Information about Asia and the Pacific Asia y el Pacífico
Chapter

4 Tax Smoothing, Financial Repression, and Fiscal Deficits in India

Editor(s):
Tim Callen, Christopher Towe, and Patricia Reynolds
Published Date:
February 2001
Share
  • ShareShare
Information about Asia and the Pacific Asia y el Pacífico
Show Summary Details
Author(s)
Paul Cashin, Nilss Olekalns and Ratna Sahay1 

Introduction

Budget imbalances are pervasive in developing countries. Yet there are few studies asking whether this outcome is consistent with optimal fiscal policy. Five decades of time series data for India are examined to answer this question.

The optimality criteria used to examine Indian fiscal policy is based on the concept of tax smoothing. Tax smoothing recognizes that in the presence of an increasing marginal social cost of raising tax revenue, it is optimal if the planned tax rate is constant (smoothed) over time (Barro, 1979).2 A smooth tax rate implies that temporary shocks to government spending and output yield fiscal imbalances, and it provides a rationale for the issuance of public debt. In this sense, the tax-smoothing hypothesis is the fiscal analog of Campbell’s (1987) consumption-smoothing model.

This paper tests for the presence of tax-smoothing behavior in India using data from 1951/52-1996/97 and the vector autoregressive approach of Huang and Lin (1993) and Ghosh (1995). This approach generates a time series for the optimal budget surplus, assuming that the government tax smooths, which is compared to the actual surplus. If smoothing is to hold, any differences between the two series should be quite small. We are also able to control for nontax-smoothing causes of fiscal deficits, enabling a more accurate test of the tax-smoothing model.

The intertemporal tax-smoothing model successfully explains the behavior of Indian fiscal deficits. Our results also confirm previous findings that financial repression traditionally made a significant contribution to Indian net revenues. The financial repression-induced overborrowing of the 1970s and 1980s has yielded a stock of liabilities that deviates from the stock of liabilities generated from the series of optimal (tax-smoothing) fiscal deficits. As of 1996/97, the government’s actual stock of public liabilities was about 18 percent of GDP higher than it would have been under optimal tax-smoothing policies, implying that fiscal surpluses (or at least smaller deficits) will need to be run in the future to ensure intertemporal solvency.

The paper presents an overview of some features of Indian public finance. Key issues involved in testing for tax smoothing are then outlined, followed by a description of the econometric methodology and relevant data. The results from tests of tax smoothing and fiscal sustainability are then presented, and are followed by some concluding comments.

Issues in Indian Public Finance

Indian Fiscal Outcomes

Figure 4.1 (top panel) plots the fiscal position (gross fiscal deficit) of the Indian central government (CENGFD) between 1951/52 and 1996/97, where the gross fiscal deficit is the excess of aggregate disbursements (net of recovery of loans and advances) over receipts (revenue receipts, including external grants, plus nondebt capital receipts of government).3 The fiscal deficits of the center can be financed by borrowing externally or domestically, chiefly through the issuance of public debt (see below for further details). Figure 4.1 (top panel) reveals that fiscal deficits (as a percent of GDP) have been large and persistent for the Indian central government. They can be characterized as growing during the 1950s, 1960s, and 1980s and contracting during the 1970s and the first half of the 1990s.

Figure 4.1.Indian Fiscal Outcomes, Central Government

(In percent of GDP)

Sources: Government of India, Report on Currency and Finance (various issues); IMF, IFS (various issues); IMF staff estimates.

The balance of payments crisis of 1991 resulted in the near exhaustion of India’s foreign exchange reserves, largely caused by the withdrawal of foreign currency deposits by nonresident Indians, While the trigger for the crisis lay in domestic political difficulties and the Persian Gulf war, concern over the sustainability of Indian fiscal policy, due to rising debt and debt servicing, was a key factor underlying the crisis (see Chopra et al., 1995, for details). Prior to the crisis, India financed its fiscal deficits largely through financial repression, with high reserve deposit requirements and statutory liquidity ratios inducing commercial banks to hold below-market-yielding public debt. Following liberalization of the financial system beginning in 1991, the government increasingly had to borrow at close to market rates of interest. This shift to market borrowing in the context of high primary deficits resulted in a sharp increase in the government interest bill. For example, central government interest payments (CENINT) rose by almost ½ of 1 percent of GDP between 1990/91 and 1996/97, even though the center’s liabilities fell by over 6 percent of GDP over the same period. Figure 4.1 (bottom panel) illustrates that central government liabilities (CENLIAB) doubled as a percentage of GDP between the early 1950s and mid-1990s.

Indian Financial Repression

As in many developing countries, governments in India have found it difficult to satisfy their intertemporal budget constraint with conventional revenue and borrowings of the type discussed above. In addition to market borrowing as a means of deficit financing, governments have also used the implicit taxation of financial intermediation, using quasi-fiscal activities such as seigniorage and financial repression as sources of fiscal revenue and reduced interest costs, respectively.

Seigniorage—the purchasing power over real goods and services that comes about due to a central bank’s monopoly over the issuance of reserve money—is typically passed on to the government either through central bank profits or via no- or low-interest loans to the government. Seigniorage taxes—the change in reserve money as a share of GDP—were an important component of Indian taxation over the period 1960/61–1994/95, representing on average 1.5 percent of GDP per year. Similarly, Fischer (1982) calculated that annual average Indian seigniorage revenues amounted to about 1 percent of GNP between 1960–76, or about 10 percent of government revenue. Click (1998) also recently calculated that annual average Indian seigniorage revenues over the period 1971–90 were about 1.7 percent of GDP, or about 12 percent of government spending.

