiments on human other types of review boards will vary. subjects during the Second World War, Depending on the project, data to be and studies such as the Tuskegee syphilis collected, and level of risk to research experiment conducted without knowledge participants, different types of review3 or consent of research subjects. Ethical apply: exempt, expedited, or full review. guidelines for protection have since been The research timeline should therefore extended to include all forms of medical, include sufficient time for the review behavioural and social scientific research. process. Once a study has received IRB approval, there is an ongoing process of 7 Consumer Finance Research Toolkit continuing review until the project is purpose and procedures for the formally closed. study. Depending on the project, research participants will give As a rule of thumb, the following kinds of signed or verbal consent. Forms materials are submitted to review boards should include the name and for clearance prior to starting the contact information of the principle research: investigator (s) and how they can be contacted when/if research  Timeline for research. This should subjects have questions or include time for review and concerns. Consent forms should research clearance. always make clear that  Study Purpose. What is the participation in the study is purpose of the study? A voluntary; research participants justification of who will be included can choose to leave a study at any time.  in the study and why? What are research subjects being asked to Plan for data protection, storage, do and why? What risks and and use. Who is responsible for benefits will accrue to study data integrity and who will have participants as well as to other access? How will identity and kinds of stakeholders? privacy be protected? What  Research instruments. These may protection and level of anonymity include questionnaires, interview can be guaranteed? What kinds of guides, survey instruments, and data will be kept, for how long, and other forms of data collection. how will it be safely archived or Questionnaires and survey destroyed during and after the study ends?  instruments should be planned carefully to think through the Dissemination of research. With implications of what questions are whom will the research be shared? asked and of whom, among other How and at what point will study considerations. Instruments can results be shared with participants, and should be adjusted in the field, research partners, and key where possible, when unexpected stakeholders? Will they be given issues must be addressed or new the opportunity to review and developments arise. comment on study findings via  Consent forms. Consent forms follow-up visits, focus group should reflect research tools and discussions, or sharing study study purpose. They should be results prior to publication? written in clear language that participants are able to The best way to begin to learn about understand. These are typically ethical practices for research with human short forms that outline the subjects is to review some of the literature 8 Consumer Finance Research Toolkit on the subject. A great deal of material is with access to practical ethnographic available online, such as the guidelines expertise from around the world.” EPIC published on Forum: Qualitative Social offers a variety of resources on ethics as Research4 or the University of Sydney well as research design. guidelines.5 These sources can give you a general idea of the issues you may In general, it is important to read up on encounter. related studies when developing a project to learn what has worked (or not) and Even without an institutional affiliation, a about the kinds of unexpected number of options for training are methodological and ethical issues available on research with human researchers have encountered on the subjects and the responsible conduct of ground. If in doubt about how best to research. For example, Collaborative proceed on more sensitive issues Institutional Training Initiative (CITI)6 and impacting research with human subjects, the NIH Human Subjects Protection consult a specialist. Training7 offer basic online training or tutorials. A number of independent8 non- Research without human academic IRBs offer training, assistance and services for biomedical, educational, subjects consulting and other types of research Research that does not involve human with human subjects. Professional subjects directly can be controversial marketing associations9 have also when it comes to ethics. The rise of “big developed ethical codes of conduct for data” has generated a great deal of conducting marketing research with excitement about its potential benefits to human subjects, from interview and create new knowledge that is beneficial to survey research to the controversial humanity. However, while “big data” “mystery shopping.”10 presents exciting opportunities to learn about human society and improve When it comes time to design your services, it poses heightened challenges project, there are hundreds of books that to individuals’ privacy, protection, and you may find useful. Some of these are consent. general, such as Mark Israel’s book Research Ethics and Integrity for Social Privacy issues are especially complex, Scientists.11 Others are specialist, such as since a great deal of data is collected Readings in Virtual Research Ethics: without people’s informed consent. For Issues and Controversies, a collection example, central banks and other financial edited by Elizabeth A. Buchanan. Another institutions commonly use bank data to great resource is EPIC12 (Advancing the analyse socioeconomic trends, such as Value of Ethnography in Industry), whose changes in the supply of credit13 or mission is “providing practitioners, consumers’ use of different payment businesses, and partner organizations mechanisms.14 Credit card and mobile 9 Consumer Finance Research Toolkit phone usage data collected by providers codes and guidelines for digital research. or financial institutions reach even deeper For example, the Data Science into the intimate realm of people’s lives. Association has produced the Data Science Code of Professional Conduct to Moreover, data sets are often shared, help researchers think through ethical especially with the current trend towards issues. open access data. Sharing data promises significant social benefits, such as Various specialized books now exist on lowering the costs of data collection and the subject, such as The Ethics of Big increasing the size and quality of data Data: Ethical Reasoning in Socio- sets. However, sharing data sets makes it Technical Informatics. However, the impossible for consumers to give consent issues are complex and are likely to for all uses of their data. Some worry that become more so in the future. While it is these data will be shared for proprietary crucial that researchers planning digital gain. There are even discussions of studies are up-to-date on current ethical paying individuals for their data which practices and standards, it is equally poses thorny ethical issues for consumer important that regulators act to protect protection, and which directly impact consumer rights. consumer finance research. It is therefore all the more important for researchers to think through issues of user privacy and Dissemination and Knowledge- design their research accordingly. Sharing “Informed consent” doesn't just mean that An important ethical dimension of respect an individual has ticked a box or signed a for research participants is building in time consent form; they should be aware of and opportunities to share and what the research involves and how it will disseminate preliminary findings and be used. In other words, they need forthcoming publications. enough information to make an informed decision. But data analysis usually Researchers can include a phase at the involves no such processes. The rationale end of a project where they return to the is that the benefits gained from creating field to hold knowledge-sharing new knowledge through data outweigh the workshops or focus-group discussions risk of harming participants. This only with key informants and clients to share holds, however, if individuals' privacy is study results. Researchers present their protected through proper anonymization main arguments and conclusions and and protection of data. elicit feedback from informants15 [See Appendix, blog post 3] who have the Researchers and professional opportunity to weigh in on the associations are responding to these appropriateness of study statements and kinds of issues by developing ethical whether and how they should be modified or expanded. 10 Consumer Finance Research Toolkit This is not only a means of following up or how financial practices occur in everyday verifying study conclusions with research life. As well as asking participants to sign participants. It is also a way of giving consent forms, they sought permission to back, if not materially, then at least in film and photograph the interviews. terms of acknowledging participants as Participants had the option to choose not co-contributors to advancing insights and to be filmed at all, to just have their hands impact in consumer finance research. filmed as they interacted with the objects they were discussing, or to also have their faces recorded. In other words, Taylor and Ethical issues in consumer Horst provided a way for interviewees to finance choose different levels of consent. The case studies in this toolkit illustrate a David Stoll's ethnographic research on wide range ethical issues that arise in debt in Guatemala highlights an ethical consumer finance research. Here we give issue with respect to the benefits of the a few examples to provide you with a study for participants [see Case Study 2 in sense of how they can play out in Ethnography]. Since most of his practice. respondents were both poor and heavily in debt, he felt that it would be unfair if his In the Ethnography section of this toolkit, research did not help them directly. He our case study of Woldmarian F. Mesfin's could not give his participants money, research in an Ethiopian market place since it would both compromise the study discusses the important issue of consent and rapidly deplete his own resources. [see Case Study 1] Mesfin's research Instead, he tried to lend assistance by shows how consent is about far more than sharing his knowledge about community reading a project description and signing a debt with organizations and policy-making consent form. He describes how his bodies working on issues relevant to the research participants were sometimes community. ashamed of how little money they had, and were therefore uncomfortable talking Data protection is always an important with him about their financial practices. issue. As Jofish Kaye's study shows, data Mesfin took the time to get to know people protection can't always wait until after data so that they would not feel pressured to collection is finished [see Case Study 1 in give him information, and instead could Object-Centred Interviews]. Kaye began agree to participate when they felt the process of data protection while comfortable. interviews were taking place by excluding personal and unnecessary information. Taylor and Horst had a similar experience When interviewees showed them their in Haiti [see Case Study 2 in Object- financial spreadsheets on their computers, Centred Interviews]. They explain how the researchers placed sticky notes over they interviewed Haitians about the information such as names, account objects they carrying with them to learn 11 Consumer Finance Research Toolkit numbers, and bank balances. This saved researchers designed the study so that them from a time-consuming and risky they would identify known service process of anonymizing data later. providers, but not individuals. By limiting their analysis to known service providers, Data protection can be more complicated and focusing on flows of Bitcoins rather for some studies than for others. When than individuals’ transactions, the projects involve a large number of researchers largely avoided issues of researchers and participants, it is more individual privacy and consent. difficult to ensure that data is adequately anonymized and protected. For example, Experimental methods, especially financial diary studies carried out by randomized controlled trials (RCTs), raise numerous researchers are likely to be important questions about how to balance more risky than a study carried out by just the requirements for study design with the one person, because there are more ways benefits to research participants. When that data can be accidentally leaked. control groups do not receive a particular service or product that is given to Digital research can produce a variety of treatment groups, how can researchers challenges around privacy. Tom address the principles of justice or fair Boellstorff reports that one big issue he treatment for all involved in a study? As has encountered in his research on Deepti Kc and Mudita Tiwari’s study payments in Indonesia is that some demonstrates [see Case Study 3 in researchers think that because something Experiments], a number of local women in is online it is not “real,” and so you don’t the village were given training to assist in have to protect people’s identities [see the delivery of the intervention – financial Case Study 2 in Digital Research]. For literacy modules. By incorporating example, data gained from a public forum capacity building into their research may be technically traceable, but if it is design, the researchers facilitated the reproduced without permission, ability of these women to share their researchers have a responsibility to knowledge beyond the treatment groups generalize the data into findings that are once the study had been completed. This not traceable. is one way in which experimental research can contribute to the community broadly in Privacy protection was a major ethical matters such as strengthening financial issue in a study of Bitcoin carried out by literacy education in ways that can endure Sarah Meiklejohn and colleagues [see well beyond the life of the study. Case Study 1 in Digital Research]. The researchers wanted to identify the extent As these examples demonstrate, every to which Bitcoin lives up to its promise of research project will generate a different pseudo-anonymity16 and to investigate set of ethical issues. Methods, topics, how people’s use of Bitcoin has changed participants, researchers, and contexts all over time. To protect users’ privacy, the vary, so for every project we design we 12 Consumer Finance Research Toolkit need to think through the risks and issues we may encounter. Beginning with basic principles of consent, privacy, benefit, and equity is a good way to start. From there you can build up an ethical approach that is relevant to your unique research project. And, if in doubt, ask an ethics expert. 13 Consumer Finance Research Toolkit Figure 2 A spreadsheet with post-it notes for anonymization. Image courtesy of Jofish Kaye 14 Consumer Finance Research Toolkit Figure 3 Summary pages for two participants, designed by Mary McQuistion. Image courtesy of Jofish Kaye 15 Consumer Finance Research Toolkit Money Methods Figure 4 IMTFI Methods Workshop (2014) Figure 5 IMTFI Methods Workshop 16 Consumer Finance Research Toolkit What is it? Ethnography  Qualitative  Combines multiple methods Ethnography has a long history of use in  Face-to-face or remote data consumer finance research. Since the late collection 19th century, anthropologists have used ethnography to discover how people Ethnography is a method of studying create, share, and manage money. people in the places where they live or Researchers have investigated topics as where the action is taking place. It is a broad as the creation of shell money as versatile method that can be coupled with currency in Melanesia, community-based both qualitative and quantitative savings associations in Africa and the techniques, and can be used in virtually Caribbean, and use of tally sticks to any setting. record debt in many parts of the world. Ethnography involves recording data As more consumers around the world using a variety of tools, including adopt formal financial products and recording devices, writing field notes, services, ethnography is being used to analysing documents, or deploying investigate an increasingly wide range of questionnaires and surveys. One behaviours and products, including important method is participant household financial management, payday observation, which involves learning about loans, mortgages, microfinance, mobile research participants’ experiences by money, Islamic banking, and remittances. doing activities with them. Ethnography as a method is also Technically, ethnography is more of an changing. New data sources and approach than a method because analytical software are being applied to ethnographers often use multiple ways to both classic anthropological problems collect data.19 However, all ethnographic (such as kinship) and new topics (such as studies have two features in common: mobile money). Technological developments have resulted in an entirely 1. Research is carried out within real- new field, known as computational life settings. This could be a anthropology.17 Ethnography is also a village, neighbourhood, workplace, core method in human-centred design,18 or online, such as an Internet an approach to research that incorporates forum. interactions with people into the design of 2. Researchers use observational a product or service. techniques alongside other methods to record actual behaviours as well as reported ones. 17 Consumer Finance Research Toolkit investigate how Islamic Ethnographers’ emphasis on living with microfinance can alleviate people and observing their behaviour in poverty25 real-life contexts distinguishes it from  Many sociologists and other qualitative methods, which tend to anthropologists use ethnography rely on either self-reporting or analysis of to study people’s use of third-party data [for an exception, see microfinance26 Experiments].  Several chapters using ethnography appear in the World Bank publication In consumer finance, it is particularly “Migration and Remittances during the Global Financial Crisis and Beyond”27  Tricia Wang shows the importance useful for understanding how consumers adapt to changing products or economic conditions, and for how people’s financial of ethnographic “thick data” as a management is interconnected with their complement to “big data” social relations. methods28 Examples of use Strengths  Professor Lisa Servon uses participant observation to Research occurs in context investigate why and how The places in which people live and customers use payday loans in a work—home, office, school, social media, check-cashing store in New York etc.—significantly affect their actions and City20 choices. While financial products are often  Dr. Zsuzsanna Vargha conducts targeted at individuals and households, ethnographic fieldwork at a their family, friends, peers, and finance Hungarian home savings and loan professionals in fact influence people’s bank to understand how banks choices. Advertising, availability of explain financial plans to clients21 technology, and financial regulation are  Dr. Erin Taylor examines how also among the contextual factors that squatter settlement residents build impact financial behaviours. homes without loans in the Ethnographers carry out research within Dominican Republic22 these contexts, collecting data on how  The IMTFI’s researchers people interact with their social and investigate mobile money use physical environments. around the globe23  Professor Bill Maurer documents ongoing efforts by Islamic bankers to remake money and finance24  Anthropologist Bridget Kustin uses participant observation to 18 Consumer Finance Research Toolkit Participation as a way of learning they feel is unimportant or because there is a cultural taboo A good way to learn about people’s on discussing it. Sometimes experiences is to join them in their people struggle to find the words activities. “Participant observation” is used to explain their physical by ethnographers to gain an insider’s view behaviours (such as why they into how people act, think, and feel. For keep documents in certain places) example, standing in a bank line with a [see Case Study 2]. customer can be a powerful method to 2. Research participants’ behaviour understand their frustration at having to may contradict what they are wait and what can be done to make saying, but this does not mean that transactions more efficient. Using a the verbal data is simply incorrect product or program, such as online and should be discarded. In all banking, helps show up design flaws that human societies there is a research participants might not think to difference between what people mention when asked for feedback feel they should do (ideal norms) and what they actually do (actual Participant observation is particularly norms). Differences in observed useful when trying to explain what and verbal data may not be due to motivates people to make choices that error or omission, but rather to this may not seem rational (such as using an variation. In fact, ethnographers expensive payday lender) or to pay special attention to such gaps understand the technical issues they because these help to identify struggle with (such as problems using cultural patterns in how people online banking). think, talk, and behave. Cross-check data sources Track behaviour over time Ethnographers collect observational data Because ethnographers generally spend as well as verbal and written data, and considerable time in the field (ranging they can cross-check these sources to from days to years), they are able to identify instances in which what people record data on repeated behaviour, such say they do differs from what people as seasonal harvests, rituals, and work actually do. This is useful for (at least) two patterns. This can be particularly useful in reasons: consumer finance research because people’s interactions with money and 1. It helps identify errors and financial products are often cyclical: omissions in data. These may have been introduced into data,  Most people get paid at specific either because people do not tell times and pay bills at regular the truth, do not accurately recall intervals information, omit information that 19 Consumer Finance Research Toolkit  Spending patterns cycle as “evidence,” they not readily throughout the year due to generalizable. The exception to this is changing seasons or religious comparative analysis, in which ranges of festivals (e.g., many people spend cases (often produced by different more money at Christmas time) researchers and at different times) are  Yearly bonuses or seasonal compared to spot patterns and build social harvests may mean that people theory. earn most of their income is within a short time frame each year A lack of representativeness is unlikely to present a problem for researchers who Tracking these cycles can yield valuable are seeking information on changing information to design products and social conditions or sentiments. However, services, or to simply understand the sectors of consumer finance that depend mechanisms of how people manage upon representative samples to make finances over time. important decisions, such as policy implementation or development programs, Facilitate access to data may need to supplement ethnographic data with quantitative data. People often consider their financial information to be private and are reluctant to share it. Ethnography can help to Usually not replicable overcome this barrier because Ethnographic studies are rarely replicable. researchers often spend enough time “in This is because: the field” to build relationships with their research participants and gain their trust.  They involve the study of complex Participants have more time than usual to social phenomena that are prone learn about the researcher and their work, to change rapidly and to ask questions about how the data  Gaining access to the same sets will be used. Access to data therefore of people that participated in an often comes about because researcher original research project can be and participants have co-negotiated difficult: they may have moved or important personal and ethical issues. no longer have time to participate in an in-depth study Limitations Studies that require replicability may be better suited to using structured interviews Not always representative of a or surveys. population at large Ethnographers do not usually collect data Resource-intensive based on representative samples. This Some ethnography can be carried out in a means that, while their findings can count matter of hours, but in most cases, the 20 Consumer Finance Research Toolkit best results come from spending  Errors made by sellers in their considerable time in the field. Whether calculations ethnography is a viable method will  How sellers divided up cash into depend on what information is needed different piles to assist their and the extent of resources available to accounting conduct research. For example, if a bank wants to investigate customer By focusing on these areas, Mesfin aimed experiences of waiting in line, then a few to identify the common practices that hours or days may well be sufficient to merchants (sellers) and buyers produce the insights needed. However, (customers) perform. observing customers’ financial management from month to month would Method require a much longer study, with many more hours spent in the presence of the Mesfin used observation techniques to customer. collect data on how people counted, stored, and transacted with cash. He recorded these observations using a Case Study 1: Transactions in combination of field notes and (where appropriate and feasible) photography Ethiopia: Direct observations of and video. use of cash in sales Mesfin explained to us the benefits of Direct observations can provide observation: information about people’s behaviour that is not accessible through any other  It helped him to quickly understand method. what respondents say about their practices IMTFI researcher Woldmariam Fikre  It enabled him and his Mesfin29 used participant observation to respondents to have common collect data on open-air cash transactions understandings, since they could in fourteen different marketplaces in point to concrete details, such the Ethiopia. Markets are busy places, and so design features of money or bags buyers and sellers depend upon shortcuts  Observation tended to raise new in order to make transactions quickly. questions that may not have come Moreover, most of Mesfin’s research to mind through other methods participants were functionally illiterate and could not read the numbers or text on In order to confirm his observations, banknotes. He paid particular attention to: Mesfin also conducted interviews with his  Shortcuts buyers and sellers use research participants. He cross-checked the data generated by each method to quickly tell currency values (known as “triangulation”) to ensure he apart had not misunderstood what was 21 Consumer Finance Research Toolkit occurring. Mesfin’s research is an helped them to keep track of product example of a multi method study. movement: Findings “Keeping these sales separate enables them to easily know the Mesfin discovered that buyers and sellers daily sales from each item, the net identify currency notes according to their profit of each item and the most colour rather than by the numbers printed profitable items in their onto them, and calculate their value portfolios.“30 accordingly. This is, in fact, a common practice in many countries, especially in While research participants could those that have brightly-coloured notes theoretically explain these behaviours in such as Ethiopia. However, these kinds of interviews, observation overcomes many shortcuts did not prevent errors of inaccuracies introduced by human recall calculation. Due to time constraints and or omission. crowding, sellers were not always sure who had paid them or whether change was due. Applications Insights such as these could be used to Observations permitted Mesfin to observe inform the development of new money- the different methods of accounting sellers management products, such as a mobile used. Some sellers preferred to use a money service that includes built-in single bag for their money, whereas other features suitable for their local contexts or sellers had a separate bag for each type particular users. For example, a mobile of item they sold. For example, money money service aimed at sellers could from the sale of coffee would go into one provide multiple wallets connected to the bag, and money from the sale of salt one account. This would allow users to would go into a second bag. When they continue to divide up cash into multiple provided change for an item, they would piles. [See Appendix, blog post 2] take it from the bag corresponding to that item. Ethical issues Research participants admitted that this As with many studies of money, Mesfin system sometimes made transactions told us that he encountered ethical issues overly complicated, but there were a few regarding data access: reasons why it was sometimes preferable. One was that when all of a seller’s money “Sometimes, as related to money, was stored in one single bag it was at respondents do not feel greater risk of theft. Another was that, in comfortable opening up their the absence of written records or cash money bags and let you see. If registers, storing money in separate bags their bag contains little money they feel shame.”31 22 Consumer Finance Research Toolkit However, since Mesfin spent a significant migration to the United States. He amount of time in each research site, he investigates the motivations behind reports that he developed trust and even people’s decisions to take on debt, inter- friendship with the sellers in the market. group lending practices, and the role of He also reported that if others saw him microfinance agencies in stimulating the looking at the bags of others, they felt community’s debt bubble. confident to open up and show him their bags. Method Stoll carried out long-term ethnographic fieldwork, beginning in 1987 and generally Case Study 2: Debt in returning to the field every year. Some of these trips were made to carry out a Guatemala: Using long-term concrete research agenda; others were ethnography to collect data simply to visit and keep in touch. Stoll reports that long-term fieldwork had two Some topics in consumer finance are advantages: difficult to research because data is private, widely dispersed, or not recorded  It permitted him to stay connected using any accessible media. Ethnographic to residents, building up trust and research can assist in overcoming these recognition barriers because researchers piece  It allowed him to observe how the together data from a wide variety of community and its institutions sources. changed over time, including generational change Moreover, ethnographers often revisit the same place for years or decades to carry Stoll’s research turned to the problem of out different research projects, meaning indebtedness in 2007 when he discovered that they get to know a place and its that residents were becoming heavily people in detail. This makes it easier to indebted to microfinance institutions, identify how financial behaviours fit into banks, moneylenders, and each other. larger social, cultural, and economic Finding out why people were taking on so patterns. much debt involved: A good example of this is David Stoll’s32  Interviewing residents about their research in Guatemala. Stoll has been borrowing and lending carrying out research on political and  Observing interactions between economic issues in the highland Maya family members, neighbours, and market town of Nebaj since 1987. In his institutions book, El Norte Or Bust!, Stoll describes  Working with a debt committee to how he watched Nebajenses become tabulate the experiences of its over-indebted as they speculated on members 23 Consumer Finance Research Toolkit Stoll’s initial research involved trying to Instead of shrinking, the bubble grew discover why people were becoming so bigger for another year. indebted. As the picture became clearer, his research questions became more Stoll found various factors that numerous and more specific, focusing exacerbated risk-taking among borrowers: especially on understanding why people would take on the risks associated with 1. People were reluctant to believe borrowing and lending. that migration was no longer a successful strategy. Findings 2. Even if they did recognize the risk, they considered it to be one worth Stoll’s initial finding was that a drop in taking. remittances from the U.S. was causing 3. Failed migrations meant that more people to default on their loans. As families would have to borrow defaults rose, the town’s lending even more money so that the institutions stopped making new loans, aspiring migrant could make and defaults rose even higher. The price another attempt at crossing the of local real estate, which had inflated border. enormously, suddenly collapsed. 4. The only way of paying back loans was to take out more loans. Many households were taking on debt to 5. Some people took loans to lend to fund illegal migration to the United States. others in the hope of making a From their onset in the late 1990s until profit. 2006, the majority of these ventures had been successful. The families who stayed Among lenders, Stoll found the following: behind received money from their relatives who had migrated. Migrants were 1. People would lend to each other to able to pay back the loans financing their fulfil social obligations or with the trip and return with enough savings to buy hope of making a profit. a used motor vehicle, land, livestock, or to 2. Loan sharks charged high rates to improve their homes. cover the cost of the risk. 3. Formal lenders, such as From 2006 onward, migrants found it microfinance agencies, were not increasingly difficult to find enough aware that their loans were being employment in the U.S. to pay for their invested in migration or third-party trip, let alone send the remittances their lending activities. families now expected. Stoll reports that while he gained initial Strangely, these increased failures did not contacts through his local networks, immediately deter other residents from strangers were also willing to talk candidly taking on debt to migrate themselves. with him about their debts, admitting to 24 Consumer Finance Research Toolkit their own faulty decisions as well as to assessment of risk, not just on instances in which they had been tricked how to choose and use products or let down by others. However, no one  Microfinance institutions could person or institution had a good sense of adjust the processes by which they the entire picture. By combining made lending decisions interviews, observations, and  Retail banks and other formal participation, Stoll pieced together the financial institutions could develop story. products that take into account residents collective financial Applications behaviours into account as well as their individual behaviours Stoll’s work has numerous applications in understanding and tackling over- indebtedness. Stoll reports that, while he Ethical issues was in the field, he was able to assist both The fact that Stoll conducted long-term individuals and institutions by: fieldwork meant that he was reasonably aware of the kinds of ethical issues that  Helping one of the debt might occur. However, this did not committees compile a list of guarantee that they could be avoided. problem cases Stoll told us:  Paying for a local radio spot warning against the resurgence of “The biggest ethical issue is my a particular swindle inability to provide financial  Paying some medical expenses for assistance to the people who told a crippled returnee me their stories. Even the smallest  Providing advice for what he of the debts that were asking for warned were unrealistic pitches to my resources were on the order of NGOs $1 to $2,000 dollars. Per my human subjects protocol, I always There are many ways in which Stoll’s introduced myself as an findings could be applied. For example: anthropologist whose job was to collect stories and publish books,  Knowledge of how debt is not an NGO who could provide accumulated can assist regulators help. But because of the aura in deciding which kinds of surrounding gringos, especially in institutions to license or in a town upon which one NGO after developing methods to combat another has descended, merely loan sharks repeating this disclaimer did not  NGOs could use his insights to vanish the hope that I might be develop financial literacy programs able to help in some way.”33 that focus particularly on the 25 Consumer Finance Research Toolkit He continued, 4. If research results in indirect assistance, participants may not “What I did do was help debtors associate it with the research and communicate with local pro-bono may feel exploited. lawyers; pay for physical therapy for a crippled returnee; and pay for It should also be kept in mind that ethical a local radio spot warning against issues change throughout the course of the resurgence of a particular research, especially in long-term studies. swindle. I have also had my book What is problematic at the beginning of translated into Spanish and data collection may cease to be an issue circulated it via PDF to down the track, and vice versa. Care Nebajenses who have enough therefore needs to be taken to regularly schooling to read it. I hope that it assess the current ethical issues. will encourage town authorities to discourage predatory money lending, crackdown on swindlers, and support victims through the More about the method legal system.”34 Research Methods in Anthropology: Qualitative and Quantitative Approaches Stoll’s assessment highlights various by H. Russell Bernard (2011, AltaMira) ethical problems that researchers studying indebtedness might encounter: A Handbook of Practicing Anthropology edited by Riall Nolan (2013, Wiley- 1. Transparency about the aims and Blackwell) outcomes of research is an ethical obligation. Practical Ethnography: A Guide to Doing 2. However, there may be a gap Ethnography in the Private Sector by Sam between what researchers claim Ladner (2014, Left Coast Press) they do, and what participants think they do. This raises the Handbook of Narrative Inquiry: Mapping a possibility of unintentional coercion Methodology edited by D. Jean Clandinin in which participants share (2006, Sage) information that they would not have otherwise given away. Doing Anthropology in Consumer 3. Research on vulnerable Research by Patricia L. Sunderland and populations that does not result in Rita Mary Taylor Denny (2007, Left Coast tangible outcomes has Press) questionable ethical value in the eyes of the population being studied. 26 Consumer Finance Research Toolkit Figure 5 Church Money Donations. Image courtesy of Woldmariam F. Mesfin Figure 6 Ethiopian Currency. Image Courtesy of Woldmariam F. Mesfin 27 Consumer Finance Research Toolkit Figure 7 Sample document to record wedding gifts. Image courtesy of Woldmariam F. Mesfin 28 Consumer Finance Research Toolkit Figure 8 Market Scenes, W.F. Mesfin 29 Consumer Finance Research Toolkit analysed, giving them a quantitative aspect. For a comprehensive guide to Verbal Interviews structured interviews, see the GAO Guide Verbal interviews are one of the most to Structured Interviews.35 widely used methods in qualitative research due to their versatility and ease In unstructured / semi-structured of execution. They are a standard part of interviews, interviewers adapt their research in both commercial and non- questions as the interview progresses. commercial research in consumer finance, This method is generally qualitative such as in the payments sector and by because there is variation between the regulators. questions asked of each interviewee. A variety of qualitative researchers are There are also many sub-types of trained to conduct interviews, including interviews, including life histories, video sociologists, anthropologists, behavioural interviews, place-based interviews, and researchers, and user experience object-centred interviews. researchers. However, researchers who are more quantitatively focused, such as Examples of use economists, may also conduct interviews.  FAIR money undertook a study of financial management among What is it? struggling individuals in Silicon  Qualitative / quantitative Valley36  Multiple techniques to choose from  In an older, but still relevant study,  Face-to-face or remote data the Joseph Rowntree Foundation collection assesses the industry in their report on moneylenders and their customers37  A Bill & Melinda Gates Foundation Interviews are a standard method in qualitative research. The main types of interviews are structured interviews and / Gallup study of sub-Saharan unstructured / semi-structured interviews. Africans examined payments and money transfer behaviour38 In structured interviews, interviewers write  IMTFI researcher Janet M. questions (the “interview schedule”) Arnardo’s researched how ahead of time and ask exactly the same indigenous communities in the questions of each interviewee. The Philippines store money informally39  Jonathan Donner and Camilo purpose of this consistency is to avoid biasing how the interviewees answer the questions. It means that data is quite Tellez examine the adoption, uniform, and interviewees’ responses can impact, and use of mobile financial be more readily “coded” (categorized) and services in their study of mobile 30 Consumer Finance Research Toolkit banking and economic location can assist in helping interviewees development40 feel comfortable.  A mixed methods approach investigates how microfinance Methodological versatility affects the development of rural women in India41 Interviews are one of the most versatile  Bruce Cahan of the Filene Institute research methods because different kinds interviewed representatives from of interviews can be used to achieve credit unions to better understand specific goals. Unstructured interviews the potential of their roles and give the interviewee the greatest control services42 over the content. An interview may begin with a single planned question, with the interviewer generating new questions in Strengths response to the interviewee’s statements. Life histories often take this format, but the approach can be used for any First-person perspective exploratory research. For example, asking Topics related to money are always a person to simply talk about their concerned to some extent with numbers, finances can lead to topics emerging that such as bank balances, prices, and the interviewer may never have thought of interest rates. This means that asking about. Moreover, it allows the researchers often need quantitative data, interviewee to discuss what they feel is and fortunately also that it is often important. available. However, while statistics are often valuable in telling us what people In contrast, semi-structured interviews do, they might not tell us why people allow deviation from the interview make certain decisions. People’s schedule to allow for the discussion of descriptions of their own behaviour can interesting points that emerge while tell us why they behave in certain ways interviews are in process. Structured and also what they think motivates their interviews, in which an interviewer sticks behaviour [see Ethnography]. to the questions on the schedule, permit interviews to be directly compared, which Flexibility of location is not the case for looser interviews. Interviews can be carried out with virtually How structured an interview is (or isn’t) is anyone and in many different locations, just part of what makes them work. Props such as at home, work, school, cafés, and context help data collection by over the phone, or over the Internet. In prompting interviewees to speak on consumer finance research, where the certain topics [see Object-Centered data being collected is often considered Interviews]. Another means of inquiry is personal, choosing an appropriate cultural domain analysis,43 which involves asking interviewees to classify items in a 31 Consumer Finance Research Toolkit list, permits researchers to understand commercial work, recruiters are often how interviewees order information used to assist with the organization. Professional interviewers are needed to Quantifiable collect the data, which will then need to be transcribed, coded, and analysed. Structured interviews that apply the same questions to each person in a sample group can work like surveys in providing Design complexity quantitative data. Developing a coding Depending on the research requirements, schema for open-ended questions and writing effective interview questions can analyzing the data using quantitative require high levels of skill. For software can achieve this. While unstructured or semi-structured interviews are generally more time- interviews, researchers need knowledge intensive to complete than surveys, of how to adapt questions while the obtaining data from a representative interview is taking place. Structured sample is possible given enough interviews are intended to produce respondents or a smaller group size. rigorous results, and questions must be carefully designed. There are many textbooks that discuss the theory and Limitations practice of writing interview questions. Self-reporting bias Interviews generally depend upon the Case “tudy : Learning good interviewee to give an accurate account of their own behaviours. However, while and ad finan ial ehaviours: people are experts in their own opinions, Semi-structured interviews studies show that we are unreliable at remembering facts. Interviews do not with Hispanic American college normally include substantial time for students observations that could be used to cross- check verbal data. It can therefore be Many consumer finance researchers have often advantageous to use interviews in observed that people often do not appear conjunction with other methods. to manage their money in ways that best suit their interests. This is partly because Resource-intensive people may not have the financial skills they need to manage their money well, Interviews generally require less research but it has also been suggested that people time than ethnography, but they still learn “good” and “bad” behaviours from present significant costs in terms of those around them. But from whom do collecting and processing data. they learn, and how can positive financial Organizing interviews and carrying them behaviours be encouraged? out is a time-consuming process. In 32 Consumer Finance Research Toolkit Marketing professor Kittichai Studying how behaviours are formed, as Watchravesringkan wanted to find out well as individuals’ motivations, deepens how Hispanic American college students our understanding of why one acquired financial skills.44 Quantitative and demographic group might differ from qualitative data from various sources others. indicates that Hispanic Americans are one of the groups that are most at risk of Method financial difficulties. However, the data did not demonstrate why this was the case. To investigate consumer socialization among Hispanic Americans, the Hypotheses have included that Hispanic researcher interviewed 11 college Americans have lower educational students of 20-25 years of age. He attainment and are reluctant to engage in focused on college students because long-term financial planning. One survey Hispanic Americans demonstrate low suggested that this demographic group levels of educational attainment due to are suspicious of advertising and are financial constraints. All interviewees were reluctant to adopt new products and either first- or second-generation Hispanic services that could assist them to manage Americans. their finances. Another study suggested that Hispanics have more present- Students were selected using a oriented attitudes and are less likely to combination of a convenience sample and engage in delayed gratification. snowballing techniques. Care was taken to ensure that interviewees represented a Out of all these possible causes, which range of majors. Monetary incentives are the most important? Underlying all of were given to interviewees to increase the them is a question of how people learn rate of participation in the study. financial behaviours in the first place: who their “socialization agents” were (family, Interviews lasted 60-70 minutes and were friends, advertising, etc.). As audiotaped. Interviewees were Watchravesringkan explains, encouraged to talk about matters that they felt were important. The interviews began “Consumer socialization refers to with “grand tour” questions to collect the process by which young general information before turning to consumers develop consumer- financial management specifically. related skills, knowledge, values, Questions were kept quite general, and attitudes throughout their focusing on how interviewees learned to different life stages via the manage their finances, their financial influence of socialization agents, values, and their current practices. They such as family and peers.”45 included:  How do you manage your finances? 