(PDF) The Ricardian equivalence hypothesis: evidence from Bangladesh

(PDF) The Ricardian equivalence hypothesis: evidence from Bangladesh
Academia.edu uses cookies to personalize content, tailor ads and improve the user experience.
By using our site, you agree to our collection of information through the use of cookies.
To learn more, view our
Privacy Policy.
About
Press
Papers
We're Hiring!
Outline
Title
Abstract
Introduction
Variables and Sources of Data
Conclusions
References
All Topics
Economics
Macroeconomics
The Ricardian equivalence hypothesis: evidence from Bangladesh
Jalal Siddiki
2010, Applied Economics
November 29, 2024
visibility
description
35 pages
Sign up for access to the world's latest research
check
Get notified about relevant papers
check
Save papers to use in your research
check
Join the discussion with peers
check
Track your impact
Abstract
This paper examines the Ricardian equivalence hypothesis (REH) and its sources of failure in the case of Bangladesh using various theoretical specifications, annual data from 1974-2001 and linear and non-linear time series techniques. The general findings tend to invalidate the REH: a finite time horizon and the presence of liquidity-constrained individuals are the sources of deviation from the REH. Empirical results reveal that real per capita private consumption (C) under various specifications is cointegrated generally at the 5% level with real per capita income (Y), government expenditure before and after interest rate repayments (G & G2), taxes (T) and the interest rate (r). Results reveal that an increase in G, G2, T and r reduces C and that that an increase in budget deficits raises trade deficits. These results highlight the importance of fiscal policies in boosting private consumption and controlling trade deficits, which are the prime goals of stabilisation policies being followed by Bangladesh.
Related papers
An Empirical Analysis of Fiscal Deficit and Inflation in Bangladesh
IJMRAP Editor
IJMRAP, 2024
This paper empirically examines the impact of fiscal deficit on inflation in the context of Bangladesh. The question addressed isare the methods of financing the fiscal deficit of Bangladesh inflationary? There are several options available to the hand of government to finance its budget deficit including bank borrowing, non-bank borrowing, and printing money. The government of Bangladesh uses all the three sources. The government's borrowing from the central bank is of special interest in this research paper. It is found by analyzing data from 1973 to 2018 that there exists a long-run relationship among inflation, fiscal deficit (FD), and total bank borrowing (TBB) as there exist more than 1 cointegrating equation. In the Johansen Cointegration Test, it is found that FD is negatively and TBB is positively related to inflation. However, FD is not significant at a 5% level of significance whereas TBB is significant at that level. It is found through the vector error correction model that changes in FD and TBB are absorbed by inflation every 8 months.
Download free PDF
View PDF
chevron_right
Exchange Rate, Interest Rate and Ricardian Equivalence Evidence from Pakistan
Muhammad Waqas
Romanian journal of economic forecasting
The aim of this study is to investigate the validity of Ricardian Equivalence Hypothesis in Pakistan by using annual data from 1973-2010. In our previous studies we investigated the Ricardian Equivalence Hypothesis and its sources of deviation in Pakistan by using Structural Consumption function and Euler Equation approach. The present study investigated this hypothesis in terms of rate of interest and exchange rate. Results of Auto Regressive Distributed Lag cointegration investigate the long run relationship among variables. By using OLS technique this study rejected Ricardian Equivalence Hypothesis in case of Pakistan. The findings of the study demonstrate the focus towards the significance of fiscal policies in raising private consumption and controlling budget deficits.
Download free PDF
View PDF
chevron_right
The Fiscal and Current Account Imbalances: An Empirical analysis of the Twin Deficits Hypothesis in Bangladesh
Muntasir Murshed
2019
This paper aims to analyze the possibility of the twin deficits hypothesis existing in the context of Bangladesh using annual data from 1980 to 2017. Vector Error-Correction approach is tapped to estimate the short and long run coefficients while the pairwise Granger causality analysis is employed to understand the long run causal associations. The results suggest that budget and current account balances in Bangladesh behave as distant cousins rather than twins as perceived from a reverse causality that is found to be stemming from budget deficit to current account deficit. Moreover, budget deficit is found to deteriorate the national trade balance in Bangladesh.
Download free PDF
View PDF
chevron_right
Government debt and budget deficit nexus in Pakistan: Evidence from Ricardian Equivalence Hypothesis
Neelma Shamsi
We checked the validity of Ricardian Equivalence Hypothesis in Pakistan by using structural consumption and saving function. By using ordinary least square method we checked the restrictions of Ricardian equivalence hypothesis that has been rejected by the vale of Wald test.
Download free PDF
View PDF
chevron_right
Testing the Ricardian Equivalence Theorem: Time Series Evidence from Turkey
Ahmet Salih İKİZ
Economies, 2020
Two of the most common measures adopted by the government to stimulate the economy are increasing government borrowings and implementing tax cuts. These tax cuts are financed through increased debt. According to the Ricardian equivalence theory, the consumers will not change their current spending when they anticipate a tax increase in the future. In order to pay high taxes in the future, the government should increase its present savings. However, the extent of applicability of Ricardian equivalence could vary across nations. In this context, the present study explores the long-running relationship between domestic borrowing and private savings in Turkey. For this purpose, the researcher collected the data for key variables, gross domestic savings, and government debt, for the period of 1980–2017. The researcher used unit root, cointegration, VECM, and the Granger causality test to examine the relationships among the variables. Apart from this, ARDL regression was used in order to ...
Download free PDF
View PDF
chevron_right
Testing the Ricardian equivalence hypothesis in the case of Ethiopia: An autoregressive-distributed lag approach
Balázs Kotosz
Hungarian statistical review, 2020
The Ricardian equivalence hypothesis (REH) suggests that when the government attempts to stimulate the economy by raising debt-financed government spending, consumption and demand do not increase but rather remain the same. The objective of this study is to test the existence of the REH in Ethiopia, using annual data from 1990 to 2011 and by employing the autoregressivedistributed lag cointegration approach. The study includes three variables (budget deficit, government consumption expenditure, and government debt) which contribute to the REH along with another variable. The results show that only the budget deficit and government consumption expenditure fulfil the REH. However, government debt fails to fulfil it. Thus, limited evidence of the existence of the REH is found in Ethiopia.
Download free PDF
View PDF
chevron_right
Examining the effect of expenditure-changing and expenditure-switching policies on the real GDP in Bangladesh
Sonia Afrin
This study examines the effects of expenditure-changing and expenditure-switching polices on the real GDP in Bangladesh. The study uses annual data set over the period of 1986 to 2015. This study uses the DF-GLS test to check whether a series suffers from a unit root problem. The Johansen maximum likelihood approach is incorporated to identify whether a long-run equilibrium relationship exists among the time series variables. The structural stability of the model is identified by CUSUM and CUSUM square tests. It is found that there is a long-run relationship among the considered time series variables and the model is structurally stable. The paper finds that expansionary fiscal policy and depreciation of the domestic currency positively affect the real GDP in Bangladesh. Therefore, the economy of Bangladesh can be stimulated by taking an expansionary fiscal policy and depreciating domestic currency. However, the paper does not support the monetarist hypothesis that higher money supply leads to higher real GDP. Therefore, the monetarist hypothesis deserves further rigorous empirical investigation in the context of Bangladesh.
Download free PDF
View PDF
chevron_right
Two Steps of Ricardian Equivalence: An Evidence in Indonesia
agus Ferdianto
2019
Ricardian Equivalence was analyzed through the influence of deficit fiscal policy on consumption, the fiscal deficit policy of Soeharto regime and reform had a difference, the difference lied in the policy as outline in the government budgets preparation. Soeharto regime applied budget concept of the T-account model, while the reform applied budget concept of I-account. The differences of concept budgetting affected the deficit fiscal policy, while the component of deficit fiscal policy consisted of budget deficits, government expenditure, government debt, and tax revenue. The purpose of this paper is to determine the effect of the differences in Soeharto regime and reform toward deficit fiscal policy component, as well as to determine the effect of deficit fiscal policy to domestic consumption, and determine the Ricardian Equivalence phenomenon and formation. Based on these objectives, an analysis tool used was Ordinary Least Squared (OLS) and Two Stage Least Squared (TSLS), with t...
Download free PDF
View PDF
chevron_right
Ricardian Equivalence, Twin Deficits and Feldstein-Horioka Hypotheses: Empirical Analysis form Pakistan
Ayesha Zahid
Kashmir Economic Review, 2018
The study attempts to investigate relationship between both budget and current account deficits for the Pakistan using annual data from 1976 to 2016. In this regard, three hypotheses, Ricardian Equivalence (RE), Twin Deficits (TD) and the Feldstein-Horioka (F-H) are tested under the framework of Multivariate Cointegration techniques (Johansen, Autoregressive distributed lag) and Error correction modeling. The cointegration results reject the RE hypothesis and support the validity of TD hypothesis for Pakistan. It implies that debt-finance tax cut raised the interest rate which attracted the capital inflow. The findings of error correction model also support the short run linkage between both deficits and confirm that long run relationship exists between them, when interest and exchange rates are included in the model. The results of cointegration test further reveal a positive relationship between the current account deficit and investment in the long run. The long run coefficient of investment indicates the high degree of foreign capital mobility and rejects the Feldstein-Horioka hypothesis. However, positive and less than one value of short-run coefficient of investment provide weak support for the Feldstein-Horioka hypothesis. So, the empirical results suggest that investment (coefficient of foreign capital inflow) has a significantly positive impact on current account deficit only in the long run.
Download free PDF
View PDF
chevron_right
An Empirical Investigation of Budget and Trade Deficits: The Case of Bangladesh
Asrafuz zaman
International Journal of Economics and Financial Issues, 2013
This study attempts to investigate the evidence of twin deficits hypothesis, that is, the simultaneous existence of fiscal and foreign trade account deficits in Bangladesh. In this regard this paper examines the causal relationship between budget deficit and trade deficit based on annual time series data from 1972-73 to 2011-12 fiscal year. Applying the VAR and Granger Causality after successfully running ADF and PP unit root test and cointegration rank test this paper reveals that budget deficit Granger cause trade deficit and vice versa but the relationship does not stand for the long-run dynamics. In fact, the long run relationship depends on overall macroeconomic environment and performances of other relevant variables. Policy implication is that Government should reduce budget deficit to improve the trade account balance.
Download free PDF
View PDF
chevron_right
The Ricardian Equivalence Hypothesis: Evidence from Bangladesh

By

Jalal U. Siddiki*, **

* School of Economics, Kingston University, Penrhyn Road, Kingston, Surrey KT1
2EE, UK. Tel:+44(0)208 547 2000 (ext. 62579); Fax: +44(0) 208 547 7388; E-mail:
[email protected]
**
I am particularly grateful to Paul Auerbach for his helpful comments on the
previous drafts of this paper. I am also thankful to Subrata Ghatak for his comments
on the previous drafts of this paper. The usual disclaimer applies.

