I. Introduction

During the last ten or so years, Latin American countries have carried out a variety of economic programs aimed at stabilizing their economies. Several of these programs have been extensively analyzed in the literature, with focus on the policy mix adopted in the programs, in particular the balance between “orthodox” instruments (in particular monetary and fiscal policies) and “heterodox” ones (i.e., direct government intervention in the wage and price formation processes). So far, relatively less attention has been paid to the specific content of the fiscal component of stabilization programs in Latin America. To be sure, a number of studies have emphasized the importance of credibility and sustainability of the fiscal measures adopted in these programs. There have been, however, few attempts at analyzing the impact that the nature and quality of fiscal adjustment have had on the success or failure of the programs.

This paper attempts a more systematic review of the role of fiscal policy in a broad range of stabilization programs in Latin America since the early 1980s. In section II, the theoretical rationale is set out for the view that not only the degree of fiscal adjustment, but also its nature and composition are important for lasting success of the stabilization effort. In section III, experiences of a number of Latin American countries are used to illustrate the theoretical agreement.

II. Why do nature and quality of fiscal adjustment matter?

There exists a broad consensus in the literature—and, increasingly also among economic policy makers in Latin America—that effective macroeconomic stabilization cannot rely primarily on wage and price interventions, as was once claimed by proponents of an extreme version of the inertial theory of inflation. Undoubtedly, this consensus has been shaped by the experiences of numerous failed stabilization attempts that were implemented in Latin America during the 1980s, and that relied predominantly on “deindexing” the economy through temporary wage and price freezes. It is now generally recognized that effective stabilization requires a core of “orthodox” policies, aimed at reducing domestic demand to levels consistent with available supply, and that heterodox policy elements (such as deindexing wages and contracts) can best play a supporting role to orthodox policies, mainly in countries with a persistent record of high inflation.

Views differ on whether stabilization programs should use as a nominal anchor the exchange rate or a monetary target. The answer to this question depends, inter alia, on the availability of foreign exchange reserves (or foreign financing) to back a fixed exchange rate, on the stability (or lack thereof) of the demand for money, and on the history of past stabilization efforts that may influence the credibility of one or the other approach.

It is generally acknowledged that fiscal adjustment plays a key role in both exchange rate-based and money-based disinflation efforts. Proponents of what Montiel (1989) called the “fiscal view of inflation,” see large and prolonged fiscal deficits, that lead to excessive monetary expansion, as the root cause of high inflation. In this framework, clearly, a substantial reduction of the size of the fiscal deficit is a sine qua non for successful disinflation. However, even under an alternative analytical framework (which Montiel called the “balance of payments view of inflation,” and which views inflation as being driven mainly by balance of payments disequilibria leading to exchange rate depreciation) reducing fiscal imbalances can be viewed as necessary to reduce inflation. By reducing domestic demand, a fiscal contraction improves the current account of the balance of payments; moreover, by strengthening confidence in government economic policies, it may also help arrest or reverse capital outflows.

While the need for fiscal adjustment in stabilization and disinflation programs can thus be considered part of “received wisdom,” it is somewhat surprising that the literature generally does not dwell on the desirable characteristics of such an adjustment. Notable exceptions in this respect are papers by Tanzi (1990), and, more recently, Perry and Herrera (1994), and Edwards (1996). As noted by Tanzi (1990), four main issues need to be addressed in designing the fiscal component of a stabilization program: (i) measuring the extent of the existing fiscal disequilibrium; (ii) determining the size of the fiscal adjustment needed; (ii) selecting appropriate high-quality fiscal adjustment measures; and, finally, (iv) sequencing correctly the implementation of the selected measures. Each of these tasks is fraught, of course, with considerable difficulties.

Measuring the degree of an existing fiscal disequilibrium is often made difficult by shortcomings in the quality and comprehensiveness of available fiscal data, due to incomplete coverage (e.g., of the operations of decentralized agencies, extrabudgetary funds, or the subnational governments), long delays in reporting by spending units, lack of information on spending commitments and arrears, or the existence of extensive quasi-fiscal operations. Also, in measuring the fiscal deficit under conditions of high inflation, there arise additional conceptual and practical issues.²

Determining the size of the fiscal adjustment needed to arrive at a fiscal balance that is consistent with the macroeconomic objectives of a stabilization program would require, in principle, the use of a full-fledged econometric model. At a minimum, it requires informed judgments on certain key macroeconomic variables, such as the demand for money consistent with the inflation and output targets.

A major difficulty in assessing the size of the required fiscal adjustment is to distinguish exogenous from endogenous factors, and temporary from lasting changes in the fiscal balance. It is a well-known fact that stabilization programs—even programs that are primarily or solely based on heterodox policies and do not have an explicit fiscal adjustment component—may, initially, have important positive effects on the fiscal balance. These so-called endogenous fiscal effects are summarized in Table 1.

Table 1.

Endogenous Fiscal Responses in Stabilization Programs and Main Variables That Determine Their Size ^1/

^1/“Ceteris paribus” assumptions are considered to hold. Some effects may differ when these assumptions are relaxed, for example, when the trade balance changes significantly in response to exchange rate adjustments.

Table 1.

