12 Effects of a Budget Deficit on the Current Account Balance: The Case of the Philippines
- Mario Bléjer, and Ke-young Chu
- Published Date:
- June 1989
Ahsan H. Mansur
The purpose of this paper is to provide an empirical study of the economy of the Philippines during the period 1970-82, with an emphasis on external developments. Since the early 1970s, the Philippines has experienced external imbalances and very low levels of external reserves, caused by both adverse external developments and expansionary domestic demand. The expansionary role of the Government in recent years is believed to have contributed to the increase in domestic demand, and, as in many other countries, the authorities tried to reduce the external deficit through policies aimed at reducing the fiscal imbalance. Notwithstanding frequent attempts to restrain government expenditure and increase tax revenue, the national government’s overall budgetary position appears to have remained expansionary, and the shift in the composition of government expenditure in favor of development expenditure with a higher import content has tended to exert more pressure on the balance of payments. In the light of these considerations, this paper examines the effects of the overall budget deficit and the composition of its financing on the current account balance of the balance of payments.1
The relationship between fiscal deficits and the current account balance has been variously examined in the literature; for example, Milne (1976), Tahari (1978), and Kelly (1982) tested the relationship between the fiscal balance and current account balance using ad hoc single-equation specifications.2 These studies test the relationship in an ad hoc fashion that does not necessarily imply causality; such tests rely on the assumption that changes in the budget deficit are autonomous, when in reality they are generally endogenous.
Different ways of financing the budget deficit are also expected to affect the trade balance in different ways; the same degree of fiscal imbalance will have different effects on the trade balance depending on how the deficit is financed—from external sources, or the domestic banking and nonbanking system, or from a combination of sources. Simple testing, short of a complete specification of the structural model, cannot capture these interactions between budget balance, price level, domestic aggregate demand, and the current (or trade) account balance. However, before attempting to establish a quantitative relationship through a structural model, one should determine whether recent developments indicate that the budget deficit and its financing can account for a large part of the current account imbalance. In this analysis of the Philippine economy, recent fiscal, monetary, and balance of payments developments are examined to determine if fiscal policy was, indeed, expansionary and whether any causal relationship running from the overall budget deficit to the current account balance can be established. Based on positive indications from the preliminary observations, a small macroeconomic model is specified to simulate the impacts of alternative fiscal policies on the current account balance and derive their quantitative effects.
The structural model evaluates the effects of fiscal policy changes on the current account balance through changes in aggregate demand and the rate of domestic credit expansion originating in fiscal operations. In addition to specifying the transmission mechanism from expansionary fiscal policy to developments in the external trade account, this study focused on the structure of the real economy—that is, private sector absorption, income, or output determination. Fiscal expansion contributes to the increase in real output through increases in the components of aggregate demand, which affects domestic private sector absorption and, consequently, the demand for imports. The model used in this paper considers the real and monetary sectors simultaneously and determines the price level endogenously.
In the following analysis of the effects of the budget deficit on the current account, two major scenarios are considered, based on differences in the sources of financing of the budget deficit. In the first, the deficit is financed by external borrowing or by banking system credit, with credit to the private sector restrained, so that total credit from the banking system remains unchanged. In the second, the deficit is financed by an equivalent net increase in credit from the banking system. The effects of alternative financing methods on the domestic price level, output, and revenue are also analyzed. Special attention is paid to macroeconomic developments in recent years (especially during 1980-82) through the use of counter-factual simulations in which the budget deficit, measured as a percentage of gross national product (GNP), is held fixed at a desirable level.
The plan of the paper is as follows. Section II presents some preliminary observations on the nature of the fiscal policy pursued during the period under analysis and on the causal relationship between the overall government budget deficit and the current account balance. Section III outlines the structure of the model. The estimated model, along with some of its empirical characteristics, is discussed in Section IV. Section V reports on a variety of simulation exercises based on the model; concluding remarks are presented in Section VI.
II. Relationship Between Current Account Balance and Budget Deficit: Some Preliminary Observations
In this section, movements in the overall government budget deficit and the current account balance over the period 1970-82 are examined; the simultaneous movements between the current account balance and the capital account balance and changes in the net claims of the banking system on the Government are also examined (Chart 1).3 A casual comparison of the movements of the various variables seems to support the contention that the Philippines’ current account balance has been significantly influenced by movements in the overall budget deficit, particularly during 1980-82.
Chart 1.Philippines: Balance of Payments and Fiscal Developments, 1970–82
Source: International Monetary Fund, International Financial Statistics, various issues.
1 Changes in net claims in the central government owing to borrowing from the domestic banking system to finance a part of the overall budget deficit.
The observed relationship between budget deficits and the current account balance may be examined more closely by subjecting them to more formal tests of causality and by determining the nature (expansionary or contractionary) of budgetary policy.
Comparison of the actual and cyclically adjusted (or neutral) fiscal balance indicates that in 7 out of 12 observations, the actual deficit was greater than the corresponding cyclically neutral balance, indicating an expansionary fiscal stance (Table 1 and Appendix II). When viewed in the context of stabilization, fiscal policy is found to be out of tune with the cyclical developments in the economy. In 7 out of 12 observations (in which trend GNP was used as the estimate for potential GNP for the Philippines), it was found that fiscal policy was not countercyclical.4 In five instances when actual output was higher than the trend, the fiscal stance was expansionary or neutral after allowing for cyclical adjustments; in two instances, fiscal policy was contractionary when actual GNP was lower than the trend level.5
|Percentage of GNP|
In several years, the net impulse from changes in revenue was expansionary, indicating a slower growth in revenue relative to actual GNP. In 1972 and 1975, however, when major revenue measures were undertaken as part of stabilization measures to contain the widening budget deficit, the revenue impulse was contractionary. In the years following the large increases, revenue increased at a slower rate relative to GNP, imparting expansionary impulses to the economy.