Abstract

This paper examines the Ricardian equivalence hypothesis (REH) and its sources of

failure in the case of Bangladesh using various theoretical specifications, annual data

from 1974-2001 and linear and non-linear time series techniques. The general

findings tend to invalidate the REH: a finite time horizon and the presence of

liquidity-constrained individuals are the sources of deviation from the REH.

Empirical results reveal that real per capita private consumption (C) under various

specifications is cointegrated generally at the 5% level with real per capita income

(Y), government expenditure before and after interest rate repayments (G & G2),

taxes (T) and the interest rate (r). Results reveal that an increase in G, G2, T and r

reduces C and that that an increase in budget deficits raises trade deficits. These

results highlight the importance of fiscal policies in boosting private consumption and

controlling trade deficits, which are the prime goals of stabilisation policies being

followed by Bangladesh.

Key words: Ricardian Equivalence; Bangladesh; Cointegration analysis; Rational

Expectations.

JEL Classification: H61; H62; O10; O11.

1. Introduction

The main aim of this paper is to examine the Ricardian equivalence hypothesis (REH)

(explained below) and its sources of failure in Bangladesh, using various theoretical

specifications, annual data from 1974-2001 and time series techniques. Bangladesh is

a less developed country (LDC), which is associated with a low level of saving,

investment, per capita income and with a high rate of both fiscal and trade deficits 1

(Siddiki (2000, 2002), where both types of deficits generally move together (figure 1

in the appendix). The sustainability and the consequences of such high deficits are a

major concern for policy makers of developed and developing countries. Very few

studies (reviewed below) investigate the consequences of fiscal deficits on private

consumption and trade deficits in LDCs. As far as the present author is concerned, no

such study on Bangladesh has been carried out. This type of analyses is particularly

important for Bangladesh since it has been following stabilisation and structural

adjustment policies2, the success of which mainly depends on the nature of the

relationships between fiscal policies and private consumption and trade deficits. One

of the novelty of this paper is to investigate the sources of failure of the REH in a

developing country such as Bangladesh, which is still an under researched area. Thus,

the findings of this paper would be important for policy makers in Bangladesh and

other LDCs.

Economic theory and empirical evidence are also not decisive for drawing a

general conclusion on the consequences of fiscal deficits on private consumption and

the balance of payments despite the central focus of macroeconomic analyses

The average of fiscal deficits during our sample periods is about 6% of GDP, ranging from 3-9% of
GDP and the average of trade deficits is more than 7% of GDP, ranging from 3.54 - 12.51% of GDP.
These policies are prescribed by international institutions such as the World Bank and the
International Monetary Fund, and are based on demand management policies which suggest reducing
the budget deficits in order to reduce trade deficits and to increase private investment, thereby
increasing in income and consumption.

concerning the effect of fiscal deficits on macroeconomic variables. In addition, most

of empirical studies (reviewed below) examining the effects of fiscal deficits on

private consumption and on trade deficits are mainly concentrated on developed

countries with some few exceptions 3.

There are mainly two types of views on the consequences of fiscal deficits on

other macroeconomic variables. The Keynesian model predicts that a rise in fiscal

deficits increases aggregate demand which in turn stimulates short-run output and

employment, raises interest rates and also causes a crowding out in private

investment. The Keynesian proposition asserts that the increase in aggregate demand

caused by fiscal deficits also widens (reduces) current account or trade deficits

(surpluses), implying that taxes should be raised in order to reduce budget deficits and

therefore trade deficits4.

The REH, in contrast with the Keynesian proposition, states that it is

government purchases and marginal taxes rather than the ratio of debt to taxes that

have an impact on private consumption and on trade deficits. That is, the mode of

financing fiscal deficits, i.e. whether fiscal deficits are financed by debt or by tax

increases, is inconsequential in its effects upon private consumption and therefore

trade balances, since economic agents consider present period’s deficit financing as a

future period’s tax liability (Barro (1974, 1989)). The stability of saving and

For example, Ghatak and Ghatak (1996); Gupta (1992); Haque (1988); Khalid (1996).
The relationship, according to the twin deficits hypothesis, between government budget deficits and
trade deficits can be summarised as follows (see Khalid and Guan (1999) for a good review). Firstly, in
a Mundell-Flemming framework, an increase in government deficits are thought to exert an upward
pressure on real interest rates, which boosts capital inflows and hence causes an appreciation in real
exchange rates and a reduction in competitiveness, causing trade (or current account) deficits
(Rosenweig and Tallman (1993), p. 580; Khalid and Guan (1999), 390). This mechanism is effective
under both fixed and flexible exchange rate regimes. Under a fixed exchange rate regime, trade deficits
deteriorate due to a positive income effect caused by the government’s excess expenditures and due to
an appreciation in real exchange rates. Secondly, the Keynesian absorption theory predicts that a rise in
budget deficits increases domestic absorption and hence an expansion in imports causing current
account deficits (see from Khalid and Guan (1999), 390)). A strong correlation between saving and

investment is crucial in supporting the REH as instability in both factors may cause

both deficits not to be correlated. Thus, according to the REH, fiscal policies do not

affect the equilibrium level of trade balances, current account, interest rates, money

demand, private consumption, investment and saving (Vamvoukas (1999)) 5. The REH

is, however, based on some strong assumptions: (a) capital markets are perfect and the

consumer does not face any borrowing constraints; (b) both the private and public

sectors have the same planning horizons; (c) taxes are non-distortionary (Barro (1974,

1989)).

Two types of empirical investigation to examine the REH have been carried

out. One is the estimation of structural private consumption models to examine the

impact of government expenditures on private consumption (Becker (1997); Ghatak

and Ghatak (1996); Khalid (1996)). Empirical evidence on this issue is inconclusive

(see Leiderman and Blejer (1988) and Seater (1993) for a survey). Authors opposed to

the REH argue that the failures of the proposition are mainly caused by the violation

of its underlying assumptions. That is, the REH fails mainly due to finite time

horizons, non-altruistic or inoperative bequest motives, childless couples, liquidity

constraints and uncertainty (see Seater (1993)).

The other type of empirical investigation explores the consequences of budget

deficits on trade deficits; some support the assertion that a budget deficit causes a

trade deficit while many oppose it (Vamvoukas (1999); Normandin (1999) and

references are therein). It is also argued that a simple violation of the REH does not

necessarily imply that the Granger causality runs from budget deficits to trade deficits

or to private consumption (Normandin (1999)). The legitimacy of strict stabilisation

investment (Feldstein and Horioka (1980)) also causes budget deficits and the current accounts of the
balance of payments to move together, supporting the twin deficits hypothesis.
Note, however, that LDCs in general, and Bangladesh in particular, are characterised by imperfect
capital markets (Siddiki (2001, 2002); Auerbach and Siddiki (2002); Ghatak (1995)) .

policies is criticised when the magnitude of the Granger causality is negligible even

when the REH is not violated6.

This paper is organised as follows: section two explains the variables used in

our analysis and sources of data. Section three surveys various specifications of the

consumption functions which are used to test the REH and to find the sources of

departures, if any, from the REH. This section also explains the link between budget

and trade deficits. In section four, models are estimated and empirical results are

explained. Section five draws conclusions.

2. Variables and Sources of Data

In this section, we explain the variables, and their data sources, which are used in

specifying the REH (section three) and empirical modelling (section four). C is

private consumption, Y is gross domestic product (GDP), T is taxes, d is budget

deficits, TD is trade deficits, G (G2) is government expenditures excluding

(including) interest payments on government debt, GI is government investment

expenditures, RB interest payments on government debt; W is wealth and A is assets:

both W and A are defined as the sum of total broad money supply and deposits in

various government sponsored saving schemes; r is real interest rates, bank rates

minus the rate of inflation measured from the consumer price index. All variables but

r are expressed in real per capita natural logarithm terms (the GDP deflator with base

1990 is used).

Data sources: Bangladesh Bureau of Statistics (various years) Statistical

Yearbook of Bangladesh, Government of Bangladesh (various issues) Bangladesh

Economic Review, Bangladesh Bank (various issues) Economic Trends.

The violation or failure of the REH imply that a government can affect trade deficits or private
consumption by changing the timing of taxes.

3.1 Various Specifications of the REH and Crowding-out Hypothesis

There are mainly two types of consumption functions used in the literature to test the

REH. One is based on ad hoc, i.e. the Buiter and Tobin (1979), consumption

functions. The second type of consumption functions incorporates the rational

expectations hypothesis which assumes the availability of perfect information about

future government fiscal policies. That is, economic agents can predict future

government fiscal policies. This type of consumption functions is also used to find the

causes or sources of the failures, if any, of the REH. In addition, analyses on the

consequences of budget deficits on trade deficits are also used to test the REH. In this

section, we will review various theoretical specifications, which will be used in the

next section to test the REH in Bangladesh.

