PROBING THE EVIDENCE OF TRIPLE DEFICIT HYPOTHESIS IN PAKISTAN

IJSSHE-International Journal of Social Sciences, Humanities and Education Volume 1, Number 1, 2017 ISSN 2521-0041 ------------------------------------------------------------------------------------------------------------------- PROBING THE EVIDENCE OF TRIPLE DEFICIT HYPOTHESIS IN PAKISTAN Wajid Ali Sustainable Development Policy Institute, Islamabad, Pakistan Azizullah Kakar International Islamic University, Islamabad, Pakistan ABSTRACT This research seeks to empirically test a third dimension to the twin deficit hypothesis and examines the co-movement of a third deficit known as the current and financial account balance. A conceptual framework based on income expenditure approach is developed for co-integration. The study uses annual data of Pakistan covering the period 1980 to 2014. The ARDL bound testing approach finds that the three deficits namely budget deficit, current account deficit and capital and financial account deficit are linearly correlated in the long-run. Further the study found that causality runs directly from current account to budget balance and financial balance, which is a strong evidence of triple deficit hypothesis. KEYWORDS Capital and Financial Account; Budget Deficit; Current Account Balance; Twin Deficits 1. INTRODUCTION The co-movement between budget deficit and current account deficit are known as twin deficit hypothesis and is discussed extensively in literature (Miller & Russek, 1989); (Cavallo, 2005, inter alia). This phenomenon was first observed in 1980s when U.S economy faced deterioration in both current account balance and budget deficit. Economists were of the view that the large deterioration in current account balance is because of record level budget deficit. Later, this mutual connection was given the name of twin deficit hypothesis. There are some traditional theories that explain the interconnection between budget deficit (BD) and current account deficit (CAD). The first basic transmission mechanism in which budget deficit causes current account deficit is explained by formal Keynesian cross approach. Per this approach, budget deficit is caused by a positive shock to government spending or decrease in taxes. This 2 Ali, Kakar ----------------------------------------------------------------------------------------------------------------------expansionary fiscal policy induces expansion in aggregate demand, then increases imports and hence current account balance deteriorates. The second relationship between government deficit and current account deficit is determined by Mundell-Fleming model. The basic assumptions in Mundell-Fleming model are perfect mobility of capital (the economy can borrow or lend as much as it wants in world’s financial markets) and small open economy. The budget deficit (caused either by increase in government expenditure or cut in taxes) increases domestic interest rate which brings on capital inflow and consequently domestic currency appreciation. The current account balance deteriorates because of appreciation of domestic currency and so does twin deficit running from budget deficit to current account deficit, (Fleming, 1962; Mundell, 1963; Kearney & Monadjemi, 1990). Finally, the twin deficit hypothesis is also explained by the Feldstein and Horioka puzzle. Feldstein and Horioka (1980) were of the view that if savings and investment are not highly correlated then budget deficit (BD) and current account deficit (CAD) can move together, because independence of savings and investment means capital mobility is high. However, this conventional view is challenged by the Ricardian equivalence hypothesis (Barro, 1974; 1989). Per this view, for a given expansionary fiscal policy the substitution of debt for taxes has no effect on aggregate demand nor on interest rate, which means that an increase in taxes cannot affect external deficit. However, if Recardian Equivalence does not hold, even then there is a scope of causal relationship between current account deficit and budget deficit. The current account balance and financial account balance moving side by side are graphed below in figure 1 over a sample period which notifies the third deficit. The motivation behind this study is the initial observation of the third deficit which lead us to search and examine the latent macroeconomic issue of the third deficit particularly in case of Pakistan. The persistent budget deficit (2.4 to 8.8 percent of GDP) along with current account deficit (0.7 to 7.2 percent of GDP) remains an important issue for policy makers in Pakistan. The fluctuations in these two indicators provide useful intuition regarding macroeconomic policy and its responsiveness to autonomous shocks. But it is also possible that this long-lasting budget and current account deficit may worsen the financial account. Keeping in view the severity and importance of the issue, several studies in Pakistan tested the causal relationship between current account deficit (CAD) and budget deficit (BD) and reached to different conclusions (Zaidi, 1995; Burney & Akhter, 