Total expenditure was more restrained in the second half of the period (1976-82); impulses owing to capital expenditure tended to be highly expansionary but were partly offset by contractionary impulses exerted through restraint in recurrent expenditure. Based on the summary measures discussed above, the stance of fiscal policy at the national government level was broadly neutral during 1970-80 and became sharply expansionary during 1980-82; this deterioration occurred despite efforts to increase revenue and contain expenditure. The shift in the composition of expenditure impulse in favor of capital expenditure, which is believed to be more import-intensive, also had adverse implications for the current account balance.
Causality tests indicate that for both yearly and biannual data, the current account deficit of the balance of payments is correlated with the past value of the budget deficit (Appendix III). The line of causality runs from the fiscal to the balance of payments deficit, and the possibility of a reverse causality is ruled out for the Philippines on empirical grounds.
III. Specifications of the Model
Once the nature of the relationship between the government budget deficit and the current account balance has been determined in quantitative terms, a model can be designed that can explicitly establish the causal relationship between them. The macroeconomic model used as the basis for the simulation analysis consists of a number of behavioral equations explaining prices, output, and fiscal, monetary, and balance of payments developments. The basic model contains macroeconomic and monetary aggregates and addresses the central issue of determining the quantitative effects of the budget deficit and its alternative forms of financing on the current account balance and on the price level. Individual behavioral relationships are estimated in order to solve the whole system, and the effects of the policy changes are simulated from the solution of this system. On the fiscal side, developments in the broad components of government revenue and expenditure are considered; the effects of budgetary adjustments on the current account balance are specified so that their contribution to aggregate demand and their effects on the supply of, and the demand for, money can be observed.
1. Price Equation
Two alternative specifications for the price equation are considered. The first is based on the widely used approach to derive the price equation from the appropriately specified functional form for the demand for real money balances.6 Demand for real money balances (broadly defined) is generally formulated as a function of the level of real income, the real balance in the previous period, and the opportunity cost of holding money. For developing economies like the Philippines, where a broad range of financial assets do not exist and the rate of interest is institutionally influenced to remain low—effectively implying a negative real rate of return—the relevant opportunity cost is the expected rate of inflation.7 Using the actual inflation rate as a proxy for the expected rate of inflation, the demand for real balances is therefore specified in log-linear form as
where M is the stock of nominal money balances, P is the price level, Y is the level of real income, and π is the rate of inflation.
The price equation may be obtained by solving equation (1)
This specification, although ad hoc, can also be derived from an adaptive expectation scheme; the desired demand for real balances (M*/P) may be expressed in logarithmic form as
Assuming that actual demand for real balances is a convex combination of the real balances of previous periods and the desired real balances of the current period,
After substitution, this system reduces to
which is equivalent to equation (1).
Other partial adjustment schemes for the demand-for-money function and adaptive expectation schemes for the expected rate of inflation were also examined, and none of the other variations, either independently or jointly, were found to be statistically significant in explaining the demand for real balances for the Philippines.
The alternative specification for the price equation is based on the notion that changes in the price level are due to three sources of shocks: the domestic real sector, foreign price movements relative to domestic prices, and domestic monetary developments.
where ADD is aggregate domestic demand, YP is potential output, and Pf is the foreign price level. This alternate form was tested with Philippine data; except for
2. Government Sector
Government budgetary developments are modeled explicitly to analyze the effects of fiscal policy under alternative formulations. For the purposes of simulation, a distinction has been made among three separate sources of fiscal revenues. Revenue from import duties is assumed to be a function of nominal imports; total domestic tax is specified to be a function of nominal gross domestic product (GDP); and non-tax revenue and export duties are assumed to be exogeneously determined:
where TM is the tax on international trade excluding export duties, IMP is the volume of total imports, PRM is the unit value of imports, and TD is total domestic-based taxes.
This separate estimation of the major components of total taxes implicitly takes care of any potential aggregation problem that may exist when a single aggregate tax function is used. Export duties have been treated exogenously, because frequent changes in the tax rates and exemptions prevent a stable relationship between export receipts and export duties in the Philippines. Revenue from export duties has been mainly derived from export of coconut products; however, as the prices of coconut products fluctuated in the export markets, duties were adjusted to stabilize the domestic producer prices. Revenue from domestic taxes (TD) is the sum of all domestic-based taxes and is collected mainly in the form of personal and company income taxes and excise and sales taxes on goods and services. These categories of taxes are independent functions of GNP; thus, their sum may also be defined as an aggregate function of GNP without any problem of aggregation bias.9
3. Demand for Imports
The demand for imports is specified as a function of aggregate demand, domestic prices relative to foreign prices, and a dummy variable to capture the shift in import demand that has taken place beginning in 1978. The aggregate demand variable (ADt) used in the import demand function is defined as the sum of domestic private sector absorption, real government expenditure, and exports. In the conventional specifications, imports are generally specified as functions of income or output but are not sufficiently explicit to enable a distinction to be made between private and public sector contributions to absorption and import demand. The indirect effects of real income or output through private sector absorption will, however, be important and are captured in the proposed specification. Income or output determines private sector absorption and so plays the same role as it plays in the conventional specification of the import demand function. Moreover, a fiscal deficit generated through excessive public expenditure (financed by external borrowing or domestic credit creation), which is not directly related to GNP but is part of aggregate demand, also affects demand for imports.
where IMP is real imports, PRM is the unit value of imports, ABSP is private sector absorption, EXP is exports, G is government expenditure, and D1 is the dummy variable.