Financial repression in India traditionally involved domestic borrowing by the government at below-market interest rates, owing to banking regulations.4 The Indian cash reserve ratio (CRR) historically ranged between 4–7 percent of bank deposits, yet was steadily raised to 15 percent by the late 1980s to bolster demand for reserve money. Similarly, the statutory liquidity requirement (SLR) was raised from about 20 percent of deposits in the early 1960s to 38.5 percent in the early 1990s. Both requirements enabled the government to garner about half of all credit extended by the banking system between the early 1960s and early 1990s, at interest rates below those required to voluntarily acquire the debt. Consequently, banks invested in assets (consistent with CRR and SLR requirements) that barely covered their cost of funds (see Joshi and Little, 1996; IMF, 1996, 1997). At end-1999, the CRR and SLR (as shares of bank deposits) stood at 9 and 25 percent, respectively.

Financial repression was also facilitated by a network of publicly controlled financial institutions. Nominal ceilings on institutional interest rates were used to limit competition from the private sector for the pool of loanable funds. Thus financial intermediaries typically set loan rates on private domestic credit that differed from the exchange-rate-adjusted world interest rate. Indian governments also required their public financial institutions to undertake additional quasi-fiscal operations, involving activities such as the promotion of subsidized credit to priority areas of the private sector (such as agriculture and small-scale manufacturing), the setting of credit ceilings and floors, exchange rate guarantees, loan rate ceilings, and loan guarantees. The liberalization of India’s financial sector in the 1990s, however, reduced the impact of many of these quasi-fiscal activities. For example, exchange guarantees were transferred to the government from the central bank, reserve requirements on commercial banks were reduced, and there occurred the removal of many of the restrictions on the setting of commercial bank interest rates (see Joshi and Little, 1996; IMF, 1996, 1997).

Sustainability of Indian Fiscal Policies

India has a low level of public saving relative to other developing countries and experienced a steady decline in public saving over the past two decades (Mühleisen, 1997). Previous work examining the historical path of India’s fiscal imbalances found that continuation of the trend of growing fiscal deficits during the 1980s was not consistent with the eventual repayment of public sector debt, and that there was little scope for seigniorage revenues to fill the fiscal gap (see, for example, Buiter and Patel, 1992). In addition, these studies argued that the positive value of India’s primary fiscal deficit was inconsistent with a shrinking present discounted value of the debt stock. That is, if nominal interest rates exceed the GDP growth rate, primary surpluses (which have not been forthcoming) would be required to stabilize the debt-to-GDP ratio and ensure the sustainability of fiscal imbalances (see Reynolds, 2000, for a discussion of fiscal sustainability and solvency).

Testing the Tax-Smoothing Hypothesis

The previous section highlighted how Indian governments have used revenues from financial repression to compensate for a shortfall in revenues from more conventional sources. This and later sections consider whether the raising of these revenues, with respect to both their magnitude and timing, has been optimal. This analysis uses the tax-smoothing model as our optimality benchmark. It also examines whether the accumulation of public liabilities, which involves both the tax-smoothing and tax-tilting components of fiscal deficits, is on a sustainable path.

The tax-smoothing model assumes that, in the absence of a first-best system of lump-sum taxes, the government seeks to minimize the welfare losses arising from its choice of tax rate. These losses are assumed to be an increasing, convex, and time-invariant function of the average tax rate. The government’s ability to minimize the tax-induced distortions is conditioned by its adherence to the intertemporal budget constraint. This requires the present value of tax receipts to be sufficient to cover all current and future government spending together with the government’s initial debt. In order to meet the intertemporal budget constraint, taxes cannot remain invariant to changes in either current or expected future expenditure. Welfare losses will be minimized, however, if, in response to newly acquired information indicating a future change in government expenditure, the government smooths the implied tax change over time.

Following the approach of Barro (1979), Ghosh (1995), and Olekalns (1997), the optimal budget surplus at time t (surt*) is given by

where it is assumed that the effective interest rate faced by the government is R; the expectations operator is E; the information set available to the government at time t is It; ∆ is the first difference operator; and gt is (exogenously given) government outlays excluding interest payments, Gt, normalized by the level of output, Yt.5

Implications of the Tax-Smoothing Hypothesis

Equation (1) states that the optimally chosen budget surplus is a linear function of expected future changes to government expenditure. The implication of an expected decline in government expenditure is that the government will reduce its budget surplus (possibly running a budget deficit), so that the tax reduction can be smoothed over time. An increase in the budget surplus is a signal that the government is anticipating an increase in its expenditure and is seeking to smooth the tax increase. The government’s behavior is analogous to that of a consumer in consumption-smoothing models, who adjusts savings based on the expectation of future “rainy days” (see Campbell, 1987).

There are four testable implications of tax smoothing that can be derived from equation (1). These are

  • The optimal tax rate changes only if there is new information concerning government expenditure. Accordingly, under rational expectations, tax changes should not be forecastable and should follow a random walk.
  • The budget surplus should Granger-cause (help predict) changes in government spending. This will be true whenever the government has better information about the future path of its expenditure than is contained in past values of the expenditure series.
  • The smoothed budget surplus should be stationary. Assuming that gt is I(1), then ∆gt will be I(0); since under the null hypothesis the actual (tax-smoothed) budget surplus is the discounted sum of ∆gt (see equation (1)), then the smoothed budget surplus will also be I(0).
  • The optimal smoothed surplus derived from equation (1) should differ from the actual, smoothed surplus by at most a random sampling error.

