Economic Modelling 37 (2014) 120–126

Contents lists available at ScienceDirect

Economic Modelling

j ourna l homepage: www.e lsev ie r .com/ locate /ecmod

Fiscal deficit and inflation: New evidences from Pakistan using a boundstesting approach

Abdul Jalil, Rabbia Tariq ⁎, Nazia BibiSchool of Economics, Quaid-i-Azam University, Islamabad, Pakistan

⁎ Corresponding author.

0264-9993/$ – see front matter © 2013 Elsevier B.V. All rihttp://dx.doi.org/10.1016/j.econmod.2013.10.029

a b s t r a c t

a r t i c l e i n f o

Article history:Accepted 22 October 2013Available online xxxx

Keywords:Fiscal deficitInflationPakistan

Leeper (1991), Sims (1994) andWoodford (2001) point out that price level is not independently determined bymonetary policy rather it is the result of inter dependence of fiscal andmonetary policies. This article aims to testthe fiscal theory of price level for Pakistan using an autoregressive distributed lagmodel framework over the pe-riod of 1972–2012. The article finds that fiscal deficit is a major determinant of the price level along with othervariables like interest rates, government sector borrowing and private borrowing. On the basis of our findings,the present article suggests that the economy of Pakistan requires an immediate correction of fiscal imbalances.

© 2013 Elsevier B.V. All rights reserved.

1. Introduction

The proponents of monetarist doctrine, for example McCallum(1999, 2001, 2003) and Niepelt (2004), contradict the fiscal theory ofprice level developed by Leeper (1991), Sims (1994), and Woodford(2001) which states that government finances must be sustainable forthe stable price level in an economy. Earlier, a well-recognized workof Sargent andWallace (1981) documented that the governments run-ning with persistent deficits have to finance those deficits with moneycreation thus causing higher inflation. However, Catao and Terrones(2005) do not neglect the importance of other determinants whichare important for the fuelling of inflation.

The fiscal view of inflation gets a special attention in the case of de-veloping countries because it is generally accepted that the developingcountries have less efficient tax collection, political instability, and a lim-ited access to external borrowing (Alesina and Drazen, 1991; Calvo andVegh, 1999; Cukierman et al., 1992) and these tend to lower the relativecost of seigniorage and increase dependence on the inflation tax. Specif-ically, De Haan and Zelhorst (1990), Metin (1998), and Domac andYucel (2005) conducted their empirical research on developing econo-mies and argue that there is a significant relationship between fiscaldeficits and inflation in high-inflation countries.

Theoretical and empirical research has developed many channelsthrough which inflation is affected. The famous statement of Friedman(1956) inflation is always and everywhere a monetary phenomenonlinks the price fluctuation to monetary policy and tomoney supply spe-cifically. The increase inmoney supply is positively linkedwith inflationin many studies including Grauwe and Polan (2005) and Komulainenand Pirttila (2002). But money supply is not independently determined

ghts reserved.

by the central banks rather it is the financial requirements of the fiscalauthorities that induce more money supply (Sargent and Wallace,1988). Thusmoney supply is endogenous and ismostly the result of sei-gniorage requirements due to fiscal deficits of the governments.

In most of the developing countries, the fiscal authorities financetheir deficits by printingmoney through central bank and hence centralbanks are not independent in forming monetary policy. As mentionedearlier, Catao and Terrones (2005) point out that the other sources of fi-nancing like imposition of tax have high political cost and are not easy toimplement. Moreover the developing countries are already involved indebt servicing that new debt issue internally or borrowing from an ex-ternal source (internal or external borrowing) stands a low chanceand is very costly. Thus this fiscal decision of financing through centralbank leads to inflation in the economy. The fiscal theory of price levelpropagates that price level is influenced by interactions of fiscal andmonetary policies. Therefore, the government deficit must be sustain-able and inter temporal budget constraint of government must be bal-anced for stable prices (Leeper, 1991, Sims, 1994).

The main objective of this study is to investigate the underlying de-terminants of inflation other than justmonetary factors. It is also its pur-pose to see whether the chronic fiscal deficit has any bearing oninflation in Pakistan in the long run and what is the role of other factorsin causing inflation. Pakistan is a good candidate for testing the infla-tion–fiscal deficit nexus because over the last decade the central bankof Pakistan is continuously saying that the budgetary borrowing is themain source of inflation in the country. But, this area is not wellresearched in the case of Pakistan. Although few studies have shownthe monetary policy behind the inflation in Pakistan (Kemal, 2006;Khan and Schimmelpfennig, 2006; Qayyum, 2006) but these papersdid not incorporate the fiscal side. There are few studies like Shabbiret al. (1994), Chaudhary and Ahmad (1995), Agha and Khan (2006)which have shown a significant relationship between fiscal deficit and

121A. Jalil et al. / Economic Modelling 37 (2014) 120–126

inflation. Analyzing the inflation in Pakistan is also important as being adeveloping country it has suffered inflation that may negatively affectthe living standards and purchasing power of the vulnerable segmentsof society. Inflation also has a political cost as political governments can-not afford to allow undue increase in prices as it would have a negativeimpact on their vote bank. This has induced the need to find the underlying causes of inflation in Pakistan economy.

The present paper is also an attempt to test the inflation–fiscal defi-cit nexus in Pakistan by using the data from 1972 to 2012. We find apositive relationship between inflation and fiscal deficit. Therefore,this paper calls for fiscal consolidation to bring down the prices and todepend on less inflationary policies of financing deficit.

The rest of the paper which proceeds as Section 2 is literaturereview. Section 3 builds the theoretical framework and estimation strat-egy while Section 4 constructs the variables based on theory. Section 5presents the empirical results based on econometric models andSection 6 concludes the study along with policy suggestions.

