Paper Prepared for Presentation at the 8

th
International Conference
on Macroeconomic Analysis and International Finance
27-29 May 2014
ethymno! Crete
C"#$%MP&I"# A#' (%')*& '*FICI&$ I# &A#$I&I"#A+
*C"#"MI*$, '"*$ &-* ICA'IA# *.%I/A+*#C* -"+'0
*/I'*#C* F"M PA#*+ 'A&A
Nicholas Apergis* Katerina Lyroudi** Ioannis Papanastasiou**
A1stract, This paper investigates the existence and direction of the relationship
between consumption and budget deficits !sing a panel from twelve different
transitional countries" from #$%& to #$$$" it applies panel cointegration and causality
methodology This study finds statistical evidence for a long run relationship between
these two variables In addition" the empirical findings do not support the hypothesis
of the 'icardian ()uivalence" since budget deficits affect consumption negatively
*urthermore" it firmly re+ects the hypothesis that consumption causes budget deficits
in the short,run
2*+ classification n3m1ers- (.#" /0."P.&" 122
4ey 5ords- 1onsumption" 3udget 4eficits" 'icardian ()uivalence" Transitional
(conomies" Panel 4ata
Ac5nowledgments- The authors wish to express their gratitude to Prof Pedroni for
providing the software that estimates the cointegration results
66666666666666666666666666666666666666666666666666666666666666666666
* Nicholas Apergis" 4epartment of International (uropean (conomic and Political
7tudies" !niversity of 8acedonia" Thessaloni5i" 9reece email-napergis:uomgr
** Katerina Lyroudi and Ioannis Papanastasiou" 4epartment of Accounting and
*inance" !niversity of 8acedonia" Thessaloni5i" 9reece email-lyroudi:uomgr
C"#$%MP&I"# A#' (%')*& '*FICI&$ I# &A#$I&I"#A+
*C"#"MI*$, '"*$ &-* ICA'IA# *.%I/A+*#C* -"+'0
*/I'*#C* F"M PA#*+ 'A&A
16 Introd3ction
3arro ;#$<=> demonstrated that budget deficits ;caused by a reduction in taxes today
in exchange for future tax increases of e)ual present value> are expected to cause no
changes in private consumption 3arro?s arguments are based on the theory of 'icardian
()uivalence" that the effect of government spending is independent of how it is financed
In other words" private agents recogni@e that the reduction in taxes today are expected to
increase future tax liabilities and thus they will save the entire tax cut" leaving private
consumption unchanged The most important assumption for the 'icardian e)uivalence
to be satisfied is that consumers are not li)uidity ;or credit> constrained
/owever" for the hypothesis to hold certain assumptions must also be absent" ie the
presence of borrowing constraints" the presence of distortionary taxes" and that
households should not be very altruistic An the other hand" the Keynesian proposition
supports that fiscal policy can affect the national output and that an increase in the
budget deficit leads to an increase in the real domestic product and in private
consumption
1ertain empirical measures of the validity of the 'icardian ()uivalence hypothesis
have concluded that consumption does not respond to budget deficit changes BKormendi
;#$%2>" 7eater and 8ariano ;#$%C>" Kormendi and 8eguire ;#$$C>D 3y contrast" studies
by *eldstein ;#$%.>" 8odigliani and 7terling ;#$$&>" 9raham ;#$$C>" (vans ;#$$2>"
1ardia ;#$$<>" Aoms ;#$$<>" (lmendorf and Liebman ;.&&&> and by 9ale and Potter
;.&&.> have found evidence against the hypothesis 1ardia ;#$$<> argues that the
conflicting empirical evidence on the hypothesis is probably due to the presence of
certain wea5nesses in the statistical methodology followed *inally" /a)ue ;#$%%> and
9upta ;#$$.> have found empirical support for the 'icardian ()uivalence behavior for
the case of developing countries
2
2
The purpose of this study is to test empirically the validity of the 'icardian
()uivalence /ypothesis" for the transitional economies and to examine any causality
effects between consumption and budget deficits The results of this test will be a
significant guide for the governments in the transition economies to help them determine
their optimal policy for the growth of their countries
The ma+or contribution of this study to the relevant literature is that for the
first time the existence and most importantly the direction of a relationship between
consumption and budget deficits in transitional economies is investigated by applying
the novel methodology of panel cointegration and causality There are strong reasons for
believing that there is significant heterogeneity in cross,country consumption,budget
deficits relationship and that no panel data estimations will lead us to misleading
inferences due to the neglect of such heterogeneity Applying developed panel