Why Run Deficits? Separating Tax Smoothing and Tax Tilting

There are two broad considerations motivating any government to run a budget deficit: tax tilting and tax smoothing. The analysis, up to this point, has assumed that only tax smoothing motivates the government to run either a budget deficit or a budget surplus. However, other intertemporal incentives for running unbalanced budgets exist. Even if we assume that government spending as a share of GDP will remain constant into the future (in which case there would be no need for tax smoothing), if the government’s subjective discount rate (reflecting the preference for current taxation over future taxation), β, differs from the effective interest rate, R, then the optimal tax rate will be affected by the government’s desire to engage in tax tilting. As noted by Ghosh (1995), the relationship between β and R is given by γ=[(1-(R/β)R)/(1-R)], where γ, the tax-tilting parameter, accounts for the fact that the optimal tax rate incorporates incentives for the government to defer taxes or enlarge surpluses, depending on the relationship between β and R, That is, when β≠R (γ≠1), the government’s optimal tax profile will be “tilted.” Tax tilting results in a bias toward either budget deficits or budget surpluses, which are created in a manner consistent with intertemporal solvency. For example, if β1), the government’s incentive is to shift taxes into the future, run fiscal deficits, increase its current level of liabilities, and gradually raise taxes over time. Such a government has a relatively high discount rate. It would choose to have a low tax rate in the present period but would raise taxes over time to service its accumulating stock of debt. Conversely, if β>R, the government has an incentive to bring tax increases forward, run fiscal surpluses, build up its stock of assets, and gradually lower taxes over time.

Since tax tilting has implications for the budget surplus that are entirely distinct from tax smoothing, it is important to ensure that the optimal surplus derived from equation (1) is compared to only that component of the budget surplus that relates to tax smoothing and not to the actual budget surplus, which potentially includes both tax-smoothing and tax-tilting components.6 This requires that tax tilting be filtered from the actual budget surplus, with the tax-smoothing component of the surplus defined as

where dt is the stock of debt (liabilities) in period t, Dt, normalized by the level of output, Yt; and τt is the average rate of tax at time t. In equation (2), when R>β (and γ <1), the tax-smoothing surplus surtsm will be larger than the measured budget surplus, since the incentive is for the government to defer tax collections into the future (and so run a budget deficit in the present on tax-tilting grounds). This paper focuses on the tax-smoothing component of budget surpluses, because without an explicit model of intergenerational welfare it is not possible to decide whether deferring/bringing forward tax collections (that is, tax tilting) is desirable. However, as long as the government’s objective function involves the minimization of the distortionary costs of taxation (which are assumed to rise quadratically with τt), then there will be avoidable deadweight costs from a failure to tax smooth (Ghosh, 1995).

Econometric Methodology

The estimation and testing procedure is carried out in four steps. The first step is to obtain an estimate of γ-1, the tilting parameter, in order to construct the stationary (tax-smoothing) component of the fiscal balance by removing from the data the nonstationary (tax-tilting) component of the fiscal balance. Given that τt and [gt + (r-n)dt] are both I(1) variables, then this estimate of γ-1 can be obtained from equation (2), as the cointegrating parameter from a regression of [gt + (r-n)dt] on τt. This relationship is best estimated using the Phillips-Hansen (1990) fully modified (FM) method, which yields an asymptotically correct variance-covariance estimator in the presence of serial correlation and endogeneity. To confirm that the regression is indeed cointegrated, the Phillips-Ouliaris (1990) residual-based cointegration test is employed; the actual (tax-smoothing) component of the fiscal balance, surtsm, is defined by the residuals of the cointegrated regression of equation (2).

The second step is to calculate the optimal tax-smoothing component of the budget surplus. The derivation of the optimal budget surplus requires a measure of anticipated future changes to government expenditure. Following Campbell and Shiller (1987), an obvious way of deriving such a measure is to exploit the fact that, under the null hypothesis that tax smoothing is valid, the budget surplus contains all the known information about future changes to the government’s spending plans and should help predict future changes in government expenditure. Because the smoothed budget surplus (surtsm) responds to expected future changes in government spending, it is a relevant information variable to forecast future changes in government expenditure. In addition, we can exploit the information concerning future expenditure plans contained in current and lagged values of Δgt. This means that forecasts of future changes to government spending can be recovered from a bivariate vector autoregression (VAR) in ∆gt and surtsm.

As a result, we estimate a first-order unrestricted VAR of the form Wt = A Wt-1 + Єt, where Wt = (∆gt, sûrtsm), Єt, is a 2x1 vector of disturbance terms, and A is a 2x2 matrix of coefficients. With the estimate of A from the VAR and using the fact that Et[Wt+j] = AjWt, an estimate of the optimal tax-smoothing component of the budget surplus can be computed as

where I2 is the 2x2 identity matrix and Λ is a 1x2 matrix of coefficients.7 Equation (3) is valid as long as both the infinite sum in equation (1) converges, and the variables appearing in the W matrix of the VAR system are stationary. If gt is I(1), Δgt will be I(0). Since under the null hypothesis the actual (tax-smoothing) budget surplus (surtsm) is equal to surt*—which from equation (1) is a discounted sum of Δgt then surtsm will also be I(0). The validity of the tax-smoothing hypothesis can be tested by comparing the estimate of the optimal (tax-smoothing) budget surplus derived from equation (3) with the estimated actual (tax-smoothing) budget surplus derived from equation (2).

The third step is to conduct a series of hypothesis tests to evaluate the validity of several implications of the tax-smoothing model. As mentioned above, these are: the tax rate should follow a random walk; the smoothed budget surplus should be stationary; and the government’s budget surplus helps predict changes in government expenditure. The final test examines whether the VAR parameters in equation (3) conform to the nonlinear restriction

This restriction implies that movements of the actual (tax-smoothing) budget surplus reflect those of the optimal (tax-smoothing) fiscal surplus; failure of this restriction implies that the government is not optimally smoothing its taxation path. Examination of whether the optimal and actual (smoothed) fiscal surpluses are similar, which would be a finding supportive of the tax-smoothing hypothesis, can be done by inspection of a plot of the respective series or, more formally, by estimation of the equation

Optimal tax smoothing implies the joint parameter restriction λ1 = 0 and λ2 = 1, and nonrejection of these joint restrictions implies that movements in sûrt*sm fully reflect movements in sûrtsm.

