Policy, Research, and Extemal Affalrs WORKING PAPERS Macroeoonomic Adjustment and Growth Country Economics Departmernt The World Bank June 1991 WPS 696 Macroeconomics of Public Sector Deficits; The Case of Chile Jorge Marshall and Klaus Schmidt-Hebbel .~~~~~~~ Chile's success suggests that fiscal stabilization is a prerequisite for structural reform - and that structural reform need not be postponed until stabilization is fully achieved. The Pdicy. Re atd Fa in Cmnpika d&t. uPREWoring Paperm to dinaatethfiicwro* inaprr d ad to ege thcu change idesu g8nk saff and en intmed in developentim.re pps ay thc naMC f th autho. reiti Iay lheir veww, ad should be used and cie accordingly. The findinp, intrpeerats. and concluajno re tbe authors' own. They waid no bt auntad to the World Bank. iu Board of Dircm, its managtt. or anyof ius mnberm ctzes. Policy, Rearch, and EswrdAtl Mak laoroo la Adjusmnt adGrowth WPS 696 This paper a product of the Macroeconomic Adjustment and Growth Division, Country Economics Department - is part of PRE's series of case studies on the macroeconomics of public se'ctor deficits. Copies are available free from the World Bank, 1818 H Street NW, Washington DC 20433. Please contact Susheela Jonnakuty, roomNI 1-039, extension 39074 (99 pages). Marshall and Schmidt-Hebl' analyze the and investment - focusing on empirical esti- structure of public deficits in Chile, distinguish- mates of the different channels through which ing between consolidated nonfinancial public public spending and taxation crowd in or crowd deficits and quasifiscal losses of the Central out private spending. Bank - focusing on the detnminants and sustnability of the deficits. Fially, they measure the spillover effects from the deficit to the real exchange rate and the In the framework of an estimated portfolio trade surplus. model, they simulate the path of domestic inflation and interest rates for money-financed Chile's successful experience suggests that and debt-financed deficits. fiscal stabilization is a prerequisite for structural reform - and that structural reform need not be Then they trace the effects of deficits, and postponed until stabilization is fully achieved. their form of finncing, on private consumption The PRE Working Paper Series disseminates the findmings of work under way in the Bank's Policy, Research, and Extemal AffaisComplex Anobjeciveoftheseries is to getthesefindings outquickly, even if presentations are less than fullypolished. Tlhe findings. intaptudons. and conclusions in these papers do not necessarily represant official Bank policy. Produced by the PRE Dissemnination Center TABLEOF CONTNTS 1. INTRODUCTION ............... 1 2. FISCAL ?OLICY, DECOMPOSIMTON OF 1974-1988 DEFIC S AND SUSTAINABILITY OFlTHE DEFIC.. 6 2.1 Fiscal Policies during 1974-1988................................ 6 2.2 Decomposition of Public Sector Deficit. 9 2.3 Quasi-Fiscal Operations and Domestic Debt .21 2.4 How Sustainatle is the Deficit? .24 3. DEFICIT FINANCING, INTEREST RATES AND INFLATION .32 3.1 The Stylized Facts .33 3.2 Public Deficit Financing and Private Sector Portfolio .39 3.3 Estimated Equations .43 3.4 Simulation Results .50 4. PRIVATE CONSUMPTION AND PRIVATE INVESTMENT. .58 4.1 Private Consumption .59 4.2 Private Investment .68 5. RELATIVE PRICES AND THE TRADE BALANCE .76 5.1 Relative Prices of Exports and Imports .79 5.2 The Trade Balance .82 6. CONCLUSIONS .88 REFERENCES APPENDIX * Useful comments were provided by Bela Balassa, Juan Carlos Lerda, and Jim Hanson to a first draft. Efficient research assistance was provided by Eduardo Saavedra. TABLES Table 1.1 Selected Macroeconomic Indicators (1970-89) Table 2.1 Chile: Non-Financial Public Sector Operations Table 2.2 Non-;inancial, Quasi-fiscal and Total Public Sector Deficit Table 2.3 Decomposed Public Sector Variables Table 2.4 Estimation Results for Behavioral Public Sector Variables Table 2.5 Decomposition of Public Sector Revenues Table 2.6 Decomposition of Public Enterprise Profits Table 2.7 Decomposition of Public Sector Expenditures Table 2.8 Summary of Public Sector Deficit Decomposition Table 2.9 Central Bank Losses Originated by Quasi-fiscal Operations Table 2.10 Central Bank Estimated deficit Table 2.1' Public Sector Budget: 1991 Table 2.12 Central Bank Deficit Financing Table 2.13 Public Debt and Primary Surplus Table 3.1 Short-Run and Long-Run Consequences of Fiscal Deficits on Interest Rates and Prices. Table 3.2 Estimation Results for the MoneLy Demand Table 3.3 Estimation Results for the Public Debt Table 3.4 Reduced-Form Estimation Results for the Interest Rate and the Price Level Table 3.5 The Effects of an Increase in Fiscal Deficit Financed by Money Table 3.6 The Effects of an Increase in Fiscal Deficit Financed by Public Debt Table 4.1 Private Consumption (Chile, 1960-1988) Table 4.2A Private Investment (Chile, 1961-1988) Table 4.2B Private Investment (Chile, 1976-1988) Table 5.1 Relative Export and Import Prices (Chile, 1960-158) Table 5.2 Trade Surplus (Chile, 1960-1988) Table 6.1 Rate of Growth of Factor Inputs Table 6.2 Decomposition of the Rate of Growth of GDP FIGURES Figure 3.1 Nominal Interest Rate (1975-1988) Figure 3.2 Inflation Rate (1975-1988) Figure 3.3 Money (MI) (1975-1988) Figure 3.4 Public Debt (1981-1988) Figure 3.5 Seignorage Revenue Figure 3.6 Interest Rate: Effect of a Money-Financed Deficit Figure 3.7 Price Level: Effect of a Money-Financed Deficit Figure 3.8 Interest Rate: Effect of a Debt-Financed Deficit Figure 3.9 Price Level: Effect of a Debt-Financed Deficit Figure 4.1 Private Corsumption to Private Disposable Income Ratio Figure 4.2 Private Investment to GDP Ratio Figure 5.1 Relative Export & Import Prices Figure 5.2 Chile: Trade Surplus to GDP Ratio 1. INTRODUCTION Chile has faced massive external shocks and policy changes during the last two decades. Public finance has been at the center of the policy shifts, and fiscal deficits varied accordingly. This paper analyzes the macroeconomic effects of fiscal deficits in Chile since 1974 and up to the present. TABLE 1.1 SELECT-ED MACROECONOMIC INDlCATORS (1970-89) Non-Financial Year GDP Growth Inflation Current Public Sector (t) (%) Account Surplus (% of GDP) (% of GDP) 1970-73 0.7 204.3 -2.4 -23.4 1974 1.0 369.2 -0.5 - 5.5 1975 -12.9 343.3 -5.2 - 2.1 1976 3.5 197.9 1.7 4.0 1977 9.9 84.2 -3.7 0.4 1978 8.2 37.2 -5.2 1.4 1979 8.3 38.9 -5.4 4.6 1980 7.8 31.2 -7.0 5.4 1981 5.5 9.5 -14.4 0.4 1982 -14.1 20.7 -9.2 -3.9 1983 -0.7 23.1 -5.4 -3.5 1984 6.3 23.0 -10.7 -4.6 1985 2.4 26.4 -8.3 -2.9 1986 5.6 17.4 -6.9 -2.0 1987 5.7 21.5 4.3 -0.2 1988 7.4 12.7 -0.7 3.6 1989 10.0 21.4 -3.1 3.8 Sources: Central Bank and Budget Office. w Includes general government and public enterprises. -2- The evolution of the economy during the period is summarized in Table 1.1. From the perspective of fiscal policy and macroeconomic adjustment this period may be broken into several subperiods. The first two years of the 1974-1989 period were characterized by a major stabilization effort aimed at correcting the macroeconomic imbalances inherited from the Allende government. During this period most prices were liberalized, real assets that had been transferred to the state during 1970- 73 were returned to the private sector, the exchange rate was sharply devalued, and contractionary monetary and fiscRl policies were carried out to restore internal and external balances. As a result of both the stabilization policies and adverse foreign shocks, GDP declined by almost 13% in 1975, the unemployment rate jumped to 15%, and real wages declined to 65% of their 970 level. Beginning in 1976, Chile experienced four years of recovery with GDP growing at an averagc annual 7.5%. However, inflation was still high, largely due to the crawling peg rule of the nominal exchange rate. During this period the government carried out radical structural reforms, liberalizing domestic markets as well as external trade. Between 1975 and 1979 tariffs were reduced from high and widely varying levels to an almost flat structure of 10%. Interest rates were freed and most banks were privatized during this period. A significant fiscal adjustment was achieved very early in this period. The key active fiscal policies were the reduction of current and capital expenditure and the increase in public enterprise tariffs. The 1980-81 years correspond to a foreign-financed, private expenditure boom. The govemment, after three years of pre-announcing the path of the nominal exchange level, decided to fix this rate indefinitely in 1979 with the aim of achieving a convergence of domestic inflation to the external level. A combination of elements -- overestimated private wealth levels, high real wage increases due to backward indexation, and the capital account liberalization during a period of 3- unlimited foreign funds' -- contnbuted to the spending frenzy of the private sector. As a result, private external debt jumped from US$2.7 billion in 1978 to $8.1 billion in 1981. The current account deficit peaked at 14.4% of GDP in 1981. As oppK. J to most other developing country experiences, the public sector in Chile refrained from running deficits during the period of easy access to foreign funds. In fact, the 1978-SI phase is characterized by a sequence of surpluses in the public budget as a result of the preceding reforms. However, during the same period a radical reform of the social security system was implemente The old scheme -- a state owned pay-as-you-go system - was replaced by a capitalization system run by private corporations. This reform had -- and still has -- strong fiscal effects because the government has to finance the transition period. This implies paying pensions to the population which did retire or will retire under the old system, without receiving the revenues from the increasing share of the active population that was affiliated to the new private social security funds. The counterpart of this higher public deficit is an increase in private sector saving, part of which is financing the public sector by acquiring domestic public debt. 1982 and 1983 are the years of the debt crisis. After (and already before) August 1982 external credits dried out and the economy started a painful adjustment path resulting from reduced domestic spending in an economy unable to generate the price adjustment required to avoid a strong recession. The required real exchange rate depreciation did not materialize until 1983 due to the combination of a fLxed exchange rate policy and full backward wage indexation. As a result of this "double nominal anchor" problem, the adjustment was highly recessionary: GDP fell sharply, by 14% during 1982, and unemployment sks-rocketed to 30% the following year. In the medium term, the real 'For a discussion of the causality between domestic spending, foreign lending and the real excharge rate during this period see Morandd and Schmidt-Hebbel (1988), in particular chapter 5. -4- exchange rate and real wage adjustment made possible an expansion of exports, activity levels recovered, and the current account deficit was significantly reduced. During these years a significant change in the public sector balance took place. An initial surplus of 5.4% of GDP in 1980 revers A into a 4.0% deficit in 982. Both government revenue and expenditure were severely hit by the debt crisis, reflecting the decline in GDP and the deterioration of the terms of trade. Total tax revenue fell from 26%o of GDP in 1980 to 22% in 1982. Also, employment emergency programs had to be financed after the burst of unemployment in 1983. The 1984-1986 period is characterized by a major fiscal adjustment program aimed at both stabilization and structural reforms. While a reduction in current and capital expenditure was achieved, a new tax reform was implemented. Direct tax rates were signiificantly reduced with the purpose of encouraging private saving and investment. However, the most important fiscal policy innovations of this period were the so-called quasi-fiscal operations. In order to avoid the breakdown of the whole financial system, the government decided to assist the private sector by providing Central Bank credits at subsidized rates and exchange rate subsidies to debtors in foreign currency. These credits attained such levels that Central Bank real liabilities grew by a factor of 3.5 during the 1985- 1988 period. The last period, 1987-1989, is characterized by relative macroecornomic stability, high growth and fiscal consolidation. At the end of the 1980s, the Chilean econowrv ' 'ws a healthy condition. During these last years GDP has grown at an average 7.7% per year, inflation has averaged 18% per yea:, and the unemployment rate has declined to 5% in 1989. With regard to fiscal policy, this final phase shows a recovery of public finances. Fiscal deficits turned into surpluses as a result of ;he economic recoveiy, an increase of public enterprise surpluses and a recovery in the price of copper. -5- Hence there are three crucial aspects of the management of fiscal policy in Chile during the 1980s which are noteworthy. First, the tax and expenditure reforms contributed to a leaner government sector, to which the privatization of public enterprises made a significant contribution. Second, the 1981 social security reform caused a large negative effect on the public sector budget, which *vill last for the period of demographic transition to the new system. This implies that it is hard to make meaningful comparisons of fiscal accounts before and after the reform. Third, in order to avoid the collapse of the financial system, the government established a series of subsidized operations to the banking system and its debtors. An implication of this is that the increase in the domestic public debt was related to transfer payments to the private sector. These three aspects of the management of fiscal deficits in Chile were crucial in the macroeconomic adjustment during the 1980s. This study is organized as follows. Section 2 describes the evolution of fiscal policy in Chile during the 1974-1988 period emphasizing the structure and decomposition of non-financial public sector deficits, the scope of quasi-fiscal operations, and the sustainability of public sector deficits. Section 3 develops a three-asset portfolio model in order to estimate the impact of domestic deficit financing on interest and inflation rates and assess the implications of alternative strategies of domestic financing of public deficits. Section 4 analyzes direct and indirect fiscal policy effects on private consumption and investment. Section 5 estimates the impact of public sector deficits on real exchange rates and the current account. Section 6 concludes. -6- 2. FISCAL POLICY, DE.COMPOSITION OF 1974-1988 DEFICITS AA:) SUSTAINABILITY OF THE DEFICIT This section analyzes the evolution of fiscal policy in Chile since 1974. A brief introduction describing the evolution of fiscal policy during the 1974-1988 starts the discussion. In subsection 2.2 a decomposition of the non-financial public sector deficit is performed, which allows to distinguish between exogenous and endogenous fiscal policy determinants of the deficits. A quantification of quasi-fiscal operations is presented in 2.3. Finally, subsect;on 2.4 looks at deficit sustainability by presenting simulation results for the financing of the puiblic sector deficit and for the primary su.plus of the public sector under alternative public debt/GDP paths. 2.1 Fiscal Policies durin2 1974-1988 Table 2.1 shows the main determinants of the consolidated non-financial public sector deficits during the 1974-1988 period. There are four well-defined subperiods regarding the evolution of the public sector deficit during these years. The first subperiod covers the 1974-75 years during w,..n a drastic process of fiscal stabilization was carried out by the new military government. The last year of the Allende administration, 1973, was characterized by a huge fiscal imbalance resulting from large public enterprises losses, a decrease in fiscal revenues due to the Keynes-Olivera-Tanzi effect, and an over-expansion of public sector wages, employment and social security payments. In 1973 the public sector deficit climbed over 20% of GDi-, being financed by a huge inflation ;i The Pinochet government cut abruptly fiscal spending, reducing real wages, the public pay roll, and public investment. Tax rates were raised and many firms - which had been transferred to the state by the Allende govemment - were returned to the private sector. Th- remaining public enterprises were required to behave as private firms maximizing profits. In 1975, a tax reform -7- Table 2.1 : Chile: Ion-Finmncial PubLic Sector Operations 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 General Goverrnent Current Revenues 30.3 34.9 37.4 38.6 33.2 32.5 32.9 32.1 29.9 27.7 28.7 28.6 28.2 28.4 28.7 Direct Taxes 5.7 8.3 7.0 5.4 5.3 5.2 5.4 5.5 4.8 3.1 3.4 3.1 3.2 3.1 2.9 Indirect Taxes 13.0 14.6 14.0 14.9 13.7 13.3 13.4 14.8 13.8 14.6 16.3 17.1 17.1 17.3 14.5 Copper Taxes 1.7 1.3 2.6 2.6 2.8 1.9 1.9 0.2 1.0 1.9 1.3 0.5 0.6 1.1 4.7 Sociat Security 3.1 3.4 3.4 3.7 3.7 5.3 5.6 4.7 3.3 2.8 2.8 2.4 2.5 2.2 1.9 Other Incomes 6.8 7.3 10.3 12.1 7.7 6.7 6.6 6.9 7.2 5.3 5.0 5.5 4.9 4.7 4 7 Current Expendit. 