cointegration techni)ues will allow us to ta5e into consideration the presence of
heterogeneity in the estimated parameters and dynamics across countries This will
enable us to generate more credible results since panel data estimation enables a
researcher to capture certain interesting time,series relations that only cross,sectional
analysis cannot do it This approach uses multi,country panel data in order to exploit
both time,series and cross,sectional information Panel unit root and cointegration tests
allow for both parameter and dynamic heterogeneity across groups" which have been
shown to generate more powerful results ;/arris and T@avalis" #$$$>
*or the empirical purposes of our paper and due to the lac5 of more reliable data a
version of the simple model proposed by 7ummers and 1arroll ;#$%<> and 7mith ;#$$&>
is adopted According to their model" consumption seems to be affected by budget
deficits" disposable income" the index that describes the age distribution of the
population" expected inflation" and the real interest rate /owever" due to certain
difficulties in having reliable data for all the above variables" our analysis steps down to
include only three variables" ie consumption" budget deficits" and income
In order to achieve our ob+ective" the paper is structured as follows- The next presents
a brief section review of literature regarding tests of 'icardian e)uivalence The third
section contains a description of the data" the model and the methodology The fourth
section presents and analyses the results The final section contains a summary and
concluding remar5s
3
3
26 icardian *73i8alence and *mpirical Findin9s
7ince our study examines a sample of the transitional economies we focus our
literature review on the effects of budget deficits on private consumption for developing
and transitional economies Ee have mentioned briefly in the introduction some ma+or
studies regarding the validity or not of the 'icardian e)uivalence for developed
economies *ull 'icardian e)uivalence entails that an increase in budget deficits should
have no impact on output and private consumption 8odified or extreme 'icardian
e)uivalence entails that an increase in budget deficits should have a negative impact on
interest rates" output and private consumption BEheeler ;#$$$>D
The absence of perfect capital mar5ets in developing countries is the most
)uoted reason for the 'icardian e)uivalence not to hold B/a)ue and 8ontiel ;#$%$>D
/owever" imperfections in the capital mar5ets are not restricted to developing countries
only *urthermore" since the basic assumption of 'icardian e)uivalence is that
consumers are not li)uidity or credit constrained" this might not hold for developing and
transitional economies /ence" it could be expected for the 'icardian e)uivalence to be
invalid in these countries
Theoretically" Kimball and 8an5iw ;#$%$> analy@ed the effects of
government debt and income taxes on consumption and saving behavior of individuals
for an infinite time hori@on instead of the usual two,period example The results
indicated that the timing of labor income taxes affects the consumption negatively or
positively 3ec5er ;#$$C> showed that as budget deficits change" private consumption
also changed depending on how ris5 averse the individuals were /is results supported
the 'icardian e)uivalence for some specific utility functions" while they supported the
Keynesian proposition for some other utility functions
'egarding small open economies" 'oc5" 1raigwell and 7ealy ;#$%$>
examined empirically the validity of the 'icardian e)uivalence proposition as a
description of consumer behavior for two similar small open economies" the case of
Trinidad and Tobago and of 3arbados" The results indicated that for the case of Trinidad
and Tobago the data did not support the 'icardian e)uivalence null hypothesis" while for
the case of 3arbados" the data supported the +oint null hypothesis of 'icardian
e)uivalence and rational expectations The authors concluded that the validity
determination of the 'icardian e)uivalence is an empirical issue and that the irrelevance
of the government?s debt,tax choice should be reconsidered and not ignored 'egarding
another small open economy" 9reece" 4ra5os ;.&&#> explored the long,run relationship
4
4
between government borrowing and private savings /is results indicated that as the
budget deficit increased" households perceived the government bonds as net wealth"
hence they increased their consumption" without considering the uncertainty of the
future tax level Famvou5as ;.&&#> and ;.&&.> also tested the 'icardian e)uivalence