1 However, some studies consider the importance ofmonetary factors in explaining theinflation of Pakistan (for example Khan and Schimmelpfennig, 2006, Malik and Khawaja,2006, Qayyum, 2006). These studies demand for tighter monetary policy to controlinflation.

2. Literature review

Indeed, Friedman (1956) documents that inflation is a monetaryphenomenon but recently Leeper (1991) and Sims (1994) give theidea of fiscal theory of price level (FTPL) which deems inflation to be afiscal phenomenon. The FTPL attributes interactions of fiscal andmone-tary policies to influencing price level and suggests that the governmentdeficit must be sustainable and inter temporal budget constraint of gov-ernment must be balanced. Therefore, the relative dominance of fiscaland monetary authorities is crucial in the inflationary impact of deficit.Importantly, Sims (1994) reinforces that in themost of cases inflation ismore of a fiscal phenomenon and is dependent upon the expectationspeople have regarding fiscal policy and fiscal deficit. However, FTPL isempirically tested for many countries that gave mixed results (Rubioet al. 2009).

The studies regardingunder developed countries showpositive rela-tionships between fiscal deficit and inflation. For instance in their casestudy of Nigeria Oladipo and Akinbobola (2011) observed a causalityrunning from fiscal deficit to inflation. While in Zimbabwe due to cur-rent and non-development expenditures of government, there is persis-tent fiscal deficit that is covered by seigniorage thus leading to inflation(Makochekanwa, 2011). For the Iranian economy, Mehdi and Reza(2011) concluded that deficit caused inflation significantly becausethe central bank of Iran is not independent in its decisions. Similarly, fis-cal dominance was observed in the case of the Italian economy andthere is an evidence of positive relation between fiscal deficits and infla-tion (Fratianni and Spinelli, 2001). Komulainen and Pirttila (2002) showthat fiscal deficit is a weak determinant of inflation in the cases ofBulgaria, Romania and Russia. But, Tekin-Koru and Erdal (2003) findno clear evidence regarding a possible relationship between inflationand budget deficit for the Turkish economy.

A panel data study of selected SAARC countries also rejects FTPL(Nawaz et al. 2012). Although the negative and significant effect of fiscaldeficit on priceswas observed using a pooled least squaremethod thereis no such evidence in fixed and random effects models. Sahan andBektasoglu (2010) test this relationship for European countries and con-clude that there is no standardized relationship however there is a longrun co-integrating relationship between deficit and inflation.

A more recent observation is that fiscal deficit is strongly related toinflation in developing countries with a long history of high inflation(Lin and Chu, 2013). Moreover fiscal deficit has an impact on long runinflation in countries with moderate inflation, however there is nolong run impact on inflation in developed countries with long historyof low single digit inflation. The weak relationship in developed coun-tries is due to greater monetary policy autonomy and credibility.While underdeveloped countries usually lack strong institutions andtheir modes of financing the deficit are inflationary. Hence there is a

dynamic non-linear and heterogeneous relationship between fiscal def-icit and inflation (Catao and Terrones, 2005, Lin and Chu, 2013).

Furthermore, the choice of fiscal deficit measure can also be crucialwhile testing the effect of budget deficit on inflation. Pekarski (2011)was of the view that fiscal deficit can be divided into two parts; onethat causes inflationary effect and other that does not. The literatureshows that it's the consumption component of government expenditurethat leads to fiscal deficit growth in long run while the investment ex-penditures are more sustainable in long run (Tiwari et al. 2012).

Fiscal deficit is not the only determinant of inflation in existing liter-ature rather the factors like oil prices, food prices, exchange rate, tradeopenness and growth rate of economy are also considered for affectingthe inflation levels. Hanif (2012) investigates the food inflation inPakistan and observed that it is half as volatile in Pakistan as globallyand the poor class is worst affected by food inflation in Pakistan.Coppin (1993) finds that level of tourism, interest rate and imported in-flation are the key determinants of inflation in Barbados. Durevall(1998) observes that inflation increases when the rate of devaluationof the exchange rate increases, and inflation decreases when outputgrowth goes up in the case of Brazil. The political instability determinesthe variation in policy about inflation and the effect of political instabil-ity is greater for developing and high inflation countries (Aisen andVeiga, 2008). Trade openness is also regarded as a good policy to controlinflation. The negative relationship between trade openness and infla-tion is suggested by Romer (1993). While the time series analysisshowed that along with fiscal and monetary policies as the main deter-minants of inflation more trade openness is also causing the inflation inPakistan (Zakariya, 2010). This might be because Samimi et al. (2012)have empirically shown that a conventional measure of trade openness,as used in the study by Zakariya (2010), is positively related with infla-tion and more openness to trade in fact raises inflation if measured bythis trade openness measure.

The high fiscal deficits have caused inflation in Pakistan’s economyand these deficits are unsustainable. The existing literature for Pakistanin this regard has more or less similar results (see for example Shabbiret al. 1994, Agha and Khan, 2006).1 Therefore, there is a need of somefresh evidences in the backdrop of high fiscal deficit and rising inflationin the case of Pakistan.

3. Analytical framework and estimation strategy

The classical quantity theory provides an explanation for the fluctu-ations in prices. For example Fisher (1911) shows that since velocity ofmoney circulation is exogenously controlled, therefore changes inmoney supply cause changes in prices and any increase in aggregatedemand is translated into inflated prices. Consequently, the major de-terminant of inflation is considered to be the monetary side and fiscalpolicy has no independent impact on price level. Keynesian theoriesare considered to be applicable in the short run. Unlike classical theoriesthe demand side policies are considered to be effective in altering theoutput level. The fiscal deficit arising from increased expenditure orcut in taxes would lead to an increase in aggregate demand. This in-creased demand would increase output only if economy is below fullemployment. If economy is already operating at full employment the re-sult will be the increase in price level.