26.4 27.6 31.0 33.0 26.8 24.8 24.5 26.7 31.9 30.6 30.7 29.5 27.4 26.3 23.2 Purchases 4.3 4.4 3.7 4.9 4.7 1,0 3.1 2.9 3.3 3.2 3.4 3.2 3.0 3.3 2.7 Wages 10.0 9.3 8.9 11.0 10.0 9.2 9.1 8.8 10.3 8.9 8.5 7.8 7.4 6.8 6.2 Domestic Interests 0.4 0.5 0.2 0.8 0.9 0.7 0.5 0.2 0.0 1.3 1.8 2.L 1.2 4.7 1.8 External Interests 0.9 2.1 2.0 0.9 0.7 0.5 0.4 0.3 0.5 0.5 0.6 0... 1.1 1.3 1.3 Total Transfers 6.4 10.3 12.3 12.2 9.8 10.8 10.9 14.1 17.3 16.3 16.2 15.0 14.3 1.0 11.2 Other expenditures 4.3 1.1 3.9 3.2 0.7 0.5 0.5 0.4 0.4 0.4 0.2 0.3 0.3 0.2 0.1 Savings 3.9 7.3 6.3 5.6 6.5 7.; 8.4 5.5 -1.9 -2.9 -2.0 -0.9 0.8 2.1 5.5 Pubtic Investment 8.7 5.7 3.1 4.2 3.5 3.2 2.6 2.5 2.1 2.1 2.3 3.1 3.3 3.0 2.9 Other Expenditures 1.9 1.6 0.1 0.5 0.8 -0.7 0.3 0.1 -1.8 -1.9 -0.7 -0.4 -0.8 -0.6 -0.8 Surplus -6.6 -0.0 3.1 0.9 2.1 5.1 5.5 2.9 -2.3 -3.1 -3.5 -3.6 -1.7 -0.3 3.4 ....... ............ .. .. .... ,......... . ... ........... ................... ................. ................................_. . ...... Public Enterprises Current Revenues 32.8 35.9 31.0 26.6 24.7 27.0 25.8 20.8 24.2 29.5 29.8 35.3 35.4 34.7 37.8 Current Expenditur 25.3 30.2 20.5 18.4 18.3 17.6 16.3 15.6 16.5 18.5 19.2 21.2 22.6 21.8 22.2 Net Savings 7.5 5.7 10.5 8.2 6.4 9.4 9.5 5.2 7.7 11.0 10.6 14.1 12.8 12.9 15.6 Taxes and Transfer 4.8 5.4 6.9 5.2 4.4 8.3 7.5 5.6 7.1 8.6 8.2 9.6 9.1 9.7 12.6 Investment 3.9 3.5 3.0 2.7 3.2 1.9 2.6 2.6 2.6 2.6 3.7 4.0 4.7 4.0 3.3 Other Income 2.3 1.1 0.3 -0.8 0.4 0.2 0.5 0.4 0.3 -0.1 0.3 0.3 0.5 0.9 0.6 Surplus 1.2 -2.1 0.9 -0.5 -0.8 -0.5 -0.1 -2.5 -1.6 -0.4 -1.1 0.8 -0.4 0.1 0.3 _ ............................... .......................................I........................................................._ Consolidated Surplus -5.5 -2.1 4.0 0.4 1.3 4.6 5.4 0.3 -3.9 -3.5 -4.5 -2.9 -2.1 -0.2 3.7 Financing 5.5 2.1 -4.0 -0.4 -1.3 -4.6 -5.4 -0.3 3.9 3.5 4.5 2.9 2.1 0.2 -3.7 Interna1 5.2 5.0 -0.9 0.q -4.0 -4.6 -5.3 -3.1 1.7 2.4 1.8 -1.1 -1.0 -2.2 -7.3 External 0.3 -2.9 -3.1 -0.3 2.7 0.0 -0.1 2.8 2.2 1.1 2.7 4.0 3.1 2.4 3.6 .e i o F ce.......................................................................... Source: Budget Office, llinistry of Ffnarice. 8- introduced the value-added tax, simplitfing notably the tax system and indexing the tax base As a result of these policies, the pubhic sector deficit was quickly reduced. In 1975, the deficit as a fraction of IDP was only 2%. The second subperiod covers the 1976-1981 years. The public sector showed inrporiant surpluses, allowing for a reduction in domestic public debt. Having stabilized the fiscal deficits in the previous years, public finances were favored by the recovery of growth, the new tax system, and the profit maximizing behavior of public enterprises. The third subperiod goes from 1982 until 1986. The debt crisis in 1982 triggered a sequence of public sector deficits. The recession that followed the debt crisis produced a fall in public sector revenues, particularly in external trade related taxes. The huge unemployment ra (30% in 1983) forced the government to undertake a series of programs that increased fiscal spending. During that period, the gcovcrnment undertook two structural reforms that contributed to the deterioration of public finances. First, a state-rut pay-as-you-go social security system was replaced by a privately run fully funded system, deepening an already existing deficit in the public social security system. Second, a tax reforn aimed at raising private investment reduced progressively direct tax rates after 1983. However, tle most importan' fiscal policy brought about by the debt crisis were the quasiGiscal operations carried out by the Central Bank. Beginning in 1983, the Central Bank rescued a technically bankrupt financial system using a package of subsidized loans. A similar operation was aimed at helping private domestic and external debtors. These operations are called quasi-fiscal to the extent that they respond to a fiscal instrument (subsidies) used by the Central Bank. This institution experimented heavy losses as a consequence of the referred policies, which were financed by transfers from the general government. Table 2.2 shows estimates of total public sector defir.its after taking into account quasi-fiscal operations. It is not simple to measure the contribution of the Central Bank to the consolidated deficit due to its accrual base accounting system. Two available estimates are shown in table 2.2. Both of them contain accrued terms, so they are not strictly comparable with thc non-financial deficit (measured on a cash ase) Anyhow, it is evident from table 2.2 that quasi-fiscal operations were an essenfial component of fiscal policy during the 1980s. Finally, the 1987-1989 period shows a strengthening of public sector finances, which not surprisingly parallels the behavior of most macroeconomic indicators. A significant contribution to this recovery was played by public enterprises, which benefited from high GDP growth, a recovery in copper prices and, running foreign exchange surpluses also were favored by the depreciated real exchange rate. 2.2 Decomogsition of Public Sector Deficits The decomposition of the non-financial public sector deficit attempts to measure the contribution of exoger.ous and endogenous fiscal policy variables to the fiscal deficit, following closely our methodology presented elsewhere (see Marshall and Schmidt-Hebbel, 1989a, b and a compact presentation in the appendix of Morande and Schmidt-Hebbel, 1991). Endogenous variables are those under the control of fiscal policymaklen: tax rates, public investment, public sector wages, and so on. Exogenous variables are domestic and external variables that affect the fiscal deficit but are beyond the direct control of the fiscal policymaker, such as inflation, output growth, interest rates, or the real exchange rate. - 10- TABLE 2.2 NON-FINANCIAL, QUASI-FISCAL AND TOTAL PUBLIC SECTOR DEFICIT (% of GDP) 1980 1981 1982 1983 1984 1985 1986 1987 1988 (1) Non financial -5.4 -0.3 4.0 3.3 4.5 2.9 1.6 -0.3 3.6 Public Sector Deficit (2) Central Bank -1.1 -1.9 -1.4 0.5 2.9 4.7 4.9 3.8 2.9 Net Interest Payments (3) Central Bank -1.0 -1.5 -0.6 6.2 6.3 21.0 4.2 0.5 5.2 Losses Total Deficit (1) + (2) -6.5 -2.2 2.6 3.8 7.4 6.6 6.5 3.5 -0.7 (1) + (3) ).4 -1.8 3.4 9.5 10.8 23.9 5.8 0.2 1.6 Sources: Table 2.1 and Central Bank Balance Sheets. There are two altematives to decompose fiscal variables, depending if they are classified as behavioral variables. For non-behavioral fiscal variables, an accounting decomposition is performed as shown by the following example. Let WL/PY be the wage bill paid by the government, as a share of current-price GDP. Then the discrete-time change in this ratio can be written as follows: (2.1) [WL/PY t- 1WL/PYJ e-1= [WL/P1 t-l(1 + + - P - + RI where hats (^) denote the corresponding percentage rates of change and R represents the algebraic residual due to the sum of the products of discrete-time rates of change. The above expression illustrates the change in a fiscal variable, decomposed by the variation of endogenous policy variables (wage rate, W, and public employment, L) and exogenous macroeconomic variables (GDP deflator, P, and real GDP, Y). Generalizing this procedure to other fiscal variables, it is straightforward to decompose the evolution of the public sector deficit into its endogenous and exogenous macroeconomic determinants. For behavioral fiscal variables the decomposition is based on their behavioral structure. Take for example the following equation for constant-price tax revenue: (2.2) (T/P) t = ao + alrt + a27t + a3e where real tax revenue (T/P) is a function of real output, the inflation rate (a) and the real exchange rate (e). The decomposition of the change in tax revenue as a fraction of GDP can be written as: (2.3) A(T/P)t = (7IPY),_ja,(PYI7),_,-l)Y + a2(P1T),_,di + a3 (eP/7),lD. In this case the decomposition uses the least squares estimators a, for the unknown coefficients a, that relate tax revenue to its structural determinants. Table 2.3 shows the list of decomposed fiscal variables, the methodology of decomposition, and the endogenous and exogenous variables that were used in the decomposition. Five public sector variables are decomposed using the estimation procedure: direct taxes, indirect taxes, copper taxes, public enterprises surplus, and total transfers to the private sector. The last variable includes all the transfers given by the government to the private sector, including social security payments. - 12 - In the case of direct and indirect taxes, real income, the inflation rate and the real exchange rate are the main determinants of tax revenue. The change in tax structures - and the corresponding change in tax rates - are reflected by changes in appropriate dunmny variables. OLS estimation results are shown in table 2.4. Taxes from the copper industry are a function of the real exchange rate and the price of copper. Total transfers to the private sector were correlated with real income and the inflation rate. Finally, the public enterprise surplus was related to the exchange rate and the copper price. The results of the non-financial public sector deficit decomposition are shown in tables 2.5 - 2.7. To avoid unnecessary detail the annual results were aggregated into four subperiods covering 1974-1988, relevant for fiscal policy (see section 2.1).2 Table 2.5 shows the decomposition results for public sector revenue, which registers an impressive improvement during the first subperiod, a slight deterioration in the second period, and only minor changes during 1982-86 and 1987-8. The massive 19.7 percentage point of GDP increase in public sector revenue during 1974-75 reflects a significant recovery in tax revenue and public firm surpluses. In turn, the improvement in tax revenues is explained by changes in effective tax rates due to the 1974 tax reform and the reduction in tax evasion. Declining effective tax rates during 1976-81 reflect higher levels of tax evasion. A negative effect on revenue is also due to the declining number of affiliates to the old social security system, as a result of the already mentioned reform. qThe figures in these tables denote cumulative changes of the corresponding variables. For instance, the cumulative variation in direct taxes in 1974-75, i.e. between 1973 and 1975 levels, is 3.2 percentage points of GDP. - 13 - Table 2.3 DECOMPOSED PUBLIC SECTOR VARIABLE Fiscal Variable Methodology Endogenous and Exogenous Variables Direct Taxes Estimation Copper price Exchange Rate Real Income Effective tax rate Direct Taxes Estimation Real Income Effective tax rate Indirect Taxes Estimaticn Real income Inflation Effective tax rate Social Security Accounting System Affiliates Revenues Effective tax rate Real income Purchases of Accounting Real purchases Goods and Services Income Wage Bill Accounting Public real wage Public employment Real Income Total Transfers Estimation Real income Inflation Interest Payments Accounting Debt Stock Interest rate Nominal income Public Investment Accounting Real Investment Real Income Pubic Enterprise Estimation Exchange Rate Surplus Copper price Real income -14- Table 2.4 Estmaetion Results of Behavioral Public Sector Variables (1973-1989) Dependent lndependent Variables Variable Const Income Exchange Inflation Copper Di D2 Rg DW Rate Rate Price 1. Direct 506.3 3.36 298 -494 .73 1.7 Taxes (.93) (1.92) (2.2) (3.0) 2. Indirect 364.9 10.65 -0.91 786 1833 .95 1.6 Taxes (0.5) (4.4) (1.5) (2.1) (2.9) 3. Copper -84.9 -6.34 7.41 27.7 .54 1.4 Taxes (0.2) (1.9) (1.6) (3.4) 4. Public -2869 43.0 23.5 .92 1.9 Firns (4.6) (10.3) (3.3) Surplus 5. Total 1121 11.2 -3.8 .65 1.6 Transfers (0.5) (1.8) (1.6) to Private Sector 1/ Direct taxes and transfers are a function of lagged income. '/ D1w1.0 for 1975-81, 0 for 1982-89 ln the case of direct taxes. 61-1.0 for 1975-83, 0 for 1984-89 In the case of indirect taxes. 3/ t-statatices are reported for each regression. XI Sample period: 1975-1989. -15- Table 2.5 Decomposition of Public Sector Revenues (Changes in Revenue caused by their Determinants, in Percentage Points of GDP)') 1974-75 1976-81 1982-86 98748 Direct Taxces 3.2 -2.8 -2.3 -0.3 Real Income -0.1 0.6 -0.1 0.3 Effective rate 3.3 -3.4 -2.1 -0.6 Indirect Taxes 4.4 0.1 2.3 -2.5 Real Income -1.3 4.7 0.0 1.4 Inflation 0.4 0.8 0.0 0.0 Effective rate 5.4 -5.5 2.4 -3.9 Copper Taxes 0.7 -1.0 0.3 4.1 Copper price -3.1 2.3 -1.2 3.9 Exchange rate 1.2 -0.6 1.5 0.1 Real income 0.7 -2.7 0.0 -0.8 Effective rate 1.8 0.0 0.0 0.8 Social Security -1.2 1.3 -2.2 -0.6 Nwmber of affiliates 0.1 -1.8 -1.1 -2.4 Real wage 0.0 1.5 0.4 0.1 Real income 0.4 -1.8 0.8 -0.3 Remainder -1.7 3.4 -1.7 0.2 Other Incomes 4.2 -0.6 -2.0 -0.1 Sector Income Public Evtervrise 8.4 -0.4 2.1 0.7 Sur2 us 3/ Total Variation 19.7 -3.0 -1.8 1.2 Notest 1/ The variation of a variable in 1974-75 is the change between the 1973 and 1975 levels. The same applies to other periods. 2/ The effective tax rate include tax evasion and other fluctuations in tax revenues. 3I Net of taxes and transfere to the government (see text). - 16 - The copper price plays an important role through its impact on copper tax revenue. Its strongly negative contribution in the first subperiod and its positive influence on public revenue during the last subperiod are particularly noteworthy. The impact of privatizations on public finances is hidden in the 'Other Public Sector Income" account. The contnbution of these asset sales to fiscal income has not been large in spire of the number of privatized enterprises. Note also that the surplus of public enterprises is net of taxes and transfers to the government. It is not possible from the available information to decompose tax revenue between private and public enterprise contributions. The same applies to "Other Public Sector Income", which includes transfers from public firms. In this respect, it is interesting to decompose the difference between public enterprise current income and expenditure, a proxy for profits obtained by this sector. The so-defined "profits" have three uses: taxes and transfers to the government, capital formation, and the surplus presented in table 2.5. Table 2.6 presents this additional information. This table reflects the importance of the real exchange rate and the copper price for Chilean public finances. The sensitivity to the real exchange rate reflects the dominance of tradable-goods producing firms among public enterprises and particularly the size of the state copper company (CODELCO) among them. Hence the strong 1982-86 real devaluations contributed to a massive 9.1 percentage point (of GDP) improvement in public enterprise profits. The important role of the price of copper reflects again the share of CODELCO in public enterprises. Table 2.6 also suggests the structural change in the behavior of public firms, which started maximizing profits in 1974. This is reflected by the residual term, which accounts for the effect of variables other than the exchange rate and the copper price in the evolution of public firms profits. - 17 - TABLE 2.6 DECOMPOSITION OF PUBLIC ENTERPRISE PROFIEM (Changes in Profits caused by their D'terminants in Percentage Points of GDI; 1974-75 1976-81 1982-86 1987.88 Total Variation 11.6 -0.4 7.6 2.8 Exchange rate 7.4 -3.5 9.1 0.8 Copper price -2.7 2.0 -1.0 3.3 Non-explained Variation 6.9 1.1 -0.5 -1.3 Table 2.7 shows the decomposition results for public sector expenditures. Looking at the variation in total expenditures, it turns out that sector public spending as a share of GDP contracted significantly during the 1976-81 and 1987-88 periods. Decreasing levels of public employment and investment caused large declines in public sector expenditures between 1974 and 1981. Fiscal adjustment during this period was associated with a decline in the size of government as measured by public employment and investment. On the other side, inflation increased public expenditure through its effect on total transfers, as a result of backward indexation rules in a period of declining inflation. Real income plays an important role in the fall of public expenditures during 1987-1988. This is simply the result of relatively constant levels of real fiscal spending during a period of strong growth. * 18- It is noteworthy that fiscal adjustment during the 1980s was not accomplished by the reduction in public investment, a common practice in other Latin American countries. Adjustment in terms of fiscal expenditure fell largely on public sector wages and transfers to the private sector. Finally, table 2.8 rearranges and summarizes the information provided by tables 2.5 and 2.7. It reports the net impact of dscal policy, macroeconomic, and external variables on the non-financial public sector deficit. The sharp decrease in the 1973-75 public sector deficit is largely explained by fiscal policy variables. The more important policies were the increase in indirect and direct effective tax rates, the decline in public employment, and the surge of the public enterprise surplus. As mentioned in section 2.1, fiscal policy during these years was aimed to restore macroeconomic balances following disequilibria during the Allende administration. The 1976-81 period is characterized by a relatively constant level of the deficit. However, the results in table 2.8 suggest that this relative flatness results from two opposite forces. On one side, fiscal policy variables - effective tax rates, public sector real wages and the external public debt - increase the public sector deficit. Also, a decreasing inCfation rate - together with a backward indexation scheme - pushes the deficit up. On the other sice, a large increase in real income and the rise of copper prices reduced the size of the deficit. During 1982-86, the source of an increasing deficit are fiscal policy variables: reduced direct tax rates, declining number of affiliates to the state-run social security system, and the rising public debt. The change in these variables reflects the tax reform of 1982, the social security reform of 1980, and the government intervention in the economy after the debt crisis, respectively. Finally, the 1987-88 recovery of the non-financial p-,.lic sector is associated with changes in macroeconomic and external sector variables. In particular, the surge in real income and copper prices explain a decreasing public sector deficit. On the other side, fiscal policy variables like indirect -19- Table 2.7 Decomposition of Public Sector Expenditure (Changes in Expenditure caused by their Determinants in Percentage Point of GDP) 1974-75 1976-81 1982-86 1987-88 Purchases goods & servs. 