The fourth and final step in this analysis concerns whether the path of public liabilities is sustainable. Sustainability focuses on whether fiscal policies could be continued indefinitely, in contrast to the tax-smoothing analysis above that focuses on the optimality of fiscal policies—that is, whether they should be continued. To address this issue, we develop a test based on a multi-period application of the single-period budget constraint and examine the time series properties of the stock of public liabilities. This is done to characterize the data-generating process and make inferences about whether the path of fiscal policies is consistent with solvency. By iterating the standard dynamic budget constraint forward we have

where the fiscal deficit is DEFt = GtτtYt. If the tax-smoothing model is valid, then we also have

where the optimal fiscal deficit is DEFt*=Gtτt*Yt, the optimal tax rate is τt*, and ρ=1/(1+r). Equation (7) states that the present discounted value of future fiscal deficits (or surpluses) must be matched by initial assets (or liabilities). Abstracting from tax-tilting causes of any change in the stock of public liabilities, since the stock of public liabilities consistent with the (tax-smoothing) model-generated path of fiscal deficits (Dt*) is sustainable by construction, the difference between the actual stock of public liabilities (Dt) and the stock consistent with the tax-smoothing model (DtDt*) must be stationary if Indian fiscal policy is to be sustainable. That is, Indian fiscal policy can be sustained without the need for reform if the series calculated as the difference in the two stocks of public liabilities (DtDt*) is stationary; if not, then the actual stock of public liabilities is not sustainable on unchanged fiscal settings, and a change in fiscal policy is required.

Data Sources and Definitions

The data are taken from official sources—definitions and descriptions of the various data manipulations are detailed in Appendix 4.1. The period covered ranges from 1951/52 (marking the beginning of India’s first five-year plan) to 1996/97. Expenditures and revenues are measured, respectively, by aggregate disbursements (current expenditure, capital outlays, and loans and advances), net of recovery of loans and advances of the central government (CENEXP), and the sum of revenue receipts (including external grants) plus nondebt capital receipts of the central government (CENREV). Accordingly, the fiscal deficit measure includes the center’s loans and grants to the states on the expenditure side and its receipt of interest and loan repayments from the states on the revenue side. This is done as the center may need to raise (lower) taxes as a result of this expenditure (revenue raising), and hence such fiscal actions will be affected by tax-smoothing considerations.8

Our measure of the budget surplus is constructed by substituting the above concepts of expenditure and revenue into the right-hand side of the government’s dynamic budget constraint written in terms of GDP, which is given by (1+n)(dtdt+1) = τt–gt + (n-r)dt. The debt stock is measured by the total liabilities of the central government (CENLIAB).9 In Indian public finance, the excess of expenditure (CENEXP) over revenue (CENREV) yields the government’s gross fiscal deficit (CENGFD).

A measure of the real interest rate and real growth rate is required to derive the optimal smoothed budget surplus. Two different nominal interest rates are tested. The first divides the central government’s interest payments (CENINT) by its liabilities (CENLIAB), and the second is a weighted arithmetic average of the interest rates at which money is accepted by selected commercial banks in Bombay (INT). The results proved to be insensitive to the choice of nominal interest rate; the reported results use the second of these two measures.10 Nominal gross domestic product at market prices (NGDP) is used to normalize the variables where appropriate, and real gross domestic product (RGDP) is used to calculate the real growth rate for the economy. Finally, the tax-smoothing component of the budget surplus is derived according to equation (2).

Empirical Results: Testing for Tax Smoothing and Fiscal Sustainability

The Phillips-Hansen (1990) fully modified OLS estimator yielded a value for γ^1 in equation (2) of 1.402, with an associated standard error of 0.040.11 The value of this estimate shows that tax tilting has been very important for the Indian government and has led deficits to be much larger than they otherwise would be. It also implies that the government has a preference for deficits falling over time. An important source of this incentive to tilt deficits toward the current period has been the extensive quasi-fiscal activities of India’s public financial institutions, chiefly the large-scale taxation of financial intermediation through seigniorage and financial repression. These quasi-fiscal activities resulted in India’s real rate of interest (r) being low (and often negative) for much of the sample period, yielding low values for the effective interest rate faced by government (R–1 ≡ (1+r)/(1+n)), indicating that the government has a high discount rate (β–1 is much greater than one.

The value of γ^1 for India far exceeds the value of this parameter in previous empirical work for developed countries of Australia (Olekalns, 1997, γ^1 = 0.96), Canada (Ghosh, 1995, 0.93), and the United States (Ghosh, 1995, 0.94).12 This result reflects the fact that tax tilting, carried out through seigniorage and financial repression, is a much more important source of net revenue for India than the other (all developed) countries that have been examined in the literature for evidence of tax-smoothing fiscal behavior. The value of γ^1 = 1.40 indicates that the component of the actual Indian fiscal deficit attributable to tax tilting is equivalent to forgoing 40 percent of taxation revenue in the near term, and subsequently raising taxes over time to clear the accumulated stock of liabilities. Some indication of the respective magnitudes of tax tilting can be gauged from Figure 4.2, which shows the actual deficit and the tax-smoothed deficit (with the tax-tilting component removed).