Various formulations of the Buiter-Tobin type approach for examining the

REH and crowding-out hypothesis used in the literature are summarised below (see,

Ghatak and Ghatak (1996) for a survey):

C t  a 0  a1Y t  a 2 G t  a 3 T t  a 4 W t (1)

C t  a(Y t  T t  d t ), 0  a 1 (2)

d  G t  RB t  T t (3a)

implies that

d  G 2t  T t (3b)

where the total government fiscal deficit (d) is the sum of primary deficits (G - T) and

interest payments (RB) on bonds; the expressions (3a) and (3b) state that an increase

in (G - T) and a resulting augmentation in RB raise d. Various forms of equation (2)

which incorporate expressions (3a) and (3b) and some other restrictions are used in

order to test the REH. For example, Buiter and Tobin (1979) estimated the following

equation:

C t  a 0  a1Y t  a 2 T t  a 3 d t (4)

subject to the following restrictions:

0  a1  1, a 2  0, a 3  0, a1  a 2 and a 2  a 3. (5)

The REH is confirmed if a1, a 2 and a 3 are statistically significant and the

restrictions in equation 5 are satisfied. The statistically significant coefficients and

equation 5 recapitulate the main assertion of the REH: the mode of financing fiscal

deficits - i.e. whether fiscal deficits are financed by debt or by tax increases - is

inconsequential on private consumption since economic agents consider present

period’s deficit financing as future period’s tax liabilities (Barro (1974, 1989)).

The restriction a 2  a 3 implies that the sign and magnitudes of the

coefficients for both taxes and government expenditures are the same: both taxes and

government expenditures exert the same effect on consumption. Rational agents with

perfect foresight, i.e. in the absence of uncertainty, would be inclined to believe that

deficits incurred by the government today will be completely offset by rising taxes in

the next period.

The coefficient a1 represents marginal propensity to consume and this, in

accordance with standard theory, is positive and less than one. The restriction

implying that the coefficients of income and taxes are equal but opposite in sign, i.e.,

a1  a 2 , indicates that consumption losses due to an imposition of taxes are equal to

consumption gains resulting from a same amount of increase in income or vice versa.

The coefficient of (Y-T) simply measures the impact of disposable income on C if the

restriction a1  a 2 is validated.

Incorporating equation (3b), Kormendu (1983) proposes the following

‘augmented consolidated approach’:

C t  a1Y t  a 2 T t  a 3 G 2 t (6)

A statistically insignificant a 2 implies that government deficits have no impact on

current consumption, lending support to the REH. This follows from the fact that the

consumption decisions of rational consumers depend on the present value of

government expenditures rather than on the timing of taxes (Barro (1989)). Using

expression (3b) and imposing the restriction that the coefficients of taxes and

government spending are equal, though opposite, in sign Boskin (1988) also provides

the following:

C t  a1(Y t  G2t )  a 2 d t (7)

a positive and statistically significant value of a 2 invalidates REH. To test REH and

crowding-out hypothesis, equation (2) can also be rewritten as

C t  a 0  a1Y t  a 2 G2t  a 3 RB t 0  a1  1. (8)

subject to the restrictions explained in equation (5) and as

C t  a1Y t  a 2 G2t 0  a1  1 . (9a)

A negative and statistically significant a 2 implies that government consumption

crowds out private consumption. The crowding out hypothesis asserts that an increase

in government expenditure or investment results in a reduction in private consumption

or expenditure. Deficit financing raises real interest rates, which in turn reduces

private or any other interest-sensitive form of private spending. Thus, we can write:

C t  a1Y t  a 2 r t  a3G2t 0  a1  1, a 2 , a3  0 (9b)

and C t  a1Y t  a 2 r t  a3GI t 0  a1  1, a 2 , a3  0 (9c)

where GI is government investment.

Incorporating the rational expectations proposition, Aschaur (1985) derived

the second type of consumption function, to test the REH, which maximises

intertemporal utility subject to a budget constraint (see also Gupta (1992) for a

review, pp. 20-21). Aschaur assumes that a representative household with a quadratic

utility function maximises the net present value of consumption in the current and

future periods. The author uses the following Eüler equation:

E t  1C *t  a  b C t  1 (10a)

where E is expectations operator and C* is the effective private consumption

described by

C *t  C t   G 2 t (10b)

where C is actual private consumption and G2 is government consumption. According

to equation (10b), government utilities influence private utilities and each unit of G2

is assumed to yield the same utility as  units of private spending. A positive value of

 implies that government spending is a substitute for private spending. On the other

hand, a negative value of  indicates government spending is a complement to private

spending7. Substitution of the lagged of equation (10b) into (10a) gives the following:

E t  1C *t  a  bC t  1  b G 2 t  1 (10c)

Assume that expectations are formed at time t-1 and taking the expectations of

equation (10b), then we can write:

E t  1C *t  C t   E t  1G 2t  C t  E t  1C *t   E t  1G 2t (10d)

Substituting equation (10c) into (10d) and incorporating the rational expectations

hypothesis, i.e. actual consumption is expected consumption plus a random error ut

which is purely a random walk, we obtain the following:

C *t  C t   G 2 t  C t  C *t   G 2 t  A positive value of  gives a negative coefficient for G2 and thus
implies that an increase in G2 reduces C, i.e. government spending is a substitute for private spending. On the other hand, a
negative value of  gives a positive coefficient of G2, implying that an increase in G2 raises C, i.e. government spending is a
complement to private spending.

10

C t  a  b C t  1  b  G 2 t  1   E t  1G 2 t  u t (10e)

Assume that E t  1G 2 t is given by

E t  1G2t     ( L) G2t   ( L) d t (10f)

where L is lag operator and  and  are two suitable polynomials, the lag operator

implies:

E t  1G 2 t     1G 2 t  1   2 G 2 t  2     1 d t  1   2 d t  2   (10g)

Substitution of equation (10g) into (10e) gives:

C t  (a   )  b C t  1   (b   1) G 2t  1    2 G 2t  2    3 G 2t  3   
(10h)
    1d t  1    2 d t  2     u t

Considering the limited number of observations and the possibility of

multicollinearity among lagged variables with a limited number of observations, we

chose one lag of G2 and d in our empirical analysis in the next section; the rational

expectations hypothesis also implies that actual government spending is expected

spending plus a random error t. Thus equation (10f) can be written as follows:

G 2t     1 G 2t  1   1 d t  1   t (10i)

In the case of one lags for G and d, equation (10h) can be written as:

C t    bC t  1  1G 2t  1  1d t  1  u t (10j)

with   ( a    ),
1   (b   1), (10k)
1    1

The cross equation restrictions in equation (10k), which are apparent from the

corresponding coefficients of equations (10h) and (10j), are based on a rational

expectations hypothesis. The acceptance of these restrictions in empirical analyses

validates the REH. Following Aschaur (1985), we first estimate (10i) and (10j) under

restrictions given by (10k) and then the unrestricted version of (10j) to test whether

11

the restrictions are violated or not. The REH is rejected when the restrictions are

violated.

There are two main types of difficulties associated with this form of

intertemporal consumption function. The first is a general one associated rational

expectations since only past values of G2 and d may not enough to estimate

E t  1G 2 t  1 . The second problem is related to the number of lags to be used for

annual data and this problem become very acute with the short time series and with

the presence of multicolinearity among lag variables as is the case for Bangladesh.

In addition, many authors use a discrete-time version of the Blanchard (1985)

model to test the REH and to find the sources of departures from the REH (see

Himarios (1995) for a survey). According to the Blanchard (1985) model, the REH

breaks down if a fraction () of the population dies in each period and transitory

consumption or preference shocks are absent:

C t   (1  r ) At  1  
 j 
 
j   1  j

0 1 r
E t Y l, t 
j
(11a)

where A t  1 is the stock of real assets outstanding at the end of period (t-1), r is

constant real returns on these assets,  is the constant probability of dying, Yl,t is the

real disposable labour income and Et is the expectations operator,  is the propensity

to consume out of total wealth. The first term in the brackets is the non-human wealth

and the second term is human wealth. This model predicts that the REH fails if  >0,

implying that a fraction of people die in each period, because a positive value of  (

> 0) causes economic agents to use different discount factors for taxes and interest

payments (see Himarios (1995), p. 166).

The aggregate budget constraint can be written as follows:

12

At  (1  r ) At  1  Y l , t  C t (11b)

Equations (11a, 11b) are used by many authors in deriving the aggregate consumption

function in the form of observable variables. For example, Evans (1988) solves the

model and derives the following consumption function in the form of non-human

wealth by eliminating human wealth from the equation:

1 r 1 r
Ct  (1   ) C t  1     t
1  t 1
A (11c)
1 

On the other hand, Haque (1988) provides the following consumption function by

eliminating (after substituting for) non-human wealth:

 1  (1  r ) 2
C t  (1  r )1     C t  1  (1   ) C t  2
 1   1 
(11d)
1 r 1 r
  Y l, t  1    t   
1  1  t 1

Hayashi (1982) incorporates both human and non-human wealth in the consumption

function:

1 r 2
Ct  1   (1   )C t  1    (1  r ) At  2
1  1 
(11e)
1 r
  Y l, t  1    t
1 

where  t   1    1  r  E t  E t  1Y l , t 
j . The presence of an infinite
j0

time horizon, i.e.  = 0, indicates that consumption in all three approaches follows a

random walk, i.e.  =1, implying that only lagged values of consumption rather than

any other variables explains current consumption (Hall (1978)).

Examining the validity or departures, if any, of the REH using equations 11c,

11d and 11e is based on whether  > 0 or  = 0; and consequently, whether all

13

coefficients other than lagged consumption are zero8. The REH breaks down if  > 0.

The difference in the time horizons of the government and of private economic agents

has been considered as a potential source of failure of the REH (Haque (1988)). A

positive value of  generates a positive coefficient of lagged income: a positive and

statistically significant coefficient of lagged income invalidates the REH. On the other

hand, a zero vale of  gives a positive coefficient of lagged consumption but a zero

value for the coefficient of lagged income: current consumption only depends on past

consumption rather than on any other variable. Thus, differences in the horizons of

the government and of private economic agents cannot be regarded as a source of

departure from the REH.

Results of the linear version of 11c, 11d and 11e encounter the following

difficulties (Himarios’s (1995)): Firstly, the equations are misspecified because of the

violation of the perfect capital market assumption. Secondly, (non-linear) restrictions

implicit in each equation are not taken into account with linear estimation. Himarios’s

(1994) (reviewed in Himarios’s (1995)) shows that the Blanchard (19885) model

gives the following three equivalent solutions, corresponding to equations 11c-11e,

when the assumption of perfect capital markets is relaxed:

 1 r   1 r 
C t   1   C t  1      At  1   Y t
1   1  
(11c’)
 1 r 
   1   Y t  1  u t
1  

 1 r  (1  r ) 2
C t   1  (1   )(1   )C t  1  1    Ct  2   Y t
1   1 
(11d’)
 1 r  (1  r ) 2
     (1   )  (1   )Y t  1   1    t
1  t  2
1  

A zero value of  supports the assumption of infinite horizon that the individual’s subjective probability of survival is unity
while a positive value of  , i.e. a fraction of population ( ) dies each period, indicates a finite horizon or survival rate.