1992; Burney & Yasmeen, 1989; Kazmi, 1992; Aqeel & Nishat, 2000; Mukhtar et al. 2007; Hakro, 2009; Siddiqui, 2009; Saeed & Khan, 2012). However, majority of the studies used linear Granger causality to test the validity of the twin deficit; a technique only capable of testing short run relationships. Figure1. Evidence of the third Deficit Int. j. soc. sci. humanit. educ. ISSN 2521-0041 Probing the Evidence of Triple Deficit Hypothesis in Pakistan 3 ----------------------------------------------------------------------------------------------------------------------Other studies, like Lau et al. (2010), emphasized on economies affected by Asian crisis and found evidence that twin deficit exists in Malaysia, Philippines and Thailand. Similarly, Lau & Tang (2009) uses cointegration and causality tests and confirms the twin deficit hypothesis for Cambodia. Nevertheless, both theoretical and empirical evidence are found in the literature regarding the relationship among budget balance (general government saving, BB), Current (trade) account balance (CA) and capital and financial account balance. This study aims to extend the traditional twin deficit hypothesis and introduces a third deficit known as financial account deficit which is defined as the sum of capital account and financial account. A deficit in current account means surplus in capital and financial account because the current account deficit is financed through sale of domestic asset to non-nationals and the other way around. Some studies (Wong & Carranza, 1999; Yan, 2005; Lau & Nelson, 2011) are conducted with the aim to find whether current account and financial account are mutually dependent are not. Although the findings of these studies and the prevailing studies on twin deficit support the interdependence of current and financial account. However, they did not find any evidence in favour of the third deficit. The prime objective of the proposed study is to explain and meet the gap by suggesting an investigative framework that challenges the tri-deficit hypothesis. 2. MATERIAL AND METHODS 2.1. Theoretical Background and Transmission Mechanism The tri-deficit hypothesis is explained properly by traditional Keynesian framework, predicting that the third deficit resulting from twin deficit can be explained by two possible channels. Per Keynesian framework, the loanable fund model and net capital outflow proposes that budget deficit has uncertain impacts on financial account. Firstly, it observes that an increase in budget deficit (negative saving) will raise the domestic interest rate which will in turn increase the expected yield of the domestic financial assets. Thus, the financial account will improve because of increase in capital inflows and decrease in outflows. The second channel through which twin deficit leads to financial account deficit can be explained by exchange rate market. The foreign (domestic) financial assets will become less-expensive (more-expensive) domestically (internationally) because of appreciation of the real exchange rate caused by twin deficit and hence deteriorate the financial account. The presence of the triple deficit hypothesis assumes that exchange rate channel is more powerful than interest rate channel. Provisionally, there exists one way causation running either from budget deficit to current account then financial account (BB -> CA -> FA, where “->” signifies “does Granger-cause”) or from budget deficit to financial account (BB -> FA), or both. The graphical representation of the two channels is shown below. Int. j. soc. sci. humanit. educ. ISSN 2521-0041 4 Ali, Kakar ----------------------------------------------------------------------------------------------------------------------Budget Deficit increases Expected Yield of Domestic Financial Asset will rise Financial Account will improve because of capital inflow Budget Deficit The Domestic Interest rate Will Will lead to deterioration of financial account Appreciate the exchange rate Domestic Financial Asset will become expensive in International market Figure 2. Mechanism of triple deficit hypothesis 2.2. Income-Expenditure Approach to Triple Deficit Hypothesis Using the income-expenditure approach in the context of general equilibrium, this subsection provides a theoretical groundwork of tri-deficits hypothesis. This is done by analysing the relationship between twin deficits and financial account (FA) deficit. The use of this approach is justified by (Tang & Fausten, 2012) stating that “the empirical finding of two-way causality reinforces the intuition that temporally based causality relationships may not reveal the true structural relationship which exists.” Tang & Lau, (2011) also use this approach to re-examine the twin deficit hypothesis in U.S. 2.2.1. Twin Deficits As we know that goods market is in equilibrium when planned expenditure (E) equals planned output (Y) per period, i.e. E = Y. Since planned expenditure (E) E = C+I+G+(X-M) and Y = C+ +T, the alternate representation of the income-expenditure approach is written as I+G+X = +T+M1. National saving denoted by is the sum of public and private saving resulting the trade balance as X-M = (1) Where C, I, G, X, M, T and represent domestic consumption of goods and services, domestic investment, domestic government expenditure, export of goods and services, import of goods and services, total tax revenue and private saving respectively. 