In the Keynesian theory when government finances deficit throughborrowing then:

And when government finances deficit by monetizing from centralbank.

Hence both channels are showingfiscal deficits that lead to inflation.

2 Typically CPI, wholesale price index (WPI) and GDP deflator are taken as measures ofinflation as commented by one of the reviewers. But we deliberately drop of idea of usingWPI and GDP deflator, because Pakistan is a net importer country and its basket of con-sumption includes a number of commoditieswhich are not domestically produced. There-fore, it can underestimate the impact of fiscal deficit on the inflation of Pakistan.

122 A. Jalil et al. / Economic Modelling 37 (2014) 120–126

The most recent theory in determining price level says that pricelevel is not independently determined by monetary policy but ratherthe result of interdependence of fiscal and monetary policies (Leeper,1991, Sims, 1994). Fiscal authorities try to satisfy inter-temporal budgetconstraints and thus in the process cause inflation. Infiscal dominant re-gimes, monetary policy acts accordingly to finance the fiscal require-ments and thus affecting the price level. Now if there is monetarydominance and central bank is independent in making policies, it willput a constraint on government financing through currency printing.Thus fiscal consolidation is possible and inflation can be targeted.Hence in fiscal dominance regimes the price level is mostly influencedby fiscal budget positions. Thus the relative dominance of fiscal andmonetary policy determines the inflationary impact of fiscal deficit. Indeveloping countries due to fiscal dominance and relative ease of fi-nancing by seigniorage, inflation is more of a fiscal phenomenon. Theexpectations regarding inflation are based upon fiscal decisions of thegovernments which also contribute to inflation.

There are several other variables which may affect the level of infla-tion in the case of Pakistan. For example, the cost of borrowing capital,which is commonly known as interest rate, is one of the importantsources of inflation from the cost push side. Specifically, Hasan (1999)tested the Fisher effect in the case of Pakistan and comes to the conclu-sion that interest rates do respond to inflation but not one to oneand the real interest rate does not remain constant as the prediction ofFisher's hypothesis. Similarly, Kandil (2005), Boschi and Girardi(2007) and Kose et al. (2012) also document that the domestic interestrates are the determinants of inflation.

Then Romer (1993) suggests that more trade openness leads tolower inflation which was empirically tested by Catao and Terrones(2005). The conclusion was that trade openness accounts for inflationin developed countries only and not for all countries. Bowdler andNunziata (2006) founded that greater trade openness reduces the prob-ability of inflation in OECDmembers. Lin and Chu (2013) also showed astrong negative relationship between trade openness and inflation inhigh inflation episodes while weak relationship in low inflation cases.Furthermore, Samimi et al. (2012) have empirically shown that a con-ventional measure of trade openness as used in the current study aswell is positively related with inflation and more openness to trade infact raises inflation if measured by this trade openness measure. Thiswas observed by Zakariya (2010) for the economy of Pakistan.

The volume of trademay be increased only due to the increase in im-ports. Therefore, the import prices are an important factor which maypredict the level of inflation. Specifically, Deme and Fayissa (1995),Darrat (1997) and Boujelbene and Boujelbene (2010) test the impactof import prices on the level of inflation and find a positive relationshipbetween two variables. This variable is also important in the case ofPakistan, because the country is a net importer. Then Deme and Fayissa(1995), Darrat (1997), Lui and Adedeji (2000), Boschi and Girardi(2007), El-Sakka and Ghali (2005), Pelipas (2006) and Boujelbeneand Boujelbene (2010) also take the exchange rate into account forexplaining the inflation.

Some important country cases, for example Naqvi et al. (1994),Hasan et al (1995) and Bokil and Schimmmelpfenning (2005) forPakistan, Callen and Chang (1999) for India, Leigh and Rossi (2002)for Turkey, Chauvet(2000) for Brazil, Sun (2004) for Thailand, Simone(2000) for Chile and Boujelbene and Boujelbene (2010) for Tunisia, oninflation provide the empirical models which incorporate both demandside and supply side factors alongwith the policy variables and adaptiveexpectations. We specify our regression keeping in view the determi-nantswhich are identifiedby the literaturewith giving special emphasisto the impact of fiscal deficit on inflation in Pakistan.

lcpit ¼ α0 þ α1lf dt þ α2lrt þ α3ltradet þ α4ler þ α5lZtμ t ð1Þ

Where, lcpi is log of consumer price index used tomeasure inflation,lfd is log of fiscal deficit, lr is lending rate, ltrade is log of trade openness,

leris log of exchange rate and Z contains the several other control vari-ables which are specifically linked with the dynamics of inflation inPakistan like government sector borrowing, real demand relative toreal supply, private sector borrowing, import prices, governmenttaxes, adaptive expectations and wheat support prices.

Eq. (1) cannot be estimated directly through Ordinary Least Squaremethod because it is the time series data and literature points out thatthose spurious results may exist when variables are specified in thelevel or non-stationary form. Therefore, the econometricians suggestthat one must test to determine whether a long run relationshipexists among the variables in the model. For this purpose we useAutoregressive Distributed Lag model (ARDL) due to its several welldocumented advantages in the literature (see Banerjee et al. 1993;Laurenceson and Chai, 2003, Pesaran and Pesaran, 1997, Pesaran andShin, 1999). For example it can be applied in the presence of I(0) orI(1) data series; it provides super consistent results in a small samplecase and endogeneity is nomore a problemwhenwe specify the regres-sion equation in the ARDL framework. The ARDL framework of Eq. (1) isas follows:

Δlcpii ¼ β0þXp

i¼1

ψiΔlcpit‐iþXp

i¼1

�iΔlf dt‐iþXp

i¼1

viΔlrt‐iþXp

i¼1

γiΔltradet‐i

þXp

i¼1

δiΔlZt‐iþθ1lcpit‐1þθ2lfdit‐1þθ3lrt‐1þθ4ltradet‐1þθ4lZt‐1þUt

ð2Þ

where β0 is the drift component and Ut white noise. Furthermore, thetermswith summation signs represented the error correction dynamics.While the second part of the equation with θi corresponds to long runrelationship.