14

 1 r  (1  r ) 2
C t   1   (1   )C t  1   At  2   Y t
1   1 
(11e’)
 1 r 
     (   )Y t  1  u t
1  

The parameter  represents the fraction of income that goes to liquidity constrained

households. If  =  = 0, then equations 11c’-11e’ reduce to a random walk

specification. Thus when equations 11c’-11e’ are estimated as unconstrained linear

models that ignore liquidity constraints and finite time horizons. If the null hypothesis

that there is no liquidity constraint (i.e.  = 0) is rejected, it could be argued that the

presence of liquidity constraints causes the REH to fail. Similarly, if the null

hypothesis implying the presence of infinite horizon ( = 0) is rejected, it could be

argued that the presence of a finite horizon causes the violation of the REH.

Similar to equation 11d above, Haque (1988) explores whether a finite time

horizon in life span, i.e.  > 0, and resulting differences in discount factors of the

private and government sectors are causes of departure from the REH. He uses

following linear model in his estimation:

C t   0 C t  1  1C t  2   2 Y t  1  T t  1  v t (12a)

A statistically insignificant 2 implies that the individual’s subjective probability of

survival is unity, supporting the assumption of an infinite time horizon, and so that the

differences in the horizons between the government and private economic agents

cannot be regarded as a source of the departure from the REH (Haque (1988), p. 328).

Khalid (1996) also uses the following reduced form equation to explore the

sources of departures the REH in 20 LDCs (p. 420):

C t   0  1C t  1   2 Y t  1   3Y t  2   4 G t  1   5 G t  2  u t (12b)

15

the coefficient of Ct-1 (1) is statistically significant and close to unity when (current)

consumption follows a random walk. On the other hand, if the lagged income

coefficients are statistically significant, then economic agents faced liquidity

constraints since the consumption of economic agents without liquidity constraints

should depend upon current income rather than past income.

3.2 The twin deficits and REH

The Keynesian proposition asserts that the government deficits resulting from excess

or increased government expenditures reduce current account or trade surpluses, and

vice versa. One of the policy implications of the Keynesian proposition is the

desirability of raising taxes in order to reduce budget deficits, which in turn will

reduce trade deficits. The REH, in contrast with the Keynesian proposition, states that

a tax increase would contract budget deficits but would not alter trade or current

account deficits.

Rearranging the accounting identity relating gross national income on an

expenditure basis and an income basis, the link between fiscal accounts and the

external balance can be expressed as (Agenor (1999)):

( I P  S P )  (G  T)  M  X  N T (13a)

Where IP is private investment, SP is private saving, G is government spending, T is

government revenue, M is imports, X is exports and NT is net current transfers from

abroad. This equation states that as long as (IP - SP) remains stable, changes in fiscal

deficits (G-T) will be closely associated with movements in current account deficits

(X–M - NT). However, the relationship between fiscal and external deficits may be

weakened if increases in government expenditures are associated with reductions in

16

private investment (the crowding out effect). This happens when economic agents can

anticipate that a current increase in public debt is associated with a future tax increase.

Thus, the following specification can be used to test whether fiscal deficits cause trade

deficits:

TD  α  α1 d (13b)

TD is trade deficits and d s budget deficits. A statistically insignificant 1 confirms

the REH while a negative and statistically significant 1 violates the REH.

17

4. Interpretation of the results of the REH and the crowding-out hypotheses for

Bangladesh, 1974-2001

Integration and cointegration analyses are used in our empirical investigation (Engle

and Granger (1987)). The integration analysis shows that data are first difference

stationary, i.e. the levels are non stationary, while the first differences are stationary

(table 1 in the appendix). Results from cointegration regression are reported in table 2

in the appendix.

The general findings of the extensive empirical exploration in this paper

confirm that the REH is violated in Bangladesh where the presence of liquidity

constrained households, i.e. the presence of imperfections in the financial markets and

finite survival rates are the sources of deviation from the REH. Empirical results show

that real per capita private consumption (C) under various specifications is

cointegrated generally at the 5% level with real per capita income (Y), government

expenditures before and after interest rate repayments (G & G2), taxes (T), interest

rate (r) and government’s interest repayments (RB) (table 2 in the appendix). The

results from the corresponding error correction models for various specifications

support the long-run relationships of private consumption with income, interest rate

and fiscal variables (table 3 in the appendix).

The cointegrated or long-run relationship of C with G or G2 and T invalidates

the REH since this proposition postulates no impact or relationship on private

consumption of G and T (equations 1, 6 and 9 in table 2 in the appendix). The results

reveal that the coefficient of G2 is negative and statistically significant, implying that

an increase in government expenditures (exclusive of interest rate repayments)

reduces private consumption. The coefficient of taxes becomes statistically significant

with a negative sign when government expenditures (G or G2) are excluded from the

18

model. This is plausible since the impact of fiscal policies could be captured by

government expenditures when both G (or G2) and T are included, causing T to be

insignificant in the model.

Results also reveal that the coefficient of budget deficits is negative and

statistically significant, implying that an increase in budget deficits reduces private

consumption (equations 4 and 7 in table 2 in the appendix). In addition, the coefficient

for interest rate is negative and statistically significant (equation 9b in table 2 in the

appendix). Deficit financing raises real interest rates, which in turn reduce private or

any other interest sensitive form of private spending. Empirical results on the

relationship between budget (d) and trade deficits reveal that budget deficits exert a

positive and statistically significant impact on trade deficits, refuting the REH

(equation 13b in table 2 in the appendix).

Thus, our results on the private consumption function estimation, and the

relationship between trade and budget deficits do not confirm the REH. The REH is

also rejected due to the violation of restrictions explained in equation 5 on equations 4

and 8: (i) a1  1 and (ii) a1  a 2 and a 2  a 3 (table 2 in the appendix) 9.

The violation of a1  1 ( a1  1.23 , in equation 4 without an intercept and

a1  1.22 in equation 8 with an intercept) is simply due to the fact that private

consumption in a developing country such as Bangladesh is influenced by many

unreported factors. There are many sources of incomes that are not included in the

national account and thus per capita income is generally underestimated. This result is

The restriction a1  1 implies that marginal propensity to consume is less than one; a1  a 2 implies that consumption
losses due to an imposition of taxes are equal to consumption gains resulting from a same amount of increase in income or vice
versa; a 2  a3 asserts that deficits incurred by the government today will be completely offset by rising taxes in the next
period.

19

also consistent with the poor accounting system in Bangladesh in which many

economic activities are left unreported.

Our results violate the restriction a1  a 2 : a1  1.23 and a 2  0.23 in

equation 4 without an intercept, a1  1.22 and a 2  - 0.376 in equation 8 with an

intercept. The violation of this restriction indicates the differential impact on private

consumption of income and taxes and thereby invalidates the REH.

Our results also give a 2  0.227 and a 3  0.137 for equation 4 without an

intercept and a 2  0.376 and a 3  0.002 equation 8 with an intercept (table 2 in

the appendix). The violation of the restriction a 2  a 3 , i.e. the differential impact of

taxes and government spending on private consumption, implies that the consumption

decision of a rational agent will be affected by government fiscal policy. The finding

of a 2  a 3 indicates that a reduction in consumption caused by a rise in taxes is

higher than a reduction in consumption due to a rise in government expenditures. This

differential impact implies that a rising deficit financing financed by issuing bonds

instead of taxation will tend to raise consumption owing to the wealth effects.

Similarly, the estimates of coefficients of equation 9 also reject the REH

because the restriction that the coefficient of Y be equal in absolute value to the

coefficient of G2 is not satisfied (table 2 in the appendix). The rejection of the REH in

our analysis in the case of Bangladesh should imply the acceptance of the crowding

out hypothesis, which is confirmed by the negative and statistically significant

coefficients of G, G2, T and r in our analysis.

The results on the rational expectations rule also tends to some extent to

violate the REH10 (table 4 in the appendix). We first consider the estimated values of

10
As explained in the footnote of table 4 below, Eviews gives somewhat unstable and implausible results, which are mainly
caused by the mis-specification of the model, since only past values of G2 and d may not enough to estimate Et-1G2t-1. In

20

b and . The results on b are contradictory: the value of b is statistically insignificant

in the unrestricted model while statistically significant in the restricted model. The

parameter  measures the extent of the ex ante crowding out of private consumption

expenditures by government expenditures.  = –0.38 and is statistically significant,

indicating a certain degree of complementarity between government and private

expenditure. This result contradicts our earlier findings 11. Having found the violation

of the REH in the Buiter-Tobin type models and contradictory results in rational

expectation models, further investigation using linear and non-linear models is carried

out in order to explore the robustness of the results from Buiter-Tobin type models.

Empirical results from all three models (equations 11c, 11d & 11d) reveal that

consumption follows a random walk, i.e.  =1 is rejected (table 5 in the appendix). On

the other hand, the empirical results support the presence of infinite horizon, i.e.  =

0, which implies that consumption should follow a random walk, i.e.  =1. These

conflicting findings, which are thought to be caused by model mis-specification and

non-linear restrictions, lead us to estimate equations 11c’, 11d’ and 11e’, which

incorporate financially constrained households (). Empirical findings from the non-

linear estimation of these models reveal that  and  are statistically significant (table

6 in the appendix). These results imply that the presence of finite horizons (i.e.  > 0)

and the presence of financial constrained households or imperfections in financial

markets (i.e.  > 0) are the sources of the failures of the REH. Both sources of failure

addition, selecting the number of lags to be used for annual data is arbitrary and difficult and such problems become very acute
with the short time series as is the case for Bangladesh.
11
As explained above, the finite horizon ( > 0)and the presence of liquidity-constrained individuals are considered as the main
sources of deviation from the REH. Estimated results from the linear and non-linear version of equations 11c, 11d & 11e are used
to explore the sources of the departures of the REH (tables 5 and 6 in the appendix). The presence of infinite horizon, i.e.  = 0,
suggests that consumption in all three approaches follows a random walk: only lagged values of consumption rather than any
other variables explains current consumption (Hall (1978)). A linear models test the hypothesis that all coefficients other than the
coefficient of lagged consumption are insignificant, i.e. consumption follows a random walk model, implying that the coefficient
of lagged consumption is one (Hall (1978)). On the other hand, the non-linear models examine whether  = 0 and  =1.