Figure 4.2.Central Government, Actual and Smoothed Budget Surpluses

(In percent of GDP)

Table 4.1 reports the results of three different tests of the unit root and stationarity hypotheses. The tests are used to evaluate the predictions made by the tax-smoothing model that the average tax rate follows a random walk, and the smoothed budget surplus is stationary. The table shows the results from the augmented Dickey-Fuller (ADF) and Phillips-Perron (1988) (PP) unit root tests, and the Kwiatkowski et al. (1992) (KPSS) test for stationarity. These results support the tax-smoothing hypothesis. The ADF and PP tests are unable to reject the unit root hypothesis for the average tax rate, and this is consistent with the KPSS test, which rejects stationarity. The respective tests also show that the first difference of government expenditure is clearly stationary, and, under the tax-smoothing hypothesis, the smoothed component of the budget surplus should also be stationary. This is confirmed by the respective tests.13

Table 4.1.Unit Root and Stationarity Tests
τtsurtsmgt∆gt
ADF Test-1.303-3.599*-2.341-7.128*
PP Test-1.317-3.639*-2.379-7.116*
KPSS Test1.401*0.1291.343*0.283
Notes: ADF and PP refer to the augmented Dickey-Fuller (1979) and Phillips-Perron (1988) unit root tests, and KPSS refers to the Kwiatkowski et al. (1992) test for stationarity. The tag length for the ADF test is determined using the lag deletion technique recommended by Campbell and Perron (1991), where lags are successively deleted until a significant lag is reached. The maximum lag was set at four. For the PP test, the lag length was set at three periods for all variables; the results did not change appreciably for other lag lengths. Both the ADF and PP test regressions include an intercept term. The results for the KPSS tests are for two lags (the results did not change appreciably for other lag lengths). A* indicates that the null hypothesis of a unit root (for the ADF and PP tests) or the null hypothesis of stationarity (for the KPSS test) can be rejected at (al least) the 5 percent significance level. The 1 percent, 5 percent, and 10 percent critical values are -3.58, -2.93, and -2.60 (for the ADF and PP tests), and 0.739, 0.463, and 0.347 (for the KPSS test), re-
Notes: ADF and PP refer to the augmented Dickey-Fuller (1979) and Phillips-Perron (1988) unit root tests, and KPSS refers to the Kwiatkowski et al. (1992) test for stationarity. The tag length for the ADF test is determined using the lag deletion technique recommended by Campbell and Perron (1991), where lags are successively deleted until a significant lag is reached. The maximum lag was set at four. For the PP test, the lag length was set at three periods for all variables; the results did not change appreciably for other lag lengths. Both the ADF and PP test regressions include an intercept term. The results for the KPSS tests are for two lags (the results did not change appreciably for other lag lengths). A* indicates that the null hypothesis of a unit root (for the ADF and PP tests) or the null hypothesis of stationarity (for the KPSS test) can be rejected at (al least) the 5 percent significance level. The 1 percent, 5 percent, and 10 percent critical values are -3.58, -2.93, and -2.60 (for the ADF and PP tests), and 0.739, 0.463, and 0.347 (for the KPSS test), re-

Table 4.2 shows the results from the Granger causality test. The hypothesis that the budget surplus Granger-causes (helps predict) changes in government expenditure is rejected by the data at the 5 percent level of significance. The hypothesis cannot be rejected, however, at the 10 percent level, and this provides some evidence that the central government’s budget surplus is informative about future changes to central government expenditure, which is consistent with tax smoothing.

Table 4.2.Granger Causality Test

Lagα1β1F
sûrtsm → ∆gt1-0.0160.2893.139*
(0.157)(0.163)
Notes: The Granger causality test is an F-test to determine if the (smoothed) budget surplus causes (helps predict) changes in government expenditure, that is, whether β1=0. The lag length, p, was chosen by minimizing the Schwari Bayesian Criterion; the maximum lag length tried was p=4. The figure in parentheses is the (heteroskedastic-consistent) standard error. A * denotes that the null hypothesis of no causation can be rejected at the 10 percent level of significance, indicating that the current budget surplus does have predictive power for future changes in government expenditure.
Notes: The Granger causality test is an F-test to determine if the (smoothed) budget surplus causes (helps predict) changes in government expenditure, that is, whether β1=0. The lag length, p, was chosen by minimizing the Schwari Bayesian Criterion; the maximum lag length tried was p=4. The figure in parentheses is the (heteroskedastic-consistent) standard error. A * denotes that the null hypothesis of no causation can be rejected at the 10 percent level of significance, indicating that the current budget surplus does have predictive power for future changes in government expenditure.

The actual (tax-smoothed) budget surplus derived from equation (2) and the optimal (tax-smoothed) budget surplus derived from equation (3) are graphed in Figure 4.3. There is a close correspondence between the actual and optimal smoothed surpluses. This is confirmed by a Wald test of the parameter restrictions implied by the tax-smoothing hypothesis, which examines whether there is a close association between movements in the actual (tax-smoothed) budget surplus and the optimal (tax-smoothed) budget surplus. The test shows that the parameter restrictions implied by tax smoothing on the VAR are not rejected by the data at the 1 percent level of significance, indicating that the differences between the actual (tax-smoothed) and optimal (tax-smoothed) surpluses observed in Figure 4.3 just represent random sampling error. In particular, the estimated coefficient on λ1 is not significantly different from zero, and the estimated coefficient on λ2 is not significantly different from one.14

Figure 4.3.Central Government, Actual and Optimal Smoothed Surpluses

(In percent of GDP)

The results, therefore, support the hypothesis that the central government of India has engaged in tax-smoothing behavior over the period analyzed. In other words, it responded to expected future changes in government spending by running budget imbalances rather than altering contemporaneous government revenue. While the observed behavior is consistent with tax smoothing, the smoothness of taxes was most likely due to the traditional inability of the government to satisfy its intertemporal budget constraint from conventional (tax and nontax) revenue sources, which resulted in changes in public borrowing as the preferred response to expected future shocks to government spending. This inability to garner sufficient revenue has stemmed largely from the narrowness of the tax base, widespread tax evasion and exemptions, weak tax administration, the poor economic performance of revenue-earning public enterprises, and the fact that a large part of economic activity is undertaken in the underground economy (see Joshi and Little, 1994, 1996).