This is an error correction representation. So, the following error cor-rection model is estimated in the third step.

Δlcpii¼β0þXp

i¼1

ψiΔlcpit‐iþXp

i¼1

�iΔlf dt‐iþXp

i¼1

viΔlrt‐iþXp

i¼1

γiΔltradet‐i

þXp

i¼1

δiΔlZt‐iþαECMt‐1þUt

ð3Þ

The error correction model result indicates the speed of adjustmentback to long run equilibrium after a short run shock.

To ensure the goodness of fit of model, the diagnostic and stabilitytests are also conducted, the diagnostic test examines the serial correla-tion, functional form, normality and heteroscedasticity associated withselected model. Pesaran and Pesaran (1997) suggest using cumulative(CUSUM) and cumulative sum of squares (CUSUMSQ). The CUSUMand CUSUMSQ statistics are updated recursively and plotted againstthe break points. If the plots of CUSUM and CUSUMSQ statistics staywithin the critical bonds of 5% level of significance, the null hypothesisof stable coefficients in the given regression cannot be rejected.

4. Data sources and variable construction

The main sources of the data are the Economic Survey of Pakistan(various issues) and Hand Book of Statistics (2010). The variables usedin the paper are consumer price index (cpi) for inflation,2 the centralgovernment deficit (fd), ratio of GDPmeasures through expenditure ap-proach for measuring the aggregate demand to GDP at factor cost formeasuring the aggregate supply (adas), the nominal exchange rate

Table 2Bounds tests for the existence of a long relationship.

Model F-statsimposing 1lag length

F-statsoptimal lag length

F-stats Lag length

lcpi = f (lfd, ltrade, lr, lcpit − 1) 5.3493 6.6030 2lcpi = f (lfd, ltrade, lr, lcpit − 1, er) 6.6307 5.5632 3lcpi = f (lfd, ltrade, lr, lcpit − 1, oil) 6.2661 6.8315 3lcpi = f (lfd, ltrade, lr, lcpit − 1, gsb) 7.4367 5.7912 2lcpi = f (lfd, ltrade, lr, lcpit − 1, psb) 5.9372 4.5700 2lcpi = f (lfd, ltrade, lr, lcpit − 1, adas) 6.4011 7.3093 2lcpi = f (lfd, ltrade, lr, lcpit − 1, imp) 4.1288 4.1288 1lcpi = f (lfd, ltrade, lr, lcpit − 1, wp) 5.8284 6.7892 3

Note: the calculated are compared with critical values: Pesaran et al. (2001).

123A. Jalil et al. / Economic Modelling 37 (2014) 120–126

(er), trade openness (trade), Government Sector Borrowing (gsb),Private Sector Borrowing (psb), lending rate (lr), import price index(imp), oil prices (oil) and wheat procurement prices (wp).

The growth rate of real GDP per capita is the annual change in thereal GDP per capita. The oil prices in the local currency are computedby the product of the average of crude price of petroleum in dollarsand the nominal exchange rate. Oil price inflation is measured by theannual change in the oil prices. Openness ismeasured by the ratio of an-nual imports plus exports to GDP. The demand pressure is measured bythe ratio of demand side GDP to supply side GDP. The supply side GDP isby GDP at factor cost and aggregate demand is the total demand in theeconomy by various economic agents. If ratio increases then it impliesthat there is demand pressure and the coefficient is expected with pos-itive sign.

Thanks to the availability of data,we use the breakup of assets side ofmoney supply instead of using the money supply as a whole. The gov-ernment sector borrowing (GB) includes the net foreign assets andprivate sector borrowing includes the borrowing of autonomous bodies.Theoretically, the government sector borrowing and private sector havethe inflationary effect therefore; we are expecting that both of thesevariables will enter positively in the regression line.

Pakistan is a net importer country of rawmaterial and energy relatedgoods. Therefore, increase in commodity prices of these goods makesour imports costlier and it puts an upward pressure on inflation. Con-sequently, it is important to include the import prices in inflationregression. We take import prices through the import price index. Thisindex should enter positively in our regression.

The role of expectations iswell established in the inflation determin-ing theory. It implies that the price rise generates the expectations forthe further price hike in the economy. Keeping the price hike in the fu-ture the people expect the higher salaries and nominal wages in the fu-ture for the compensation of inflation. The credit of the manufacturingsector diverts to real estate and stock markets. Furthermore, the profitseeking agents become more active in the higher inflation situation toearn the profit; therefore an untargeted bidding has a devastating im-pact on the prices. These expectations can be captured through the lagvalue of dependent variable, that is, CPI.

Wheat crises and then the huge rise in the wheat prices have beenwitnessed over the last five years in Pakistan. Being an overpopulateddeveloping countryWheat prices have a significant impact on inflation,specifically the food inflation, in Pakistan. As a robustness check we in-sert the procurement prices of wheat in the inflation regression and ex-pect a positive sign.

5. Empirical results

The starting point in estimating the time series data sets is to test thelevel of integration of the variables under consideration. However, ARDLworks irrespective of the fact whether the data series are I(0) or I(1) but

Table 1Unit root tests.