0.2 -1.4 0.1 -0.3 Real Purchase -0.3 0.2 0.0 0.1 Real income 0.5 1.7 -0.0 -0.4 Non-explained 0.0 0.1 0.1 0.0 Wage Bill -0.2 -0.5 -1.4 -1.2 Public Employment 4.3 -2.1 -0.1 0.0 Real wages 0.2 3.3 -1.5 0.4 Real income 1.2 -4.2 0.1 -0.9 Non-explained 2.7 2.4 -0.1 -0.6 Eoreien Int. Payments 1.4 -1.8 0.9 -1.1 Debt Stock 1.7 2.8 2.9 2.5 Interest rate -0.2 0.5 -0.5 0.2 Real income 0.1 -0.5 -0.1 -0.2 Exchange rate 0.6 -0.3 0.4 0.1 Non-explained -0.8 -3.8 -1.9 -3.8 Domestic Paments 0.5 -0.3 1.1 -1;2 Debt stock 2.7 0.1 Interest rate 0.2 0.2 Real income -0.1 0.8 Non-explained -1.7 -2.4 ota Transfr 2.6 3.8 -0.5 -1.5 Inflation 2.5 5.4 -0.1 0.1 Real income 0.2 -2.1 -0.6 -0.3 Non-Explained 0.1 0.5 0.2 -1.3 Public In_es j -0.2 0.8 -0.3 Real investment -0.8 -1.8 0.8 0.0 Real income 1.1 -1.5 0.0 -0.4 Non-Explained -0.4 0.1 -0.1 0.0 Other Expenditures -3.5 -2.3 -0.9 -0.2 Total Expenditures 0.8 -5.7 -0.2 -4.3 l -20- TABLE2. SUMMARY OF PUBLIC SECTOR-DEF2CI DECOMPOSITO-N (Changes in Public Sector Deficits Compared by Groups of Determinants, in Percentage Points of GDP) 1974-75 1976-81 1982-86 1987-88 Fiscal Policy Variables -14.1 13.2 5.7 5.2 Indirect tax rate -5.4 5.5 -2.4 1.8 Direct tax rate -3.3 3.5 2.2 0.7 Copper tax rate -1.8 0.0 0.0 -0.8 Public employment -4.3 -2.1 -0.1 0.0 Affiliates soc. sec. -0.1 1.8 1.1 2.4 Public investment -0.8 -1.8 0.8 0.0 Public s. real wage 0.2 3.3 -1.5 0.4 Public external debt 1.7 2.8 2.9 0.5 Public domestic debt 2.7 0.1 Real purchases -0.3 0.2 0.0 0.1 Macroeconomic Variables 4.7 -7.5 -2.8 -2.7 Real income 3.4 -10.8 -1.3 -2.9 Real exchange rate -0.7 0.3 -1.2 0.0 Inflation rate 2.1 4.6 -0.1 0.1 Real wage -0.1 -1.6 -0.4 -0.1 Domestic interest 0.2 0 2 External Seetor Variables 2.9 -1.8 0.8 -3.6 Copper price 3.1 -2.3 1.3 -3.9 External interest -0.2 0.5 -0.5 0.3 Non-Decomposed Accounts Public Enterprise Surplus -8.4 0.4 -2.1 -0.7 (after transfers) Other Public S. Income -4.2 0.6 2.0 0.1 Other Public S. Expenditure -3.5 -2.3 -0.9 -0.2 Explained Deficit -22.6 2.3 2.7 -1.9 Non-explained Deficit 4.1 4.7 -0.9 -3.9 Total Deficit -18.5 -2.4 1.8 -5.8 - 21 - tax rates (a decrease in the VAT rate) and the continuing decline of affiliates to the social security system had a negative impact on public finances during this period. 2.3 Ouasi-Fiscal Operations and Domestic Debt An essential feature of the Chilean adjustment process after 1982 was the quasifiscal operations to support a bankrupt financial system, in effect subsidizing private debtors and financial institutions. Quasi-fiscal operations are carried out by central banks. Nonetheless, they correspond to a fiscal policy action rather than a purely monetary or financial one. Examples of quasi-fiscal operations are subsidized credits, exchange rate subsidies, free-of-charge guarantees given by the Central Bank, or any other policy containing an element of subsidy to private sector agents. The nature of these operations often makes its estimation difficult. Moreover, the accounting systems utilized by central banks and general governments differ. The Central Bank uses an accrual base system whereas government transactions are registered on a cash basis. This feature complicates the joint treatment of both institutions. The main quasi-fiscal programs developed by the Chilean Central Bank after the debt crisis were: 1. Credit to Bankrupt Banks: In 1981 the Central Bank offered emergency credits to eight financial institutions facing insolvency. Later all these institutions went bankrupt and the total amount of these credits turned into Central Bank losses. 2. Purchase of Bad Loans: The Central Bank acquired bad loans from the private banks, with the latter's commitment to purchase back the bad portfolio in the future. In this sense, the Central Bank provided liquidity to the financial sector. However, the operation also involved a subsidy from the Central Bank as the credits were given at subsidized rates. - 22 - 3. Preferential Exchange Rate: Private external 'ebtors were allowed to purchase dollars at a subsidized exchange rate for servicing their debt, after the massive 1982-83 devaluations which raised the domestic-currency cost of foreign debt service. 4. Debt Rescheduling: The Central Bank rescheduled domestic debts and financed this operation through loans at negative spreads. 5. Exchange Rate Guarantees: The Central Bank purchased foreign exchange with the commitment to sell it back after a year at the purchasing exchange rate adjusted by the inflation rate. In a period of an appreciating real exchange rate, the above operations produced losses to the Central Bank. 6. Foreign Exchange Capital Losses: The acquisition of private external debt pushed the Central Bank into a net external debtor position. Large real devaluations during ithe 1982-88 period caused significant capital losses to this institution. Table 2.9 reports estimated losses incurred by the Central Bank due to quasi-fiscal operations. Losses were estimated as the present value - at December 1989 - of total Central Bank disbursements minus the economic value of the corresponding assets. The total estimated loss is about US$ 9.0 billion. This figure represents almost 40% of 1989 GDP. Its equivalent cash-flow is about US$ 540 million, considering an average 6% interest rate on Central Bank liabilities. The Central Bank financed most quasi-fiscal operations by issuing domestic debt. As a result, its outstanding domestic debt increased by US$ 5.7 billion during the 1982-89 period, while its extemal debt increased by US$ 4.0 billion. -23- TABLE 2.9 CENTRAL BANK LOSSES ORIGINATED BY QUASI-FISCAL OPERATIONS (US$ million, December 1989) Qu5asi-Fiscal Present Value Estimated Operation of Disbursement Asset Value Net Loss Credit to Bankrupt Banks 1,930 0 1,930 Purchase of Bad Loans 3,114 2,513 601 Preferential Exchange Rate 3,320 0 3,320 Debt Rescheduling 1,570 1,180 390 Exchange Rate Guarantee 1,585 0 1,585 Foreign Exchange Capital Losses 1,227 0 1,227 TOTAL 12,746 3,693 9,053 Source: Eyzaguirre and Larranaga (1990) - 24 - In order to finance Central Bank losses, the general govemment transferred to this institution approximately USS 7.2 billion in Treasury bonds during 1983-86. In this form, the government recognized that the Central Bank was his financial agent during the debt crisis. Nonetheless, the Central Bank exhibits a large cash-flow deficit as shown in table 2.10. This deficit is rooted in quasi-fiscal operations and is not financed by the above mentioned fiscal transfers to the extent that Treasury bonds in hands of Central Bank yield a (minimum) real return of 2% while the average real interest on Central Bank liabilities is 6%. 2.4 How Sustainable is the Deficit? This sub-section analyzes two issues related to the sustainability of public deficits. First, the focus is on deficit financing under the fiscal policies carried out now in Chile. Second, the issue of public debt is addressed. 2.4.1 Fiscal Policy and Deficit Sustainability under the New Govemnment The drastic structural reforms - as well as the deep stabilization programs - that allowed the Chilean economy to start a dynarnic growth path were undertaken under military rule. On March 1990, a new president took over the government after having been elected by a broad political coalition. The new administration has stressed that the basic framework - a stabilized, free-market, open economy - will be preserved. However, the new government has also stressed that social demands, largely postponed under the previous government, will strongly emerge. Political stability wculd require at least a partial satisfaction of these demands. The relationship between political and economic stability is critical for the new government. Fiscal policy is at the center stage. On the one side, the satisfaction of social demands requires an increase in govemment spending, particularly in education, health, housing, and social security. On -25 - TABLE 2.10 ESTMAED CENTRAL BANK DEFICIT (As % of GDP) Expenditure 2.95 Domestic Liabilities 1.66 External Liabilities 0.83 Dollar-Denominated Domestic Liabilities 0.48 Revenue 1.05 International Reserves 0.58 Loans to the Financial Sector 0.47 Treasury Bonds 0.43 Deficit 1.47 Source: Eyzaguirre and Larranaga (1990) -26. the other side, the financing of higher expenditure can be a delicate matter: tax hikes are politically resisted and can jeopardize investment and growth; debt financing is seriously limited given the already large stock of public debt; and money financing is out of the question given the stabilization commitment of the new government. Anyhow, since the end of 1989 the Central Bank is autonomous so that the government would not have access to inflationary financing. During its first four months in power, the new government obtained parliamentary approval for a tax increase. The new law established an increase in the VAT rate from 16% to 18%, an increase in the profit tax rate from 10% to 15% together with an expansion of the tax base, and a rise in personal income taxes. An estimated additional US$600 million (2.3% of GDP) is expected to be collected as a result of the tax reform. Table 2.11 shows the estimated public sector budget for 1991. Calculations include the effect of the tax reform and increased social expenditure. Overall the non-financial public sector shows a balanced budget. However, the Central Bank estimated deficit would reach 1.5% of GDP. Assuming that macroeconomic conditions do not change too much, the 1991 budget turns out to be a good proxy for future public sector budgets as fiscal policy is expected to remain unchanged at least until the next election year (1994). Hence the estimates from Table 2.11 focus our attention on the financing of the Central Bank deficit. The Central Bank deficit reflects the quasi-fiscal debt issue analyzed in the previous sub- section. As mentioned, the autonomous Central Bank can finance its deficit by printing money or by issuing debt. Along the lines of Buiter (1988) we define a sustainable deficit level as one that avoids increasing debt/GDP ratios over time. Table 2.12 shows simulation results for Central Bank debt/GDP ratios for different levels of monetary financing. -27- TABLE 2.11 ESTlMATED 1991 PUBLIC SECTOR BUDGET (% of GDP) General Government Current Revenue 24.7 Operatacnal 1.6 Social Security 1.9 Income Tax 3.8 Value-Adcded Tax 8.9 Trade Taxes 3.0 Other Taxes 2.1 Copper Surplus 1.7 Other Income 1.7 Current Expenditure 20.8 Wages 3.8 Purchases Z6 Social Security 7.0 Transfers 5.3 Interest Payments 2.1 Saving 3.8 Investment 3.1 Other Capital Expenditure 0.7 Surplus 0.0 2. Public Enterprises aving (after transfers)2.4 Investment 2.4 SuM] 0.0 3. Central Bank Current Income 1.5 Domestic Assets 0.5 Extemal Assets 0.6 Current Extenditure 3.0 Domestic Uabilities 2.2 External LUabilities 0.8 Surplus .1.5 Consolidated Total Public Sector Srnslus -1.5 Source Budget Office and Table 2.10. -28- As expected, monetary financing - compromising seignorage and the inflation tax - is a direct function of GDP growth and the inflation rate. Total revenue for the Central Bank fluctuates between 0.6% to 1.1% of GDP depending on the macroeconomic scenario. The Central Bank has to finance the remaining deficit by issuing domestic debt. The initial domestic pubHic debt/GDP ratio is 30%o and the initial total public debt/GDP ratio is 60%o. Table 2.12 shows what happens to these ratios after five years of debt-financed (after-money-financed) deficits.3 The evolution of debt/GDP ratios depends on two factors. On one side, large revenue from monetary finance requires a low amount of debt fnancing, so that the debt/GDP ratio decreases over time. On the other side, high GDP growth reduces the debt/GDP ratio. Total public debt/GDP ratios show a more favorable evolution than domestic public debt/GDP ratios as the former include external Central Bank debt, whose path is given exogenously. The simulations results in Table 2.12 suggest that the Central Bank defic,t - and hence the total public sector deficits' - is sustainable under reasonable macroeconomic conditions as reflected by declining debt/output ratios. However, some difficulties might arise from the fact that the deficit is concentrated in the Central Bank. In particular, the autonomous character of this institution could be jeopardized if the Central Bank depends on the government to finance its deficit. 2.4.2 Public Debt and the Sustainability of Deficits Apart from financing social expenditures, public finances will be stressed by the large levcl of public debt. Total public debt increased from US$7.9 billion in 1981 to more than US$22 billion -Table 2.12 simulations are computed assuming constant real interest rates. 'Recall from Table 2.11 that the estimated 1991 deficit is entirely concentrated in the Central Bank. -29- in 1989. ITis figure includes both domestic and external debt, held by the financial and non-Ginancial public sector. The steep increase in public debt reflects the government intervention in the economy after the debt crisis. Leaving aside public firms and the large state bank (Banco del Estado), an estimation of the net public debt of the consolidated general government and Central Bank is about 60% of GDP. To illustrate the implications of public debt for public finances, let's start with the public sector budget constraint. If the public deficit is financed through monetarv creation and issuance of domestic and foreign public debt, financing of the primary deficit, as a fraction of GDP (d), can be expressed as: (2.4) d=mh +m(i +n) +b6+b* -b(r-n) -b*(r*-n+ p) where a is the inflation rate, n is the GDP growth rate, r is the real interest rate paid on domestic debt, r is the real interest paid on foreign debt, and /u is the rate of real devaluation.' The outstanding stocks of base money, domestic public debt, and external public debt - as shares of GDP - are denoted by m, b, and b*; respectively. A dot over a variable denotes the time derivative. Table 2.13 shows simulations results for primary public surplus levels required for (i) maintaining a constant debt/GDP ratio, and (ii) decreasing this ratio by S percentage points of GDP per year. A clear implication from Table 2.13 is that the large public debt outstanding makes public finances highly sensitive to macroeconomic fluctuations. Under an optimistic scenario, with output 5'be real domestic (foreign) interest rate r (r*) is defined as the difference between the nominal domestic (foreign) interest rate and domestic (foreign) inflation, and the rate of real devaluation is defined as the rate of increase of the nominal exchange rate plus foreign inflation minus domestic inflation. -30 - growing at 5% and interest rates at 6%, the public sector can maintain its debt ratio by running a 0.8% of GDP deficit. Alternatively, it may reduce significantly the debt/GDP ratio by 5 percentage points of GDP, running a primary surplus of 3.6% of GDP. However, in a bad year with 2% growth and 10% interest rates, the public sectcr would have to generate a primary surplus of 4.2% to maintain constant the debt/GDP ratio, required to jump to 8.6% of GDP if this ratio were to be lowered by 5%. TABLE 2.12 CENTRAL BANK DEFICIT FINANCING AND DEBT/GDP RATIOS6 Inflation Rate: 15% GDP Growth: 2% 5% 7% Monetary Financing 0.6 0.7 0.8 (% of GDP) Domestic Debt/GDP 32.1 27.2 24.4 Total Debt/GDP 59.2 50.1 45.8 Inflation Rate: 30% GDP Growth: 2% 5% 7% Monetary Financing 0.9 1.0 1.1 (% of GDP) Domestic Debt/GDP 30.4 25.8 23.1 Total Debt/GDP 57.6 49.4 44.4 6Debt/GDP ratios are shown for year 5 of deficit financing, starting at levels of 30% for domestic debt and 60% for total public debt at year 0. * 31 - TABLE 2.13 P!RIMARYPUBLIC SURPLUSES UNDER ALTERNAITIVEPUBLIC DEBT: PATHS Real Primary Interest GDP Surplus Rate Growth (% of GDP) (%) (%) (i) Constant Public DebtGDP Scenario: Optimistic 6.0 5.0 -0.8 Pessimistic 10.0 2.0 3.6 (ii) Decreasing Public Debt/GDP by 5% Per Year Scenario: Optimistic 6.0 5.0 4.2 Pessimistic 10.0 2.0 8.6 Note: Simulations assume an inflation rate of 15%, a constant real exchange rate, and equal domestic and extemal real interest rates. Base values correspond to 1987. - 32 - 3. DEFICIT FINANCING, INTEREST RATES AND INFLATION This section develops a model for assessing the inflation and interest rate effects of alternative strategies of fimancing public sector deficits. The past years have witnessed a revival of interest in the relation between fiscal deficits and macroeconomic variables. The growing literature on fiscal deficits has dealt with a variety of issues such as the relation between fiscal deficits and economic growth, the sensitivity of macroeconomic variables to changes in the public sector budget and the channels through which fiscal deficits are transmitted to the rest of the economy. This literature emphasizes the implications of the increased reliance on domestic financing of public deficits as a result of the sharp reduction in extemal financing to most high-debt countries7. In this section a framework is proposed for analyzing empirically the macroeconomic consequences of alternative strategies of domestic financing. The model introduces modifications to the framework proposed in Easterly (1989) along a model developed by Mujica (1990) in order to capture specific features of the Chilean economic experience during the 1980s8. The rest of this section is organized as follows. Subsection 3.1 provides the empirical background for the issues discussed in this section. Subsection 3.2 presents the basic structure of a model that relates explicitly the means of domestic financing to the behavior of interest rates and inflation. In subsection 3.3 we estimate asset demand equations and reduced-form equations for the nominal interest rate and the price level. Fmally, subsection 3.4 discusses some simulation exercises for alternative deficit financing policies. 