The above results point to the broad consistency of India’s fiscal data with the predictions made by the tax-smoothing hypothesis. It is important to remember when interpreting these results, however, that they relate only to the smoothed component of India’s fiscal imbalances. The magnitude of the tax tilting that has occurred is sufficiently large that there can still be concerns about the sustainability of India’s overall fiscal imbalances, even if the stationary component of the budget surplus adheres to the tax-smoothing hypothesis. To gain some insight, the consistency of the actual stock of public liabilities with intertemporal solvency is investigated. The rationale is to see whether the stock of liabilities consistent with the optimal path of fiscal deficits generated by the tax-smoothing model, (Dt*), evolves in tandem with the actual stock of public liabilities, (Dt), as set out in equations (7) and (6), respectively. This test is conducted by examining whether (DtDt*), the implied excess accumulation of public liabilities, is stationary. The tax-smoothing model generates conditions under which the stock of public liabilities can be repaid, and fiscal deficits derived under the model are, by definition, sustainable. Accordingly, if the actual stock of public liabilities (which includes a tax-tilting component) is rising more rapidly than the stock of liabilities implied by the tax-smoothing model, then the current path of fiscal deficits under unchanged policies is unsustainable.

To formally test for the presence of nonstationarity in (DtDt*), we use the ADF test, and construct Dt* assuming that the 1953/54 actual stock of liabilities equals the stock of liabilities consistent with the optimal path of fiscal deficits generated by the tax-smoothing model. The result for the liabilities of the central government (CENLIAB) indicates that the ADF test statistic has a value of -2.305, which fails to reject the null hypothesis of nonstationarity in the difference between the actual and tax-smoothing-based stocks of liabilities at the 5 percent level of significance. Accordingly, the difference between the actual and tax-smoothing-based stocks of liabilities contains a unit root, implying that the two series deviate and have no tendency to follow one another. That is, under unchanged fiscal policies, India’s stock of public liabilities is not sustainable.15

This result can also be seen in Figure 4.4, which shows (after normalizing the debt stock by GDP) the actual (dt) and tax-smoothing-based (dt*) stocks of liabilities, as well as the implied excess accumulation of public liabilities (dt-dt*). Over the period 1953/54–1996/97, the excess accumulation of liabilities has been rising, indicating that public borrowing is in excess of what expected future fiscal surpluses can service. While the excess accumulation was relatively small until the early 1970s (the stock of actual liabilities was less than 5 percent of GDP greater than its optimum level), the difference between the two stocks of liabilities grew rapidly during the 1970s and 1980s, peaking in the late 1980s at about 25 percent of GDP. During the first half of the 1990s, the excess accumulation declined as a result of the central government’s program of fiscal consolidation, so that in 1996/97 the actual stock of public liabilities was about 18 percent of GDP higher than the level consistent with the optimal path of fiscal deficits generated by the tax-smoothing model.

Figure 4.4.Central Government, Actual and Implied Optimal Debt Stocks

(In percent of GDP)

The actual stock of public liabilities reflects both tax smoothing and tax-tilting considerations. Given that the Indian central government was found to tax smooth, then the bulk of its excessive public borrowing can be attributed to tax tilting, with the government levying low taxes in the present and (implicitly) higher taxes in the future so that intertemporal solvency can be satisfied. This requires that at some future point in time taxes will need to be raised and fiscal surpluses (or smaller fiscal deficits) will need to be run to service the government’s stock of liabilities.

Conclusion

This paper examined evidence for tax-smoothing behavior in India over the period 1951/52-1996/97. In response to a temporary increase in government spending, the tax-smoothing approach predicts that the tax burden of funding this expenditure will be spread over time, and so the government will run a fiscal deficit in the short-run. Conversely, a permanent increase in spending should be financed by raising contemporaneous taxes, resulting in no fiscal deficit. The intertemporal tax-smoothing model is successful in explaining the behavior of the fiscal deficits of the Indian government, and so the government does keep its tax rate relatively constant (smooth) in the presence of temporary shocks to expenditure. We argue that the traditional inability of the government to satisfy its intertemporal budget constraint from conventional (tax and nontax) revenue sources resulted in public borrowing being its preferred response to shocks to government spending—behavior consistent with the tax-smoothing hypothesis. Moreover, this same inability to garner sufficient receipts from conventional revenue sources results in tax-tilting behavior by the government, with quasi-fiscal activities such as seigniorage and financial repression being important sources of net revenue.

An analysis based on long-term trends indicates that during the period 1953/54-1996/97, under unchanged policies, India’s stock of liabilities was not on a sustainable path. In particular, the tax-tilting-induced over-borrowing of the 1970s and 1980s yielded a stock of liabilities that deviates significantly from the stock of liabilities generated from the series of optimal (tax-smoothing) fiscal deficits. By 1996/97, the actual stock of public liabilities was about 18 percent of GDP higher than it would have been under tax smoothing. This implies that fiscal surpluses (or at least smaller fiscal deficits) will need to be run in the future to ensure intertemporal solvency. This result emphasizes the importance of enhancing the process of fiscal consolidation that began in the early 1990s.

    BarroRobert J.1979“On the Determination of the Public Debt,”Journal of Political EconomyVol. 87 (October) pp. 94071.

    BarroRobert J.1995“Optimal Debt Management,”Working Paper No. 5327 (Cambridge, Massachusetts: National Bureau of Economic Research) pp. 136.

    • Search Google Scholar
    • Export Citation

    BuiterWillem H. and Urjit R.Patel1992“Debt, Deficits and Inflation: An Application to the Public Finances of India,”Journal of Public EconomicsVol. 47 (March) pp. 172205.

    • Search Google Scholar
    • Export Citation

    CampbellJohn1987“Does Saving Anticipate Declining Labor Income, An Alternative Test of the Permanent Income Hypothesis,”EconometricaVol. 55 (November) pp. 124973.

    • Search Google Scholar
    • Export Citation

    CampbellJohn and RobertShiller1987“Cointegration and Tests of Present Value Models,”Journal of Political EconomyVol. 95 (October) pp. 106288.