33 Consumer Finance Research Toolkit  Are you currently satisfied with the to learn “bad” financial behaviours from way you handle your finances? outside their families (peers and media).  How do you learn to develop financial skills? Applications  What kind of values may impact your financial management do you While this study was small, its findings are believe that you should or should potentially useful for further research or not have? program development. The interviewer then coded and compared In particular, the suggestion that family the interviews to spot emergent themes. provides a positive overall influence is They showed a selection of their interesting because it contradicts common conclusions to the interviewees to attain assumptions. Many social studies tend to their responses. Overall they found that assume that, if a behaviour is specific to students mostly agreed with the study’s an ethnic group, then it must be culturally conclusions. learned through parent-child transfer. Instead, the results of this study suggest Findings that inter-group socialization may not be a The researchers found that the students’ problem, but instead may contribute financial behaviour was strongly towards improving financial management, influenced by their family members. such as by developing programs that Specifically, interviewees reported that enhance the influence of families on they had learnt the importance of financial financial behaviour. As the researchers management and saving from their explain, fathers. Students also reported being influenced by watching other people get “The results may aid academic into debt. administrators, financial counselors, and consumer In some cases, students watched others educators in gaining a greater become over-indebted and copied their understanding of this particular behaviour; in other cases, students college segment and finding learned not to follow the footsteps of means to develop effective people who became over-indebted. To a outreach programs geared toward certain degree, students also learnt from this growing segment.”46 their peers, television, or their religious communities. Key to uncovering this insight was that interviews were relatively long and open Overall, the data suggested that whereas form. This allowed the students the time the interviewees learn “good” financial and space to cover issues that they felt behaviours from their families, they tend were relevant to them. In a more structured approach or a shorter time 34 Consumer Finance Research Toolkit frame, these insights may not have been ethnic group. This is known as implicit able to emerge. bias or ingroup bias.49 Poorly designed research can sometimes inadvertently Ethical issues reinforce stereotypes, such as when questions are poorly thought through. This This study raises various ethical issues study, by keeping questions open, that merit consideration. One of these is maximized the ability of students to share snowballing. a method of recruiting their own perspective and limit bias in the participants to a study by asking already- research design. enrolled participants to suggest people they know for recruitment. It is especially useful when working with populations that are difficult to access. However, there are Case Study 2: Interviews for some reservations about this method because there is a danger that social network analysis: Mobile participants may not feel that they can money in Kenya refuse to assist the researcher. Moreover, it can result in breaches of privacy. For An interesting application of interview more information on the ethics of techniques is in the analysis of social snowballing, see the University of Sydney networks. Interviews can be used to guidelines.47 collect data on networks, which can be analysed either qualitatively or Monetary payments are often used to quantitatively. Essentially, social network encourage participation in interviews. analysis is a self-contained mixed method. Some researchers argue that this is ethical because participants should be Quantitative interview data can be used to compensated for their time. Others raise map nodes and connections in social concerns that financial stress may drive networks. The resulting visualisations are people to participate in research that an excellent way to see clearly who is works against their interests. This is of connected to whom, and whether a social particular importance in medical research network is open (loose connections) or that can have negative physical outcomes closed (close ties among group for participants. It is less of an issue in members). research that focuses on opinions, values, and self-reported behaviour (such as this Qualitative interview data can be used to study). An article posted at The Hastings explain what drives social networks. For Centre describes the ethical issues example, interviewees can be asked to involved in paying research subjects.48 explain why particular connections exist, how they are maintained, and how they Cultural stereotypes can be reinforced or have changed over time. In other words, challenged by research that focuses on a whereas the quantitative data tells the particular social demographic, such as an “what,” qualitative data tells the “how.” 35 Consumer Finance Research Toolkit Social network analysis is particularly Method useful for studying patterns of circulation, Kusimba and her team undertook such as remittances, conditional cash research in rural Kenya in 2012. They transfers, gifts, and other forms of conducted participant observation, payments. It is handy for analysing mobile research interviews, and survey money transactions, in which users are questionnaires with more than 300 often individuals who send and receive Kenyans, 80% of whom were farmers. money for social purposes as much as for The team also conducted interviews with economic ones. It has the advantage of a smaller sample of Kenyans living in showing not only who is connected to Chicago. whom, but also in demonstrating how and why money moves across large The researchers carried out different kinds geographic areas. of interviews to elicit qualitative and quantitative data. “Intercept interviews” Sibel Kusimba and her colleagues were accomplished by walking through conducted a study of mobile money in the main marketplace. Kusimba Kenya, where at least 60% of adults are “intercepted” people, asked them if they unbanked.50 Mobile money was launched want to answer a few questions, and then in Kenya in 2007 and is widely recognized saw how much they wanted to talk about as the world’s most successful mobile the topic. While the resulting sample was money service. It offers person-to-person not representative, Kusimba notes that it transfers, a merchant payment service, is a good way to canvass general opinions and a basic means of saving money. In and identify people who are interested in November 2012, a related service called taking part in a more formal study. M-Shwari was launched that offers basic savings accounts and microloans. Because Kusimba had been working in the geographic area for 20 years, she Kusimba and her team were interested in already had contact with a number of discovering how rural Kenyans were families who were receiving remittances. networked through mobile money and the She therefore recruited people for reasons why they sent money. The interviews from both this familiar group researchers wanted to find out whether and her new contacts. common assumptions about mobile money—that it empowers individuals, In-depth interviews provided background stimulates entrepreneurship, and reflects and contextual information about people’s rural-urban migration patterns—reflect experiences, feelings, social lives, and Kenyans’ experiences of using mobile economic practices. Kusimba notes that money services. the advantage of her in-depth interviews was that the quality of information she received were high. A disadvantage was that the interviews often required a great 36 Consumer Finance Research Toolkit deal of time and several visits in order to Kusimba notes that, for her research, the achieve rapport. During interviews, the primary benefits of social network analysis researchers drew up kinship charts. They were: asked interviewees to tell them to whom  Visualisations help them to clearly they had sent money in the last year, and see and analyse the connections who had sent them money to draw the in a way that is difficult or networks. impossible with interviews  They could tell which networks For the quantitative part of the study, the were “open” or “closed” team interviewed between 3-10  They obtained a sense of who was individuals from 14 families. Each “brokering” the gaps in networks interviewee was asked to name all of the  As SNA is a statistical technique, relatives that they had sent money to, or the networks could be examined in received money from, in the previous terms of size, number of ties, and year. Most interviewees had sent money other parameters to 5-9 people. Where possible, the researchers then contacted the individuals She explained to us, that had been mentioned, and approached them for an interview as well. “Social network analysis is good They entered the resulting matrices into because it reveals different kinds R, statistical computing software that can of social relationships. It also be used to draw social networks provides quantitative assessments diagrams. in terms of size and number of ties. These can also become Kusimba and her team chose to ask apparent through ethnographic people to list the names of people they interviews but SNA makes it had transacted with rather than the clearer. We need both because amounts of money they had sent. There the ethnographic interviews give were two reasons for this. First, people context. It's also good to follow up tended to be inaccurate in recalling SNA and do another study in a few quantities of money. years (or other appropriate time frame) because then you can see Second, many people did not like to talk the social network change.”51 about money directly. This was especially the case with men who would organize Social network analysis has limitations as large ritual ceremonies that could cost up well as benefits. Like any model, it to 26,000 Kenyan shillings. Whereas simplifies reality, collapsing a lot of women would admit that they asked family information about family ties and and friends for financial assistance, men obligations. People send money for a preferred to say that they had collected variety of reasons, including deep kinship debts owed to them. ties, social obligation, or as a debt, but 37 Consumer Finance Research Toolkit these differences are not generally visible First, there is the assumption that in social network models. Whereas social primarily “individuals” use mobile money network modelling shows what people do, to conduct person-to-person transfers or in-depth interviews demonstrate why they for their own particular purposes, such as do it. saving money. In contrast, Kusimba argues that it is better to conceptualize Methodologically, networks drawn from mobile money as created by collectives interview data need to be treated as and groups. samples. People forget or intentionally omit their connections for various reasons. Remittances sent by mobile money are Like any other kind of ethnographic used to strengthen social ties, especially information, information needs to be among siblings and mothers, and as a verified wherever possible by talking to way of contributing to social rituals such the people who an interviewee says they as funerals, weddings, and coming of age have sent money to or received money ceremonies. Moreover, a person who from. This sometimes yields contradictory receives money will often re-circulate a information, but can also improve certainty portion of it to other family members and as to the accuracy of data if different friends. [see also appendix, blog post 5] interviewees’ accounts agree. Mobile money helps to equalize access to SNA has grown by leaps and bounds over resources within a family, rather than the years. Kusimba notes that simply contribute to an individual’s wealth. collaborations with experts in the method Indeed, while remittances are often and in consumer science can introduce a assumed to flow from urban to rural areas, range of models and approaches that Kusimba and her team found that money social scientists might not be aware of. flows in both directions. She says that it can also help to have a data scientist as a co-author, since this Second, promoters of mobile money for helps with the production of journal development often represent the service articles and also peer-review. as a tool that empowers people both socially and economically. This is often Findings true, since sending money via a mobile phone can present a significant reduction The team’s combination of qualitative and in economic and transaction costs quantitative analysis of social networks compared to other kinds of financial resulted in a wide array of discoveries. services. Many of the findings contradict common assumptions about how mobile money However, contrary to the idea that mobile operates as a social and economic tool. money changes the playing field, Kusimba Others illuminate how mobile money states that, “For the majority, mobile interacts in the context of rural Kenya. money is a way of reaching out to 38 Consumer Finance Research Toolkit traditional economic support networks.” purchases or the inability to say no to Moreover, its functions and uses are requests for money. sufficiently different from those of mainstream banking that it does not act as In fact, Kusimba explains, there is a a close supplement. Kusimba argues that growing sentiment that Kenyan social life mobile money is better understood “not as is becoming overly monetized. Aside from banking but as adjunct to the mobile the burden of giving, some urban workers phone,” since the practice of sending and will send money to their rural homes receiving money is closely connected to rather than return in person to participate that of speaking or texting. in rituals. This is fundamentally altering the structure of social life. Third, mobile money is often seen to benefit women. Mobile money incurs The social network visualizations that the advantages for women because it study generated provide a valuable provides a way to make transactions complement to the in-depth interviews. privately, and this can help women gain They showed clearly that people’s some autonomy from their husbands and networks are relatively dense and that other men. Mobile money gives there are many transactions between the individuals more control over their social people within them. Family networks are networks, allowing them to both create based around siblings and mothers, and and sever connections. show a preponderance of matrilineal ties. Moreover, they are reciprocal networks, Yet, as Kusimba notes, while women tend with money moving backwards and to receive a large share of remittances, forwards rather than in one direction. they often view mobile money as These findings contribute to conversations something that helps them cope rather about families, informal finance, social than that empowers them. This is because relationships, and ideas of reciprocity. productive wealth is tied up in land and stock, which are predominantly controlled Applications by men. The combination of in-depth interviews Fourth, an ethic of generosity places with social network analysis has many pressure on people to recirculate potential applications. For example, when remittances, and this can be seen as a we think of remittances we often picture burden. People grow weary of constant urban migrants sending money home to requests for money and may stop their rural families. However, social answering their phones at times when network analysis can bring this requests have a high frequency, such as assumption into question and show the before the beginning of the school year. underlying logic of juggling many ties that People avoid storing money on their informs people’s money decisions. phones out of fear that it will lead to large 39 Consumer Finance Research Toolkit Comparison with other areas of the world More about the method would be fascinating and may uncover contrasting cultural dynamics around InterViews: Learning the Craft of money sending. This approach is likely to Qualitative Research Interviewing by be of interest to policy makers, especially Steinar Kvale and Svend Brinkmann since it emphasizes the importance of (2009, SAGE) groups in mobile money usage. Inside Interviewing: New Lenses, New Ethical issues Concerns by Jaber F. Gubrium and James Holstein (2003, SAGE) Privacy is always an issue with networks. When an interviewee named a contact, Handbook of Narrative Inquiry: Mapping a Kusimba and her team generally tried to Methodology by Jean D. Clandinin (2006, follow up to see if the named person was SAGE) also interested in participating in the study. Some follow-ups declined to take Analyzing Social Networks by Stephen P. part, and it was important not to make Borgatti, Martin G. Everett and Jeffrey C. them feel pressured into participating. Johnson (2013, SAGE) Some people gave contradictory information in terms of size and frequency of remittances. There were people in the study who sent remittances secretly and did not want them to be a part of the network visualisations. While such problems are anticipated, they mean that interview data on networks is subject to some uncertainty and also entails privacy concerns. In general, says Kusimba, East Africans are quite open with information, but some of their mobile money use can be illicit. They often omitted information if it could not be verified by at least one other person, usually the recipient/sender, or if the inclusion of personal information had the potential to cause harm to the participant. 40 Consumer Finance Research Toolkit Figure 9 Dorcas (Broker) Social Network Analysis. Image Courtesy of Sibel Kusimba 41 Consumer Finance Research Toolkit contents of the interviewee’s wallet or the devices they use for banking. Object-Centred Interviews Objects may be static or interactive. Static objects are things that interviewees are Researching objects used in consumer asked to comment on, but not interact finance practices is growing in popularity, with. Questions may focus on an object’s particularly in commercial research that use, design, appeal, or functionality. Flash informs product design and marketing. cards, photographs, video, money, credit User experience researchers, designers, cards, and even sound can all be used as and ethnographers carrying out consumer static objects in interviews. research are leading the development of this method. Interactive objects are things that the interviewee handles, manipulates, or While object-centred interviews are produces during the interview. Asking commonly used in product development participants to handle objects, sort cards, and marketing, the general principles or draw diagrams can elicit feedback on have been applied to many kinds of design and prompt interviewees to share consumer research problems. As the useful information that may not emerge following case studies show, object- during a verbal interview. centred interviews have been used in research that focuses on how objects (money, credit cards, wallets, etc.) prompt Examples of use specific behaviours, including decision-  IMTFI researcher José Ossandón making, household financial planning, and showed how people use credit financial consequences. card invoices to track funds by way of “accounting in the margin”52 [see also appendix, blog post 4]  An anthropologist used a “pile sort” What is it?  Qualitative technique to classify expenses in  Combines interviews with props household budgets in Mexico53  Face-to-face or remote data  CGAP’s human-centred design collection included a simulated income and expense activity54 In object-centred interviews, props are  In his book, Dr. Joe Deville shows incorporated into a verbal interview with how credit card statements and the goal of prompting conversation around debt collections letters prompt particular topics. The interviewer may customers to make payments55 introduce objects, such as a product  IMTFI researcher Charmaine 'Ilaiu prototype or flash cards, or objects may Talei studied hand-woven mats belong to the interviewee, such as the alongside mobile money in Tonga 42 Consumer Finance Research Toolkit to understand how the two are spreadsheets for budgeting, being able to commercially connected56[see point directly to a particular item in the appendix, blog post 3] spreadsheet and ask a question about it can save time and confusion. This technique may be especially useful when Strengths interviewing children as it can help them to understand what the interviewer is Assist in generating discussion asking of them. People are often not accustomed to talking in detail about their use of financial Discuss specific features of tools and services. Props can help the objects interview to open up, and to stay on topic Props allow for the specific features of an by providing a point of focus and by object to be discussed. For example, an encouraging people who may feel shy or interviewer might ask a customer to uncertain. For example, interviewees may demonstrate how they use their mobile find it easier to show the interviewer the phone to send money and comment on contents of their wallet, pointing out the steps involved as they are carried out. various objects that relate to money This gives the interviewee a chance to practices, than to recall all the money- explain what they do and don’t like about related items they carry. the functionality and aesthetics of the object, and to identify any stumbling The method also provides the researcher blocks. with an opportunity to ask about objects that the interviewee has overlooked. For example, asking a person to Whereas the interviewee may only point physically count money may assist in out items they associate with financial understanding precise aspects of financial management, such as cash and credit literacy, so long as the interviewee is not cards, an interviewer may also need to made to feel that they are under know about other kinds of money, such as examination. Asking an interviewee to store cards, or secondary items such as explain which of their bank cards they like identification documents. the most can demonstrate whether aesthetic appearance is important to the Clarify meaning of interview user, or whether their preferences are questions shaped by other considerations. The meaning of interview questions is not always clear to interviewees, and using Assist in information recall and visual / textual aids can help the accuracy interviewee understand exactly what is Consumer finance research often requires being asked. For example, when interviewees to recall details about their interviewing a person about their use of 43 Consumer Finance Research Toolkit financial transactions and the products Limitations they use, but much of this information is either never memorized or is difficult to recall. Props such as wallets, credit cards, Self-reporting bias and bank statements help us to recall Like ordinary verbal interviews, object- information that is recorded on the props based interviews can involve self- themselves or that is in our memories. For reporting bias. However, incorporating example, asking an interviewee who observational techniques can sometimes regularly sends money overseas to show mitigate these. you their receipts will confirm the dates on which they made transactions. Resource-intensive Similarly, interviewees may have multiple Object-centred interviews can sometimes retirement funds and forget where their be more resource-intensive than ordinary money is saved; going through physical verbal interviews, depending upon what documents can help them to reconstruct a needs to be achieved. First, more picture of their finances. Using objects to planning needs to go into selecting, and improve information accuracy therefore possibly creating, appropriate props. helps to counter a range of limitations that Second, object-centred interviews are may be present in verbal interviews. more likely to require additional recording means, such as video cameras or screen capture software. Produce user-created data Some use of props involves asking interviewees to draw maps, diagrams, and other illustrations that can be used as Case Study 1: Combining data. This user-created data can be highly multiple object-centred valuable in cases where interviewees feel they are better able to represent their methods to understand thoughts and behaviours visually rather financial management than verbally. Maps and other diagrams drawn by the interviewee can be saved for Untangling personal finances can be a later analysis, either by photographing complex research task. People often have them or keeping the physical object [see multiple income streams, combine Case Study 1] incomes, or help manage the financial situations of family members. Business records may or may not be kept separate from personal finances. Moreover, people do not necessarily keep all their financial information in the one place. As a result, trying to gain a comprehensive view of a 44 Consumer Finance Research Toolkit household’s or individual’s overall study was that, because finances are financial picture can be difficult to achieve. generally considered to be private in the United States, it was not possible to test As part of a study of financial practices in the methods beforehand with their friends, the San Francisco Bay area, Jofish Kaye family, and colleagues. This meant that from Yahoo Labs and his team conducted the interviews became the testing ground a preliminary study with fourteen for the method. interviewees, aged 26-29, with incomes The team interviewed 15 people, ranging from US$18,000 - US$150,000 endeavouring to recruit a range of people per year.57 They incorporated multiple with different incomes, employment object-centred exercises to try to piece situations, and demographic factors. To together a picture of their financial begin, they sent an email58 to people who management. Their paper “Money Talks: had previously volunteered in their studies Tracking Personal Finances” describes and asked them to fill out a short survey. the study in detail. Here we summarize Participants were selected from this their methods and findings. group. Method Each interview was attended by at least two researchers, and they took place in Kaye and his team wanted to explore the the interviewees’ homes. The interviews range of ways in which people keep track took around 90 minutes each. They began of their finances. They devised an with general questions about the interview structure that incorporated a interviewee’s employment, sources of range of static and interactive objects, income, household composition, and so including: on (see the study protocol59 for more  Financial maps drawn by details). interviewees  Financial calendars filled out by Given that privacy is a major concern in financial research, the team wanted to interviewees  Index cards with text for quickly find ways to break the ice and make interviewees feel comfortable. To interviewees to choose and achieve this, they introduced an exercise discuss that was “something that people couldn’t  The contents of interviewees’ get wrong”: drawing a “map” of their wallets  Guided tours of interviewees’ finances. homes  Computers and mobile devices The researchers emphasized that there was no “right” way to map their finances used for financial management and that it was entirely up to them what they drew. Interviewees drew a variety of Kaye and his co-authors note in their illustrations, including pie charts, pictures, article that a limitation with designing this 45 Consumer Finance Research Toolkit and cartographic-style maps of where expenses’) it ended up being really their financial transactions took place. The useful.”62 researchers report that this helped people feel that they were in control of the This is a good example of how props can interview. be an effective way to encourage participants to overcome their shyness Next, interviewees were presented with and prompt discussion of topics that may sixteen handwritten index cards. In a book not occur to them. chapter60 on the study, Kaye explains: Next up, interviewees were asked to “…each of [the index cards] had a empty out the contents of their wallets or event with financial consequences purses and talk about the items in them. written on it: college, They found a wide variety of payment debt/bankruptcy, unexpected cards in people’s wallets, including credit windfall, unexpected expense, cards, debit cards, store cards, and gift employment, move, family change, cards. Where necessary, they retire, travel, birth, divorce, anonymized objects by placing pieces of marriage, death, medical bills, post-it notes over identifying details, then buy/sell home, graduate. We put photographed them in order to have a these out in no particular order and reference for analysis. asked our interviewee if any had had a financial impact on them This was followed by guided tours of recently.”61 interviewees’ homes. Financial management can occur in many places: Interviewees were asked to select and people keep documents in filing cabinets comment on two of the cards, although and in places where their presence will they report that some interviewees prompt them to take action; they use decided to discuss all of the cards. Kaye computers and mobile devices to manage commented, accounts, budgets, and spreadsheets, and they may store cash in common or “We found the index card method secret places. Interviewees were also remarkably powerful. I’ve never asked to show the researchers the digital seen something that was quite so tools that they use to manage their generative. We were concerned finances, such as a mobile device or a that maybe it meant that we’d limit computer, including online banking, people’s discussions to only the spreadsheets, applications, and financial things we expected to talk programs. about, but I think because we were really careful to make sure that we Looking at where people keep financial included some pretty open-ended information and money-related objects categories (‘unexpected helped the researchers to get a sense of 46 Consumer Finance Research Toolkit how people tracked their finances and to world. However, in terms of specific what extent they integrated management details, their work provides insights into of different aspects of their finances. the range of ways that people manage their finances and is a useful basis on Throughout the interviews, interviewees which to build a larger research project. were asked a range of specific, general, and hypothetical questions. These Few of the interviewees had a included how they had learned about comprehensive idea of their own financial finances (financial socialization), what situation. In fact, many reported keeping they would do if they received an their financial information in their head—a unexpected windfall, what they wouldn’t location that is certainly not suitable for want anyone to know about their finances, object-based interviews. As Kaye and his and finally, what else the researchers co-authors recount, “The most common should have asked them about. tool that people used to keep track of the overall state of their finances was nothing Data were collected using a voice at all.”63 recorder, a still camera, and through old- fashioned note taking. Interviews were Even in cases where interviewees used externally transcribed and photos were computer programs, mobile device further anonymized if necessary. To applications, computer spreadsheets, and analyze the data, multiple reviewers paper-based accounts to track financial tagged themes in the transcripts. In order flows, they rarely tracked every aspect of to aggregate the themes, they wrote them their finances. For example, one up onto post-it notes, accompanied by photographer tracked her business sketches and photographs. They then expenses but not her personal ones, and sorted and summarized the notes, a mother tracked her college-aged generating a list of fourteen meta-themes children’s credit card use but did not track and design opportunities. details of her own expenditures. Findings The researchers point out that interviewees engaged in behaviours that Overall, Kaye reports that this study gave seem “irrational” if considered from a them a “big picture” view of people’s purely financial perspective, but which practices, provided a talking point to make sense when other social norms and return to throughout interviews, and values were taken into account. They perhaps most importantly helped people explain, “People make financial decisions to feel comfortable with the research. based on their emotional, historical, familial and personal backgrounds in The researchers note that the study addition to financial considerations.”64 sample was small, with just 15 interviewees, and is not representative of the United States let alone the rest of the 47 Consumer Finance Research Toolkit As an example, one of their interviewees sticking paper labels onto credit cards so was a woman called Bonnie whose that she can easily tell which are for parents both survived the holocaust. She personal use and which are for business is so thrifty that she will rarely spend use. The second woman keeps most of $1.25 on a cup of coffee, but she is her accounts in her head and retains considering spending $3,000 to give her some fluidity between her personal and $5,000 1985 Nissan 300ZX a new paint business funds. The third has a salaried job. This move would be “based on some job but has multiple other small income factor other than optimizing financial gain. streams from what are essentially informal It seems clear this car has emotional micro-businesses. value that exceeds its financial value.”65 These examples suggest that while Another interviewee, Doug, used a system people do separate funds, they may not of index cards to cross-check the interest do so in ways that we would expect. he was earning on his $1.2 million in Some people may be less concerned with municipal bonds. In fifty years, he had separating business and personal only ever found one small error. The expenses, and more concerned with opportunity costs of the time spent maintaining divisions that are meaningful checking for errors had never paid off, nor to them, such as separating out money to would it likely pay off in the future. pay for a wedding or for college. Some of the most interesting findings Overall, this study indicates that people’s concerned how people divide up their methods of financial management are money into separate pots. Many social dictated by what is important to them, and scientists, including Viviana Zelizer and may have little or nothing to do with Wolmarian Mesfin [see Ethnography], optimal financial decision-making. When have noted that this practice is common in trying to understand people’s financial many places around the world. It can take behaviours, then, it is important to explore the form of dividing personal funds from their motivations. business funds, or engaging in goals- oriented savings for personal use. For Applications example, some people might have a bank account specifically to save for holidays, Although the sample size was small, this or a coin jar to spend on leisure activities. study has numerous potential applications. Given that the interviewees As examples of dividing funds into pots, were not using existing systems for Kaye and his co-authors describe three financial integration, the authors suggest female interviewees who both run small that there are opportunities “for rethinking businesses. One keeps a strict divide those systems, but also for the design of between her personal and business novel financial interaction experiences.”66 expenses, keeping paper accounts and 48 Consumer Finance Research Toolkit Most of the interviewees in this study during the interview, rather than wait until checked their bank balances on a daily it was completed. All objects with personal basis, often multiple times per day. information on them, including Features could potentially be built into spreadsheets and the objects in people’s online accounts that capitalize on this wallets and purses, were anonymized behaviour, such as showing “prompts for before photographing them by sticking good financial behaviour.” Another idea post-it notes over names and any other emerging from the study is to allow people identifying details. The researchers to have multiple savings accounts that contend that this is a far more reliable and they can earmark for different purposes. efficient way to protect participants’ privacy than trying to anonymize This kind of research could also be photographs later. Moreover, they do not incorporated into the design of “financial believe that it reduced the quality of their literacy” programs and other interventions data, as the details lost were not integral aimed at improving financial behaviour. to the insights they wished to gain. Clearly, people’s financial concerns do not line up neatly with what financial planners would advise them to do. Case “tudy : Porta le kit Learning formal financial management is studies on poverty and a useful, and perhaps indispensable, life skill, but financial well-being does not personal finances in Haiti arise from financial health alone. Rather, A portable kit study is a method of financial well-being occurs when people interviewing people about the objects they are able to achieve the financial goals that carry with them away from home. matter to them personally. Studying the financial tools that people carry with them on a daily basis as they As well as teaching people generic work, shop, and socialize helps us financial skills, financial literacy programs understand how they meet their financial could help people to achieve their goals. needs. People need to have financial tools In fact, focusing on goals may help on hand to navigate everyday and motivate people to learn financial unusual transactions. This can be as management techniques that would simple as carrying cash or cards, but otherwise not interest them. there are many other financial transactions that people undertake away Ethical issues from home, including debt payments, As with all consumer finance research, money transfers, and re-charging phone privacy issues can be difficult to navigate credit. in object-centred interviews. In this study, the researchers decided to anonymize as Dr. Erin B. Taylor and Professor Heather much personal information as possible A. Horst undertook a study of mobility and 49 Consumer Finance Research Toolkit finances among poor Haitians living on border on a regular basis. Their income the border of Haiti and the Dominican ranged from zero to around 5,000 pesos Republic between 2010 and 2012.67 The per month (US$111). towns of Anse-à-Pitres (Haiti) and Pedernales (Dominican Republic) are All of the portable kit interviews were right next to each other and many Haitians conducted in Spanish, as it was the main cross the border daily to shop, work, shared language among the research socialize, and access services. team. While it would have been preferable Haitians living in the region access to conduct a number of the interviews in financial and other services on both sides Haitian Creole, all but one of our of the border. One of the goals of the interviewees spoke fluent Spanish. This is research was to learn how this financial common among most Haitians living on binationalism affected their economic and the border of the Dominican Republic. social lives. After three months conducting interviews, observations, and a In preparation for the study, the questionnaire in the region, Taylor and researchers asked participants to bring Horst carried out a “portable kit” study with with them everything they carried on a a subset of ten interviewees. “normal” day. Most of the portable kit studies took place in a quiet area of a Method local hotel on the Dominican side of the border, as this created greater privacy for Participants in the study were all of interviewees and less disruptions than Haitian nationality, and one interviewee carrying out interviews in their homes. had dual Haitian / Dominican nationality. Participants for the portable kit study were The portable kit interviews began with the mainly recruited from among people the researcher explaining the purpose of the researchers had already interviewed and / research and the method. If an initial or surveyed. In most cases, background interview had not previously conducted, interviews had focused upon employment, they started the portable kit interview with family, impressions of the border region, a set of background questions. and access to technology and financial services. Completing the portable kit study To begin the object-centred interview, the at the end of the research meant that the researchers requested that participants team had built up rapport with a sizeable take all of the objects they carry with them number of people and could be sensitive out of their bags, pockets, and wallets, to issues of privacy. and display them on a flat surface. This included objects related to consumer Approximately half of the interviewees finance (currency, cards, receipts) and were living in the Dominican town of other items (identity documents, Pedernales, and the other half were living photographs, face cloths, pens, etc.) in the neighbouring Haitian town of Anse- à-Pitres. Most respondents crossed the 50 Consumer Finance Research Toolkit Interviewees were invited to discuss the topics based on what their interviewees objects generally. The interviewers often told them. began by asking about the bag or wallet itself: where it had been purchased, why Interviews were transcribed by research they chose that particular item, and assistants and translated into English. The whether they own other bags that they researchers then went through the switch between depending upon their interviews to tag key themes. Video plans for the day. The benefit of this initial footage was summarized, particularly step was that it gave participants a noting visual aspects such as the chance to relax and the researchers an condition of money, branding on cards, opportunity to build rapport. mobile phone personalization, and so on. Then the interviewees were asked to Findings divide the objects into two piles: one of objects that they must carry with them The portable kit study enabled the every day, and a second pile of objects researchers to understand how people that could be left at home. The living on the border used financial interviewees were asked to explain why services in their everyday lives. The some items were more important than objects people carried with them reflected others. This step was crucial in identifying both this formal financial environment and what people felt they needed to navigate alerted the researchers to a range of their everyday lives. As the interviews other, less visible financial practices. progressed, the interviewers formulated Interviewees carried a range of items questions that would tell them how relating to finances from both sides of the different objects were involved in personal border, including different currencies, and household finances. credit cards, receipts for money transfers, bank documents for debts, and mobile The portable kit interviews were recorded phones used in transactions. They also using an audio device, a video camera, carried items such as house keys, and a still camera. Interviews ranged from photographs of family members, bibles, forty-five minutes in the case of one pens, paper, and mobile phones. participant who had already been interviewed multiple times and did not Most interviewees reported that their most carry many items, to three hours for important items were identity documents, people with many items or an unusually followed by keys, cash, and their mobile extensive collection of phones and SIM phones. Without identity documents, we cards. The interviews were semi- were told, “you cannot do anything”: it is structured: while the researchers followed impossible to collect a money transfer, the same general process for each buy a SIM card, or even use many interview, they also followed particular services. Moreover, interviewees felt that the Dominican police could harass them if 51 Consumer Finance Research Toolkit they did not carry their identification with At the time of the research (February - them. May 2012), a microcredit bank called Fonkoze had just opened in Anse-à- Cash was the second most important Pitres, Haiti. Before this, there was not a item. Whereas Dominican pesos are single bank in Anse-à-Pitres, while there accepted throughout the Haitian town of was a bank and a credit union in Anse-à-Pitres, Haitian gourdes are not Pedernales, Dominican Republic. Using accepted in the Dominican town of one of these Dominican banks required Pedernales. This means that Haitians Dominican residency, so few Haitians had must carry Dominican currency, whereas any bank account at all. Dominicans will rarely (if ever) need to One exception was a woman called use Haitian currency. All of the Bronte, who had both Haitian and interviewees carried Dominican pesos, but Dominican citizenship, and was thus able few carried Haitian gourdes. Interviewees to open an account in the credit union. reported that, when they crossed the During her portable kit interview, Bronte border, they left their Haitian gourdes at showed the interviewers her bank card home. and some papers. She explained that she had taken out a loan to buy her Haitian Interviewees were also asked, “How much husband, Emmanuel, a motorbike so that money do you carry with you on a daily he could start a microenterprise ferrying basis?” Responses ranged from 50 pesos passengers and fetching goods for people to 500 pesos, but most were around 200- (known as “motorconcheando”). She 300 pesos (roughly US$4-7). This was explained that they had nearly paid off the enough money to pay for transport, food loan, mostly from her income as a and drink, and other small expenses, such receptionist in a local hotel. Bronte carried as topping up phone credit. her loan documents with her because she considered it safer than keeping them at This question provided a sense of the cost home, where multiple people (including of living as well as how much people felt children) had access to her house. they could afford to pay for items. It also led to further discussion of how the By focusing on all objects that people interviewees made a living. For example, carry rather than just money-related ones, some interviewees explained to us that the study helped to identify the reasons Haitians working in the Dominican why loans and capital investment might Republic are often paid significantly less not result in significant increases of than their Dominican citizens, although income. Emmanuel’s income remains low this depended somewhat on the time they despite their investment in his motorbike, had been living in the area and whether largely because there are more men they had formed strong relationships with employed in this profession than there are Dominican employers. customers to support their businesses. 