21

of the REH are consistent with the existing literature on developing countries (Ghatak

and Ghatak (1996), Khaled (1996), Haque (1988)).

The results of linear estimation of equations 11c’ and 11d’ reveal that the

coefficients of lagged income or lagged disposable income in all three models are

positive and statistically significant (table 6 in the appendix). The positive coefficient

of past income implies that a group of individuals is faced with liquidity constraints,

so that their consumption decision is also influenced by past income. Thus, these

results in 11e’ are consistent with the non-linear estimation results. Similar results are

derived when the Khaled (1996) model, which includes income and government

expenditures, is estimated. The results in the Khaled (1996) model reveal that lagged

government expenditures exert a positive impact on current consumption. This result

is consistent with the fact that (lagged) government expenditures increase (lagged)

private income, which in turn raises (current) consumption.

22

5. Conclusions

This paper examines the Ricardian equivalence hypothesis (REH) and its sources of

failure in the case of Bangladesh using various types of theoretical specifications,

annual data from 1974-2001 and linear and non-linear time series techniques. The

empirical findings tend to invalidate the REH and reveal that a finite time horizon and

the presence of liquidity-constrained individuals are the sources of deviation from the

REH. Empirical results show that real per capita private consumption (C), under

various specifications, is cointegrated generally at the 5% level with real per capita

income (Y), government expenditures before and after interest rate repayments (G &

G2), taxes (T), budget deficits (d) and the interest rate (r).

The results reveal that the coefficients of G2, d and r are is negative and

statistically significant, implying that an increase in these variables reduces private

consumption: deficit financing raises the real interest rate which in turn reduces

private or any other interest sensitive form of private spending. The coefficient for the

variable taxes becomes statistically significant with a negative sign when government

expenditures (G or G2) are excluded from the model. This result is plausible, since the

impact of fiscal policies is captured by government expenditures when both G (or G2)

and T are included, causing T to be insignificant in the model.

Empirical findings on the relationship between the budget (d) and trade

deficits imply that budget deficits exert a positive and statistically significant impact

on trade deficits, refuting the REH. Thus, our results on private consumption function

estimation, and on the relationship between trade and budget deficits do not confirm

the REH.

The finding of the differential impact of taxes and government expenditures

violates the REH and indicates that a reduction in consumption caused by a rise in

23

taxes is higher than a reduction in consumption due to a rise in government

expenditures. This differential impact implies that a rising deficit financed by issuing

bonds instead of taxation will raise consumption owing to wealth effects. The

violation on this restriction indicates the differential impact on private consumption of

income and taxes and hereby invalidates the REH.

Results from non-linear estimation methods imply that the presence of finite

horizons and the presence of financial constrained households or imperfections in the

financial markets are the sources of the failures of the REH. The results from the

linear model reveal that the coefficients of lagged income or lagged disposable

income positively affect current consumption, implying that some individuals are

faced with liquidity constraints, therefore their current consumption decision is also

influenced by past income. Thus both linear and non-linear methods provide

consistent results which confirm the existing literature.

In short, our extensive empirical exploration confirms that the REH is violated

in Bangladesh where the presence of liquidity constrained households, i.e. the

presence of imperfections in the financial markets and finite survival rates are the

sources of the deviation of the REH. Thus, fiscal policies should be used as important

policy instruments in order to boost private consumption and control trade deficits,

which are the prime goals of stabilisation policies being followed in Bangladesh.

24

References

Agenor, Pierre-Richard (2000) The Economics of Adjustment and Growth, Academic

Press, London, chapter three, ‘Fiscal Deficits, Public Debt, and the Current Account’.

Aschaur, D.A. (1985) Fiscal Policy and Aggregate Demand, American Economic

Review, Vol. 75, pp. 117-127.

Auerbach, Paul and Siddiki (2002) Financial Liberalisation and Economic
Development: An Assessment, Journal of Economic Surveys, forthcoming.

Barro, R. (1974) Are government bonds net wealth?, Journal of Political Economy,

Vol. 82, no. 3, pp. 1095-1117.

Barro, R. (1989) The Ricardian approach to budget deficits, Journal of Economic

Perspectives, Vol. 3, No. 2, pp. 37-54.

Becker, T. (1998) An Investigation of Ricardian Equivalence in a Common Trens

Model, Journal of Monetary Economics, Vol. 39, p. 405-431.

Blanchard, O. (1985) Debt, Deficits and Aggregate Demand, Journal of Political

Economy, 93(2), 223-247.

Boskin, M.J. (1988) Concepts and Measures of Federal Deficits and Debt and Their

Impact on Economic Activity, in: K. Arrow and M.J. Boskin, eds., The Economics of

Public Debt, Macmillan, London.

Buiter, W. and Tobin, J. (1979) Debt Neutrality: A Brief Review of Doctrine and

Evidence, in G.M. von Furstenburg, ed., Social Security Versus Private Saving,

Ballinger, Cambridge, MA.

Charemza, Wojciech W. and Deadman, Derek F. (1997) New Directions in

Econometric Practice, Edward Elgar, Second Edition, Uk.

Engle, R. F. and Granger, C. W. J. (1987) Cointegration and Error Correction:

Representation, Estimation and Testing, Econometrica, 52, 251-76.

25

Evans, Paul (1988) Are Consumers Ricardian? Evidence for the United States,

Journal of Political Economy, 96, 983-1004.

Feldstein, M. and Horioka, C. (1980) Domestic Saving and International Capital

Flows, Economic Journal, 90, 314-329.

Ghatak, Anita and Ghatak, Subrata (1996) Budgetary deficits and Ricardian

equivalence: The case of India, 1950-1986, Journal of Public Economics, Vol. 60, pp.

267-282.

Ghatak, Subrata (1995) Monetary Economics in Developing Countries, Macmillan

Press Ltd, Second Edition.

Hall, Robert E (1978) Intertemporal Substitution in Consumption, Journal of Political

Economy, 86, 339-357.

Haque, Nadeem U. (1988) Fiscal Policy and Private Sector Saving Behavior in

Developing Countries, Staff Papers, International Monetary Fund, Vol. 35, p. 316-35.

Hayashi, Fumio (1994) The Permanent Income Hypothesis: Estimation and Testing

by Instrumental Variables, Journal of Political Economy, 90, 895-916.

Himarios, Daniel (1994) Non-Ricardian Consumers: Finite Horizons or Liquidity

Constraints?, Mimeo, University of Texas at Arlingtson.

Himarios, Daniel (1995) Eüler equation tests of Ricardian equivalence, Economic

Letters, Vol. 48, p. 165-171.

Khalid, Ahmed M. (1996) Ricardian equivalence: Empirical evidence from

developing economies, Journal of Development Economics, Vol. 51, pp. 413-432.

Khalid, Ahmed M. and Guan, T. W. (1999) Causality Tests of Budget and Current

Account Deficits: Cross-Country Comparisons, Empirical Economics, Vol. 24, p.

389-402.

26

Kormendi, Rodger C. (1983) Government Debt, Government Spending, and Private

Sector Behaviour, American Economic Review, Vol. 73, no. 5, p. 994-1009.

Leiderman, Leonard and Blejer, Mario I. (199) Modeling and testing Ricardian

equivalence: A survey, IMF Staff Papers, March, pp. 1-35.

Normandin, Michel (1999) Budget deficit persistence and twin deficits hypothesis,

Journal of International Economics, Vol. 49, pp. 171-193.

Pesaran, M. H. and Pesaran, B. (1997) Microfit 4.0, Oxford University Press, Oxford.

Seater, J.J. (1993) Ricardian equivalence, Journal of Economic Literature, Vol. 31,

no.1 pp. 142-190.

Rosenweig, Jeffrey A. and Tallman, Ellis W. (1993) Fiscal Policy and Trade

Adjustment: Are the Deficits Really Twins? Economic Inquiry, 31(4), p. 580-594.

Siddiki, J.U. (2000) The Demand for Money in Bangladesh: A Cointegration

Analysis, Applied Economics, 32(15), pp. 1977-1984

Siddiki, J.U. (2002) Unofficial Exchange Rates in Bangladesh: A Cointegration

Analysis, Contemporary Economic Policy, 20(3), June, 288-300.

Vamvoukas, George A. (1999) The twin deficits phenomenon: evidence from Greece,

Applied Economics, Vol. 31, pp. 1093-1100.

27

Appendix

Figure 1
Budget (TBD) and Trade Deficits (TD) in Bangladesh

14

12

10

TBD &TD

1970 1975 1980 1985 1990 1995 2000 2005
YEAR

TDGDP BDGDP

Table 1: Augmented Dicky-Fuller Test for Unit Roots (check the results again:

do tests again using updated data

Variables Levels First Difference Variables Levels First Difference
B -2.7025 -5.97 T -1.8274 -7.7743
TD -1.6986 -10.4779 Y 2.7543 -5.6563
C -0.19880 -5.4599 Y-G2 2.2529 -5.1653
G -1.2947 -7.1405 Y-T 2.0148 -4.8938
G2 -1.1432 -6.8550 W -1.6169 -66.2374
‘r’ -2.5044 -7.6366 RB -2.9009 -4.8115

The Schwartz Bayesian Criterion (SBC) is selecting number of lags. In all cases, the number
of lags based on SBC appears to be sufficient to secure the lack of autocorrelation of error
terms. Critical value with 22 observations is –2.9750.