and the Keynesian proposition for 9reece using first cointegration analysis and then
7!'( analysis In both cases he found support for the latter
'egarding developing economies" /a)ue ;#$%%> and 9upta ;#$$.> have found
empirical support for the 'icardian ()uivalence behavior for the case of developing
countries 9hata5 and 9hata5 ;#$$0> examined the validity of the 'icardian e)uivalence
for India and found no support for this proposition An the other hand" Khalid ;#$$0>
found support for the 'icardian e)uivalence for #. out of the #< developing countries in
his sample 4alamagas ;#$$.a> and ;#$$.b> found no support for this proposition for a
sample of both developed and developing economies /owever" when he split his
sample into two groups according to the si@e of their government debt" his results for the
group with the high indebtness gave support to the validity of the 'icardian ()uivalence
This group was consisted of the developing countries in his sample /owever" we cannot
generali@e based on these results
In contrast" we should mention here that for Gapan" a well developed economy" during
the past decade #$$&,.&&&" there was observed the highest budget deficit of any
industrial country Eal5er ;.&&&> found that with respect to taxes" there was strong
support for the 'icardian proposition since" the changes in taxes had @ero or a negligible
effect on output An the other hand" increases in government spending and budget
deficits affect positively both output and private consumption The implications for
Gapan are that a tax cut would provide a low stimulus or none on output An the
spending side" an increase in government debt will cause a positive impact on output and
consumption" but at a diminishing rate
Astry ;#$$<> applied a consumption smoothing model to five countries from
Asia and 8iddle (ast /is results indicated that the widening of external imbalances was
not influenced by excessive private consumption for most of the selected countries" with
the exception of Indonesia and 8alaysia" to a small degree The variables influencing
large external deficits were found to be the level and composition of external liabilities"
the flexibility of macroeconomic policies and the health of the country?s ban5ing system
Issler and Lima ;.&&&> examined the effects of public debt on consumption behavior in
3ra@il Their results indicated that the behavior of a HrationalI consumer in 3ra@il may
5
5
be consistent with 'icardian e)uivalence The budget deficits are financed and balanced
entirely through changes in taxes An the other hand" 4omenech" Taguas and Farela
;.&&&> tested the 'icardian e)uivalence for a panel of A(14 countries and their results
invalidated it" since private savings compensated only a small portion of the budget
deficit
8ore recently" 9iorgioni and /olden ;.&&2> assessed whether the 'icardian
e)uivalence held for ten developing economies ;3urundi" (l 7alvador" (thiopia"
/onduras" India" 8orocco" Nigeria" Pa5istan" 7ri Lan5a and Jimbabwe> Their results
indicated that for these countries the 'icardian e)uivalence was valid" since there was
observed a negative relationship between consumption and budget deficits" although not
always significant
The empirical evidence is ambiguous for both developed and developing
economies" ma5ing the results of this study very significant by enriching the relevant
literature" as well as by presenting insights for the policy ma5ers of the examined
economies
:6 &he 'ata
The time period under examination is from #$%& to #$$$ for the following countries-
Albania" 3ulgaria" 1roatia" (stonia" /ungary" 1@ech 'epublic" Latvia" Poland" 'omania"
'ussia" 7lovenia" and !5raine Annual data on the following variables were obtained-
private consumption is proxied by consumer expenses on goods and services" while real
income is proxied by the value of per capita 94P ;94P divided by total population> The
variable of consumption was also set in real terms ;c> by dividing it by the consumer
price index ;1PI> *inally" data on public deficits ;or surpluses" 34> were also used
They were measured as a percentage of 94P The real variables are calculated according
to #$$C prices and exchange rates" and the unit is expressed in millions of !7 dollars All
data come from the Eorld 3an5 Throughout the paper" small letters indicate variables in
logs All the estimates were done using 'AT7 software
46 *mpirical Analysis
A. Panel Integration Analysis
The null hypothesis of non,stationarity versus the alternative that the variable is
stationary is tested using the group mean panel unit root test ;or Kt,barK test> of Im" et al.