52 Consumer Finance Research Toolkit In order to operate in Pedernales, he had The interviews also provided insights invested in a fake Dominican ID card. about who sends money to whom and Emmanuel reported that because he lived how they send it. These cross-island links on the Dominican side of the border and became evident through looking at lived with a Dominican-Haitian, the people’s receipts. Physical records are authorities assumed that he was also of useful because they provide a fool proof dual nationality. This identification also record of transaction details, such as the saved him money, as he did not have to date and amount received. This can help pay bribes to cross the border. However, interviewees give interviewers a better his fake ID did not fool the locals, who idea of interviewees’ income and how it is knew he was Haitian and would not let affected by long-distance relationships. him join the Dominican union who Money circulates between Haitians and coordinated this business in Pedernales. their family members or friends living in This meant that he could not wait for different parts of the island, on both sides customers at designated motorbike of the border. The reasons why people stands, so he relied on a regular clientele send money are various, and the (mostly of Haitians) that he had built up circulation occurs in multiple directions. over time. For example, a woman called Fredelina received money regularly from her In the absence of banks, the interviewees boyfriend in the east of the Dominican used other formal and informal services to Republic. In another case, a Haitian manage their money. The main formal woman living on the border sent money financial products used by Haitians living regularly to her children who were in the border region tended to be the studying at university in Santo Domingo. remittance services available on both sides of the border. Until 2011, there was Applications no formal money transfer service in Anse- à-Pitres, so Haitians had to use the Poor people often have complex financial Western Union or Caribe Express offices lives and they use a range of financial in Pedernales. If they were receiving tools to manage difficulties such as money from family living in Haiti, they shortage of funds, lack of access to would sometimes use informal service infrastructure, and irregular income. providers, such as the boats that carry However, just as Jofish Kaye and his goods, people, and money between Anse- team discovered in the first case study, à-Pitres and Marigot, located eighty people do not usually keep their accounts kilometres away. Interestingly, even when in one place. more money transfer services became available, some people continued to use This combination of diverse financial the slow boat service, as it was free. practices and dispersed recording of transactions presents a research challenge as it makes data collection expensive. However, combining methods 53 Consumer Finance Research Toolkit such as verbal and object-centred Commercially, product design can clearly interviews can help to build an integrated benefit from understanding how people picture of people’s financial transactions, incorporate objects into their everyday and the insights generated could assist lives. with the design of larger research projects. Ethical issues A problem with all research into poverty is The ethical and practical issues raised in that researchers can find it difficult to this study are similar to those faced by know what questions to ask, since their Jofish Kaye and his team in Case Study 1. lives often do not resemble those of the However, the fact that this research people they are studying. Researchers focused on poor and marginalized people are often from wealthy countries or raises further issues. demographics, own multiple technological devices, are highly mobile, and have a One research problem is whether to pay greater degree of financial security than interviewees for their participation. Ideally, their interviewees. They are not interviewees should be compensated for necessarily familiar with the range of their time, but it does pose a number of financial services that poor people use, problems. First, research on personal and they may not be aware of other issues such as finance and everyday life factors that impact the lives of poor and requires the building of relationships. This marginalized people, such as security is difficult to achieve if residents view the concerns or the need use fake researchers as providing employment or identification. Object-centred interviews welfare. However, building personal can help reduce the gap between relationships can be difficult regardless of interviewer and interviewee because the whether money has changed hands, since things that people carry and use are a researchers often have very different solid basis for asking questions. backgrounds and levels of wealth compared with their interviewees. Object-centred interviews have clear applications in policy work, socioeconomic Second, paying participants means that development, and commerce. Making those with the most need for cash may be good policy requires good information, more likely to participate in the study. This and object-centred interviews can be used introduces a research bias towards one in an exploratory way to shape research, segment of the population, and it also or to clarify issues arising from other data raises questions about whether collection methods. Microfinance and interviewees are participating voluntarily. other lending institutions can benefit from knowing more about their customers to To mitigate these problems, participants reduce the risks involved in lending and were only paid for the portable kit study at extend loans to more people. the end of the research period. They received 100 pesos (around US$6) for 54 Consumer Finance Research Toolkit each hour of the interview, an amount that Participant mapping: An innovative is above the Dominican minimum wage. sociological method (Morgan Centre) Most portable kit participants were people who had already contributed to the study, Situating Everyday Life: Practices and so paying them for the portable kit study Places by Sarah Pink (2012, SAGE) was a way of compensating them overall. It also meant that most of the participants Personal, Portable, Pedestrian: Mobile were people who participated willingly in Phones in Japanese Life edited by Mizuko the study because they were interested in Ito, Misa Matsuda and Daisuke Okabe the research. (2006, MIT Press) Another ethical issue was how to provide people with sufficient opportunities to ask questions about the research. At the beginning and end of each interview, participants were asked if they had any questions or if there was anything else they would like to tell the researchers. Often they wanted to know more about what the research would be used for. Although the researchers explained the study’s aims and nature during the recruitment process and at the beginning of each interview, it was often not until after the interview had taken place that participants thought of questions to ask. This is a good illustration of how important it is to provide participants with opportunities to speak up during different phases of the research. More about the method Using participatory visual methods (Morgan Centre) Participant produce video toolkit (Morgan Centre) 55 Consumer Finance Research Toolkit Figure 10 Common items in a portable kit. Image courtesy of Erin B Taylor and Heather A Horst Figure 11 Object Centred Interview, Haiti (Erin Taylor) 56 Consumer Finance Research Toolkit Figure 12 Financial event cards. Image by Jofish Kaye 57 Consumer Finance Research Toolkit Figure 13 Financial event cards. Image by Jofish Kaye 58 Consumer Finance Research Toolkit Figure 14 Financial event cards. Image by Jofish Kaye 59 Consumer Finance Research Toolkit The financial diaries method has since been applied by numerous social Financial Diaries scientists working in different parts of the world. The method has proved useful for What we currently recognize as the understanding households in wealthy financial diary methodology was nations as well as in poor ones (such as pioneered in the late 1990s by a group of the US Financial Diaries), and has proved researchers with expertise in economics, useful for commercial projects as much as finance, anthropology, development, and ones with non-profit goals (see the case architecture. studies in this section). In the late 1990s, David Hulme (credited Depending on a research project’s goals with the diary idea), Stuart Rutherford, and resources, the financial diary method Jonathan Morduch, Daryl Collins, and is adaptable and can be used to collect Orlanda Ruthven developed financial any range of data, from small, exploratory diary research methods based on three studies to large, statistically representative very different projects (qualitative and ones. The case studies in this section give quantitative) that involved returning to examples of two very different kinds of households at intervals (usually every two implementation. weeks) to collect data. Using questionnaires and observations, they What is it? would record households' cash flows in the intervening period, including where  Qualitative / quantitative their money came from, their spending  Combines self-reporting with and savings patterns, and the financial interviews instruments they used to manage their  Face-to-face or remote data money. collection In their book, Portfolios of the Poor, the Financial diaries are a method of authors explain why the diary method is collecting data on financial behaviours by such a powerful tool: using a “diary” to record transactions. Financial diaries are not technically diaries “…finance is the relationship in the way we normally understand them. between time and money, and to Respondents do not simply write down understand it fully, time and money their thoughts, but rather are given a must be observed together.”68 structured set of questions that record both qualitative and quantitative responses [see Case Study 1]. 60 Consumer Finance Research Toolkit Questions are generally designed to elicit Examples of use the reporting of both formal and informal financial activities and how these fit into  The US Financial Diaries tracked the context of a participant’s life. They can 235 low- and moderate-income include open questions and closed households over the course of a questions, thus producing both qualitative year to collect data on how and quantitative data. Financial diaries families manage their finances on can also be combined with other methods, a day-to-day basis69 such as different kinds of interview  Microfinance Opportunities used techniques, observations, or even field financial diaries in their research in experiments such as RCTs [see several countries, including Kenya, Experiments]. Malawi, and Zambia, as a learning tool to enhance a savings group A standard approach to using financial program70 diaries is to choose a sample population  The Southern Africa Labour and of households or individuals, design a Development Research Unit financial diary that suits the research (SALDRU) has information on questions, and apply the diary at regular Daryl Collins’s study using intervals (e.g., every week or two weeks). financial diaries as a learning tool While the data collection period varies for a specific MFI (SEF), targeting from study to study (usually between a their clients71 month and a year), all financial diary  CGAP’s Smallholder Households studies capture participants’ changing Financial Diaries Project, also financial behaviours over time. implemented through Bankable Frontier Associates, is examining Participants may fill out the diary how smallholder households themselves, or an interviewer may fill it combine agricultural and non- out for them (especially in cases where agricultural sources of income and literacy might be an issue). Diaries may employ a range of financial tools in be recorded on paper, using a mobile three countries (Pakistan, application, or directly into a laptop Tanzania, and Mozambique)72 database. They can be implemented in  FSD Kenya with Bankable Frontier person or using online programs. Associates undertook a financial diaries project between 2012 and Depending on a research project’s goals 2013 that took a broad look at the and resources, the financial diary method financial behaviour of a diverse is adaptable and can be used to collect sample of Kenyans73 any range of data, from small, exploratory  Researchers from the Australian studies to large, statistically representative National University and the ones. University of Burdwan in India used financial diaries to study the 61 Consumer Finance Research Toolkit financial behaviours of people Commercial work can also benefit from living on river islands in Bengal74 financial diary studies. It is often useful for  Freedom from Hunger used financial services providers to find out how financial diaries in Mali and customers use their own products and Ecuador to better understand those of competitors. New financial young people living in poverty75 products and services are developing  A Bankable Frontier Associates rapidly around the globe, and keeping up and The MetLife Foundation-led with customers’ changing preferences and Mexican financial diaries project in use patterns can assist companies to Mexico City and Oaxaco offers adapt their products to these changes. crucial insights into the ways that families are juggling borrowing and Track financial behaviour over instalment credit to bridge gaps in volatile income patterns.76 time Financial diary studies are carried out over a period of time, usually ranging from a few weeks to a year. Because Strengths participants are generally asked the same questions at intervals over the entire time Shows range of financial product period, it is possible to observe how individual households or families address use many different kinds of needs, Financial diary studies can yield insights opportunities, and challenges. These into financial management by individuals observations can include juggling volatile and households everywhere, and of all incomes, coping with crises, spending on levels of financial means. However, they a major event, or investing in capital. are best known for transforming our understanding of money management Record information creatively among poor families in “developing” countries. The diary format makes it possible to build in creative ways for respondents to One of the most striking findings of the answer questions. This is particularly the financial diary studies described in case for self-reported diaries. Portfolios of the Poor is it demonstrated Respondents can be asked to provide an that poor people use a wide range of array of low-tech or high-tech information formal and informal financial products. types, including written answers, This insight, now confirmed by multiple numerical answers, choosing from a studies, contests the assumption that poor scale, drawing pictures, generating maps, people are not active money managers. adding photos and videos, and attaching This insight can potentially lead to new documents such as bank statements. This policies and interventions.77 is also important for identifying differences 62 Consumer Finance Research Toolkit in local concepts and ways of categorizing policies or design financial products that savings vs. investments, for example. better fit people’s existing strategies and needs. Increase financial awareness A potential side effect of financial diaries Limitations is that they may help participants to become more aware of their financial Representativeness behaviours, leading them to make changes in their lives. This might be When researchers carry out diary studies achieved passively, such as when a in person, they face constraints on how participant’s involvement prompts them to many diaries they can feasibly collect. think more carefully about their financial This makes it difficult to achieve a behaviour. representative sample. If a research group wanted to find out about the financial As is the case with ethnographic and any behaviours of, say, Haitians living on less other research methods, it is important to than $5 per day, then they would require keep in mind that respondents act an enormous amount of researchers to differently when being observed. They carry out the study with a representative don’t always tell researchers what they sample of Haitians. However, if the want to hear and they may work to keep research focuses on a small geographic up appearances to meet standards of area, such as a village, then achieving a politeness or respectability. representative sample is possible. In practice, most financial diary studies ask However, further research needs to be questions about broader populations, but undertaken before it can be confirmed that need to cluster the sample geographically such changes have taken place in the due to resource restrictions. This means long term. Moreover, a potential problem that their ability to extrapolate their with viewing financial diary participation as findings to an entire population is limited. having an educational effect is that it rests on the assumption that families were Many financial diaries studies do not aim doing something “wrong” to begin with.78 to be representative. Instead (like ethnography) they provide fine-grained Financial Diaries researchers see their detail, qualitatively and quantitatively, role as one of listening and recording the about peoples financial lives – not only details of people’s financial lives. It is not what people “earn or spend, but in what about telling respondents to do things frequencies, amounts, and modes.”79 This differently or to give them lessons in can help to generate new questions and money management. Rather, financial hypotheses that can be tested with diaries, done well, can show when and complementary RCTs, for example, to how people use different types of tools. show a specific cause and effect Findings can be used to recommend 63 Consumer Finance Research Toolkit relationship or capture an effect at a more they are dependent upon participants representative scale.80 being at home or available, and this is not always achievable. When participants Self-reporting bias self-report, they may be better placed to fill out their diaries on time; alternatively, All research runs the risk of incorporating they may forget or have other core the biases of the researchers and activities that take precedence over their participants involved. Self-reported data research participation. can be particularly problematic because there are fewer opportunities for According to Julie Zollmann, data researchers to control the data collection accuracy is not particularly affected by process. minor deviations in timing. However, timing issues can strain the research For example, participants may not project's resources, since costs increase understand a question fully, but if a when researchers have to spend time researcher is not present when the diary searching for respondents. is filled out then there may be no opportunity for them to check the In some cases, data collection can be question’s meaning. Whether this made more efficient when the target group compromises a study will depend on the meets regularly for an important event, project’s aims. such as a meeting or a social activity. Scheduling data collection for times that fit Generally speaking, the more oriented a with the participants’ collective schedule research project is to collecting data on can be effective, so long as the group is people’s values and opinions, the less comfortable with the research taking place such bias will interfere with the results. within that time slot. Self-reporting is generally unreliable in cases where accurate numerical information is required.81 With more qualitative studies, however, Case “tudy : The portfolios of misunderstanding bias often goes away over a few interviews. This is an the poor in Bangladesh, India, advantage of diary studies over surveys. and South Africa The Portfolios of the Poor authors Timing of data collection developed the financial diary “research One important feature of financial diaries technique”82 because they wanted to is that they are generally collected at better understand the diversity of financial regular intervals. Participants are asked to tools and money management strategies give responses on particular dates or of the poor in “granular detail” generally even at set times of day. However, when missing from studies of the poor at that researchers fill out the diaries themselves, time. Most studies depicted poor people 64 Consumer Finance Research Toolkit as having few financial tools available and Method gave the impression that the poor lived Between 1999 and 2005, the researchers purely hand-to-mouth.83 undertook financial diary studies with over 250 families in Bangladesh, India, and In fact, as the financial diaries showed, South Africa. Teams of researchers poor people often have a greater range of visited the homes of participants every financial tools at their disposal than two weeks for the course of a year in each people who live in wealthy countries. This country, recording information about their is because they use a range of informal as saving, spending, lending, and insuring well as formal tools. practices. Part of the reason why that knowledge Participants were generally residents of a about poor people’s use of financial tools small number of communities. In order to was limited was because financial service choose research locations, the providers who undertook many of these researchers made use of national studies were only concerned with how surveys, but they were also guided by their customers used their own products. practical considerations, such as choosing According to Rutherford, banks in communities that were within reasonable Bangladesh acted as though they worked travel distance. in a vacuum, as though the poor had no financial partners other than themselves. To choose households in India and South Africa, the researchers used a technique In order to address these problems, the known as “wealth ranking.” This involves researchers set about developing a new asking residents to rank the wealth of their method. They explain: neighbours and compare the results. The logic is that people are likely to misreport “What was needed was a method their own financial position, but in small that would capture the richness communities they often have a good and complexity of poor people’s understanding of the financial position of financial lives while being their neighbours. Wealth rankings allowed systematic enough in its data the researchers to select participants from collection to prevent it from being the bottom, middle, and top of the list. dismissed as a set of mere They report that an additional bonus was ‘anecdotes.’”84 that it also gave participants a sense of ownership of study. The researchers were The team wanted to retain the richness of not able to use this technique in data that qualitative research produces, Bangladesh because people moved while also generating sufficient around so much that they did not know quantitative data of high quality to show enough about each other. general patterns in behaviours across their research sites. 65 Consumer Finance Research Toolkit Each study began with an initial two cash flow and compared data across the interviews that allowed researchers to get three research sites. to know a household. Then, the researchers returned to the households Findings every two weeks for a year, using a variety of forms to record information on What makes these studies remarkable is financial practices and new developments. that they demonstrate clearly how poor people cope with unpredictability by using In order to increase the accuracy of self- a wide range of financial tools. They use reporting, after every visit the researchers different tools in tandem to achieve calculated the “margin of error” in savings targets and pay off debts. A table responses by comparing incoming and of all the formal, semi-formal, and informal outgoing expenditures. On the next visit, financial instruments that the team the interviewer would ask further discovered can be found in Appendix 1 of questions to try to find out where the Portfolios of the Poor. difference had arisen. Researchers explored the emotions that accompanied By visiting interviewees regularly and interviewees’ transactions, as well as the following up on points that were previously characteristics of the transactions unclear, the team made various findings themselves. that they were not expecting. One of these was the habit of “moneyguarding,” a Interviews often took place while practice of leaving money with neighbours interviewees carried on with their work, and friends for safekeeping. Sometimes such as cooking lunch or feeding cows. people chose this option because it was Visitors often interrupted them. These more convenient than storing money in a conditions were not always favourable for bank. At other times, people’s distrust of data collection, but gave the researchers banks was the driving factor. One man in a chance to observe everyday life. The Bangladesh had sizeable savings that he researchers took care to listen to their used to keep in a bank. However, he participants, but not to offer advice or eventually gave them to a friend to mind opinions. because he had an overdue loan and did not want the bank to know that he had Once data collection finished, the data savings. was analysed using both qualitative and quantitative methods. The results were Another case is that of Thabo, a man compiled into “portfolios” consisting of the living in South Africa. Thabo received balance sheets of the households and money from time to time through his bank. qualitative data detailing their At first the researchers supposed that circumstances and experiences. The someone must be sending him money, researchers also conducted analyses of but after many conversations it turned out that these were interest payments on a deposit. Thabo had become retrenched 66 Consumer Finance Research Toolkit some years previously and had deposited Microfinance agencies are most prevalent his lump sum payment in the bank. He in countries where levels of formal would usually reinvest the interest he banking are low, national data collection is earned, but sometimes he would withdraw scarce, and credit bureaus do not exist. it to spend. Agencies have therefore found it difficult to assess risk, anticipate how households Much like with ethnography, the long-term will spend the money they receive, where relationships developed between else they are getting loans from, and find interviewers and interviewees in financial out what other financial instruments diary studies improved data quality people use that may impact their financial because they enabled the development of health. trust and provided time to explore research themes. As the authors note, Financial diaries help microfinance agencies to improve client information “None of this is peculiar to poor because they provide a way to collect people: in developed economies data on every aspect of a household’s people may also be unclear about financial position. More broadly, financial their financial actions and may diaries can help identify a range of possibly be even more reticent. household financial behaviours that may But the strength of the diaries otherwise go unnoticed, and they can approach is that it can, over time, provide a wealth of recommendations for break down much of this reticence all kinds of organizations working with and confusion.”85 poor households. In many respects, individual’s stories have Researchers at Digital Divide Data and generated more ground-breaking Bankable Frontier Associates revelations than have the project’s implemented the Kenya Financial Diaries, quantitative outputs. Individual cases which tracked the detailed cash flows of capture behaviours that are rarely 300 low-income families for a full year. recorded in surveys, bank statements, or One of this study's most valuable one-off interviews. contributions was to shed light on how poor people use social networks to Applications provide a safety net. While it is well known that poor people depend upon such Since Portfolios of the Poor was published networks, we rarely have a nuanced view in 2006, the financial diaries method has of their benefits and limitations. been widely used by development organizations and companies working with In a blog post on the CGAP website,86 the poor, especially in the area of Julie Zollmann points out that many poor microfinance. people prioritize investment-related saving over short-term liquidity. While this is a 67 Consumer Finance Research Toolkit sound strategy for improving one's They undertook specific analyses of financial position in the long run, it can health and education financing, and these create cash shortages in the short run, helped providers to think about service since people's savings might be locked options that might be more effective in away in savings groups. This means that helping people finance key life needs. people need to be able to borrow quickly. With respect to risk, they have been This explains why, in Kenya, M-Shwari is helping providers look beyond insurance, so popular despite its short loan period given that people are often unwilling to tie (30 days) and high fees (7.5%). up their funds in purpose-specific risk mitigation. In another post,87 Zollmann explains how the financial diaries showed that networks Financial diaries are equally useful in the are often inadequate channels to raise commercial world. As the Portfolios of the money for unforeseen expenses (such as Poor authors note, collecting data on the medical costs), may take too long to financial practices of wealthy people can deliver, and often serve women more than be just as difficult, if not more so, than men. Moreover, households that give collecting data on the poor. Banks can money away may find themselves short of benefit just as much as microfinance cash to cover their own expenses. agencies from knowing more about their customers’ financial practices. This is These findings echo Sibel Kusimba's especially the case today given the rapid observation, uncovered through her social changes sweeping the banking and network interviews, that Kenyans grow payments industries. Financial diaries can weary of being asked to contribute funds show how people adopt new services and within their social networks, and that they what factors influence their decisions. find ways to avoid financial reciprocity [see Case Study 2 in Interviews]. Government bodies also stand to learn much through applying the financial Zollmann provides a number of diaries method. For example, using suggestions for how we may harness financial diaries to learn about the these insights to better meet the needs of financial behaviour of socially the poor, including marketing financial disadvantaged groups may help to show services for households that give money how issues such as financial literacy and away regularly, and improving information decision-making contribute to flow so that households in need can find disadvantage or can help overcome it [see support more efficiently. Case Study 1 in Verbal Interviews] Importantly, FSDK funded an additional According to Zollmann,88 they also year of work after the study was published suggest that financial literacy programs to help service providers and funders are not as useful as we may think, since incorporate key insights into their work. they show how people are highly skilled at 68 Consumer Finance Research Toolkit managing their own money. A lack of When reporting analyses, care must be resources, not a lack of skills or taken to hide participants’ identities. This knowledge, is the real problem faced by is more difficult than people often imagine, the poor. Driving down the costs of since it is relatively easy to identify a banking is one specific way in which person based on very little information. financial tools can become more Moreover, funders may apply pressure to affordable. share participants’ personal stories and information in ways that compromise their Ethical issues anonymity and privacy. According to Julie Zollmann, these can in include: Financial diaries raise ethical issues similar to those that are found in any other “…pressure to make the data study. These include confidentiality, ‘open’ including really deep privacy, and coercion. They are time- qualitative responses that can intensive and may be burdensome to compromise confidentiality pretty participants who much must make time in easily; pressure to do more their otherwise busy schedules and social geospatial analysis, where our obligations. Indeed, the fact that financial funders really want people’s GPS diary studies often involve repeated points and to make maps that interviews could also exacerbate the betray respondents’ locations— difficulties of ensuring the privacy of including some who tell us about participants. crimes during the diaries; pressure from funders to add visuals, like Researchers who are undertaking repeat photos and videos. It takes a lot of studies often gain participants’ trust to a work to ensure images and stories greater degree than is usual because they are kept separate. It’s much easier return to households on a regular basis to give in to donor pressure, and come to know their participants well. particularly for institutions who Participants may become confident don’t answer to IRBs [internal enough to share sensitive information that review boards].”89 could work against them if it were widely known. For more information on protecting participants’ identities, see the guidelines Extra care must therefore be taken to published on Forum: Qualitative Social protect participants’ identities and data. Research90 or Mark Israel’s book This includes being aware during Research Ethics and Integrity for Social interviews that other people may be Scientists.91 listening in to the interview, and providing interviewees with opportunities to interrupt the interview. After each interview, care must be taken to store data securely. 69 Consumer Finance Research Toolkit Case Study 2: Online financial Mack’s financial diary was conducted entirely online, using software called diaries: A short-term Revelation. Participants were able to record their diaries in their own time over commercial project the course of a week. The first time they Financial diaries can be used for logged in, participants were asked to commercial research as well as in the agree to the terms and conditions of the area of socioeconomic development. project, and to choose a screen name and Given that commercial organizations a password. might have fewer resources to spend on research, and often require fast Over the next week, participants were turnarounds, shorter versions of financial required to log in to the site each day and diaries can be particularly useful. complete a variety of activities. These included answering questions, keeping Alexandra Mack, a Research Fellow at logs of some financial interactions, and Pitney Bowes, conducted financial diaries having group discussions with other as part of a study of financial participants. They would also take communications management in the pictures using their digital camera or United States.92 Mack was interested in camera phone and post them to the how “financial communications” impacted project site. financial management within a household. She had already used other methods, Mack individually emailed participants including interviews and scrapbooking, to before the start of the study using her collect data on financial behaviour. The work email address (displaying all her financial diaries were an opportunity to dig contact information). This email welcomed deeper into some of the issues she had the participant and told them specifically discovered, such as how financial what to expect from the study. Mack management varies by life stage, and reports that an additional advantage of factors that impact attitudes toward new emailing participants directly is that technologies for managing finances. participants know who the researcher is and how to get in contact, and that they Method are interacting with a real person. Participants were recruited through a Each day of the study, Mack sent an email professional recruiting form and were in the morning via a group email, using informed that, if they successfully BCC so that the participants could not see completed the study, they would receive a each other’s contact information. In that one-off payment of $150. All participants email she reminded them of whatever were over 21 years of age, had household daily tasks they should include in their incomes over $50,000 a year, and lived in diary. This was a useful way of reminding the United States. participants to fill out their diary because 70 Consumer Finance Research Toolkit they were all using email every day in the course of their everyday lives. “This lets them know that they aren’t communicating into a black Mack observes that methods to prompt hole. Also, the follow-up questions participants may change depending upon help to get more details and clarify the characteristics of the demographic and encourage more engagement taking place in the study. In some cases, and longer answers.”93 such as where people have mobile lives and occupations, text message reminders This follow-up technique is similar to that might work better. The key point is to outlined in the Portfolios of the Poor case remember you are asking them to go to study. However, whereas the Portfolios of something “new” for the period of the the Poor studies used teams of study. researchers and took place in diverse locations, Mack worked alone. She notes The first question in the diary study asked that, while resource-intensive, a reason the participant to talk about them. As with why this study worked well was because the Portfolios of the Poor studies, and doing the diaries online permitted the indeed most other qualitative research, recruitment of a greater number of asking general questions is essential to participants than would have been help researcher and participant get to possible if the study were carried out in know each other, and to give participants person. a chance to communicate their own point of view. Mack found people to be quite willing to share information online, although this is Participants were asked to report every at least partly because participants were day on communications they received not asked to share financial details such from banks and billers, as well as on as bank balances or account numbers. financial interactions other than shopping. However, the online nature of the study Other questions asked participants to also posed a disadvantage, in that by not discuss their use of mobile applications, being physically present in their space it practices around bill payments, and their was not possible to observe their experiences with fraud. In group- behaviours. This made it difficult to know discussions, participants were asked what other possible questions should be questions such as, “What annoys or asked. bothers you most about your financial communications?” Findings Mack tried to give some feedback to their Mack found the method to be suitable to responses daily, whether in the form of a draw a broad picture of people’s financial “thank you” note or a follow up question. behaviours, the products they use, and She explains: their financial communications. While not longitudinal, she was able to ask 71 Consumer Finance Research Toolkit questions about changing practices, and While online tools are not public in what prompted individual’s shifts in their the same way as social media own behaviours. Because the interactions sites or blogs, it is important for lasted over several days, Mack could participants to understand that query the subjects on different topics that their words and pictures may be might have felt disconnected if asked back preserved on a third party to back in an interview. What began as a server.”94 study of financial communications evolved based on participant responses into a Protecting participants’ privacy can larger project around financial require somewhat different procedures in management. online diaries compared to research carried out in person, since data is Applications transferred through third party programs. For more on ethics and data, see Digital Online diaries can be a useful method for Research. targeted data collection on a range of topics not limited to financial research. They allow the participants to engage in their own time, and they provide More about the method participants with space for deeper The Portfolios of the Poor and US reflection as well as some dialog with the financial diaries websites have extensive researcher. They are particularly useful for information about the method subjects that might require closer documentation than the participant’s Microfinance in India: A Primer on the memory, as they provide a location to Financial Diaries Methodology by R. capture information. Online tools can also Kamath, S. Ramanathan and S. Rathna be used to gather a wide range of (2009, College of Agricultural Banking) feedback from participants, whether in the form of an idea, a picture, or a slogan. A financial diary study is described in Chapter 13 of Understanding Your Users: Ethical issues A Practical Guide to User Research Online financial diaries face many of the Methods by Kathy Baxter, Catherine same ethical issues as diaries collected in Courage, and Kelly Caine (2015, Morgan person. Mack told us: Kaufmann) “As with any qualitative research, it FSD Kenya provide data sets from their is crucial to have participants’ financial diary study in 2012-2013 informed consent, and to make very clear to them their ability to CGAP has also released a report and stop the study or simply opt out of data sets in February 2016 for their year- any parts that are uncomfortable. 