Fiscal Discipline and Exchange Rate Regimes
Evidence From the Caribbean
  • 1 0000000404811396https://isni.org/isni/0000000404811396International Monetary Fund

Contributor Notes

Author(s) E-Mail Address: rduttagupta@imf.org, gtolosa@imf.org

This paper assesses the nature of fiscal discipline under alternative exchange rate regimes. First, it shows in a simple theoretical framework that fiscal agencies under a currency union with a fixed exchange rate can have the largest incentive to overspend or "free-ride" (compared to those under other exchange rate regimes) owing to their ability to spread the costs of overspending in terms of the inflation tax across both time-given the fixed exchange rate-and space-given the currency union. In contrast, such free-riding behavior does not arise under flexible regimes owing to the immediate inflationary impact of spending. Next, empirically, it shows that fiscal stances in countries with fixed pegs and currency unions regime demonstrate greater free-riding behavior than countries with more flexible regimes in 15 Caribbean countries during 1983-2004.

Abstract

This paper assesses the nature of fiscal discipline under alternative exchange rate regimes. First, it shows in a simple theoretical framework that fiscal agencies under a currency union with a fixed exchange rate can have the largest incentive to overspend or "free-ride" (compared to those under other exchange rate regimes) owing to their ability to spread the costs of overspending in terms of the inflation tax across both time-given the fixed exchange rate-and space-given the currency union. In contrast, such free-riding behavior does not arise under flexible regimes owing to the immediate inflationary impact of spending. Next, empirically, it shows that fiscal stances in countries with fixed pegs and currency unions regime demonstrate greater free-riding behavior than countries with more flexible regimes in 15 Caribbean countries during 1983-2004.

I. Introduction

Some countries have time and again challenged the received wisdom that fixed exchange rates induce fiscal discipline by pursuing fiscal expansion under currency pegs, while others have even allowed fiscal profligacy under a currency union. Thus, a fundamental question arises regarding the relationship between fiscal and monetary regimes, that is, if and how fiscal incentives are affected under alternative exchange rate regimes.

This paper contributes to the understanding of the relationship between exchange rate regimes and fiscal stances by focusing on the mechanisms by which the former affects the latter. In particular, it studies the scope for moral hazard or “free-riding” behavior in fiscal policies under various exchange rate regimes, such as fixed pegs, currency unions, a combination of both—that is, a fixed peg within a currency union (FPCU)—and flexible regimes in both a theoretical and an empirical setup. The focus on this specific relationship does not mean to disregard other equally important determinants of fiscal outcomes, such as institutional factors,2 but the purpose is to demonstrate that the exchange rate regime can also have a very important bearing on fiscal discipline.

The paper develops a simple conceptual framework to show that the scope for free riding in fiscal policy differs under alternative exchange rate regimes depending on the ability of fiscal agencies to postpone the inflationary costs of overspending over time, or transfer it to other countries. In fact, under some conditions, a FPCU can induce the largest incentive to free-ride, while a flexible regime the least. Empirically, we show that fiscal stances worsen where there is more scope for free riding. We analyze the factors underlying fiscal efforts in 15 Caribbean countries after controlling for institutional and other macroeconomic factors, and find that FPCU regimes are indeed associated with exacerbated free-riding behavior in fiscal policy, whereas flexible exchange regimes are not.

The conceptual framework developed in the paper draws on past studies that have analyzed, separately, the impact of fixed exchange rate regimes and currency unions on fiscal policies (Tornell and Velasco, 1995, 2000; Sun, 2003; Beetsma and Bovenberg, 1999; and Chari and Kehoe, 2004). These two separate strands of work are integrated to derive the conditions under which the scope for free-riding intensifies under a FPCU relative to that under other regimes. Specifically, we show that under certain conditions, including the presence or expectation of fiscal dominance, a FPCU can allow member governments to transmit costs of fiscal overspending—that is, the inflation tax—across time (to future governments) as well as space (to other countries within the union), and thus generate scope for greater fiscal indiscipline compared to other regimes. Conversely, the scope for such free riding does not arise under flexible regimes, given the immediate inflationary impact of overspending and the inability of fiscal agencies to share the costs with other countries.

Empirically, the paper addresses a key weakness in existing studies in a cross-sectional setup (Tornell and Velasco, 2000; Fatas and Rose, 2000). These cross-sectional analyses document the relation between the choice of exchange rate regimes and fiscal stance, but are not convincing in establishing the presence of causality or its direction: first, because they do not take into account possible unobservable heterogeneity; and second, because they disregard any potential simultaneity problems, in that fiscal performance could be influencing the choice of exchange rate regimes. Our empirical analysis controls for unobservable heterogeneity by using a panel fixed-effects regression. Also, it reduces potential simultaneity problems by identifying specific channels of free-riding opportunities that could affect fiscal behavior differently under alternative exchange rate regimes.

The sample of countries is motivated by a number of compelling factors. First, the Caribbean region has a variety of exchange rate arrangements—from the fixed peg currency union (FPCU) in the Eastern Caribbean Currency Union (ECCU) to more flexible regimes— allowing one to compare fiscal stances across the spectrum of exchange rate regimes over time.3 Second, all the countries in the region are developing countries in which fiscal dominance—a crucial component in our theoretical setup—cannot be ruled out. Third, the generally high political and institutional stability in the Caribbean relative to many other developing countries makes fixed-effects estimation (which assumes stable country specific structural characteristics over time) an acceptable empirical approach. Fourth, given its high exposure to exogenous shocks (e.g., natural disasters, global shocks to tourism, erosion of trade preference for agriculture exports), the Caribbean region serves as an ideal platform to assess whether fiscal expansions in some countries were merely in response to GDP downturns or indeed reflected fiscal indiscipline or free-rider problems.

The empirical results confirm the presence of exacerbated free-riding behavior in fiscal policies under the FPCU regimes. The data indicates that among the 15 Caribbean countries, fiscal outcomes in the countries with the fixed peg currency union (FPCU) regime were on average the worst, followed by countries with fixed or crawling pegs; while countries with flexible regimes were the best fiscal performers. Most importantly, the regression analysis reveals important results with respect to the presence of free-riding in fiscal behavior under alternative exchange rate regimes. Fiscal policies under both FPCU and other pegged regimes suffer from “intertemporal” free riding arising from the ability to postpone the potential costs of fiscal overspending (that is, the inflation tax) given the fixed exchange rate, while fiscal performance under flexible regimes is not affected by intertemporal free riding. In addition, fiscal stances deteriorate with increasing systemic importance for countries with the FPCU regime. Finally, improved bailout capacity of the currency union central bank is associated with a worsening of fiscal outcomes under FPCU regimes, and this is not the case in the countries which do not belong to a currency union. These results are robust to a number of sensitivity tests.