;#$$C" #$$<> This test is based on the Augmented 4ic5ey,*uller ;A4*> statistic for each
6
6
country ;4ic5ey and *uller" #$%#> and allows each member of the cross section to have a
different autoregressive root and different autocorrelation structures under the alternative
hypothesis The results are reported without and with a trend and are presented in Table
# The hypothesis that variables c" y" and 34 ;in levels> contain a unit root cannot be
re+ected at the #L significant level Ehen first differences are used" unit root
nonstationarity is re+ected at the #L significant level" suggesting that the variables under
study are I;#> variables These results open the possibility of cointegration among certain
variables
B. Dynamic Heterogeneity
An issue that it is of ma+or concern is the heterogeneity of the countries included in
this data set In particular" through time and across countries" the effects on the
consumptionMincome,budget deficits relationship of the different macroeconomic
policies implemented" as well as the effects of the institutional framewor5s established in
each country should be expected to be diverse Although there has been a paucity of
relevant studies for the transitional economies certain explanations could be offered
+ustifying the presence of consumption heterogeneity in transitional economies
/eterogeneity could be generated by the fact that countries" especially transitional
countries" are characteri@ed by heterogeneous sensitivity of consumption to budget
deficits This latter differentiation of sensitivity is probably attributed to the fact that
government debt levels should be either low or high B4alamagas ;#$$2>D Table .
presents data ;available from International *inancial 7tatistics of I8*> on the percentage
;with respect to 94P> of borrowing from capital mar5ets These data show the above
mentioned type of heterogeneity among the transitional economies under study This
type of heterogeneity is due to the differentiation of debt leverage levels The presence of
different levels of government debt tends to contribute to the departures from the
predictions of the 'icardian e)uivalence hypothesis" albeit on a differentiated manner In
other words" consumption is expected to be responding differently in a country that
borrows heavily from capital mar5ets compared to a country that borrows less from
capital mar5ets 3ecause of this" the coefficients in the estimated relationship will be
biased due to a heterogeneity bias
Another potential explanation could be the different degree of imperfections in the
transitional economies? capital mar5ets B3acchetta and 9erlach ;#$$<>" /ahm ;#$$%>D In
other words" cross,country differences could reflect the degree of li)uidity constraints"
7
7
eg indicating differentiation of the degree of availability of consumer credit" sector
credit" interest rates charged on consumer groups and so on In addition" following the
change in the political regime in the countries under study from early $&s" it is highly
li5ely that the effects of li)uidity constraints might have changed over time *inally" a
potential explanation for the heterogeneous nature of consumption is the differentiated
degree of li)uidity,constrained consumers vis,N,vis the rule,of,thumb consumers The
former save when their current income is high even though they are constrained from
borrowing when current income is low" while the latter consume +ust their current
income without borrowing or saving to smooth consumption B7eater ;#$$<>D
In the statistical framewor5 of this study" these issues can be resolved by first testing
for heterogeneity and then by controlling for it through appropriate techni)ues The