1 Int. j. soc. sci. humanit. educ. ISSN 2521-0041 Probing the Evidence of Triple Deficit Hypothesis in Pakistan 5 ----------------------------------------------------------------------------------------------------------------------Which is planned net national saving. The trade balance i.e. X-M is identical to the sum of net private saving and budget surplus i.e. X-M = . Rewriting the above construction and denoting the trade balance and budget balance as CA = X-M and BB = T-G respectively show: CA = ( -I) + BB (2) Equation (2) tells us that budget deficit is counterbalanced by an increase in private saving and fall in domestic investment or exports. This adjustment process will plunge the trade balance as seen in twin deficit hypothesis. 2.2.2. The Mutual Dependence of Current, Capital and Financial Account The alternative measure of current account in the balance of payment account is the capital and financial account (FA), which per Coughlin et al. (2006) measures worldwide the stream of capital assets. Fausten (1989) provided theoretical basis for the mutual dependence between current and capital accounts. How financial and real sectors respond to systemic turmoil and their collaboration during the adjustment process is the impression by these mutual dependences of accounts. The seminal work of Fausten (1989) is extended by Tang & Fausten (2012) and found evidence in favor of interdependence of current and financial accounts. The subsequent balance of payment identity is written as BoP = CA +FA ≡ 0. Ultimately the financial sector deficit (-FA) resulting from floating exchange rates is financed totally by the excess exports (X > M) in the goods and service sector while the trade deficit (M > X) is funded either by net capital flows or foreign inflows. The national saving in open economy could be either invested at home or overseas to write the saving-investment relation as Sn = Id + I* where I* is foreign investment equivalent to I* ≡ CA = -FA. Tang & Fausten (2012) noted that to satisfy following equation, (Sn -I) = CA = -FA (3) it is necessary to acquire the foreign assets (FA < 0) equivalent to the domestic real assets and transfer to foreign investor. 2.2.3. Tri-Deficits With this background, we are finally going to construct an investigative framework of tri-deficit phenomena based on the Keynesian cross approach. This will be based on the central idea of twin deficit hypothesis and the mutual dependence of current and financial account. By substituting equation (3) into (2), the balancing position of current account (CA = -FA + BB) assumes that Sn ≡ Sp. The cointegration approach will assure the triple deficit hypothesis after re-organizing the above counterbalancing position and assigning a negative sign to respective balance given as under; -BB = -CA -FA (4) In other words, the co-movement of the three series BB, CA, and FA will reinforce the tri-deficit hypothesis. Speaking differently, the budget deficit in the long-run will be offset exactly by the combined deficit of current and financial accounts. It is also observed that the deficit in mutually dependent current and financial accounts is because imports are funded by real sector resources (instead of financial) as financial account and fiscal account are already in deficit. Int. j. soc. sci. humanit. educ. ISSN 2521-0041 6 Ali, Kakar ----------------------------------------------------------------------------------------------------------------------- 2.3. Data Using Pakistan’s annual data from 1980 to 2014, this section provides the evidence for triple deficit hypothesis in Pakistan. The series included in the study for analysis is budget balance (general government saving, BB) in million rupees and is obtained from State Bank’s Handbook of statistics on Pakistan’s economy. Current (trade) balance (CA) defined as the difference between exports and imports is collected from economic survey of Pakistan on various issues; and capital and financial account balance is defined as all transactions between an economy and the rest of the world in goods, services; and income including change in reserves (FA) is retrieved from World Economic Outlook Database (accessed on April 2015). 