The rest of the paper is organized as follows. Section II provides a brief summary of the related theoretical literature and draws on it to show how fiscal incentives can be distorted under alternative regimes. Section III describes some stylized facts on the nature of national fiscal policies, then presents the estimation analysis. Section IV concludes.

II. Fixed Exchange Rates, Currency Unions, And Fiscal Discipline

This section first reviews the theoretical literature on the relationship between fiscal discipline and exchange rate regimes. It then integrates the relevant studies and analyzes the nature of fiscal incentives arising under alternative exchange rate regimes, with a particular focus on the FPCU regime.

A. Literature Review

The existing literature has analyzed both the “traditional” and the “unconventional” association between fiscal incentives and exchange rate regimes. Earlier studies supported the traditional view that a fixed exchange rate is an effective policy for fiscal discipline, since fiscal profligacy is deterred by the risk of losses in foreign reserves or buildup of public debt, resulting ultimately in a costly abandonment of the peg.4 However, country experiences with realignment or collapse of fixed exchange rates caused in part by fiscal deterioration (for example, the CFA franc zone in January 1994 and Argentina in December 2001) have time and again questioned the conventional wisdom.

Recent studies have shown that the conventional view can be overturned by explicit consideration of fiscal incentives induced by the exchange rate regime. Tornell and Velasco (1995, 2000) show that fiscal discipline is not always maintained under a fixed exchange rate. The authors assume that a government can finance fiscal deficits by issuing debt for a temporary period, but eventually has to rely on the inflation tax (as with Krugman, 1979). Thus, different exchange rate regimes influence fiscal incentives differently, depending on when observable costs start to bite. Under a fixed exchange rate, observable costs will not materialize until inflation takes place at some time in the future. Conversely, under a flexible regime, inflation is observed in the present, owing to the consequence of anticipated future inflation (in the spirit of Sargent and Wallace, 1981). If governments are shortsighted and dislike inflation, they spend more under fixed exchange rates, as they can postpone the costs of higher spending.

In a similar vein, Beetsma and Bovenberg (1999) and Chari and Kehoe (2004) show that fiscal discipline is not necessarily upheld under a currency union. In their model, the supranational central bank faces a tradeoff between the benefits of greater debt deflation and the output costs of higher inflation, and reneges on its commitment to low inflation when the benefits exceed the costs. 5Consequently, decentralized fiscal authorities in a currency union have the incentive to overspend, given that the benefits of spending accrue solely to its own country while the inflation cost of higher fiscal deficits can be shared with other members of the union.

Thus, the combination of the two exchange rate arrangements—a fixed peg within a currency union (FPCU)—can indeed give rise to perverse fiscal incentives, a fact that has not been explicitly explored before.6 Under the traditional setup, fixed rates and currency unions reinforce each other, making the monetary arrangement an ideal environment for fiscal discipline. However, considering also the elements of the alternative view, the scope for free riding can intensify. Following Tornell and Velasco (2000) and Sun (2003), and assuming that: (i) there are no enforceable rules for fiscal deficits and no policy coordination between member governments; (ii) governments eventually rely on inflationary financing of fiscal deficits; (iii) and governments are biased toward spending and are shortsighted, that is, they discount the future more heavily than the present, then fiscal policies under a FPCU can induce greater free-riding opportunities by allowing a member government to transmit costs of fiscal slippages to the future and to other member governments. These arguments are formalized in the following subsection. Readers who are mainly interested in the intuitive explanation can skip this subsection and move to the next one that summarizes the main results of the formal analysis shown here.

B. The Conceptual Framework

This subsection provides the theoretical underpinnings that support the case that the scope for free-riding in fiscal policy—under some conditions—is present in all fixed exchange rate regimes (while not so under flexible regimes), and is maximized under a currency union with a fixed exchange rate (FPCU). 7

A simple conceptual framework is developed, drawing on Tornell and Velasco (2000) and Sun (2003), which allows a clear understanding of the optimal level of government spending under different exchange rate regimes.8 The following propositions will be proved with respect to free-riding behavior in fiscal policies:

  1. “Pure regional free-riding”: g flex,cu > g flex,ic, akin to a “tragedy of commons” problem.

  2. “Pure Intertemporal free-riding”: g fix,ic > g flex,ic

  3. “Exacerbated free-riding”: (a) g fix,cu > g fix,ic and (b) g fix,cu > g flex,cu,

where,

  • g flex,ic is the net present value of the level of government spending in a country with a flexible exchange rate regime;

  • g flex,cu is the net present value of the level of government spending in a country in a currency union with a flexible exchange rate regime;

  • g fix,ic is the net present value of the level of government spending in a country with a fixed exchange rate regime; and

  • g fix,cu is the net present value of the level of government spending in a country in a currency union with a fixed exchange rate regime (FPCU).

The environment

The model takes place in two periods in a world with perfect capital mobility and price flexibility. It assumes an economy in which a set of n identical countries populated by identical individuals pursue decentralized fiscal policy with no enforceable rules. The policy makers enjoy spending but dislike inflation. The countries belong to a currency union characterized by fiscal dominance, that is, with a common regional central bank that eventually accommodates fiscal authorities and distributes seignorage revenue.9 The relevant functional forms, timing assumptions and institutional framework are standard in this literature and chosen for ease of computation.

The model is based on some convenient assumptions. First, the household problem can be solved independently from that of the public authorities. Second, the solution for government spending—the main focus of this analysis—can be solved analytically and independently from budget constraints and equilibrium conditions. Third, the propositions (I to III) can be proved without explicit consideration of the strategies of other countries.

Households

The representative household (same across the currency union) is assumed to receive a constant endowment y and a transfer from its own government gi. It has to pay inflation tax on money holdings and consumes the only good in the economy, for which the law of one price and a unit international price is assumed (P=E). It is able to save through an internationally traded bond fi,t or by holding nominal domestic currency Mi,t. Real money holdings are defined as mi,t=Mi,t/Et.