dynamic heterogeneity" ie variation of the intercept over countries and time" across a
cross,section of the relevant variables can be investigated as follows In the first step" an
A4*;n> e)uation for each relationship in the panel is estimatedO then" the hypothesis of
whether regression parameters are e)ual across these e)uations is tested Next" a similar
test of parameter e)uality is performed by estimating a n,order autoregressive model for
each of the relationships under investigation 7tandard 1how,type * tests under the null
of parameter e)uality across all relationships are also performed /eterogeneity in cross,
sectional parameters is indicated if the results re+ect the null hypothesis *inally"
homogeneity error variance across groups is also examined as another measure of
dynamic heterogeneity EhiteKs tests for group,wise heteroscedasticity are employed to
serve this ob+ective
The results of this procedure are reported in Table 2 for the relationship between
consumption" income and budget deficits The empirical findings indicate that the
relationship under investigation is characteri@ed by heterogeneity of dynamics and error
variance across groups" supporting the employment of panel analysis
C. Panel Cointegration Analysis
Ance the order of stationarity has been established" one can move to a panel
cointegration approach" developed by Pedroni ;#$$<" #$$$> The panel cointegration
techni)ue ma5es use of a residual,based A4* test The specific cointegrating
relationships estimated are-
8
8
Cons3mption! income! and 13d9et deficits relationship
c
it
P
&i
Q
#i
y
it
Q
.i
34
it
Q
it
;#>
where i P # R N countries and t P # R T year observations The term
it
is the
deviations from the modelled long,run relationship If the series are cointegrated" this
term will be a stationary variable Thus" stationarity can be achieved by establishing
whether
i
in-

it
P
i

i;t,#>
Q
it
;.>
is unity The null hypothesis" associated with PedroniKs test procedure is that
i
P # This
implies that the null hypothesis associated with PedroniKs test procedure is e)uivalent to
testing the null of nonstationarity ;no cointegration> for all i Pedroni ;#$$<" #$$$>
developed four panel cointegration statistics and three group mean panel cointegration
statistics
The Pedroni cointegration results are reported in Table = The results re+ect the null
hypothesis of no cointegration" confirming that in both testable relationships the panel is
stationary Thus" the results indicate that the variables share a long,run cointegrating
relationship 1onclusively" these findings provide support to a strong long,run
relationship between consumption and budget deficits and income and budget deficits A
similar result was found by /o ;.&&#> using also cointegration analysis on panel data for
.= A(14 countries An increase in government spending caused a decrease in private
consumption and income
9iven cointegration" we estimate the long,run relationship through the 4ynamic AL7
;4AL7> approach provided by 7toc5 and Eatson ;#$$2> This approach regresses a I;#>
variable on other I;#> variables plus lags and leads of the first,differences of the I;#>
variables The inclusion of the first,differenced variables eliminates any possible bias
resulting from correlation between the error term and the I;#> variables Ee also
calculate corresponding robust standard errors through an ad+ustment suggested by
Newey and Eest ;#$%<>" with 3artlett weights and a truncation lag of 2 The 4AL7
regression is employed" by adding one lag and one lead of the first difference of the
right,hand side variable to the e)uation
c
it
P &&C< Q &20= y
it
M &=<% 34
it

;2<0>* ;=C%>* ;2%.>*
'
.