3. RESULTS AND DISCUSSION The results of ADF test presented in Table 1 inform us about the order of integration of the given series. Results reveal that BB is I (0) while CA and FA is I (1) which suggests that ARDL is an appropriate technique to be used for empirical analysis (Pesaran, et al., 2001). This approach to cointegration is suitable for analysis when the series contains both the I (0) or I (1) variables. In addition, this technique also gives efficient estimate in case of small sample (Pesaran & Pesaran, 1997). The error correction specification of ARDL, used in this study, is as follows: (I) (II) (III) If the three series are co-integrated in the long run, it will be the evidence in support of triple deficit hypothesis. In other words, in long run the budget deficit should be equal to the deficit of trade balance and financial sector balance together. The existence of triple deficit hypothesis can be tested statistically by F-statistic by putting the lagged levels parameter equal to zero. Rejecting the is the evidence that the underlying series are co-integrated null hypothesis of in the long run. In ARDL bound testing approach to co-integration, if the computed F-statistic value is greater than the upper bound of the critical value then it can be concluded that the series under consideration are co-integrated in the long run. The opposite will be the case if the computed F-statistic value is less than the lower bound of critical value. Estimates will be in the inconclusive zone if the computed F-statistic value lies in between the upper bound and the lower bound of critical values. For this study, the computed F-statistic values 8.78, 3.87 and 9.10 are greater than the upper bound critical values (-4.53, -3.95 and -3.8) for F (FA/BB, CA), F (BB/FA, CA) and F (CA/BB, FA) respectively (Table 1). These statistics suggest that all the three series (BB, FA, CA) are co-integrated in the long run and hence we have strong evidence in support of triple deficit hypothesis in Pakistan. Int. j. soc. sci. humanit. educ. ISSN 2521-0041 Probing the Evidence of Triple Deficit Hypothesis in Pakistan 7 ----------------------------------------------------------------------------------------------------------------------Table 1. Augmented Dicky Fuller Test. BB TB CB I(0) I(1) I(1) Level -5.23 -2.01 -2.49 Test equation with constant and trend Difference 0.008 0.57 -6.19 0.00 0.33 -5.02 0.001 Table 2. ARDL Bounds Testing Approach to Cointegration. F-Statistic F(CA/BB, FA) F(BB/FA, CA) F(CA/BB, FA) 8.78 3.87 9.10 Critical Values Bounds I(0) I(1) -3.96* -4.53* -3.41** -3.95** -3.13*** -3.84*** ARDL based on AIC ARDL (5,5,1) ARDL (5,1,5) ARDL (2,4,1) Notes: The bounds critical values are taken from Table CI(v) Case V: Unrestricted intercept and unrestricted trend (Pesaran et al, 2001, p. 478) with k =2. *, **, *** represent significant at 1, 5 and 10 percent respectively. The triple deficit hypothesis among BB, FA and CA can also be validated by identifying the direction of causality through Granger causality time series approach. The Toda & Yamamoto (1995) approach to causality too is compatible with the case when the order of integration of the series is not uniform. The Toda & Yamamoto (1995) non-causality test results are given in Table 3. The results show that causality is not running from BB to CA and from FA to CA; nevertheless, for the existence of triple deficit hypothesis the causality running from CA to BB and CA to FA is sufficient. Table 3. Tri-variate VAR (m+1) Causality Tests (Toda & Yamamoto, 1995) Null Hypothesis FA Does Not Granger-Cause BB BB Does Not Granger-Cause FA BB Does Not Granger-Cause CA CA Does Not Granger-Cause BB FA Does Not Granger-Cause CA CA Does Not Granger-Cause FA F-Statistic (P-Value) 6.35 (0.0955) 15.01 (0.0018) 2.19 (0.533) 32.47 (0.00) 1.29 (0.73) 6.52 (0.088) We take 1 to 3 lag intervals based on AIC information criteria. After estimating the ECM version of ARDL model given in equations (I) to (III), the stability of the model parameter is tested by cumulative sum of recursive residuals (CUSUM) and the CUSUM square (CUSUMSQ) basing upon the Pesaran & Pesaran (1997). The results are presented in Figure 1 and 2. It is evident from the graph that the plot of the CUSUM and CUSUMSQ statistics falls inside the critical bands of the 5% confidence interval and there is strong evidence that the model parameters are stable. Int. j. soc. sci. humanit. educ. ISSN 2521-0041 8 Ali, Kakar ----------------------------------------------------------------------------------------------------------------------20 15 10 5 0 -5 -10 -15 -20 84 86 88 90 92 94 96 98 CUS UM 00 02 04 06 08 10 12 14 08 10 12 14 5% S ignific anc e Figure 3. Results of CUSUM Test 1.4 1.2 1.0 0.8 0.6 0.4 0.2 0.0 -0.2 -0.4 84 86 88 90 92 94 96 98 CUSUM of S quares 00 02 04 06 5% Signific anc e Figure 4: Figure 3: Results of CUSUM square Test 2. CONCLUSION The aim of this exercise was to extend the well-known hypothesis known in the literature as twin deficit hypothesis to triple deficit hypothesis i.e. to include the financial account (FA) deficit. This new established phenomenon of triple deficit hypothesis for Pakistani data is validated by finding co-integrated relationship among current account balance (CA), budget balance (BB) and financial account balance (FA). The Toda & Yamamoto (1995) non-causality test results also support the triple deficit hypothesis. This work will add to the existing literature of twin deficit hypothesis and have some policy implication too. It is important to mention that there is persistent increase in Pakistani current account deficit since the early 1990s which is financed by borrowing. This study indorses generally that Pakistan should