28

Table 2: Results on the REH: Equations 1 and 4: sample period: 1974-2001

Equation no. Estimated t-values R2, DW, ADF and its critical value(CV) and
and variables coefficients other diagnostic tests12
(1) (2) (3) (4)
1. C = f(INPT, Y, G, T, W); Sample periods: 1974-2001 (28 observations)
INPT -0.118 -0.12 R2 = 0.94; DW = 1.69; ADF =-4.5963 (CV=-
Y 1.300** 9.12 4.9527); AR2-2(2) = 0.883[0.643]; RESET-
G -0.340** -5.21 F(1, 22) = 0.6578[0.426]; NOR- 2(2) =
T -0.037 -0.68 12.1909[0.002]; H-F(1, 26) = 3.18[0.086]
W -0.019 -0.45
1a. C = f(INPT, W, G, T)
INPT 8.515 19.02 R2 = 0.76; DW = 0.99; ADF =-2.7157 (CV=-
W 0.276** 4.64 4.5276); AR2-2(2) = 7.27[0.026]; RESET-
G -0.402** -2.95 F(1, 23) = 6.2451[0.020]; NOR- 2(2) =
T 0.132 1.21 2.2609[0.323]; H-F(1, 26) = 0.13439[0.717]
1b. C = f(INPT, W, G, T) with AR(1)
INPT 9.00 17.57
W 0.290** 3.92 R2 = 0.82; DW = 1.996 (the coefficient of
G -0.482** -3.42 AR(2) is not significant)
T 0.127 1.45
AR(1) 0.576** 3.73
1. C = f(Y, G, T) without intercept and W
Y 1.288** 43.22 R2 = 0.95; DW = 1.78; ADF =-4.9352
G -0.346** -5.75 (CV13=-4.17); AR2-2(2) = 0.62589[0.731];
T -0.054 -1.44 RESET-F(1, 24) = 0.16354[0.689]; NOR-
2(2) = 7.14[0.028]; H-F(1, 26) = 0.08[0.071]
4. C = f(INPT, Y, T, d)
INPT -0.32 -0.52 R2 = 0.94; DW = 1.78; ADF =-4.9570 (CV= -
Y 1.284** 12.63 4.5276); AR2-2(2) = 0.523[0.770]; RESET-
T -0.246** -5.51 F(1, 23) = 0.24121 [0.628]; NOR- 2(2) =
d -0.137** -4.87 3.522[0.172]; H-F(1, 26) = 2.6964[0.113]
4. C = f(Y, T, d) without intercept14
Y 1.233** 49.28 R2 = 0.94; DW = 1.74; ADF =-4.7405 (CV= -
T -0.227** -9.54 4.17); AR2-2(2) = 0.76572[0.682]; RESET-
D -0.137** -4.92 F(1, 24) = 0.25744[0.617]; NOR- 2(2) =
5.2754[0.072]; H-F(1, 26) = 2.1681[0.151]

12
Throughout our analysis, t-statistics are reported in the parentheses, ** and * represent 1% and 5% significance levels,
respectively. AR2- 2(2) is chi square tests for second order residual joint autocorrelation; RESET-F is the F test for mis-
specified functional form; NOR- 2(2) is the chi square statistic for testing normality; H-F is the F statistics for testing
heteroscedasticity; probability values are reported in the square brackets.
13
Estimation is carried out using Microfit 4.0, which provides critical values (CVs) of Dickey-Fuller and Augmented Dickey-
Fuller (ADF) tests when a constant is included with a model; we use CVs from Charemza and Deadman (1997) (p. 288) if a
model is estimated without a constant. There is no significant difference between the CVs obtained from both sources. The CVs
of ADF tests reported in this paper are based on 30 observations.
14
Wald Statistic 2( 1) = 2480.6 [.000] for a 1=|a2 |; Wald Statistic 2( 1) = 5.9936[.014] for a2=a3; where for a1, a2 and a3 are the
coefficients of Y, T and d, respectively.

29

Table 2: continued (equations 6, 7, 8 and 9a)

Eq. No. and Estimated t-values R2, DW, ADF and its critical value(CV) and
variables coefficients other diagnostic tests
(1) (2) (3) (4)
6. C = f(INPT, Y, T, G2)
INPT -0.250 -0.44 R2 = 0.95; DW = 1.74; ADF =-4.8158(CV= -
Y 1.315** 14.06 4.5276); AR2-2(2) = 1.3736[0.503]; RESET-
T -0.037 -0.69 F(1, 23) = 0.002[0.962]; NOR- 2(2) =
G2 -0.356** -5.79 5.1942[0.074]; H-F(1, 26) = 3.6864[.066]
6. C = f(Y, T, G2) with out intercept
Y 1.275** 46.46 R2 = 0.95; DW = 1.72; ADF =-4.6720(CV= -
T -0.023 -0.54 4.17); AR2-2(2) = 1.6297[0.443]; RESET-
G2 -0.356** -5.87 F(1, 24) = 0.19454[0.663]; NOR- 2(2) =
7.1521[0.028]; H-F(1, 26) = 3.06[0.092]
7. C = f(INPT, (Y-G2), d)
INPT 1.936** 3.72 R2 = 0.87; DW = 1.26; ADF =-3.6494(CV= -
(Y-G2) 0.851** 12.64 4.0706); AR2-2(2) =6.2157 [0.045]; RESET-
d -0.124** -3.12 F(1, 24) = 0.0845[0.774]; NOR- 2(2) =
5.40[0.067]; H-F(1, 26) = 1.1065[0.303]
7. C = f(INPT, (Y-G2), d) with AR(1)
INPT 1.91** 2.66 R2 = 0.89; DW = 1.75
(Y-G2) 0.86** 9.56
d -0.13** -3.10
AR(1) 0.35 1.96
8. C = f(INPT, Y, G2, RB)15
INPT 0.465 1.0422 R2 = 0.95; DW = 1.67; ADF =-4.4421(CV= -
Y 1.221** 15.60 4.5276); AR2-2(2) =1.31 [0.519]; RESET-
G2 -0.376** -7.95 F(1, 23) = 0.000[0.999]; NOR- 2(2) =
RB 0.002 0.35 7.24[0.027]; H-F(1, 26) = 3.04[0.093]
8. C = f(Y, G2, RB) without intercept
Y 1.297** 43.38 R2 = 0.94; DW = 1.57; ADF =-4.2484(CV= -
G2 -0.404** -10.32 4.17); AR2-2(2) =1.4411[.486]; RESET-F(1,
RB 0.004 0.82 24) = 1.09[0.308]; NOR- 2(2) = 0.98[0.613];
H-F(1, 26) = 3.97[0.035]
9a. C = f(INPT, Y, G2)
INPT -0.007 -0.02 R2 = 0.95; DW = 1.63; ADF =-4.4125(CV= -
Y 1.28** 15.94 4.0706); AR2-2(2) =1.6780 [0.432]; RESET-
G2 -0.38** -8.29 F(1, 24) = 0.004[0.951]; NOR- 2(2) =
3.98[0.136]; H-F(1, 26) = 3.41[0.076]
9a. C = f(Y, G2) without intercept16
Y 1.28** 52.02 R2 = 0.95; DW = 1.63; ADF =-4.4142(CV= --
G2 -0.38** -12.27 3.82); AR2-2(2) =1.6832 [0.431]; RESET-
F(1, 25) = 0.003[0.98]; NOR- 2(2) =
4.07[0.131]; H-F(1, 26) = 3.39[0.077]

15
Wald Statistic 2( 1) = 2.1201[.145] for a 1=|a2 | ; Wald Statistic 2( 1) = 175.3211[.000] for a2=a3; where for a1, a2 and a3 are the
coefficients of Y, G2 and RB, respectively.
16
Wald Statistic 2( 1) = 18077.4[.000] for a1=|a2 |.

30

Table 2: continued (equations 9b and 13b)

Eq. No. and Estimated t-values R2, DW, ADF and its critical value(CV) and
variables coefficients other diagnostic tests
(1) (2) (3) (4)
9b. C = f(INPT, Y, G2, r)
INPT 0.457 1.09 R2 = 0.96; DW = 1.62; ADF =-4.7638 (CV= -
Y 1.1334** 12.86 4.5276); AR1-2(2) =2.02 [0.363]; RESET-
G2 -0.26** -4.42 F(1, 23) = 0.03[0.866]; NOR- 2(2) =
‘r’ -0.002** -2.86 4.5466[0.103]; H-F(1, 26) = 4.27[0.049]
9b. C = f(Y, G2, r) without intercept17
Y 1.223** 38.87 R2 = 0.96; DW = 1.51; ADF =-4.5417 (CV= -
G2 -0.310** -7.76 4.17); AR1-2(2) =1.94 [0.379]; RESET-F(1,
‘r’ -0.002* -2.63 24) = 1.189[0.286]; NOR- 2(2) =
1.5589[0.459]; H-F(1, 26) = 4.5879[0.042];
13b. TD = f(INPT, d)
INPT 3.86** 2.93 R2 = 0.11; DW = 1.82; ADF =-5.0352(CV= -
d 0.38 1.74 3.5804); AR1-2(2) =0.73 [0.694]; RESET-
F(1, 24) = 0.22[0.64]; NOR- 2(2) =
7.33[0.06]; H-F(1, 25) = 1.39[0.25]

17
Wald Statistic CHSQ( 1)= 11680.1[.000] for a1=|a2|; Wald Statistic CHSQ( 1)= 57.9031[.000] for a2=a3

31

Table 3: Error correction models

Eq. No. and Estimated t-values R2, DW, ADF and its critical value(CV) and
variables coefficients other diagnostic tests
(1) (2) (3) (4)
ECM 1. Error correction (EC) model of equation 1: C = f(INPT, Y, G, T, W); statistically
insignificant intercept and T are excluded.
Y 1.651** 5.88 R2 = 0.85; DW = 1.52; AR1-F(1, 22)=
G -0.343** -6.08 7.49[0.012]; RESET-F(1, 22) =
W -0.175* -2.37 1.1982[0.286]; NOR-  (2) = 1.0270[0.598];