P &0=%
*
2
c
P<<&$B&&&D
9
9
where the *,test indicates that the coefficients are +ointly significant across countries
*igures in brac5ets indicate p,values" while an asteris5 denotes significance at #L The
empirical findings show that both the income coefficient and the budget deficit
coefficient are statistically significant at the #L level 8oreover" the budget deficit
coefficient is negative" indicating that consumption responds negatively to budget
deficits" which in itself is supportive of 'icardian ()uivalence behavior In other words"
the empirical findings provide support for the expected positive correlation between
consumption and income as well as for the negative correlation between consumption
and budget deficits

for the entire sample. In terms of the latter" budget deficits exert
lower consumption expenditures for the transition economies *inally" the fact that the
coefficient of budget deficits is negative implies that the impact of debt,ridden countries
prevails In other words" the consumers in those countries fully reali@e the unfavorable
prospects of their future standards of living and they set downward their consuming
behavior in order to avoid drastic cuts in their future spending
D. Panel Causality
As cointegration is confirmed" we proceed to estimate causality using the Pooled
8ean 9roup ;P89> estimator of Pesaran" 7hin and 7mith ;#$$$> to account for the
panel data causal relationships This estimator is suitable when variables are
cointegrated This provides +ustification for examining the direction of the causal lin5s
among the variables under consideration through an error correction FA' ;(1FA'>
model The P89 estimator was obtained through the methodological software analysis
proposed by Pesaran" 7hin and 7mith ;#$$$>
/aving established that consumption is cointegrated with income as well as budget
deficits" it is also appropriate to examine the associated multivariate causality
relationship 1onsidering that the cointegrating e)uation is-
c
it
P
&
i Q
#i
y
it
Q
.i
34
it
Q u
it
;2>
and the associated autoregressive distributed lag A'4L e)uations are described by a
;#"#"#> model-
c
it
P
i
Q
#&i
y
it
Q
##i
y
i"t,#
Q
.&i
34
it
Q
.#i
34
i"t,#
Q
i
c
i"t,#
Q
#it
;=>
and
34
it
P
i
Q
#&i
c
it
Q
##i
c
i"t,#
Q
.&i
y
it
Q
.#i
y
i"t,#
Q
i
34p
i"t,#
Q
.it
;C>
the error correction e)uations yield-
10
10
The error,correction coefficients ;s> are negative but statistically significant only in the
case where consumption is the dependent variable" indicating" that budget deficits do
have causal effects on consumption but not the other way around There are not any
feedbac5 effects between consumption and budget deficits The results imply that
consumption seems to respond to budget deficit changes" thus" providing evidence
against the 'icardian ()uivalence hypothesis BEheeler ;#$$$>D
;6 Concl3sions and Policy Implications
This paper investigates the existence and direction of a relationship between
consumption and budget deficits since it is a topic with ma+or political and economic
implications !sing a panel from twelve different transitional countries" from #$%& to
#$$$" it applies panel cointegration and causality methodology The study finds statistical
evidence for a long,run relationship between consumption and budget deficits In terms
of causality" the empirical findings seem to invalidate the 'icardian ()uivalence
hypothesis 8oreover" our results firmly re+ect the hypothesis that consumption causes
budget deficits The empirical findings suggest that the invalidation of the 'icardian
()uivalence hypothesis offsets a portion of the effectiveness of fiscal policy" rendering it
not so powerful to encourage aggregate demand In other words" these new transitional
economies should learn from the experience of the developed economies and should
attempt to enlarge their arsenal of affecting the course of the real sector" eg
unemployment" and not +ust giving fiscal policy too much credit to do the +ob
11
11
REFERENCES
3acchetta" P and 9erlach" 7 ;#$$<> 1onsumption and 1redit 1onstraints-
International (vidence" Journal of Monetary Economics" =&" .&<,.2%
3arro" 'obert ;#$<=> Are 9overnment 3onds Net EorthS" Journal of Political
Economy" %." #&$C,###<
3ec5er" T ;#$$C> 3udget 4eficits" Tax 'is5 and 1onsumption" Eor5ing Paper No <="
Working Paper eries in Economics an! "inance# 7toc5holm 7chool of
(conomics
1ardia" (manuela ;#$$<> 'eplicating 'icardian ()uivalence Tests with 7imulated
7eries" American Economic $e%ie&" %<" 0C,%<
4alamagas" Fasilios ;#$$.a> /ow 'ival are the 'icardian ()uivalence Proposition
and the *iscal Policy Potency FiewS" cottis' Journal of Political Economy" 2$" ="
=C<,=<0
4alamagas" Fasilios ;#$$.b> Testing 'icardian ()uivalence- A 'econsideration"
Applie! Economics" .=" #" C$,0$
4alamagas" Fasilios ;#$$2> /ow (fficient is the 7ubstitution of 4ebt for Taxes in
Influencing 4emandS" Applie! Economics" .C" .$C,2&2
4ic5ey" 4avid A and *uller" Eayne A ;#$%#> Li5elihood 'atio 7tatistics for
Autoregressive Time 7eries Eith a !nit 'oot" Econometrica" =$" #&C<,#&<.