ADF k ADF K

cpi 0.7231 0 cpi −2.9939 1fd −0.2548 1 fd −3.5508 1lr −3.1433 1 lr NA NAtrade −0.8078 1 trade −9.3880 1er 0.9848 2 er −5.5095 1oil −1.7092 1 oil −6.4859 0gsb 0.3594 1 gsb −4.8835 0psb 0.4216 2 psb −2.4178 0adas 0.7994 1 adas −2.9061 1imp 0.1752 1 imp −4.8736 1wp 0.3381 0 wp −3.4895 0

Notes: ADF augmented Dickey–Fuller test. k is the degree of augmentation that isautomatically determined by following the procedure of Campbell and Perron (1991).

ARDL does not work in the case of I(2) data series. Therefore the testingprocedure of stationarity is still valid. For this we use the conventionalaugmented Dickey–Fuller test. The results are presented in Table 1.

The results are showing that all the variables are non-stationary atlevel except for lending rate. The lending rate is stationary at level.Therefore, it is a strong justification for ARDL as an estimation techniqueto test the existence of long run relationship among the variables.

After establishing the order of integration, we test the long run rela-tionship among the variables. Prior to testing for the existence ofcointegration, we chose the optimal lag length. However, Ang (2010)imposes the 1 lag length in the presence of short span of data. Here,we by imposing the 1 lag length found strong evidences for the longrun relationship when compared with the Pesaran et al (2001) criticalvalue. We also estimate the F-stats for cointegration test taking theoptimal lag length for each model. The results of F-stats and optimallag length are presented in Table 2. The new results provide a strongsupport to a previous result that there exists a long run relationshipamong the variables.

Next, we estimate Eq. (2) following the ARDL co-integration tech-nique for the long run estimates. We estimate 8 different models.

The results are showing that the coefficient offiscal deficit enters sig-nificantly positive in all regressions. However, the level of significanceand magnitude of the coefficient varies from case to case. Specifically,the coefficient 0.6466 implies that the inflation may be increased by0.65% by 1% increase of fiscal deficit. Our findings are in linewith earlierfindings. The positive relationship between fiscal deficit and inflation isin accordance with the preposition of fiscal theory of price level whichattributes inflation as fiscal phenomenon (Leeper, 1991 and Sims,1994). Our result that the fiscal deficit affecting inflation positively inPakistan has also been observed previously by Shabbir et al (1994)who finds positive and significant results.

The rate of interest measured by the lending rate also enters signifi-cantly positive in all cases. This result was predicted by the Fisher effectandwas tested for the Pakistan economy byHasan (1999). Being robustwe also test the base model by using the deposit rates as a measure ofinterest rate. That model also shows the positive relationship betweeninflation and deposit rates.3 It implies that the increase in inflation dueto increase in nominal interest rate means lower real interest ratewhich discourages savings in the economy. This will lead to increaseddemand of loanable funds and decreased supply and hence bid up thenominal interest rate which will further aggravate inflation. The Fishereffect predicts a one to one relationship between nominal interest rateand inflation but we have observed a weak relationship, except models4 andmodel 5,whichwas also observed for the Turkish economywherethere was a long run interdependence between interest rate and infla-tion (Kose et al., 2012).

3 The results of deposit rates are not shown in Table 3, however available on request.

124 A. Jalil et al. / Economic Modelling 37 (2014) 120–126

The trade openness is also observed to be significantly affectinginflation in the long run. This positive relationship is against the obser-vation of Lin and Chu (2013) and Catao and Terrones (2005) who em-pirically tested the suggestion of Romer (1993) that more tradeopenness leads to lower inflation. They argue that the trade revenuegenerated by a country via tariffs will increase and this will decreasethe other inflationary financing methods of deficit like seignioragethus reducing inflation. However, a small country cannot earn huge rev-enue from tariffs and maintain the increased imports at the same timerather the openness would mean lesser tariff revenue and hence morereliance on inflationary financingmodes. Our result of positive relation-ship between trade openness and inflation is in accordance with theconclusion of Zakariya (2010) that the two variables are positively relat-ed. This is because Pakistan is a net importer and the imports largelyconsist of oil and manufacturing goods which have observed increasedprices over the last several years. This has resulted in imported inflationissue andmore expansion of trade has actually caused inflation. Samimiet al. (2012) have empirically shown that trade is positively relatedwithinflation and more openness to trade in fact raises inflation.

Thenwe put the nominal exchange rate into the inflation regression.The exchange rate of Pakistan is continuously depreciating over the lastthree decades except in the early 2000s. The depreciation of thePakistani currency implies that the more rupees for a dollar furtherimply costly imports. Therefore, we are in line with a priori expectationthat the exchange rate will enter in regression significantly positive.

Then we take, oil prices, government sector borrowing, private sec-tor borrowing, ratio of real demand to real supply, import price index,wheat prices and lagged dependent variables in model 3, model 4,model 5,model 6, model 7, andmodel 8 respectively. All these variables

Table 3ARDL log run estimate of fiscal deficit and inflation.