'Evidence of this interest is reflected in Buiter (1988), van Wijnbergen (1988) and Dornbusch (1984,1987). 'For a reference on the Chilean experience in the 1980s see Corbo (1985,1989), Edwards and Edwards (1987), Ffrench-Davis and de Gregorio (1987) and Morande and Schmidt-Hebbel (1988). - 33- 3.1 The Stylized Facts A brief review of the policies and evidence carried out in Chile during the 1980s from the perspective of their effects on interest and inflation rates will help to introduce this section. The stylized facts are depicted in Figures 3.1 - 3.4. Figure 3.1 shows the evolution of the short-run nominal interest rate (30 days) using quarterly data for the 1975.1-1988.IV period. Figure 3.2 depicts the evolution of the 30 - day inflation rate for the 1975.11-1988.IV period. Figures 3.3 and 3.4 summarize the evolution of money and domestic public debt (as percentages of GDP) for the 1976.1- 1988.IV and 1983.I-1988.IV periods, respectively. Comparing figures 3.1 and 3.2, a similar time pattern emerges for both the interest and the inflation rate during two distinct periods. In the first period, from 1975.1 through 1978.IV, both rates decline systematically. In the second period, from 1979.! through 1988.IV, the two rates remained stable with the exception of a sharp upward jump in the fourth quarter of 1982. The relative stability of interest rates and prices in Chile during the 1980s -- by Latin American standards - contrasts with the erratic behavior of money and particularly with the upward trend of the stock of domestic public debt. Figure 3.3 illustrates the significant changes of Ml velocity of circulation: M1 as a share of GDP ranged between 13% and 31%. Demand for real M1 as measured by its velocity of circulation peaks in the first quarter of 1983 and thereafter starts a downward trend. Between 1983.1 and 1987.IV Chile's real balances -- as a percentage of GDP -- decreased by 12%. Figure 3.4 shows the evolution of the domestic public debt as a share of GDP. From a stable low level of around 10% before the debt crisis in 1982, public debt rises sharply to 80% of GDP in 1983.1, starting an upward trend until 1986.IV. As many other countries in Latin America, Chile faced in the early 1980s an adverse external environment which triggered a crisis of vast proportions forcing major adjustments in macro policy, particularly in the management of fiscal deficits. However, what seems to be of particular interest - 34 - in the Chilean case is that the sharp decline in extemal financing was preceded by the implementation of a set of economic reforms aimed at increasing the role of market mechanisms and reducing existing barriers to international trade and capital movements. As discussed in Section 2, the reduction in the government's role in economic activity was the main factor behind the sequence of surpluses in the public sector budget before the debt crisis in 1982. As a result of these policies, the government was able to finance the deterioration in the public sector budget after the debt crisis in 1982 without destabilizing the economy. On the other hand, a social security reform implemented during the same period and the management of the debt crisis modified drastically the nature of the effects associated with a change in the public sector budget. In the first place, the reform implemented in the social security system had strong fiscal consequences and was related with an increase in private saving, part of which was transferred to the public sector as domestic public debt issues. Second, as a result of the debt crisis and the overexpansion of domestic credit the financial system was in a very fragile position. In order to avoid the breakdown of the financial system, the Central Bank provided credit to the private sector at subsidized rates. ITis subsidy accounted for the quasi-fiscal content of the credit policy and implied a financial transfer from the public sector to private corporations. Both the social security reform and the quasi-fiscal operations of the Central Bank must be- considered in any attempt to explain the inflation and interest effects related to changes in fiscal deficits. The empirical modei presented in this section will emphasize these specific features of the fiscal adjustment undertaken in Chile. FIGURE 3.1 NOMINAL INTEREST RATE ( 1975.1 - 1988.4 ) 0.15 - . . ._ 0.14 - 0.13 - 0.12 0.11 z 0.1 0 0.09 - a. 0.08 w O 0.07 U' 0- Z 0.046 0.05 0.01 0 76. 7 78.1 I 8 81.1 8 8 8 8 8 87.1 8 1 I I Tllll l 75.1 76.1 77.1 78.1 79.1 80.1 81.1 82.1 83.1 84.1 85.1 86.1 87.1 88.1 FIGURE 3.2: INFLATION RATE ( 1975.2 - 1988.4 ) 0.7 0.6 0.5 F- z 0, 0.4 a. w 0.3 I-n - z W ~0.2 0~~~ 70. I 76.1 77.1 78.1 79.1 80.1 81.1 82.1 83.1 84.1 85.1 86.1 87.1 88.1 FIGURE 3.3 MONEY ( Ml ) ( 1975.1 - 1988.4 ) 31 -. 30 29 28 27 26- a. 25- 0 ~ 24- La. o 23- O ~22 z 21- O 20- iL ~19- 17 1 6 15 14 13- 75.1 76.1 77.1 78.1 79.1 80.1 81.1 82.1 83.1 84.1 85.1 86.1 87.1 38.1 FIGURE 3.4: PUBLIC DEBT ( 1981.4 - 1988.4 ) 230 190 180 170 1 6Og 150- a. 140- 0 C] 130 - hl 1210 -2 0 110 - 10 - z O i 90 - C-) 80 - Li 70 - 60 50- 40- 30 20 10 r 1 v 82.1 83.1 84.1 85.1 86.1 87.1 88.1 - 39 - 3.2 Public pefinai nvate Sector Portfolio The point of departure of the model is the basic government financing identity, which may be stated by rewriting (2.4) as an equation for the nominal total consolidated public sector deficit: (3.1) G + i + + e where G, De, D, H, i, i' and e are respectively the primary deficit, the stock of foreign public debt, the stock of domestic public debt, the stock of high-powered money, the domestic nominal interest rate, the foreign nominal interest rate, and the nominal exchange rate. The government budget identity states that the total public deficit consisting of the primary deficit plus foreign and domestic interest payments must be financed by changes in foreign and domestic public liability holdings. However, since for most developing countries after 1982 external borrowing was quasi-exogenous, an increase in the primary deficit must be basically financed by issuing bonds or by creating high-powered money. Therefore, the model will emphasize the trade-off associated with the choice between money and domestic debt financing for given levels of the primary deficit and foreign financing. This choice has serious implications for both real and nominal variables. Hence the focus in this section will be on inflation and interest rates. Consider a small open economy that faces an exogenous flow of external credit. In each period individuals allocate their wealth, which is fixed at any point in time, among three broad assets: domestic money, public sector bonds and interest bearing foreign assets. The long-run portfolio demands for these three assets are given by the following expressions: dm -p, = a+ al y, + 2 It + a (it + J + W 3a <>O a20 U3 O -40- (3.3) dd - Pt = p P I Y, + P2S + pP(i' + 8) + W P&< 02> P3 T2<0 V3>O where w is real financial wealth, y is domestic real output, p is the domestic price level, i is the domestic nominal interest rate, i is the foreign nominal interest rate, m is the domestic nominal money supply, m is foreign nominal asset holdings, d is public sector bonds, e is the nominal exchange rate, and 6 is the expected rate of change of the exchange rate. All variables, except i, i and 6, are expressed in natural logarithms. The subscript t refers to time and the superscript d denotes demand. Expected coefficient sigps are denoted below each equation. Equations (3.2) - (3.4) specify long-run portfolio demands - as a fraction of real wealth - as depending on domestic real output, the domestic interest rate and the foreign interest rate adjusted by the expected rate of change of nominal exchange rate depreciation. Total wealth is given by the sum of the three asset holdings. Equations (3.2) - (3.4) assume implicitly unit elasticities with respect to wealth. A more general form would allow for non-unit wealth elasticities. The wealth adding-up constraint implies that only two of the three demand equations are independent and therefore from now on we will focus on the equilibrium conditions in the domestic money and public sector bonds markets. Let's focus now on the nature of short-run stock adjustment for the holdings of real balances and domestic public debt. Assuming a partial adjustment mechanism due to the existence of implicit - 41 - adjustment costs, the change in the stock of real balances and real public bonds holdings can be written as: (3.5) (m,-ps) a[(m,d-p,) - (m _ -p,.)] + OS O -P2 -' 24 Run Effect Short 1P2( 6) > O 2( ?-1) 1L1 - 14 I P - 09 Run Effect PRICE LEVEL P2 >o P2 - O'2 U2 Long P2 - 2> Run Effect -46- 3.3.2 Reduced-Form Equations for Interest Rates and Prices In order to assess the ability of the model to explain directly the behavior of interest rates and prices, we estimated their reduced-form equations for a 1983.1-1988.IV sample. Table 3.4 reports the estimated coefficients for both the interest rate and the price level equations (3.9) and (3.10). The results are quite consistent with the hypothesis implied by the model and with the results obtained from the estimation of the structural equations (3.7) and (3.8). All variables with an unambiguous effect on the interest rate and the price level present the expected signs. The only exception is lagged real public bond holdings, which has the wrong sign in both the interest rate and the price level equations, but is not significant. This result can be explained in terms of the values of the estimated coefficients of the structural equations. For a high value of the coefficient related to the adjustment speed in the public bonds market (u) -- which is the case according to our estimations -- the model implies small values for the coefficients of lagged real public bonds in the reduced-form equations (see Appendix 1). In addition, for those variables with en ambiguous effect in the reduced-form equations, the signs of the estimated coefficients coincide in all cases with the signs implied by the values obtained from the structural equations. On the other hand, the overall fit for both the price level and the interest rate equations is quite satisfactory, in particular for a period dominated by frequent changes in government interventions and private expectations. The results also tend to confirm the hypothesis, implicit in the model, that the nominal interest rate and the price level have been sensitive to both foreign and domestic influences. An interesting result concerns the effect of lagged real balances in the interest rate equation. As opposed to Edwards (1988) findings, here is clear evidence of a positive relationship between the interest rate and lagged real balances. - 47 - TABLE 3.2 ETMMTION RESULT EOR ThE 1'OM DEMAND (1976.1 - 1988.1V ) mt - P ' + OLy, + 02, + 03(it+6) + 04(mt.I-P,.1) + OsS, + r1TME OLS ARI NLS 00 -1.582 -1.959 (1.3) (1.5) , Q0.415 0.486 (3.5) (3.9) 02 -3.101 -3.595 (4.1) (4.6) 03 -0.165 -0.102 (0.9) (0.6) 904 0.702 0.660 (15.2) (12.6) t5 0.148 0.151 (3.5) (4-1) r 4-0.004 -0.004 (4.1) (4.0) cr 0.298 0.340 a1 1.390 1.430 1.394 a2 -10.410 -10.590 -10.414 a3 -0.550 -0.300 -0.555 A2 0.972 0.979 0.972 h 1.947 0.570 1.947 rho 0.285 Note: t-values are in parenthesis. h is Durbin's h and rho is the residual first-order correlation coefficient. The estimation methods were: Ordinary Least Squares (OLS). Maximum Likelihood for ARI residuals (ARI), and Nonlinear Least Squares (NLS). -48- TABLED3.3 ESTIMATION RESULTS FOR THE PUBLIC DEBT DEMAND (1983.1 - 1988.) d- - P, = 00+ Ot, + OA + 03(i,+6j + . 4(dt.%-pl.,)+ o4 + r2TIME OLS AR1 NLS 00 55.005 S1.932 (3.8) (3.7) 4, -4.249 -3.956 (3.3) (3.2) 102 19.024 19.175 (2.6) (2.7) 03 -2.816 -2.820 (2.1) (2.1) 04 0.305 0.281 (4.6) (4.2) 41$ T2 Q.081 -0.082 (4.5) (4.4) p 0.695 0.719 BS -6.115 -5.502 -6.117 B2 2730 26.668 27.381 B3 4.052 -3.922 -4.054 0.872 0.816 0.872 h 0.375 0.570 rho 0.285 NgW: see table 3.2. -49- TABILE 3.4 REDUCED-FORM ESIATON RfESULTS FOR THE INTEREST RATE AND THE PRIE3EEL (1983.11 - 1988.IV) (OLS) INTERFST PRICE RATE LEVEL CONSTANT -1.170 -8.810 (2.5) (1.3) y, 0.084 0.170 (1.9) (0.3) *(i,+61) 0.106 0.632 (4.8) (1.9) 0.055 -0.110 (2.8) (0.4) 0.00026 0.064 (0.03) (0.6) m, 0.014 0.514 (1.9) (2.0) ci, 0.004 0.182 (0.5) (1.6) S, 0.0039 -0.068 (0.5) (0.6) i Q0.757 0.977 DW 1.720 1.500 NVote: t-values in parenthesis. DW is the Durbin-Watson statistic. -50- 3.4 Simulation Resulta The results reported in Tables 3.2 and 3.3 were used to simulate the impact of inflation on seignorage revenue and the effects on the interest rate and the price level of a change in the fiscal deficit under alternative strategies of domestic financing. In the simulation exercises that follow we used the multipliers summarized in Table 3.1 and the estimated coefficients of the first columns of Tables 3.2 and 3.3. 