Ut-1 -0.650** -3.25 H-F(1, 25) = 0.001[0.974]
ECM 4. EC model of equation 4: C = f(INPT, Y, T, d); statistically insignificant intercept
is excluded.
Y 1.150** 4.93 R2 = 0.79; DW = 1.73; AR1-F(1, 22)=
T -0.223** -5.49 1.78[0.196]; RESET-F(1, 22) = 0.114[0.739];
d -0.170** -6.55 NOR- 2(2) = 1.33[0.513]; H-F(1, 25) =
Ut-1 -0.907** -4.81 0.181[0.674]
ECM 6. EC model of equation 6: C = f(INPT, Y, T, G2); statistically insignificant
intercept and T are excluded.
Y 1.24** 5.95 R2 = 0.83; DW = 1.84; AR1-F(1, 22)=
G2 -0.434** -9.03 0.38[0.542]; RESET-F(1, 23) = 0.15[0.702];
Ut-1 -0.881** -4.8 NOR- 2(2) = 2.74[0.254]; H-F(1, 25) =
0.17[0.683]
ECM 7. EC model of equation 7: C = f(INPT, (Y-G2), d); statistically insignificant
intercept and b are excluded.
Y 0.653* 2.18 R2 = 0.57; DW = 1.78; AR1-F(1, )=
(Y-G2) -0.165** -4.89 0.02[0.889]; RESET-F(1, 23) = 3.37[0.079];
Ut-1 -0.760** -3.985 NOR- 2(2) = 0.355[0.837]; H-F(1, 25) =
2.24[0.084]
ECM 8. EC model of equation 8: C = f(INPT, Y, G2, RB); statistically insignificant
intercept and RB are excluded.
Y 1.225** 5.61 R2 = 0.81; DW = 1.85; AR1-F(1, 23)=
G2 -0.448 -8.96 0.167[0.686]; RESET-F(1, 23) =
Ut-1 -0.829** -4.35 0.202[0.657]; NOR- (2) = 4.1857 [0.123];
H-F(1, 25) = 0.17[0.689]
ECM9b. EC model of equation 9b: C = f(INPT, Y, G2, r)
INPT 0.01 1.01 R2 = 0.89; DW = 1.78; AR1-F(1, 21)=
Y 1.04** 4.83 1.48[0.237]; RESET-F(1, 21) = 0.265[0.612];
G2 -0.366** -6.81 NOR- 2(2) = 0.248[0.883]; H-F(1, 25) =
r -0.001 -1.40 0.418[0.524]
Ut-1 -1.05** -5.52
ECM 13b. EC model of 13b: TD = f(INPT, b)
INPT 0.02 0.39 R2 = .48; DW = 2.28; AR1-F(1, 22) = 14.87
d 0.157 0.73 [.001]; RESET-F(1, 22) =.006[.939]; NOR-
Ut-1 -0.93** -4.6388 2(2) = 6.886[.032]; H-F(1, 24) =
0.21410[.648]
Ut-1 is the EC term, i.e. the lag value of residual of the corresponding equation.

32

Table 4: Estimates of Aschauer model18:

Constrained Unconstrained Hypothesised
 = -0.29 (-0.67)  = -0.29 (-0.00)  = -0.4385
b = 1.03** (27.30) b = 1.03 (0.04) b = 1.03
 = 0.47* (4.06) 1 = 0.47 (0.03) 1 = -0.0235
1 = 0.42 (0.03) 1 = 0.1645
 = 1.55** (3.24)  = 1.55 (1.38)  (C1) = 1.55
1 = 1.08* (6.27) 1 = 1.08 (1.90) 1(C2) = 1.08
1 = -0.35 (-1.66) 1 = -0.35 (-1.35) 1 (C4) = -0.35
Log likelihood (Lr) = 78.95819 Log likelihood(Lu) = -61.42386

The Wald statistics = -2log(Lr/Lu) = - not significant

18
We use Eviews to estimate this non-linear model. The full information maximum likelihood and three-stage least squares
methods are used to estimate both restricted and unrestricted models. Results obtained from the full information maximum
likelihood methods are reported here. Both methods give somewhat unstable results. The full information maximum likelihood
method in some cases gives unexpected positive values of log likelihood. Thus, further investigation will be made using other
software packages in order derive stable results.

33

Table 5: Sources of the deviation of REH

Eq. No. and Estimated t-values R2, DW and null hypotheses
variables coefficients (4)
(1) (2) (3)
Evans (1988) Model) (eq. 11c): non-linear estimation (with r=4)
 0.81** 190.49 R2 = 0.82; DW = 2.03; Wald Statistic 2(1) =
 0.01 -1.69 2085.931 [0.000] for  =1.
Haque (1988) Model (eq. 11d): non-linear estimation (with r=4)
 0.70** 5.67 R2 = -4.75; DW = 2.01; Wald Statistic 2(1) =
 -0.47 -0.77 5.657[0.017] for  =1.
Hayashi (1982) Model) (eq. 11e): non-linear estimation (with r=4)
 0.80** 344.35 R2 = 0.77; DW = 1.91; Wald Statistic 2(1) =
 0.000 0.42 7527.538[0.000] for  =1.
Himarios (1994) Model (eq. 11c’) non-linear estimation
 0.98** 159.54 R2 = 0.86; DW = 1.18; Wald Statistic 2(1) =
 -0.01 -1.40 14.88[0.000] for  =1. Wald Statistic 2(1) =
 0.89 25.52 10156.45[0.000] for  ==0.
Himarios (1994) Model (eq. 11d’) non-linear estimation
 1.18** 346.11 R2 = 0.29; DW = 1.51; Wald Statistic 2(1) =
 -2.67** -6.82 2704.80 [0.000] for  =1. Wald Statistic
 3.31** 6.94 2(1) = 432.7557 [0.000] for  ==0.
Himarios (1994) Model) (eq. 11e’) non-linear estimation
 0.79** 122.41 R2 = 80; DW = 1.84; Wald Statistic 2(1) =
 0.002 1.70 1096.407 [0.000] for  =1. Wald Statistic
 -0.07 -1.91 2(1) = 3.70[0.157] for  ==0.

34

Table 6: Sources of the deviation of REH

Eq. No. and Estimated t-values R2, DW, ADF and its critical value(CV) and
variables coefficients other diagnostic tests
(1) (2) (3) (4)
Evans (1988) Model) (eq. 11c) with AR(2)
Ct-1 0.976** 118.34 R2 = 0.91; DW = 2.02; 2(1) = 9.7601[.002]
At-1 0.030** 3.12 (to test the coefficient of At-1 equal to zero.
AR(1) -0.047 -0.53 2(1) = 8.63(00.003) (to test the coefficient of
AR(2) -0.891** -10.23 Ct-1 equal to one).
Haque (1988), Equation 11d
Ct-1 0.50* 2.67 R2 = 0.86; DW = 1.79; AR1-F(1, 23)=
Ct-2 -0.02 -0.14 13.95[0.001]; RESET-F(1, 23) = 3.58[0.071];
Yt-1 0.51** 4.01 NOR- 2(2) = 31.2455[0.000]; H-F(1, 25) =
3.02[0.095]
Haque (1988) model when lagged disposable income is included
Ct-1 0.374* 2.04 R2 = 0.88; DW = 1.69; AR1-F(1, 23)=
Ct-2 -0.053 -0.36 4.08[0.055]; RESET-F(1, 23) = 2.85[0.105];
(Yt-1 - Tt-1) 0.672** 4.81 NOR- 2(2) = 45.48[0.000]; H-F(1, 25) =
2.36[0.137]
Hayashi (1982) model eq. 11e
Ct-1 0.216* 2.07 Linear estimation Sample 1975-1997; R2 =
At-2 -0.049** 5.03 0.93; DW = 2.04 ; 2(2) = 58.93(0.000) (to
Yt-1 0.810** 7.63 test the coefficient of Wt-2 and Yt-1 equal to
zero.
Khalid (1996), equation 12c
INPT 1.56 2.00 R2 = 0.89; DW = 1.71; AR1-F(1, 20)=
Ct-1 0.01 0.03 1.53[0.230]; RESET-F(1, 20) = 0.02[0.898];
Yt-1 1.52* 2.41 NOR- 2(2) = 25.9077[0.000]; H-F(1, 25) =
Yt-2 -0.63 -1.06 2.1432[0.156]; Wald Statistic 2(1) =
Gt-1 -0.26 -1.55 8.5888[.003] for the csoefficient of Ct-1.
Gt-2 0.13 1.39