Endogenous Fiscal Responses in Stabilization Programs and Main Variables That Determine Their Size ^1/

Type of Endogenous Fiscal Response	Variables that Determine the Size of the Endogenous fiscal Response
Endogenous responses that improve the overall fiscal balance
• Reverse Tanzi effect: a decline in inflation tends to affect positively the ratio of tax revenue to GDP.	The reverse Tanzi effect will be larger the stronger the decline in inflation, the lower the frequency of tax collections, and the greater the share of taxes that are less frequently collected or have long collection lags.
• A devaluation of the nominal exchange rate at the outset of a stabilization program increases revenue from taxes on foreign trade. This effect is reversed if the real exchange rate appreciates as stabilization proceeds. In addition, the devaluation would increase the domestic currency value of foreign currency-denominated credits of the public sector vis-à-vis the private sector. It would also increase the domestic currency value of the return on foreign exchange reserves of the central bank, and therefore the transfer of profits of the latter to the budget.	The first effect will be stronger the larger the devaluation of the nominal exchange rate, and the higher the share of foreign trade taxes in total revenue. The strength of the others will reflect the size of foreign currency credits of the public sector and of the foreign exchange reserves.
• In the initial phase of an exchange-rate based stabilization, the decline in inflation may lead to a consumption boom, through the real wealth effect and, in some cases, through a shift in the income distribution in favor of groups with a higher marginal propensity to consume. To the extent that revenue depends on consumption-based taxes, the tax revenue to GDP ratio will increase during the boom period	The effect will be stronger the faster the growth of consumption relative to the growth of GDP, and the larger the share of consumption taxes in total revenue.
• An appreciation of the real exchange rate in stabilization programs (particularly in programs based on a fixed nominal exchange rate) reduces the real domestic currency value of interest payments on foreign currency-denominated debt.	The effect will be stronger the larger the foreign currency denominated debt, and the greater the real appreciation.
• A reduction in inflation leads to lower nominal domestic interest rates and, hence, interest payments on the stock of domestic currency-denominated debt outstanding. This, in turn, reduces the overall (but not the operational) fiscal deficit.	The effect will be stronger the larger the stock of domestic currency debt outstanding, and the greater the decline in inflation.
Endogenous responses that worsen the overall fiscal balance
• A devaluation of the nominal exchange rate at the outset of a stabilization program increases the local currency equivalent of external public debt service payments and other expenditure denominated in foreign currencies. The effect is reversed if the real exchange rate appreciates as stabilization proceeds.	This effect is stronger the larger the foreign currency-denominated public debt and other foreign currency expenditures, and the greater the devaluation.
• The maintenance of a backward-looking indexation of wages, salaries, and social benefits, as inflation rates decline under a stabilization program, increases real government expenditures (and the expenditure to GDP ratio). Of course, the reverse is true for forward-looking indexation.	This effect depends on the extent to which wages, salaries, and social benefits continue to be indexed; the weight of indexed expenditures in the budget; and the extent of the decline in inflation.
• An appreciation of the real exchange rate during the course of an exchange-rate-based stabilization program reduces the real domestic currency value of revenue from foreign trade taxes.	This effect is stronger the larger the size of the real depreciation, and the larger the share of foreign trade taxes in total revenue.
Endogenous responses with ambiguous effects on overall fiscal balance
• Changes in inflation tax revenue	Inflation tax revenue is maximized where money demand is unit elastic. Usually, inflation tax revenue increases up to relatively high levels of inflation, but eventually collapses when a country moves into hyperinflation (or becomes effectively “dollarized”). Hence, a stabilization program may or may not reduce inflation tax revenue depending on when during the inflationary process a stabilization is carried out, and the extent of stabilization.

^1/“Ceteris paribus” assumptions are considered to hold. Some effects may differ when these assumptions are relaxed, for example, when the trade balance changes significantly in response to exchange rate adjustments.

View Table

Table 1.

Endogenous Fiscal Responses in Stabilization Programs and Main Variables That Determine Their Size ^1/

Type of Endogenous Fiscal Response	Variables that Determine the Size of the Endogenous fiscal Response
Endogenous responses that improve the overall fiscal balance
• Reverse Tanzi effect: a decline in inflation tends to affect positively the ratio of tax revenue to GDP.	The reverse Tanzi effect will be larger the stronger the decline in inflation, the lower the frequency of tax collections, and the greater the share of taxes that are less frequently collected or have long collection lags.
• A devaluation of the nominal exchange rate at the outset of a stabilization program increases revenue from taxes on foreign trade. This effect is reversed if the real exchange rate appreciates as stabilization proceeds. In addition, the devaluation would increase the domestic currency value of foreign currency-denominated credits of the public sector vis-à-vis the private sector. It would also increase the domestic currency value of the return on foreign exchange reserves of the central bank, and therefore the transfer of profits of the latter to the budget.	The first effect will be stronger the larger the devaluation of the nominal exchange rate, and the higher the share of foreign trade taxes in total revenue. The strength of the others will reflect the size of foreign currency credits of the public sector and of the foreign exchange reserves.
• In the initial phase of an exchange-rate based stabilization, the decline in inflation may lead to a consumption boom, through the real wealth effect and, in some cases, through a shift in the income distribution in favor of groups with a higher marginal propensity to consume. To the extent that revenue depends on consumption-based taxes, the tax revenue to GDP ratio will increase during the boom period	The effect will be stronger the faster the growth of consumption relative to the growth of GDP, and the larger the share of consumption taxes in total revenue.
• An appreciation of the real exchange rate in stabilization programs (particularly in programs based on a fixed nominal exchange rate) reduces the real domestic currency value of interest payments on foreign currency-denominated debt.	The effect will be stronger the larger the foreign currency denominated debt, and the greater the real appreciation.
• A reduction in inflation leads to lower nominal domestic interest rates and, hence, interest payments on the stock of domestic currency-denominated debt outstanding. This, in turn, reduces the overall (but not the operational) fiscal deficit.	The effect will be stronger the larger the stock of domestic currency debt outstanding, and the greater the decline in inflation.
Endogenous responses that worsen the overall fiscal balance
• A devaluation of the nominal exchange rate at the outset of a stabilization program increases the local currency equivalent of external public debt service payments and other expenditure denominated in foreign currencies. The effect is reversed if the real exchange rate appreciates as stabilization proceeds.	This effect is stronger the larger the foreign currency-denominated public debt and other foreign currency expenditures, and the greater the devaluation.
• The maintenance of a backward-looking indexation of wages, salaries, and social benefits, as inflation rates decline under a stabilization program, increases real government expenditures (and the expenditure to GDP ratio). Of course, the reverse is true for forward-looking indexation.	This effect depends on the extent to which wages, salaries, and social benefits continue to be indexed; the weight of indexed expenditures in the budget; and the extent of the decline in inflation.
• An appreciation of the real exchange rate during the course of an exchange-rate-based stabilization program reduces the real domestic currency value of revenue from foreign trade taxes.	This effect is stronger the larger the size of the real depreciation, and the larger the share of foreign trade taxes in total revenue.
Endogenous responses with ambiguous effects on overall fiscal balance
• Changes in inflation tax revenue	Inflation tax revenue is maximized where money demand is unit elastic. Usually, inflation tax revenue increases up to relatively high levels of inflation, but eventually collapses when a country moves into hyperinflation (or becomes effectively “dollarized”). Hence, a stabilization program may or may not reduce inflation tax revenue depending on when during the inflationary process a stabilization is carried out, and the extent of stabilization.