72 Consumer Finance Research Toolkit long Smallholder financial diaries project in Mozambique, Tanzania and Pakistan95 73 Consumer Finance Research Toolkit Figure 15 Bills on microwave—put there until you have to deal with them (Photo by Alexandra Mack) 74 Consumer Finance Research Toolkit Figure 16 Donation receipts for taxes (Photo by Alexandra Mack) 75 Consumer Finance Research Toolkit cards. Charge cards have existed since the late 1800s, but they did not become Digital Research widespread until after EFTPOS (Electronic From paying with cheques to trading Funds Transfer at Point of Sale) was stocks, digital technology has transformed launched in 1981. how consumers and professionals do their banking. As David L Stearns describes in Today, this field is huge, and growing. It his book, Electronic Value Transfer: encompasses everything from how people Origins of the VISA Electronic Payment use gift cards to how they purchase items System (2011, Springer), the history of in online games. It is made even more payment systems can be traced back as huge by the vast quantities of data now far as the 19th century, when wire available about people's transactional services made it possible to send money histories, although many of those data quickly over long distances. Wire services sets are proprietary and often off limits to provided the foundation for the interbank researchers. associations that became the Visa and MasterCard networks later on. In order to gain access to such data, researchers must approach the institution In the early 1960s, the banking industry that owns the data they want and request was one of the largest users of paper, a Data Use Agreement (DUA). A DUA is a struggling under the weight of deposit legal document that allows researchers to slips and cheques.96 Coding cheques access, analyse, and publish data under magnetically was an early move to use certain conditions, usually focused on technology to reduce both paper usage protecting the confidentiality of and processing time. participants. Whether the owner is legally permitted to share their data with another Over the next few decades, as the cost of party depends upon the legislation that is computing technology waned, these early in operation in both the institution and technological developments were followed recipient’s jurisdictions. Researchers by the automation of other processes wishing to access a particular data set using On-Line Real Time (OLRT) should contact the institution that owns it computing.97 These included the use and request a DUA. computers at the point of sale to process transactions, and in the back end to keep As digital consumer finance products and records. data have proliferated, the range of research topics and the tools we use to However, digital finance did not take off in investigate them have also expanded. a large way until the World Wide Web Service providers, such as banks and became accessible in the early 1990s. payments companies, have historically Instead, ATMs were joined by an been geared towards identifying customer increasing array of credit, debit, and store needs and preferences in order to develop 76 Consumer Finance Research Toolkit better products and better delivery risks103 using digital services than mechanisms. transacting in cash. In terms of methods, service providers Social scientists working in universities have been particularly instrumental in and government bodies have applied a developing survey techniques and ways to broad range of qualitative and quantitative test user preferences. While earlier methods to digital finance research. commercial research tended to focus on These have included ethnographic product development, later studies have observations of credit card use by shifted focus to customer relationships, anthropologists, interviews and surveys by including investigating how digital devices sociologists, and lab experiments by can be harnessed to create greater economists and psychologists. They cover intimacy with customers.98 With respect to a broad range of topics, including financial consumer welfare, psychologists and literacy, digital service uptake, privacy sociologists have investigated the impact issues, security risks, how digital services of digital money on people’s spending affect consumer choices, how digital habits, especially focusing on whether it interaction shapes financial decisions, the increases indebtedness.99 Other studies uptake of digital currencies, use of digital find new ways to gather data, such as services for illicit activities, and so on. Joshua Blumenstock’s work analysing mobile data in Rwanda and Afghanistan to discern wealth distribution and migration What is it? patterns.100  Qualitative / quantitative  Multiple methods possible Non-profit organizations have tended to  Face-to-face or remote data be more focused on how digital consumer collection finance can increase the “financial inclusion” of the world’s poorest people, Digital research is not a method in its own including Information and Computing right, since technically all of the studies in Technology for Development (ICT4D) and this toolkit can potentially have “digital” the use of digital devices for microfinance aspects. Interviews, surveys, focus and money transfers.101 groups, ethnography, experiments, financial diaries, and user data analysis Field experiments have been used are just some of the methods that can be extensively as a way of testing program adapted to in digital research. effectiveness [see Experiments]. However, commentators have pointed out However, digital research deserves that unequal access to technologies102 specific attention because it changes the (the “digital divide”) means that some ways in which we can carry out these groups are unable to benefit from digital classic methods, by providing new tools services, and may even face greater and avenues for communication with 77 Consumer Finance Research Toolkit research participants and data. Moreover, Digital Research therefore often blurs the consumers are using an increasingly wide boundary between “online” and “offline” array of digital payments services, and worlds. Indeed, as the anthropologist Tom these are an important topic of study. Boellstorff comments, They include: “One thing that this kind of  Using ATMs research demonstrates is that  Multichannel banking online interaction can be “virtually”  Shopping online for insurance face-to-face, and digital  Making mobile phone payments technologies are changing what it  Use of digital currencies means to be “remote” in the first  Playing the stock market online place.”104  Creating a household budget on a computer spreadsheet An exciting outcome of the spread of digital consumer finance is that it has Digital Research can take place online, expanded the range of methods, tools, face-to-face, or both: and techniques available to researchers. The fact that most people use computing  Studies carried out online may technology (especially mobile phones) include interacting with participants means that researchers can shift away in games, conducting participant from classic ways of collecting data, observation in forums, or instead using online surveys, mobile apps, examining patterns of usage chat programs, video interviews, blogging across social media programs, and other tools.  Studies carried out in person may include moderating a focus group about mobile money use, Examples of use conducting participant observation  Chapter 7 in the Handbook of of people using online banking, or Consumer Finance Research interviewing a person about their discusses laggards in e-banking use of a financial literacy uptake105 application on a mobile phone  A research paper called Zap It to  Some studies incorporate both Me: The Short-Term Impacts of a modes, incorporating data Mobile Cash Transfer Program collected online and offline, reports on the first randomized including users’ interactions with evaluation of a cash transfer researchers online and in person, program delivered via the mobile and data collected by service phone106 providers subject to data use  A World Bank working paper, agreements, confidentiality, and South-South Migration and ethical considerations Remittances, uses World Bank 78 Consumer Finance Research Toolkit data sets to analyse and estimate usually be able to take part due to remittance flows107 geographical distance from the research  ING supplements their Europe- site. Recruiting and data collection can wide surveys with ongoing polling take place through a variety of platforms, on single questions through their including social media, electronic mailing computer portal, eZonomics108 lists, third party websites, games, and  Researchers examined the extent video calls. This reduction of geographic to which anonymized data from constraints can assist with increasing the mobile phone networks can be representativeness of a sample, and can used to predict the poverty and facilitate studies that compare wealth of individual subscribers in geographically distant groups (say, in Rwanda109 different countries or regions), which can  The Institute of Network Cultures’ then be mapped with other variables. Money Lab Reader includes various studies of “alternative Expand the range of tools currencies” such as Bitcoin110  A European Parliament report, available for research Consumer Behaviour in a Digital Digital researchers are innovating new Environment, investigates how ways to collect and analyse qualitative consumers benefit from the digital and quantitative data. There is now a wide environment and whether and how array of web-based tools for data they change their purchasing collection and analysis, including software behaviour111 that were deliberately developed to collect  Researchers from the National data (such as Revelation, a program that Bureau of Economic Research participants use to share information with studied how “digital assistants” can researchers114), platforms that were be used to remind people to save designed for other purposes but that money112 researchers use for data collection (such  CGI’s report Understanding as social media sites, discussion groups, Financial Consumers in the Digital or third party data sets), or analytical Era presents findings from their software that can be downloaded or used survey of consumers in the U.S., online. Canada, France, Germany, Sweden, and the U.K.113 Facilitate follow-up studies Locating participants after a research Strengths phase is completed can be difficult. People may change their address or phone number, and locating people in Reduces geographic constraints person is expensive and time-consuming. Internet-based studies can enable Social media and email make it easier to participation by people who would not 79 Consumer Finance Research Toolkit contact participants for follow-up research Case Study 1: Studying Bitcoin or to share a study’s results. using qualitative and Limitations quantitative methods The evolution of digital payments has Limited access to some groups generated some interesting forms of financial transactions. Digital currencies, Digital research methods can only such as Bitcoin and Dogecoin, are increase a sample’s representativeness if the target population are online or own a perhaps the most controversial of these, mobile device. While mobile phone since they are not created by government entities or banks, provide a means to ownership is increasing rapidly, rates of circumvent currency control, and can be Internet connectivity are far from uniform. used (within limits) to hide transactions This is especially true in developing from the law. countries, and sometimes also true of “wealthy” economies. Research design Bitcoin has unique characteristics that therefore needs to consider what the best make it markedly different from other method is to reach the target population. payment systems in how it operates and how it can be used. Unlike credit cards Remote data collection can transactions, Bitcoin payments are reduce data quality anonymous; no user information is Relying on remote data collection can recorded in the transaction. Bitcoin reduce the quality of the data that transactions are not cleared through researchers collect. This is because banks or any “central” location. Instead, participants tend to share more they are cleared by a network of information with people they trust, and it participants who compete with one can be difficult to develop this trust online. another for a reward for authorizing them. This can present problems for both quantitative research (e.g., convincing Bitcoin is a matter of interest to policy- people to take a survey or complete it makers, central banks, and researchers accurately) and qualitative research (e.g., who are interested in the technical conducting an interview with a stranger). aspects of this new form of money These problems are not unique to digital creation and transaction. It is also of research, as telemarketers have faced interest to exchanges and investment them for decades. Ultimately, whether or banks interested in the possibilities of its not remoteness presents a problem distributed nature for increasing depends on the type and depth of settlement speed.115 However, Bitcoin’s information that the research aims to novelty and technical complexity means collect. that researching it requires both creativity 80 Consumer Finance Research Toolkit and expert knowledge of computational In this study, the researchers exploited the mathematics.116 visibility of Bitcoin flows to cluster and positively identify Bitcoin service providers A team of researchers from the University through applying algorithmic analysis and of California, San Diego and George participating in transactions themselves. Mason University have pioneered new They used public keys to make 344 ways of understanding Bitcoin and its purchases from a range of sellers on users through combining qualitative and Bitcoin, including mining pools, wallet quantitative techniques.117 The services, bank exchanges, non-bank researchers were interested in identifying exchanges, vendors, and gambling sites. the extent to which Bitcoin lives up to its promise of pseudo-anonymity118 and to These exchanges allowed them to cluster investigate how people’s use of Bitcoin users, which in turn enabled them to has changed over time. identify major institutions in the Bitcoin marketplace and the interactions that The researchers state explicitly that they occurred between them. The researchers did not aim to identify individual users. were only able to identify those users that Rather, they used a combination of they interacted with, which were almost participation (using Bitcoin themselves) entirely third-party services like and algorithmic analysis of transactions to exchanges, not individual people. cluster users and the transactions between them. They identified service The researchers began by carrying out a providers, but not users themselves. They “re-identification attack” in which the used this information to identify the factors researchers opened accounts and made that compromised their pseudo- purchases from a variety of Bitcoin anonymity. merchants and service providers whose identities are already public (such as Mt. Method Gox and Silk Road). Since the researchers knew which public key they In contrast to most payment systems, used themselves, they were able to Bitcoin’s users are pseudo-anonymous, positively label the public key on the other but flows of value around the network are end as belonging to a particular service publicly visible. When a user makes a provider. transaction, they use a public key119 that encodes their identity so that it is not The researchers then turned to Bitcoin passed on to anyone else in the chain. forums to locate cases in which vendors However, the Bitcoin block chain120 had identified their own particular key. encodes all transactions, past and They explain that many users list their present, and it is this feature that makes addresses (or “tags”) publicly. For value flows publicly visible. example, charities list their donation addresses, and a company called LulzSec 81 Consumer Finance Research Toolkit publishes their address on their Twitter account clustering heuristics.123 This account. enabled the researchers to identify 1.9 million public keys belonging to service The researchers did not attempt to collect providers or identities. They examined all addresses available, but did amass a interactions with known Bitcoin service collection of 5,000 in total. They also providers, and were able to identify searched Bitcoin forums (such as 500,000 addresses as controlled by Mt. bitcointalk.org) to look for Bitcoin Gox, and more than 250,000 addresses addresses of defunct organizations or as controlled by Silk Road. This did not ones that are associated with major thefts. allow them to identify the individuals According to lead author Sarah making transactions per se, but it did Meiklejohn, allow them to observe interactions with particular services, such as deposits and “The thefts show how criminal withdrawals. In other words, the flow of actors are engaging with Bitcoin; Bitcoins in and out of the service was de- i.e., are they using it in a naive anonymized. way, in which our attacks could be easily applied, or are they doing Findings something more sophisticated like using mix services? Are they The main finding of the research was that, cashing directly out of the system despite widespread belief that bitcoin is using exchanges, or are they pseudo-anonymous Bitcoin users can in keeping the stolen funds in fact be identified. The authors state: bitcoins? Basically, thieves were the most motivated users we could “Even our relatively small think of in terms of wanting to experiment demonstrates that this maintain anonymity, so it seemed approach can shed considerable natural to study their behaviour for light on the structure of the Bitcoin this problem.”121 economy, how it is used, and those organizations who are party The researchers warn that these self- to it.”124 identified tags are not as reliable as the ones they collected themselves through While they did not identify real-world making transactions, so they accounts directly, their analysis de- “consequently labelled users only for anonymized users to a significant degree. addresses for which we could gain some In particular, the researchers analysed confidence through manual due certain highly publicized thefts to see if diligence.”122 they could track the bitcoins to known services. In most cases they found that After this collection phase, the this was quite straightforward. This has researchers analysed the data using major implication for law enforcement: 82 Consumer Finance Research Toolkit “…demonstrating that an agency Policy development can also benefit from with subpoena power would be the study’s innovations and insights. well placed to identify who is Research such as this advances our paying money to whom.” 125 understanding of how Bitcoin works and provides us with new methods with which This is largely because a small number of to study it. These kinds of studies could Bitcoin institutions (mostly services prove crucial to shaping public and performing currency exchange) are monetary policy to take digital currency becoming dominant, but it is also due to use into account. the public nature of Bitcoin transactions and the ability to label monetary flows to In some countries, there is currently a major institutions. Pseudo-anonymity tentative move towards incorporating therefore: Bitcoin into mainstream payments services, such as through contracting “…ultimately makes Bitcoin vendors to accept Bitcoin payments or unattractive today for high-volume installing Bitcoin ATMs.127 However, illicit use such as money governments are legislating against laundering.”126 Bitcoin use as much as they are legislating in favour of it. Understanding The researchers suggest that a follow-up Bitcoin’s potentials and pitfalls will help quantitative study could help to identify legislators decide its public value. the scale of the issue. This study also increases our Applications understanding of consumption patterns and of factors that lead to consolidation in Studies of Bitcoin and other digital payment service provision. One of the currencies have clear applications for law researchers’ findings was that fewer, but enforcement, policy development, and larger, sellers are coming to dominate the understanding changes taking place in Bitcoin market. It would appear that online trade. existing consumers share information with potential consumers regarding how to use With respect to law enforcement, this Bitcoin and which sellers to choose. study suggests that it would be easier to confirm identity and, therefore, prosecute This has the potential to drive customers illegal activity carried out using Bitcoins towards Bitcoin from other payments than people tend to believe. Users may systems and marketplaces. The wish to think twice about whether Bitcoin increasing monopolisation of the Bitcoin really does protect their identity, and law marketplace by particular companies has enforcers may develop new approaches the additional effect of decreasing based on the findings of the study. anonymity, since large sellers are more readily identifiable. In other words, the 83 Consumer Finance Research Toolkit Bitcoin marketplace is changing, and this for what purpose? Will it be used to changing market changes Bitcoin itself. increase company profit or to deny rights to certain individuals? Ethical issues People are right to be concerned: there Digital research, whether carried out in are plenty of ways in which data can person or remotely, presents a broad potentially be misused. In September array of ethical challenges. Privacy 2015, a story broke about how Facebook protection was the major ethical issue to secured a patent that would allow banks arise in this Bitcoin study. To protect to make loans based upon the credit users’ privacy, the researchers designed history of their entire social network.129 the study so that they would identify Theoretically, a bank could deny credit if known service providers, but not an individual’s friends had a bad credit individuals. Sarah Meiklejohn comments: history, even if the individual themselves was in good financial standing. While such “Our thinking was that we used fears may be unrealistic, they point to the public data, and as you say we fact that few of us really understand the identify users only by their Bitcoin implications of data sharing and its effects addresses, and (intentionally) on our financial lives. didn't identify any users who don't have a public-facing element (e.g., Health insurance is another area of individuals rather than concern, and it is a good example of how services).”128 social benefits can clash with social risks. Big data has the potential to provide By limiting their analysis to known service enormous social benefits in the area of providers, and focusing on flows of health care. If health care providers such Bitcoins rather than individuals’ as the NHS in the UK are able to access transactions, the researchers largely large amounts of data on public health, avoided issues of individual privacy and then they will be far better equipped to consent. plan services for the future.130 In many cases, however, judgements as However, this kind of data is intensely to when social benefit outweighs issues of private, and needs to be well protected. consent are subjective and problematic. An example of misuse would be if Even where consent is given, it is not insurance companies could use the always clear that individuals will datasets to positively identify individuals understand what they are agreeing can be with chronic illnesses or who need done with their data. How will it be expensive treatments and deny them analysed? Will the data be adequately coverage.131 anonymized and stored? What will the findings be used for? Will the data or results be shared with third parties, and 84 Consumer Finance Research Toolkit Researchers and professional their use for online shopping have grown associations are responding to these rapidly in Indonesia over the past two kinds of issues by developing ethical decades. While many Indonesians codes and guidelines for digital research. continue to be left out of the digital For example, the Data Science revolution, the gap is fading fast and those Association132 has produced the Data who are connected often have multiple Science Code of Professional Conduct133 devices. to help researchers think through ethical issues. Due to the low price of SIM cards, many Indonesians have multiple smart phones Various specialized books now exist on with different providers so that they can the subject, such as The Ethics of Big obtain the cheapest calls possible. By Data: Ethical Reasoning in Socio- October 2012, Indonesia had over 64 Technical Informatics134. However, the million active Facebook users, making it issues are complex and are likely to one of the top 5 nations in the world. become more so in the future. While it is crucial that researchers planning digital Indonesians use Facebook to connect studies are up-to-date on current ethical with friends, but also to buy and sell practices and standards, it is equally consumer items through a variety of important that regulators intervene to online stores and mobile apps. Boellstorff protect consumer rights. and his co-authors write that the International Data Corporation (ICD) showed that the value of internet-based trade in Indonesia reached $3.4 billion in Case Study 2: Combining online 2011, and that a MasterCard survey in and offline data collection on 2012 indicated that online shopping had increased 15% in six months. payments in Indonesia The American anthropologist Tom However, credit cards played a relatively Boellstorff and his team carried out a small role in trade: a Nielsen Online report study in 2012-13 examining how online showed that 57.4% of respondents were behaviours are affected by offline lives. using online transfer methods for They collaborated with two Indonesian payment, but only 11.5% were using research teams to learn how Indonesians credit cards, and 13.1% preferred cash on were combining social media, mobile delivery. phone use, and payments systems given that use of devices and the Internet was Boellstorff and his teams combined face- increasing rapidly in Indonesia.135 to-face interviews with analysis of the online purchasing and payments The researchers chose this focus because environments that people were using to device ownership, Internet access, and gain insights into how people made 85 Consumer Finance Research Toolkit purchasing and payment decisions. They dimensions of device use. For example, to focused on both the technology that identify time discrepancies between when people used and the social relations that people began to use devices and when shaped people’s actions. they began actually shopping online, interviewees were asked the following Method questions: The research took place in Surabaya  When did you begin using (Java) and Makassar (Sulawesi). The gadgets? researchers used qualitative methods  When did you begin using the including participant observation, Internet? individual semi-structured interviews, and  When did you begin making online focus groups. They supplemented these transactions? with the analysis of websites, mobile apps, and advertisements. To investigate users’ behaviours, including how people decided to become In Makassar, the researchers interviewed resellers and any problems with addiction 54 respondents and conducted two focus to online shopping, the researchers groups with ten participants in each group. asked: In Surabaya they interviewed 52 respondents and conducted four focus  What motivated you to make groups. All data were collected in transactions online? Indonesian or in local languages and then translated into English by members of the In response to other studies that argue research teams. that online shopping almost always happens in combination with other In each location, the researchers sought activities (such as socializing or working), to recruit a diversity of people from they asked: different social groups, including those that they thought would give a range of  Are there particular times when perspectives on mobile social media and you cannot shop online? mobile payments (such as university  What are you doing when online students and “housewives”). The best- shopping? represented group were heterosexual women, who are active in the world of People’s reasons for using online online shopping in Indonesia. transaction services tend to differ according to context. For example, while The research teams held an initial shopping may be popular in one country meeting in September 2012, before or among a particular social group, other research began to decide on key interview services such as remittances may be questions to be included in all of the more widely used elsewhere. To find out studies. The questions covered multiple 86 Consumer Finance Research Toolkit what people were using digital financial “The key thing to remember is that services for, without biasing their answers, the phenomenon being studied is the interviewers asked: already “synthesizing” the online and offline before we ever got  What kinds of transactions do you there. So synthesizing the online make online (shopping, sending and offline in this case (and in money, paying bills, etc.)? practically every case of such digital research nowadays) is not To find out how people saw themselves an artificial imposition. Ideally it using services in the future, they asked: should reflect the specific ways that in the case at hand, the online  Have you ever thought about and offline are shaping each stopping shopping online? other.”136  Have you ever thought about reselling things that you purchase In other words, the “online” and “offline” online? research should not be treated as separate data sets that are collected Finally, in order to find out how people independently of one another and then actually pay for goods, why they choose brought together for analysis. Rather, the one payment mechanism over another, online and offline data are inextricably and whether they set aside special funds linked. For example, interviewees might for online shopping, they asked: tell stories in which events take place in the home and simultaneously through  What funds do you use for social media. Similarly, an observation of shopping online? How do you or a person using an ATM must take into your friends pay for online account the “real” world, because factors shopping if you don’t have money such as time constraints and safety at hand? considerations will affect how they use that digital device. Moreover, people use For the online part of data collection, the multiple devices to achieve particular researchers analysed websites, mobile goals. Treating each digital interaction as apps, and advertisements. They collected distinct does not reflect how people use and assessed this online data by going devices in the course of their everyday directly to the websites in question. lives. However, they note that it is possible to do interviews and focus groups online as Findings well. Data analysis was synthesized as part of The research was designed to study the the overall research process. Boellstorff intersection of mobile phones, social comments, media, and payments. As a result, some of their findings address digital consumer 87 Consumer Finance Research Toolkit finance directly, while other findings Buyers often knew sellers personally or at address it indirectly through describing the least lived in same city, meaning that they context in which transactions take place. could choose sellers based on personal knowledge or recommendations. This also Boellstorff and his team found that online made it easier to complain if there was a shopping in Indonesia is made possible by problem. Buyers often did not have to pay the prevalence of devices. All of the shipping because items would be hand- interviewees owned more than one delivered. Their buying practices therefore device, often a laptop, a BlackBerry, and often mimicked physical shopping. at least one other smartphone. Facebook was often the first pathway to Due to the low cost of SIM cards and the online shopping because respondents advantages of using multiple providers, would see sponsored advertisements, many respondents had multiple information posted by friends on their own smartphones or SIMs; for example, one to timelines, or comments customers had keep in touch with a romantic partner and posted on the Facebook pages of sellers. one for other friends, or one for personal Alternatively, people would be introduced use and one for business use. to online shopping through the BlackBerry store app. At the time, BlackBerry was still the most commonly used handset, and all of their Once interviewees had begun to shop respondents had a BlackBerry and often online they identified numerous other kinds of smartphones as well. One advantages. The five primary reasons that interviewee had 5 smartphones, each with respondents gave for wanting to shop a different provider. online were 1) it is easy; 2) interesting things are sold in online stores; 3) it The reasons why people began to shop avoids the hassle of going to a physical online rather than in physical retail stores store; 4) it is often cheaper; and 5) some varied. The researchers found that items are hard to find in physical stores. respondents became interested in online shopping after seeing items their friends We often think of online shopping as had purchased. The researchers explain: something done by individuals or households, but the researchers found “This reflects a broader pattern in that groups of friends would also make which friends and acquaintances collective purchases. For example, they play an influential role in online discovered that groups of university shopping practices not just as students would purchase food (such as recommenders, but increasingly as snacks) in bulk to receive lower prices. customers and sellers.”137 The amount of money interviewees spent online monthly varied from less than $1 to 88 Consumer Finance Research Toolkit around $50. This sum reflects people’s were paying for online shopping by disposable incomes, but also their feelings making a transfer at an ATM or a bank about online shopping. Concerns about counter. The main banks they used were losing money tended to limit how much BCA and Mandari; only 5 respondents money people were willing to spend used BNI. online. Why use an ATM rather than making an As well as concerns about fraud, people online transfer? Some interviewees said worried that the items they bought would that they were worried about credit card not match the online photographs or fraud, but the main reason cited was to description and that they would be avoid bank fees. If the buyer and seller disappointed with their purchases. To used the same bank, the transfer could be counter this, one woman set a 500,000 made for free. rupiah ($50.50) limit on purchases. Beyond this limit she felt uncomfortable In fact, some sellers had accounts at and preferred to make the purchase in a multiple banks to ensure that their buyers physical store. could transfer them money with no extra charge. Some buyers reported that if they Some people also earmarked funds to be didn’t have the same bank account as the spent on online shopping. One seller, they would ask a friend or relative interviewee, a man called Eska, gave his to complete the transaction for them. Note monthly salary to his wife to manage (a the pattern: we see a cost-savings common practice). He had his own, behaviour taking place at both the ATM separate bank account for his own and the SIM card level, since, as noted, expenditures (“men’s money”) that was many Indonesians have multiple SIMs in funded primarily by workplace bonuses order to save money by switching SIM rather than his salary. cards depending on whom they are calling. When he shopped online he always used this account, even when his wife asked People sometimes used other people’s him to buy something for her. In practice, credit cards, but that could create then, it wasn’t just his “men’s money,” it secondary problems. For example, one was also the household’s online shopping gay man reported that he preferred account. shopping for makeup online because he felt safer than when visiting a physical In terms of making payments, the store. In order to pay for his purchases he researchers discovered that only 6 used his mother’s credit card. However, respondents used Internet banking he was concerned about his privacy here, services to make a payment, and only 5 because his mother could then see on her respondents used a credit card. Instead, statements what he had been buying. the vast majority of their respondents 89 Consumer Finance Research Toolkit Online BlackBerry shops were sometimes overcome some of these costs by used to both make and complete smoothing out the transaction process. transactions. Interviewees stated that they preferred these shops because sellers However, as other brands are displacing would be identifiable by their BlackBerry BlackBerry handsets, customers will PIN. They also tended to be more familiar require other channels to make payments. with the sellers and had friends in the Knowing where the pain points lie for system that could provide customers, and which device to use when recommendations. Moreover, BlackBerry making a transaction, can assist in the Money allowed peer-to-peer cash identification of appropriate payment transfers, which interviewees perceived as channels. We also see cost-saving safer than using a credit card. behaviours that cross over between devices and services—in this case, ATMs Applications and mobile phone SIMs. Risk is another important area for product The team’s combination of online and design and marketing. This study shows offline data collection has many potential that consumers assume they are taking a applications in product design and certain level of risk when shopping online. marketing. For example, there could be They attempt to offset risk by using known opportunities to develop consumer finance providers and channels for their products that reduce the costs of making purchases, or through seeking advice transactions. from people they trust. Consumer decisions are influenced by Building social proof into product design, various factors: they may identify with the such as through peer-to-peer transfers or brands of their devices, their mobile permitting product recommendations from phone carriers, or be swayed by the friends could be leveraged to increase preferences of their friends and family. trust and encourage the use of particular These identifications influence consumer payment channels. decisions in different ways, and at different times. The combination of in- person ethnography and examination of Ethical issues digital sources used in this study is useful Most of the ethical issues raised in this in finding out how people use products study are the same as in any other face- within real-life contexts. to-face study. These include: Many of the study’s participants reported  Giving potential participants having to put in a significant level of effort sufficient information about the to complete payments, often maintaining study so that they can make an multiple bank accounts to lower costs. informed decision about whether The BlackBerry store presents a way to they would like to take part 90 Consumer Finance Research Toolkit  Ensuring, as much as possible, Safety issues may also be a consideration that participants are not placed in when studying money use. Following physical or emotional harm during participants as they go about their daily the study tasks can yield valuable insights into their  Anonymizing data so that use of consumer products, but a foreign participants are not identifiable, researcher accompanying a participant to unless they have given specific an ATM may attract unwanted attention. permission These are not necessarily problems that were raised in this particular study, but However, as with the Bitcoin study they need to be considered at the outset described above, the “digital” nature of the and built into project design. research raises extra points of consideration.138 Boellstorff reports that one big issue he has encountered is that some researchers think that because More about the method something is online it is not “real” and so Internet Communication and Qualitative you don’t have to protect people’s Research: A Handbook for Researching identities. Online by Chris Mann and Fiona Stewart (2000, SAGE) For example, data gained from a public forum may be technically traceable, but if Advancing Digital Humanities: Research, it is reproduced without permission, Methods, Theories by Katherine Bode and researchers have a responsibility to Paul Longley Arthur (2014, Palgrave generalize the data into findings that are MacMillan) not traceable. Ethnography for the Internet: Embedded, Another important issue in digital research Embodied and Everyday by Christine Hine is that social media tends to make people (2015, Bloomsbury) more visible, and so it can make anonymization and consent more difficult. Ethnography and Virtual Worlds: A Say an interviewee is demonstrating how Handbook of Method by Tom Boellstorff, they make a purchase in a BlackBerry Bonnie Nardi, Celia Pearce and T.L. store. When they show the researcher Taylor (2012, Princeton University Press) their own information, they are also likely to expose information about the people Readings in Virtual Research Ethics: they are transacting with. These third Issues and Controversies edited by parties have not given consent to take Elizabeth A. Buchanan (2004, IGI Global) part in the research, and extra effort must be taken to discuss and record this data in How Would You Like to Pay? How the most general terms. Technology is Changing the Future of 91 Consumer Finance Research Toolkit Money by Bill Maurer (2015, Duke University Press) 92 Consumer Finance Research Toolkit Figure 17 Bitcoin ATM, Irvine (Photo by Ursula Dalinghaus) 93 Consumer Finance Research Toolkit Figure 18 Screenshot of Bitcoin Wallet (Photo by Ursula Dalinghaus) 94 Consumer Finance Research Toolkit Figure 19 Photo by Tom Boellstorff Figure 20 Photo by Tom Boellstorff 95 Consumer Finance Research Toolkit success of a particular program or product. Experiments The use of experiments in consumer What is it? finance research grows out of behavioural psychology and experimental  Quantitative economics.139 Consumer finance  Conduct experiments under experiments reflect the concerns of both controlled conditions those disciplines, especially regarding  Data is collected in a lab, or in the how people make choices and whether field financial behaviours fit with standard economic theory. Experiments involve testing hypothesis under controlled conditions through the An important precursor to consumer manipulation of key variables. They can finance experiments was the development take the form of lab experiments, which of modern game theory, the study of are carried out under controlled strategic decision-making, by John von conditions, or field experiments, which Neumann and Oskar Morgenstern in the take place in a context that is largely 1930s. Game theory lent itself more natural. Field-based experiments (RCTs) readily to experimentation than theoretical follow criteria for randomization of the microeconomics or macroeconomics study population and testing of specific because it was concerned with how variables to measure impact of a product people make choices under specific or intervention. Field experiments scenarios. generally involve running interventions or treatments on people as they live their Experimental researchers try to model the daily lives. It is more difficult to control different ways in which people make experimental conditions in field decisions. These models can help with experiments, but they can provide a more things like structuring financial literacy realistic picture of behaviour than lab programs, identifying social sectors that experiments. might be at risk of getting into debt, and providing people with better information so that they can make better choices. Lab experiments Experimental methods, in particular Randomized Control Trials (RCTs), are In consumer finance research, lab also the cornerstone of evaluations, which experiments are often concerned with provide a way to rigorously assess the testing the psychological components of 96 Consumer Finance Research Toolkit consumer finance, especially risk-taking experiments are carried out under behaviour and how people make choices. everyday circumstances and test the impact of an intervention, product, or In lab experiments, the researcher(s) first policy. devise a hypothesis; then they design an experiment to test it. They identify at least In the field of socioeconomic one independent variable (an event, e.g., development, a particular kind of field opening a bank account) and one experiment is used widely: the dependent variable (measurable effects, “randomized controlled trial” (RCT).140 e.g. saving more money). RCTs are considered to be the “gold standard”141 of hypothesis and impact Lab experiments take place in a location testing.142 Derived from biomedical and that is under the control of the epidemiological clinical trial methods, researchers. This is normally a laboratory RCT’s evaluate and measure impact by or room in the researcher’s institution or comparing the effect of an intervention in workplace, but lab experiments can be a treatment or target group(s) against a carried out anywhere the researchers can control group that does not receive the gain a reasonable degree of control over intervention. the environment. They are valuable because they can For example, a café or park would reliably test the effectiveness of consumer probably not be suitable for a lab finance initiatives, especially in experiment, but an empty room in a microfinance or the introduction of new school, house, or public building might be financial products by banks and other sufficient. This is important because it financial institutions. As noted in the IPA means that researchers are able to travel RCT toolkit, experimentation can be built to the people they would like to study in into the design of products and order to undertake lab experiments, just services.143 Some financial RCTs such as as with field experiments and natural experiments with cash grants or financial experiments. Experimental research is literacy training may include physical and therefore not always limited to recruiting psychological health and well-being participants whom the researcher might components. need to pay to travel to the lab. A standard method to carry out an RCT is Field experiments to plan an intervention in a particular place, such as lending money to women Field experiments involve real-life testing in a particular town, and set up a control of a hypothesis or intervention. Like lab group in another town. Careful experiments, field experiments are about randomization of the study population to testing and measuring behaviour. But be studied is crucial to the design and is unlike lab experiments, which take place “the core RCT methodology.”144 This is under controlled conditions, field 97 Consumer Finance Research Toolkit important for claims that can be made A variation on the field experiment is the about measurable outcomes due to the natural experiment.149 These resemble effects of a product or intervention and not field experiments, but there is no other factors or individual/group intervention. Researchers simply measure characteristics. Field-based experiments the things that people are already doing. may use other experimental methods for One such study in India used loan comparison than randomization, applying repayment data to investigate the optimal a variety of statistical techniques to control structure of a microfinance loan.150 They for differences, but these often involve compared people who had repaid making assumptions that are more difficult individual loans in full (the “treatment” to test than if randomization is used in group) with people who had an ongoing advance.145 Determining the appropriate individual liability loan, but who would sample size in order to observe a eventually convert to group liability due to particular effect is also crucial to the a change in policy in the lending institution technical design. The larger the sample (the “control” group). By watching how size, the more visible a small effect will customers changed their borrowing be.146 A simplified explanation of the basic practices as their loan type transitioned, setup is as follows.147 At the beginning of they were able to infer which loan the evaluation, both groups will be structure worked the best. [see also surveyed. The test group will be given the appendix, blog post 7] intervention, but the control group will not. At the end of the intervention, both groups are surveyed again. The results are Examples of use analysed to show the effects of receiving  For a range of experiments on or not receiving the intervention. choice, risk, reactions to pricing structures, and other aspects of For example, a field experiment in Malawi, economic behaviour, visit the conducted by Xavier Giné and Dean Centre for Economic Learning and Yang, tested whether the provision of Social Evolution (ELSE)151 insurance induces farmers to take out  IMTFI fellows Deepti Kc and loans.148 The researchers selected a Mudita Tiwari carried out an sample of 800 farmers and offered them experiment to test ways to improve credit to buy high-yielding seed. Half of the financial capability of women in the farmers were required to purchase India152 (see Case study 3 below) insurance to receive this credit. They  Johannes Haushofer and Jeremy found that farmers were less likely to take Shapiro used a randomized up credit if they were offered insurance controlled trial to measure the with their loan. These kinds of response of poor rural households experiments are valuable because they in rural Kenya to large temporary involve people making real-life decisions. income changes in the form of unconditional cash grants.153 98 Consumer Finance Research Toolkit  A study implemented by Strengths Innovations for Poverty Action shows how a commitment savings product led to increased savings Permit control of variables and more decision making power Because experiments take place under in the household for women154  A natural experiment carried out in controlled conditions, and with lab experiments normally in the researcher’s India investigates the optimal institution or place of work, it is usually structure of a microfinance loan155  Economists undertook a natural possible to limit the number of variables that impact the experiment. Again, this is experiment in Singapore using a particularly true for lab experiments. It can panel dataset of consumer be difficult to control for variables in field financial transactions to study how experiments, since they take place in consumers responded to an “natural” settings. unanticipated income shock156  A field experimental in Malawi, The control of variables means it is conducted by Xavier Giné and possible to make accurate measurements Dean Yang, tests whether the and identify within certain statistical provision of insurance induces parameters “cause and effect” farmers to take out loans157  In Pakistan, another field relationships (e.g., that opening a bank account leads to greater savings). These experiment measured demand for kinds of observations are generally not microinsurance after severe possible with non-experimental methods flooding in 2010158  An article called The Social because the effect (in this case, increased savings) could be caused by any number Dilemma of Microinsurance: Free- of factors. riding in a Framed Field Experiment explores incentives for However, it should be noted that this free riders in joint liability health ability is dependent upon solid research insurance in Tanzania159  A large natural field experiment to design and execution. It is often difficult to isolate variables and limit external identify the effects of formal influences in a study. savings on inter-household transfers and the safety nets of the Experiments are replicable poor in villages in Malawi160  The Institute for the Study of Experiments are generally more replicable Labour published an article on how than non-experimental research because two different insurance products they operate under controlled conditions, and a secret saving device impact use limited variables, and allocate solidarity among rural villagers in participants randomly to test groups and the Philippines161 control groups. Again, this is more true for 99 Consumer Finance Research Toolkit lab experiments than it is for field RCT field experiments experiments, since real-life conditions can demonstrate impact change rapidly and make repeat experiments impossible. An advantage of randomized controlled trials (RCTs) is that they can be used to This means that is possible for other demonstrate whether (or to what extent) researchers to confirm or challenge a an intervention does or does not have the study’s results. It also means that studies hoped-for impact. This is why they are can be readily compared across different used extensively for the purpose of groups. evaluating programs carried out with people in the area of socio-economic These features contribute to the development.163 development of general models and theories (such as the effects of In fact, David Roodman, in his book Due asymmetric information on decision- Diligence: An Impertinent Inquiry Into making), the behaviour of specific Microfinance,164 argues that randomized populations (such as stock market use by trials are the best way to test whether elderly investors), and dynamic microfinance programs work. Monitoring inconsistency (how people’s preferences and evaluation is a large and important change over time). field with an entire methodology of its own, and people who are interested in Nevertheless, researchers need to be developing their skills in this area have a aware that, while it is relatively wide range of materials and courses to straightforward to replicate experiments, choose from. IPA’s toolkit also offers the results can differ significantly helpful guidelines for identifying and depending upon the variables tested and developing partnerships for evaluating sample choices. This can limit the financial products in the US based on its “external validity” of an experiment—how expertise in RCT research in developing generalizable is it to the results of other nations.165 studies, to other situations and other people? For instance, evaluation may be Limitations limited to places or regions where there is “visible variation,” that is, where a new program is expanding into, and not within Design flaws can invalidate a particular region. This means that you experiments can usually only evaluate a kind of The effectiveness of experiments is marginal impact, not infra-marginal dependent on a solid research design. impacts.162 It is also important to conduct Confounding effects or confounding careful pilot studies when setting up an variables are variables that the RCT. Testing results through multiple experimenter failed to control and which studies can strengthen generalizability. compromise the validity of the experiment. 