In period 1 the household holds assets mi,0 and fi,0, which are chosen at the end of the previous period. It receives the endowment and government transfer gi,1; receives (or pays) interest, r, for bond holdings; and consumes ci,1. It also decides the amount of assets to carry over to period 2 (mi,1 and fi,1). In period 2 it might accumulate assets (debt) by choosing a lower (higher) ci,2 than the sum of endowment and transfers it receives. The budget constraint faced by the household in each period is given by10, 11:

Period 1:(fi,1+mi,1fi,0mi,0)=y+gi,1+rfi,0ci,1π1mi,0(1)
Period 2:(fi,1mi,1)=y+gi,2+rfi,1ci,2π2mi,1(2)

Summing up over both periods and rearranging yield the following intertemporal budget constraint, in which initial liabilities (left-hand side) are equal to the present value of the surpluses (right-hand side):12

(1+r)(fi,0+mi,0)=y+gi,1+y+gi,21+rci,1(i1mi,0)ci,2(i2mi,1)1+r(3)

The household’s utility function is defined as:

u(ci,1)+(εε1)(mi,0)(ε1)/ε+β[u(ci,2)+(εε1)(mi,1)(ε1)/ε](4)

where, u(c) has the standard properties, the discount rate β is equal to the world interest rate, and ε is a parameter which is between 0 and 1.13

The representative household chooses ci,1,ci,2, mi,0, mi,1 and maximizes the objective function (equation (4)) subject to the budget constraint (equation (3)), for which the associated Lagrange multiplier is λ. The first order conditions follow:

u(ci,1)=β(1+r)u(ci,2)=λ(5)
mi,01/ε=λi1(6)
mi,11/ε=λi2.(7)

It is assumed that β(1 + r) = 1 and λ is normalized to 1. If the government transfers all resources to the household, the solution for the above equations is:

c1=c2=c=f(y).

While money demand becomes:

mi,j1=mj1=ii,jεj=1,2.(8)

Note that money demand is independent of transfers received from the local government, so the country subscript can be dropped.

Regional central bank

The central bank is assumed to be able to commit to price stability in period 1 but to abandon its commitment in period 2 and provide seignorage to monetize budget deficits, and that the central bank provides for each government according to its needs to remain solvent.14 The basis for this assumption can be substantiated both from a theoretical and empirical stand point. Theoretically, inflation is seen as the result of the tradeoffs of the central bank between benefits to fiscal accounts versus costs of output decline.15 Even in regimes with strong commitment to a peg where the likelihood of inflation may appear to be low, at some point the benefits of inflation will reach a threshold when it is optimal to inflate.16 Moreover, country experiences have proven that currency crises can take place (and inflation occur) even when the central bank had apparently no incentive (and not even legal capacity) to devalue (e.g., Argentina). Fears of fiscal insolvency usually spur self-fulfilling mechanisms resulting in a widespread sudden plunge in the demand for government liabilities, including the currency. 17

Under fixed exchange rates the monetary authority controls the exchange rate, and given the law of one price, it controls inflation of the union as a whole. It will set π1 = 0, and π2 will be determined by the aggregate government budget constraint. Under flexible exchange rates, it controls the growth rate of money supply (μ) and thus has no direct control on inflation.18 It will set μ1 = 0, and μ2 will be determined by the aggregate government budget constraint.

Fiscal agents

Government operations comprise giving transfers to the citizens of its country and collecting inflation tax from them. They can also incur debt (b), which has to be repaid by the end of the second period. In the context of a currency union, government budget constraints must hold in the aggregate.19 The budget constraint is:

(i=1n(bi,1bi,0)+ i=1n(mi,1mi,0)=i=1ngi,1+ri=1nbi,0π1i=1nmi,0(9)
(-i=1nbi,1i=1nmi,1)=i=1ngi,2+ri=1nbi,1π2i=1nmi,1(10)

Rearranging and using money demand equation (8), we get the traditional expression where the initial liabilities (left-hand side) are equal to the net present value of surpluses (right-hand side).

(1+r)(i=1n(bi,0+mi,0))=nm011/ε+nm111/ε/(1+r)i=1ngi,1i=1ngi,2/(1+r).(11)

Note that a social planner who cares exclusively about the welfare of its citizens would choose government spending equal to zero. Government transfers have no effect on the level of consumption, but create inflation, which is socially costly. Three distortions are introduced, which give the analysis more realism in a developing country setup and makes the decision on spending meaningful and contingent on different regimes.

The first fundamental distortion is that the fiscal authorities care not only about the utility of their constituencies but also about their own transfers, which are assumed to improve the chance of being re-elected. A second distortion is that governments are shortsighted, that is, the discount rate is higher than the interest rate, (i.e δ[1 + r] < 1). The third distortion, which provides a rationale for the existence of this harmful form of government, is that the private sector cannot create fiat money. The utility function of the government can then be expressed by:

αu(gi,0)+(1α)u(mi,0,ci,1)+δαu(gi,1)+(1α)u(mi,1,ci,2),(12)

where α represents the weight governments assign to their own spending, and

u(m,c)=u(c)+(εε1)(m)(ε1)/ε.

Solution under fixed exchange rates

Under fixed exchange rates by construction, π1 = 0 and thus m0=r. Therefore, inflation in the first period is outside government influence. However, the government indirectly affects π2 (and consequently m1), as higher spending would result in higher inflation in period 2. The fiscal authorities’ problem becomes maximizing the objective function (equation (12)), subject to solvency condition (equation (11)), for which the associated Lagrange multiplier is ψ. The first-order conditions yield:

u(gi,1)=δ(1+r)u(gi,2)=ψ(13)
δmi,11/ε=ψn(1ε)(1+r)εmi,11/ε.(14)

The left-hand side of equation (14) reflects the costs of additional spending (a reduction in welfare associated with higher inflation and less money holdings) and on the right-hand side reflects the benefits of it: marginal utility times the amount of spending in terms of inflation tax. Note that money balances nicely cancel in this type of utility function, which allows one to solve for government spending analytically while disregarding the budget constraint and equilibrium conditions. The fundamental implication is that decisions of individual governments do not depend on the actions of other governments, that is, the model has a single possible solution. Solving these two equations yields:

u(gi,1fix,cu)=δ(1+r)εn(1ε).(15)

If n =1 (individual country):

u(gi,1fix,ic)=δ(1+r)ε(1ε),(16)

and given that u(g)<0gi,1fix,ic<gi,1fix,cu.

where g is the intertemporal government spending of a government:

g=gi,1+gi,21+r

Solution under flexible exchange rates

The central bank exogenously fixes money growth 1 = 0, μ2). Thus, both π2 and π1 become endogenous, for which the system requires an extra equation. The following identity holds:

m1=m0(1π1)(17)

Using equation (8) and rearranging yields:

m1=m0(1+r)m0(ε1)/ε(18)

The fiscal authorities’ problem amounts to maximizing the objective function (equation (12)), subject to the solvency condition (equation (11)) —for which the associated Lagrange multiplier is ψ—and money dynamics (equation (18)), with the associated multiplier θ. The first-order conditions are (equation (13)) and:

δmi,11/ε=ψn(ε1)(1+r)εmi,11/ε+θ(19)
mi,01/ε=ψn(1ε)εmi,01/ε+θ[(1εε)m01/ε+(1+r)](20)

Combining equations (19) and (20) yields:

ψ=δ(1+r)ε(1ε)nΛ(21)

where, Λ=[mi,01/εδ(1+r)(1ε)ε(1+r)mi,01/εmi,11/ε+mi,11/εmi,01/ε(1ε)ε(1+r)mi,01/εmi,11/ε+mi,11/ε]

First, note that

u(gi,1flex,cu)=δ(1+r)εn(1ε)Λ<δ(1+r)ε(1ε)Λ=u(gi,1flex,ic)

In this case, the analytical solution for government spending depends on the level of inflation and thus of the level of money balances. However, it is clear that for any level of those variables it could be proved that the marginal utility under “flexible” exchange rates is bigger than in the “fixed” exchange rate case.