The aggregate demand variable (ADt) has a positive effect on import demand and the price term has a negative effect. An increase in the government budget deficit through increased nominal expenditure leads to higher import demand through its effect on aggregate demand; financing the deficit through domestic credit creation increases the domestic price level to partially offset the initial expansionary thrust and its indirect effects on the demand for imports.
In an effort to establish a link between import-intensive development expenditure and import demand, real imports were regressed on real output (or aggregate demand excluding government development expenditure), relative price, real development expenditure, and a dummy variable. The statistical fit of real development expenditure was highly significant, but contrary to its expected sign. One possible explanation of this perverse relationship may be that the Philippines was under successive adjustment programs during most of the period covered in this study. During periods of high imports relative to exports, the Government, to reduce domestic and external imbalances, constrained current and development expenditure within certain targets, leading to a conceptually perverse, but statistically significant relationship.
4. Real Income
Reflecting the short-term nature of the model, real income is determined by the movements in aggregate demand for domestically produced goods in combination with capacity constraints. The previous peak level of output is assumed to be a proxy for the capacity output; capital accumulation, growth of the labor force, and technical progress, which are more important in the long run, are not explicitly considered. For the Philippines, however, this measure of capacity output coincides with output in the previous period. To take into account the effects of aggregate demand on output, two specifications are considered. In one form, aggregate demand for domestically produced goods (ADD) affects real output directly, so that
This specification allows for variations in real output through demand management policies, although the coefficient of log (ADDt) is expected to be smaller than the coefficient of log (Yt − 1).
In the second specification, real output depends on the output of the previous period and on the excess of aggregate demand over the trend level of real output (Y*), where the trend for real output level is derived by regressing output on time
This specification states that aggregate domestic demand in excess of normal output will temporarily increase real income, and any slackness in aggregate domestic demand arising from reductions in private and government demand will reduce output. The degree of responsiveness of real output to aggregate demand depends on the parameter 66 and the size of domestic aggregate demand relative to the trend for real output.
5. Absorption Function
Private sector aggregate consumption and investment behavior are represented in terms of a single absorption function:
6. Fiscal Policy, Domestic Credit, and Money Supply
Government fiscal operations and the money supply are linked through the financing of budgetary deficits. In the absence of adequate nonbank sources, most of the domestic financing of the budget deficit tends to be in the form of borrowings from the domestic banking system. Changes in domestic credit (ADCt) can take place through changes in the banking system’s claims on the private sector (ACPt) and on the government sector
where DC is total domestic credit, NFF is net foreign financing of the budget, DNB is domestic nonbank financing, and CP is credit to the private sector.
In this formulation, an expansion of the fiscal deficit not financed from external and nonbank domestic sources results in an equivalent increase in domestic credit. In economies where the financial markets are not developed, the option of nonbank domestic financing may be fairly limited.
The broad definition of the supply of money—which includes currency, demand deposits, and time deposits—is identically equal to the sum of the level of net domestic credit extended by the banking system and the net stock of international reserves (NRt)—that is,
where OTM is the other monetary instruments not included in the definition of money in the present analysis.10 Net international reserves should optimally be treated as endogenous to be perfectly consistent with the rest of the model. However, given the high volatility of the Philippines’ net external reserves (arising from various economic and non-economic factors), a simple specification would be inappropriate. Since a detailed modeling for the determination of international reserves is beyond the scope of this paper, it is treated here as exogenous.
7. The Complete Model
The complete model has been modified slightly to capture some specific factors; dummies have been introduced in the tax functions to capture the shifts in revenue collection caused by substantive revenue measures introduced since 1975. The full structural model along with a glossary of the variables is shown in Table 2.
|1. Price Equation|
|2. Government Sector|
|3. Demand for Imports|
|4. Real Income|
|5. Private Sector Absorption|
|Definition of Variables|
|P||= price level|
|TM||= nominal import duty|
|TD||= nominal domestic taxes|
|IMP||= real private sector imports|
|Y||= real income|
|Y*||= normal or potential income|
|ABSP||= real private sector absorption|
|ADD||= aggregate real demand for domestically produced goods|
|R||= total revenue in nominal terms|
|DC||= domestic credit|
|M||= broadly defined supply of money|
|TB||= trade balance|
|G||= nominal government expenditure|
|PRM||= import price index|
|t||= time period|
|NONT||= non-tax revenue (nominal)|
|TEX||= export duty (nominal)|
|NFF||= net foreign financing (nominal)|
|DNB||= domestic nonbank financing (nominal)|
|CP||= credit to the private sector (nominal)|
|OTM||= sum of bonds and other items (net)|
To elucidate the operation of the model, consider the effects of an increase in the government budget deficit, resulting from increased government expenditure and partly financed by an increased supply of central bank credit to the Government. This increase will have two direct effects: aggregate real domestic demand will increase, and so will the nominal supply of money. These direct effects will tend to increase the price level, real income, and real level of imports through different channels and feedback effects of other endogenous variables. Imports will be affected indirectly through three channels: (i) higher aggregate demand owing to an increase in real government expenditure; (ii) increased private sector absorption resulting from a higher income effect; and (iii) import prices falling relative to the general price level. The process will gradually reverse as increased import duties and taxes with domestic bases and a decreased stock of real balances reduce the budget deficit and private sector absorption; if the system is mathematically stable, imports and the current account balance will stabilize.