100 Consumer Finance Research Toolkit This is true for all kinds of experiments, experiments, even though they take place whether they take place in a lab or in the in real-life settings. It is not the case for field. natural experiments, since they are by definition the study of real-life behaviours. While design flaws are a problem in all research, qualitative methods tend to be One critical reason why experimental more forgiving when problems arise. For behaviour may not reflect real-life example, if analysis of a set of interviews behaviour is that experimental subjects shows that a crucial question has been may be conscious that they are being omitted, researchers may be able to make watched, and this may make them more further inquiries to fill the gap in their likely to follow moral norms or make more knowledge by asking participants directly rational decisions than in real life.168 This for clarification. This is not generally is known as the “Hawthorne Effect.”169 possible with experiments; for reasons of validity, the entire experiment may need to The Hawthorne Effect changes depending be run again if an error or omission upon whether your research participants occurs. are interacting with each other or if they are not. Say that in a lab-based study However, in rigorous lab as well as RCT each of your participants completes an experimental studies, researchers check experiment alone, entering responses for lack of balance, spillover effects166, or anonymously on a computer. Their other unexpected issues in the answers may be affected by what they randomization process over the duration think the researchers are looking for, but of the study. [see appendix, blog post 6] they are unlikely to be directly affected by Researchers adjust or re-randomize peers, since their fellow participants are accordingly, where possible. Other not observing their choices. design “threats” to the internal validity of an RCT study include noncompliance in In contrast, lab-based experiments that the treatment group(s), spillover effects require participants to interact with each that impact the control group, or other give rise to a number of participants who drop out of the study, methodological problems that have been which affects project outcomes and can closely observed and are well lead to “missing data” in the final phase of understood.170 In cases where participants evaluation.167 are anonymously interacting with other participants, it is often observed that Results may not reflect real life decisions can be anti-social: people will often act in their own benefit, not for the behaviours benefit of the group. The artificial nature of experiments can produce results that are unlikely to occur In cases where interactions between in real life. This can even be true of field participants are not anonymous, the same 101 Consumer Finance Research Toolkit experiment can produce very different $1 usually means more to a student than results, as people are often more likely to to someone on a stable income, and so cooperate and behave generously when experiments can be run for less money. they have to interact directly. Overall, However, students sometimes access the experimenters find that repeated experiment several times to earn more interactions cause subjects to eventually money, even though this is against the start cooperating, and that their rules. This undermines the fundamental cooperation increases mutual benefit. This assumption of the experiment and is also something we observe in the real reduces its representativeness. world. Another common method of recruiting The degree to which lab-based subjects is to give study credit to college experiments do or do not reflect real life students who take part. While this is a behaviours can be mitigated to some great way to gain the required number of extent by a careful design that considers participants, it does not solve the problem these kinds of influences. of representativeness. This is particularly the case for experiments carried out with so-called WEIRD subjects (Western, It can be difficult to form a Educated, and from Industrialized, Rich, and Democratic countries).171 However, representative sample using a control group can eliminate this Sometimes it is difficult to recruit bias. If participants are distributed participants that are representative of a randomly between each group, then sample of the population under scrutiny. differences between results from each This is equally true for lab experiments group will have nothing to do with them and field experiments. being students. This is partly because the resources While moving a lab to the field can help required to run lab experiments are often offset some of these concerns, anyone limited, but also because it can be difficult wishing to perform a lab experiment is to persuade people to take part in such advised to read up on the many ways that studies, especially when they are required biases can be introduced in real life to travel to the site of the experiment. settings. One way of lessening recruitment issues is to offer monetary awards, but these can Limitations specific to interfere with representativeness. Some behavioural economics experiments use Randomized Control Trials financial incentives to attract participants, such as through playing games with real A number of issues specific to RCTs money. Students are often used because should be given particular attention when 102 Consumer Finance Research Toolkit RCT experimental design is part of the regarding what information and services research methodology. RCTs can be can be withheld or offered to some, but costly or require multiple studies in not all members/clients of a financial different places in order to provide institution, as well as what incentives can persuasive results for service providers or be offered.173 policy circles. Practitioners of this method are often asked about the ethics of RCTs may also be supplemented by withholding information, resources, or qualitative research, including services from a control group. Whereas ethnography. Researchers can resources are often scarce with our (individually or in partnerships) without the intervention and it is generally incorporate participatory fieldwork, not possible to give everyone access, for multiple visits, and other forms of those who do receive the treatment there interaction into the study design. Because is also an “intervention fatigue.” This is ethnography brings researchers into the especially true where the same study daily spaces and concerns of research populations are targeted for multiple and subjects, researchers’ assumptions about frequent interventions because of where how things are working or not—and why— they are located or due to the services may be challenged and their conceptual being rolled out. Other kinds of tensions categories unsettled. These kinds of data may be generated in treatment groups are often missed in quantitative data where interventions can be perceived as collection and often yield crucial insights yet more handouts or interference from for understanding issues of reception, powerful outsiders. And what comes after uptake, and changing interrelationships a study has finished? Long-term follow-up that are relevant to the intervention that is is not always possible, limiting study being tested in an RCT. participants’ ability to give feedback and researchers’ opportunities to analyse impact over a longer period of time. Case Study 1: Understanding There are a number of ways to address Risk Preferences and Time these limitations within and outside the RCT research design. For instance, many Preferences RCTs phase in a particular product or Risk and time are important topics in service as part of the experimental consumer finance. Whether people are design.172 The group that receives the risk-takers, risk-averse, or loss-averse intervention at a later time serves as the impacts all kinds of decisions, including control group for the treatment group in taking loans, buying insurance, and the first phase of the study. In the case of making investments. Consideration of financial products and services such as in time frames is just as important: financial the US, there are legal, regulatory and planning involves thinking ahead, and how consumer protection requirements 103 Consumer Finance Research Toolkit people perceive time is crucial for making However, the economists James Andreoni good plans. and Charles Sprenger ran a series of lab Less discussed is the fact that risk and experiments that contests this finding. time are intertwined. For researchers They present their results in an article wanting to understand financial behaviour, “Risk Preferences are Not Time it is crucial to be aware of how time can Preferences” (2012).174 In this article they affect people’s judgement of risk. give an alternate explanation for why people might choose to take a smaller Most studies of financial decision making benefit now rather than a larger benefit in over time, including prospect theory, claim the near future. that people are so biased towards the present that they will make decisions that Method are counter-productive in the longer term. Andreoni and Sprenger ran experiments For example, say you are offered a choice with 80 undergraduate students at the between $100 now or $120 one week University of California, San Diego. from now. The latter choice is generally Students participated in four experiments, the most rational, but many people which took an hour each. The researchers choose to take the money now rather than used a method called “convex time wait. Why might this be the case? One budgets” (CTBs) in which participants reason why people make this decision allocated a budget of tokens towards might be that their assessment of their receiving an early payment (in 7 days’ current needs and desires outweighs their time) and a later payment (in 28 or 56 assessment of their future needs and days’ time). desires. This is called “present bias.” The researchers varied the probability of The most famous experiment of this kind payments being delivered and the interest was the Stanford marshmallow paid on later payments. Participants had experiment into delayed gratification in the to choose between money sooner (to be late 1960s and early 1970s. In this delivered to them in a week) and money experiment, researchers gave children later (to be delivered to them in either 28 marshmallows and told them that they or 56 days). Students also received a could eat it straight away, but if they basic participation payment, of which half waited for 15 minutes they would receive was delivered in the first payment and half two marshmallows. The researchers then in the second payment. Andreoni and left the room. A minority of participants ate Sprenger explain, their marshmallow immediately. In follow- up studies, the researchers found that “For all payments involving children who waited had better life uncertainty, a ten-sided die was outcomes. rolled immediately after all decisions were made to determine whether the payments would be 104 Consumer Finance Research Toolkit sent. Hence, p1 and p2 were risk, distributing the payment between the immediately known, independent, two delivery times: and subjects were told that “…if a sooner reward will be different random numbers would realized 100 percent of the time determine their sooner and later and a later reward will be realized payments.”175 80 percent of the time, then intertemporal allocations should be An important part of research design was identical to when these to minimize uncertainty based on probabilities are 50 percent and 40 confounding variables, such as whether a percent, respectively.” payment would accidentally go missing. To achieve this, experimental participants However, they found that their participants were chosen from among students living only behaved in this way under conditions on campus who had 24-hour access to of uncertainty. For example, when two locked personal mailboxes in their dorms. options have the same degree of uncertainty (for example, a 50% chance of The researchers took care to explain the Payment 1 being delivered and a 50% process of payment delivery thoroughly so chance of Payment 2 being delivered), that students would be confident that they then participants would allocate their would receive their payments. In fact, a payment between these two events. companion survey showed that students had 100% confidence that their payments When conditions were certain, participants would be delivered. So, it is reasonable to behaved differently. In fact, “85 percent of assume that confounding variables were subjects violate common ratio predictions limited, and the decisions that the and do so in more than 80 percent of students made during the experiment opportunities.” There was little were not affected by extraneous consistency in how participants allocated influences. the delivery of money under conditions of certainty, and they did not seem to prefer A major advantage of this design is that it sooner payments. allowed the researchers to test whether students made decisions based on risk or Instead, it appears that people were time: that is, were they failing to delay responding to changing levels of risk. gratification, or were they taking risk into Andreoni and Sprenger point out that, for account? most of us, the present is certain because it is already happening, while the future is Findings risky because it is difficult to say what will happen. They explain, Andreoni and Sprenger note that, according to discounted expected utility “Allais (1953, p. 530) argued that (DEU) models, participants should when two options are far from allocate their money according to relative 105 Consumer Finance Research Toolkit certain, individuals act effectively Experiments like these have also been as expected utility maximizers, valuable in testing the validity of economic while when one option is certain models. They have clear real-life and another is uncertain a implications in consumer finance, such as “disproportionate preference” for for understanding how people will be certainty prevails. This intuition affected by time considerations, the risk of may help to explain the frequent receiving or not receiving a payment, and experimental finding of present- the effect of interest rates. biased preferences when using monetary rewards (Frederick, Ethical issues Loewenstein, and O’Donoghue 2002). That is, perhaps certainty, Many of the ethical issues that arise in lab not intrinsic temptation, may be experiments are the same as in all leading present payments to be research involving human subjects. disproportionately preferred.”176 Psychological harm is the most common type of potential harm in non-medical Hence people may not be biased towards research, that is, creating situations that the present at all, but instead risk averse. lead to embarrassment or anxiety. Researchers can help to reduce harm by providing sufficient details of the study Applications and giving participants the option to skip This case study has valuable implications questions they are not comfortable with or for experimental design. In experiments, if to leave the study altogether. you don’t control for the fact that the future looks more risky than the present, then Issues can also arise from the people will make decisions that appear to objectification of research participants, be “present-biased” when they are that is, treating them as merely research actually making a risk-averse decision. material rather than as human beings. Maintenance of privacy and confidentiality For example, say you are one of the is another issue, and steps need to be children in the famous Stanford taken to protect privacy during all phases marshmallow experiment. How do you of the research, including data collection, know that the researchers will really give analysis, data storage, and the publication you another marshmallow? If you invest of the results. money, how do you know that it will pay off? The future is uncertain and anything Experiments also involve some could happen: a financial crisis may wipe considerations that are generally not out your investment or the marshmallow present in other kinds of research.177 supply. People who are risk-averse may Experiments with human subjects depend decide that it is better to take an upon isolating a variable that is tested immediate reward than to depend upon a under laboratory conditions. Participants bigger reward in the future. are often not told exactly what the study is 106 Consumer Finance Research Toolkit trying to test, because they may change as the researchers are able to control the their behaviour to fit in with the experimental conditions to a satisfactory experiment. This lack of information degree. makes it difficult for participants to give informed consent. In the mid-2000s, a group of researchers working for the Financial Access Initiative Deception can also harm experimental and Innovations for Poverty Action carried research in a more general sense, since it out ten microfinance games in an can erode trust and make people unwilling experimental economics laboratory in to volunteer for the study. Moreover, some urban Peru.178 The resulting article by researchers claim that it can alter Xavier Giné, Pamela Jakiela, Dean participants’ behaviour in future studies, Karlan, and Jonathan Morduch describes thus compromising results for other the experimental process and results. researchers. Many microfinance agencies only engage Another issue with experiments on human in group-lending because it significantly subjects is that they often depend upon lowers the risk of making loans within low- students to participate. Apart from the fact income communities. In group-lending, that students are not usually a individual borrowers guarantee each representative sample of the population at other’s loans. Rates of repayment are large, there are also issues of coercion to generally high, at around 95%. consider. In cases where students are required to participate in experiments as And yet the fact that group lending part of their course evaluation, or are regularly out-performs individual lending is offered extra credit, their choice to puzzling because group lending comes participate or not has essentially been with problems of its own. For example, removed. Moreover, paying students to group lending is vulnerable to free riding participate can be problematic, given that because it is potentially easier for an poverty can drive people to accept options individual to default against their group that they would otherwise reject. (who will cover for them) than against a bank. Whereas an individual who defaults runs the risk that the bank will not lend to them again, with group liability it is easier Case Study 2: Microfinance to maintain access to loans. games: Group lending versus Does joint liability really encourage such individual lending in Peru “moral hazards”? To find out, the Lab experiments do not have to be carried researchers set up a series of out in the headquarters of a company or experiments that explored the impact of organization. They can be carried out in individual and group lending mechanisms settings that resemble the “field,” so long on investment decisions. The purpose of 107 Consumer Finance Research Toolkit these experiments was to show how In each round of the games, experimental liability affects whether people made risky subjects explained the rules in Spanish. or safe investments. They were given "loans" of 100 points and were asked to invest their points into one Method of two projects: either a safe project with a certain return of 200 points, or a risky The team set up a makeshift experimental project that paid 600 points with a economics lab in an empty room in a probability of one half. They were given marketplace in urban Lima, Peru. They game sheets on which to mark their chose the location to attract participants choices. If a borrower’s project succeeded whose profiles resembled those of they would have to repay their loan, but if microfinance customers. their project failed they would not be able to repay. At the end of each session, The researchers recruited participants participants were paid a fee for showing using two methods: employing delegates up and another fee for every treatment from the local association of micro- they had taken part in. entrepreneurs to invite vendors to specific game sessions, and allowing participants The researchers also conducted a census to bring friends to subsequent of the vendors in the market, which experimental sessions. allowed them to compare their experimental group with the general Over seven months, the team ran ten market demographic and work out experimental games an average of 29 whether they were representative of this times each. The games consisted of broader population. multiple rounds of borrowing and repayment. The researchers observe that playing a sequence of games with the Findings same individuals allowed them to control The researchers found that subjects were for individual’s risk preferences and more likely to make riskier investments assess the impact of each lending when they had joint liability because, in mechanism on risk-taking and loan the event that their investment failed, the repayment. other members of the group would look after their debt. They observe: The researchers changed the variables in each of the ten games in order to assess “Risk-taking broadly conforms to the effects of different circumstances that theoretical predictions, with mimic the actual conditions of dynamic incentives strongly microfinance programs. These included reducing risk- taking even without individual versus joint liability, dynamic group-based mechanisms. Group incentives or no incentives, and the lending increases risk-taking, amount that players were allowed to especially for risk-averse communicate or to observe each other. borrowers, but this is moderated 108 Consumer Finance Research Toolkit when borrowers form their own The researchers also point out that the groups. Group contracts benefit question of whether contract structure borrowers by creating implicit inhibits risk-taking is important for policy insurance against investment development. Evidence suggests that losses, but the costs are borne by most microfinance loans have a limited other borrowers, especially the effect on the growth of businesses. If this most risk averse.”179 is the case, perhaps contract structure could be altered in such a way that it However, cutting off defaulting borrowers encourages a level of risk-taking that is from future loans greatly reduced risk- suitable for setting up a successful taking behaviour. business. Based on their observations and the work Field-based lab experiments can be of other researchers, the authors suggest fruitfully combined with other methods to that joint liability is not always necessary broaden their findings and applicability. to maintain high repayment rates. They Whereas the lab experiments in Peru state: demonstrate the effects of collective action on individual behaviour, “Given large enough incentives to ethnographic studies could describe the avoid default, borrowers will mechanisms by which collectives operate. choose safe projects and repay their loans.”180 For example, anthropologist David Stoll’s research on debt in a Mayan town in Hence it is not possible to conclude that Guatemala unravelled the puzzle of how joint liability is better than individual an entire town became heavily indebted to liability or vice versa. Rather, how each lending institutions and to each other [see kind of loan structure affects repayment Case Study 2 in Ethnography]. Similarly, and risk-taking depends upon how the Caroline Schuster,181 in her ethnographic contracts are structured. research on microfinance in Paraguay, describes how joint liability loans, in which Applications the entire group is responsible for paying back their debt, uses social relations as Microfinance experiments have clear collateral in the absence of other viable implications for policy, commercial forms of guarantee. operations, and the design of development programs. In fact, field- based lab experiments such as these Ethical issues have influenced microfinance institutions Because this experiment is essentially a (MFIs), which are increasingly shifting lab experiment that takes place in the towards individual liability loans with time- field, the ethical issues it raises are largely based incentive structures. the same as in the first case study in this section. However, the fact that it takes 109 Consumer Finance Research Toolkit place in the field does raise some Research shows that women are key additional ethical issues. actors in making financial decisions for the household.183 Women use much of their To recruit participants, the researchers available income for household used a technique known as “snowballing;” consumption, children’s school fees and that is, asking participants to bring along education, while also striving to put small their friends to participate in the study. amounts aside. But women often have Methodologically, non-random selection of limited power in allocating husbands’ participants did not pose a problem to the income or may struggle to safeguard internal validity of the study. Ethically, income and what little they can save from snowball sampling can sometimes pose the demands of husbands, family, and problems if new participants feel social networks. pressured to participate.182 But here, as for all of the methodologies discussed in Formal financial inclusion initiatives are this toolkit, Institutional Review Boards using field-based experimental design and place special emphasis on the provision of randomized control trials (RCTs) to test information and the justification of which the impact of formal financial tools and subjects are recruited to participate in a products on women’s ability to channel study. The researchers provided recruited savings. What is the relationship between participants with detailed information the introduction of a new financial tool about the study and ensured that (such as a savings device) and improved participants individually consented to savings? participating in the research. IMTFI researchers Deepti Kc and Mudita Tiwari conducted a field-based experiment with poor women who were part of Self Case Study 3: Randomized Help Groups (SHGs) in Bihar, eastern Controlled Trial: Innovative and India, to test if a simple savings tool – a lock box and key – could improve poor Interactive Ways to Improve women’s capacity to save.184 the Financial Capability and Kc and Tiwari’s ongoing work185 on Savings of Women in India gender and financial literacy has spread In research and policy circles increasing across multiple projects that served in attention is being directed at closing the some respects as “pilot” studies for the still significant gap between women and case discussed here. In Dharavi slum in men globally in terms of income, wealth, Mumbai they examined women’s and access to formal financial services. practices of storing and hiding money in Improved savings is a priority for many the home using a variety of informal women and of particular interest for means. Women did not trust banks and consumer finance research. 110 Consumer Finance Research Toolkit often were unaware of the formal financial The researchers incorporated quantitative services or products available to them. and qualitative methods within the randomized control testing design. The To address these issues, KC and Tiwari study population was randomly selected developed context-specific financial from seven villages in the Gaya district in education modules using a comic book the state of Bihar, eastern India, each of story-telling format. Modules Illustrated which had community SHGs (external financial concepts and scenarios of validity). Two hundred and three women, spending and saving using characters and all of whom belonged to socio- real life challenges women could relate to. economically disadvantaged groups, were The researchers then tested the financial divided into 4 treatment groups and one literacy tools in New Delhi with migrant control group (40 women each). The labourers. Both projects demonstrated the majority of households depended primarily effectiveness of the financial education on non-agricultural labour, with tools in women’s perception of banks and agricultural labour, smallholder farming their desire to save. and government payments also representing sources of income. All of the However, the ability to meet savings goals women participated in SHG’s and 85% often remained out of reach for lack of a deposited savings with their groups. But more formal saving device. This led KC women were only saving Rs 10, which and Tiwari to design a field-based was the minimum amount required for experiment in Bihar to test if financial membership eligibility. With membership education accompanied by an alternative in the SHG, women were eligible for other savings tool might positively impact benefits from government promoted women’s capacity for building up savings. programs, such as bank loans; women were therefore saving in order to be part of the group. Method A. (Control) no intervention B. Women received an alternative saving Inspired by a randomized control trial in tool: a lock box and a key. Kenya186 that introduced a lock box to C. Women received financial literacy women market vendors and male bicycle training. taxi drivers, the experiment set out to test D. Women were provided with a financial the effectiveness of both financial diary to track their expenses, but they education and savings tools beyond the were neither given any financial question of bank access. They literacy training nor any alterative hypothesized that financial education was savings tool. important, but insufficient if not E. Women received financial training, a accompanied by appropriate savings lock box and a key as well as a products. financial diary. (p. 5-6) 111 Consumer Finance Research Toolkit Women who received the lock box were The baseline survey was used to test allowed to keep the key, but asked to relative differences between the control make a “soft commitment” to open it only and treatment groups. Each respondent when they deposited money at the bank was assigned a literacy score based on or with their SHGs. “on-the-spot” reading and quantitative tests. Out of a maximum score of 120, The researchers hired women from the 11% of women achieved the highest score village to run the financial literacy training. of 80 while over 50% scored 0.187 It was important to the project design that Women’s financial status was also given a instructors were women in terms of score based on variables such as access capacity building and because gender to formal financial services and ability to dynamics played a significant role in the make household financial decisions was savings and literacy interventions. determined and given a score.188 During Teachers also needed to be good the randomization process, statistical tests instructors to whom study participants were performed on the scores and could relate; the researchers spent ten baseline data to ensure that there was no days training the instructors in facilitating meaningful difference between control the financial literacy modules. and treatment groups. On the second visit each of the randomly selected groups In order to understand how savings received their assigned treatments. behaviour changed over time as a result of the treatment, the researchers The third visit included a mid-line survey designed the study around five key visits to collect data after the intervention and with the women over a three-month focus group discussions were organized period, with a two-week interval between for the fourth visit. Women receiving the each visit. financial education treatment were tested for understanding and retention of story During the first visit, a baseline survey content. Local language, locations, and was created collecting data on important situations in the modules illustrated socioeconomic factors such as savings realistic strategies for reducing spending strategies with informal and formal on temptation goods and saving with financial services and challenges women SHGs and formal financial institutions. experienced in their money management practices. Women were asked to provide At the time of the final visit and the end of a detailed account of savings sources and 3 months the researchers collected data locations (including at home) and their on women’s savings after the intervention total amount of savings at the time of had been completed. Women were also interview was recorded. SHGs were the asked how their use of the savings tools most popular and important mode of impacted spousal and household saving, but many women also saved relationships to understand how the use of money at home. (p. 21) 112 Consumer Finance Research Toolkit formal and alternative financial tools were the lock box, 19% hid it from their affecting women’s lives during the study. husbands (p. 17). Follow-up visits tested the impact of the Findings financial literacy training for those women who received it. Additionally, survey and interview methods were used to elicit The researchers found that women who responses on how the lock box and were given the lock box with key financial training influenced intra- dramatically improved their savings, with household dynamics. Women’s financial the context-specific financial story-telling literacy skills were re-tested at the end of tools enhancing women’s attitudes toward the intervention and compared to the saving for those who received the results from the original baseline scores treatment. from the first visit. Women with a higher literacy score were Results showed that the intervention more likely to visit banks. Households that increased household dialogue about the used formal financial channels showed budget and spending decisions. Women more significant daily savings than those shared their new knowledge of financial households that relied on informal concepts and management with family channels. While the savings tool led to members and friends. Using a regression increased savings, the financial training analysis, researchers also saw that for had a ripple effect because women often each additional unit of improvement in the shared their new knowledge with their financial literacy score there was an husbands and children. increase in savings (Rs.3 or $0.05). Perceptions about formal financial Women experienced intra-household institutions were also measured at each challenges to their savings practices visit; at the beginning of the study only during the time of the study. Over 60% of 42% felt that banks were “customer- the women earned their own income, but centric,” with 76% reporting this to be the often had little say over its use in the case at the time of the fifth visit.189 household. In general, husbands limited women’s decision-making power over Women’s bi-weekly savings after finances and women devised numerous receiving the treatment were recorded at strategies to hide and therefore protect the third and fifth visits. Each savings the money they did save. During the time channel was noted, from savings hidden of the intervention, 50% of women at home or deposited in banks, to SHGs, reported that husbands were not to the savings in the lock box. Of the 94% supportive of their participation in the of women who received the lock box, 83% study and a small percentage reported used it to save during the time of the domestic disputes. Of those who received study. Notably, women who received only 113 Consumer Finance Research Toolkit the lock box saw a percentage increase in gap between informal and formal financial savings of 51%, with women receiving service options. both the lock box and financial education experiencing a 42% increase. In contrast, Isolating the effect of a financial tool on the savings of the control group increased financial behaviour can often prove by only 1% and those receiving only difficult where many variables are at play financial education increased by 8%.190 in influencing people’s adoption (or not) of The researchers noted some spillover a new product. Even more difficult to effect with women who received the determine is how precisely financial financial training sharing some of their literacy programs change not only knowledge with neighbours or their fellow knowledge but also behaviour. For SHG members. instance, even after learning about more cost effective means for remitting money, The study showed that financial education studies have often shown minimal impact could have an impact on its own and of this changed awareness on remitter’s change attitudes about formal financial choice of providers or use of services.192 institutions. Yet as Deepti KC qualified, “Financial literacy alone is not enough By building in different control groups for without good teachers. The financial the financial education training, the study literacy training worked because teachers made visible the impact of financial were well trained and women could education relative to the introduction of a identify with them.”191 Moreover, the savings tool. In agreement with similar but provision of a simple and appropriate larger-scale studies,193 the results show savings tool was essential to improving the importance of combining financial women’s ability to save. literacy initiatives with relevant savings or financial products to improve women’s access and help close the gender gap. Applications The financial literacy tools used in the This study has important policy study are based on qualitative research applications. RCTs are considered the with the study participants. They are gold standard of research for measuring designed with the target audience’s lives impact and evaluating outcomes. With a and experiences in mind and can be relatively limited budget and project further modified and adapted based on duration, the RCT design allowed the the interactive engagement between the researchers to isolate the impact of a researchers and the study population. The simple financial tool appropriate to the use and development of these stories local context. It also suggested a enabled ongoing dialogue throughout all sustainable interim solution to formal of the phases of the research with financial access because it is low cost and participants. This approach not only low maintenance; it can help bridge the improved savings but also changed 114 Consumer Finance Research Toolkit perceptions over the relatively short limitations include the challenges of duration of the study. following up with women over the longer term due to time and resource constraints. Field experiments that creatively combine RCT with qualitative, participatory Recently there have been increased calls methods over a longer time and with more for evaluating the concrete benefits of resources can help to sharpen the tools research to study subjects when for understanding the impact of financial conducting big data and RCT studies.195 consumer products. But as this study Critics argue that the data collection shows, even with time and resource process has become overly burdensome constraints, careful study design can go a on participants who are treated not as long way in showing impact and participants but as “data sources,” to be transforming circumstances on the mined for information while receiving little ground. concrete benefit from the research. Deepti Kc notes that particularly when research is about people’s finances and money, Ethical Issues research participants are forced to discuss and reveal intimate details about their The researchers noted that incorporating personal wealth and well-being. “We need multiple visits into the study design was to be extremely careful and considerate. essential to building trust with the That is one reason why we decided to participants. This was important because make five visits, rather than follow the women experienced new tensions in their standard three: baseline, intervention, and relationships with husbands and altered endline.”196 household dynamics as a result of their participation in the study intervention. Deepti Kc emphasized that the village Deepti Kc explained, women whom they trained to facilitate the financial literacy intervention were “Talking about money is always essential to the successful impact of the difficult. Why would I speak about intervention. “Financial training alone my money with a stranger? With cannot do wonders without efficient and multiple visits, we were able to well trained teachers.”197 By developing gain trust and women were more financial literacy and savings tools forthcoming about their money interactively with teachers and study management practices and their participants, the researchers treated savings.”194 women as collaborators in producing the context-specific modules. Financial “Multiple visits allowed the researchers to education about tracking expenditures, follow women’s well-being in the study in reducing spending on temptation goods, addition to building a more holistic and using banking services was a form of understanding of their situation. Potential expertise sharing and “giving back” that 115 Consumer Finance Research Toolkit the researchers and village teachers, and Toolkit for the Evaluation of Financial in turn, study participants, shared with Capability Programs in Low- and Middle- family and friends in the community. Income Countries (2013, International Bank for Reconstruction and Development / The World Bank) More about the method Experimental Economics: Rethinking the Rules by Nick Bardsley et al. (2010, Princeton University Press) Laboratory Experiments in the Social Sciences edited by Murray Webster and Jane Sell (2007, Academic Press) Evaluating Financial Products and Services in the US: A Toolkit for Running Randomized Controlled