Given that u(g)<0gi,1flex,cu>gi,1flex,ic.

Using equation (13),gflex,cu>gflex,cuPropositon I

Also, given the assumption that δ(1 + r)<1, it is clear that Λ > 1. Thus:

ψflex=u(gi,1flex,cu)>δ(1+r)εn(1ε)=u(gi,1fix,cu)

which implies gi,1fix,cu>gi,1flex,cu.

Using equation(13),gfix,cu>gflex,cu.Proposition III(b)

It is straightforward that u(gi,1flex,ic)>δ(1+r)ε(1ε)=u(gi,1fix,ic)

which implies gi,1fix,ic>gi,1flex,ic.

Using equation(13),gfix,ic>gflex,ic.Proposition II

Appendix I presents some interesting extensions to this simple model, which also has implications for the variables used to proxy for free-riding in the empirical section.

C. Summary of the Theoretical Results

Table 1 illustrates how fixed exchange rate regimes and currency unions can spread the burden of the inflation tax across time and space, and hence induce fiscal incentives that are at odds with the conventional wisdom. Four cases are highlighted:

Table 1.

Allocation of the Inflation Tax Under Alternative Exchange Rate Regimes

article image

Case I, represented by the upper-left panel of Table 1, shows the situation when a country has a flexible exchange rate regime. Fiscal overspending would be translated into depreciation of the exchange rate and inflation in the same period as demand for money decreases in anticipation of future inflation. This is the benchmark case with no free-riding in fiscal policy.

Under Case II, a country is a member of a currency union that operates a flexible exchange rate. While fiscal overspending would generate costly present inflation, this is now shared with all union members. This case is labeled as “regional free-riding” since the costs of spending are diluted for the country undertaking fiscal expansion.

Under Case III, a country has a fixed exchange rate. In this case, future inflation does not lead to present inflation as the current exchange rate is fixed. Deferring the costs of the inflation tax amounts to free-riding on future governments by spending today, a phenomenon that can be called “intertemporal free-riding.”

Finally, under Case IV, the common currency of the union—adopted by all union members—is fixed vis-à-vis a major international currency (that is, the FPCU regime). The outcome in this case follows naturally from the other three cases. Actual inflation or even the probability of higher inflation in the future has no consequences for money demand or inflation today, given the fixed exchange rate. Inflation is expected at some point in the future, and the cost is expected to be shared by future member governments, given the currency union. Thus, the inflationary costs of fiscal expansion are minimal at present— future governments end up bearing them and member governments end up sharing them. Consequently, incentives for fiscal slippages at present are the highest.

III. Empirical Analysis

This section explores the empirical validity of the conclusions of the conceptual framework by presenting some stylized facts and then undertaking a more formal regression analysis. The sample comprises 15 Caribbean countries over 1983–2004. The sample starts in 1983 because of lack of data on the fiscal stance prior to that year. The Caribbean featured a unique spectrum of exchange rate regimes during the sample period, including fixed pegs, various forms of flexible regimes, and one of the only two “currency union cum fixed peg” regimes in the world.20

A. Stylized Facts

The data suggests that fiscal policies in the sample countries were in line with the predicted outcome of the conceptual model (Figure 1). In particular, the average primary deficit of the six ECCU countries with a FPCU regime was the worst in the Caribbean during 1990–2004 (a sufficiently long period over which short-run determinants of fiscal policy can be expected to net out), followed by countries with fixed or crawling peg regimes. The countries with various forms of flexible regimes—including floats—were the best fiscal performers in the sample.21

Figure 1.
Figure 1.

Fiscal Stances of Caribbean Countries, Average 1990–2004

(primary balance in percent of GDP)

Citation: IMF Working Papers 2006, 119; 10.5089/9781451863796.001.A001

Source: Authors’ calculations.

The deterioration of primary balances in the six ECCU countries with the FPCU regimes was mainly due to a worsening in government expenditures, which increased sharply during the 1990s (Figure 2, Panel (a)). Fiscal expenditure growth generally surpassed GDP growth irrespective of the nature of the business cycle (Panel (b)), suggesting that fiscal stances were influenced by other factors besides the growth slowdown. The rise in primary expenditure over time characterized every ECCU country, and in each case, exceeded the increase in fiscal revenue during the same period (Panel (c)). Also, the composition of primary spending did not change in a major way, implying that fiscal policies were not driven by a sharp rise in government preference towards a particular item (Panel (d)).

Figure 2.
Figure 2.

Eastern Caribbean Currency Union: Nature of Fiscal Stance, 1990–2004

Citation: IMF Working Papers 2006, 119; 10.5089/9781451863796.001.A001

Sources: Eastern Caribbean Central Bank; ECCU member country authorities; and Fund staff estimates.

The ECCU governments had access to foreign financing even when other emerging market countries faced a turnaround of net capital inflows (Figure 3). Also, unlike other developing countries, where capital flows are usually procyclical, non-FDI capital inflows continued to the ECCU countries even during periods of low economic activity.22 Possible reasons for their ability to borrow externally could be their good repayment record, relatively low GDP volatility, the perception that the central bank of the currency union would serve as a lender of last resort in the event of potential liquidity shortages faced by member governments, and the gradual elimination of transaction costs with financial innovation in capital markets over time.23

Figure 3.
Figure 3.

Private Net Capital Inflows (less FDI)

(In percent of GDP)

Citation: IMF Working Papers 2006, 119; 10.5089/9781451863796.001.A001

Sources: Country authorities; WEO, and authors’ calculations.

The relatively poor fiscal performance of pegged regimes in the Caribbean, and in particular in the FPCU regimes, supports related empirical findings. Sahay (2005) analyzes the public debt dynamics of a sample of 15 Caribbean countries and finds that the ECCU countries are among the most highly indebted emerging market economies. The author shows that most of the increase in public debt is accounted for by a deterioration in primary balances.24 Alberola and Molina (2004) find in a sample of 32 emerging market economies that countries with fixed exchange rates during 1990–2001 were worse fiscal performers compared to those with flexible regimes. Tornell and Velasco (2000) find in a sample of countries in Africa that countries with the CFA Franc FPCU regimes had higher fiscal deficits than those with flexible regimes.