Dependent variable is natural log of consumer price index

Regressors Model 1 Model 2 Model 3

Fiscal deficit 0.6466** 0.3049* 0.4087***(0.3108) (0.1800) (0.1635)

Interest rate 0.6990*** 0.5162*** 0.3386***(0.1417) (0.2358) (0.0752)

Trade openness 0.4524* 0.2129** 0.3850***(0.2450) (0.0943) (0.0951)

Exchange rate NA 0.5489* NANA (0.2867) NA

Oil price NA NA 0.4799*NA NA (0.2790)

Government sector borrowing NA NA NANA NA NA

Private sector borrowing NA NA NANA NA NA

Real demand to real supply NA NA NANA NA NA

Price index of import NA NA NANA NA NA

Wheat price NA NA NANA NA NA

Lagged CPI 0.6775*** 0.6318* 0.2406**(0.1762) (0.3365) (0.1035)

Constant 0.4854*** 0.1056* 0.4966**(0.1234) (0.0586) (0.1765)

ECMt − 1 −0.3724*** −0.5511** −0.4092***(0.1598) (0.2830) (0.1549)

Diagnostic testNormality 0.2437 0.3435 0.2698Functional form 0.3898 0.4099 0.1850Heteroscedisactity 0.4254 0.8831 0.3660Serial correlation 0.9108 0.6200 0.1137CUSUM Stable Stable StableCUSUMSQ Stable Stable Stable

Note: *, ** and *** represent 10, 5 and 1% level of significance, respectively.

enter significantly positive in the regression which implies that thereare many other factors as well which are contributing in the inflationin Pakistan besides the conventional variables like fiscal deficit, interestrate and trade deficit.

We also measure the short run effect besides the long run effectthrough the error correctionmechanism(ECM). The short run estimatesare also in linewith the theory however, we are not presenting the shortrun estimates keeping brevity in mind. The important outcome ofECM is the ecmt − 1term. The negative and significant estimate of theecmt − 1coefficient is showing that there exists a short run relationshipaswell. This means that whenever there is any shock that brings the re-lationship out of long run equilibrium the model will adjust back andthe ecmt − 1 term is the speed of adjustment from short run disequilib-rium toward the long run equilibrium. The negative sign implies thatdisequilibrium will converge toward long run equilibrium. The magni-tude−0.3724 in the first model implies that it will take approximately3 years before converging back to the equilibrium path.

All eight models pass through the conventional diagnostic tests. Thereported values in the lower panel of Table 3 are the p-values of chi-square which test the null hypothesis of normal distribution of error,correctly specified functional form, no serial correlation and homosce-dastic errors. The p-values are evident of well specification of themodels, acceptance of the null hypothesis of normality assumption ofthe residuals, no serial correlation and homoscedastic errors. Further-more, the plots of CUSUM and CUSUMSQ statistics are well within thecritical bounds, implying that all coefficients in the error-correctionmodel are stable. Therefore, the coefficients of regressions can be usedfor policy decision-making purposes, such that the impact of policychanges considering the explanatory variables of inflation equation

Model 4 Model 5 Model 6 Model 7 Model 8

0.4555** 0.8204** 0.1476** 0.5401*** 0.4454*(0.2163) (0.4577) (0.0710) (0.1744) (0.2396)0.9443*** 0.8839*** 0.3915* 0.2110*** 0.2972**(0.2864) (0.3038) (0.2184) (0.0837) (0.1475)0.5007** 0.3094* 0.4465*** 0.5864*** 0.8774***(0.2293) (0.1639) (0.1372) (0.1750) (0.2714)NA NA NA NA NANA NA NA NA NANA NA NA NA NANA NA NA NA NA0.9790** NA NA NA NA(0.2037) NA NA NA NANA 0.7208* NA NA NANA (0.3767) NA NA NANA NA 0.8025*** NA NANA NA (0.1465) NA NANA NA NA 0.1295* NANA NA NA (0.0663) NANA NA NA NA 0.8853***NA NA NA NA (0.2639)0.1929* 0.7647** 0.1356*** 0.6054*** 0.5695***(0.1082) (0.1727) (0.0236) (0.1942) (0.2390)0.9394** 0.5096* 0.7336*** 0.6236*** 0.4128**(0.4096) (0.2976) (0.2651) (0.1613) (0.1947)−0.318** −0.2018* −0.2125** −0.7682*** −0.575***(0.1785) (0.1090) (0.1074) (0.2865) (0.1409)

0.7135 0.1159 0.3609 0.3329 0.74200.6964 0.7160 0.8494 0.2969 0.53070.4015 0.7558 0.5224 0.5663 0.39340.2596 0.4620 0.3710 0.7996 0.5849Stable Stable Stable Stable StableStable Stable Stable Stable Stable

125A. Jalil et al. / Economic Modelling 37 (2014) 120–126

will not causemajor distortion in the level of inflation, since the param-eters in these equations seem to follow a stable pattern during the esti-mation period.

6. Conclusion

This article uncovers the possible determinants of inflation giving aparticular emphasis on fiscal deficit using ARDL estimation procedureover the period of 1972 to 2012. The long run estimates show that fiscaldeficits have a positive impact on inflation along with other variablestaking into account the various models. Furthermore, the study doesnot refute the inflationary impact of money supply as this is an es-tablished hypothesis but the conclusion is that the positive long run re-lationship can be attributed to fiscal dominance. Furthermore, the hikein wheat procurement, interest rate and oil prices also contribute fromthe supply side.

On the basis of our quantitative analysis, we argue that fiscal imbal-ances are leading to inflation and require immediate fiscal consolidation.Therefore, there is a need to set some threshold level for fiscal authoritiesand monetary policy should be less dependent. In particular, the growthin the flow of government sector credit and private sector credit willhave further pressure on the prices. Furthermore, a positive relationshipbetween trade openness and inflation shows that Pakistan has also facedthe imported inflation. The need is to develop the manufacturing indus-try so that imports reduce and exports increase and the country becomesmore export oriented. This will lead to more trade along with lowerprices as suggested by Romer (1993).

Acknowledgment

We thankHammadHussain, Falak Sher and two anonymous refereesfor the helpful comments.

References

Agha, A.I., Khan, M.S., 2006. An empirical analysis of fiscal imbalances and inflation inPakistan. SBP Res. Bull. 2, 342–366.