Trials (2015, Innovations for Poverty Action) Field Experiments: Design, Analysis and Interpretation by Alan S. Gerber (W.W. Norton & Company) Field Experiments and Their Critics: Essays on the Uses and Abuses of Experimentation in the Social Sciences by Dawn Langan Teele (2014, Yale University Press) Natural Experiments in the Social Sciences: A Design-Based Approach by Thad Dunning (2012, Cambridge University Press) Running Randomized Evaluations: A Practical Guide by Rachel Glennerster and Kudzai Takavarasha (2013, Princeton University Press) 116 Consumer Finance Research Toolkit Figure 21 Financial Training. Photo courtesy of Deepti Kc 117 Consumer Finance Research Toolkit Figure 22 Lock Box. Photo courtesy of Deepti Kc. 118 Consumer Finance Research Toolkit statistics. Relatively few studies collect information about all the financial tools Challenges in people use, and sometimes one mode of payment, such as a credit card, may be Consumer Finance used to purchase further consumer finance products (e.g., buying online Research insurance, sending money, depositing We hope that you have enjoyed the money in an e-wallet). Consumer Finance Research Methods In a similar vein, financial diary studies Toolkit and the many interesting and can help us understand how people innovative applications we discuss. combine financial products, but may shed Just as consumer finance itself is in a relatively little light on the social state of flux, so is research in the area interactions that guide product use. And changing rapidly. Since the first wire ethnography's suite of methods can transfer of money occurred in the 19th provide deep insights into social and century, consumer finance researchers cultural influences on behaviour, but it and practitioners have had to continually generally does not contribute the kind of adapt to new technologies, infrastructures, numerical analysis that helps us to human mobility, consumer needs, and, of understand behaviour at a large scale. course, financial globalization. To better understand what consumers do Rapid changes in consumer finance and how they do it, we need to be aware present significant methodological issues, of the limitations of our chosen field and and it is becoming increasingly difficult to be ready to use other methods when tackle these problems using one method necessary. This may well entail forming alone. Instead, issues are often best collaborations with professionals who are compensated for through combining experts in their particular field or quantitative and qualitative research.198 methodology. This toolkit gives This means that collaboration and practitioners a starting point to think about teamwork are just as important as what other methods they might use and methods and numbers in building a with whom they might collaborate. foundation for understanding human As we move forward, what are likely to be financial behaviour in its ever-changing some of the major challenges we face in complexities. accessing good data and producing For example, data sets can be a fantastic adequate analyses? The studies way to learn about consumer finance discussed in this toolkit provide some practices and trends. However, new telling clues. financial products are appearing every day, meaning that statistical data are not up-to-date. Moreover, diversity in product use may not be accurately reflected in 119 Consumer Finance Research Toolkit The north / south divide as much as with product developments. As the IMTFI's research network attests, Changes in consumer finance have some there is a wealth of valuable information interesting implications for how we try to being produced by researchers from all understand the global north / south divide. over the world. Companies seeking to invest in developing economies are keen to find out Research findings are also challenging how consumers think, act, and most of all, how we view “the poor.” For example, how they spend their money. For financial diary studies challenge common example, microfinance institutions, mobile assumptions about people living in money providers, and money transfer poverty, such as that they lack financial services have all sought to capitalize on tools or don't know how to manage their the “fortune at the bottom of the money. Similarly, the increasing use of pyramid”199 while delivering social benefits randomized controlled trials (RCTs) of greater consumer choice and more represents a shift in the development affordable services to broad segments of world from assuming that an intervention populations. in one country will be successful everywhere, to actually testing what works The dismantling of the north / south divide and adjusting accordingly. is only set to increase as more markets come online. Not only is a degree of Just because an intervention or program “product democratization” taking place, it is successful in one poor country does not is problematic to view the “Global North” mean it will be equally successful in as leading the charge with the “Global another. Mobile money has met with South” struggling to keep up. extraordinary success in Kenya, but has struggled to get off the ground in dozens Many innovations in consumer finance of other countries for various reasons, begin in the “Global South.”200 China, including scaling, branding, and Brazil, Mexico, India, and Kenya are just insufficient product education.201 some examples of markets that are ahead of the global curve when it comes to many Globalization and migration also challenge consumer finance products. Safaricom's the north / south divide. As various social mobile money service in Kenya, M-PESA, scientists have noted,202 the “transnational is one such example. capitalist class” in one country (say, Tanzania) may have more in common Researchers and practitioners may find with the same class in another country that, if they want to keep to keep up-to- (say, the USA) than with their own fellow date with developments in consumer citizens. But it is not just the wealthy that finance, they need to look beyond the form social groups across national usual WEIRD (Western, Educated, borders. Industrialized, Rich, and Democratic) nations. This point holds true for keeping Countless studies have pointed out that all up-to-date with research advancements kinds of migrants maintain family and 120 Consumer Finance Research Toolkit commercial ties across borders, often for populations can help us understand why many decades after they originally people might respond to changes migrated (including David Stoll’s work on differently and how they adapt to changing Guatemalan migrants, discussed in Case markets. This can help us to better design Study 2 in Ethnography). and target our products, programs, and policies in the future. These movements have crucial implications for consumer finance, since people send money backwards and Human welfare and well-being forwards across borders. Remittances to Human welfare and well-being will developing countries are estimated to continue to be critical issues in consumer have reached $436 billion in 2014,203 not finance, especially with respect to how because of wealthy people sending consumers everywhere make decisions. money, but because people of far fewer To figure out what kinds of welfare issues means are geographically mobile and are important, we need to think about maintain links home to their families and what choices and conditions different friends. Private remittances often make up demographics are facing. Simply living in a significant proportion of national GDP (in a wealthy country with good financial places like the Philippines) and are seen services is no guarantee that people will increasingly as potential resources for benefit from financial innovation. local/national investment.204 For example, the Eurozone boasts a high And as more and more consumer finance level of integration in regulation, currency, products became globally available and banking infrastructure (notably the through the Internet, we have greater Single Euro Payments Area, SEPA). choice as to how we send money around However, there are significant socio- the world. We are no longer limited to economic differences between using our local bank or a product that is households and nations, including income offered domestically. The north / south levels, product availability, demography, divide increasingly makes little sense as a and social/cultural specificities. Poverty, research framework, either from a debt, literacy, and inequality affect people consumer or product perspective. everywhere and shape their ability to take However, globalization does not equal advantage of consumer finance homogenization. Humans maintain innovations.205 diversity even as we migrate and The benefits of consumer finance can also communicate around the world. Just as be difficult to measure. One it’s worth looking beyond the west in our methodological problem is that global research, then, we could benefit disparity in relative income levels shapes enormously from conducting comparative how researchers conceptualize consumer studies. Comparing the consumer choice. It is very difficult to compare the practices and financial repertoires of quality of choices available to, say, a nations or other subsections of Haitian trader earning $2000 per year with 121 Consumer Finance Research Toolkit an American grocery store owner earning comprehensive understanding of how $40,000 per year. digital finance impacts consumer well- being. Research design needs to take into Whether the choices available to these account the fact that people are shopping consumers are adequate depends on their for consumer finance products in a global needs and the context in which they are market, and that this global market operating. In fact, sometimes people who presents new challenges to consumer have no access to formal banking facilities well-being and welfare. have more choice between consumer finance products (formal and informal) than someone living in a wealthy city. Inequalities When assessing choice, we therefore At the same time that digital finance need to take into account the range of promises to reduce inequalities, it can services available, their quality, consumer exacerbate them in other ways. needs, and avenues for consumer redress. Digital finance has mixed implication for inequality at the household level. For Another problem related to well-being is example, Sibel Kusimba’s research shows introduced by the fact that more choice is that a benefit of mobile money for Kenyan not always better. Even though we may women is that they receive more have access to more product information remittances than men [see Case Study 2 than ever before (thanks to the Internet), in Verbal Interviews]. However, the flip the sheer number of products available side to this success story is that women makes it difficult to compare them and continue to control less capital than make the right choice. This conundrum men—and hence women are still very has been long observed in the insurance much unequal. industry, where people particularly struggle to understand what different Good research and policy design can policies will actually cover.206 illuminate both sides of that story and help to close the “gender gap,” such as through This has also been the case with credit the development of products targeting products, since people often do not women's needs and priorities, or the understand the conditions they are implementation of laws regarding equal agreeing to (such as how much interest pay and property rights. they will end up paying in total). Problems of product proliferation and asymmetric Similar problems of inequality exist in information is compounded today, since larger society. As we discussed in the products are both more numerous and Interviews section of this toolkit, research come from a far greater range of on Hispanic college students [Case Study companies and geographic locations. 1 in Interviews] showed how ethnicity can be correlated with financial practices, but Without a clear picture of how people are not necessarily in ways that we might using products, we cannot develop a expect. 122 Consumer Finance Research Toolkit Professor Watchravesringkan found that Yes, mobile phone access is nearly family members had a positive influence universal these days, and it is not on Hispanic college students’ behaviours, particularly difficult to learn how to use and suggested that this insight could be mobile money. But whether mobile money used to develop financial management and other “products for the poor” decrease learning programs. Good research can poverty and inequality is questionable.207 productively challenge our assumptions There is a risk that people who have about how inequalities affect financial technological literacy and access will gain practices. an advantage over others, thus increasing the digital divide. Moreover, access to Another methodological problem is that financial tools is only useful if people have social categories such as “race” and money to manage (such as income or “ethnicity” can be methodologically receipt of remittances) in the first place. problematic, not least because our definitions of these terms change over But transformations in consumer finance time. How do you define a research do not inevitably exacerbate inequalities. population based on ethnicity or race? By There are many excellent and study participants’ own self-reporting? encouraging examples of people Based on government statistics or census benefitting from changes in consumer categories? finance. Some of these can be classed as “interventions,” such as the provision of It is difficult to point to a “right” way to microfinance services. define a population that experience inequality, since this definition will depend As David Roodman points out in his book upon what the research is trying to show. Due Diligence, there are just as many This is particularly true when conducting examples of microfinance resulting in net international comparative studies, since benefits as there are of it causing social categories in one country may be problems such as indebtedness. Access completely irrelevant in another. to reasonably priced credit can indeed be highly beneficial—but it must be applied How might cutting-edge changes in wisely. Many financial diary studies have consumer finance exacerbate now shown, people living in poverty are inequalities? One obvious answer is that perfectly capable of managing their own new products and platforms feed a “digital complex finances and, given sufficient divide” between people who have access resources, will competently manage their to technology (and know how to use it) own futures. and people who do not. The digital divide can be self-reinforcing if it helps people The trick with managing inequalities is who are already in the know increase their being aware of the multifaceted effects wealth and social status, and pass these that consumer finance can incur. Product lessons on to their children. innovation and mobility create wonderful opportunities, but can also put consumers at risk of greater indebtedness,208 reduced 123 Consumer Finance Research Toolkit consumer protection, fraud, and even simply being overwhelmed with choices (a phenomenon well-recognized by economists and psychologists). Researchers, practitioners, and regulators need to be aware that consumer finance product consumption today is taking place in a complex global market, and design our research and policy accordingly to protect consumers. Solid research will inform us of what kinds of interventions are needed, how regulation and policy might respond, and in what ways we need to accommodate people’s existing practices. Erin B. Taylor and Gawain Lynch The Hague, 4 February 2016 124 Consumer Finance Research Toolkit James Andreoni, University of California Acknowledgements San Diego Tom Boellstorff Production Michael Chibnik, University of Iowa Joe Deville, Lancaster University Erin Taylor, Canela Group Ruth Fortmann, Universidade Nova de Gawain Lynch, Canela Group Lisboa Bill Maurer, IMTFI Heather Horst, RMIT Jenny Fan, IMTFI Jofish Kaye, Yahoo Labs Ursula Dalinghaus, IMTFI Deepti Kc, IFMR John Seaman, IMTFI Sibel Kusimba, American University Nandita Badami, IMTFI [Matthew Levasseur, Dimagi] Nima Lamu Yolmo, IMTFI Alexandra Mack, Pitney Bowes Bill Maurer, IMTFI Project Leads Sarah Meiklejohn, University College Erin Taylor and Gawain Lynch London Woldmarian Mesfin, Addis Ababa Authors University Erin Taylor, Gawain Lynch Todd Sanderson, University of Sydney with Ursula Dalinghaus Charles Sprenger, Rady School of Foreword by Bill Maurer Management David Stoll, Middlebury College Administration and Coordination Kittichai Watchravesringka, University of Bill Maurer North Carolina Greensboro Jenny Fan Jing Jian Xiao, University of Rhode Island John Seaman Julie Zollmann, Bankable Frontier Associates Research and Editorial Assistance Anonymous reviewer Ursula Dalinghaus Nandita Badami Nima Yolmo Advisors The production of this toolkit benefitted Layout from consultations with a number of Ursula Dalinghaus experts: Reviewers Robin Beers, Wells Fargo Mollie Bell, Filene Institute The following people generously reviewed Ian Bright, ING London the sections and case studies in this Gerhard Coetzee, CGAP toolkit: Bruce Davis, Abundance Generation Melissa Fisher, NYU Lara Gilman, GSMA 125 Consumer Finance Research Toolkit Arturo Gonzales, Board of Governors of the Federal Reserve Greta Greathouse, USAID Tanaya Kilara, Bridge International Academies Simon Lelieveldt, independent consultant Liz McFall, Open University Martha Mckenzie-Minifie, ING London Ellen Merry, Board of Governors of the Federal Reserve Loretta Michaels, US Treasury Emory Nelms, Center for Advanced Hindsight Sascha Noé, Cordaid Rafael Perez Ribas, University of Amsterdam Simon Roberts, Stripe Partners Barbara Robles, Board of Governors of the Federal Reserve Ben Rogers, Filene Institute Ronald Wiekencamp, Cmotions Figure 23 Money Lei, photo by Bill Maurer 126 Consumer Finance Research Toolkit Appendix I. In the following section we present blog posts, originally prepared for the IMTFI Blog, which we think effectively illustrate some of the key takeaways and approaches presented in the toolkit. All of these examples showcase approaches to disseminating research results (sometimes preliminary) for broader public consumption. Other key takeaways in these examples include:  Knowledge sharing with research  participants Adapting project design and evaluation to  unexpected situations on the ground Opportunities for further research using  mixed-methods approaches  Visualizing networks or processes Ways of telling stories with data IMTFI Researcher Workshop 2011 A brief note that spotlights points of connection to methods, concepts and/or issues presented in the toolkit precedes each post. IMTFI Researcher Workshop 2010 127 Consumer Finance Research Toolkit Blog Post 1 We just took them out of the closet to pass them on to This post by Jacobo Menajovsky illustrates her younger brother who’s recently started walking, but how shoes, like data, can be read for when I took a closer look at them, I wondered if their best days weren’t behind them. Call me crazy, but I indicators of use. Shoes leave footprints immediately started decoding all the signs and for data analysis. But inversely, we can indicators of their usage. Yes, to me, data is also think about how big data might also everywhere. leave footprints for qualitative research We are constantly gathering, interpreting and acting on questions and modelling. data. Think about it. Every time you walk into a new situation, your “decision support system” starts to process past data to help you adjust to the new experience. Your brain is actually modeling those signs Originally published April 16, 2012 and symbols (data), building connections and classifying them into categories. Where is the data? Analyzing customer footprints for better product design What if you wanted to understand how these shoes were used? Do you think you could reconstruct the By guest blogger Jacobo Menajovsky, Senior Data Analyst – past simply by looking at them? There are lots of signs Grameen Foundation and indicators: a broken ankle wrap, a heavily-rubbed toe cap, and many holes. Now let’s move from data gathering to data modeling. When we put all this data to work we can build a great profile of how the shoes have been used. It looks like they went through a lot of kicking and dragging, and plenty of crawling. If you look at the soles, though you’ll see that they’re unworn. So, it seems the upper parts of the shoes were used more than their bottoms. This observation might even give you a few ideas about how to improve shoes like these and make them more durable. This is exactly what we call improving your product using customer footprints. In this case, the footprints are literal! As a data scientist and microfinance practitioner, I am always searching for signals and indicators that show how poor people are using products and services. I believe the best way to understand their behavior is by analyzing their footprints. This data can come to me in various formats (e.g. digitally or on paper) and platforms. I often have to put in a lot of work before I begin analysis, but if done correctly, it gives me a lot in return. These are my daughter’s old shoes. 128 Consumer Finance Research Toolkit In an ideal world, records would have unique customer Blog Post 2 IDs and information about the products that each uses IMTFI Researcher, W. F. Mesfin, whose work as well as past transactions. If you are really lucky you is featured in Ethnography Case Study , may also have some socio-demographic information like age, gender, rural/urban indicator, branch or shows how his qualitative study of cash location, household composition, family size, and usage in the market place might translate to poverty level. digital research and digital product design. The more data the better. Your data set can help you answer some key business questions: What’s the penetration of product A at different locations? Is this affected by poverty level or household composition? Originally published May 20, 2014 What about understanding our customers’ lifecycle? Do we see differences in outstanding balances at different One researcher's thoughts on money and metadata customers’ tenures? Our recent study on implementing based on fieldwork in Ethiopia data analytics provides an exhaustive list of business questions and analytic approaches. By IMTFI Researcher W. F. Mesfin At Grameen Foundation we are working towards As a researcher I tried to unpack what cash money helping pro-poor organizations crunch their numbers, really is. I am an information systems expert and my understand their customers better, and make more analysis of mobile money is primarily pragmatic and informed decisions. We are also refining our data depends on my readings across different domains, collection process, using the right set of mobile data collections tools and state of the art analytics to better particularly sociology and money, anthropology and understand the challenges and needs of the poorest. money, behavioural economics and money. I get After all, it’s only by gaining better insights that we will knowledge and information from these domains and contribute to the development of more tailored combined with my own areas of expertise I am able to products and services – from baby shoes to reflect on what mobile money or digital money should microfinance products– and that’s essential if we want look like and the nature or elements of the kind of to improve the lives of the most vulnerable people in our planet. system that can handle mobile money. Here are my reflections: http://blog.imtfi.uci.edu/2013/04/where-is-data- First what is money and how it is changing? Even analyzing-customer.html though there are different categories of money from social and anthropological points of view, in this writing I am interested in looking at cash as money and thinking about how its changed forms affect system design. To me cash money is simply information with some specified metadata, like color, image or icons, numbers, and some other hidden security controlling means (see the figure below). People agree to accept this information as money because they know they can give it to others without concern. What makes this 129 Consumer Finance Research Toolkit piece of paper (cash money) and or coins is the design implications of my ethnographic research I did information (metadata inscribed on them). in 2012 in Ethiopia among open air market participants. Purposes of money’s metadata: • Images or icons: The icons or images on money bills have different purposes like: as national identity (collective national values). For example, Ethiopian money bills have pictures of coffee plants, farmers, a map of Ethiopia, historical buildings, the signature of Metadata the governor, and a tractor machine. Some countries This is to say that if we remove these metadata, the also inscribe photos of elite individuals and their remaining piece of paper cannot be considered as sacrifice for the nation, for example photos on USD. In money. And thus, these metadata are making money Ethiopia, the icons and images inscribed on money bills valuable and make people develop trust and enable illiterate users to identify money bills. confidence. But with the digitization of money, in its • Security tools: These are features inscribed on money current state of research and development, these bills to identify real money from forgery. The problem metadata are excluded from existing platforms and with these tools is that, when the bills get old, these solutions. In the current platforms or solutions, money features usually fade away and may not be visible, is represented as a simple positive rational number of which makes differentiating forgery from valid money the form say 2.89 USD, 0.89 USD, 247 USD etc, bills difficult. excluding the metadata as well as the different money • Serial numbers: These have also a very important role. denominations. For example, in Ethiopia our When all or parts of these numbers are lost (faded currencies are denominated as (5 cent, 10 cent, 25 away), individuals do not accept. Such money bills have cents, 50 cents, 1 birr, 5 birr, 10 birr, 50 birr, and 100 to be taken to banks for replacement, as per the birr notes). This indicates that we cannot pay (get paid) respondents from my field study. Otherwise, they for example exactly 12 cents, 11 cents, 9 cents, 9.87 etc. become worthless, as people, particularly illiterate users However, we know that with the digitization of money, will not accept for change during transaction. it is possible to accommodate any amounts and thus, • National identifier: This is written in both English and unlike in the case of cash based transactions, in digital Amharic and identifies the legal issuer (governor). transactions making changes is not an issue. That is the • Color: All Ethiopian money bills are color coded, bonus of current computing and mobile money which enable illiterate users to identify different bills. technologies. But such money representation with Color and images on money bills are used for counting positive rational numbers and removal of money’s and computational purposes. For example, illiterate metadata elements faces challenges when it comes to individuals know the sum of 10 birr and 5 birr will give people that are illiterate. For example, illiterate users 15 birr and yet do not know how to spell these know their balance by counting the material money. numbers. When they are also asked to pickup money They do not know what numbers 20, 40, 12 etc mean. bills of say 50 birr from a lump sum of bills with Rather, they know these figures when they are different denominations, they easily identify them physically handed them and able to count them by through their color. Thus, it could be said that color of hand. The following paragraphs outline some of the money bills is a means to identify them. 130 Consumer Finance Research Toolkit • Economic value: These are numbers written in terms of One thing we learned is that current mobile money roman numbers as 1 birr, 5 birr, 10 birr, 50 birr, and platforms and solutions did not consider these 100 birr as well as Ethiopian numbering systems and metadata in their design. My reflections and proposed the Amharic language. Birr has denominations of 1, 5, solution for this problem will be presented in two 10, 50, and 100. forthcoming papers to be shared on the IMTFI Blog once they are published. Money organization From my open air market study I have also observed and understood that illiterate merchants and customers distribute their money among different bags or pockets. They also give different labels or names like bag for sales from coffee, sales from salt, and sales from other materials. When they need a change, they try to look into the respective bag. In fact if there are no changes in the necessary bag, they take from another bag and return the money later. It is kind of a “loan” from the other bag. I call this “distributed cognition”. They remember from which bag they took change and want to replace the amount they took. Thus, it appears that from a technology design perspective, designers need to be aware of such cognitions and practices and its implication for design. Problems with cash The material nature of money bills also has an added I also observed that even though the material property value for illiterate and visually impaired people. of cash helped illiterate people count and know their Illiterate people make some simple mathematical balance as well as the difference between different computations (additions, subtractions, multiplication, currency notes, (based on its color), it has some and divisions) by moving money bills here and there as limitations. It can get old and stick together, part of it they cannot accomplish these through writing numbers can be torn and people are not willing to accept for on paper or calculating machines. For example, in change, some individuals also make forgery money and order to make payment or receive payments people easily cheat illiterate rural people. It also creates count money bills and in order to count them, people difficulty for making changes, if there are no changes. usually sort and arrange them according to the But, with the upcoming mobile or digital money denominations (from smallest money bills to the systems, even though it appears that the issues of largest) and then hold the stack in one hand and count forgery, sticking together, and changes can be with the other hand. In this context digitization can addressed, the issues of operationalizing technologies make illiterate people frustrated, unless there are for these issues can not be an easy task. Addressing one solutions for this issue. of the issues will come at the expense of another. 131 Consumer Finance Research Toolkit Final thoughts Finally, based on the nature of money digitization and money handling experiences of illiterate people in Blog Post 3 developing counties, I feel that new technologies need )n this post, Charmaine )laiu Talei describes to have capability like audio in order to embed how she incorporated a knowledge-sharing money’s metadata, and enable individuals’ to fieldtrip into the project design. Research experience “physicality within digital environment”. I participants were given the opportunity to also recommend interested readers to consult the work hear and vote on the accuracy of the of (Balen et al 2009) for more reflections about money study s main conclusions. She also explains digitisations as related to ease of usability, security, and auditing. In this case I want to make a note that the how study participants helped to sharpen work of Balan et al. (2009), focuses on literate people the research framework itself during the while my reflection is in the context of illiterate people data collection phase of research. See the who have different money practices. toolkit entry, A Note on Ethics, for how this applies to questions of ethics and informed For further details on the project see here for the final consent. report. You can also read Mesfin's working paper from 2012: "Understanding Social Relationships and Payments Among the Poor in Ethiopia" or email him with any comments and Originally published January 22, 2015 suggestions:

. Understanding the transformative value of Tongan http://blog.imtfi.uci.edu/2014/05/one-researchers- women’s kau tou lālanga: mobile mats, mobile thoughts-on-money-and.html phones, and money transfer agents Project Page: By IMTFI researcher Charmaine 'Ilaiu Talei Year: 2010 Kau tou lālanga is a group of Tongan women who The Impact of New Technologies on Social collectively weave one another’s fine pandanus mats to Payments barter and sell. Their prime customers are Tongan http://www.imtfi.uci.edu/research/2010/fikre.ph women living in diasporic communities around the p Pacific Rim. Our research has determined two business negotiations of kau tou lālanga: firstly, to weave per lineal foot, also known as‘iate, and secondly, to weave towards a ‘gathering’, orkātoanga. An ‘iate negotiation starts with a customer, usually a local person, making an order to a collective to weave one or two mats— only a small quantity. The second negotiation, kātoanga, is a gathering between a number of weavers from a collective and a group of customers, who are mostly Tongan women from overseas. Before 132 Consumer Finance Research Toolkit a kātoanga, the parties involved negotiate the large central island is Tongatapu, where the capital number of mats to be exchanged, the sum of cash for Nuku’alofa is located. Three fieldwork trips were the order, and the date and venue of their gathering is conducted within the project timeframe. A total of also agreed upon.Kātoanga agreements reach higher eight weeks in two remote islands groups: Vava’u and annual returns than ‘iate negotiations. Ha’apai. Being remote islands they generally have less access to a wider range of cash making opportunities in comparison to the main island. Vava’u and Ha’apai are also mat making epicenters. The fieldwork sites in Vava’u include the rural village of Leimatu’a and Kameli in Neiafu town. In Ha’apai, the sites include the rural villages of Fangale’ounga on Foa Island and Pukotala on Ha’ano Island, and finally, the urban village of Pangai. These village sites were selected based on existing contacts and referrals of where to best find kau tou lālanga groups. A total of 24 participants were interviewed over the course of three trips, 14 from Vava’u and 10 from Ha’apai. Some 20 were sole weavers and all were women and 4 acted as trading A kau tou lālanga in Kāmeli, Neiafu Vava’u, agents, of which 2 were men. Tonga. Photo: C. ‘Ilaiu Taleidd Tongan mats, or fala, are part of a wider system of customary gift exchanges within Tongan society. In this customary sense, Tongan mats are cashless forms of value storage. Such value is traditionally activated during a Tongan occasion, such as funerals, weddings and birthdays. A traditional gift from guests to hosts or vice versa could consist of tapa cloth and several types of fine mats. Thus, fine mats—the products of kau tou lālanga businesses—are highly prized items in Tongan material culture. For this reason, this work is part of a A participant displaying her mats (fala hinehina) for a wider discussion of gifts of exchange studied by anthropologists: Adrienne Kaeppler (1999), Phyllis kātoanga with her United States of America based Herda (1999), Ping-Ann Addo and Niko Besnier Tongan customer, Tongatapu, Tonga. (2008), and Fanny Wonu Veys (2009). However, their Photo: C. ‘Ilaiu Talei analyses exceed the purposes of this blog. The first objective of this research was to understand The Kingdom of Tonga is located in the South Pacific what is the transformative value ofkau tou lālanga in Tonga and secondly how mobile phones and money (20 00 S, 175 00 W) and is an archipelago of 169 islands transfer platforms help to achieve this transformation. stretched throughout 747 square kilometres. The main 133 Consumer Finance Research Toolkit The transformative value was first defined by the The transformative value of this business for you as a weaver is authors at the project’s conception as, ‘the potential to not about receiving money to spend or save, but being enabled to move away from an uncontrollable financial situation financially satisfy the needs of your family, at the time of need and to a position where one can manage financial every time of need. Importantly, it is from this position that you challenges with confidence’. The findings have refined gain emotional and mental confidence. It is for this significant and elaborated on this statement. reason why you choose to weave and partake in the business of mats. The first survey investigated (a) motivations for joining a kau tou lālanga (b) understanding one’s financial role Moreover, this study broadens our focus on in their family and (c) what one spends their profits or transformations that take place because of the business wages on, the survey attempted to shape an initial of kau tou lālanga, such as socio-cultural changes in the understanding about the transformative value of kau tou role of female weavers in their families. Traditionally, lālanga. Preliminary findings show that paying childrens’ mats have been considered the woman’s domain in education fees, maintaining one’s home through utility Tongan society and it was shameful for a man to bills, feeding dependents and donations to church dabble in women’s work. Evidently, the financial offerings and supporting village fundraisers are reasons appeal of the business has normalized the dual- why participants weave and join kau tou lālanga. The gendered activity of making mats and has helped to three top reasons ranked in order are (1) to pay remove the male shame of helping, especially when household utility bills, (2) to pay childrens’ education, many hands do make the work easier. Undoubtedly, and (3) to make church donations. kau tou lālanga is changing the female weaver’s role and consequently others in her family and within the wider It became clear after the first survey that the original Tongan society. definition was limiting the emotional motivations of the weavers. For this reason, the interviews of the second In answering the second objective of this project, the fieldtrip included asking weavers why they chose to do findings reveal that there is a technology knowledge this business and in other words what value they see in gap for older members of collectives, who are also the this kau tou lālanga? Their responses certainly leaders and decision-makers in the group. This gap highlighted that it is not about creating a huge savings delays business innovations like mobile money and the account but instead creating a sense of personal use of banking applications to facilitate transactions. satisfaction when one has met the needs of their The very establishment of kātoanga negotiations families. The outcome of providing therefore embeds stresses that weavers and customers alike still prefer to value into what they do as weavers. transact face-to-face, so customers can thoroughly check mats before closing a deal and weavers can count Our final fieldtrip allowed us to present back a their money before releasing the mats. In this way summary of their responses. They were asked to vote money transfer agencies have not displaced the practice yes or no if they agreed with this statement as of kātoanga, which explains why money exchange accurately describing transformative value of kau tou platforms were less important for some collectives and lālanga and why they choose this business. All the less used by such weavers than what was first assumed. weavers present at all presentations answered in confidence ‘yes’ to this statement of transformation. The transformative value of kau tou lālanga has been an Translated into English: invaluable investigation because now we can begin to 134 Consumer Finance Research Toolkit understand the ‘livelihood’ of Tongan weavers; Blog Post 4 revealing how kau tou lālanga affects the weavers and While this post is oriented around a more their dependents but also how weavers shape this academic audience, it shows how simple tools business to achieve their customary and financial goals can be used to map card use distinct from card ownership. Researchers wanted to see the http://blog.imtfi.uci.edu/2015/01/understanding- networks created by the practice of card transformative-value-of.html lending, which was otherwise invisible to retail companies that only tracked transaction data connected to a single owner. String and Project Page: push pins help to visualize qualitative data Year: 2013 relevant to the data sets collected and Understanding the transformative value of analyzed by retail entities. The study findings Tongan women’s kau tou lalanga: mobile presented here can be useful for policy makers mats, mobile phones, and money transfer evaluating credit scoring and indebtedness, or agents for retail outlets and banks in adapting http://www.imtfi.uci.edu/research/2013/talei.ph product design to existing use or unmet credit p needs. Originally published November 13, 2012 The Economy of the Quota: The Financial Ecologies and Commercial Circuits of Retail Credit Cards in Santiago, Chile The following is a translation of preliminary results from research conducted by José Ossandón, with the support of Tomás Ariztía, Macarena Barros, and Camila Peralta, and funded by IMTFI. 