B. Estimation Results

The estimation method used to examine the presence of free-riding in fiscal stances of Caribbean countries represents a marked improvement over previous studies on the same issue. Past studies (such as Fatas and Rose, 2001; and Tornell and Velasco, 2000) have generally used cross-section regression techniques to estimate this relationship between fiscal stance and exchange rate regimes. Two key weaknesses of past studies have been the inability to take into account unobserved heterogeneity, and not satisfactorily addressing potential simultaneity problems. Panel data analysis helps tackle unobservable heterogeneity problems—for instance a fixed effects estimation allows one to estimate the response of fiscal policy to changes in economic conditions under the prevailing exchange rate regimes after taking into account country-specific, time-invariant characteristics. 25Moreover, the problem of potential simultaneity is avoided by identifying specific proxies for free-riding opportunities that can arise under alternative exchange rate regimes, rather than attempting to establish a direct one-on-one relationship from exchange rate regimes to fiscal performance.

Two alternative estimation methods are used to investigate the factors underlying fiscal stances in the Caribbean. First, we estimate fiscal stances using a simple ordinary least squares (OLS) regression with binary dummy variables corresponding to alternative exchange rate regimes as right-hand side variables, to assess whether fiscal stances are associated with alternative regimes in any significant way. This method suffers from the same problems identified above as the key weaknesses in past empirical methodologies. Then, fiscal policies are estimated using a fixed effects (FE) estimation method with specific time-varying proxies of intertemporal and regional free-riding behavior under alternative exchange rate regimes, which aims to address both the weaknesses of the OLS estimation method.

The estimated equation has the following form:

yit=α+βxit+γzit+vit

where:

(i) yit is a measure of fiscal stance of country i at time t, expressed as the primary fiscal balance (expressed as a percentage of GDP). Since the primary balance is unaffected by interest payments on accumulated public debt, it serves as an appropriate indicator of fiscal policy stance;

(ii) xit comprises a number of control variables for country i at time t, the description of which (and their expected signs in the regression) is given in Box 1;

(iii) vt is the error term in the regression; and

(iv) under the OLS regression, zit comprises dummy variables corresponding to the pegged and FPCU regimes. Under the FE regression, zit is a group of three indicators representing proxies for intertemporal and regional free-riding as described below.

Control Variables Used in the Regression26

  1. Economic performance, measured by the annual real GDP growth rate.

  2. Trade openness, expressed as the sum of exports and imports of goods and services as a percentage of GDP, as a proxy for trade policies.

  3. Terms of trade, measured by the ratio of export price to import price, in dollars. Improvement in the terms of trade would improve fiscal revenues, reduce the need for expansionary fiscal policy, and help improve the primary balance.

  4. A dummy for an IMF program controls for the effect of existing IMF programs on the fiscal stance.

  5. Time dummies, to control for time-specific events and also account for innovations in the financial markets over time that ease borrowing constraints for member governments.

Intertemporal free-riding is proxied by the closeness to election under alternative exchange rate regimes. The conceptual framework established that the more shortsighted the government is, the more incentives it will have to spend under fixed exchange rates (see proof of Proposition II in previous section.). Elections are natural situations when governments’ shortsightedness generally increase, that is, the closer the elections, the more governments care about the present—in which the chances of winning the election are decided—and the more they discount the future. However, a fixed or pegged exchange rate regime can conveniently postpone the costs of fiscal overspending to the future, while under flexible regimes the costs would have to be paid upfront. Thus, the tendency to free ride might be inversely related to the time to election under fixed regimes.

Three variables are used to fully explore the impact of all exchange rate regimes on intertemporal free-riding: (i) the product of the time remaining to the next election and a dummy for all ECCU countries to capture the effect under the ECCU; (ii) the product of the time remaining to the next election and a dummy for countries that maintained fixed peg regimes to capture the effect under these exchange rate regimes; and (iii) the time remaining to the next election in years for all other regimes to assess the effect under flexible regimes. In the presence of intertemporal free-riding, there would be a negative relationship between fiscal stance and proximity to election for all countries with fixed exchange rates, including the ECCU, and no such relationship for countries with flexible regimes.

Regional free-riding is proxied by the level of official foreign reserves, relative to base money under different exchange rate regimes. While countries not belonging to any currency union have access to external reserves only at their central banks, each ECCU member country has access to the entire pool of foreign reserves at the Eastern Caribbean Central Bank (ECCB).27 In the presence of regional free-riding, the increase in foreign reserves at the ECCB would induce fiscal slippages, resulting in a worsening of fiscal balances in ECCU countries. In the non-ECCU countries, foreign reserves are not expected to have a negative bearing on fiscal stances.

Two variables are used in the regression to explore the impact of all exchange rate regimes on regional free-riding: (i) the product of a dummy for all ECCU countries and the level of foreign reserves at the ECCB (as a percent of reserve money) that captures the effect under the quasi currency board arrangement; and (ii) the level of foreign reserves under all other exchange rate regimes.

The relative size of an ECCU member country, reflecting its systemic importance in the ECCU, is used as an alternative proxy for regional free-riding. However, as shown in the theoretical model, the relationship between this proxy and fiscal stance is ambiguous. On the one hand, the more systemically important a country becomes, the greater can be the perceived prospects of being bailed out by the regional central bank to maintain the stability of the FPCU.28 On the other hand, the expectation of being bailed out could be seen to be higher if a country is small, since the associated costs are relatively small.

C. Results

The estimation results of the OLS regression indicate that both, fixed peg and FPCU regimes adversely impact fiscal stances (Table 2, column (a)). After controlling for time specific effects and the impact of other macroeconomic variables on fiscal stances, fiscal stances appear to be worse in countries with fixed peg and FPCU regimes.29

Table 2.

Determinants of Fiscal Policy in the Caribbean, 1983-2004

article image
Source: Authors’ calculations.

Each coefficient represents the impact of a change in a given explanatory variable on the fiscal stance in percentage. The parentheses contain probability values. Results that are statistically significant at 5 percent and 10 percent are marked by “**” and “*” respectively.

Coefficients for the statistically significant time dummies are presented only.