^1/“Ceteris paribus” assumptions are considered to hold. Some effects may differ when these assumptions are relaxed, for example, when the trade balance changes significantly in response to exchange rate adjustments.

The strength of these endogenous effects on the government’s budget will of course vary from country to country, depending on the characteristics of the latter, and on the types of stabilization policies implemented. While some combinations of the two are likely to result in strong endogenous effects (such as a large exchange rate devaluation in a country that relies heavily on taxes on foreign trade), it is very difficult to know ex ante the duration and strength of such endogenous fiscal effects. For example, while an increase in tax revenue due to the reverse Tanzi effect may be considered permanent (as long as inflation stays down), it is more difficult to predict accurately when a consumption boom will end, or when a real exchange rate appreciation will be halted or reversed. Uncertainties about the duration and strength of fiscal improvements resulting from such endogenous factors are, in turn, likely to affect adversely the credibility of a disinflation program, and may contribute to the emergence of self-fulfilling expectations of their reversal. These adverse credibility effects are likely to be stronger the more heavily a disinflation program relies on endogenous policy elements.

There is now growing consensus in the literature that successful stabilization requires high-quality fiscal adjustment, meaning that not only the extent but also the nature and composition of fiscal adjustment matter. There are at least two broad sets of reasons for this.

First, a reduction of the fiscal deficit can have different effects on aggregate demand and supply, depending on the nature of the specific adjustment measures taken. Clearly, increases in different taxes and/or reductions in different categories of public expenditures affect differently the various components of aggregate demand, the incentives to work, save and invest, and thereby, over time, the growth potential of an economy.

Second, fiscal policy packages of the same magnitude can differ profoundly in their sustainability, and in the perception thereof by economic agents, and can therefore have different effects on expectations and market behavior. Specifically, fiscal measures that are—and are viewed by economic agents, both within and outside the country—as politically feasible, socially balanced, and of a lasting nature, are likely to carry greater credibility—and to affect appropriately market behavior—than measures which, because of their temporary character or their lack of political viability, can be expected to be reversed soon.

There is, of course, a good deal of variability from country to country in what makes a fiscal policy package to be of high quality and sustainable. An assessment of the political and social viability, as well as of the administrative feasibility, of specific fiscal measures needs to take into account the circumstances of each individual country at a particular point in time. For example, even a sound and well-balanced package of structural fiscal adjustment measures may not be feasible or credible if it is introduced during the last months of the expected tenure of a government, unless it is endorsed by the likely successor government.

Nevertheless, some broad generalizations on high quality fiscal measures can be drawn from a range of country experiences. As concerns tax policies, it is widely recognized that sustainable and high quality reforms are those that aim at simplifying the tax system by eliminating relatively low-yielding taxes; broadening the tax base, while streamlining the rate structure and reducing high marginal tax rates; and at establishing a level playing field, by curtailing special treatments and incentives, and thereby reducing the scope for tax planning and avoidance. In the area of tax administration, reform efforts need to focus on promoting taxpayer compliance through strengthened procedures for monitoring, detecting and reacting to noncompliance, including the establishment of tough but credible penalties. These broad tenets of policy and administration have been reflected to some extent in tax reforms worldwide—which have included the introduction of the VAT in a large number of countries and reductions in the top marginal rates of most income taxes—and in ongoing efforts to strengthen and modernize tax administrations. Often, however, resistance by organized interest groups and a misplaced confidence in the effectiveness of tax incentives, have constrained progress.

On public expenditure policy reform, there is also a consensus, at least in theory, that it is desirable to eliminate or phase out unproductive expenditures (such as “white elephant” investment projects or excessive military expenditures); make adequate provision for operation and maintenance expenditures; replace generalized price subsidies with targeted income support to the poorest and most vulnerable segments of the population; give priority, in the allocation of scarce resources for health and education, to basic/preventive health services and primary education; and reduce the overstaffing of civil services, while promoting their productivity and professionalism. Unfortunately, this consensus does not always translate into concrete policy actions, also because of resistance by powerful interest groups that would be adversely affected by these reforms.