135 Consumer Finance Research Toolkit Luisa and her husband have informal jobs, neither has access to checking accounts or bank loans. Luisa is, however, an active participant in three informal financial institutions, two pollas (rotating savings organizations) and a caja común (“common fund”) that functions as a Christmas savings club. The caja can also be used as a source of credit, but under certain restrictions. Loans must be repaid with interest, there are fines for late payments, and if a member misses her quota for three consecutive dates, she is removed from the group and the money she has contributed up to that point is not returned to her. Luisa is a 54-year-old housewife who lives in the municipality of La Pintana, south of Santiago. Seven Luisa also has access to loans offered by retail others live in her household: her husband Patrick, her companies. Ten years ago, she acquired her first credit children Nacho, Paty, and Andrea, her son-in-law card in the department store La Polar and a few years Rafael, and her grandchildren Camila and Cristian. later, ended up with cards from Paris, Corona, Tricot, Luisa also has a fourth daughter Katia, who lives with Fashion Park, and Salcobrand. Two years ago, her husband Rodrigo in the same neighborhood. however, after feeling that her debts were spiraling out Luisa’s husband Patrick works as a freelance painter of control, she closed the Tricot, Fashion Park, and sporadically, and he earns on average 150,000 pesos Salcobrand cards, and, a year ago, a loan renegotiation (US$312) a month. In addition to housework, Luisa ended with La Polar blocking her card; today, then, she manages a kiosco or small shop in her home, which only holds cards from Paris and Corona. Luisa uses earns her between 20-30,000 pesos a month. Andrea these store cards in various ways. For example, and her husband work and take care of their own between September 2011 and February 2012, she used expenses. Paty, in turn, is unemployed and so receives the AlmacenesParis card to purchase two pairs of help from her parents to cover her expenses and those sneakers for her son Nacho, a cell phone for herself, of her daughter. Nacho is studying nursing with the and merchandise in a supermarket, which is part of the support of a loan (called a Crédito Aval del Estado, or same business group as the store and where this card is State-Guaranteed Credit) and recently has begun to also accepted. Each purchase was paid for in six receive his first income as an occasional worker in installments. She also used the Corona card for a cash construction. Luisa and her family maintain their home advance. with the money that Luisa and her husband earn and with the financial support they receive from their But Luisa did not use her cards only for her own children. purchases. For example, between December 2011 and January 2012, Luisa lent her Paris card nine times to With regards to her financial life, Luisa has a savings her daughter Andrea—three times for installment account, an emergency fund of 40,000 pesos cobbled purchases of merchandise, another five times to together with money from the kiosk in the purchase goods in the shop (shoes, an iron, an oven, BancoEstado, a state-owned financial institution. Since and pants for her son), and once for a cash advance 136 Consumer Finance Research Toolkit consisting of six installments of 15000 pesos each. In common for Chileans to lend health insurance addition, in February 2010, Luisa lent her card from La vouchers (bonos) to one another, they also lend retail Polar to Andrea to buy a refrigerator; she also lent her credit cards to one another. What we did not expect Corona card to Andrea for an advance of ten was that in every home we visited, we would find such installments of 10,000 pesos. Luisa has also lent her La lending and borrowing and the degree of complexity Polar card to her daughter Katya for a cash advance that emerged as a result. In this way, little by little, card and for a furniture purchase, and the Corona card to lending came to constitute the main object of our her daughter Paty to buy an iron in ten installments. attention. In this post, we use some elements of recent On two other occasions, moreover, Luisa's son-in-law work in economic sociology to begin to unravel what Rafael used her Corona card, once to buy himself this is all about. sneakers and another time to buy a cell phone for his son. In our qualitative study of the financial practices of Sociology of the Quota thirteen households in three municipalities in low- income sectors of Santiago, we encountered many In short, it is not difficult to associate the practice of stories like this one. These stories, much like those card lending to the basic principle of the new economic recently described by Ariel Wilkis, do not necessarily fit sociology, as formulated by Mark Granovetter in into the traditional categories associated with studies of 1985 (.pdf). At first glance, credit cards might appear to popular finance. Here there is no clear demarcation be private property, owned and managed by the person between formal and informal financial inclusion. There whose name is in the card, but closer observation of is exclusion, without a doubt, in the sense that these card lending reveals a parallel and collective circuit of stories are about people without access to bank credit, debt—that is, a network. Still, how and what might we often because they do not have access to formal see were we to examine the lending of credit cards as a employment, or to so-called créditos sociales (“social network? What are the nodes and types of relations? credits”) from cajas and cooperatives. These stories How to classify the types of actors involved? also speak to popular finance, since like Luisa, many of our respondents are active participants in “pollas” or Inspired by the description offered by one of our other rotating savings and credit associations. Yet, such favorite sociologists John Law (2007) of the advantages stories also evince formal inclusion and economic of “pin boards” or bulletin boards as a way to think rationality in the traditional sense. The credit cards of visually, we felt we should pursue a more experimental retail businesses are almost universally present in the path to understanding such lending networks. Instead homes we studied, and we encountered many stories of of working directly with relational data visualization people who, over the years, have become experts on software (such as Pajek), we, therefore, decided to interest rates, loan renegotiations, and installments—a embrace the flexibility afforded by using our own complexity that is multiplied if we also consider that hands. Armed with the necessary materials—cork many deal with several cards at once. bulletin board, yarn, and pushpins of different sizes and colors—we met to search for a way to think visually It is in this context that we encountered the practice of about what we were finding. lending and borrowing retail credit cards. Obviously, we were aware before our investigation that, just as it is 137 Consumer Finance Research Toolkit We completed similar exercises with the other twelve households. Seen below are the networks of retails card use in the households of Carmen, Marisol, and Yeni. After several attempts, the images you can see here have left us satisfied as a good way of visualizing our data. But the central question remains: So what? What is this all about? As has been already mentioned and as can be seen clearly above, credit card lending practices produce networks. But what kind of collective or social formation are we talking about? At what level do these networks operate? The Scale of the Quota Economy This is the case of Luisa. The red pins represent her and her husband, and blue pins below, her daughters One option is to focus on the nuclear family or and sons. The large pins represent retail store cards. In household as the fundamental unit of such networks. this case, Luisa is the only one with cards—one from As we saw in the same case of Luisa, however, card Paris, one from Corona, and another from La Polar. lending can span different family units living together The yarn threads represent uses of a card involving in the same residence. A second option is the family or some form of credit, and they connect the person who networks of extended kinship. Luisa not only lends her receives the loan of the card with the credit card used cards to her family members living in the same for the transaction. As can be seen, Luisa has used her property, but also her daughter Paty, who lives in three cards for personal transactions, but the same another house. In several of the cases studied, however, cards have also been used by her daughters and sons- we found card lending extending beyond the family to in-law. friends and neighbors. The case of two homes that 138 Consumer Finance Research Toolkit were connected via two friends turns out to be “Imagine, for instance, each installment is instructive in this way. 10,200 pesos. I give my mom eleven or twelve lucas [Chilean slang for 1,000 pesos], I always give her a bit more, because they always charge my mom for the mail service, the use of the cards, whatever damn fee they add—a thousand of this, fifteen hundred of the other. It’s the same with my dad, they are always charging him five lucas extra, so I always give some money on top [of the installment amount] to my mom.” (Patricia) Moreover, as the second quotation, taken from another case, illustrates, such circuits entail a system of parallel calculations, which are in fact often drawn in the margins of the monthly bill. A useful concept for expressing this particular type of But what is lent? What is the medium of this circuit? social formation is that of “commercial To understand this, it is important to consider that here circuits” developed by sociologist Viviana Zelizer we are not talking about just any network. In more (2010), which refers to circuits of economic transfers technical terminology, these are 2-Mode networks, among a delimited group of actors, who bestow upon since the actors are not connected directly among each these transactions a shared meaning. These circuits of other, but through nodes of a different type. What we transactions establish a clear line of belonging and have here are people connected among each other make use of a particular medium of payment. We through the use of a common card. The cards, in turn, suggest that each of these networks of credit card are not just any node. They correspond to what Callon, lending functions as a commercial circuit that frames or Millo and Muniesa (2007) have called “market connects to existing collectives—neighborhoods, devices”—that is, objects that do not only mediate families, or households—but that also has its own between humans, but have an active role in the emergent character and forms of inclusion and transformation of the relations that they connect. exclusion. Indeed, as shown in the first of the following citations, an important part of the interviews revolved Unlike a traditional commercial exchange, whose mode around the edges or boundaries drawn when a of payment is cash and for which there is no record commitment is broken and how the limits of these besides the receipt, every transaction realized with a circuits can be re-established. card is recorded, as the bills we receive at the end of each month repeatedly remind us. This information is “Flor, my neighbor, was slow to pay, so now I key to the operation of retail companies, which don’t lend them [cards] to her, because then statistically evaluate the behavior of each of its she takes a long time to pay and I have to pay customers. The type of evaluation performed by everything myself. And afterwards, they screw Chilean retailers is called “sowing” (Ossandón 2012), you with the card.” (Luisa) and it consists of extending credit, or increasing the 139 Consumer Finance Research Toolkit credit limit associated with each card, not only based Text prepared by José Ossandón for 7th Chilean Congress of on external variables (such as income, age, or state of Sociology, University of La Frontera, Pucón—24, 25 and 26 employment), but also on payment behavior. In other October 2012. From the work of the research team led by words, if the card lending among family members and Ossandón, with the participation of Macarena Barros, Camila friends is “managed” well, in the sense of paying off Peralta, and Tomás Ariztía. Study developed in Programa de the debts as they are generated, the increased use of the Estudios del Consumo y los Mercados, Instituto de Investigación card increases the credit limit assigned to it. en Ciencias Sociales (ICSO), Universidad Diego Portales, and funded by the Institute for Money, Technology and Financial This brief “socio-technical” detour is important, Inclusion at the University of California, Irvine. because it allows us to understand something key, that Text in Spanish originally posted at the blog Estudios de la the medium of the commercial circuit described here is Economia. Translated by Taylor Nelms, with assistance the credit limit itself. People do not lend money, but from Smoki Musaraj. the capacity to borrow and go into debt, their quota (or "cupo" in Spanish, a word that suggests both a bound and an allotment), that is delimited by the algorithms of http://blog.imtfi.uci.edu/2012/11/the-economy-of- retailers’ risk analysis systems. It is, therefore, no quota-financial.html coincidence that at the center of our networks we find “dueñas de casas” (female heads of household). Their centrality is directly related to the “discovery” by stores’ Project Page: analysts that adult women are statistically better Year: 2011 “pagadores” (payers), despite sometimes lacking their The Financial Ecologies and Circuits of own income. What stores probably do not know—or Commerce of Retail Credit Cards in Santiago do not care to know, since their concern is timely de Chile payment, not a sociological understanding of what lies behind each customer—is that behind each card is the http://www.imtfi.uci.edu/research/2011/ossando emergence of a new financial circuit, an entire n.php “economy of the quota.” In this post, we have used some visual and conceptual tools of economic sociology to begin to understand this economy. But this is only a first step. We believe that deepening our understanding of this economy of the quota (economía del cupo) is not only academically relevant, but also may help us to understand better how the boundaries between financial inclusion and exclusion are being reformulated today ... * The names in the cases and quotations have been changed to maintain anonymity. 140 Consumer Finance Research Toolkit The social network maps show that women have various positions of advantage – what sociologist Pierre Blog Post 5 Bourdieu called social capital. Women often have many This post by Sibel Kusimba was a connections, sending and receiving money from preliminary report on the findings of her siblings, children and grandchildren this helps make research, featured in Verbal matrilineal kinship ties important in social network )nterviews, Case study . Kusimba uses graphs. Finally, women are often brokers – connecting groups or “cliques” of others who would an ethnographic writing style to present not otherwise be connected. her data. )t is a great example of telling a story with data in the presentation of Figure 1 shows mobile money flows among a family findings. This approach for explaining living in the rural hamlet of Naitiri, Nairobi, and data could be essential to disseminating Chicago. The two main groups in Figure 1 are the findings to various audiences children of a pair of sisters. A 67-year old farmer I will (stakeholders, research participants, and call Ruth (red circle) receives remittances from many of her eight children and 44 grandchildren, in turn policy makers). It also illustrates how circulating them back to her children but also brokering qualitative methods like ethnography or these resources with her deceased sister’s oldest interview research are not just a method daughter (green circle), who in turn sends and receives of data collection but a particular form of remittances with her siblings (also in green) and their communicating and writing about the children (in yellow). Ruth’s network also receives fieldwork itself. international remittances from the United States from one son and one daughter (light blue). Originally published November 11, 2013 Grandmothers as Mobile Money Brokers in Kenya By Sibel Kusimba based on her IMTFI-funded research In Kenya the use of mobile money sending systems has become a part of daily life. In 2012 my research team and I used semi-structured interviews to collect information on the connections forged by sending and receiving mobile money. From this information we created social network maps which reveal the paths of mobile money that circulate in families and show the Writing in the July 2013 issue of American Anthropologist, connections of love, reciprocity, and obligation among Matthew Peeples and W. Randall Haas review theories members. of brokerage. The “individualist” model associated with 141 Consumer Finance Research Toolkit Burt’s classic study assumes brokerage is a position of sisters. At the deceased sister’s funeral (often a time advantage that implies control over flows from one when social groups and generations reconstitute group to another. |The alternative “collectivist” theory themselves, and when discord is displayed and argues that brokerage can be an ambivalent, risky or assuaged), the children of these two women, who live disadvantageous position and this can be more in Naitiri, Kimilili, Chicago, and Nairobi, discussed the appropriate in settings where group or collective high cost of education. They then formed a credit and interests are valued more highly than those of savings group in which each of them agreed to individuals. Where the interests of the group in general contribute 1000 shillings a month to a common savings are more valued, individuals will in fact use their broker account from which school fees would be paid on a position to close the “structural hole” and create more rotating basis. The members meet once a month for a ties. These cultural settings may value collective meal, where they also contribute 1000 Ksh. each interests- the trust or frequent contact allowed by the towards a banked fund for school fees. Mobile money dense network- above those of individuals. services are used by some at the meeting to send mobile money to the treasurer - from Chicago, a There may be aspects of both individualist and daughter uses Western Union. collectivist approaches to brokerage in Kenyan women’s use of mobile money services. In another In many families, siblings and other close relatives use example, Sister Lucida is a 47-year old nun and student mobile money to contribute regularly to informal in Chicago who sends about $300.00 a month to her savings or insurance plans in anticipation of funerals, mother in Homa Bay County, Western Kenya. Sister weddings, medical care and educational fees. Matrilineal Lucida is too busy to hear and assess numerous flows of mobile money often give mothers and requests for school fees and business investments from grandmothers the potential to be brokers; but often her relatives and allows her mother, a widow, to use the women use these positions to recirculate funds, and money she sends as she sees fit to help the family. create new, close connections of benefit to the group as After her father died, Sister Lucida explained that her a whole. mother became vulnerable among her in-laws, who chased her away from her home and stole her dishes Dense networks may also reflect the limited social and home furnishings. Sister Lucida and her siblings prestige and authority of mothers and grandmothers built and furnished her a new home on land they who, on maps of mobile money, are so central in purchased that is the envy of the village. Sister Lucida collecting remittances. Several months after my field sends her money for “upkeep,” and is aware that the research in the summer of 2012, Ruth was hospitalized money ends up helping others in the family as her with typhoid. Her son in Chicago told me that her mother sees fit. Both she and her mother enjoy her children in Kenya had been unable to raise the 40,000 status as a mother of a child in America. Kenya shillings (about $600.00) necessary for her release. Ruth continued to accumulate a bill for two As a broker, the Naitiri grandmother Ruth has used her weeks after her treatment while she waited to be let go influence in the family to fill the structural hole and from the hospital; her son and daughter in Chicago had further the connections between her children and her refused to pay the bill, arguing that the large number of sister’s children. The children of Ruth and her deceased siblings in Kenya should raise the money by sister have created a family association to collect school contributing amongst themselves. As the impasse fees for the children and grandchildren of this pair of continued, Ruth called her oldest son in Nairobi, and 142 Consumer Finance Research Toolkit during their heated conversation she told him that the family’s economic security was in large part a result of Project Page: the remittances from America. He responded by telling Year: 2012 her to henceforth inform her Chicago children of her Mobile Kin and Mobile Money: The needs. After another week, her Chicago daughter sent Anthropology of International Remittances in the money, and contact between the siblings was discontinued for several weeks as hard feelings were Kenya nursed. http://www.imtfi.uci.edu/research/2012/kusimb a.php Dense networks of mobile money result from strategies that value reciprocity, communication and trust. Reciprocity and closure are the most valued forms of social capital in the mobile money networks of Kenyans, and these often leave individuals, even brokers, unable to accumulate the resources they need. A request cannot be denied; instead one’s phone must be shut off or “lost” to avoid pressure for remittances. In a working paper for IMTFI, we examined how gender is reflected in mobile money networks in matrilineal ties, in the centrality of mothers and grandmothers, in the strong patterns of reciprocity that make receivers also senders, and in polygynous families where money is from those with resources to those without: from men to women. Gendered flows of mobile money provide a safety net and reduce risk, but often the resources sent are not enough, or cannot be held on to long enough, to prevent a mother from languishing in the hospital while her children squabble over the bill. Mobile money may simply provide a new technological means through which women cope in societies where they are marginalized from resources and wealth: by relying on their children, siblings, and others in their social network for the emergencies and needs of daily life. Read Sibel Kusimba's full working paper, "Social Networks of Mobile Money in Kenya". http://blog.imtfi.uci.edu/2013/11/grandmothers-as- mobile-money-brokers-in.html 143 Consumer Finance Research Toolkit Blog Post 6 When we decided to work on mobile savings in What happens when the separation of Tanzania, we had an initial understanding of the value control and treatment groups in a of formal savings from the economics literature. We randomized control trial is complicated knew that savings accounts — from banks or mobile by developments in the field? In this money operators — were becoming increasingly common in many unbanked areas of the world, and preliminary analysis of study results that adopters derived some important benefits from from an experimental project in their accounts (Dupas and Robinson 2013; Prina 2015). Tanzania, the researchers describe how What was less clear, and became the focus of our work, they were able to problem-solve in their was the extent to which savings accounts generate analysis when marketing visits were not social benefits (or costs). In particular, we became targeted as planned in their research interested in measuring the extent to which savings design. The researchers show how accounts generate “network spillovers”— that is, impose benefits or costs on the adopters’ social interview techniques, social network network. The questions we posed to ourselves were: analysis, and marketing surveys can be Do savings spillovers exist? Are they positive? effectively combined within an RCT Negative? Or both? Who gains and who loses? Finally, experimental design to establish the does savings generated through mobile money impact of mobile accounts on obligations accounts lead to such spillovers? within social networks. They illustrate some of the challenges of working with research partners in implementing the research as designed, but also the ways in which the combination of different methods and data can still yield meaningful results despite some minor setbacks. See the toolkit entry on Experiments for further background on experimental research design. Originally published February 1, 2016 Mobile Money Uptake, Savings and Social We have several reasons to believe that spillovers from Networks in Tanzania: A Lesson in Methods savings exist and have important policy implications. Conceptually, positive spillovers may arise because By Alfredo Burlando, Cynthia Kinnan, Silvia Prina savings access appears to raise the incomes of adopters, and a long line of economic research indicates that income gains may be shared among social networks. 144 Consumer Finance Research Toolkit Mobile money facilitates transfers, and therefore could also facilitate this type of spillover. The case for negative spillovers is more nuanced, but also plausible. Existing economic theory indicates that savings could encourage “shielding” cash from social networks. In practice, we have evidence from sub-Saharan Africa that households appear to attempt to avoid “kin taxes,” (see Baland et al. 2011; Dupas and Robinson 2013; Jakiela and Ozier 2012) and savings accounts may play a role in this avoidance. Given this premise, we hypothesize that different members of a social network may experience different effects. “Altruistic” ties (those characterized by “I want to share”) may experience positive benefits while “obligatory” ties (characterized by “I have to share” may experience negative effects). If obligatory ties tend to be poorer to start with, inequality may widen and, in turn, inequality may aggravate negative effects through a vicious cycle. Once we decided on the hypothesis, we set on creating a randomized control trial that would allow us to generate, measure and ultimately study these spillover effects. At the time of writing this blog, we have Our work began in September of 2013. Our research completed all aspects of the trial and are working on partners in Tanzania advertised a community meeting the analysis. where a study tied to savings accounts was carried out. We then surveyed those individuals who showed up to Our intervention: Zantel’s EZY-Pesa in Tanzania the meeting and had expressed interest in opening a mobile money account. We collected baseline data on In our research, we distributed mobile savings accounts their income, occupation, and other socioeconomic to approximately 1,500 participants in 33 rural and peri- markers. In addition, we collected information about urban areas of Zanzibar (Tanzania), and then studied their “key network partners”: financial partners, the effects of adoption of these accounts on the friends, advice-givers, etc. The key partners were adopters’ social networks through follow-up interviews. ranked by respondent on the basis of questions such as This work was done in partnership with Zantel, “Imagine you got a surprise payment of Tsh 100,000 Zanzibar’s leading mobile phone operator whose (~$65). Who would you most want to keep this secret mobile money product, called EZY-Pesa, allows from? Who least?” This way, we were able to identify deposits and withdrawals through local retailers. At the other people in the community who were likely to be start of our study, this product was not widespread either an “altruistic” tie, or an “obligatory” tie. outside Zanzibar Town, and Zantel agreed to implement a marketing campaign tied to the study. Following the interviews, we coordinated a mass marketing campaign with Zantel in October through 145 Consumer Finance Research Toolkit December 2013. We randomly assigned some of our Despite this setback, we find that the marketing visits baseline respondents (“focal respondents”) to receive were effective at signing up new users: 40% of visited help signing up for EZY-Pesa. The marketing visit treatment individuals opened an account. This is higher included a one-on-one explanation on how to operate than take-up rates of microfinance (e.g. Banerjee et al. EZY-Pesa to accumulate savings. Importantly, 2015). Of those who signed up, 52% used the account marketers were not allowed to talk about the transfer “actively” (two or more times), a figure similar to features of EZY-Pesa. Finally, to encourage usage, we findings for traditional savings accounts in Kenya also made a Tsh 2000 (~$1.25) initial deposit in their (Dupas and Robinson 2013). new EZY-Pesa accounts. The remaining households were randomly assigned to receive Tsh 2000 in cash, no marketing visit, and no sign-up assistance. In summer 2014, we identified baseline respondents (“focal households”) plus one tie we deemed to be “altruistic” and one tie we deemed to be “obligatory”. Ultimately, we were able to gather information from a large number of participants—over 1,500 focal individuals and 2,500 of their social ties. In this endline survey we collected information on demographics, informal transfers, and demand for savings. Respondents were also asked whether they received a marketing visit in October through November from Fortunately, a much smaller proportion of visited Zantel, and whether they opened a mobile account. We control individuals (24%) and of visited ties (19%) used the answers to these two questions to determine opened a savings account. We interpret this to mean how well the marketing intervention went. that many Zantel marketers indeed declined to sign up those not formally assigned to the treatment. Because Stumbling blocks on the road to analysis of this, we do observe a statistically significant (albeit small—only 5%) difference between the fraction of all Marketers were instructed to enroll in EZY-Pesa only treatment individuals opening an account and the rest those focal individuals who were randomly assigned to of our respondents. the treatment. Unfortunately, the marketing visits were not targeted as agreed, complicating our efforts to The road ahead study their causal impacts. We found that only one quarter of treatment individuals reported being visited Just because marketers did not follow the instructions by Zantel representatives; moreover, one fifth of focal provided does not mean we cannot uncover the individuals assigned to the control reported receiving a spillover effects. By pairing each focal individual who visit. (A much smaller fraction, 13%, of ties reported a received a marketing visit with another focal individual visit from marketers). Since there is not much who was skipped by the marketers but has very similar difference in the likelihood that a focal person received characteristics, we can use “matching” techniques to a marketing visit, we could not use standard methods of analysis for experiments. 146 Consumer Finance Research Toolkit study both positive and negative spillovers on the field experiment in Kenya. American Economic paired ties. Journal: Applied Economics 5(1), 163–92. While we now working through this analysis having Jakiela, P. and Ozier, O. (2015) Does Africa Need a derived certain lessons from our intervention. First, Rotten Kin Theorem? Experimental Evidence from (mobile) savings are not for everyone: of a group that Village Economies. Review of Economic Studies. expressed interest in EZY-Pesa, 40% of those who received a marketing visit signed up for an account and Prina, S. (2015) Banking the Poor via Savings among these only half of those who signed up used the Accounts: Evidence from a Field Experiment. Journal account actively. While there is an active discussion on of Development Economics, 115: 16-31. the importance of reducing fees and entry costs, our study suggests that reducing those costs to zero is not enough. Costs are not the only (or perhaps even the http://blog.imtfi.uci.edu/2016/02/mobile-money- main!) barrier to adoption. uptake-savings-and-social.html Second, we think that our partnership with Zantel had an important benefit: having done almost 700 Project Page: marketing visits was certainly a big accomplishment, Year: 2013 and it increased “external validity” (and hence the How does Mobile Money affect adopters’ scalability of the intervention). On the other hand, the social networks? scale did come at a cost: there was noncompliance in http://www.imtfi.uci.edu/research/2013/burland the way marketers approached local communities, and this reduces the study’s “internal validity” (i.e. ability to o_kinnan_prina.php estimate causal impacts without additional assumptions). It will be interesting to see if future studies will emerge that can address the latter. Works Cited Baland, J.M. , Guirkinger C., and Mali, C. (2011). Pretending to Be Poor: Borrowing to Escape Forced Solidarity in Cameroon, Economic Development and Cultural Change, 60(1), 1 - 16. Banerjee, A. V., E. Duflo, R. Glennerster, and C. Kinnan (2015). The miracle of microfinance? Evidence from a randomized evaluation. American Economic Journal: Applied Economics 7(1): 22-53. Dupas, P. and J. Robinson (2013). Savings constraints and microenterprise development: Evidence from a 147 Consumer Finance Research Toolkit Blog Post 7 that in the last 15 years, there have probably been This project also serves as an example of hundreds of frauds in Mexico which led poor savers to experimental research design (See the lose all the money they deposited in financial toolkit section, Experiments). It nicely institutions. Researchers have also noted that poorer illustrates a natural experiment, where a clients who received assistance in surmounting the initial costs of opening bank accounts often ended up product has already been introduced but using the account to merely withdraw money from certain cause-and-effect relationships can be transfers and let the account remain idle the rest of the tested using existing data sets. The research time. Along with transaction costs, low trust frequently described here generated new research figures as the main reason cited by the poor for not questions and hypotheses about people s using savings account. Overall, this state of affairs motivations behind account use, levels of might help us to explain why there is little savings in trust and distrust, and the relation between formal accounts. traditional and new savings techniques with In our study we were interested in finding out whether the account. These new questions could be ATM debit cards and mobile banking could potentially further examined, for instance, using alleviate this problem since these technologies ethnographic, interview and financial diaries potentially lower the cost of monitoring movements in methods. the account and simultaneously increase convenience and access to savings. Ideally, this could be tested using a randomized experiment where we award savings accounts with and without ATM cards randomly. Originally published March 28, 2016 Unfortunately, such a randomized experiment does not yet exist in the given context. Thus, in our study, we Using the ATM Debit Card to Build Trust and look at the effect of giving debit cards to the Savings: A Study through Mexico's Oportunidades beneficiaries of a Mexican state conditional cash transfer program Progresa/Oportunidades by Enrique Seira, from the study "Paying Conditional (now Prospera). The beneficiaries had already received Cash Transfer Programs in Bank Accounts" their transfer at a government development bank Bansefi but had not been given a debit/ATM debit Trust is an essential element of economic transactions. card connected to the account. For the study, we have This is especially true for savings where transactions used account level information on more than 300,000 take the form of a promise to future returns. accounts some of which received an ATM debit cards Unfortunately trust in financial institutions appears low in a staggered fashion. across the world and even more so among the poor and less educated. In Mexico, for instance, 25 percent of those with primary school education admitted to having "no trust at all" in banks, while 18 percent of those with more than primary school education expressed some trust in the banks (Gallup World Values Survey). This is not entirely inexplicable, given 148 Consumer Finance Research Toolkit checked 57 percent less frequently per bimester. This evidence strongly suggests that initial use of ATM debit cards to frequently check and monitor savings account helps to build up trust. Additionally, the increase in saving was gradual and coincided with the increase in directly elicited trust in the bank account and was not related to the learning curve for the debit card use. We found that in the groups using the debit cards, the savings increased dramatically, almost tripling in a span of 2 years after receiving the ATM card vis-à-vis the control group (without the ATM card). We noted that the increase in trust was contemporaneous with increases in savings amount in the bank account. Photo: Will Kay, Flicker, Creative Commons http://bit.ly/1OVvdtK In addition, the use of a consumption-income survey showed that the increase in savings in the account At the beginning of the research, we were intrigued in comes from new savings and not just from savings part, by the stories of Oportunidades beneficiaries who shifting. We found that although income did not complained about money disappearing from their change, consumption decreased after getting the ATM account. We found that this was due to the substantial card. One interpretation of this finding is that the fees they were incurring by frequently checking their account allows for some commitment to saving and account balances. This was so particularly in the avoidance of "temptation goods" like high sugar foods beginning when they did not have any trust in the that would otherwise be consumed. In our final report banks and resorted to making sure that their money we lay out some alternative explanations such as was there by checking their accounts several times in a mechanical savings where saving increases related to single day. Since there was a transaction cost involved more frequent but lower amount withdrawals, and in the process, the beneficiaries lost a substantial differential changes due to changes in the amount of amount of their funds in this way. We focused on the the Oportunidades transfer. increase in the savings account of the group that had received the debit cards. For purposes of comparison, All in all these findings indicate good news. Trust and we also had a control group that had not received any how to increase trust has not received enough attention debit cards as a part of the program. For analysis, we in the academic literature. Our results suggest that an undertook a differences-in-differences empirical design existing and simple technology – namely the debit card to approximate the causal effect of receiving the card could increase trust, account use, and savings. on savings in the account. Furthermore our findings show that in contexts such as the one examined in this study, the experiment can be As mentioned in the beginning, we found that when easily scaled up. Increasingly, the world over, tens of they first got the debit card, customers checked their millions of poor households in dozens of countries saving balance frequently, but the frequency of this currently receive cash transfer programs, more and checking declined over time. Most notably, more into bank accounts. Given the possible benefits, beneficiaries with more than 6 months with the card 149 Consumer Finance Research Toolkit providing ATM cards for these accounts appears to be a worthwhile measure worth ensuring. The final report of the study on how the ATM card has an effect on trust, savings, and the use of formal savings accounts can be found here. http://blog.imtfi.uci.edu/2016/03/using-atm-debit- card-to-build-trust-and.html Project Page: Year: 2014 Paying Conditional Cash Transfer Programs in Bank Accounts http://www.imtfi.uci.edu/research/2014/seira_ag uilar.php 150 Consumer Finance Research Toolkit Blog Post 8 In this post we see how study design can play close attention to generational and gender differences in mobile money use. Researchers interviewed and surveyed Tigo money users, service point attendants, and professional experts. We include samples of their research instruments in part II of the appendix as examples of how to design research instruments for specific target populations included in a study, as well as the For this research project, we chose to focus on the informed consent process. The acceptance of the mobile money service "Tigo Money" researchers conclude their post by noting by the rural population of Bolivia and the effects of this a regulatory change that occurred after the new technology on the development of rural areas. completion of their study. Legal and 40% of the Bolivian rural population lives in extreme poverty on less than US$1.25 per day. However, regulatory changes can have important despite the country's poor infrastructure and dispersed consequences for studies of new consumer population, an estimated 98% of Bolivians have access finance products like mobile money. In the to mobile technology. It may be for this reason that, in context of this study, new research might January of 2013, Tigo--a brand of the international track how the mobile operator is adapting telecommunications and media company Millicom-- to an altered competitive environment, or launched its nationwide mobile money platform Tigo how users negotiate changing fee Money in order to allow for further “financial inclusion that will allow overcoming of barriers and distance structures. between people, especially among rural populations.” Our research area consists of two municipalities-- Originally published May 4, 2015 Urmiri and Chayanta--located in the northern area of Potosi, one of Bolivia’s nine departments. The Hand Held Wealth?: A Case of Tigo Money in municipalities chosen have two of the highest indices Bolivia of poverty in all of Bolivia and most of their economic By IMTFI Researchers Maria Isabel Balderrama and activity is derived from subsistence agriculture and/or Oscar Rocabado livestock. These two municipalities are also fairly difficult to access: The inhabitants of Urmiri must travel one hour and forty-five minutes by bus through very rough terrain to reach Potosi, the nearest urban center, while the inhabitants of Chayanta are thirty minutes away by bus from Llallagua, a medium-sized 151 Consumer Finance Research Toolkit mining town. We were able to gather information from However, Bolivian Tigo Money users face a big hurdle these two communities regarding their use of Tigo in that Tigo's mobile service coverage does not reach Money through participant observation and by the most rural areas of Bolivia, including Urmiri and conducting over 500 surveys and 69 interviews with its Chayanta. The government-owned telecommunications inhabitants over a period of four weeks. We also spoke company ENTEL, on the other hand, is mandated by to Tigo Money Service Point attendants and General Law 164 (2011) to provide coverage to all professional experts in the areas of communication and areas of Bolivia, both urban and rural. The cost of a technology in Bolivia. phone call via ENTEL is US$0.17 per minute and the cost of a text message is US$0.02. For this reason, the Tigo Money: How does it work? Tigo Money users in our study almost always utilize What makes Tigo Money appealing to the Bolivian ENTEL as a carrier when engaging in Tigo Money rural population is that its users do not have to be transactions rather than Tigo's own mobile service. formally affiliated or fill out any paperwork, something Family members located in the urban centers of Bolivia that requires a great deal of time and effort. Tigo let their relatives in Urmiri and Chayanta know about Money only requires its clients to present their national the money transaction through a phone call or a text identification cards and visit one of Tigo Money’s 1,174 message sent to their ENTEL cell phones, and then the service points, 39% of which are located in rural areas recipient travels to his or her nearest Tigo Money point throughout the country. and picks up the money. As a result, Tigo describes Tigo Money as “a mobile payment system that allows [customers] to transfer money in a simple and secure way through a mobile wallet that is activated through a cell phone.” It is via this mobile wallet that users may transfer mobile money to other mobile wallets and then convert the electronic money into physical and legal tender, but in reality, users in rural areas see Tigo Money as mobile money service used primarily for receiving informal cash transfers. Women and Migration Our results show that 20% of respondents in the research areas (24% in Urmiri and 17% in Chayanta, We found that most of the recipients of payments respectively) utilize Tigo Money to receive money. using Tigo Money are women. Why? The data show There was only one recorded case of a user utilizing that there is a relationship between Tigo Money, Tigo Money to both send and receive, but all of the migration, and care networks. In fact, we found that other respondents in our study reported having never there were very few women between the ages of 30 and used the service to send money to anyone, anywhere in 39 living in the research area, primarily due to the Bolivia. Of these, 100% use the service to receive internal migration flows from rural to urban centers. If money. these female migrants to urban centers are the ones to 152 Consumer Finance Research Toolkit send money, who receives it? Based on interviews and 28 of the 68 Tigo Money users whom we interviewed observations, we concluded that the recipients of this for this study indicated that their spending habits money are mainly women older than 39, namely improved since they began using Tigo Money: grandmothers left caring for the children of migrant parents. “Now, with the money my father sends we buy things for the family” (Female, 18) “I can spend in studies for my children thanks to the money I receive” (Male, 35) And 23 indicated that they felt that their saving habits improved: “I now save for when something really bad happens” (Female, 49) “I save more money for when I go to Potosi [City]” (Male, 35) Tigo Money vs. Financial Funds and Banks How have mobile money flows improved the lives of the people of Northern Potosi? In a number of ways: 87% of all Tigo Money users surveyed thought that Tigo Money’s fees were too high, with one respondent 1) An improvement in the eating habits of the recipient even suggesting that “the government should provide households through the diversification of their daily this service for free.” So now we must ask the question: diet related to having more income to spend on food Why Tigo Money and not formal banking? The other than the few staple foods locally available and research suggests that it might have to do with rural grown. inhabitants' overall distrust of formal banking coupled 2) An improvement in education. A significant number of with the lack of familiarity with new financial rules and interviewees pointed out that they use the money regulations. Tigo Money launched a very aggressive received through Tigo Money to pay for college. marketing campaign that has reached, with its flyers Others also pointed out that they are saving the money and advertising posters, even the most remote to send their children to school. communities of the Altiplano region, something that 3) Increased short and long term investing. In the short term, traditional banks have failed to do. commercial activities such as purchasing and reselling canned goods, pasta, flour, sugar, and others have been While we were in the midst of doing our research, Law incorporated into the economic activities of the 393: Financial Services Law was put into effect in 2014. communities' inhabitants. In the long term, the The law specifies that formal banking institutions must purchase of agricultural technology such as sprinkler make all of their services universally acceptable. systems and greenhouses aims to improve agricultural Consequently, banks must now be free for customers productivity and set a goal beyond subsistence farming. to use without incurring any additional maintenance 153 Consumer Finance Research Toolkit fees. Although clients need a minimum of US$8 to open an account with PRODEM--the bank with the most presence in rural areas--no minimum balance is needed to maintain an account. These changes mean that PRODEM, and similar formal banking entities, are just as convenient, if not better and perhaps cheaper than Tigo Money, as Tigo Money requires users to pay a fee for each transaction. It will be interesting to see how this new law affects the mobile money landscape in Bolivia in the years to come. Read more in Balderrama and Rocabado's Final Report Project Page: Year 2014 Hand Held Wealth? Mobile Money & Food Production in Rural Potosi http://www.imtfi.uci.edu/research/2014/bald errama_rocabado.php 154 Consumer Finance Research Toolkit Appendix II. Sample Research Instruments From the project, “Hand Held Wealth? Mobile Money and Food Production in rural Potosi” by M. Isabel Balderrama and Oscar Rocabado, IMTFI Final Report, June 1, 2015 http://www.imtfi.uci.edu/files/docs/2015/Balderrama%20and%20Rocabado%20Final%20Report.pdf Survey and Interview Templates QUESTIONNAIRE: TIGO MONEY USERS DATE ________________________ LOCATION ___________________________ NAME OF QUESTIONNAIRE-TAKER ______________________________________________ Mobile Money & Food Production in Rural Potosi Hello! My name is ______________, and I would like to ask you to participate in a research project about how you save, spend and store your money. This research is being supported by the University of California in Irvine, California, in the United States of America. The research is being conducted by Isabel Balderrama. I am from UMSA’S CIDES in La Paz. [Hand subject own business card, with UC Irvine IRB address, phone and email written on the back]. You may also contact the University of California if you have any questions about this research [Show back of card]. I would like to ask you a few questions from this questionnaire about the way you utilize, or the reason why you do not utilize Tigo Money, and if this service has been useful or not in bettering your quality of life. It will not take more than 15 to 20 minutes of your time. The only risk of completing this questionnaire is the possibility that if someone else gets a copy of it, they might learn about your saving and spending habits, and how you utilize Tigo Money. But I will not collect any information that identifies you by name, and we will make sure that no one has access to this information other than me and the researchers I am working with. Your participation in this research is totally voluntary, and you can refuse to participate if you wish. If you do not want to participate, please let me know and I will leave. You can also stop the questionnaire at any time, or refuse to answer any question as we go along. May I also photograph or take a videotape of you? If so, may I use these images if I present my research to other people in a lecture presentation or published book or article? Please sign the following photo/video release form. We hope that this research will help us better understand how people save, store and spend money in your community. May I ask you these questions on this questionnaire? YES NO START OF SURVEY 1. Respondent’s Gender MALE FEMALE 2. Age 3. Highest education level reached 4. Primary Occupation 5. Estimated monthly income 6. Number of family members in 155 Consumer Finance Research Toolkit household? 6. Do you have a TIGO phone line? YES NO 7. Do you use TIGO Money? YES NO 8. Why or Why Not? *If respondent answers “NO” this is the end of the survey. If he or she answers “YES” then proceed to question 8* 9. Do you send or receive money through TIGO SEND RECEIVE BOTH Money? *If respondent answers “SEND” fill out section “A.” If he or she answers “RECEIVE” then fill out section “B.” If “BOTH” fill out both sections.* SECTION A 10. Where do you send the RURAL LOCATION URBAN LOCATION money? 11. Who do you send the money FAMILY MEMBER FRIEND OTHER to? 12. Occupation of the recipient 13. How often do you send money? 