IV. Estimates of the Structural Model
squares method. The estimated equations are shown in Table 3, and the ratio of the coefficients to their respective standard errors are noted in the parentheses below the coefficients.
The price equation derived from the original demand-for-money function is statistically significant and has all the coefficients with the expected signs. The estimated coefficients indicate that an increase in real output and the stock of real money balances in previous periods reduces the price level, whereas increases in the rate of inflation reduce demand for real balances and raise the price level.
The equation determining taxes on international trade is statistically significant and indicates that the elasticity of import duties with respect to the value of imports is less than unit to the value of imports is less than unity; this explains the slow growth of import duties in the Philippines as the value of imports increased over time. The statistical significance of the dummy variable (D1) represents a shift in the import duty since 1975, when a number of new measures were adopted to increase import duty collection. The estimated buoyancy of taxes with domestic bases also appears to be less than unity, indicating a slackness in tax collection relative to the growth of income. These estimates also support the previous finding that expansionary fiscal impulses were being exerted on the economy through slower growth in revenue (Section II).
The specification of real output as a function of the demand for domestically produced goods and of real output in the previous period yields a better statistical fit for the Philippine data set, compared with the alternative specification containing excess domestic demand over the capacity output as an explanatory variable. Equation (4a) in Table 2 provides a better fit as a single-equation specification and also as an equation of the complete model, since the predicted values in the baseline simulations with equation (4a) are much closer to the actual values, compared with the alternative specification. Thus, in the subsequent analysis, equation (4a) is considered as the specification for real output. By allowing aggregate domestic demand to influence real output, this model allows for government real expenditure to influence real output directly. However, the elasticity of output with respect to capacity output (or the output in the previous period) is more than twice that of real output with respect to the aggregate demand for domestically produced goods and services, and thus, in the long run real output is determined mostly by the rate of growth of potential output.
The import demand function has statistically significant estimators for the coefficients of log AD and the dummy variable; the coefficient of the relative price term is not significantly different from zero, although the coefficient has the correct sign. The equation is statistically significant and the overall statistical fit is good. The estimated elasticity of imports with respect to AD is less than unity, and real government expenditure, which is a component of AD, affects the demand for imports directly.
V. Simulation Exercises
This section describes how the model simulates the impact of a change in government expenditure on the trade balance, prices, and output; the simulations consist of changing government expenditure in nominal or real terms by a certain percentage, with appropriate specifications for its financing, and tracing the impact of these changes on the rest of the model. Before conclusions can be drawn regarding the quantitative significance of fiscal adjustments, however, it must be determined how equations (1a)-(6e) in Table 2 perform as a system, compared with the observed historical data. Statistical information about the simulated values of endogenous variables indicates that all the equations perform well, both as individual behavioral relations and as a part of the complete model (Table 4). For all the equations, the means of the actuals and predicted values of the endogenous variables are very close to each other, and the standard deviations of the predicted values of endogenous variables are generally not higher than the standard deviations of the corresponding actual values. The goodness of fit of the baseline simulations, which is evident from the statistics reported in Table 4, is also reflected in Chart 2.
|log M||0.05||0.06||0.06||0.06||0.015||0.025|Chart 2.Philippines: Actual and Predicted Values of Selected Variables in the Basic Simulation, 1970–82
Starting from the baseline simulations discussed above, two general types of simulations are performed: (1) the case of a hypothetical increase in nominal government expenditure by 10 percent financed by borrowing from the domestic banking system or from external sources; (2) a simulation of what would have happened to the trade balance, prices, and output during 1978-82, if the overall budget deficit as a percentage of GNP had been at the 1978 level. The differences arising from alternative sources of financing are also considered.
1. Increase in Government Expenditure
The first type of simulation examines the sensitivity of the trade balance, output, and price level in the Philippines to arbitrary changes in the budget deficit through changes in expenditure. Consider two scenarios: government expenditure is increased by 10 percent in real and nominal terms. In the case of a real increase, the price level is assumed to be constant; in the case of a nominal increase, domestic credit will change, depending on the financing of the budget deficit. When nominal government expenditure is increased, the ultimate effect depends on the final real increase based on the endogenously determined price level. The effect of an expansionary fiscal policy on the price level depends on the financing of the increased deficit; it is most inflationary if the deficit is financed by borrowing from the domestic banking system, as the consequent increase in money supply increases the price level. The partially offsetting effect coming from the output response is also minimal in the money-financed case, because the final real increase in government expenditure is less than the initial increase. If the increased government expenditure is fully financed by external borrowing, domestic credit and the money supply will not change, and increases in output, other things being equal, will depress the domestic price level somewhat. The effects of an externally financed government budget deficit are qualitatively and quantitatively similar to the simulations where the price level is held fixed and real government expenditure is increased by 10 percent. In general, any increase in the budget deficit owing to increased expenditure adversely affects the trade balance, irrespective of how the deficit is financed; a certain nominal increase in government expenditure is likely to cause a deterioration in the trade balance in real terms when the deficit is externally financed or the price level is held constant (Tables 5 and 6 and Chart 3).
|Case A1||Case B2||Case C3|
|Trade deficit||Billion constant pesos|
|Price index||1970 = 100|
|Scenario II||100.0||114.7||108.7||129.7||174.7||172.2||198.2||231.5||264.5||314.3||366.3||370.7||430.4|Chart 3.Philippines: Impact of a Permanent 10 Percent Increase in Government Expenditure Under Different Forms of Financing, 1970–821
1 Original simulation refers to the basic run. Case A represents a 10 percent increase in government expenditure financed by external borrowing. The scenario in which the deficit is financed by a net increase in borrowing from the banking system is shown as Case B.