3.4.1 Inflation and the Inflation Tax Figure 3.5 shows the steady state relationship between inflation rates and the inflation tax revenue derived from our estimated coefficients of the money demand equation and assuming no growth in the exogenous variables. The revenue-maximizing monthly inflation rate is 20%, which yields an inflation tax revenue of about 2% of GDP. Little tax gain is obtained from raising inflation. At an inflation rate of 2% per month the revenue is about 0.5% of GDP. Raising the inflatioui rate from 2% to 20% per month, inc:eases seignorage revenue by only 1.5% percentage points of GDP. 3.4.2 Effects of a 100% Increase in the Fiscal Deficit To simulate the effects of a one-period increase in the public sector deficit we have assumed that before the fiscal shock occurs the economy is at a steady-state equilibrium, with a 1% of GDP deficit and a 0.2 money/GDP ratio." Table 3.5 summarizes the magnitude of the short and long-run effects of a money-financed, transitory 100% rise in the fiscal deficit. The Sal shock results in a fall in the interest rate by 0.39 percentage points in the first period and a long run decline of about 0.27 percentage points. The effect of the fiscal shock on the price level is of about 7% both in the short run and long run. This " Recall that flow variables, such as GDP, are quarterly figures. - 51 a exercise suggests that the main effect of a change in the deficit financed by money creation is in the first period. In addition, since the increase in the price level never matches the increase in the stock of money required to finance the fiscal shock, this simulation suggests that money has real effects. The dynamic behavior of the interest rate and the price level after a money-financed 100% increase in the fiscal deficit are shown in Figures 3.6 and 3.7. Let's consider now the case of a 100% increase in the fiscal deficit financed by issuing public bonds. In this case, the fiscal shock requires only a 2% increase in the stock of public debt. The short and long-run effects are reported in Table 3.6. The results indicate that the increase in the fiscal deficit will induce a slight increase in the interest rate and the price level. Because the interest rate rises, the rise in inflation required to maintain the portfolio Nquilibrium is much lower than under the money-financed deficit. The dynamic behavior of the interest rate and the price level after a bond-financed fiscal shock are depicted in Figures 3.8 and 3.9 respectively. In summary, these results suggest that the main impact of a change in the fiscal deficit occurs in the first quarter. The only exception is the gradual rise in prices when the fiscal deficit is financed by domestic borrowing. In addition, the exercises show a small impact on interest rates and prices of a debt-financed deficit. This result reflects the specific features of the sample period of our estimates. In particular, the transfer payments to the private sector implied by the social security reform and the quasi-fiscal operations are the main factors behind the small impact of debt-financed fiscal deficits on interest rates and prices. -52 - TABLE 3.5 BEE3 OEM-ONEY-FINANCED INC-REASE IN kISCAL DEF!CIT SHORT RUN LONG RUN Change in Interest Rate -0.385 -0.265 (Percentage Points) Change in Price Level 7.326 7.245 (100 d log p) TABLE 3.6 EFFECTS OF PUBLIC DEBT-FINANCED INCREASE IN FISCAL DEFICIT SHORT RUN LONG RUN Change in Interest Rate 0.090 0.053 (Percentage Points) Change in Price Level 0.280 0.551 (100 d log p) FIGURE 3.5: SEIGNORAGE REVENUE 2.2 - 2 1.8 1.6 a. 0~0 1.4- 1.2 - 00 0.8 0 w 0.6- 0.4- 0.2 0.01 0.04 0.07 0.1 0.13 0.16 0.19 0.22 0.25 0.28 0.31 0.34 0.37 0.4 INFLATION RATE FIGURE 3.6: INTEREST RATE EFFECT OF A MONEY FINANCED DEFICIT 0.1 1 0 z -0.1 0 L&\ a:\ 0. -0.2 z ix -0.3- at. -0.4- 0 1 2 3 4 5 6 7 8 9 10 If TIME FIGURE 3.7: PRICE LEVEL EFFECT OF A MONEY FINANCED DEFICIT 0.1 0.090 0.08 - 0.07 - w i 0.06- w a: 0.05- a. U I&. II 0.04- 0 0.03 - 0.02 - 0.01 , 0 0 1 2 3 4 5 6 7 8 9 10 TIME FIGURE 3.8 : INTEREST RATE EFFECT OF A DEBT FINANCED DEFICIT 0.1- I 0.09 0.08- ZL 0.07- z 0 0.06- 0.05 - 0 10 -8 w a. 0.03 0.02- 0.01 0 1 2 3 4 5 6 7 8 9 10 I11 12 TIME FIGURE 3.9 PRICE LEVEL EFFECT OF A DEBT FINANCED DEFICIT 0.006 0.005 - ^ 0.004 - w -i U 2 0 0.003 a.~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~U 6. 0 0 4 0.002- 0.001 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 TIME - 58 - 4. PRIVATE CONSUMPTION AND PRIVATE INVESTMENT This section goes a step further in analyzing the macroeconomic implications of public sector deficits by assessing the impact of the public sector on private sector spending in Chile. Thereforc the focus is on the sensitivity of private consumption and investment to fiscal variables, in addition to indirect effects of them via interest rates, inflation or private disposable income. How private saving arid capital formation are affected by fiscal policies has significant implications for both short- term stabilization issues and long-run growth prospects. Figures 4.1 and 4.2 show the evolution of the private consumption to private disposable income ratio and the private investment to GDP ratio during the last three decades. Not surprisingly, private consumption has tended to be counter-cyclical, i.e. consumption does not decline as strongly as private disposable income during recessions, such as those of 1975 and 1982-1983. Two high consumption episodes coincided with the fiscal-monetary expansion of the early seventies and the foreign-financed private spending boom of 10 years later. A distinct structural change takes place after 1984: private consumption as a ratio of disposable income reaches the lowest values in three decades, and stays there throughout the 1985-1988 recovery - an important counterpart to the significant current account adjustment which took place in the 1980s. The private investment rate shows wild swings during the 1960-1988 period in Chile. After a historical decline to 5% of GDP during the early 1970s, it starts a continuous recovery, to boom in the early 1980s. Then private capital formation took a dive coincident with the 1982-83 recession and subsequent policy uncertainty. Investment remained low in 1984-1986 before picking up again in 1987 and 1988. It is estimated that the private investment rate could surpass 16% of GTOP in 1990. - 59 - 4.1 Private Conumpt This subsection, significantly based on a framework developed by Corbo and Schmidt-Hebbel (1991), addresses the effects of public policies on consumption in Chile. Direct effects of fiscal policies on consumption or saving operate through public saving (or the deficit) and its composition. If the stringent conditions required for Ricardian equivalence are satisfied, a rise in public saving, if it is done via lower public spending, is exactly offset by an increase in consumption. As disposable income does not change, the reduction in private saving matches the increase in public saving. A rise in public saving does no affect consumption at all if it is made possible via higher taxes. But as disposable income is reduced by the size of the tax, the reduction in private saving also matches the increase in public saving (the latter being macroeconomically equivalent to issuing more public debt). The opposite results are predicted by the Keynesian (and permanent income without Ricardian equivalence) hypotheses: current (permanent) taxes matter for consumption, not current (permanent) public spending levels. The Ricardian hypothesis has been widely rejected in empirical studies for industrialized countries. Most of these studies identify the existence of pervasive borrowing constraints as the main cause for its rejection.' A study for a set of developing countries by Montiel and Haque (1987) tests for two different causes which could explain a deviation from Ricardian equivalence: higher private than govemment discount rates (due to Blanchard-Yaari infinite-lived households facing a mortality probability) and liquidity constraints. They find significant evidence for the latter causing a deviation from Ricardian equivalence without much support for the former. 'For Ricardian equivalence to hold, the main conditions to be satisfied jointly are: absence of liquidity constraints, equal rates of discount for the public and private sectors, and certainty on future income, tax and public sector expenditure flows. For a further discussion of these conditions and surveys of empirical studies see Hayashi (1985), Hubbard and Judd (1986), Bemheim (1987), and Leiderman and Blejer (1989). FIGURE 4.1 Chile: Private Consumption to Private Disposable Income Ratio (C/YD) 1,.8 1 .04 0.96 0.84~~~~~~~~~~~~~~~~~~ 60 61 62L 64 64 566 67 B8- 69 70 71 7;Z 7. 74 7S 76 77 78 79 90 81 CS;Z 94 84 SO 96 Y7 ur Yr r FIGURE 4.2 Chile: Private Investment to GDP Ratio (O/Y) 0.17 . 0.1 6 0.10 0.14 0.1$ ')AZ 02.; 0.07 0.04 0.04 0.p0- i I 1 I II I I 1I I 1 1 1 1 1 1 1 i i II l l l 60 61 ZS 65 B465 66 67 B 697071 72Z73747$76777879 8081 82Z8$84iUS864768 Ycir -62- Borrowing constraints, proxied by current income or financial asset holdings, are also major determinants in the cross-developing country studies for private consumption and household saving by Rossi (1989) and Schmidt-Hebbel, Webb and Corsetti (1991), respectively. Indirect effects of public (fiscal-monetary) policies on consumption operate via the impact of public deficits and their financing on major prices affecting private consumption: real interest rates, inflation, and the real exchange rate. Real interest rates affect private consumption only if the well- known substitution, income and wealth effects do not cancel each other. This seems not to be the case, judging from the growing evidence which shows that consumption is not sensitive to real interest rates.' While inflation's first order effect is on the composition of the savings stock and not on saving or consumption flows, it may have second-order effects reducing saving if it is associated to capital flight or increasing (precautionary) saving if it raises uncertainty. An expected devaluation reduces the consumption-based real interest rate and leads to capital flight - the first effect could increase actual consumption while capital flight raises the measured consumption to income ratio if capital flight reduces measured income."4 Substitution effects from fiscal policies could arise when fiscal spending on privately appropriated services (education, health, nutrition) and direct transfers to households reduce the necd for private consumption expenditure on these categories, hence leading to possible declines in aggregate private consumption (see Easterly et al., 1989). '3Among the studies on the interest-insensitivity of saving in developing countries see, for instance, Giovannini (1985), Corbo and Schmidt-Hebbel (1991), and Schmidt-Hebbel, Webb and Corsetti (1991). For an alternate view, see Fry (1988) and Balassa (1989). "For the theory and the Latin American experience on the role of consumption-based real mterest rates on intertemporal consumption allocation see Dornbusch (1983, 1985), and for the relation between saving and capital flight see Dornbusch (1989). - 63 Monetary policy and financial reform affect tbe aggregate of mor:tmry and quasi -n.oetary assets. Higher holdings of broad money tend to increase consumption foi two reasons: first they are a major component of private financial wealth and second they tend to reduce the incidence of bfrrowing constraints faced by at least a subgroup of consumers. Another form of liquidity constraints are those imposed by foreign lenders. Particularly after 1982 foreign saving tends to reflect the foreign resource constraint imposed by foreign lenders and not the excess of income over unconstrained domestic absorption. Let's start the discussion of a relevant framework for assess;ng the impact of public policies on private consumption by referring to a standard linear form for testing the Ricardian equivalence proposition, based on Bernheim's (1987) survey: (4.1) Cp, = ao + a, (YN, - T, + i, D,) + a2 (T, - EG, - i, D,) + + a4 D, + as W, + va where Cp is private consumption, YN is national income, T is tax revenue net of transfers and subsidies to the private sector, i is the nominal interest rate, D is government (domestic) debt, EG is public spending, W is private wealth, and v2 is a stochastic error term. Hence the first right-hand determinant is private disposable income and the second is the nominal or total government surplus. Three simple null hypotheses can be tested with this specification: (i) Keynesian hypothesis: ao > 0, a1 > 0, other coefficients 0; (ii) Permanent income hypothesis without Ricardian equivalence: a4 > 0, a5 > 0, other coefficients 0; (iii) Ricardian equivalence: a, = a2 > 0, other coefficients equal to 0. -64- However, the specification in equation 4.1 presents various shortcomings.,, Frst, public saving and not the public surplus should be the relevant determinant in 4.2, because public investment adds to real capital and therefore constitutes 'net wealth' (Barro, 1974). Under inflation, only the real component of domestic interest payments should enter both the private disposable income and the government surplus (saving)'. Tbird, permanent, not current private disposable income and public surplus (saving) should enter equation (4.1) for a fair test of the Ricardian proposition. Finally, additional potential consumption determinants, such as foreign payments, the real interest rate or inflation, should be included. The following specification for the private consumption to private disposable income ratio takes care of the above mentioned shortcomings. In addition, the scaling of non-stationary variables to private disposable income reduces the incidence of non-stationarity problems: '42' C PDY P 4.2) DY =Po + P1 DYp + P2 Ps | P3 rc + P4g a + 155p + '' DYptDp DYppe +a m + P, He + P FS V - DY +t 7DYw DYz where DYP is current private disposable income, PDYp is permanent private disposable income, PS, is permanent public saving, r. is the consumption-based real interest rate, x, is consumption inflation, P. and P. are prices for imported and national private consumption goods, respectively, CPTR is 15Bernheim (1987) addresses the third and fourth limitations discussed below. "Tbe inflation component of interest payments on the public debt, which compeisates for the loss in principal due to inflation, is put back into public bond holdings by private investors aiming at maintaining constant real holdings of public debL -65 - the sum of public expenditure on privately appropriated services and direct transfers to consumers, H is base money, FS is foreign saving, and v is a stochastic error term. Again, B, 01, and B2 are associated to the corresponding Keynesian, permanent-income and Ricardian hypotheses. Note, however, that I2 could be positive due to a financial market structure giving preferential access to bank credit to the public sector and hence allocating the residual to private consumption (and investment). This direct form of crowding out of private saving by public sector saving behavior has obviously nothing to do with Ricardian unconstrained, forward-looking private consumers internalizing the govemment's intertemporal budget constraint. Expected signs of the coefficients are the following: 130,0, a2, a%, Ba > 0; B6 < 0; B3, 84, 55 < > 0. Permanent private disposable income and permanent public saving are consistent with the following definitions for their corresponding current values: (4.3) DYp E GDP, - NFp, - T, + r, D, (4.4) SGt T, - CG, - NFPG, - r, D, where GDP is gross domestic product, NFPp is net foreign payments made by the private sector, SG is current public saving, CO is public consumption, and NFPG is net foreign payments made by the public sector. Note that D refers now only to the domestic public debt. For the 'permanent' values of any variable (private disposable income and public saving in this section, and other variables in the investment section below) we specify two alternatives. The first is defining expectations of the permanent values consistent with partial perfect foresight, defined - 66- as the simple average of the contemporaneous variable and two periods into the future, for any variable x: (4.5a) Px, = [x, + x,. + x1*2j/3 where Px defines the expected permanent value of variable x The second alternative is the simple static expectations specification which allocates a 100% weight to the contemporaneous value in (4.5a), as follows: (4.5b) Px, = x, Similar assumptions are made with respect to expected consumption inflation (and expected investment inflation in section 4.2 below). A fust alternative takes actual inflation between today and tomorrow as the relevant proxy for rationally expected inflation. The second alternative is static, specifying the expected price change to be equal to the actual price change between yesterday and today. Equation 4.2 was tested using annual data covering the 1960-1988 period, which is the sample period used in most of sections 4 and 5 of this study. This is a relatively long period in Chilean recent economic history, characterized by significant structural breaks and policy regime changes, which due to their intensity introduces a note of caution when interpreting the statistical results presented below. Table 4.1 presents a selected number of empirical results for the consumption function, separately for the static ai,J partial perfect foresight (PPF) versions for all expectations. Lines 1.1 and 2.1 of the table present results for the complete specification of equation 4.2. In general, the -67. results of the static expectations version dominate statistically those related to the partial perfect foresight case.17 The average Chilean consumer appears to be partly Keynesian, partly PIH forward looking: under the static version, her marginal propensity to consume out of current private disposable income is approximately 0.70 and her marginal propensitv to consume out of permanent income is around 0.30. Under the PPF version the corresponding propensities are of a similar magnitude, close to 0.50. Interestingly, she does not react to permanent public saving; i.e., the consolidated public sector deficit (abstracting from public investment) has no effect on private consumption. The policy implication is clear: an increase in public sector saving has a strong effect on national saving, not neutralized by a decline in private saving. Of all relevant prices, only the real exchange rate (given by the relative price of imported to national consumption goods components of aggregate consumption), affects private consumption. A 10% real devaluation, which would affect directly this relative price, would reduce aggregate consumption by approximately a 1 percentage point of private disposable income. However, neithcr the real interest rate nor the inflation rate have any consistently significant, separate effects on private consumption, corroborating previous studies on the interest insensitivity of consumption in Chile (Schmidt-Hebbel, 1987, 1988, Arrau, 1989). In addition, there is no evidence of a substitution effect of public spending on privately appropriate services (education, health) and transfers on aggregate consumption. Monetary holdings and foreign saving flows exert significant positive influences on private consumption. While higher base money holdings tend to relax domestic borrowing constraints faced by consumers, higher foreign saving is associated to weaker foreign resource constraints. The effect "7Ihis results mimics the quarterly structural consumption function results for Chile in Sclimidt-Hebbel (1988), which are stronger for the backward than for the forward-looking expectaw1ons specifications. - 68 - of the latter variable on private consumption is very strong: from the results one can infer that between 42% and 7T % of the dramatic current account correction which took place after 1981 was borne by private consumers. The policy implications of our results are quite clear. In the first place, there is no evidence of a direct effect of the public deficit on private consumption; consumers are neither Ricardian nor directly crowded out by public spending in Chile. A transitoly deficit reduction financed by a transitory tax hike reduces private disposable income, and hence private consumption and saving. If the tax hike is permanent, private consumption will take the brunt of the reduction in (current and permanent) private disposable income, without private saving being significantly affected. On the other side, there are no indirect effects of domestic financing via changes in the real interest and inflation rates on consumption. However, if inflation reduces real base money, private consumption would be affected. Foreign adjustments have strong effects on private consumption in Chile. The combination of lower foreign saving and a higher real exchange rate has a strong negative effect on private consumption expenditure, as evidenced during the post-1982 adjustment period. 