14. What use does the recipient DON’T give the money, generally? KNOW 15. Amount you usually send? SECTION B 16. Where do you receive the money RURAL LOCATION URBAN LOCATION from? 17. Who do you receive the money FAMILY MEMBER FRIEND OTHER from? 18. Occupation of the sender 19. How often do you receive money? 20. What use do you generally give the money? 21. Amount you usually receive? SECTION C 22. Are you happy with the services TIGO Money provides? YES NO 23. Why or why not? 156 Consumer Finance Research Toolkit 24. Are you content with the costs related to using TIGO Money? YES NO 25. Do you have any other comments or observations? INTERVIEWS: TIGO MONEY USERS DATE ________________________ LOCATION ___________________________ NAME OF INTERVIEWER ______________________________________________ Mobile Money & Food Production in Rural Potosi Hello! My name is ______________, and I would like to ask you to participate in a research project about TIGO Money usage in Rural Potosi. This research is being supported by the University of California in Irvine, California, in the United States of America. The research is being conducted by Isabel Balderrama. I am from UMSA’S CIDES in La Paz. [Hand subject own business card, with UC Irvine IRB address, phone and email written on the back]. You may also contact the University of California if you have any questions about this research [Show back of card]. I would like to ask you a few questions from this questionnaire about Tigo Money usage, It will not take more than one hours of your time, probably less. Your participation in this research is totally voluntary, and you can refuse to participate if you wish. If you do not want to participate, please let me know and I will leave. You can also stop the questionnaire at any time, or refuse to answer any question as we go along. May I also photograph or take a videotape of you? If so, may I use these images if I present my research to other people in a lecture presentation or published book or article? Please sign the following photo/video release form. We hope that this research will help us better understand how people save, store and spend money in your community. May I ask you these questions on this questionnaire? YES NO May I also digitally record this interview? YES NO START OF INTERVIEW SECTION A – General Information 1. Respondent’s Gender MALE FEMALE 2. Age 3. Highest education level reached 4. Main occupation 5. Estimated monthly income 157 Consumer Finance Research Toolkit 6. Marital Status SECCION B – TIGO Money Use 7. How did you first hear about TIGO Money? 8. What was the main reason you started using TIGO Money? 9. Did you experience any difficulties accessing the services provided by YES NO TIGO Money? 9a. If yes, please explain. 10. Since you began using TIGO Money, your SPENDING GOT STAYED IMPROVED habits… WORST THE SAME 10a. Please explain. 11. Since you began using TIGO Money, your SAVING GOT STAYED IMPROVED habits… WORST THE SAME 11a. Please explain. 12. Since you began using TIGO Money, your levels of DECREASE NO INCREASED financial INVESTMENT... D CHANGE 12a. Please explain. 158 Consumer Finance Research Toolkit 13. Who makes the majority of financial decisions in your household? 14. Have you recommended TIGO Money to any of your acquaintances, YES NO friends or family members? 15. Do you have any final comments? 159 Consumer Finance Research Toolkit INTERVIEWS KEY INFORMANTS: TIGO MONEY EMPLOYEES DATE ________________________ LOCATION ___________________________ NAME OF INTERVIEWER ______________________________________________ Mobile Money & Food Production in Rural Potosi Hello! My name is ______________, and I would like to ask you to participate in a research project about TIGO Money usage in Rural Potosi. This research is being supported by the University of California in Irvine, California, in the United States of America. The research is being conducted by Isabel Balderrama. I am from UMSA’S CIDES in La Paz. [Hand subject own business card, with UC Irvine IRB address, phone and email written on the back]. You may also contact the University of California if you have any questions about this research [Show back of card]. I would like to ask you a few questions from this questionnaire about Tigo Money usage, It will not take more than one hours of your time, probably less. Your participation in this research is totally voluntary, and you can refuse to participate if you wish. If you do not want to participate, please let me know and I will leave. You can also stop the questionnaire at any time, or refuse to answer any question as we go along. May I also photograph or take a videotape of you? If so, may I use these images if I present my research to other people in a lecture presentation or published book or article? Please sign the following photo/video release form. We hope that this research will help us better understand how people save, store and spend money in your community. May I ask you these questions on this questionnaire? YES NO May I also digitally record this interview? YES NO START OF INTERVIEW SECTION A – General Information 1. Respondent’s Gender MALE FEMALE 2. Age 3. Highest education level reached 4. Position at TIGO Money 5. Estimated monthly income SECTION B – QUESTIONS ABOUT TIGO MONEY USAGE 6. About how many people utilize Tigo Money? 7. How much money gets transferred per day on average? 160 Consumer Finance Research Toolkit 8. Do you assist your clients by teaching them how to utilize TIGO Money’s services and answering any questions they may have? 9. On average, do you usually have enough cash on hand to be able to make the payments out to customers in a timely manner? 10. Can you tell me of any difficulties that you are aware of that customers have run into while trying to access TIGO Money’s services? 11. What are the main advantages to using TIGO Money, from your point of view? 12. From your point of view, are URBAN – RURAL – RURAL – URBAN – most transactions… URBAN RURAL URBAN RURAL 13. Is coverage adequate in rural areas where TIGO Money is utilized? YES NO 13a. If not, what, if anything, is being done to correct this situation that you are aware of? 14. Has TIGO Money usage grown in the past six months? By how much? 15. Do you think TIGO Money is accessible to people who are low-income earners? 16. What is your opinion on the transactions costs related with the usage of TIGO Money? 17. Do you think, generally speaking, that TIGO Money is easy for people to use? 161 Consumer Finance Research Toolkit 18. Why do you think people choose TIGO Money over other money transfer services? 19. Do you have any other comments? 162 Consumer Finance Research Toolkit INTERVIEWS KEY INFORMANTS: ACADEMIC EXPERTS DATE ________________________ LOCATION ___________________________ NAME OF INTERVIEWER ______________________________________________ Mobile Money & Food Production in Rural Potosi Hello! My name is ______________, and I would like to ask you to participate in a research project about TIGO Money usage in Rural Potosi. This research is being supported by the University of California in Irvine, California, in the United States of America. The research is being conducted by Isabel Balderrama. I am from UMSA’S CIDES in La Paz. [Hand subject own business card, with UC Irvine IRB address, phone and email written on the back]. You may also contact the University of California if you have any questions about this research [Show back of card]. I would like to ask you a few questions from this questionnaire about Tigo Money usage, It will not take more than one hours of your time, probably less. Your participation in this research is totally voluntary, and you can refuse to participate if you wish. If you do not want to participate, please let me know and I will leave. You can also stop the questionnaire at any time, or refuse to answer any question as we go along. May I also photograph or take a videotape of you? If so, may I use these images if I present my research to other people in a lecture presentation or published book or article? Please sign the following photo/video release form. We hope that this research will help us better understand how people save, store and spend money in your community. May I ask you these questions on this questionnaire? YES NO May I also digitally record this interview? YES NO START OF INTERVIEW SECTION A – General Information 1. Respondent’s Gender MALE FEMALE 2. Name 3. Age 4. Position SECTION B – QUESTIONS ABOUT MOBILE TECHNOLOGIES IN RURAL LATIN AMERICA AND RURAL BOLIVIA 5. In your opinion, do you think the current socioeconomic and regulatory environment is conducive to the introduction and/or expansion of new mobile money technologies? 163 Consumer Finance Research Toolkit 6. What impact do issues such as marketing, user education, literacy and cultural mistrust of formal financial services have on the uptake of mobile-money services in the rural Bolivia and in Latin America in general? 7. What is your general opinion on the importance of the adoption of new mobile banking services by low-income rural populations throughout Bolivia? 8. Do you think mobile-money services have or will gain much momentum in rural Bolivia, and, if so, what could be or is the catalyst/s that can stimulate growth? 9. What is your outlook for mobile-money services in rural Bolivia? In Latin America? 11. What are the main advantages to using TIGO Money, from your point of view? 164 Consumer Finance Research Toolkit 19. Do you have any other comments and/or observations? 165 Consumer Finance Research Toolkit Timeframe of research activities PROCEDURES CONDUCTED  August 4th, 2014: Start of the research project; Initial team meeting  August 22 – August 29: First visit to research areas; Initial meetings with key informants; Pilot interviews  August 24 & 27: Team meeting with key informants to assign tasks for upcoming events; Suggestions on how to improve research tools based on the results of the pilot interviews  September 1st: Second visit to research area: Chayanta  September 2 – September 7: Quantitative and qualitative data collection on-site  September 5: Team meeting to discuss observations  September 7: Third visit to research area: Urmiri  September 8 – September 14: Quantitative and qualitative data collection on-site  September 12: Team meeting to discuss observations  October 6: Fourth visit to research area: Chayanta  October 7 – October 12: Quantitative and qualitative data collection on site  October 10: Team meeting to discuss observations  October 12: Fifth visit to research area: Urmiri  October 13 – October 20: Quantitative and qualitative data collection on site  October 17: Team meeting to discuss observations  October 22 – November 3: Completion of Semi-annual Technical Report  November 4 – November 28: Collection of interviews with, key informants and academic experts  October 22 – November 30: Data analysis; preparation medium term report  December 1: Fifth official team meeting to discuss progress on final report and prepare presentation for the IMTFI Conference in Irvine, CA  January 1, 2015 – May 31, 2015: Data analysis, report writing  June 1, 2015: Presentation Final Report; End of project 166 Consumer Finance Research Toolkit Activities Aug Sep Oct Nov Dec Jan Feb Mar Apr May Initial team meeting Development of research tools (Surveys & Interviews) Collection of interviews with key informants First visit to research areas/Initial meetings with key informants; pilot interviews Second visit to research area: Chayanta Quantitative and qualitative data collection on-site Third visit to research area: Urmiri Quantitative and qualitative data collection on-site Fourth visit to research area: Chayanta Quantitative and qualitative data collection on-site Fifth visit to research area: Urmiri Quantitative and qualitative data collection on-site Data analysis and completion of Semi-annual Technical Report Preparation of final report and project completion 167 Consumer Finance Research Toolkit 2014; See also: MRA Code of Marketing Footnotes Research Standards http://www.marketingresearch.org/issues- policies/mra-code-marketing-research- standards and “Ethical Considerations of Marketing Research,” by Terry Masters, Houston Chronicle 1 The CFRM Project is funded by the http://smallbusiness.chron.com/ethical- Institute for Money, Technology, and Financial considerations-marketing-research-43621.html Inclusion and by the Canela Group. Erin B. 10 Taylor and Gawain Lynch serve as project According to the AAA code of ethics, leads. Information can be found on the IMTFI’s researchers must be open and honest about website, http://www.imtfi.uci.edu/cfrmp their work. It is unethical for researchers to 2 The Belmont Report: Ethical engage in misleading or deceptive behavior to Principles and Guidelines for the Protection of collect data, and does not comply with Human Subjects of Research, The National standards of informed consent of human Commission for the Protection of Human subjects. (http://ethics.aaanet.org/ethics- Subjects of Biomedical and Behavioural statement-2-be-open-and-honest-regarding- Research, 1978, your-work/) Alternatives to mystery shopping http://videocast.nih.gov/pdf/ohrp_appendix_bel could include product testing and participant mont_report_vol_2.pdf observation of enrollment, usage, and provider See also: interactions with customers when researching new consumer finance products or services. http://www.hhs.gov/ohrp/humansubjects/guida 11 Research Ethics and Integrity for nce/belmont.html Social Scientists: Beyond Regulatory 3 http://oprs.usc.edu/review/typesofirb/ Compliance by Mark Israel, SAGE 4 “Confidentiality and Informed Publications, 2014. Consent: Issues for Consideration in the 12 https://www.epicpeople.org/ Preservation of and Provision of Access to 13 See, for example, “Loan Supply, Qualitative Data Archives” by Louise Corti, Credit Markets and the Euro Area Financial Annette Day and Gill Backhouse, Forum: Crisis” by Carlo Altavilla, Matthieu Darracq Qualitative Social Research 1(3), Paries and Giulio Nicoletti, European Central http://www.qualitative- Bank, Working Paper No.1861, October 2015, research.net/index.php/fqs/article/view/1024/2 https://www.ecb.europa.eu/pub/pdf/scpwps/ec 207 bwp1861.en.pdf 5 “Snowball Recruitment—Active and 14 For example, the Dutch Payments Passive,” Research Support, University of Association aggregates data on use of debit Sydney, cards and cash based on information provided http://sydney.edu.au/research_support/ethics/ by their members. See their website, human/guidelines/snowballing.shtml http://www.betaalvereniging.nl/en/ 6 See: CITI Program 15 See Charmaine 'Ilaiu Talei’s https://www.citiprogram.org/index.cfm?pageID discussion of presenting findings to research =22 participants, ”Understanding the 7 See: NIH Human Subjects Protection transformative value of Tongan women’s kau Training tou lālanga: mobile mats, mobile phones, and https://grants.nih.gov/grants/policy/hs/training. money transfer agents,” IMTFI blog, January htm 22, 2015 8 See WIRB Services http://blog.imtfi.uci.edu/2015/01/understanding https://www.wirb.com/Pages/Default.aspx -transformative-value-of.html 9 See CASRO Code of Standards and 16 Bitcoin is pseudo-anonymous Ethics because users have public addresses that http://www.casro.org/?page=TheCASROCode 168 Consumer Finance Research Toolkit anyone can see, but it is very hard to deduce experiences-product-innovation-and-the- who or what controls a public address. potential-for-poverty-alleviation 17 For an overview, see “The Emerging 26 Livelihood and Microfinance: Science of Computational Anthropological and Sociological Perspectives Anthropology,”Technology Review, 10 June on Savings and Debt by Hotze Lont and Otto 2014, Hospes, Eburon Publishers, 2004. http://www.technologyreview.com/view/52821 27 Migration and Remittances During the 6/the-emerging-science-of-computational- Global Financial Crisis and Beyond, edited by anthropology/ Ibrahim Sirkeci, Jeffrey H. Cohen, and Dilip 18 CGAP’s work on the topic is Ratha, The World Bank, 2012. introduced in “What Human-Centered Design 28 “Why Big Data needs Thick Data,” by Means for Financial Inclusion” by Yanina Tricia Wang, Ethnography Matters, 2013. Seltzer and Claudia McKay, CGAP Blog, 16 https://medium.com/ethnography-matters/why- October 2014, big-data-needs-thick-data- http://www.cgap.org/publications/what-human- b4b3e75e3d7#.bd0ay4ngb centered-design-means-financial-inclusion 29 You can read a description of his 19 So argues Sam Ladner in her book project and find links to his papers on the Practical Ethnography: A Guide to Doing IMTFI website at Ethnography in the Private Sector, Left Coast http://www.imtfi.uci.edu/research/2010/fikre.ph Press, 2014. p. Also see “Mobile Money Information System 20 For an overview, see “‘Poor People Architecture for Open Air Market in Ethiopia” Aren't Making Ends Meet': Inside Payday by Woldmarian Fikre Mesfin, IMTFI, 2014, Lending” by Jana Kasperkevik, The Guardian, http://www.imtfi.uci.edu/files/docs/2014/mobile 4 November 2013, s_in_open_air_market_report.pdf http://www.theguardian.com/money/2013/nov/ 30 "Mobile Money Information System 03/poor-people-arent-making-ends-meet- Architecture for Open Air Market in Ethiopia" inside-payday-lending by Woldmariam Fikre Mesfin, IMTFI website, 21 See, for example, “From Long-Term 2014, p.4, Savings to Instant Mortgages: Financial http://www.imtfi.uci.edu/files/docs/2014/mobile Demonstration and the Role of Interaction in s_in_open_air_market_report.pdf Markets” by Zsuzsanna Vargha, Organization, 31 W. Mesfin, personal communication, March 2011, 18(2). pp. 215-235. March 2015. 22 As described in the book 32 David Stoll is a Professor of Materializing Poverty: How the Poor Anthropology at Middlebury, USA. His staff Transform Their Lives by Erin Taylor, profile is located at AltaMira, 2013. http://www.middlebury.edu/academics/soan/fa 23 For information on a range of culty/node/25831. David’s book on debt is El projects, see the IMTFI’s website, Norte Or Bust!: How Migration Fever and http://www.imtfi.uci.edu/. Microcredit Produced a Financial Crash in a 24 Mutual Life, Limited: Islamic Banking, Latin American Town, Rowman & Littlefield, Alternative Currencies, Lateral Reason by Bill 2012. Maurer, Princeton University Press, 2011. 33 D. Stoll, personal communication, 25 Kustin has written about her work in a April 2015. short article, “The Promise of Islamic Finance: 34 Ibid. User Experiences, Product Innovation, and the 35 “Using Structured Interviewing Potential for Poverty Alleviation,” Impatient Techniques,” Program Evaluation and Optimists, Bill & Melinda Gates Website, 10 Methodology Division, United States General October 2015, Accounting Office, June 1991, http://www.impatientoptimists.org/Posts/2015/ http://www.gao.gov/assets/80/76090.pdf. 10/The-promise-of-Islamic-finance--user- 36 “Good with Money: Getting By Silicon Valley,” FAIR Money, 169 Consumer Finance Research Toolkit https://fairmoneytest.files.wordpress.com/2015 edited by Jing Jian Xiao, Springer, 2008, /03/fair-money_good-with-money- pp.271-285. report_final_3-15-15.pdf 45 Ibid., p.275-276. 37 “Moneylenders and Their 46 Ibid., p.281. Customers,” Social Policy Research 68, 47 “Snowball Recruitment—Active and December 1994. Passive,” Research Support, University of 38 “Payments and Money Transfer Sydney, Behavior of Sub-Saharan Africans” by http://sydney.edu.au/research_support/ethics/ Johanna Godoy, Bob Tortora, Jan human/guidelines/snowballing.shtml Sonnenschein, and Jake Kendal1, Gallup, 48 “Money, Coercion, and Undue June 2012, Inducement: Attitudes about Payments to http://www.gallup.com/poll/155132/payments- Research Participants” by Emily A. Largent, money-transfer-behavior-sub-saharan- Christine Grady, Franklin G. Miller, and Alan africans.aspx Wertheimer, IRB: Ethics & Human Research 39 “Hidden in a Coke Bottle: Modernity, 34(1), 2012, pp.1-8. Accessed through The Gender and the Informal Storing of Money in Hastings Center, Philippine Indigenous Communities” by Janet http://www.thehastingscenter.org/Publications/ M. Anardo, IMTFI Working Paper 2012-5, IRB/Detail.aspx?id=5705&utm_source=Eurek http://www.imtfi.uci.edu/files/blog_working_pa Alerts&utm_medium=email&utm_campaign=P pers/2012-5_arnado.pdf R022212IRB 40 “Mobile Banking and Economic 49 “The 12 Cognitive Biases That Development: Linking Adoption, Impact, and Prevent You from Being Rational” by George Use” by Jonathan Donner and Camilo Tellez, Dvorsky, io9, 2013, Asian Journal of Communication 18(4), http://io9.com/5974468/the-most-common- pp.318-332, cognitive-biases-that-prevent-you-from-being- http://www.jonathandonner.com/donner_tellez rational _mbanking_use.pdf 50 This case study is drawn from 41 “"Micro-finance and Development of Kusimba’s IMTFI working paper, a Rural Women: A Mixed Methods Approach" by conversation with her, and her mobile money Anand Jerard Sebastine and P. Ilango, Quality video; see also “Family Networks of Mobile Enhancement of Social Work Profession in Money in Kenya” by Sibel Kusimba, Yang India: Issues and Challenges, Allied Yang, and Natesh V. Chawla, Information Publishers, 2009, pp.511-516. Technology in International Development, https://www.academia.edu/3275077/Micro- 2015, 11(3), pp.1-21. finance_and_Development_of_Rural_Women 51 S. Kusimba, personal communication, _A_Mixed_Methods_Approach August 2015. 42 "Choosing Relevance: How Credit 52 “'My Story Has No Strings Attached': Unions Can Harness Transparency and Show Credit Cards, Market Devices, and a Stone Impact" by Bruce B. Cahan, Filene Institute, Guest" by José Ossandón, IMTFI Working Publication #346 (12/14), Paper 2014-3, https://filene.org/research/report/choosing- http://www.imtfi.uci.edu/files/blog_working_pa relevance-how-credit-unions-can-harness- pers/2014- transparency-and-show-impa 3_Ossandon_Working%20Paper%202.pdf 43 For a brief description of cultural 53 “Household Organization and Budget domain analysis, see Wikibooks, Structures in a Purepecha pueblo" by James https://en.wikibooks.org/wiki/Cultural_Anthrop M. Acheson,American Ethnologist, Vol. 23, ology/Anthropological_Methods#Domain_anal No. 2 (May, 1996), pp.331-351. ysis 54 “Insights Into Action: What Human- 44 “Financial Behavior of Hispanic Centred Design Means for Financial Americans” by Kittichai Watchravesringkan, Inclusion," CGAP, October 2014, Handbook of Consumer Finance Research, 170 Consumer Finance Research Toolkit http://www.cgap.org/publications/what-human- Heather A. Horst and Erin B. Taylor, The centered-design-means-financial-inclusion Australian Journal of Anthropology, 2014, 55 Lived Economies of Default: 25(2), pp.155-170. Consumer Credit, Debt Collection and the 68 Portfolios of the Poor: How the Capture of Affect by Joe Deville, Routledge, World's Poor Live on $2 a Day by Daryl 2015. Collins, Jonathan Morduch, Stuart Rutherford, 56 “Understanding the Transformative and Orlanda Ruthven, Princeton University Value of Tongan Women’s Kau Tou lālanga: Press, 2009, p.187. Mobile Mats, Mobile Phones, and Money 69 The U.S. Financial Diaries website is Transfer Agents" by IMTFI researcher located at http://www.usfinancialdiaries.org/ Charmaine 'Ilaiu Talei, IMTFI Blog, 22 January 70 See the Microfinance Opportunities 2015, website, http://blog.imtfi.uci.edu/2015/01/understanding https://www.microfinanceopportunities.org/3- -transformative-value-of.html approach/looking-for-something/how/financial- 57 “Money Talks: Tracking Personal diaries/ Finances” by Joseph “Jofish” Kaye, Mary 71 See the SALDRU website, McCuistion, Rebecca Gulotta, and David A. https://www.saldru.uct.ac.za/projects/complete Shamma, Proceedings of the 32nd Annual d-projects/financial-diaries-project ACM Conference on Human Factors in 72 “Financial Diaries with Smallholder Computing Systems, 2014, pp.521-530, Families" by Jamie Anderson and Wajiha http://dl.acm.org/citation.cfm?id=2556288.255 Ahmed, CGAP, 25 February 2016, 6975 http://www.cgap.org/publications/financial- 58 The screener they sent to participants diaries-smallholder-families is available at 73 “An Overview of the Kenya Financial http://lessonsfromhome.org/materials/07/kaye- Diaries Research Program," FSD Kenya, 13 screener.pdf October 2015, http://fsdkenya.org/financial- 59 The study protocol is available at diaries/ http://lessonsfromhome.org/materials/07/kaye- 74 “How Do the Poor Handle Money? study-protocol.pdf What Do the Financial Diaries of Char 60 “The Financial Tour: Methods for Dwellers Tell Us About Financial Inclusion?" Studying Sensitive Financial Questions" by by Kuntala Lahiri–Dutt and Gopa Samanta, Jofish Kaye, in Studying and Designing ASARC Working Paper 2012/16, Technology for Domestic Life: Lessons from https://crawford.anu.edu.au/acde/asarc/pdf/pa Home, edited by Tejinder K. Judge and pers/2012/WP2012_16.pdf Carman Neustaedter, Morgan Kaufmann, 75 “Advancing Integrated Microfinance 2014, pp.111-134. for Youth (AIM Youth): Financial Diaries 61 Ibid. Research Report" by Megan Gash and Bobbi 62 J. Kaye, personal communication, Gray, Freedom from Hunger, May 2014, November 2014 https://www.freedomfromhunger.org/sites/defa 63 Ibid. ult/files/documents/AIMyouth_FinancialDiaries 64 Ibid. _Eng.pdf 65 Ibid. 76 “Mexico Financial Diaries Reveal How 66 Ibid. Households use Credit to Stretch Budgets and 67 For a brief article describing the Bridge Gaps in Income” CFI Blog, March 15, study, see “What do the things you carry say 2016. http://cfi-blog.org/2016/03/15/mexico- about you?” by Erin B. Taylor, PopAnth, 7 financial-diaries-reveal-how-households-use- April 2014, http://popanth.com/article/what-do- credit-to-stretch-budgets-and-bridge-gaps-in- the-things-you-carry-say-about-you; For a full- income/ length article, see “The Role of Mobile Phones 77 See “Tailoring Formal Financial in the Mediation of Border Crossings: A Study Products to the Poor” by Daryl Collins and of Haiti and the Dominican Republic” by Claudia McKay, CGAP Blog, 22 October 2012, 171 Consumer Finance Research Toolkit http://www.cgap.org/blog/tailoring-formal- 92 Alexandra Mack generously told us financial-products-poor about her work in personal communications. 78 Thanks to Julie Zollmann for this As well as her role at Pitney Bowes, she is insight. Treasurer at the Ethnographic Praxis in 79 Daryl Collins, “Why Financial Diaries Industry Conference (EPIC), to understand the needs of Smallholders?” https://www.epicpeople.org/people/ CGAP Blog, April 15, 2014 93 A. Mack, personal communication, http://www.cgap.org/blog/why-financial-diaries- November 2014. understand-needs-smallholders 94 A. Mack, personal communication, 80 ibid. April 2015. 95 “New CGAP Report provides unique 81 An overview of research biases by Martyn Shuttleworth can be found at insights into the financial lives of smallholder Explorable: Psychology Experiments, families,” CGAP Blog, February 25, 2016. https://explorable.com/research-bias http://www.cgap.org/news/new-cgap-report- 82 Portfolios of the Poor, p. 4 smallholder-diaries 83 See Portfolios of the Poor: How the 96 “Computer Aspects of Technological World's Poor Live on $2 a Day by Daryl Change, Automation, and Economic Progress” Collins, Jonathan Morduch, Stuart Rutherford by Paul Armer, Rand Corporation, 1966, and Orlanda Ruthven (2009, Princeton https://www.rand.org/content/dam/rand/pubs/p University Press). See also the Portfolios of apers/2013/P3478.pdf the Poor website, 97 Building Bankomat: The Development http://www.portfoliosofthepoor.com/ of On-line, Real-time Systems in British and 84 Ibid., p.186 Swedish Savings Banks, c. 1965-1985 by 85 Ibid., p.204. Bernardo Bátiz-Lazo, Tobias Karlsson, and 86 “When Saving for Tomorrow Björn Thodenius, 2009, Necessitates Borrowing for Today” by Julie http://papers.ssrn.com/sol3/papers.cfm?abstra Zollman, CGAP Blog, 28 August 2014, ct_id=1734004 http://www.cgap.org/blog/when-saving- 98 "A Literature Review on Digital tomorrow-necessitates-borrowing-today Transformation in the Financial Service 87 “The Kenya Financial Diaries: 4 Industry" by Timo Cziesla, BLED 2014 Flaws in the Social Network" by Julie Proceedings, https://domino.fov.uni- Zollmann, CGAP Blog, 27 August 2014, mb.si/proceedings.nsf/0/9e0b8c345e951a62c http://www.cgap.org/blog/kenya-financial- 1257cf5002cce7c/$FILE/03_Cziesla.pdf diaries-4-flaws-social-network 99 See, for example, Joe Deville’s book 88 J. Zollmann, personal Lived Economies of Default: Consumer Credit, communication, December 2015. Debt Collection and the Capture of Affect 89 J. Zollman, personal communication, (2015, Routledge). For an early study of credit December 2015. cards and spending, see "Credit Cards as 90 “Confidentiality and Informed Spending Facilitating Stimuli: A Conditioning Consent: Issues for Consideration in the Interpretation" by Richard A. Feinberg, Journal Preservation of and Provision of Access to of Consumer Research 13, 1986: 348-356. Qualitative Data Archives” by Louise Corti, 100 See his website, Annette Day and Gill Backhouse, Forum: http://www.jblumenstock.com/ Qualitative Social Research 1(3), 101 See the report “The Opportunities of http://www.qualitative- Digitizing Payments,” IBRD / The World Bank, research.net/index.php/fqs/article/view/1024/2 2014. 207 102 “A Conceptual Framework for 91 Research Ethics and Integrity for Examining Digital Inequality” by Lynette Social Scientists: Beyond Regulatory Kvasny, AMCIS 2002 Proceedings, Compliance by Mark Israel, SAGE http://aisel.aisnet.org/cgi/viewcontent.cgi?articl Publications, 2014. e=1605&context=amcis2002 172 Consumer Finance Research Toolkit 103 “Doing Digital Finance Right" by s/etudes/join/2011/464441/IPOL- Katharine McKee, CGAP Blog, 17 June 2015, IMCO_ET(2011)464441_EN.pdf http://www.cgap.org/publications/doing-digital- 112 “Getting to the Top of Mind: How finance-right Reminders Increase Saving" by Dean Karlan, 104 T. Boellstorff, personal Margaret McConnell, Sendhil Mullainathan, communication, August 2015. and Jonathan Zinman, NBER Working Paper 105 "E-Banking" by Jinkook Lee, Jinsook No.16205, 2010, Erin Cho, and Fahzy Abdul-Rahman, in http://www.nber.org/papers/w16205 Handbook of Consumer Finance Research, 113 “Understanding Financial Consumers 2008, pp.105-123. in the Digital Era: A Survey and Perspective 106 "Zap it To Me: The Short-Term On Emerging Financial Consumer Trends,” Impacts of a Mobile Cash Transfer Program" CGI, 2014. by Jenny C. Aker, Rachid Boumnijel, Amanda 114 See the Revelation website, McClelland and Niall Tierney, Center for https://www.focusvision.com/products/revelati Global Development Working Paper no.268, on/ 2011, 115 According to financial experts such as http://papers.ssrn.com/sol3/papers.cfm?abstra Blythe Masters and the Bank of England, ct_id=1931641 blockchain technology has the potential to 107 "South-South Migration and create an “Internet of finance” that will Remittances" by Dilip Ratha and William revolutionize still-clunky payments systems; Shaw, World Bank Working Paper No.102, see “Blythe Masters Tells Banks the Washington, DC, 2007, Blockchain Changes Everything” by Edward http://siteresources.worldbank.org/INTPROSP Robinson, Bloomburg, 1 September 2015, ECTS/Resources/334934- http://www.bloomberg.com/news/features/201 1110315015165/SouthSouthMigrationandRem 5-09-01/blythe-masters-tells-banks-the- ittances.pdf blockchain-changes-everything 108 See the ING International Survey, 116 For a guide to Bitcoin, see “Bitcoin http://www.ezonomics.com/ing_international_s Internals: A Technical Guide to Bitcoin” by urvey/sharing_economy_2015, and the Chris Clark, 2013. Ezonomics website, 117 Meiklejohn, S., Pomarole, M., Jordan, http://www.ezonomics.com/ G., Levchenko, K., McCoy, D., Voelker, G. M., 109 "Predicting poverty and wealth from & Savage, S. (2013). “A Fistful of Bitcoins: mobile phone metadata" by Joshua Characterizing Payments Among Men with No Blumenstock, Gabriel Cadamuro and Robert Names” by Sarah Meiklejohn, Marjori On, Sciencemag 350(6264): 1073-1076, Pomarole, Grant Jordan, Kirill Levenchko, November 27, 2015 Damon McCoy, Geoffrey M. Voelker, Stefan http://www.sciencemag.org/content/350/6264/ Savage, Proceedings of the Internet 1073.full.pdf?keytype=ref&siteid=sci&ijkey=jl1 Measurement Conference 2013 (IMC ’13), FOo2RaNJQk pp.127-140. 110 MoneyLab Reader: An Intervention in 118 Bitcoin is pseudo-anonymous Digital Economy, edited by Geert Lovink, because users have public addresses that Nathaniel Tkacz and Patricia de Vries, pp.244- anyone can see, but it is very hard to deduce 256. Amsterdam: Institute of Network who or what controls a public address. Cultures, 2015, 119 https://en.wikipedia.org/wiki/Public- http://networkcultures.org/blog/publication/mon key_cryptography eylab-reader-an-intervention-in-digital- 120 For an explanation of public keys and economy/ block chains, see 111 “Consumer Behaviour in a Digital https://en.wikipedia.org/wiki/Block_chain_(data Environment,” Directorate-General for Internal base) Policies, European Parliament, 2011, 121 S. Meiklejohn, personal http://www.europarl.europa.eu/RegData/etude communication, December 2015. 173 Consumer Finance Research Toolkit 122 Ibid. eport- 123 For a simplified explanation of how landscaping_mobile_media_in_indonesia.pdf account clustering heuristics work, see the 136 T. Boellstorff, personal authors’ five-page explanation, “A Fistful of communication, August 2015 Bitcoins: Characterizing Payments Among 137 “Landscaping Mobile Social Media Men with No Names,” ;login:, December 2013, and Payments in Indonesia” by Tom 38(3), pp. 10-14, Boellstorff, IMTFI, 2013, p.12, http://www0.cs.ucl.ac.uk/staff/S.Meiklejohn/file http://www.imtfi.uci.edu/files/docs/2013/final_r s/login13.pdf eport- 124 Meiklejohn et al, p.12. landscaping_mobile_media_in_indonesia.pdf 125 Ibid., p.2. 138 For a short overview of digital 126 Ibid., p.2. research ethics, see “Digital Research 127 CoinDesk maintains a Bitcoin ATM Ethics—Some Considerations” by Sarah- map at http://www.coindesk.com/bitcoin-atm- Louise Quinnell, Social Science Space, 24 map/ October 2011, 128 S. Meiklejohn, personal http://www.socialsciencespace.com/2011/10/di communication, December 2015. gital-research-ethics-%E2%80%93-some- 129 “Could a Bank Deny Your Loan considerations/. For a more comprehensive Based on Your Facebook Friends?” by review of using data, see “Guidelines for the Robinson Meyer, The Atlantic, 25 September Ethical Use of Digital Data in Human 2015, Research” by Karin Clark et al, The University http://www.theatlantic.com/technology/archive/ of Melbourne, 2015, 2015/09/facebooks-new-patent-and-digital- http://carltonconnect.com.au/wp- redlining/407287/ content/uploads/2015/06/Ethical-Use-of- 130 “How Big Data Could be Used to Digital-Data.pdf Predict a Patient's Future” by Wayne Parslow, 139 "On the Early History of Experimental The Guardian, 17 January 2014, Economics" by Alvin E. Roth, Journal of the http://www.theguardian.com/healthcare- History of Economic Thought, 15, Fall, 1993, network/2014/jan/17/big-data-nhs-predict- pp.184-209; see also Experimental illness Economics: Rethinking the Rules by Nicholas 131 “City Watchdog to Look Into the Use Bardsley, Robin Cubitt, Graham Loomes, of ‘Big Data’ by Insurers” by Michael Bow, Peter Moffatt, Chris Starmer and Robert Independent, 25 November 2015, Sugden, Princeton University Press, 2010. http://www.independent.co.uk/news/business/ 140 See Randomized Controlled Trial news/city-watchdog-to-look-into-the-use-of- (RCT), Better Evaluation, big-data-by-insurers-a6747716.html http://betterevaluation.org/plan/approach/rct 132 http://www.datascienceassn.org 141 See Daryl Collins discussion of RCTs in 133 http://www.datascienceassn.org/code-of- relation to the financial diary method: conduct.html http://www.cgap.org/blog/why-financial-diaries- 134 understand-needs-smallholders http://www.amazon.com/gp/product/33192842 142 Evaluating Financial Products and 07/ref=as_li_qf_sp_asin_il_tl?ie=UTF8&camp Services in the US: A Toolkit for Running =1789&creative=9325&creativeASIN=331928 Randomized Controlled Trials (2015, 4207&linkCode=as2&tag=pop08f- Innovations for Poverty Action) 20&linkId=OQVN7I2S2UUWHWGD https://www.poverty- 135 “Landscaping Mobile Social Media action.org/sites/default/files/publications/Evalu and Payments in Indonesia” by Tom ating%20Financial%20Products%20and%20S Boellstorff, Khanis Suvianita, Dédé Oetomo, ervices%20in%20the%20US%20- and Nurul Ilmi Idrus IMTFI, 2013, %20A%20Toolkit%20for%20Running%20RCT http://www.imtfi.uci.edu/files/docs/2013/final_r s%20(1).pdf 174 Consumer Finance Research Toolkit 143 ibid. p. 7. They go on to note, "Finance Princeton University, 2013, related RCTs are unique in that they seek to https://www.princeton.edu/~joha/publications/ not only answer questions related to Haushofer_Shapiro_UCT_2013.pdf improvements in financial well-being for 154 “Commitment Savings in the customers, but also to identify how these Philippines” by Nava Ashraf, Dean Karlan and solutions can improve a financial institution's Wesley Yin, Innovations for Poverty Action, bottom line” (8). 2000-2003, 144 Ibid. p. 11 http://www.povertyactionlab.org/evaluation/co 145 For an excellent comparison of mmitment-savings-products-philippines methods, see Table 1, p. 10, in the IPA toolkit. 155 “Liability Structure in Small-Scale 146 ibid. p. 33. See also the handout, Finance: Evidence from a Natural Experiment” “What is a Randomized Control Trial?” in the by Fenella Carpena, Shawn Cole, Jeremy IPA toolkit appendix. Shapiro, and Bilal Zia, The World Bank 147 See the IPA toolkit for more detailed Economic Review 27(3), 2013, pp.437-469, and technically specific guidelines on setting http://hbswk.hbs.edu/item/liability-structure-in- up an RCT. small-scale-finance-evidence-from-a-natural- 148 “Insurance, Credit, and Technology experiment Adoption: Field Experimental Evidence from 156 “Consumption and Debt Response to Malawi” by Xavier Giné and Dean Yang, Unanticipated Income Shocks: Evidence from Journal of development Economics 89(1), a Natural Experiment in Singapore” by Sumit 2009, pp.1-11. Agarwa and Wenlan Qian, American http://siteresources.worldbank.org/DEC/Resou Economic Review, 104(12), pp.4205-30, rces/Insurance_Credit_and_Technology_Adop https://www.aeaweb.org/articles.php?doi=10.1 tion_Malawi.pdf 257/aer.104.12.4205 149 For an explanation of a variety of 157 “Insurance, Credit, and Technology methods including natural experiments, see Adoption: Field Experimental Evidence from Simply Psychology, Malawi” by Xavier Giné and Dean Yang, http://www.simplypsychology.org/experimental Journal of development Economics 89(1), -method.html 2009, pp.1-11. 150 “Liability Structure in Small-Scale https://ideas.repec.org/a/eee/deveco/v89y200 Finance: Evidence from a Natural Experiment” 9i1p1-11.html by Fenella Carpena, Shawn Cole, Jeremy 158 “Microinsurance Demand After a Shapiro, and Bilal Zia, The World Bank Rare Flood Event: Evidence From a Field Economic Review 27(3), 2013, pp.437-469, Experiment in Pakistan” by Ginger Turner, http://hbswk.hbs.edu/item/liability-structure-in- Farah Said and Uzma Afzal, The Geneva small-scale-finance-evidence-from-a-natural- Papers on Risk and Insurance—Issues and experiment Practice 39, 2014, pp.201-223, 151 Accessible at http://www.palgrave- https://else.econ.ucl.ac.uk/orsee/public/ journals.com/gpp/journal/v39/n2/full/gpp20148 152 “Innovative and Interactive Ways to a.html Improve the Financial Capability and Savings 159 “The Social Dilemma of of Women” by Deepti Kc and Mudita Tiwari, Microinsurance: A Framed Field Experiment IMTFI, 2015, on FreeRiding and Coordination" by Wendy http://www.imtfi.uci.edu/files/docs/2015/IMTFI Janssens and Berber Kramer, Tinbergen %20Report- Institute Discussion Paper TI 2012-145/V, %20Women%20study_KC%20and%20Tiwari_ http://papers.tinbergen.nl/12145.pdf Final.pdf 160 “Micro-Savings and Informal 153 "Household Response to Income Insurance in Villages: How Financial Changes: Evidence from an Unconditional Deepening Affects Safety Nets of the Poor, A Cash Transfer Program in Kenya" by Natural Field Experiment” by Jeffrey Flory, Johannes Haushofer and Jeremy Shapiro, The Milton Friedman Institute for Research in 175 Consumer Finance Research Toolkit Economics, MFI Working Paper Series May 2013, No.2011-008, http://www.slate.com/articles/health_and_scie http://bfi.uchicago.edu/RePEc/bfi/wpaper/BFI_ nce/science/2013/05/weird_psychology_social 2011-008.pdf _science_researchers_rely_too_much_on_we 161 “Insurance Versus Savings for the stern_college.html Poor: Why One Should Offer Either Both or 172 See for example Building Women’s None” by Andreas Landmann, Björn Vollan Economic and Social Empowerment through and Markus Frölich, IZA Discussion Paper No. Enterprise: An Experimental Assessment of 6298, 2012, http://ftp.iza.org/dp6298.pdf the Women’s Income Generating Support 162 Anonymous reviewer’s comment (WINGS) Program in Uganda, by Christopher 163 See “A Toolkit for the Evaluation of Blattman, Eric Green, Jeannie Annan and Financial Capability Programs in Low and Julian Jamison, Logica Study Series No. 1, Middle-Income Countries,” World Bank, April 2013 http://www.poverty- http://documents.worldbank.org/curated/en/20 action.org/sites/default/files/publications/buildi 13/01/18054632/toolkit-evaluation-financial- ng-womens-economic-and-social- capability-programs-low-middle-income- empowerment.pdf countries 173 IPA toolkit, p. 23 164 Due Diligence: An Impertinent Inquiry 174 “Risk Preferences Are Not Time Into Microfinance by David Roodman, CGD Preferences” by James Andreoni and Charles Books, 2012. Sprenger, The American Economic Review 165 IPA toolkit, ibid. 102(7), 2012, pp.3357-3376. 166 See IMTFI fellows Alfredo Burlando, 175 Ibid., p.9-10 Cynthia Kinnan, and Silvia Prina’s discussion 176 Ibid., p.26. of spillover effects and challenges of design 177 “Ethics and Experiments” by Karen A. implementation in their project blog post, Hegtvedt, in Laboratory Experiments in the “Mobile Money Uptake, Savings, and Social Social Sciences, edited by Murray Webster Networks in Tanzania: A Lesson in Methods,” and Jane Sell, Elsevier, 2007, pp.141-172. IMTFI blog, February 1, 2016 178 “Microfinance Games” by Xavier http://blog.imtfi.uci.edu/2016/02/mobile- Giné, Pamela Jakiela, Dean Karlan and money-uptake-savings-and-social.html Jonathan Morduch,American Economic 167 IPA toolkit, pp. 34-35 Journal: Applied Economics 2(3), 2010, pp.60- 168 See “A Toolkit for the Evaluation of 95. Financial Capability Programs in Low and 179 Ibid., p.1. Middle-Income Countries,” World Bank, 180 Ibid., p.33. http://documents.worldbank.org/curated/en/20 181 Social Collateral: Women and 13/01/18054632/toolkit-evaluation-financial- Microfinance in Paraguay’s Smuggling capability-programs-low-middle-income- Economy by Caroline E. Schuster, University countries of California Press, 2015. 169 For an explanation of the Hawthorne 182 Effect, see http://research.boisestate.edu/complia http://psychology.about.com/od/hindex/g/def_h nce/institutional-review-board-irb- awthorn.htm home/guidelines-for-researchers/guidelines- 170 “Public Goods Experiments Without for-investigators-using-snowball-sampling- Confidentiality: A Glimpse Into Fund-Raising” recruitment-methods/ by James Andreoni and Ragan Petrie, Journal 183 See for example of Public Economics 88(7), 2004, pp.1605- https://www.womensworldbanking.org/news/bl 1623. og/if-technology-is-so-great-why-hasnt-it- 171 “Psychology Is WEIRD: Western done-more-for-womens-financial-inclusion/ College Students Are Not the Best “184 Innovative and Interactive Ways to Representatives of Human Emotion, Behavior, Improve the Financial Capability and Savings and Sexuality” by Bethany Brookshire, Slate, 8 of Women” by Deepti Kc and Mudita Tiwari, 176 Consumer Finance Research Toolkit IMTFI Final Report, 2015 198 Central banks and other financial http://www.imtfi.uci.edu/files/docs/2015/IMTFI institutions routinely compensate for gaps in %20Report- their data by conducting qualitative research, %20Women%20study_KC%20and%20Tiwari_ especially in times of rapid change. See Final.pdf “Economy of words” by Douglas Holmes, 185 Cultural Anthropology 24.3: 381-419, 2009. http://www.imtfi.uci.edu/research/2014/tiwari_k 199 The Fortune at the Bottom of the c.php Pyramid: Eradicating Poverty Through Profits 186 “Savings Constraints and by C.K. Prahalad, Wharton School Publishing, Microenterprise Development: Evidence from 2010. a Field Experiment in Kenya” by Pascaline 200 See How Would You Like to Pay? Dupas and Jonathan Robinson, American How Technology is Changing the Future of Economic Journal: Applied Economics 5 (1): Money by Bill Maurer, Duke University Press 163-192, 2013 2015; see also Monetary Ecologies and https://www.aeaweb.org/articles.php?doi=10.1 Repertoires: Research from the Institute for 257/app.5.1.163 Money, Technology and Financial Inclusion, 187 Kc and Tiwari, ibid. p. 7 January 2010, 188 bid. p. 16 http://www.imtfi.uci.edu/files/articles/IMTFI_Fir 189 ibid. pp. 17-18 stAnnualReportDesignPrinciples.pdf 190 ibid. p. 19 201 See, for example, “Scaling mobile 191 Deepti Kc, personal communication money” by Ignacio Mas and Dan Radcliffe, 192 The Impact of Financial Literacy Journal of Payments Strategy & Systems 5.3 Training for Migrants, CReAM Discussion (2011): 298-315. Paper Series, No 16/2012, by John Gibson, 202 The Transnational Capitalist Class by David McKenzie and Bilal Zia Leslie Sklair, Wiley-Blackwell Publishing, https://ideas.repec.org/p/crm/wpaper/1216.htm Oxford, UK, 2001. l 203 Reported by the World Bank, 193 Dupas and Robinson, ibid. See also http://web.worldbank.org/WBSITE/EXTERNAL Banking the Poor via Savings Accounts: /TOPICS/0,,contentMDK:21924020~pagePK:5 Evidence from a Field Experiment, by Silvia 105988~piPK:360975~theSitePK:214971,00.h Prina, Journal of Development Economics, tml 2015, 115: 16-31 204 See Jeremaiah Opiniano and Alvin Ang’s http://econpapers.repec.org/article/eeedeveco/ on-going research of remittance investment in v_3a115_3ay_3a2015_3ai_3ac_3ap_3a16- the Philippines: 31.htm; What are the Headwaters of Formal http://www.imtfi.uci.edu/research/2014/opinian Savings? Experimental Evidence from Sri o_ang.php Lanka. NBER Working Paper 20736, 2014 by 205 For a recent analysis of the unbanked Michael Callen, Suresh de Mel, Craig and under-banked in the US and how banks McIntosh and Christopher Woodruff can design products and services differently http://www.hks.harvard.edu/index.php/content/ for these segments, see the Oliver Wyman download/73265/1670499/version/1/file/Callen and Ideas 42 report, Reimagining Financial _289.pdf Inclusion (Nov 2015) 194 Deepti Kc, personal communication, http://www.oliverwyman.com/content/dam/oliv March 2016 er-wyman/global/en/2015/nov/Reimagining- 195 See for example Financial-Inclusion-Final.pdf http://nextbillion.net/evaluation-that-creates- 206 See “Online insurance” by Robert N. value-for-participants-a-client-centric- Mayer, Handbook of Consumer Finance approach/ Research, edited by Jing Jian Xiao, Springer, 196 Kc, personal communication, ibid. 2008, pp.271-285. 197 ibid. 207 Due Dilligence: An Impertinent Inquiry into Microfinance, by David Roodman, Center 177 Consumer Finance Research Toolkit for Global Development, 2012. http://www.amazon.com/Due-Diligence- Impertinent-Inquiry- Microfinance/dp/1933286482 208 "Taking On Debt: Perspectives from across the Social Sciences" by Erin B. Taylor and Gawain Lynch, IMTF Blog, 24 & 26 August 2015, http://blog.imtfi.uci.edu/2015/08/consumer- finance-research-global.html 178 Consumer Finance Research Toolkit