Frequently, achieving high quality and sustainable fiscal adjustment requires wide-ranging institutional and structural reforms. Although they are unlikely to yield short-term budgetary savings, such stabilization-enhancing structural reforms are of crucial importance to improve the longer-term outlook for public finances, and may include, inter alia:

• privatizing state enterprises and certain services traditionally provided by the government;

• reforming the social security systems, including health care and pensions, in order to improve their long-term viability;

• overhauling public expenditure management systems in order to improve the choice of expenditure priorities, the control and supervision of expenditure execution, and the ex-post evaluation of the cost-effectiveness of spending programs; and

• strengthening the system of intergovernmental fiscal relations, which, especially in federal countries, is important to ensure that ongoing decentralization efforts do not weaken macroeconomic management or worsen distributional equity;

• implementing banking sector reforms, including through a strengthening of bank supervision, also to reduce the fiscal costs associated with rescue operations for private and public financial institutions.

In general, stabilization programs that do not undertake to address these problems are likely to carry limited credibility. Still, many stabilization programs, including those implemented in Latin America, have tended to rely more on “quick fixes” (i.e., measures that are easily implementable and relatively less costly politically, albeit frequently nondurable and distortive), than on the high-quality measures discussed above. A nonexhaustive list of such quick-fix fiscal policy measures would include (Tanzi, 1990):

• emergency tax measures, including the introduction of temporary surcharges (e.g., on import duties, consumption taxes, or income taxes) or distortionary taxes (e.g., taxes on exports, or on financial transactions);

• incentives to prepay future tax liabilities, which tend to boost current revenues at the expense of future ones;

• tax amnesties, which weaken incentives for future tax compliance, particularly if used repeatedly;

• forced savings schemes and/or forced conversions of public debt instruments;

• across-the-board cuts in non-entitlement spending programs;

• suspensions of investment projects underway, and reductions in operation and maintenance expenditures on completed investments;

• delays in civil service wage adjustments, to compress real wages of civil servants;

• forced unpaid “administrative” leave for government employees;

• accumulation of payment arrears, including to suppliers, civil service salaries, or social security contributions;

• postponement of interest payments on public debt through financial engineering operations;

• reliance on quasi-fiscal operations of the central bank or of state-owned financial institutions, in order to “park” the fiscal deficit outside the government’s budget;

• use of nonfinancial public enterprises for quasi-fiscal purposes (e.g., by levying extraordinary taxes on their profits, allowing their decapitalization, or preventing them from adjusting their workforce as needed).

Frequently, stabilization requires policymakers to confront complex issues of sequencing fiscal adjustment measures. Short-term macroeconomic and financial considerations often argue for giving priority, in timing and effort, to measures that can be expected to yield budgetary savings quickly. Political considerations, often related to preserving the limited political capital and goodwill enjoyed by elected governments in representative democracies, also often argue in the same direction. As a result, structural reform measures, that are frequently politically controversial and often have less immediate (albeit more lasting) economic effects, are often postponed, and, when finally implemented, may not begin to produce benefits until it is too late for them to sustain an ongoing stabilization effort.

III. Selected experiences in latin american Stabilization programs

A comprehensive and detailed review of the fiscal adjustment components in the many stabilization efforts undertaken by Latin American countries in the 1980s and early 1990s is beyond the scope of this paper. Nevertheless, an analysis of salient successful and failed stabilization attempts in Latin America during that period may help illustrate the considerations put forward in section II above.

A first observation that can be made on the basis of a broad-brush review of stabilization efforts throughout Latin America, is that the quality and sustainability of stabilization programs seems to have improved in recent years. Lasting success in stabilization and disinflation during the 1980s was, by and large, more the exception than the rule. Chile in the 1980s, Bolivia (with the New Economic Policy Program of 1985), and Mexico (with the Solidarity Pact of 1987) constituted the main examples of successful stabilization. By contrast, the histories of Argentina, Brazil, Peru and, to a lesser extent, Uruguay and Venezuela during the 1980s, show a succession—in some instances at an accelerating pace—of attempts at stabilizing their economies, with shorter and shorter success records. In Argentina alone one can identify at least 15 such stabilization programs between mid-1985 and the end of 1990. In Brazil, there were at least 9 major and minor stabilization attempts between the beginning of 1986 (the Cruzado Plan) and 1991 (the Collor II Plan). In the more recent past, however, these countries have implemented more comprehensive and sustainable stabilization programs.

While the short-lived stabilization efforts of the 1980s obviously differed in a number of specific characteristics, they also exhibited certain common traits. Their foremost common characteristic is perhaps the fact that they all failed to signal convincingly a fundamental change of the economic policy regime, and therefore lacked credibility. This lack of credibility can, in turn, be viewed as a reflection of a number of factors:

• inconsistent policy mixes, like excessive reliance on heterodox policies, especially repeated and increasingly ineffective price and wage freezes; or the use of a fixed nominal exchange rate anchor unsupported by adequate monetary and fiscal restraint; or an imbalance in the fiscal/monetary mix, resulting in very high levels of real interest rates, speculative capital inflows, and excessive real appreciations of the exchange rate;

• excessive reliance on endogenous factors, which failed to provide a lasting improvement of the fiscal balance;

• failure to implement fundamental fiscal reforms, and sole or main reliance on revenue and expenditure measures of a quick-fix type;

• lack of needed complementary structural reforms.