With the FE regression, five important results are obtained. First, fiscal policies in the fixed peg and FPCU regimes are significantly influenced by intertemporal free-riding (Table 2, column (b)). In other words, for all countries with fixed peg regimes, including the ECCU members, fiscal stances worsen as the election year draws closer, reflecting that the cost of fiscal expansion is deferred to the future. Second, the negative association is not observed between fiscal stances and proximity to election under flexible regimes, reflecting that the immediate inflationary consequences of fiscal expansion under flexible regimes deter free-riding. Third, fiscal policies under the FPCU regime are also affected by regional free-riding. In particular, fiscal stances of the ECCU countries worsen with an increase in foreign reserves at the ECCB, consistent with the expectation of being bailed out rising with an improvement of the bailout capacity of the central bank. Fourth, the negative association between the central bank’s foreign reserves and fiscal stance is not observed for countries that do not belong to a currency union. Fifth, fiscal stances under the ECCU also worsen as the relative size of a member country in the ECCU rises, confirming that countries’ expectation of being bailed out rises with an increase in their systemic importance in the union.

A positive relationship is observed between primary balance and the prevalence of an IMF program, implying that countries that adopted fiscal reforms under IMF supported economic programs were able to improve their fiscal stances.

Finally, the results show that fiscal policies in Caribbean countries deteriorated significantly since the late-1990s. Four of the time dummies from 1998–2004 are individually significant in worsening the fiscal balances of Caribbean countries. Also, the hypothesis test at the bottom panel of Table 2 confirms the joint significance of the years after 1998 in adversely affecting fiscal stances. A possible explanation could be that with innovations in financial markets, Caribbean countries had better access to external financing, which exacerbated their fiscal imbalances.

D. Robustness Tests

The majority of results were robust to a number of sensitivity tests, which were as follows.

Additional control variables

Natural disasters—dummies corresponding to the year in which a natural disaster hit a particular country⎯were added to assess whether fiscal effort was significantly affected by these exogenous shocks (Table 3, column (a)).30 Table 3 presents the statistically significant dummies only. Note that some of the natural disaster dummies had a positive effect on fiscal balance.31 The only natural disasters that adversely affected fiscal balances were those in Antigua and Barbuda (1995), and St. Vincent and the Grenadines (1992). Also, the five key results corresponding to fiscal free-riding continue to hold, that is: presence of inter-temporal free-riding under both FPCU and fixed peg regimes; absence of intertemporal free-riding under flexible regimes; deterioration of fiscal stances with increase in bailout capacity of the regional central bank and with increase in size of the member country in the ECCU (both results reflecting the presence of regional free-riding); and the absence of regional free-riding in countries that do not belong to a FPCU.

Table 3.

Determinants of Fiscal Policy in the Caribbean, 1983-2004

article image
Source: Authors’ calculations.

Each coefficient represents the impact of a change in a given explanatory variable on the fiscal stance. The parentheses contain probability values. Results that are statistically significant at 5 percent and 10 percent are marked by “**” and “*” respectively. Coefficients on time dummies are not shown.

Other controls—e.g., availability of external financing (proxied by total private sector capital flows from industrial countries to emerging market economies); world oil prices; world interest rates (proxied by the three-month U.S. treasury bill rate); and institutional development (proxied by real GDP per capita) were also added to the regression but did not have any systematic or significant influence on fiscal policy.

Alternative proxy for fiscal stance

Primary expenditure (as a percentage of GDP) is used as an alternative proxy of fiscal stance (Table 3, column (b)).32 The results on intertemporal free-riding and size-based regional free-riding under the ECCU continue to hold, while the result of intertemporal free-riding under fixed exchange rate regimes does not. Note also that, primary spending does not increase in response to the reserve coverage at the ECCB. Using fiscal revenue as an alternative proxy for fiscal stance, the results show that increase in the bailout capacity of the ECCB reduces fiscal revenues significantly, that is, regional free-riding is manifested through an increase in governments’ laxity in generating fiscal revenues. 33Also, as expected, fiscal policies under non-FPCU regimes are not affected by regional free-riding.

IV. Conclusion

This paper explores how fiscal incentives are distorted under alternative exchange rate arrangements. It first develops a conceptual framework to show that different monetary arrangements provide different degrees of scope for fiscal free-riding behavior, depending on fiscal agencies’ abilities to defer the costs of fiscal profligacy over time (to future governments), or over space (to neighboring governments). The empirical analysis, based on a sample of 15 Caribbean countries during 1983–2004, confirms the conclusions of the conceptual framework. In particular, fiscal free riding is the most intense in countries with fixed pegs under currency unions, followed by countries with fixed pegs; while such free-riding behavior is not evident under flexible exchange rate regimes. The approach used here to analyze the relationship between fiscal and exchange rate policies addresses the problems of sample unobservable heterogeneity and endogeneity issues that have been the main weaknesses of previous empirical studies on this issue.

The results raise an important concern regarding the choice of exchange rate regimes. Indeed, while many countries opt for fixed exchange rate regimes to control fiscal spending, at the same time, many eventually renege on their own implicit commitment on fiscal constraint, and prefer to postpone the costs of spending to the future. Under a FPCU, this result is intensified by the additional dimension of free riding whereby member countries expect to share the costs of country-specific spending with other member countries. Thus, the consequence of choosing restrictive exchange rate regimes backfires, resulting in looser fiscal discipline than before, and often causing costly adjustments.

It is imperative to ensure the consistency of fiscal policies with prevailing monetary arrangements. Under flexible regimes, the immediate inflationary consequence of fiscal overspending is one way to contain fiscal agencies’ urge for overspending. Under fixed and FPCU regimes, in the absence of the stated inflationary cost, other ways to enforce fiscal discipline are critical, especially when an impeccable credit history tend to mitigate possible disciplinary effects of capital markets. Under a currency union, members’ tendencies to free-ride on other member countries can be discouraged by clearly demonstrating the regional central bank’s unwillingness to yield to fiscal dominance and bail out members—either directly by financing fiscal deficits, or indirectly by bailing out banking systems in individual countries—that would over time boost the credibility of its commitment to exchange rate stability. In this regard, the current practice by the Eastern Caribbean Central Bank in the ECCU region of not bailing out governments facing intermittent debt servicing problems (as in Antigua and Barbuda during several recent years and Dominica in 2002) has helped establish the credibility of the regional central bank. Under fixed exchange rates, mechanisms to eliminate fiscal dominance—such as an independent central bank that can build a strong reputation over time—would also eliminate the scope for intertemporal free riding.

APPENDIX I

Some Extensions to the Theoretical Model

We consider here the consequences of introducing simple extensions to the basic conceptual framework that was developed in the text. The extensions do not change the main results of the model in terms of the differences in the scope for free-riding under different exchange rate regimes, but help explore how to relate the main results with the empirical estimation of free-riding in fiscal policies.

Asymmetries in Country Size

The model assumes that all governments are identical and the solution for government spending is symmetric. Now consider how the results might change if country sizes are different. The analysis will reveal two offsetting forces affecting the incentives for spending. On the one hand, larger countries tend to have greater bargaining power with the regional central bank, and thus could sustain higher deficits based on greater access to inflationary financing. On the other hand, large countries have less capacity to dilute the costs of their misbehavior across the union, as they represent an important share of it.