Aisen, A., Veiga, F.J., 2008. The political economy of seigniorage. J. Dev. Econ. 87, 29–50.Alesina, A., Drazen, A., 1991. Why are stabilizations delayed? Am. Econ. Rev. 81,

1170–1188.Ang, J.B., 2010. Does foreign aid promote growth? Exploring the role of financial liberali-

zation. Rev. Dev. Econ. 14, 197–212.Banerjee, A., Dolado, J., Galbraith, J.W., Hendry, D.F., 1993. Co-integration, Error-

correction, and the Econometric Analysis of Non-stationary Data. Oxford UniversityPress, Oxford.

Bokil, M., Schimmmelpfenning, A., 2005. Three attempts at inflation forecasting inPakistan. Pak. Dev. Rev. 45, 341–368.

Boschi, M., Girardi, A., 2007. Euro area inflation: long run determinants and short rundynamics. Appl. Financ. Econ. 17, 9–24.

Boujelbene, T., Boujelbene, Y., 2010. Long run determinants and short run dynamics ofinflation in Tunisia. Appl. Econ. Lett. 17, 1255–1263.

Bowdler, C., Nunziata, L., 2006. Trade openness and Inflation episodes in OECD. J. MoneyCredit Bank. 38, 553–563.

Callen, T., Chang, D., 1999. Modelling and Forecasting Inflation in India. IMF workingpaper, WP/99/119.

Calvo, G., Vegh, C., 1999. Inflation stabilization and BOP crises in developing countries. In:John, T., Woodford, M. (Eds.), Handbook of Macroeconomics. Vol. C. North-Holland,Amsterdam, pp. 1531–1614.

Catao, L.A.V., Terrones, M.E., 2005. Fiscal deficits and inflation. J. Monet. Econ. 52, 529–554.Chaudhary, M.A., Ahmad, N., 1995. Money supply, deficit, and inflation in Pakistan. Pak.

Dev. Rev. 34, 945–956.Chauvet, M., 2000. Real Time Leading Indicators of the Brazilian Inflate. BancoCenral Do

Brasil Working Paper Series 7.Coppin, A., 1993. Recent evidence on the determinants of inflation in a tourism-oriented

economy: Barbados. Soc. Econ. Stud. 42, 65–80.Cukierman, A., Edwards, S., Tabellini, G., 1992. Seigniorage and political instability. Am.

Econ. Rev. 82, 537–555.Darrat, A.F., 1997. Domestic and international sources of inflation in some Muslim

countries: An empirical inquiry. Middle East Bus. Econ. Rev. 9, 1–27.De Haan, J., Zelhorst, D., 1990. The impact of government deficits on money growth in

developing countries. J. Int. Money Financ. 9, 455–469.Deme,M., Fayissa, B., 1995. Inflation, money, interest rate, and causality, the case of Egypt,

Morocco and Tunisia. Appl. Econ. 27, 1219–1224.Domac, I., Yucel, E.M., 2005. What triggers inflation in emerging market economies? Rev.

World Econ. 141, 141–164.

Durevall, D., 1998. The dynamics of chronic inflation in brazil, 1968–85. J. Bus. Econ. Stat.16, 423–432.

El-Sakka, M.I.T., Ghali, K.H., 2005. The sources of inflation in Egypt: A multivariatecointegration analysis. Rev Middle East Econ. Finan. 3, 257–269.

Fisher, I., 1911. Recent Changes in Price Levels and their Causes. Bull. Am. Econ. Assoc.Fourth Ser. 2 (1), 36–45.

Fratianni, M., Spinelli, F., 2001. Fiscal dominance and money growth in Italy: The longrecord. Explor. Econ. Hist. 38, 252–272.

Friedman, M., 1956. The quantity theory of money: A restatement. In: Freidman, Milton(Ed.), Studies in the Quantity Theory of Money. University of Chicago Press, Chicago,pp. 3–21.

Grauwe, P.D., Polan, M., 2005. Is inflation always and everywhere a monetary phenome-non? Scand. J. Econ. 107, 239–259.

Hanif, M.N., 2012. A note on food inflation in Pakistan. MPRA Paper 45009.UniversityLibrary of Munich, Germany.

Hasan, H., 1999. Fisher effect in Pakistan. Pak. Dev. Rev. 38, 153–166.Hasan, A., Khan, A.H., Pasha, H.A., Rasheed, A., 1995. What explains the current high rates

of inflation in Pakistan. Pak. Dev. Rev. 34, 927–943.Kandil, M., 2005. Money, interest, and prices: Some international evidence. Int. Rev. Econ.

Financ. 14, 129–147.Kemal, M.A., 2006. Is inflation in Pakistan a monetary phenomenon? Pak. Dev. Rev. 45,

213–220.Khan, M.S., Schimmelpfennig, A., 2006. Inflation in Pakistan. Pak. Dev. Rev. 185–202.Komulainen, T., Pirttila, J., 2002. Fiscal explanations for inflation: Any evidence from

transition economies? Econ. Plan. 35, 293–316.Kose, N., Furkan, E., Sezgin, A., 2012. The interest rate–inflation relationship under an

inflation targeting regime: The case of Turkey. J. Asian Econ. 23, 476–485.Laurenceson, J., Chai, C.H.J., 2003. Financial Reform and Economic Development in China.

Edward Elgar, Cheltenham, UK.Leeper, E.M., 1991. Equilibria under “active” and “passive” monetary and fiscal. J. Monet.