Each of these points can be illustrated with a few country examples. More or less inconsistent policy mixes plagued many stabilization programs in Latin America during the eighties. The extent and the timing of the consequences of such imbalances for the stabilization performance has varied depending on the severity of the imbalance. Stabilization attempts like Brazil’s Cruzado Plan or Peru’s Inti Plan, both in the mid−1980s, which did not support heterodox policies with financial and fiscal restraint, had only a short-lived success. By contrast, programs like Mexico’s Solidarity Pact, Argentina’s Convertibility Law, or Brazil’s Real Plan, which combined incomes policies with fiscal and monetary tightening, succeeded in securing more lasting disinflation.

Based on the hypothesis that inflation was largely inertial many of the earlier Latin American stabilization programs experimented with temporary prize/wage freezes that were thought to break inflationary expectations. Indeed, some stabilization programs boiled down to little more than a temporary freeze.³ While the effectiveness of freezes in stabilizing prices in the short run explains their recurrent use, they became increasingly ineffective, and, particularly in countries with a history of repeated freezes such as Brazil or Argentina, the very hint or expectation of a new freeze encouraged private sector firms to increase prices preemptively. In addition, once the freeze went into effect, popular expectations that it would soon have to be lifted, and a new bout of inflation would follow, resulted in further delays of tax payments.

The importance of supporting a nominal anchor, like a fixed exchange rate, with adequate monetary and fiscal restraint, in order to achieve a consistent policy mix, can be illustrated, for example, by contrasting Argentina’s failed attempts at fixing the rate in some of its stabilization programs during the 1980s with the success to date of the Convertibility Law. A nominal exchange rate anchor can generally be maintained for a short period, particularly when prices and wages are frozen. However, Argentina’s experience in the 1980s shows that a nominal anchor unsupported by appropriate financial policies tends to quickly lose credibility, thereby jeopardizing the sustainability of stabilization. In contrast to earlier stabilization attempts, the Convertibility Law was backed by substantial financial, fiscal, and structural reforms. Also, the Argentine government demonstrated that it was prepared, in the wake of speculative attacks on the currency, to take prompt fiscal measures—as it did in March 1995, shortly before the presidential elections. These policies were essential for ensuring the success to date of the nominal exchange rate anchor, the key element of the Convertibility Law.

Still, there are numerous examples of imbalances in the fiscal/monetary policy mix, even in programs with a relatively long record of success. Brazil’s recent experience under the Real Plan falls in this category, as an initial improvement in the fiscal position in 1994 was reversed in 1995 through excessive wage increases of civil servants and overspending by state and local governments. Even today, after three years of successful stabilization, further fiscal adjustment and the early implementation of structural fiscal reforms would help strengthen the credibility of the Real Plan and secure its durability over the longer term

Excessive reliance on endogenous factors has been another problem that has beset stabilization programs in a number of Latin American countries. In many cases, fiscal balances showed marked improvements in the early phases of stabilization programs, largely reflecting endogenous factors (see Table 2), like the reverse Tanzi effect or disinflation-induced consumption booms. Usually, these improvements were short lived. The reverse Tanzi effect, for example, played a substantial role in stabilization programs in, among others, Argentina, Bolivia, and Chile, as well as in the initial phase of the Brazil’s Cruzado Plan. Disinflation-induced consumption booms boosted revenues in the early stages of stabilization programs in Brazil (the Cruzado Plan and also the Real Plan), Mexico, and Uruguay. In the case of the latter, it is estimated that, during 1991−93, the cyclically adjusted budget deficit exceeded the actual deficit by a cumulative 9 percentage points of GDP⁴ (Talvi, 1995). The reversal of these temporary factors (as the economies moved into recessions or the exchange rate weakened) exposed the fragility of the fiscal improvement.

Table 2.

Estimated Size of Some Endogenous Fiscal Effects in Various Stabilization Programs

Table 2.

Estimated Size of Some Endogenous Fiscal Effects in Various Stabilization Programs