The setup of the problem changes in two different dimensions. First, we assume the representative consumer of each country is scaled by a factor s, which represents the relative size of each country. Thus, the money holdings and consumption are scaled by the same factor and the aggregate government budget constraint becomes:

(1+r)(i=1n(fi,0+mi,0))=i=1nsimi,011/ε+i=1nsimi,111/ε/(1+r)i=1ngi,1i=1ngi,2/(1+r)(22)

Second, the central bank treats countries differently according to their size. Large countries typically have more bargaining power with the central bank, as their importance for economic success of the union as a whole tends to be higher.34 To capture this effect we now introduce the possibility that the central bank decides to extend bailouts only to certain subsets of countries based on their relative size. We are thus going to assume two types of countries: “small” (with no access to central bank bailout) and “large” (with access to central bank bailout).

The solution to government spending for small countries now becomes

gi,1fix,cu,small=0(23)

For large countries, under fixed exchange rates, first-order condition (equation 14) for each country i becomes:

δsimi,11/ε=ψ(1ε)j=1nsj(1+r)εmi,11/ε(24)

Thus, the solution for government spending is:

u(gi,1fix,cu,large)=δ(1+r)sij=1nsjε(1ε)(25)

The solutions show offsetting effects on government spending. On the one hand, the bigger the share of the country in the union (sij=1nsj) the smaller government spending, as there are less free-riding opportunities available. On the other hand, the smaller the share of the country the more likely it will not have access to bailouts, and its government spending financed by the inflation tax would have to be zero.

Role of pooled international reserves

The regional central bank’s bailout decision can be further complicated by introducing the role of international reserves. If international reserves are high enough so that the central bank can extend a bailout without generating inflation in period 2, every individual country will have an extra incentive to spend, given the possibility of a non-inflationary bailout. Even if governments do not expect to be directly bailed out by the central bank, the latter’s commitment to bail out the financial system from systemic crises reduces the urgency for governments to prepare for potential liquidity shortages in the banking system, and also benefits governments that have significant ownership in the banking system. This implies that in times of higher reserves, government spending also increase as fiscal authorities endeavor to exploit this opportunity.

Finally, to approach the data, it is desirable to take into account that the model has assumed that there is full access to capital markets and no uncertainty in debt repayment. However, full free-riding opportunities may not be fully exploited if capital markets imperfections are introduced.

APPENDIX II

Data Sources

Fiscal stance proxies. (i) Primary balance divided by nominal GDP: For the ECCU countries, data for primary balance and GDP during 1983–1990 was obtained from the Eastern Caribbean Central Bank (ECCB), while data after 1990 was obtained from IMF, Western Hemisphere Department. For the non-ECCU countries (except The Bahamas) data was obtained from IMF, Western Hemisphere Department. The data of the Bahamas was obtained from IMF, World Economic Outlook (series GCBXI for primary balance, and series NGDP for nominal GDP); (ii) Primary expenditure, divided by nominal GDP: For the ECCU countries, the primary expenditure series before 1990 was obtained from the ECCB, while that after 1990 was from IMF, Western Hemisphere Department. For the non-ECCU countries the data was from IMF, World Economic Outlook (series GCENL–series GCEI); (iii) Overall balance, divided by nominal GDP: For the ECCU countries, the overall balance before 1990 was obtained from the ECCB, while that after 1990 was from IMF, Western Hemisphere Department. For the non-ECCU countries, the data was from IMF, World Economic Outlook (series GCB).

De facto exchange rate regime. Reinhart-Rogoff (2002) classification of exchange rate regimes, the IMF’s Annual Report on Exchange Arrangements and Exchange Restrictions, various issues, and IMF, Western Hemisphere Department.

Real gross domestic product. For ECCU countries from IMF, Western Hemisphere Department. For the rest of the Caribbean countries from WEO (series W_NGDP_R).

Election dates. From Database of Political Institutions, World Bank.

ECCB foreign reserves coverage was measured by the ratio of foreign assets at the ECCB in terms of reserve money (lines 1L. DZF and 14..ZF in IMF’s International Financial Statistics, IFS). Nominal exchange rate between EC$ and US$ (series AE.ZF in IFS) was used to convert foreign assets of the ECCB in US$ to that in EC$.

Terms of trade. WEO, Series W_TT.

Openness. Defined as the sum of exports and imports of goods and services, divided by nominal gross domestic product. For ECCU countries, these series were obtained from the IFS, series codes 90C..ZF (exports), 98C..ZF (imports) and 99B..ZF (nominal GDP). For rest of Caribbean, the series were obtained from WEO: WEO W_NX (exports), W_NM (imports) and W_NGDP (nominal GDP).

References

  • Abiad, A., and T. Baig, 2005, “Underlying Factors Driving Fiscal Effort in Emerging Market Economies,” IMF Working Paper 05/106 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • Alberola, E., and L. Molina, 2004, “What Does Really Discipline Fiscal Policy in Emerging Markets? The Role and Dynamics of Exchange Rate Regimes,” Banco de España Working Paper No. 0402, Madrid.

    • Search Google Scholar
    • Export Citation
  • Alesina, A., R. Hausmann, R. Hommes, and E. Stein, 1999, “Budget Institutions and Fiscal Performance in Latin America,” Journal of Development Economics, Vol. 59, No. 2, pp. 25373.

    • Search Google Scholar
    • Export Citation
  • Beetsma, R.M.W.J., and A.L. Bovenberg, 1999, “Does Monetary Unification Lead to Excessive Debt Accumulation?Journal of Public Economics, Vol. 74, No. 3, pp. 299325.

    • Search Google Scholar
    • Export Citation
  • Berger, H., G. Kopits, and I.P. Szekely, 2004, “Fiscal Indulgence in Central Europe: Loss of the External Anchor,” IMF Working Paper 04/62 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • Bergin, P., 2000, “Fiscal Solvency and Price Level Determination in a Monetary Union,” Journal of Monetary Economics, Vol. 45, No. 1, pp. 3753.

    • Search Google Scholar
    • Export Citation
  • Calvo, G.A., and F.S. Mishkin, 2003, “The Mirage of Exchange Rate Regimes for Emerging Market Countries,” NBER Working Paper No. 9808 (Cambridge, Massachusetts: National Bureau of Economic Research).

    • Search Google Scholar
    • Export Citation
  • Chari, V.V., and P. Kehoe, 2004, “On the Need for Fiscal Constraints in a Monetary Union,” NBER Working Paper No. 10232 (Cambridge, Massachusetts: National Bureau of Economic Research).