    • Search Google Scholar
    • Export Citation

    CampbellJohn Y. and PierrePerron1991“Pitfalls and Opportunities: What Macro-economists Should Know about Unit Roots,”Research Memorandum Econometric Research Program Princeton University No. 360 (April) pp. 167.

    • Search Google Scholar
    • Export Citation

    CashinPaulNilssOlekalns and RatnaSahay1998“Tax Smoothing in a Financially Repressed Economy: Evidence from India,”IMF Working Paper WP/98/122 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation

    CashinPaulNadeemUl Haque and NilssOlekalns1999“Spend Now, Pay Later? Tax Smoothing and Fiscal Sustainability in South Asia,”IMF Working Paper WP/99/63 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation

    Central Statistical Organization1996National Accounts StatisticsMinistry of Planning and Program ImplementationNew Delhi.

    ChopraAjaiCharlesCollinsRichardHemming and KarenParker1995“India: Economic Reform and Growth” IMF Occasional Paper 134 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation

    ClickReid W.1998“Seigniorage in a Cross-Section of Countries,”Journal of Money Credit and BankingVol. 30 (May) pp. 15471.

    • Search Google Scholar
    • Export Citation

    DickeyDavid A. and Wayne A.Fuller1979“Distribution of the Estimators for Autoregressive Time Series with a Unit Root,”Journal of the American Statistical Association No. 74 pp. 42731.

    • Search Google Scholar
    • Export Citation

    FischerStanley F.1982“Seigniorage and the Case for a National Money,”Journal of Political EconomyVol. 90 (April) pp. 295313.

    • Search Google Scholar
    • Export Citation

    GhoshAtish R.1995“Intertemporal Tax Smoothing and the Government Budget Surplus: Canada and the United States,”Journal of Money Credit and BankingVol. 27 (November) pp. 103145.

    • Search Google Scholar
    • Export Citation

    GiovanniniAlberto and Marthade Melo1993“Government Revenue from Financial Repression,”American Economic ReviewVol. 83 (September) pp. 95363.

    • Search Google Scholar
    • Export Citation

    HuangChao-Hsi and Kenneth S.Lin1993“Deficits, Government Expenditures, and Tax Smoothing in the United States: 1929–1988,”Journal of Monetary EconomicsVol. 31 (June) pp. 31739.

    • Search Google Scholar
    • Export Citation

    International Monetary Fund1996“India—Selected Issues,”IMF Staff Country Report 96/132 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation

    International Monetary Fund1997“India—Selected Issues,”IMF Staff Country Report 97/74 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation

    JoshiVijay and Ian M.D.Little1994India: Macroeconomics and Political Economy 1964–91 (Washington: World Bank).

    JoshiVijay and Ian M.D.Little1996India’s Economic Reforms 1991–2001 (Oxford: Clarendon Press).

    KwiatkowskiDenisPeterPhillipsPeterSchmidt and YoncheolShin1992“Testing the Null Hypothesis of Stationarity Against the Alternative of a Unit Root: How Sure Are We That Economic Time Series Have a Unit Root?”Journal of EconometricsVol. 54 (October–December) pp. 15978.

    • Search Google Scholar
    • Export Citation

    Ministry of Finance1966Budget for 1965–66 and various issues (Delhi, India: Government of India Press).

    MühleisenMartin1997“Improving India’s Saving Performance,”IMF Working Paper WP/97/4 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation

    OlekalnsNilss1997“Australian Evidence on Tax Smoothing and the Optimal Budget Surplus,”Economic RecordVol. 73 (September) pp. 24857.

    • Search Google Scholar
    • Export Citation

    OlekalnsNilss and PaulCashin2000“An Examination of the Sustainability of Indian Fiscal Policy,”Research Paper 748Department of Economics, University of Melbourne.

    • Search Google Scholar
    • Export Citation

    PhillipsPeter C.B. and Bruce E.Hansen1990“Statistical Inference in Instrumental Variables Regression with I(1) Processes,”Review of Economic StudiesVol. 57 (January) pp. 99125.

    • Search Google Scholar
    • Export Citation

    PhillipsPeter C.B. and SamuelOuliaris1990“Asymptotic Properties of Residual-Based Tests for Cointegration,”EconometricaVol. 56 (January) pp. 16593.

    • Search Google Scholar
    • Export Citation

    PhillipsPeter C.B. and PierrePerron1988“Testing for a Unit Root in Time Series Regression,”BiometrikaVol. 75 (June) pp. 33546.

    • Search Google Scholar
    • Export Citation

    Reserve Bank of India1997Report on Currency and Finance 1996–97 and earlier issues (Mumbai, India: Reserve Bank of India).

    ReynoldsPatricia2000“Fiscal Adjustment and Growth Prospects in India,” in India at the Crossroads; Sustaining Growth and Poverty Reduction Ten Years After CrisisM.Blejer and C.Toweeds. (Washington: International Monetary Fund) Chapter 5.

    • Search Google Scholar
    • Export Citation

    SrinivasanT.N.2000“India’s Fiscal Situation: Is a Crisis Ahead?,”mimeoYale University.

Appendix 4.1: Data Appendix

All data used have been derived from official sources, and are annual in frequency. It should be noted that they are for financial years ending March 31; for example, 1994/95 refers to the year ending March 31, 1995.