2. Effects of Keeping the Budget Deficit at 1978 Level
The budget deficit in the Philippines remained at about 1.0 percent of GNP during 1970-80, but in 1981 and 1982 it increased sharply to 4.0 percent and 4.3 percent, respectively. In light of the deterioration in the trade balance that arises when government expenditure increases relative to revenue, as pointed to in the simulations in Section V.1, it may be argued that a smaller deterioration in the trade balance could have been attained if the overall budget deficit had been held at the 1978 level as a percentage of GNP.
To examine the quantitative effects of any such adjustment in the budget deficit, government expenditure during 1979-82 is adjusted so that the measured deficit as a percent of GNP remains at the level of 1978 (in terms of actual data).11 Qualitatively, the simulated impact of such a reduction in government expenditure is opposite to a hypothetical increase in government expenditure, as discussed earlier in this section. The hypothetical reduction would have led to a consequent reduction in the trade account deficit by about 10 percent during 1981-82, if prices remained unchanged or the financial surplus of the government were used to reduce external financing without any net effect on domestic credit; in this situation, the trade deficit in 1982 would have been 13.8 percent higher than that of 1978, compared with 25.4 percent without the adjustment (Table 7 and Chart 4). In a simulation where the reduced budget deficit in nominal terms implies a corresponding reduction in total domestic credit, the deflationary effect on the price level would have caused a smaller real reduction in government expenditure, and the consequent reduction in the trade deficit would also have been smaller. The hypothetical reduction in government expenditure would have improved the trade balance and exerted a significant deflationary effect on the price level, depending on the composition of the financing of the adjusted deficit. In any case, the simulations indicate a significant quantitative relationship between a change in real government expenditures and the balance of trade in the Philippines.
|Billion constant 1980 pesos|
|1980 = 100|
|Scenario II3||72.0||85.6||100.0||102.2||119.6|Chart 4.Philippines: Effects on Trade Balance and Price Level Under Alternative Scenarios, 1970–821
Source: Fund staff estimates.
1 The original case represents the baseline simulation. In the other two scenarios (Case A and Case B), government expenditure is adjusted during 1979-82 so that the budget deficit measured as a percentage of GNP remains constant at the 1978 level.
2 The reduced budget deficit is offset by a net reduction in external borrowing.
3 The reduced budget deficit is offset by an equivalent reduction in net borrowing from the banking system.
It should be noted that the sharp deterioration in the trade balance of the Philippines during 1980-82 was mainly attributable to the reduced value of exports caused by external developments. Merchandise exports remained stagnant in 1981 and declined by 12.3 percent in 1982 in U.S. dollar terms, with most of the decline attributed to lower export prices. In this paper, recent developments in exports have not been analyzed in any detail; they have been treated as being exogenously determined to avoid possible distortions arising from a simplistic or inadequate examination of the complex issues related to the determination of exports from the Philippines. The deterioration in the trade balance may be attributable to a number of external and domestic factors not covered in this study, but it can be argued that, apart from the adverse developments in exports, expansionary fiscal policy also contributed significantly to the deterioration in the trade and current accounts of the Philippines.
3. Relationship Between Simulated Real Budget Deficit and the Trade Balance
In the simulations considered above, expenditure was adjusted to certain levels and its impact on the trade balance, output, and prices was examined. In the basic model, tax revenue is endogenous to the system, depending on the nominal value of output and imports, so the overall budget balance is also endogenously determined. However, if the resultant trade balances are examined under alternative hypothetical situations, the simulations with a higher budget deficit in real terms, under all forms of financing, yield a higher deficit on the trade account in real terms (Table 8). When interpreting the results reported in Table 8, one should note that an increase in the real budget deficit over time is not necessarily reflected as a corresponding increase in the external trade deficit, because other developments not explicitly considered in the model may offset the effect of expansionary fiscal policy. If the data for any single year are observed and the impacts on the trade balance of variations in the budget deficit under different simulations are compared, a higher budget deficit (in real terms) corresponds to a higher deficit in the trade account; this is the only way the simulated budget deficit is allowed to change between the scenarios; other factors remain the same for the same year. It may be noted, however, that the effect of an expansionary fiscal policy on the trade balance is generally not as high as may be expected under the fiscal approach to the balance of payments; an increase in the budget deficit results in a less-than-equal deterioration in the trade account balance for the Philippines.
Through empirical observations, testing, and simulations, this study has examined the responsiveness of deficits in the Philippines’ trade and current accounts to changes in the central government budget deficit, with special emphasis on developments during 1980-82. Preliminary observations of the movements in the current account deficit and the budget deficit indicate a strong direct relationship between the two, which is also supported by causality tests. Fiscal policy was not necessarily counter-cyclical during 1970-80; during 1981-82, when fiscal policy assumed a countercyclical role, the fiscal stance was highly expansionary even after adjusting for cyclical variations. Based on these preliminary observations, a simple macroeconomic model was specified for the Philippines, which focused on the interactions of the government budget deficit and its associated financing with domestic absorption and imports.