4.2 Private Investment Following Easterly et al. (1989) and empirical investment studies for Chile by Solimano (1990) and Morocco by Schmidt-Hi o)bel and Mueller (1990), we specify a behavioral function for private investment which will depend on neoclassical profit and cost variables, liquidity constraints, and risk determinants. To avoid again spurious correlation, we scale all non-stationary variables to GDP. Therefore we specify the following generic equation for the private investment to GDP ratio: TABLE 4.1 PRIVATE CONSUMPTION (CHILE. 1960-1988) Dependent Variable Private Consumption to Private Disposable Income Ratio (CplDYp) Equation c PD Yp ff re Pcm Cp7 N FS Rh RFA DW DYr DY, Pcn DY, DY, DYr 1. Static Ez,ectations 1.1 OLS 0.71 0.28 0.18 0.05 -0.02 -0.14 0.12 0.29 0.24 - 0.79 1.77 (10.4) (4.1) (0.8) (1.7) (-O5) (-3.0) (0.4) (3.2) (1.7) 1.2 OLS 0.74 0.25 0.36 -0.13 0.23 037 - 0.72 1.52 (9.8) (3.2) (1.6) (-5.2) (26) (2.5) I12A ML 0.66 0.26 029 - 40.08 0.45 0.41 0.59 0.90 1.75 (10.0) (3.7) (1.2) (-2.2) (3.8) (2-8) (35) 13 OLS 0.72 0.27 - - 40.13 0.24 0.48 - 0.70 138 (9.4) (3.5) (-5.0) (2.6) 3.6) 13A ML 0.67 0.27 - - - -0.09 - 0.44 0.42 0.57 0.89 1.50 (10.0) (3.9) (-24) (3.7) (2.8) (13.4) 2 Partial Perfec Foresight 2.1 OLS 0.49 0.49 0.002 0.0004 0.08 -0.19 0.04 0.28 0.87 0.63 132 (2.6) (2.8) (0.005) (0.0008) (1.2) (-3.6) (0.09) (1.5) (4.0) 2.IA ML 0.18 0.50 0.38 .0.03 0.03 .0.11 0.24 0.52 0.68 0.67 0.89 1.39 (2.9) (3.6) (0.8) (40.6) (0.4) (-:LI) (0.5) (3.0) (4.1) (4.1) 2.2A ML 0.42 0.48 0.14 - - 0.10 0.58 0.71 0.62 0.89 1AI (3.4) (3.7) (0.4) (-2.4) (4.8) (4.6) (3.9) 2.3A ML OA.3 0.48 - 40.10 0.57 0.73 0.60 0.90 .43 (3.5) .3.80 (-2.6) (4.9) (4.9) (3.7) Note Rho is the rnz-order residual correlation cefiient. - 70 - 6) Ip I (PUCK1, P(&- K K=, , PRO, FC, (4.6 yt=it - yP0p 7y I * 1 y H_ (- (? H- (+) 0+ ( H Hel- FSt VUCK, VY (+) (+) ( ) ( where Ip is private fLxed-capital investment, Y is GDP, UCK is the user cost of capital and PUCK is the estimated permanent UCK, K.J/Y is the private sector capital output ratio and PK.,/Y is its permanent estimate, P,FM,, is the price ratio of imported and national private investment components, COT is corporate tax revenue and PCOT is its permanent estimate, KB is the public sector capital stock, PRO is corporate profits, FC is banking credit flows to firms, H is base money, FS is foreign saving, VUCK is the coefficient of variation of UCK, and VY is the coefficient of variation of GDP. Expec.d signs of partial derivatives are denoted below each variable in equation (4.6). The current real user cost of capital is defined as: (4.7) UCK, = (P1,/Ps) 1(iF; - P 1)( I +P 1e) + 6] where P, is the private investment deflator, iF is the nominal interest rate on banking loans to firms, P^el, is the expected rate of change of the private investment deflator, and 6 is the (real) capital depreciation rate. The private sector capital output ratio, the inverse of the average product of capital, stands for both the neoclassical marginal product of capital (which is a linear transformation of the marginal - 71 X product under a Cobb-Douglas technology) and for the Keynesian potential to actual output ratio. Note that private and public capital add up to obtain the total domestic capital stock: (4.8) K, = K c + K , Expected investment inflation is based on an estimated AR structure. All expected permanent variables are specified according to the two hypotheses mentioned in section 4.1: the partial perfect foresight alternative of equation (4.5a) and the static version of equation (4.5b). Finally, the two coefficients of variation are defined as five-period moving variances, two periods back, the current period, and two into the future. The empirical results for the private investment functions are presented in tables 4.2A and 4.2B, also for the static and PPF expectation alternatives for all relevant right-hand variables. The final results reported in the tables exclude three variables appearing in equation (4.6): the relative price of the two aggregate investment components, corporate taxes, and firm credit. The exclusion of the first variable from preliminary results was decided on its implausibly high coefficient, which affected many other parameters. The latter two variables were not included due to lack of data. The permanent user cost of capital (PUCK) appears to be positive in all and significant in most equations reported in table 4.2A. This seems to be a reflection of the extremely strong structural breaks occurred during the 1960-1988 period with regard to the functioning of Chilean financial market, the determination and the levels of the real interest rate, and hence the dependence of investment on the cost of domestic financial capital. TIwo attempts were done to face this problem. First, multiplicative dummies for the PUCK were specified, separately for the sixties (D60s) and early seventies (D70s) in table 4.2A. Second, the original specification was tested for the 1976-1988 period, which starts after the 1974-75 domestic financial market and interest liberalization took place. The - 72 - corresponding results for the latter sub-period are shown in table 4.2B. However, the brevity of this more homogeneous period suggests to exercise caution in interpreting the results of this table due to the smali number of degrees of freedom. The coefficients of the period-specific dummies in table 4.2A tend to present the correct negative s.gn, and are significant in some of the reported results. Take for instance line 2.2 While the coeffirent of PUCK is still positive for the 1974-1988 period (0.16), it is close to zero for the early 1970s (0.16-0.15) and it is negative for the sixties (-0.03 = 0.15 - 0.18). However, for the static expectations alternative and the 1976-88 period, the user cost of capital is negative and significant although of small magnitude in the best results presented in table 4.2B. The lagged capital to current output ratio appears to be negative and highly significant in most reported results. This is not a surprising result: private investment is strongly procyclical (see also Solimano, 1990). More surprise causes the consistently negative influence of the public sector capital stock on private investment. Due to the fact that we only have a theory for a positive effect of this variable (crowding-in of private investment due to positive externalities of higher public past investments in infrastructure and other public goods), we excluded this variable from the final results. Among liquidity constraint variables, their influence depends strongly on the chosen period. Tlhis is absolutely consistent with the change in the role of interest rates and hence of the user cost of capital during the 1960-1988 period: while quantity constraints were of higher importance before domestic financial liberalization, their influence was weakened afterwards. Precisely this is the case of own firm profits and foreign saving, which have a strong influence on investment during the 1960- 88 period as a whole, while disappearing from the scene during the 1976-1988 period. On the other side, base money plays an ambiguous and unstable role: while in most sub-periods and specification TABLE A PRIVATE INVESrMENr LCHL2. 1961-198) Dpedet Varble Prite wtmail to GWP RatIo IpY) elao c D6f D7% PU 4PMi si 7V VY Rbo MW OW t. Sic eeatlm tiOLS 0.22 do6 410 Q.12 405 411 Q.17 411 407 401 4101 . U2 129 (33) (08) (L5) (23) (-3) (t2.2) (2.0) (.1.4) (45) (-Ls) (41) U2ML(ARI) l17 403 Q 05 01 406 407 Q 26 411 0.0 4002 427 .7 05 L93 (27 (45 (01) (LI) (-8) (-L6) (28) (-L.) (A4) (47) - (59) 13ML (ARI) at1 -dolS QS Q01 407 * 25 0-13 .426 la a5t L72 (L0) (.2) (6(9) (2.$) (04 (.2) (62) 2 Pul Pafee FPadra 1 OIS Q21 4a27 409 0.06 0.04 418 0.03 Q.22 411 482 43 - 01 L19 (14) (4) ( ( (08) (11) (03) 9) (49) (4) (47) 22 ML(ARI) 0.1s 421 415 Q16 402 412 0.16 4Q0 405 4-00 48 Oa 0.67 Li (Z7) (O (-7) (&I) (42) (-2.0 (L) (44) (40) (.1.) (4 (12 22 ML(ARI) 020 409 4001 06 412 1 77 - 446 OA 037 L.6 (5) (L) (4M (2. (23 (Ls) (. (42) (53) TABLE 4.2B PRIVATE INVESTMENT (CHILE. 1976-1988) Dependent Variabe- Private lnvestment to GDP Ratio (Ip/Y) Equation C PUCK dr!R PtRO H- VUCK VY Rho R2A DW 1.1 Static Expectations 1.1 OLS 035 0.05 4.09 -0.14 0.02 0.51 4.02 0.003 -0.26 - 4a8 3.69 (4.1) (0.7) (-1.8) (-2.2) (0.14) (1.7) (4.1) (0.5) (-1.0) I )LS 0.36 40.06 40.17 - - 0.26 - - 056 - 0.82 2.95 (05) (-2.0) (-7.3) (1.1) (.2.3) 13 ML(ARI) 0.37 -0.07 -0.18 - 0.36 - - 4.65 .51 0.95 219 (16.9) (-3.1) (-11.7) (1.9) (-3-3) (-1-9) 1.2 Partial Perfect Foresizht 1.1 0.52 0.16 -0.19 -0.07 40.24 -0.32 40.05 0.001 *0.27 - 0.86 291 (Z6) (1.3) -1.0) (-0.5) (-1.2) (4.5) (4.02) (0.1) (-05) 1.3 ML (ARI) 0.45 0.10 40.24 -0.49 - -0.55 -0.37 0.95 2.09 (13.) (3.2) (-11.1) (-Z5) (-2.9) (-1.1) 75 - alternatives it exerts a negative, non-significant role, it presents a positive and significant, although minor influence only in line 1.3 (static expectations) for the 1976-88 period (table 4.2B). With, regard to the risk variables (the coefficients of variation of thc user cost and of GDP), the first one plays a negative although rarely significant role during the complete 1976-88 period, which disappears during the second sub-period. Much more important is the variability of GDP, which affects private investment consistently and negatively during the entire period, and particularly strongly during the more recent years. From the empirical results we can draw the following policy implications. The financing of the public deficit via debt issuance is increasingly felt by private investors in Chile due to its upward pressure on domestic real interest rates (see section 4), which affects negatively private capital formation since the mid-1970s. The other side of the coin is the weakening role of both domestic and foreign liquidity constraints: while firm profits and foreign saving where important investment determinants before financial liberalization, they vanished afterward. No evidence was found on public investment crowding in effects on private investment. The relevant variable, which is the public capital stock, does not affect positively private capital formation. The foreign adjustment has affected private investment negatively via the relative price of inivestment goods, which has increased due to the real devaluations, hence increasing the user cost of capital. Finally, investment is negatively affected by the variability of GDP. Therefore a more stable external environment and stable domestic macroeconomic policies exert a positive influence on private capital formation. -76- S. RELATIVE EXTERNAL PRICES AND THE TRADE BALANCE his section focuses on the impact of fiscal policies on the external sector, especially with regard to the trade deficit and the real exchange rate. 'he framework applied here is derived in two steps. The first involves determining the relative prices of exportables and importables from the equilibrium condition in the non-traded goods market, and the second implies deriving an equation for the trade balance. Public spending, deficits and debt stocks are important codeterminants of the behavior of these external sector variables. Based on an appropriate framework, this section assesses the impact of fiscal variables on Chile's external sector based on empirical findings related to the 1960-1988 period. Figures 5.1 and 5.2 depict the evolution of relative export and import prices, and the trade surplus to GDP ratio, respectively, during the last three decades in Chile. Tbe relative price of exports to non-tradeables is strongly correlated with the international copper price, which is the main determinant of the country's extemal export prices and terms of trade. The price of imports to non-tradeables is a more useful measure of the real exchange rate relevant for spending and production (other than copper) decisions. However, both measures of the real exchange rate tend to move together during 1975-1988, a period of smaller terms of trade fluctuations than those experienced during 1960-1975. The two periods of massive real exchange rate appreciations - the early 1970s and the early 19BOs - coincide with the policy-induced spending frenzy of the Allende government and the foreign- financed private spending spree of Pinochet's 'plata dulce. Trade deficits, depicted in figure 5.2, reach their highest levels during those years - the most dramatic observation being the 109%o (of GDP) trade deficit observed in 1981. FIGURE 5.1 Chile: Relative Export and Import Prices (Px/Pn, P'n/Pn) 2- 1 S 1.4 1.1 Q.4 ('.7 N. 60 61 62 63 64 i5 $;6 67 S8S 6 70 71 72 7 74 75 76 77 7T 79 8 9Q61 IZ2 94 4 I6I Y 6 85 YPmgr El PxfFPn + fmlFn FIGURE 5.2 Chile: Trade Surplus to GDP Ratio (TS/Y) 0.1 - -0.06 Q*Q4 ,. 69 61 62 6x 64 65 66 67 6 3 6X 70 71 72 7 74 75 76i 77 79 79 a-0 01 42 63 64 4 86 87 98 Yoa r - 79- The required current account corrections after 1973 and 1981, respectively, led to fast reversals of the trade deficits as a result of massive exchange rate depreciations and reductions in domestic spending. However, a distinct feature of the post-1984 extemal adjustment is the achievement of massive trade surpluses, exceeding 6% of GDP in 1988. These are the counterpart of the significant decline in private consumption as a share of income (discussed in Section 4) and a progressive correction in public sector deficits (discussed in section 2). 5.1 Relative Prices of Exoorts and Imports Following Easterly et al. (1989) and Rodriguez (1989), the real exchange rate, defined as the relative tradable/non-tradable goods price, is derived from the continuous market-clearing condition for non-tradable goods in the Salter-Swan-Corden-Dombusch small open economy tradition. Extending this paradigm to a three-sector distinction between expoftables, importables, and non- tradeables, the market clearing condition can be rewritten as either one of the following functions for the relative prices of exportables and importables:8 '"For the same reason discussed in the preceding sect;on (to avoid spurious correlation) we scale non-stationary variables (such as the trade surplus, go-ernment expenditure, or net foreign assets) co appropriate scale variables in equations 5.1 - 5.3. -80- (5.1) CX , = ex (77T,tst, Ut9 YU' Mt (+) (?) (+) (? (- -P 7T' ty' y'GM (5.2) emCa-Al = em ( M[TT tMt _ _ G Nt pN1 yt I 7 " ) {_) H (+M+ ? (-) wheie ex and em are the relative prices of exportables and importables, respectively; Px P.M and P, are the absolute prices of exportables, importables, and non-tradables; TP is the foreign terms of trade, tM is the ad valorem tariff rate for imports, iS is the current-price trade surplus; G is current- price (general) govemment spending (public consumption plus public investment); GN is current-price (geueral) govemment spending on nontradeable goods; and Y is current-price GDP. The above framework was app!ied to Chile using annual data for the 1960-88 period. The results are reported in table 5.1 for both equations. They tend to favor the relative export to non- traded gor ;s price interpretation as the relevant reduced-form equation for the equilibrium in the non-tradable good& market. The extemal terms of tmde are an important determinant of both export and import prices. although their numerical influence is much stronger on the former. The average tariff rate TABLE 5.1 RF.LATIVE EXPORT AND IMPORT PRICFS (CHtLE. 