4omenech" ' " Taguas" 4 and Farela" G ;.&&&> The (ffects of 3udget 4eficit on
National 7aving in the A(14" Economic (etters" 0$" 2" 2<<,2%2
4ra5os" Kostas ;.&&#> Testing the 'icardian ()uivalence Theorem- Time 7eries
(vidence from 9reece" Journal of Economic De%elopment" .0" #=$,#0&
(lmendof" 4ouglas" E and Liebman" Geffrey" 3 ;.&&&> 7ocial 7ecurity 'eform and
National 7aving in an (ra of 3udget 7urpluses" Brookings Papers on Economic
Acti%ity" #,<#
(vans" Paul ;#$$2> 1onsumers are not 'icardian- (vidence from Nineteen 1ountries"
Economic In)uiry" 2#" C2=,C=%
*eldstein" 8artin ;#$%.> 9overnment 4eficits and Aggregate 4emand" Journal of
Monetary Economics" $" #,.&
9ale" Eilliam" 9 and Potter" 7amara" ' ;.&&.> An (conomic (valuation of the
(conomic 9rowth and Tax 'elief and 'econciliation Act of .&&#" *ational +a,
Journal" C=" #22,#%0
9hata5" A and 9hata5" 7 ;#$$0> 3udgetary 4eficits and 'icardian ()uivalence- The
1ase of India" #$C&,#$%0" Journal of Pu-lic Economics" 0&" ." .0<,.%.
9iorgioni" 9 and /olden" K ;.&&2> 4oes the 'icardian ()uivalence Proposition
/old in Less 4eveloped 1ountriesS" International re%ie& of Applie! Economics"
#<" ." .&$,..#
9raham" *red" 1 ;#$$C> 9overnment 4ebt" 9overnment 7pending" and Private 7ector
3ehavior- 1omment" American Economic $e%ie&" %C" #2=%,#2C0
9upta" K L ;#$$.> 'icardian ()uivalence and 1rowding Aut in Asia" Applie!
Economics" .=" #$,.C
/ahm" G / ;#$$%> 1onsumption Ad+ustment to 'eal Interest 'ates- Intertemporal
7ubstitution 'evisited" Journal of Economic Dynamics an! Control" .." .$2,2.&
/a)ue" N ! ;#$%%> *iscal Policy and Private 7ector 7aving 3ehaviour in 4eveloping
(conomies" IM" taff Papers" 2C" 2#0,22C
/a)ue" N ! and 8ontiel" P ;#$%$> 1onsumption in 4eveloping 1ountries - Tests for
Li)uidity 1onstraints and *inite /ori@ons" +'e $e%ie& of Economics an! tatistics"
<#" 2" =&%,=#C
/arris" ' 4 * and T@avalis" ( ;#$$$> Inference for !nit 'oots in 4ynamic Panels
where the Time 4imension is *ixed" Journal of Econometrics $#" .&#,..0
/o" T ;.&&#> The 9overnment 7pending and Private 1onsumption- A Pan
1ointegration Analysis" International re%ie& of Economics an! "inance" #&" $C,
#&%
Im" Kyung 7" Pesaran" 8 /ashem" and 7hin Tonacheol ;#$$C> Testing for !nit
'oots in /eterogeneous Panels" Eor5ing Paper $C.0" 4epartment of Applied
(conomics" 1ambridge !niversity
Im" Kyung 7" Pesaran" 8 /ashem" and 7hin Tonacheol ;#$$<> Testing for !nit
'oots in /eterogeneous Panels" revised version of Eor5ing Paper $C.0"
4epartment of Applied (conomics" 1ambridge !niversity
Issler" G F and Lima" L ' ;.&&&> Public 4ebt 7ustainability and (ndogenous
7eigniorage in 3ra@il- Time 7eries (vidence from #$=<,#$$." Journal of
!e%elopment Economics" 0." #" #2#,#=<
Khalid" A 8 ;#$$0> 'icardian ()uivalence- (mpirical (vidence from 4eveloping
(conomies" Journal of De%elopment Economics" C#" =#2,=2.