  • INT is the call money rate for scheduled commercial banks in Bombay for the period 1951/52-1956/7, then the money market rate (rate offered in the Bombay interbank market) from 1957/58 onwards, both taken from IMF, International Finance Statistics (IFS), line 60b.
  • NGDP is nominal GDP at market prices, in billions of rupees (Rs. crore), taken from IMF, IFS line 99b and Central Statistical Organization (1996).
  • RGDP is real GDP at market prices, in billions of 1990/91 rupees (Rs. crore), taken from IMF, IFS line 99b and Central Statistical Organization (1996).
  • GDPDEF is the GDP deflator (base 1990/91 = 100), derived from NGDP and RDGP.
  • CPI is the consumer price index for industrial workers for 50 centers of India, taken from IMF, IFS line 64 and Central Statistical Organization (1996).
  • CENREV is the sum of revenue receipts (including external grants) plus nondebt capital receipts of the Government of India, GOI, in billions of rupees (Rs. crore), taken from Budgetary Position of GOI, Revenue Receipts of GOI and Capital Receipts of GOI tables of the Reserve Bank of India’s Report on Currency and Finance, and IMF staff estimates.
  • CENEXP is aggregate disbursements (revenue expenditure, capital outlays, and loans and advances), net of recovery of loans and advances of the central government, in billions of rupees (Rs. crore), taken from Budgetary Position of GOI, Revenue Expenditure of GOI and Capital Disbursements of GOI tables of Report on Currency and Finance, and IMF staff estimates.
  • CENGFD is the gross fiscal deficit of the central government, and is calculated as the excess of CENEXP over CENREV. It is financed by external borrowing and domestic borrowing, where the latter comprises market borrowing (chiefly from publicly owned financial institutions), treasury bills, changes in cash balances with the Reserve Bank of India (RBI), small savings scheme, and state provident funds.
  • CENINT is interest payments made by the central government, in billions of rupees (Rs. crore), taken from Revenue Expenditure (nondevelopment expenditure) table of Report on Currency and Finance. This measure is for total interest payments, involving interest payments on internal debt, external debt, small savings and provident funds, reserve funds, and other obligations.
  • CENLIAB is total liabilities of the central government on March 31, in billions of rupees (Rs. crore), taken from Liabilities and Capital Investments and Loans Advanced by Central Government table of Report on Currency and Finance and Ministry of Finance, Budget Papers, various issues. Total liabilities includes public debt, small savings scheme, provident funds, and reserve funds and deposits.
1

The authors are grateful to Eduardo Borensztein, Timothy Callen, Ajai Chopra, Pietro Garibaldi, Nadeem Ul Haque, Mohsin Khan, Paul Masson, John McDermott, Xavier Sala-i-Martin, Patricia Reynolds, Parthasarathi Shome, M.R. Sivaraman, Peter Wickham, and especially Martin Mühleisen for their valuable comments and suggestions on earlier drafts, and Manzoor Gill for valuable research assistance. Any remaining errors are our responsibility.

2

As noted by Barro (1979, 1995), for a given amount of public expenditure, if taxes are lump sum and the other conditions for Ricardian equivalence are present, there are no real effects from shifts between taxes and the issuance of public debt as modes of financing fiscal imbalances. However, if taxes are distorting then the timing of taxes will matter, and it will be desirable to smooth tax rates over time, financing any temporary difference between public revenue and public expenditure by creating public debt.

3

The definition of the gross fiscal deficit follows that of the publications of the government of India, as it includes the proceeds from disinvestment in public sector enterprises in the center’s revenue.

4

Annual average revenue from financial repression in India has been estimated by Giovannini and de Melo (1993) at a sizable 2.86 percent of GDP and over 22 percent of government revenue (excluding revenue from financial repression) for the period 1980–85.

5

When the rate of real output growth, n, is positive, the effective interest rate faced by the government (R-1 ≡ (1+r)/(1+n)) will be smaller than the actual market interest rate, (1+r), where r is the assumed (constant) real rate of interest.

6

The tax-tilting (nonstationary) component of the actual fiscal surplus is removed to construct the tax-smoothing (stationary) component of the fiscal surplus. Beyond our desire to focus on tax smoothing, this is necessary to ensure the validity of standard statistical inference techniques, which will be used for hypothesis testing below.

7

The assumption of a constant real interest rate (r) assists in the derivation of equation (3), by allowing for the summation of a matrix geometric series. It also implies that the tax-tilting parameter is constant, which allows for stochastic detrending of the actual budget surplus data to focus on the stationary tax-smoothing component of the budget surplus.

8

See Cashin, Olekalns, and Sahay (1998) for additional details.

9

As with many developing countries, there are two main reasons why India’s stock of public debt may not be willingly held by market agents. First, part of India’s external debt was obtained on concessional terms from official bilateral and multilateral sources, and, second, part of India’s domestic debt is held by financial institutions at below-market rates of return to satisfy liquidity requirements.

10

When calculating the surplus, r and n are set equal to their respective average values. The consumer price index (CPI) is used to convert these nominal rates to real rates.

11

Phillips-Perron unit root tests (using an intercept and trend) reveal that both τt (-1.317) and [gt + (r-n)dt] (-2.037) are integrated of order one (the null hypothesis of a unit root cannot be rejected at the 5 percent level of significance), and so the possibility of cointegration exists.

12

In contrast, the estimate of γ-1 for India is close to that found for Pakistan, where γ^1 = 1.23 (Cashin, Haque, and Olekalns, 1999).

13

Using the critical values from the Phillips-Ouliaris (1990)Z(t) residual-based coinregration rest, we find that the null hypothesis of a unit root for surtsm can be rejected at the 5 percent significance level in favor of stationarity. Accordingly, we accept that equation (2) is a cointegrated regression.

14

The coefficients λ1^ and λ2^ are the estimated parameters from equation (5), and their values (with heteroskedastic standard errors in parentheses) are λ1^ =0.002 (0.125) and λ2^ =0.878 (0.455).

15

Recent work by Olekalns and Cashin (2000) and Reynolds (2000) also finds that, under unchanged fiscal policies, India’s fiscal imbalances are not consistent with intertemporal solvency. See also Srinivasan (2000) for a recent analysis of India’s fiscal situation.

    Other Resources Citing This Publication