Country and Period	Reference	Endogenous Responses Considered	Major Results
Uruguay following the December 1990 stabilization program	Talvi (1995)	Cyclically adjusted overall budget balance, i.e., the balance without the temporary effects generated by the stabilization program.	Cyclically adjusted overall budget balance during 1991–93 was, respectively, 1.2, 3.7, and 4.1 percent of GDP worse than the unadjusted balance. In 1991, for example, the unadjusted overall budget deficit was 0.9 percent of GDP, whereas the cyclically adjusted overall budget balance was estimated at a deficit of 2.1 percent of GDP.
Argentina during 1978−85 before the Austral I stabilization plan	Kiguel and Neumeyer (1989)	Inflation tax revenue	Empirical evidence suggests that inflation tax revenue is maximized at a monthly inflation rate of about 20 percent and falls unambiguously if monthly inflation rates exceed 22 percent. At its maximum level, inflation tax revenue amounts to about 7.5 percent of GDP. Any serious stabilization effort requires an alternative source of revenue to replace such a large inflation tax revenue.
Argentina and Brazil in the early-mid 1980s	Dombusch and Simonsen (1987)	Tanzi effect and other elements that cause budget deficit to be high because of inflation	Stabilization automatically reduced budget deficits. The Tanzi effect alone accounted for 1-2 percent of GDP in revenue lost; this revenue was regained automatically during stabilization.
Argentina during 1985−87 under the Austral plan	Machinea and Fanelli (1988)	Tanzi effect	Before the Austral plan, revenue losses on account of the Tanzi effect amounted to about 2.5 percent of GDP. When the Austral plan was in effect, these losses dropped to 0.5 percent of GDP.
Argentina under the Austral plan during June 1985−March 1996	Canavese and Di Tella (1988)	Reverse Tanzi effect	Revenue increased from 23 to 28 percent of GDP; almost 4 percentage points of this increase was due to the fact that real taxes were no longer eroded by inflation.
Argentina under the Austral plan	Kiguel (1991)	Transitory versus permanent effects	The budget deficit was cut from 11.7 percent of GDP in the first quarter of 1985 to 2.3 percent of GDP in the third quarter, largely due to changes in tax revenue. This increase in tax revenue was due to four main elements, increases in trade taxes, higher public sector prices, the introduction of a forced savings scheme and improvements in tax collections resulting from the reduction in inflation (reverse Tanzi effect). The latter two effects were transitory.
Bolivia during 1983−85 before introducing the stabilization plan	Kiguel and Liviatan (1995)	Tanzi effect and inflation tax revenue	Government revenue fell from around 85 percent of expenditure in 1980 to around 50 percent during 1983–85, and coincided with the rise in inflation, which suggests the presence of an extreme form of the Tanzi effect. While inflation tax revenue increased up to 1984, when it reached a level of over 15 percent of GDP, it dropped during the hyperinflation of 1985, and continued to drop further during the ensuing stabilization program.
Argentina, Brazil, and Peru during 1989−90	Kiguel and Liviatan (1995)	Inflation tax revenue	In Argentina and Brazil, inflation tax revenue increased (rather than fall) in response to the onset of hyperinflation; hence, an extreme form of the Tanzi effect was not present. In Peru, the hyperinflation outburst of 1990 was preceded by a large increase in inflation tax revenue; however during much of the period leading up the hyperinflation, which was characterized by high but relatively stable inflation, inflation tax revenue was decreasing.
Nicaragua in the second half of the 1980s	Saborio (1990)	Tanzi effect	Tax revenues plummeted by about 18 percent of GDP during the second half of the 1980s, largely as a result of the Tanzi effect.
Chile during 1950−82	Larraín and Meller (1991)	Tanzi effect	Tanzi effect helps to explain the low level of tax revenue during 1970–83; in particular, regression estimates for 1950–82 suggest that a 10 percent rise in inflation increased the overall fiscal deficit by 1.2 percentage points of GDP.

View Table

Table 2.

Estimated Size of Some Endogenous Fiscal Effects in Various Stabilization Programs

Country and Period	Reference	Endogenous Responses Considered	Major Results
Uruguay following the December 1990 stabilization program	Talvi (1995)	Cyclically adjusted overall budget balance, i.e., the balance without the temporary effects generated by the stabilization program.	Cyclically adjusted overall budget balance during 1991–93 was, respectively, 1.2, 3.7, and 4.1 percent of GDP worse than the unadjusted balance. In 1991, for example, the unadjusted overall budget deficit was 0.9 percent of GDP, whereas the cyclically adjusted overall budget balance was estimated at a deficit of 2.1 percent of GDP.
Argentina during 1978−85 before the Austral I stabilization plan	Kiguel and Neumeyer (1989)	Inflation tax revenue	Empirical evidence suggests that inflation tax revenue is maximized at a monthly inflation rate of about 20 percent and falls unambiguously if monthly inflation rates exceed 22 percent. At its maximum level, inflation tax revenue amounts to about 7.5 percent of GDP. Any serious stabilization effort requires an alternative source of revenue to replace such a large inflation tax revenue.
Argentina and Brazil in the early-mid 1980s	Dombusch and Simonsen (1987)	Tanzi effect and other elements that cause budget deficit to be high because of inflation	Stabilization automatically reduced budget deficits. The Tanzi effect alone accounted for 1-2 percent of GDP in revenue lost; this revenue was regained automatically during stabilization.
Argentina during 1985−87 under the Austral plan	Machinea and Fanelli (1988)	Tanzi effect	Before the Austral plan, revenue losses on account of the Tanzi effect amounted to about 2.5 percent of GDP. When the Austral plan was in effect, these losses dropped to 0.5 percent of GDP.
Argentina under the Austral plan during June 1985−March 1996	Canavese and Di Tella (1988)	Reverse Tanzi effect	Revenue increased from 23 to 28 percent of GDP; almost 4 percentage points of this increase was due to the fact that real taxes were no longer eroded by inflation.
Argentina under the Austral plan	Kiguel (1991)	Transitory versus permanent effects	The budget deficit was cut from 11.7 percent of GDP in the first quarter of 1985 to 2.3 percent of GDP in the third quarter, largely due to changes in tax revenue. This increase in tax revenue was due to four main elements, increases in trade taxes, higher public sector prices, the introduction of a forced savings scheme and improvements in tax collections resulting from the reduction in inflation (reverse Tanzi effect). The latter two effects were transitory.
Bolivia during 1983−85 before introducing the stabilization plan	Kiguel and Liviatan (1995)	Tanzi effect and inflation tax revenue	Government revenue fell from around 85 percent of expenditure in 1980 to around 50 percent during 1983–85, and coincided with the rise in inflation, which suggests the presence of an extreme form of the Tanzi effect. While inflation tax revenue increased up to 1984, when it reached a level of over 15 percent of GDP, it dropped during the hyperinflation of 1985, and continued to drop further during the ensuing stabilization program.
Argentina, Brazil, and Peru during 1989−90	Kiguel and Liviatan (1995)	Inflation tax revenue	In Argentina and Brazil, inflation tax revenue increased (rather than fall) in response to the onset of hyperinflation; hence, an extreme form of the Tanzi effect was not present. In Peru, the hyperinflation outburst of 1990 was preceded by a large increase in inflation tax revenue; however during much of the period leading up the hyperinflation, which was characterized by high but relatively stable inflation, inflation tax revenue was decreasing.
Nicaragua in the second half of the 1980s	Saborio (1990)	Tanzi effect	Tax revenues plummeted by about 18 percent of GDP during the second half of the 1980s, largely as a result of the Tanzi effect.
Chile during 1950−82	Larraín and Meller (1991)	Tanzi effect	Tanzi effect helps to explain the low level of tax revenue during 1970–83; in particular, regression estimates for 1950–82 suggest that a 10 percent rise in inflation increased the overall fiscal deficit by 1.2 percentage points of GDP.