made every effort to recover the current account deficit on priority basis by finding ways and means to increase the revenues and increase the scope of direct taxation as well as reduce the un-necessary current expenditure as the current account is the main cause to fiscal and financial account deficits in the short-run. It is important to note that this initial but influential piece of research has some limitations. Firstly, this approach did not consider the financial sector in theoretical framework and focused only on real sector i.e. goods and services sector. Therefore, to establish a theoretical linkage among the Int. j. soc. sci. humanit. educ. ISSN 2521-0041 Probing the Evidence of Triple Deficit Hypothesis in Pakistan 9 ----------------------------------------------------------------------------------------------------------------------three series, an alternative approach based on portfolio balance is needed. For comparison purposes the study suggests to include other countries or use panel estimation for further findings. proceedings. REFERENCES 1. Aqeel, A., Nishat, M., & Qayyum, A. (2000). The Twin Deficits Phenomenon: Evidence from Pakistan [with Comments]. . The Pakistan Development Review, , 535550. 2. Barro, R. J. (1974). Are government bonds net wealth?. Journal of political economy,. (82(6)), 1095-1117. 3. Barro, R. J. (1989). The Ricardian approach to budget deficits. . National Bureau of Economic Research Cambridge, Mass., USA. 4. Bartolini, L., & Labiri, A. (2006). Twin deficits, twenty years later. Current Issues in Economics and Finance,. 12 (7), 1-7. 5. Burney, N. A., Yasmeen, A., & Niazi, M. K. . (1989). Government Budget Deficits and Interest Rates: An Empirical Analysis for Pakistan [with Comments]. . The Pakistan Development Review,, 28(4), 971-980. 6. Burney, N. A., and N. Akhter. (1992). Government budget deficits and exchange rate determination. Evidence from Pakistan, 31:4, 871–882. 7. Cavallo, M. (2005). Understanding the twin deficits: new approaches, new results. FRBSF Economic Letter (2005-16), 1-4. 8. Coughlin, C. C., Pakko, M. R., & Poole, W. (2006). How dangerous is Current account deficit? The Regional Economist. 9. Fausten, D. K. (1989-90). Current and capital account interdependence. . Journal of Post Keynesian Economics, 273-292 . 10. Feldstein, M. & C. Horioka. (1980). Domestic savings and international capital flows. The Economic Journa, 90, 314-329. 11. Fleming, J. M. (1962). Domestic Financial Policies under Fixed and under Floating Exchange Rates. . Staff Papers-International Monetary Fund, 369-380. 12. Hakro, A. (2009). Twin Deficits Causality Link-Evidence from Pakistan. International Research Journal of Finance and Economics. 13. Kazimi, A. A. (1992). Ricardian Equivalence: Some Macro-econometric Tests For Pakistan. The Pakistan Development Review, 31:4, 733–758. 14. Kearney, C and Monadjemi M. (1990). Fiscal Policy and Current Account Performance: International Evidence on the Twin Defiits". Journal of Macroeconomics, 12, 197-220. 15. Lau, E., Abu Mansor, S., & Puah, C.-H. (2010). Revival of the twin deficits in Asian crisis affected countries. Economic Issues, 29-53. Int. j. soc. sci. humanit. educ. ISSN 2521-0041 10 Ali, Kakar ----------------------------------------------------------------------------------------------------------------------16. Lau, E., & Nelson, F. (2011). Financial and current account interrelationship: an empirical test. Journal of Applied Economic Sciences, 34-42. 17. Miller, S. M., & Russek, F. S. (1989). Are the twin deficits really related? . Contemporary. 18. Mukhtar, T., Zakria, M. and Ahmed, M. (2007). An Empirical Investigation for The Twin Deficits Hypothesis In PAKISTAN. Journal of Economic Cooperation, 28(4), 63-80. 19. Mundell, R. A. (1963). Capital mobility and stabilization policy under fixed and flexible exchange rates. . Canadian Journal of Economics and Political Science, 475485. 20. Pesaran, H. M., & Pesaran, B. (1997). Working with Microfit 4.0 Interactive Econometric Analysis. New York: Oxford University Press Inc. 21. Saeed, S., & Khan, M. A. (2012). twin deficits hypothesis:the case of pakistan. 22. Tang, T. C., & Fausten, D. K. . (2012). Current and capital account interdependence: an empirical test. International Journal of Business and Society, 229-244. 23. Tang, T. C., & Lau, E. (2011). General equilibrium perspective on the twin deficits hypothesis for the U.S.A. Empirical Economics Letters, 245 – 251. 24. Toda, H. Y., & Yamamoto, T. (1995). Statistical inference in vector auto regressions with possibly integrated processes. . Journal of Econometrics, 225-250. 25. Wong, C. H., & Carranza, L. . (1999). Policy responses to external imbalances in emerging market economies: further empirical results. IMF Staff Papers, 225-237. 26. Yan, H.-D. (2005). Causal relationship between the current account and financial account. International Advances in Economic Research, 149-162. © 2017 by the authors. Submitted for possible open access publication under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/). Int. j. soc. sci. humanit. educ. ISSN 2521-0041