Simulations based on the specified structural model indicate that an increase in the budget deficit worsens the trade balance significantly. A permanent increase in nominal government expenditure financed by borrowing from external sources causes the most deterioration in the trade balance; an equivalent increase in expenditure financed through borrowing from the domestic banking system is reflected in a sharply higher domestic price level and some deterioration in the trade deficit. The higher the inflation, the lower will be the real increase in government expenditure (given a fixed nominal increase in expenditure); the deterioration in the trade balance will be correspondingly lower in real terms. Simulations also indicate that significant reductions in the trade account deficit could be made if the overall budget deficit as a percentage of GNP were held at the level of 1978 by reducing government expenditure. If the financial savings were used to reduce net external borrowing, the trade deficit could be significantly reduced; if the savings were used to reduce the borrowing from the domestic banking system, the inflation rate could be significantly lowered and some modest improvement in the trade balance could also be achieved. Notwithstanding the strong relationship between the fiscal deficit and the trade balance, it should nevertheless be emphasized that the deterioration in the Philippines’ trade and current account balances during 1981-82 was mainly attributable to adverse external developments, which reduced export receipts. The expansionary fiscal stance during 1981-82 further worsened the deteriorating balance of payments position. Fiscal adjustments in the form of a reduced overall deficit alone were probably not sufficient to reverse the developments in the current account deficit during 1981-82, but in any event, fiscal restraint was necessary to achieve a sustainable balance of payments position in the medium term.
In this appendix, budgetary, monetary, and balance of payments developments during 1970-82 are discussed in detail. This period was characterized by moderately high and stable real growth (5.8 percent a year), with successive subperiods of adverse external economic developments; to attain financial stability with a sustainable balance of payments position, the Philippine authorities undertook a number of short-term stabilization programs, which were supported by a series of stand-by arrangements with the Fund. The measures included limits on the banking system’s credit to the public sector. Subceilings were introduced to limit government deficits to stated targets without recourse to unduly large external borrowing. Various revenue measures were adopted to improve the overall resource position of the public sector and to reduce recourse to domestic banks to finance the budget deficit. The overall budget deficit remained small during the 1970s, owing to these measures, but deteriorated rapidly during 1980-82.
1. Fiscal Developments
During the period 1970-80, the fiscal deficit was stable, averaging about 9.5 percent of expenditure, or 1.0 percent of GNP (Table 9). Revenue grew at a compound rate of about 25.6 percent, averaging 12.4 percent of GDP during the same period. Taxes with domestic bases increased at a relatively slower rate (22.5 percent) than taxes on international trade (32.5 percent). Thus, the share of taxes with domestic bases in total tax revenue decreased from about 78 percent in 1970 to about 62 percent in 1980; this decrease occurred despite a number of measures (e.g., increases in the sales tax, domestic excise duty, and income tax, and the introduction of a number of new tax measures) undertaken by the authorities to enhance the responsiveness of the tax system to domestic economic activity and to reduce the increasing dependence on taxes on international trade.
|Total tax revenue||3.2||4.5||6.2||8.5||10.2||14.3||15.1||16.5||20.6||26.0||30.6||31.4||33.8|
|Income and profits||(0.9)||(1.2)||(1.0)||(2.4)||(2.8)||(3.2)||(3.7)||(4.5)||(5.1)||(6.3)||(8.3)||(7.8)||(8.3)|
|Goods and services||(1.4)||(1.7)||(1.7)||(1.7)||(3.0)||(3.9)||(6.2)||(6.5)||(8.7)||(11.8)||(1.2)||(11.6)||(12.2)|
|Total revenue and grants||3.6||4.9||7.0||9.5||12.0||16.8||18.3||20.0||24.0||29.3||34.4||35.9||38.2|
|Percentage of GNP|
Expenditure (including equity contributions and net lending) increased at a compound annual rate of 23.5 percent, averaging 13.5 percent of GNP. However, current expenditure increased at a slower rate (20.5 percent) than capital expenditure (34.5 percent), and the ratio of capital expenditure to total expenditure increased from 15.2 percent in 1970 to 35.7 percent in 1980. This change in composition indicated increased development efforts in the form of infrastructure and other investment projects and efforts to contain current expenditure growth through expenditure controls on maintenance and subsidy outlays.
The budgetary position deteriorated rapidly during 1981-82, reflecting both the operation of built-in stabilizers on the revenue side and expansionary shifts in expenditure. Growth of tax revenue averaged only 5.1 percent, and expenditure grew by about 15.8 percent. Taxes on income and profits grew slowly, owing to a squeeze in the profits of major firms engaged in the export of primary products; import duties declined as a result of a sharp fall in the growth of imports. Expenditure policies were partly influenced by the sluggish economy; although strict economy measures were imposed on current expenditure, equity contributions to the financially depressed public enterprises and accelerated implementation of the infrastructure program increased capital expenditure sharply.
Reflecting developments in the overall budget balance, bank credit to the Government increased rapidly. Net claims on the Government, which amounted to only 6.4 percent of total domestic credit at the end of 1980, increased to 14.6 percent by the end of 1982 (Table 10). During 1982, more than 36 percent of domestic credit expansion went to the Government; the corresponding share in 1980 was less than 8 percent.
|Net foreign assets||–0.6||–0.8||0.1||4.8||6.6||2.8||0.2||—||6.7||–14.1||–20.8||–26.3||–44.7|
|Total domestic credit||12.8||14.3||16.5||18.6||26.1||34.3||43.5||52.3||65.0||81.4||99.1||122.1||151.0|
|Net claims on Government||(2.0)||(2.1)||(2.1)||(1.1)||(—)||(0.9)||(2.3)||(4.1)||(4.7)||(4.9)||(6.3)||(11.6)||(22.0)|
|Change in total domestic credit||1.4||1.5||2.2||2.1||7.5||8.2||9.2||8.7||12.7||16.4||17.7||23.1||28.8|
|Net claims on Government||(—)||(0.1)||(—)||(–10)||(1.1)||(0.9)||(1.4)||(1.8)||(0.6)||(0.2)||(1.4)||(5.3)||(10.4)|
2. Balance of Payments and Monetary Developments
Developments in the Philippine balance of payments have been characterized by steadily increasing trade and current account deficits. Deterioration in the trade balance during most of the period (except in 1973) was attributable to the rapid increase in imports (averaging 19.4 percent a year) relative to the growth of exports (averaging 13.8 percent a year) (Table 11). Notwithstanding increasing receipts from remittances, the deficit in the services account rose sharply during 1976-82, owing to increasing payments on freight and insurance and net investment income. Reflecting these developments in the trade and services accounts, the current account had deficits equivalent to 6.1 percent and 8.6 percent of GNP in 1981 and 1982, respectively. A steadily rising net inflow of capital tended to offset the developments in the current account balance, and the overall balance remained in surplus during most of the period.