Kimball" 8 and 9N 8an5iw" ;#$%$> Precautionary 7aving and the Timing of Taxes"
Journal of Political Economy" $<"=" %02,%<$
13
13
Kormendi" 'oger ;#$%2> 9overnment 4ebt" 9overnment 7pending" and Private 7ector
3ehavior" American Economic $e%ie&" <2" $$=,#&#&
Kormendi" 'oger and 8eguire" Philip ;#$$C>" 9overnment 4ebt" 9overnment
7pending" and Private 7ector 3ehavior- 'eply" American Economic $e%ie&" %C"
#2C<,#20#
8odigliani" * and A 7terling ;#$$&> 9overnment 4ebt" 9overnment 7pending and
Private 7ector 3ehavior- A *urther 1omment" American Economic $e%ie&# %&" 2"
0&&,0&2
Newey" E K and Eest" K 4 ;#$%<> A simple positive semi,definite
heteros5edasticity and autocorrelation consistent covariance matrix Econometrica"
.." <&2,<2.
Aoms" F4 ;#$$<> (conomic 9rowth" 3udgetary 3alance and #$$< *iscal Policy"
Business Economics" 2."=" ##,#<
Astry" G ;#$$<> 1urrent Account Imbalances in A7(AN 1ountries- Are They a
ProblemS" International Monetary "un! Working Paper- EPU$<UC<" A" pages .2
Pedroni" Peter ;#$$C> Panel 1ointegration- Asymptotic and *inite 7ample Properties
of Pooled Time 7eries Tests Eith an Application to the PPP /ypothesis" Indiana
!niversity" Eor5ing Paper in (conomics" No $C,&#2
Pedroni" Peter ;#$$<> Panel 1ointegration- Asymptotic and *inite 7ample Properties
of Pooled Time 7eries Tests Eith an Application to the PPP /ypothesis" New
'esults" mimeo" Indiana !niversity
Pesaran" /ashem" 8" 7hin" Tongcheol" and 7mith" 'on" P ;#$$$> Pooled 8ean 9roup
(stimation of 4ynamic /eterogeneous Panels" Journal of t'e American tatistical
Association" $=" 0.#,02=
7eater" Gohn" G ;#$$<>" An Aptimal 1ontrol 7olution to the Li)uidity 1onstraint
Problem" Economics (etters" C=" #.<,#2=
7eater" Gohn" G and 8ariano" 'oberto" 7 ;#$%C>" New Tests of the Life,1ycle and Tax,
4iscounting /ypotheses" Journal of Monetary Economics" #C" #$C,.#C
7mith" ' 7 ;#$$&> *actors Affecting 7aving" Policy Tools" and Tax 'eform- A
'eview" IM" taff Papers" 2<" C&=,C#<
7toc5" G and Eatson" 8 ;#$$2> A simple estimator of cointegrating vectors in higher
order operational systems Econometrica" /0" <%2,%.&
7ummers" L / and 1arroll" / ;#$%<> Ehy is !7 National 7aving so LowS"
Brookings Papers on Economic Acti%ity" 0&<,02C
14
14
Famvou5as" 9 A ;.&&#> The 'elationship 3etween 3udget 4eficits and 1urrent
Account" Arc'i%es of Economic History" #." C<,<0
Famvou5as" 9 A ;.&&.> 3udget 4eficits and Interest 'ates in a 7mall Apen (conomy,
Another Loo5 at the (vidence-'eply" International Economic Journal" #0" ." 2#,20
Eal5er" 1 ;.&&.> 'icardian ()uivalence and *iscal Policy (ffectiveness in Gapan" Asian
Economic Journal" #0" 2" .%C,2&.
Eheeler" 8 ;#$$$> The 8acroeconomic Impacts of 9overnment 4ebt- An (mpirical
Analysis of the #$%&s and the #$$&s" Atlantic Economic Journal" .<" 2" .<2,.%C
15
15