35
References (31)
Agenor, Pierre-Richard (2000) The Economics of Adjustment and Growth, Academic Press, London, chapter three, 'Fiscal Deficits, Public Debt, and the Current Account'.
Aschaur, D.A. (1985) Fiscal Policy and Aggregate Demand, American Economic Review, Vol. 75, pp. 117-127.
Auerbach, Paul and Siddiki (2002) Financial Liberalisation and Economic Development: An Assessment, Journal of Economic Surveys, forthcoming.
Barro, R. (1974) Are government bonds net wealth?, Journal of Political Economy, Vol. 82, no. 3, pp. 1095-1117.
Barro, R. (1989) The Ricardian approach to budget deficits, Journal of Economic Perspectives, Vol. 3, No. 2, pp. 37-54.
Becker, T. (1998) An Investigation of Ricardian Equivalence in a Common Trens Model, Journal of Monetary Economics, Vol. 39, p. 405-431.
Blanchard, O. (1985) Debt, Deficits and Aggregate Demand, Journal of Political Economy, 93(2), 223-247.
Boskin, M.J. (1988) Concepts and Measures of Federal Deficits and Debt and Their Impact on Economic Activity, in: K. Arrow and M.J. Boskin, eds., The Economics of Public Debt, Macmillan, London.
Buiter, W. and Tobin, J. (1979) Debt Neutrality: A Brief Review of Doctrine and Evidence, in G.M. von Furstenburg, ed., Social Security Versus Private Saving, Ballinger, Cambridge, MA.
Charemza, Wojciech W. and Deadman, Derek F. (1997) New Directions in Econometric Practice, Edward Elgar, Second Edition, Uk.
Engle, R. F. and Granger, C. W. J. (1987) Cointegration and Error Correction: Representation, Estimation and Testing, Econometrica, 52, 251-76.
Evans, Paul (1988) Are Consumers Ricardian? Evidence for the United States, Journal of Political Economy, 96, 983-1004.
Feldstein, M. and Horioka, C. (1980) Domestic Saving and International Capital Flows, Economic Journal, 90, 314-329.
Ghatak, Anita and Ghatak, Subrata (1996) Budgetary deficits and Ricardian equivalence: The case of India, 1950-1986, Journal of Public Economics, Vol. 60, pp. 267-282.
Ghatak, Subrata (1995) Monetary Economics in Developing Countries, Macmillan Press Ltd, Second Edition.
Hall, Robert E (1978) Intertemporal Substitution in Consumption, Journal of Political Economy, 86, 339-357.
Haque, Nadeem U. (1988) Fiscal Policy and Private Sector Saving Behavior in Developing Countries, Staff Papers, International Monetary Fund, Vol. 35, p. 316-35.
Hayashi, Fumio (1994) The Permanent Income Hypothesis: Estimation and Testing by Instrumental Variables, Journal of Political Economy, 90, 895-916.
Himarios, Daniel (1994) Non-Ricardian Consumers: Finite Horizons or Liquidity Constraints?, Mimeo, University of Texas at Arlingtson.
Himarios, Daniel (1995) Eüler equation tests of Ricardian equivalence, Economic Letters, Vol. 48, p. 165-171.
Khalid, Ahmed M. (1996) Ricardian equivalence: Empirical evidence from developing economies, Journal of Development Economics, Vol. 51, pp. 413-432.
Khalid, Ahmed M. and Guan, T. W. (1999) Causality Tests of Budget and Current Account Deficits: Cross-Country Comparisons, Empirical Economics, Vol. 24, p. 389-402.
Kormendi, Rodger C. (1983) Government Debt, Government Spending, and Private Sector Behaviour, American Economic Review, Vol. 73, no. 5, p. 994-1009.
Leiderman, Leonard and Blejer, Mario I. (199) Modeling and testing Ricardian equivalence: A survey, IMF Staff Papers, March, pp. 1-35.
Normandin, Michel (1999) Budget deficit persistence and twin deficits hypothesis, Journal of International Economics, Vol. 49, pp. 171-193.
Pesaran, M. H. and Pesaran, B. (1997) Microfit 4.0, Oxford University Press, Oxford.
Seater, J.J. (1993) Ricardian equivalence, Journal of Economic Literature, Vol. 31, no.1 pp. 142-190.
Rosenweig, Jeffrey A. and Tallman, Ellis W. (1993) Fiscal Policy and Trade Adjustment: Are the Deficits Really Twins? Economic Inquiry, 31(4), p. 580-594.
Siddiki, J.U. (2000) The Demand for Money in Bangladesh: A Cointegration Analysis, Applied Economics, 32(15), pp. 1977-1984
Siddiki, J.U. (2002) Unofficial Exchange Rates in Bangladesh: A Cointegration Analysis, Contemporary Economic Policy, 20(3), June, 288-300.
Vamvoukas, George A. (1999) The twin deficits phenomenon: evidence from Greece, Applied Economics, Vol. 31, pp. 1093-1100.
Jalal Siddiki
Papers
19
Followers
View all papers from
Jalal Siddiki
arrow_forward
Related papers
Ricardian Equivalence in Developing Countries: Fiscal Policy, Private Consumption and Investment in Indonesia
Yuniarto Hadiwibowo
This study reviews the impact of fiscal policy on private consumption and investment in Indonesia. Ricardian-Barro proposition suggests fiscal policy does not have real impact on consumption and investment. On the other hand, Keynesian perspective believes fiscal policy can influence consumption and investment. The dataset is built according to the budget periods. We find that private consumption and investment are cointegrated with their determinants. Using Vector Error Correction Model and specification derived from Augmented Consolidated Approach by , we find long-run relationship among private consumption, income, government spending, tax, wealth and government external debt. In addition, long-run relationship can also be seen among investment, income, government spending and tax. Interestingly, fiscal policy seems to be insignificant in the short-run. The results suggest that Keynesian perspective can be observed in the long-run, while Ricardian-Barro proposition cannot be rejected in the short-run. Short-term government spending and tax measures cannot significantly influence consumption and investment. Implication is fiscal policy should be directed toward medium and long-term objectives rather than focusing only for the short-term measures.
Download free PDF
View PDF
chevron_right
Financing the Fiscal Deficit in Pakistan: Evidence on Ricardian Equivalence
Neelma Shamsi
Muhammad Waqas
This study investigated the validity of Ricardian equivalence hypothesis in case of Pakistan. Annual time series data from 1978 to 2013 has been used for analysis. Two models have been estimated; structural consumption function and structural saving function. The results of unit root tests explored that all the variables have different order of integration. The results of ARDL co-integration approach verified the existence of a long run relationship among variables for both models. By using OLS, wald test rejects the restrictions on both models and concludes that Pakistan is a non-ricardian economy. Keywords: Ricardian Equivalence Theory, Budget Deficit, Pakistan
Download free PDF
View PDF
chevron_right
Government expenditure and household consumption in Bangladesh through the lens of economic theories: an empirical assessment
Nahid Mahmud
2012
The relationship between government and household consumption remains to be one of the contentious issues in both theories and empirics, though its' immense importance in fiscal policy formulation. Like theories, the empirical studies regarding the relationship between government and household consumption provide opposing results. In this backdrop, the present study examines public-private consumption relationship for Bangladesh economy through the lens of economic theories using the cointegration and error correction modeling strategies to tackle the problem of non-stationary data. Two different variant of cointegration technique have been employed and in either case a valid long run positive relationship has been found. However, the error correction model has found an inverse relationship between public and private consumption in the short run. Finally, we test for Granger causality and find no long run causal relationship between government consumption and household consumption. In general, our finding goes with the Barro-Ricardian equivalence hypothesis of government spending that household consumption is unrelated with government consumption decision in the long-run.
Download free PDF
View PDF
chevron_right
The Effect of Fiscal Policy on the Indonesian Household Consumption: The Application of the Ricardian Equivalence Hypothesis
Setyo tri wahyudi
2018
The Keynesian perspective suggests that fiscal policy through tax reduction affects the economy through its influence on public consumption. However, the Ricardian Equivalence perspective argues that fiscal policy through tax reduction will not have an impact on household consumption because it is likely that the household will respond to tax reduction policy by increasing household savings to anticipate future tax increase. The level of Ricardian equivalence varies by country, depending on the household characteristics and fiscal situation of each country. These imply that the Ricardian Equivalence cannot remain continuous over time. Using the Error Correction Model (ECM) method to analyze the Indonesian data of the period of 1990-2015, this study aims to detect the existence of the Ricardian Equivalence in Indonesia as an indicator of the fiscal policy effectiveness in Indonesia. Our results indicate that fiscal policy through tax instruments and government expenditures do not aff...
Download free PDF
View PDF
chevron_right
Checking the Validity of Ricardian Equivalence Hypothesis: Analysis from Developed and Developing Countries
Rakia Nasir
Journal of Development and Social Sciences
The existence of the Ricardian Equivalence Hypothesis (REH) in the case of the Group of Seven (G7) and the South Asian Association of Regional Cooperation (SAARC) nations is investigated in this study. The objective of this study is check the idea that combining the cases of industrialized and emerging economies will demonstrate REH. General Government Revenue, Government Budget Deficit, General Government Gross Debt, Household Final Consumption Expenditure, Disposable Income, General Government Total Expenditure, and Wealth are all measured using a panel data set from 2001 to 2021.Five distinct unit root tests were used to validate the stationarity of selected variables. Considering the Hausman test, researcher estimate the random effect model before Panel Least Squares (PLS). The Wald test has rejected all restrictions applied to verify the presence of REH in developed and emerging economies after applying PLS. According to the Findings the Fiscal policy should be one of the stabilizing measures in this understudied economic world to regulate income and expenditures.
Download free PDF
View PDF
chevron_right
Budget deficit and growth: in search of ceiling for Bangladesh
Salina Siddiqua
Business and Economic Horizons, 2018
The impact of fiscal deficit measured by deficit in national budget on the growth of respective economy has been a widely researched area with plenty of debatable results. Shedding light in search of the optimum level of budget deficit, the current paper tried to contribute to the field of literature on this issue which is perhaps inadequate as far as Bangladesh economy is concerned. A total of 40 years of time series data spanning form 1975 -76 to 2014 -15 has been employed. Identification of integration order of the variables was examined performing Augmented Dickey Fuller (ADF), Phillips-Perron (PP) and Kwiatkowski-Phillips-Schmidt-Shin (KPSS) tests. Establishing the existence of cointegration among variables following the Johansen's procedure, long-run cointegrating vector has been estimated depending on VECM. The threshold has been identified solving the estimated long-run cointegrating relationship for a local maximum. Findings can be summarized by saying that the long-run impact of budget deficit on growth would remain positive; nevertheless, there would be no short-run adjustment. Depending on the model definition and the particular exogenous variable(s), the threshold budget deficit has been measured to range between 4.55 to 5.0 percent of GDP.
Download free PDF
View PDF
chevron_right
Related topics
Economics
Public Finance
Economic policy
Applied Economics
Public health systems and servic...
Stabilization policy
Private Consumption
Finance and Investment Banking
consumption sociology
Ricardian equivalence
Cited by
Review of Ricardian Equivalence in Theory and Practice: Empirical Data from Nigeria
Ridwan Dairo
Applied Journal of Economics, Management and Social Sciences
This study investigates the Ricardian Equivalence (RET) in theory and practice particularly as it relates to Nigeria economy. The study employed Autoregressive Distributed Lagged (ARDL) model to establish both the long-run and short-run relationship between deficit financing and consumption. The study found no strong evidence to reject the Ricardian Equivalence using data from Nigeria economy contrary to most literatures reviewing RET in Nigeria. Specifically, the study found that deficit financing variables like debt, tax revenue, and government expenditure have significant impact on consumption when the strict assumption of RET is not introduced in the model but became insignificant when ratio of tax revenue to changes in debt is introduced in the model. The study therefore concludes that Ricardian Equivalence is valid in the case of Nigeria when strict assumption of RET is maintained but insignificant when the major assumption of RET is relaxed or when deficit finance variables e...
Download free PDF
View PDF
chevron_right
Explore
Papers
Topics
Features
Mentions
Analytics
PDF Packages
Advanced Search
Search Alerts
Journals
Academia.edu Journals
My submissions
Reviewer Hub
Why publish with us
Testimonials
Company
About
Careers
Press
Content Policy
580 California St., Suite 400
San Francisco, CA, 94104