|Errors and omissions and other||–99||140||103||28||69||–182||–178||–214||–113||–143||–117||118||–672|
Monetary policy was expansionary during most of the period. Total domestic credit increased at an annual average rate of about 23 percent during 1970-80; most of this rise was attributable to the nongovernment sector, since claims on the Government increased at a slower rate (12.2 percent). Government borrowing from the banking system increased sharply during 1980-82 from
This examination of the fiscal, monetary, and balance of payments developments over the period 1970-82 indicates that during most of the 1970s, the budget deficit remained low, as did the current account deficit. During 1980-82, financial stabilization programs were discontinued, the fiscal deficit and the associated domestic and external financial imbalances increased sharply, and the current account balance plunged to record deficit levels.
Before one attributes the deteriorating balance of payments position to budgetary developments, it is desirable to establish that fiscal policy was expansionary during the period under consideration; even if the overall fiscal deficit is not expansionary, a shift in the composition of government expenditure toward higher import intensity may also cause a higher current account deficit. If none of these is true, deterioration in the current account balance may be attributable to private sector demand or to other factors.
The approach considered to determine the stance of fiscal policy involves adjusting the actual fiscal balance for cyclical effects in an effort to obtain a cyclically neutral measure of the budget balance. The summary measure used in this section was originally proposed by the German Council of Economic Experts and employed in the Fund’s World Economic Outlook (1985).12 The actual budget balance (Bt) may be decomposed into a cyclically neutral component
t0 = TQ/Yq, base-year ratio of tax (t0) to GNP (Y0)
g0 = G0/Y0, base-year ratio of expenditure (G0) to GNP
Y = actual output in nominal prices
YP = potential output in nominal prices
T = government revenue
FIS = fiscal stance measure
The first two terms in equation (7)—normal balance and cyclical components—can be merged to define the cyclically neutral balance; the actual deficit in excess of the cyclically neutral deficit is deemed expansionary, relative to the base-year fiscal stance. To determine if the thrust of fiscal policy has been more expansionary relative to the previous year, the fiscal impulses (FI) may be examined by taking the first differences of the fiscal stance measure:
The expansionary or contractionary nature of total budgetary expenditure and its components are also examined. Government expenditure is termed cyclically neutral if it increases proportionately with increases in nominal potential output; a more-than-proportionate increase is considered expansionary, and vice versa.
The causality test used in this paper is based on the method developed by Pierce (1977) and Pierce and Haugh (1977). In the context of this analysis, the budget deficit is viewed as causing the current account deficit if it leads the current account deficit over time; leads and lags between the two series, when both are transformed to their stationary forms, determine the direction of causality. Possible transformations include ordinary or seasonal differencing, lags, or, more generally, the power transformation of Box and Cox (1964), and certain types of detrending. The transformations remove the effect on a series of its own past values and convert it into a form in which consecutive values of the series are uncorrelated.13 The cross-correlation coefficients between the current account deficit and the budget deficit in the previous and current periods after the transformation of this data are shown in Table 12. The coefficients indicate that for both yearly and biannual data, obtained from the Fund’s International Financial Statistics, the current account deficit of the balance of payments is correlated with the past value of the budget deficit; for annual data, the contemporaneous relationship is also statistically significant.
(t − 1)
(t − 1)
|Detrending (ordinary differencing)|
|Detrend (moving average)||0.3471||0.091|
These significantly positive correlation coefficients indicate that, in a plausible behavioral model for the Philippines, a widening fiscal deficit caused by weak revenue performance or excessive government spending has led to a deterioration in the current account deficit; the line of causality runs from the fiscal to the balance of payments deficit.
An argument has sometimes been made for reversed causality, according to which a reduction in exports may lead to a fall in imports, which may harm government revenue and lead to a wider fiscal deficit. Examination of the data on the Philippines indicates that exports increased steadily in all years except 1975, when recession in the industrial countries caused exports to decline (Table 11). Imports, however, continued to increase, as did international-trade-based tax revenue. In fact, revenue increased sharply in 1975, owing to new revenue measures adopted as a part of the stabilization program. Thus, reverse causality running from lower exports and the consequent current account deficit to the overall fiscal deficit is not relevant for the Philippines.
Aghevli, BijanB., and Mohsin S.Khan, “Government Deficits and the Inflationary Process in Developing Countries,”Staff Papers, International Monetary Fund (Washington), Vol. 25 (September1978), pp. 383-416.
Box, G.E.P., and D.R.Cox, “An Analysis of Transformations,”Journal of the Royal Statistical Society (London), Series B, Vol. 26 (No. 2, 1964), pp. 211-43.
Chand, SheetalK., “Summary Measures of Fiscal Influence,”Staff Papers, International Monetary Fund (Washington), Vol. 24 (July1977), pp. 405-49.