190-19fl) Eqti C T tm TS G D75 Rho R2A DW r ~y 1. ,athe rau (a) 1.1 OLS -021 0.63 4.04 248 .3.64 0.51 - 06 129 (-1.2) (8.5) (-0.8) (4.0) (3.0) (4.5) 1.2 ML(ARI) 0.01 0.59 .0.05 2.22 2.50 0.38 052 0.77 1.76 (0.04) (6.9) (.0.9) (3.4) (1.7) (4.0) (3.0) CD 13 ML (ARI) -0.02 0.56 -2.9 2.80 037 0.51 0.77 1.74 (.0.0i) (7.7) (3.8) (7.0) (4.0) (3.0) 2 Reative Imo lPrIe (CM) 2.1 OIS 120 ^0.15 ^0.09 .10 4.62 0.61 - 0.7S 1.03 (6.7) (-2.0) (-1.9) (3.4) (O.5) (5.4) 2.2 ML(ARI) 1.26 -0.14 *0.09 2.00 -109 0.47 0.62 0.72 1.65 (5A) (-1.8) (-1.7) (3.3) (4.8) (5.7) (3.9) 2.3 ML (ARI) 1.09 -0.13 .0.09 2.04 0.47 0.60 0.72 1.62 (10.7) (-1.7) (-1.6) (3.5) (5.7) (3.6) 24 ML (ARI) 1.08 O.18 2.26 - 0.46 0.66 0.70 158 (9.8) (-2.5) (.8) (5.5) (4.6) -82- has a negative (although not significant) influence on both relative prices, which is surprising in the case of the import price, whici should be positively affected bv tariffs. The trade surplus to oLput ratio has an extremely high, consistent and symmetric effect on both relative prices: a one-percent of GDP increase in the trade surplus implies a 2.0 - 2.5% devaluation of the real exchange rate. Aggregate government spending has a strongly positive impact on the relative export price and a non-significant negative influence on the relative price of imports - in general, its influence could be of either sign. Unfortunately, it was not possible to include as an additional determinant the non-traded component of public spending due to lack of ata. Finally, 0.38-0.61 of the massive 1975 devaluation could not be explained by any of the preceding variables, and hence was treated as an outlier. 5.2 The Trade Balance The trade balance is both the difference between output and absorption and the goods markets counterpart to the accumulation of net foreign assets. Depending on which is the underlying paradigm, the trade surplus is specified as reflecting optimal output and absorption decisions or optimal foreign asset accumulation decisions. In Rodriguez (1989) and Easterly et al (1989), the trade surplus is directly related to the accumulation of net foreig- assets (NFA). Private net foreign asset accumulation depends on the difference between desired and actual (that is, actual lagged) private NFA holdings, the former substituted by its main determinants: the covered interest differential between domestic and foreign rates, domestic public debt, the terms of trade, aud income. Public NFA accumulation will reflect directly, with a negative sign, the (cperational) public sector deficit, for given stocks of domestic - 83- public debt and base money. Hence under this foreign asset accumulation version, the generic equation for the trade balance can be written as follows:1' (5.3a) zst 7Y( C i-(i_*+t'+ite) NFA,, BD, TZ * 7Y, YS C l+i+'i ) Y. D' ' t ( ) H ( )() ( ) where Y is current-price GDP, C is a constant, i is the average annual domestic nominal interest rate (average of active and passive rates), il is the average annual nominal external interest rate paid on net foreign assets, E" is the expected rate of nominal devaluation (defined below), B., is current- price domestic public sector debt (at the end of the preceding period), NFA.1 is current-price net foreign assets (total foreign debt less international reserves at the end of the preceding period), and OD, is the current-price operational public sector deficit. Expected signs of partial derivatives are denoted in parenthesis below each variable in equation (5.3a). An alternative -,view of the trade balance is to derive it from the macroeconomic equilibrium condition, and hence reliecting production, consumption and investment decisions which are behind the determinat;on of the excess of output over absorption. This view links more explicitly the determination of the trade surplus in this section with that of private consumption and investment in the preceding section. However, in order to maintain a structure as close as possible to that postulated by th1e asset accumulation version above, we focus now mostly on those variables deternining production and absorption which also appear in equation (5.3a). '9Current-price GDP is used as the relevant scale variable. Hence the positive sign of the constant term C reflects the hypothesized positive effect of income on the trade balance, when multiplying equation (3a) by the latter, and abstracting from the presence of multiplicative terms involving income and all non-scaled right-hand variables in eq. 5.3a. - 84- The trade surplus in current prices, defined by the excess of exports over imports, equals the difference between current-price output and total absorption (A): (5.4) 1S = Px - P,m = Y- A = Y + [r B1 + i*NFA_X - 71 - OD, - Ap where r is the domestic real interest rate paid on government bonds, T is total taxes net of transfers paid by the private sector, A, is private absorption (private consumption and investment expenditure), As is public sector absorption, and OD,, is the operational public sector deficit defined as: (5.5) OD$ = As + rB, + i NFAg, - T Substituting functional forms for output and absorption into (5.4) obtain: (5.4) 7 = Y(i, r ,...) + [rB_X + iLNFAg_l - 71 - OD# - (-) (- -A,(NFAp, B ,NFAg, i , r 7',,YP) (+) (+.) (0,+) (?,-) (?-) H+ (+) where the negative signs of the interest rates in output supply denote the effect of the cost of working capital on current production decisions (the Cavallo effect), and the two signs below dom,tic public debt and public NFA in private absorption reflect the non-Ricardian (the first sign) and the Ricardian (the second sign' hypotheses on the role of these assets (are they wealth?) in private consumption, while the two signs below the two interest rates in private absorption reflect their ambiguous effect on private consumption (the first) and their unambiguously negative effect on - 85 - private investment (the second). Finally, current terms of trade and permanent income affect positively Ap via private consumption. Rewriting eq. (5.4') as a functional expression for the trade surplus after scaling all relevant variables to currezt-price GDP, obtain the following form, as an alternative to equation (5.3a): (5.3b) 7 _ C i-(i'+&+i*+ ') NFAt, BD, ye Y, Yo ' +(i+ +i-&)' Y ' Y' (+) (- (?)) 7TT ODt YP .yt I y ) Two differences arise between the asset accumulation version 5.3a and the macroeconomic equilibrium version 5.3b of the trade surplus: first, the effect of domestic public debt, net foreign assets, and the interest rate differential are ambiguous in eq. 5.3b, and second, the ratio of permanent to current output (an inverse measure of the business cycle) appear in the latter equation. The subsequent empirical application will test the relative relevance of these different approaches to the trade surplus for the Chilean economv. The empirical results for the Chilean trade balance are shown in table 5.2. Lines 1.1 and 2.1 present results for the complete specification. Of the right-hand determinants, the interest rate differential (IRD), the total net foreign asset holdings, and the stock of domestic public debt holdings present positive, mostly significant signs under both versions for expectations. To capture a possible break of the relation between the trade surplus and the stock of net foreign asset holdings in 1981, the multiplicative dummy D81 for NFA was included for 1981. The positive signs of the former three variables contradict the asset accumulation version of the trade -86, balance in eq. (5.3a), but are consistent with the ambiguous signs postulated by the output less absorption version of equation 5.3b. Tle current extemal terms of trade have a consistently significant but low positive influence on the trade balance. However, the influence of the business cycle is not significant. A dummy for 1979-1982 (D7982) signals the particular regime of high access to foreign crAdit which characterized those years as opposed to all other subperiods. It reflects that the trade deficit was 2-3 percentage points of GDP higher during those years of easy access to foreign lending, unhindered by either domestic or foreign limits imposed to foreign borrowing. Finally, the operational public sector deficit has a negative though not significant effect on the trade surplus under the static expectations complete specification for the dependent variable (line 1.1). To test for the influence of this variable when most others are omitted, we run the specification of line 1.2. Having in mind the variable exclusion bias, we may conclude that a I percent of GDP increase in the public deficit tends to reduce the Chilean trade surplus by a maximum of 0.29 percentage points of GDP. TABLE 5.2 TRADE SURPLUS (CHILE. 19(&9]8 Dependent Variable Trade Surplus to GDP (lS/Y) Equation C IRD D81 FA_l B- 1P OD, YP D7982 Rho IPA DW -I. y 7 7 Staf- Expectations 1.1 OLS 0.04 0.01 -1.76 050 0.04 0.03 40.12 0.0.03 40.03 - 0.68 1.84 (4.3) (1.5) (-31) (1.9) (1.0) (2.0) (.1.0) (-1.0) (-2.1) CD 1.1 OLS 0.006 40.29 40.06 - 0.39 1.84 (.0) (4.1) (-2.6) Paurial Pereet Foresight 21 OLS .0.08 0.03 -1.61 0.50 0.07 0.04 0.03 .0.02 -0.02 - 0.73 1.97 (.05) (1.9) (-2.6) (2.4) (2.0) (3.3) (0.2) (4.1) (-1.4) - 88- 6. CONCLUSIONS This section summari7es the main findings and lessons that emerge from the study of causes and consequences of budget sector deficits in Chile. The first, unavoidable observation is about the wild gyrations of public sector deficits during the last two decades in Chile. In fact, each decade comprises a complete cycle of fiscal and macroeconomic crisis, recovery and consolidation. Tle 1970s started with consolidated public sector deficits of 23.4% in 1970-73, which in conjunction with adverse foreign stocks forced a major stabilization effort, which was very successful on the fiscal side, culminating in a 5A% surplus in 1980. In the early 1980s the conjunction of negative external developments and misguided domestic policies - this time, however, not on the fiscal side - caused a major macroeconomic crisis, which deteriorated public finances of both the financial and non-financial public sector. Quasi-fiscal rescue operations of domestic debtors and the private financial system by the Central Bank and emergency programs combined with a recession-induced fall of revenue in the case of the non-financial public sector were behind the large deficits of both public sector, running at 4.9% of GDP in 1986 for the former and at 4.6% of GDP in 1984 for the latter. Public sector nationaiization and reform are major causes of the variation in the non-financial public sector deficit: the tax reforms and public sector nationalization/privatization of the mid-1970s increased revenue, while the 1981 social security reform and the tax reductions in the mid-to-late 1980s destabilized public finances. The strong swings in the copper prices were traditionally the most destabilizing factors outside the control of policy makers, until the copper stabilization fund in.;oduced a stronger separation between public spending and copper revenue. An implication of the high quasi-fiscal deficits was the huge build-up of total public debt, which rose from $7.9 billion in 1981 to more than $22 billion. Nonetheless, our calculations show - 89 - that the solvency of the total consolidated public sector is not jeopardized by current deficit levels: a 1.5% deficit level - caused entirely by the Central Bank - is sustainable under normal macroeconomic conditions. Section 3 of the paper explored the implications of domestic financing of public deficits for the interest rate and the price level in Chile. The effects of alternative strategies of domestic financing were discussed and contrasted in the context of a portfolio model with a partial adjustment structure for both the money and the public bonds markets. The simulations based on the estimated portfolio model showed relatively conventional results: money-financed public deficits reduce interest rates and are strongly inflationary, while debt-finaiiced deficits raise interest rates and are only weakly inflationary. Less conventional was the result that the moderate fluctuations of interest rates and price levels during the 1980s were the result of quasi-fiscal operations and a social security reform that implied a net transfer of resources to the private sector, which in turn provoked an increase in the demand for public debt. This nrivate sector response is caused by institutional regulations of the social security funds and the banking system. Therefore, one should not generalize the experience with fiscal deficits in Chile during the eighties. Section 4 of this paper went a step further in analyzing the macroeconomic implications of public sector deficits by assessing the impact of the public sector on private sector spending in Chile. Here the focus was on the sensitivity of private consumption and investment to fiscal variables, in addition to indirect effects of fiscal policies via interest rates, inflation, or private disposable income. Clear policy implications can be drawn from our empirical results on private consumption. In the first place, there is no evidence of a direct effect of the public deficit on private consumption; consumers are neither Ricardian nor directly crowded out by public spending in Chile. A transitory deficit reduction financed by a transitory tax hike reduces private disposable income, and hence private consumption and saving. If the tax hike is permanent, private consumption will take the brunt of the -90- reduction in (current and permanent) private disposable income, without private saving being significantly affected. On the other side, there are no indirect effects of domestic deficit financing via changes in the real interest and inflation rates on consumption. However, if inflation reduces real base money, private consumption will be affected. Foreign adjustments have strong effects on private consumption in Chile. The combination of lower foreign saving and a higher real exchange rate has a strong negative impact on private consumption expenditure. This was clearly observed during the post-1982 adjustment period, when private consumption fell to historical lows, while private investment reached historical heights. From the empirical results for private investment behavior in Chile the following policy implications can be drawn. The financing of the public deficit via debt issuance is increasingly felt by private investors in Chile due to its upward pressure on domestic real interest rates, which affects negatively private capital formation since the mid seventies. The other side of the coin is the weakened role of both domestic and foreign liquidity constraints: while firm profits and foreign saving were important investment determinants before financial lberalization, they are not more significant investment determinants since the mid seventies. No evidence was found on public investment crowding in effects on private investment. The relevant variable, which is the public capital stock, does not affect positively private capital formation. The foreign adjustment has affected private investment negatively via the relative price of investment goods, which has increased due to the real devaluations, hence increasing the user cost of capital. However, the recent external adjustment period coincided with a strengthening of the development strategy and stable policy rules, reducing significantly policy uncertainty and actual output fluctuations. This reduction of systemic risk and of uncertainty of main macroeconomic variables affecting frequently irreversible investment decisions has had a major beneficial impact on private capital formation in Chile, as evidenced by our results and the record levels which this variable is achieving in the present. -91 - The foreign adjustment in Chile during the 1980s has implied a massive real exchange rate depreciation and a 19 percentage point correction of the trade surplus to GDP ratio (from a 10% of GDP deficit in 1981 to a 6% surplus in 1988), required to ensure continuous debt servicing and supported by a massive fiscal correction culminating in a 4.9% of GDP public sector surplus in 1989. Our results reflect this sensitivity of the trade surplus to the public sector deficit; however, its low parameter also is consistent with the fact that a significant share of the trade surplus adjustment was borne by private consumption. Which are the lessons from the Chilean experience with regard to fiscal policy management? Despite the fact that the Chilean experience