Econ. 27, 129–147.Leigh, D., Rossi, M., 2002. Leading indicators of growth and inflation in Turkey. Interna-

tional Monetary Fund, Washington Working Paper No 02/231.Lin, H.Y., Chu, H.P., 2013. Are fiscal deficits inflationary? J. Int. Money Financ. 32,

214–233.Lui, O., Adedeji, O.S., 2000. Determinants of inflation in the Islamic Republic of Iran — A

macroeconomic analysis. International Monetary Fund, Washington, DC, WorkingPaper No. 00/127. 1–29.

Makochekanwa, A., 2011. Impact of budget deficit on inflation in Zimbabwe. Econ. Res.Guardian 1, 49–59.

Malik, W.S., Khawaja, M.I., 2006. Money, output, and inflation: evidence from Pakistan.Pak. Dev. Rev. 1277–1286.

McCallum, B.T., 1999. Issues in the design of monetary policy rules. In: Taylor, J.B.,Woodford, M. (Eds.), Handbook of Macroeconomics. vol. 1C. North Holland,Amsterdam, pp. 1483–1530.

McCallum, B.T., 2001. Indeterminacy, bubbles, and the fiscal theory of price level determi-nation. J. Monet. Econ. 47, 19–30.

McCallum, B.T., 2003. Is the fiscal theory of the price level learnable? Scott. J. Polit. Econ.50, 634–649.

Mehdi, S., Reza, M., 2011. Relationship between government budget deficits and inflationin the Iran's economy. Inf. Manag. Bus. Rev. 2 (5), 223–228.

Metin, K., 1998. The relationship between inflation and the budget deficit in Turkey. J. Bus.Econ. Stat. 16, 412–422.

Naqvi, S.N.H., Khan, A.H., Ahmed, A.M., Siddiqui, R., 1994. Inflation in Pakistan: CausesAnd Remedies. Pakistan Institute of Development Economics, Islamabad.

Nawaz, M., Iqbal, M.M., Ali, A., Zaman, K., 2012. Fiscal theory of price level: A panel dataanalysis for selected SAARC countries. Rom. J. Econ. Forecast. 3, 152–170.

Niepelt, D., 2004. The fiscal myth of the price level. Q. J. Econ. 119, 277–300.Oladipo, S.O., Akinbobola, T.O., 2011. Budget Deficit and Inflation in Nigeria: A Causal

Relationship. J. Emerg. Trends Econ. Manag. Sci. 2, 1–8.Pekarski, S., 2011. Budget deficits and inflation feedback. Struct. Chang. Econ. Dyn. 22, 1–11.Pelipas, I., 2006. Money demand and inflation in Belarus: evidence from cointegrated

VAR. Res. Int. Bus. Finan. 20, 200–214.Pesaran, M.H., Pesaran, B., 1997. Working with Microfit 4.0: Interactive Econometric

Analysis. Oxford University Press, Oxford.Pesaran, M.H., Shin, Y., 1999. An Autoregressive Distributed Lag Modelling Approach to

Cointegration Analysis. In: Strom, S. (Ed.), Econometrics and Economic Theory in20th Century: The Ragnar Frisch Centennial Symposium. Cambridge UniversityPress, Cambridge (Chapter 11).

Pesaran, M.H., Shin, Y., Smith, R.J., 2001. Bounds testing approaches to the analysis of levelrelationships. J. Appl. Econ. 16, 289–326.

Qayyum, A., 2006. Money, inflation, and growth in Pakistan. Pak. Dev. Rev. 45,203–212.

Romer, D., 1993. Openness and inflation, theory and evidence. Q. J. Econ. 8, 870–903.Rubio, O.B., Diaz-Roldan, C., Esteve, V., 2009. Deficit sustainability and inflation in

emu: an analysis from the fiscal theory of the price level. Eur. J. Polit. Econ. 25,525–539.

Sahan, F., Bektasoglu, Y., 2010. A panel co-integration analysis of budget deficit and infla-tion for Eu countries and Turkey. Empirical Studies In Social Sciences. Prepared forthe 6th International Student Conference (Izmir, Turkey).

Samimi, A.J., Ghaderi, S., Hosseinzadeh, R., Nademi, Y., 2012. Openness and inflation: newempirical panel data evidence. Econ. Lett. 117, 573–577.

Sargent, T.J., Wallace, N., 1981. Some unpleasant monetarist arithmetic. Quarterly Review,Federal Reserve Bank of Minneapolis, issue Fall.

Sargent, T.J., Wallace, N., 1988. Some unpleasant monetarist arithmetic. Fed Reserv. BankMinneap. Q. Rev. 5, 1–17.

126 A. Jalil et al. / Economic Modelling 37 (2014) 120–126

Shabbir, T., Ahmed, A., Ali, M.S., 1994. Are government budget deficits inflationary?Evidence from Pakistan. Pak. Dev. Rev. 33, 955–967.

Simone, F.N.D., 2000. Forecasting inflation in Chile using State-space and RegimeSwitching Model. IMF working paper No. 00/162.

Sims, C.A., 1994. A simple model for study of the determination of the pricelevel and the interaction of monetary and fiscal policy. Economic Theory 4,381–399.

Sun, T., 2004. Forecasting Thailand’s Core Inflation. IMF working paper no. 04/90.

Tekin-Koru, A., Erdal, O., 2003. Budget deficits, money growth and inflation: the Turkishevidence. Appl. Econ. 35, 591–596.

Tiwari, A.K., Tiwari, A.P., Pandey, B., 2012. Fiscal deficit and Inflation: What causes what?The case of India. J. Int. Bus. Econ. 13, 57–81.

Woodford, M., 2001. Fiscal requirements for price stability. J. Money Credit Bank. 33,669–728.

Zakariya, M., 2010. Openness and inflation: evidence from time series data. Doğuş Univ.Derg. 11, 313–322.