Dernburg, ThomasF., “Fiscal Analysis in the Federal Republic of Germany: The Cyclically Neutral Budget,”Staff Papers, International Monetary Fund (Washington), Vol. 22 (November1975), pp. 825-57.
Hansen, Bent, and Wayne W.Snyder, Fiscal Policy in Seven Countries, 1955–1965: Belgium, France, Germany, Italy, Sweden, United Kingdom, United States (Paris: Organization for Economic Cooperation and Development, 1969).
Heller, PeterS., Richard D.Haas, and Ahsan H.Mansur, “A Review of the Fiscal Impulse Measure, with Estimates of the Structural Budget Balance”(unpublished, International Monetary Fund, March21, 1985). This was later published as A Review of the Fiscal Impulse Measure, IMF Occasional Papers, No. 44 (Washington: International Monetary Fund, 1986).
International Monetary FundWorld Economic Outlook: A Survey by the Staff of the International Monetary Fund (Washington, 1985).
Kelly, MargaretR., “Fiscal Adjustment and Fund-Supported Programs, 1971–80” (unpublished, International Monetary Fund, September28, 1982).
Khan, MohsinS., and Malcolm D.Knight, “Stabilization Programs in Developing Countries: A Formal Framework,”Staff Papers, International Monetary Fund (Washington), Vol. 28 (March1981), pp. 1-53.
Milne, Elizabeth, “The Fiscal Approach to the Balance of Payments”(unpublished, International Monetary Fund, December14, 1976).
Otani, Ichiro, and Yung ChulPark, “A Monetary Model of the Korean Economy,”Staff Papers, International Monetary Fund (Washington), Vol. 23 (March1976), pp. 164-99.
Pierce, DavidA., “Relationships—and the Lack Thereof—Between Economic Time Series, with Special Reference to Money and Interest Rates,”Journal of the American Statistical Association (Washington), Vol. 72 (March1977), pp. 24-26.
Pierce, DavidA., and Larry D.Haugh, “Causality in Temporal Systems: Characterizations and a Survey,”Journal of Econometrics (Amsterdam), Vol. 5 (May1977), pp. 265-93.
Riha, TomásJ.F., An Evaluation of Fiscal Performance in the Philippines: 1947–73, Discussion Paper No. 75–7 (Quezon City, Institute of Economic Development and Research, School of Economics, University of the Philippines, June1975).
Snyder, WayneW., “Measuring Economic Stabilization: 1955–65,”American Economic Review (Nashville, Tennessee), Vol. 60 (December1970), pp. 924-33.
Tahari, Amor, “Budget Deficits, Credit Creation, and the Balance of Payments: Empirical Evidence for Brazil, Philippines, Sri Lanka, Thailand, and Venezuela”(unpublished, International Monetary Fund, October24, 1978).
Throughout this paper, current account balance (deficit) refers to the current account balance (deficit) of the balance of payments, unless stated otherwise.
Milne (1976) used time-series data to estimate the equation TBt = α + β(Gt - Tt) where the trade balance (TB) is regressed on the budget deficit; Kelly (1982) regressed the changes in the current account on the change in the overall budget deficit using cross-sectional data for a number of Fund program countries. Using a modified version of this type of testing, Tahari (1978) found a statistically significant relationship for some countries, including the Philippines, whereas for many other countries the result was negative.
For an overview of recent fiscal, monetary, and balance of payments developments, see Appendix I.
Estimates of trends in GNP have been used as proxies for GNP, since official estimates of potential gross domestic product (GDP) are not available for the Philippines. An alternative measure of potential GNP based on the “linked peak” method was used by Riha (1975). Conceptually, trend GNP represents output levels that are attainable; and from a fiscal point of view, government expenditure should be tied to a level of economic activity that is attainable, rather than to some hypothetical full employment or potential level of output that may not be attainable in the medium term.
Using an alternative approach based on Hansen and Snyder (1969) and Snyder (1970), Riha (1975) noted that in many instances during 1947-73 the estimated budget impact was destabilizing for the Philippines. Riha concluded that in periods when budget changes favorably influenced stabilization, much of the stabilization could be ascribed to fortuitous events rather than to cognizant policy.
This assumption is valid for most of the sample period (1970-82); a process of interest rate deregulation was initiated in 1980 and was completed in 1983. Since then interest rates have been fully market determined.
The estimated price equation in this form is
The t-ratios are shown in parentheses below the coefficients.
The sum of the aggregate parameter vectors is equal to the estimate of the parameter vector of the aggregate equation, and the expected values from the disaggregated equations also sum to those given by the aggregate equation if each kind of tax is explained by the same set of variables. For more on the problem of aggregation bias, see Theil (1954).
Based on the classification shown in International Financial Statistics (IFS) for the Philippines, the sum of net foreign assets and domestic credit equals the sum of money, quasi-money, bonds, and other items (net); the term OTM, which is used in this equation, is the sum of bonds and other items (net).
The simulated budget deficit, however, will not necessarily remain constant as a percentage of simulated GNP, since figures for both revenue and GNP will be different from the corresponding historical data.
One might suggest that causal relationships between the budget deficit (BD) and the current account deficit (CAD) can be determined through their sample crosscorrelations, or by regressing CAD on past and present values of BD and performing an F-test on the appropriate set of regression coefficients. Both these procedures (correlation and regression), however, can be misleading if autocorrelation is not appropriately taken into account; the types of autocorrelations normally found in the time series, if unattended, overestimate the significance of the tests and indicate relationships that do not exist. Consequently, it is desirable to make appropriate transformations of the time series to remove any existing autocorrelation prior to a cross-correlation analysis.