Another shortcoming of many Latin American stabilization programs has been a failure to implement fundamental fiscal reforms and their reliance on revenue and expenditure measures of a quick-fix type. For example, Argentina during most of the 1980s, Chile during 1984−87, and Colombia during 1985−88 all made extensive use of foreign trade taxes, despite their well known allocative shortcomings. In some cases, reliance on the quick fixes was in stark contrast to the declared general policy objectives. For example, Argentina’s 1985 Austral Plan raised export taxes, notwithstanding the declared policy objective of fostering export-led growth. Some countries used multiple exchange rates as an implicit form of export taxation. Such distortive policies, once in effect, are difficult to change: for example, in 1989, export taxes still amounted to over 1.5 percent of GDP in Argentina, before being eliminated in the context of the Convertibility Law plan.

Another popular quick-fix stabilization measure has been the use of taxes on financial and exchange transactions, which tend to increase financial disintermediation. For example, especially during the 1980s, Argentina and Brazil, repeatedly resorted to taxes on financial transactions, because of their relative ease of collection and also made significant use of forced savings schemes, forced conversions of public debt instruments, and outright freezes of financial assets. To some extent, the cost of such schemes in terms of public confidence still manifests itself today in the relatively high real rates of interest demanded by financial asset holders, especially in Brazil.

Similarly, public sector price adjustments were frequently viewed as a quick-fix remedy in stabilization programs. Frequently, governments had failed to adjust some public sector prices (for example, public transportation or household energy tariffs) in line with inflation or cost developments. Many stabilization programs relied heavily on adjusting public sector prices to reduce enterprise subsidies or to halt full decapitalization of these enterprises. Bringing public sector prices closer in line with cost often required rather large adjustments. Instead of implementing systematic adjustments of public sector prices to costs, these adjustments were, in many instances, implemented haphazardly in large steps, after which prices were kept constant and left again to erode in real terms. In some stabilization programs, particularly in the 1980s, public sector price adjustments were decreed along with, or immediately prior to, a general price freeze. However, these adjustments fed into prices of other sectors, thereby increasing cost pressures on prices and contributing to the demise of the price freeze. Public sector price adjustments were often implemented in a way that magnified their adverse impact on inflationary expectations, wage claims, and, in some cases (for example, Venezuela in 1989), even led to social unrest.

Up to very recently, the tax system of many Latin American countries resembled Heymann’s (1991) description of the Argentine tax system of the late 1980s: “the general picture is that of a system without clear design, with complicated legislation that is not enforced, which cannot collect broad-based taxes and has to rely on a diversity of rather primitive taxes.” Some of the more successful stabilization programs made remarkable progress precisely in reforming tax policies and strengthening tax administration. The Bolivian program of 1985, for example, replaced a complicated and ineffective system of over 450 different taxes, levied mainly on income, with a streamlined system of nine taxes, levied mainly on consumption and wealth. Substantial tax reforms were also part of the Mexico program of 1987 and of the more recent stabilization programs in Argentina (1991) and Peru (1990). The Mexico program included reductions in tax rates (especially import tariffs and income taxes), together with steps to broaden the tax base and to strengthen enforcement. In the context of Argentina’s Convertibility Law of 1991, about 20 distortionary taxes were abolished, including some taxes on exports and on financial assets, and many exemptions and incentive schemes were curtailed; at the same time, a substantial effort was launched to strengthen tax administration. Peru’s Fujimori Plan of 1990 included major reforms both in tax policy (with a drastic simplification of the tax system and reductions of exemptions, as well as the introduction of some emergency taxes), and in tax administration. A significant beginning in needed tax reforms has also been made recently in Brazil, with a streamlining of personal and corporate income taxes, and with proposals to replace the narrowly-based federal level VAT (the IPI) with one levied on the same broad base as the state-level VAT (the ICMS).

The public expenditure systems of many Latin American countries were also plagued by severe shortcomings. Again, Heymann’s (1991) description of Argentina in the late 1980s can serve as a good illustration of a fairly generalized problem throughout the region: “expenditure allocations depended more on the amount of pressure that various groups could exert than on considerations of economic performance or overall equity,” and since the budget “was always approved late in the year, it validated actions that had already been taken.” Also, expenditure management systems were frequently weak, and, as a result, expenditure controls remained fairly ineffective.

Within this context, it is not surprising that many of the earlier stabilization programs in the 1980s either did not pay much attention to expenditure reform or relied on rather crude policy measures that amounted to little more than “cuts without adjustment.” Hausmann’s (1990) description of expenditure policy reforms in Venezuela in the mid−1980s seems applicable to many Latin American stabilization programs of the time: “no strategic criteria were established to redirect spending: cuts were made in those areas where they were easy to achieve.” Usually, the initial brunt of expenditure cuts fell on public investments, and on operation and maintenance spending, as expenditures on wages and entitlement programs (such as social security payments and transfers to subnational levels of government) were relatively inflexible in the short run. However, in several cases, when fiscal adjustment needs were especially large, governments resorted to more drastic but equally short lived measures, such as freezes in civil servants wages and in public sector hiring. In a few instances, they even ran up payment arrears for wages, pensions, or interest on the public debt. For example, the issuing of recognition bonds (Bocones) in Argentina, which, during 1991−95 totaled Arg$20.5 billion (or 8 about percent of the country’s 1993 GDP),⁵ was a legacy of an era when running up payment arrears to pensioners and suppliers was used extensively as a fiscal policy tool.