does not provide an example of 'quick fLx" policy packages, there are some general conclusions which are noteworthy. First, a stable policy environment provides a solid stage for adjusting to internal and external shocks. Although difficult to quantify, there is little doubt that the economic reforms implemented in Chile before the 1982 debt crisis were an important stabilizing factor in the fast recovery of the Chilean economy during the 1980s. Second, another important lesson which emerges from the Chilean experience is related to the need for consistent policies. In particular, the macroeconomic environment must be supportive of economic reforms. Finally, investment in both physical and human capital is a key to economic growth: public sector focus on high-return investment in human capital and physical infrastructure, complementary with private investment which responds vigorousiy to a stable macroeconomy and market-induced incentives, is showing to be an effective high-growth road to development for Chile. .92v REFERENCES Acevedo, . and Vial, J. (1979): "Demanda por Dinero y Expectativas de Inflaci6n: Chile 1976- 1979", Estudios de Ecnomfa- 14. Aedo, C. and Lagos, L (1984): "Protecci6n efectiva en Chile 1974-1979", in Documento de Trabajo #94, Universidad Cat6lica de Chile. Arrau, P. (1989): "Mtertemporal Monetary Economics: Evidence from the Southern Core of Latin America", unpublished Ph.D. thesis. University of Pennsylvania. Balassa, B. (1989): 'The Effects of Interest Rates on Saving in Developing Countries", manuscript World Bank, Washington, D.C. Banco Central (1990): Indicadores Econ6micos y Sociales 1960-1988. Barro, R. (1974): 'Are Govemment Bonds Net Wealth?, Journal of Political Economy. vol. 82. Bernheim, B.D. (1987): 'Ricardian Equivalence: an Evaluation of Theory and Evidence", in NBER Macroeconomics Annual 1987, MIT Press. Buiter, W. (1988): "Some Thoughts on the Role of Fiscal Policy in Stabilization and Structural Adjustment in Developing Countries", NBER Working Paper 2603. Carr, J. and M. Darby (1981): 'The Role of Money Supply Shocks in the Short-Run Demand for Money", Journal of Moneta.y Economics & (1985): "Reforms and MacroeconomicAdjustments in Chile during 1974-1984", World Development 13, No 18. Corbo, V. (1982): "Desequilibrio de Stocks, Shocks Monetarios y la Estabilidad de la Demanda por Dinero", Cuadernos de Economfa 48. (1989): "Public Finance, Trade and Development: the Chilean Experience", in V. Tanzi, ed., Public Finance and Economic Development. Corbo, V. and J. de Melo (1989): "External Shocks and Policy Reforms in the Southem Cone: A Reassessment", in G. Calvo (ed.): Debt. Stabilization and Development, Basil Blackwell. Corbo, V. and K Schinidt-Hebbel (1991): 'Public Policies and Saving in Developing Countries" Journal of Development Economics forthcoming. Dornbusch, R. (1983): "Real Interest Rates, Home Goods, and Optimal External Borrowing", Journal of Political Economy, vol.1, No.1, February. 93 - (1985): 'Overborrowing: Three Case Studies", in G.W. Smith and J.T. Cuddington: Intemational Debt and the DeveloDing Coun_, World Bank. (1985): 'External Debt, Budget Deficits and Disequilibrium Exchange Rates", in G.W. Smith and J.Cuddington, eds., International Debt and the aeveloping Countries, World Bank. (1988): "Mexico: Stabilization, Debt and Growth", in G. de Menil and R. Portes, eds., Economic Policy: a European Forum- Cambridge University Press. (1989): "Capital Flight: Theory, Measurement and Policy Issues", manuscript M.I.T., Cambridge, Mass. Easterly, W.R. (1989): "Fiscal Adjustment and Deficit Financing during the Cost Crisis", in Husain, I. and I. Diwan (eds.): Dealing with the Debt Crisis, The World Bank, Washington D.C. Easterly, W.R., C.A. Rodrigucz, and K Schmidt-Hebbel (1989): "Research Proposal: The Macroeconomics of the Public Sector Deficits", manuscript, The World Bank, Washington, D.C., March. Edwards, S. and AC. Edwards (1987). Monetarism and Liberalization: the Chilean Experiment, Ballinger Publishing Co. Eyzaguirre, N. and C. Larranaga (1990): "Macroeconomia de los Operacione. uasi-Fiscales en Chile", manuscript, ECLAC, Santiago, Chile. Ffrench-Davis, R. ard J. de Gregorio (1987): "Origenes y Efectos del Endeudamiento Externo en Chile", El Trimestre Econ6mico, 54. Ffrench-Davis, R. (1984): "Indice de Precios Externos", in Coleccion de Estudios CIEPLAN #13, June. Fry, MJ. (1988): Monev. Interest and Banking in Economic Developme;,t Johns Hopkins University, Baltimore. Giovannini, A. (1985): "Saving and the Real Interest Rate in LDCs", Journal of Development Economics, vol. 18, August. Haindl, E. (1986): 'Trimestralizaci6n del PGB por origen y destino", ir. Estudios de Economfa. Hayashi, F. (1982): "The Permanent Income Hypothesis: Estimation and Testing by Instrumental Variables," Journal of Political Econom Vol. 90, No. 5, October. Hayashi, F. (1985): "Tests for Liquidity Constraints: A Critical Survey", NBER Working Paper 1720, Cambridge, Mass., October. Hubbard, R.G. and KL Judd (1986): "Liquidity Constraints, Fiscal Policy, and Consumption", Brookings Papers on Economic ActiMty. 1. -94- Jadresic, E. (1990): "Trimestralizaci6n sin Uso de Series Relacionadas' manuscript, CIEPLAN. Larranaga, 0. (1989): "El Deficit del Sector Publico y Ia Politica Fiscal en Chile", manuscript, ECLAC. Larrafn. F. and Larrafn, A. (1988): "El Caso del Dinero Desaparecido, Chile 1984-1986", Cuadernos de Economia. 75. Leidennan, L and M. Blejer (1988): "Modeling and Testing Ricardian Equivalence: A Survey", IMF Staff Papers, vol. 35, No. 1, March. Le Fort, G. and Ross, C. (1985): "?", in Serie Ln #72 Depto. de Economia, Universidad de Chile. Maddison, A. (1989). The World Economy in the Twentieth Century. Development Center Studies, OECD, Paris. Marshall, J. and K. Schmidt-Hebbel (1989b): 'Economic and Policy Determinants of Public Sector Deficits", PRE Working Paper No. 321, The World Bank, Washington, D.C. Marshall, J. and K Schmidt-Hebbel (1989a): 'Un Marco Analftico Contable para la Evaluaci6n de la Poliftica Fiscal en America Latina", Serie Polftica Fiscal 1, CEPAL Matte, R. and Rojas, P. (1976): "Evaluaci6n del Mercado Monetario Chileno y una Estimaci6n de la Demanda de Dinero", manuscript. Banco Central de Chile. Montiel, P. and N.U. Haque: 'Ricardian Equivalence, Liquidity Constraints, and the Yaari- Blanchard Effect: Tests for Developing Countries, IMF Working Paper IMF, Washington, D.C. Morande, F. and K. Schmidt-Hebbel (1991): "Macroeconomics of Public Sector Deficits: The Case of Zimbabwe:, PRE Working Paper The World Bank, Washington, D.C., forthcoming. Morande, F. and Schmidt-Hebbel, K (eds.) (1988): Del Auge a la Crisis de 1982: Ensavos sobre Estabilizaci6n Einanciera y Endeudamiento en Chilt, ILADES-Georgetown and Instituto Interamericano de Mercados de Capital, Santiago, Chile. Mujica, P. (1990): "Financiamiento Domestico del Deficit Fiscal y sus Efectos Macroecon6micos: Un Modelo para Chile, manuscript ECLAC, Santiago, Chile. Rodriguez, C.A. (1989): "The External Effects of Public Sector Deficits_ PPR Working Paper. The World Bank, Washington, D.C. Rossi N. (1988): "Government Spending, the Real Interest Rate, and the Behavior o' Liquidity- Constrained Consumers in Developing Countries", IMF Staff Papers vol. 35, No. 1, March. Schmidt-Hebbel, K. (1987): "Terms of Trade and the Current Account under Uncertainty," 6n1lisis Econ6mico Vol. 2, No. 1, June. -95. Schmidt-Hebbel, K. (1988): "Consumo e Inversi6n en Chile (1974-1982)- Una Interpretaci6n 'Real' del Boom", in Morande and Schmidt-Hebbel (1988). Schmidt-Hebbel, K and T. Mueller (1990): 'Private Investment and Saving in Morocco", E Working Paper, forthcoming, The World Bank, Washington, D.C. Schmidt-Hebbel, K, S. Webb and G. Corsetti (1991): 'Household Saving in Developing Countries", World Bank Economic Review, forthcoming. Schmidt-Hebbel, K, F. Castro, and I. Leng, (1990) 'Una Base de Datos Trimestrales para la Economfa Chilena", Serie Investigacion #24, Ilades-Georgetown University. Serven, L (1990): "A RMSM-X Model for Chile", PRE Working Paper, The World Bank, Washington, D.C. Solimano, A. (1990): "How Private Investment Reacts to Changing Macroeconomic Conditions: The Case of Chile in the 1980s", PRE Working Paper The World Bank, Washington, D.C. Summers, L H. (1985): "Issues in National Savings Policy", NBER Working Paner 1710, Cambridge, Massachusetts. Tanzi, V., M. I. Blejer, and M. 0. Teijeiro (1987): "Inflation and the Measurement of Fiscal Deficits", IMF Staff Papers, Vol. 34, No.4, December. van Wijnbergen, S. (1989): "External Debt, Inflation and the Public Sector: Towards Fiscal Policy for Sustainable Growth", World Bank Economic Review 3. Vial, J. and Marin, C. (1986): "Series Monetarias Chilenas 1960-1985", in Estudios de Economia. World Bank (1990): Second ReRort on Adjustment Lending, Washington, D.C. World Bank: World Development Report. Oxford University, various issues. -96- APPENDIX 1: REDUCED-FORM COEFFICIENTS This appendix defines the short-run coefficients of assets demand equations (3.7)-k3.8) and the coefficients of the reduced-form equations for the nominal interest rate and the price level (equations (3.9)-(3.10)) in terms of the structural coefficients of equations (3.2), (3.3), (3.5) and (3.6). The coefficients of the asset demand are determined by the following combinations of structural parameters: 00 = aoa OPo = a Bo o1 = ca, 01 = 8"13 02 = oa2 02 = 4 2 0)3 = Cra3 'tp3 = PUB3 04-= ' 4> = I -. S5 = E) 45 = n 06 = a 4D6 = U The reduced-form coefficients are defined determined by the following combinations of structural parameters: oa - oB q(a2Bo,B2ao) Qo~ ~ ~ ~~ ~~~~ = Bo no~~~~~~1= o02 - t2 (a82 -'-123 aa,-,u1 qu(a2B1--a2) M02 -482 - c2 - 97 - oa, - /sBt oru(a2B3-aA2) 92 = X2 = /sB2 - 4a2 ya - oa2 1 - a (a-lpB2 Qt = -- 3 = #82 -oa, 10132 - 0a2 f I (I-,Ua2 Q4 = n4 =- #132 - aa2 #82 - a2 #B2 -Qa2 42 - Oa2 - n aa(n-1) 96 = 6= #B2 - oa=2 U'B2 - Ora, n - cOna2 527 = JT7 = #132 - oa2 JU32 - Oa. et - 0#12 08 = f8= #132 - ora2 #82 -va, or - /4uoj(a2-B2) '19= p832 - 2 /482 - oa -98- APPENDIX 2: DATA SOURCES This appendix gives major sources of data used in the paper. Chapter 3 The basic sources were: Indicadores Economicos y Sociales del Banco Central Chile (IES), Instituto Nacional de Estadistias (INE), International Monetary Fund Financial Statistics (IFS), Cortazar and Marshall (1980), Haindl (1986), Vial y Marin (1986), Le Fort y Ross (1985), Larranaga (1989). i Domestic nominal interest rate. The average of the effective monthly interest rate paid on short-term deposits (30-89 days). Source: IES. i* Foreing interest rate. Monthly average London Interbank offered rate in U.S. dollars for 180 days deposit. Source: IFS. Expected devaluation rate of the nominal exchange rate. Source: for the 1975:1 - 1984:3 period Le Fort and Ross (1985) and for the 1984:4 - 1988:4 a univariate method based on the nominal exchange rate of the banking market (IES). Ml Narrow (MI) money balances. The original series was deseasonalized using the SAMA command in TSP. Source: Vial and Marin (1986), IES. D Quarterly domestic public debt. Source: Annual data were obtained from Larranaga (1989) and correspond to the debt of the Central Bank with the private sector. The quarterly interpolation was performed following the method described in Jadresic (1990). Y Quarterly real GDP. Sources: Haindl (1986) and IES. Anticipated money was estimated assuming that actual money balances followed an autoregressive process of the fourth order. Chapter 4 The basic sources were Indicadores Economicos y Sociales (IES), Solimano and Zucker (1989), Solimano (1989). Cp Private consumption. Source: IES. CC Public consumption. Source: IES. .99- Y Annual real GDP. Source: IES. I Private flx,'d-capital investment. Source: from 1960 to the source was Solimano and Zucker (1989) and for the rest of the period Solimano (1989). UCK User cost of capital. Source: based on IES data base. MPK Marginal product of capital. Source: based on IES data base. K This series was elaborated based on Solimano and Zucker (1989) and Solimano (1989), using the pivotal method assuming a ratio between capital and GDP of 2.5. In addition, we assumed that the ratio between public and private capital was similar to the ratio between public and private investment. H Nominal base monc, tock. Source: IES. The rest of the variables were obtained directly from the YES or derived from the variables defined above according to the definitions provided in the text. Chapter S The basic sources were: Indicadores Economicos y Sociales (IES), Schmidt-Hebbel, Castro, Leng (1990), French Davis (1984), Lagos and Aedo (1984). Px/Pn The relative price of exportable and non-tradable goods. Source: IES for the exportable goods and Schmidt-Hebbel, Castro and Leng (1990) for the non- tradable goods index. Pm/Pn The relative price of importable and non-tradable goods. Source: price of importable goods, YES. tm Nominal average tax rate on the importable goods. Source: 1960-1969: French Davis (1984). 1970-1973: current at the 1979 level. 19741979: Labos and Aedo (1984). 1980-1988: elaborated by the authors. The rest of the variables were obtained from the YES. Chapter 6 The basic source were French Davis and Munoz (1990), Madison (1989) and ECIAC data base. The data on GDP and Investment were obtained from French Davis and Munoz (1990) and the data on factor inputs from Madison (1989) and ECLAC data base PRE Working Paper Series Contact T" Author lDo fgr pa,per WPS675 Are Buybacks Back? Menu-Driven Ishac Diwan May 1991 S. King-Watson Debt-Reduction Schemes with Mark M. Spiegel 33730 Heterogenous Creditors WPS676 Higher Education in the Republic Viswanathan Selvaratnam May 1991 C. Cristobal of Yemen: The Univers:ty of Sana'a Omporn L. Regel 33640 WPS677 On Economic Transformation in East- Andres Solimano May 1991 E. Khine Central Europe: A Historical and 39361 International Perspective WPS678 enomic Growth: A Review of the David Renelt May 19511 R. Luz rheoretical and Empirical Literature 34303 WPS679 Poverty Alleviation in Mexico Santiago Levy May 1991 M. Stroude 38831 WPS680 On Hunger and Public Action Martin Ravallion May 1991 C. Spooner 30464 WPS681 Political-Economy Arguments for Arvind Panagariya May 1991 K. Cabana UniformTariff Dani Rodrik 37947 WPS682 Intertemporal Substitution, Risk Patricio Arrau May 1991 S. King-Watson Aversion, and Private Savings in Sweder van Wijnbergen 31047 Mexico WPS683 Vocational Schooling, Occupational Shoshana Neuman May 1991 C. Cristobal Matching, and Labor Market Earnings Adrian Ziderman 33640 in Israel WPS684 The Value of Intra-household Survey Lawrence Haddad May 1991 J. S~weeney Data for Age-based Nutritional Ravi Kanbur 31021 Targeting WPS685 Children and Intra-household Ravi Kanbur May 1991 J. Sweeney Inequality: A Theoretical Analysis 31021 WPS686 Lending for Learning: Twenty Years Adriaan Verspoor May 1991 C. Cristobal of World Bank Support for Basic 33640 Education WPS687 Brazilian Frozen Concentrated Carlos Alberto Primo Braga May 1991 N. Artis Orange Juice: The Folly of Unfair Simao Davi Silber 37947 Trade Cases WPS688 Macroeconomics of Public Sector Felipe Morande May 1991 S. Jonnakuty Deficits: The Case of Zimbabwe Klaus Schmidt-Hebbel 39074 PRE Working Paper Series Contact TM AUhz Da afo = WPS689 Do Tax Policies Stimulate Investment Anwar Shah May 1991 A. Bhalla in Physical and Research and John Baffes 37699 Development Capital? WPS690 The Terms-of-Trade Effects from the Gabor Oblath May 1991 J. Smith Elimination of State Trading in Soviet- David Tarr 37350 Hungarian Trade WPS691 Can Debt-Reduction Policies Restore Jacques Morisset May 1991 S. King-Watson Investmei t and Economic Growth 31047 in Highly Indebted Countries? A Macroeconomic Framework Applied to Argentina WrS692 Health Financing in the Poor J. Brunet-Jailly May 1991 0. Nadora Countries: Cost Recovery or Cost 31091 Reduction? WPS693 Report on Adjustment Lending II: Vittorio Corbc May 1991 A. Oropesa Lessons for Eastern Europe 39075 WPS694 Labor Markets in an Era of Susan Horton May 1991 M. Schreier Adjustment: An Overview Ravi Kanbur 36432 Dipak Mazumdar WPS695 Long Term Prospects in Eastern Ishac Diwan June 1991 S. King-Watson Europe: The Role of External Finance Fernando Saldanha 33730 in an Era of Change WPS696 Macroeconomics of Public Sector Jorge Marshall June 1991 S. Jonnakuty Deficits: The Case of Chile Klaus Schmidt-Hebbel 39074 WPS697 Volatility Reversal from Interest Paul D. McNelis June 1991 S. Jonnakuty Rates to the Real Exchange Rate: Klaus Schmidt-Hebbel 39074 Financial Liberalization in Chile, 1975-82 WPS698 Tax Policy Options to Promote Andrew Feltenstein June 1991 A. Bhalla Private Capital Formation in Pakistan Anwar Shah 37699 WPS629 Regulation and Deregulation in Ralph Bradburd June 1991 E. "-drona Industrial Countries: Some Lessons David R. Ross 37496 for LDCs WPS700 Trade Liberalization and the Oleh Havrylyshyn June 1991 N. Castillo Transition to a Market Economy David Tarr 37961 WPS701 Education and Adjustment: A Review Andrew Noss June 1991 C. Cristobal of the Literature 33640