Expenditure management during stabilization often took the form of cash rationing, based on revenue availability. Use of cash limit characterized also some of the more successful stabilization programs, such as the Bolivian plan of 1985 and the more recent programs in Argentina and Peru. Cash limits can be an effective instrument of spending control when they are part of a realistic government financial plan, including the planning and timely monitoring of spending commitments. Over the last several years, a number of Latin American countries have made substantial progress in strengthening financial control over expenditures, particularly of the central government. For example, already in 1986, Brazil implemented a major reform and modernization of the central government financial management information system. Similar systems have been, or are being, set up in a number of other Latin American countries in recent years. Most of these systems, however, so far do not cover adequately the operations of decentralized agencies, and especially of the subnational levels of government. Moreover, these systems are primarily geared toward financial control. More limited progress has been made to date in strengthening other aspects of public expenditure management, including budgeting techniques and ex-post evaluation of the quality and cost-effectiveness of public expenditures (value-for-money analysis).

The evidence on the distributional impact of successful stabilization programs in Latin America to date is mixed. A number of studies have suggested that the distributional impact of expenditure policies adopted in the context of stabilization programs has often been negative.⁶ However, it is also clear that the lower income groups tend to be the ones most adversely affected by high inflation (Amadeo, 1995): backward looking indexation does little to protect their wages when inflation is rising; they frequently do not hold indexed financial assets; high real interest rates imply a major redistribution of wealth from the government to its creditors (who tend to be in the middle to higher income brackets). On the other hand, it is also true that the impact of budgetary cuts, necessitated by stabilization efforts, has often fallen sharply on social spending. Pressures from powerful interest groups and the lack of political clout of the poor have often prevented expenditure reforms that would be more likely to benefit the poor, such as redirecting health and education expenditures toward basic/preventive health care and primary education, and improving the targeting of social safety nets.

These expenditure allocation problems continue to beset stabilization programs, and more generally expenditure policies. For example, in the context of Mexico’s recent PARAUSEE stabilization program, an important initial priority was to provide significant budgetary support to bank debtors (mostly for mortgage and credit card debt), also to prevent large-scale debtor defaults from adding to the severity of the banking crisis. Only in a second step the Mexican government started to reform its social safety nets, for example by transforming generalized transfers to all households into transfers targeted to the poor. Securing substantial improvements in the income distribution without endangering macroeconomic stability remains a major challenge throughout Latin America in the years ahead.

Finally, some of the early stabilization programs in Latin America were characterized by a lack of complementary structural reforms. For example, while privatization was often talked about in the 1980s, it usually remained nothing but a promise; none of the unsuccessful Argentine stabilization attempts of the 1980s even envisaged privatizing state enterprises to the extent that it was carried out in the context of the Convertibility Law a few years later. By contrast, in the last several years, a number of Latin American countries have made substantial progress in structural reform. Argentina, Bolivia, Chile, Mexico, and Peru have made major strides in restructuring and privatizing state enterprises, thus reducing their drain on the budget and improving the overall efficiency of the economy. Similarly, major social security reforms have been already implemented in Chile and Argentina, and are currently underway in other Latin American countries. These reforms entail substantial short term cost to the budget but will have major beneficial effects in the longer run. Available evidence already suggests that, in general, those Latin American countries that implemented deeper structural reforms experienced greater reductions in macroeconomic volatility than did countries in which reforms were less pronounced (Gavin, 1997 and Lora, 1997).⁷

Still, it seems fair to say that the structural reform agenda in areas related to the public finances remains large for most Latin American countries. Major items in this agenda are: (i) continuing progress in restructuring and privatizing public enterprises, including public banks; (ii) continuing the reform of social security systems, to ensure their financial viability over the long term with moderate payroll tax rates; (iii) advancing with civil service reform to improve government efficiency; and (iv) undertaking, especially in federal states, major reforms of the system of intergovernmental fiscal relations. While these reforms are unlikely to lead to quick improvements in the fiscal balance—indeed some, like the restructuring of state enterprises and banks, civil service reforms, or pension reforms are usually costly, also because they bring into the open hidden liabilities—failure to carry them out would likely undermine, over the longer term, hard-won gains in fiscal adjustment.

Political economy considerations suggest that conditions for wide-ranging structural reforms are probably better today than during the 1980s. First, Latin America’s fledgling democracies of ten years ago have gained strength and credibility. Second, despite setbacks in the wake of the Mexico crisis, most countries in Latin America have experienced some years of significant economic growth, and, helped by lower international interest rates, have substantially reduced their external debt burden. Most importantly, after prolonged high inflation, a political and social constituency has emerged for low inflation and financial stability. Electorates in several major Latin American countries have rewarded governments that introduced durable stabilization plans and did not shy away from the fiscal reforms necessary to make stabilization policies stick. Finally, Latin America, through trade and financial sector reforms, is becoming more and more closely integrated in the global economy. Its economic policymakers, and its citizens are becoming increasingly aware that structural reforms— although inevitably involving losses for various, often powerful, interest groups—are necessary to enable their countries to fully exploit the potential gains from globalization.

All these considerations give cause for cautious optimism that the progress made in recent years by most Latin American countries in stabilization will not be lost, that the maintenance of low rates of inflation will go hand in hand with strong and balanced economic growth, and that governments will persevere with, and expand the scope of, reforms of the public finances.