    • Search Google Scholar
    • Export Citation
  • Da Rocha, J.M., E.L. Gimenez, and F.X. Lores, 2002, “Devaluation Beliefs and the Argentinean Debt Crisis” (Unpublished; Vigo, Spain: Universidade de Vigo).

    • Search Google Scholar
    • Export Citation
  • DeBrun, X., 2000Fiscal Rules in a Monetary Union: A Short Run AnalysisOpen Economies Review, Vol. 11, pp. 32358.

  • Duttagupta, R., and G. Tolosa, 2005, “Fiscal Policy in a Regional Currency Union,” in Eastern Caribbean Currency Union: Selected Issues, IMF Country Report No. 05/305 (Washington: International Monetary Fund), pp. 5172.

    • Search Google Scholar
    • Export Citation
  • Fatas, A., and A. Rose, 2001, “Do Monetary Handcuffs Restrain Leviathan? Fiscal Policies in Extreme Exchange Rate Regimes,” IMF Staff Papers, Vol. 47, Special Issue, pp. 4061.

    • Search Google Scholar
    • Export Citation
  • Frenkel, J., M. Goldstein, and P. Masson, 1991, “Characteristics of a Successful Exchange Rate System,” IMF Occasional Paper, No. 82 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • Giavazzi, F., and M. Pagano, 1988, “The Advantage of Tying One’s Hands: EMS Discipline and Central Bank Credibility,” European Economic Review, Vol. 32, pp. 105575.

    • Search Google Scholar
    • Export Citation
  • Kaminsky, G. L., C.M. Reinhart, and C.A. Vegh, 2004, “When it Rains it Pours: Procyclical Capital Flows and Macroeconomic Policies,” NBER Working Paper No. 10780 (Cambridge, Massachusetts: National Bureau of Economic Research).

    • Search Google Scholar
    • Export Citation
  • Krugman, P., 1979, “A Model of Balance of Payments Crisis,” Journal of Money, Credit and Banking, Vol. 11, pp. 31125.

  • Kufa, P., A.J. Pellechio, and S. Rizavi, 2003, “Fiscal Sustainability and Policy Issues in the Eastern Caribbean Currency Union,” IMF Working Paper 03/162 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • Manasse, P., N. Roubini, and A. Schimmelpfennig, 2003, “Predicting Sovereign Debt Crisis,” IMF Working Paper 03/221 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • Rasmussen, T., 2004, “Macroeconomic Implications of Natural Disasters in the Caribbean,” IMF Working Paper 04/224 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • Rasmussen, T., and G. Tolosa, 2005, “Islands of Stability? Determinants of Macroeconomic Volatility in the ECCU,” in Eastern Caribbean Currency Union: Selected Issues, IMF Country Report No. 05/305 (Washington: International Monetary Fund), pp. 1834.

    • Search Google Scholar
    • Export Citation
  • Reinhart, C., 2002, “Default, Currency Crises and Sovereign Credit Ratings,” NBER Working Paper No. 8738, (Cambridge, Massachusetts: National Bureau of Economic Research). Also published in The World Bank Economic Review, Vol. 16, pp. 151–70.

    • Search Google Scholar
    • Export Citation
  • Reinhart, C., K. Rogoff, and M. Savastano, 2003, “Debt Intolerance,” NBER Working Paper No. 9908 (Cambridge, Massachusetts: National Bureau of Economic Research).

    • Search Google Scholar
    • Export Citation
  • Sahay, R., 2005, “Stabilization, Debt and Fiscal Policy in the Caribbean,” IMF Working Paper 05/26 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • Sargent, T., and N. Wallace, 1981, “Some Unpleasant Monetarist Arithmetic,” Federal Reserve Bank of Minnesota Quarterly Review, Vol. 5, No. 3, pp 117.

    • Search Google Scholar
    • Export Citation
  • Sun, Y., 2003, “Do Fixed Exchange Rates Induce More Fiscal Discipline?,” IMF Working Paper 03/78 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • Tytell, I., and S.J. Wei, Does Financial Globalization Induce Better Macroeconomic Policies?IMF Working Paper 04/84 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • Tornell, A., and A. Velasco, 2000, “Fixed or Flexible Exchange Rates: Which Provides More Fiscal Discipline?Journal of Monetary Economics, Vol. 45, pp. 399436.

    • Search Google Scholar
    • Export Citation
  • Tornell, A., and A. Velasco, 1995, “Fixed or Flexible Exchange Rates: Which Provides More Fiscal Discipline?NBER Working Paper No. 5108 (Cambridge, Massachusetts: National Bureau of Economic Research).

    • Search Google Scholar
    • Export Citation
  • van Beek, F., J.R. Rosales, M. Zermeno, R. Randall, and J. Shepherd, 2000, The Eastern Caribbean Currency Union, IMF Occasional Paper 195 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • von Hagen, J., and I. Harden, 1996, “Budget Processes and Commitment to Fiscal Discipline,” IMF Working Paper 96/78 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • Wildasin, D.E., 1997, “Hard and Soft Budget Constraints in Intergovernmental Fiscal Relations,” Mimeo, Vanderbilt University.

  • Williams, O., T. Polius, and S. Hazel, 2005, “Reserve Pooling in the ECCU and the CFA Franc Zone: A Comparative Analysis,” Savings and Development, Vol. 29, No. 1, pp. 3960.

    • Search Google Scholar
    • Export Citation
  • Woo, J., 2003, “Economic, Political, and Institutional Determinants of Public Deficits,” Journal of Public Economics, Vol. 87, pp. 387426.

    • Search Google Scholar
    • Export Citation
  • Woodford, M., 1998, “Control of the Public Debt: A Requirement for Price Stability?,” in G. Calvo and M. King, eds., The Debt Burden and its Consequences for Monetary Policy (New York City: St. Martin’s Press), pp. 11754.

    • Search Google Scholar
    • Export Citation
1

We are grateful to Tobias Rasmussen for his contributions at early stages of this paper and Paul Cashin for his insightful comments. Thanks are also due to Roberto Alvarez, Jingqing Chai, Pablo Druck, Robert Flood, Rishi Goyal, Arnold Harberger, Juan Manuel Jauregui, Amin Mati, Guy Meredith, Prachi Mishra, Garth Nicholls, Patrick Njoroge, Sanjaya Panth, Catherine Pattillo, David O. Robinson, Ratna Sahay, Garnett Samuel, Sebastian Sosa, Carlos Végh, Sir K. Dwight Venner, and Patricia Welsh-Haynes; seminar participants at the Eastern Caribbean Central Bank; the Caribbean Center for Monetary Studies; the IMF’s Western Hemisphere Department; and University of California at Los Angeles for helpful comments and suggestions. We are responsible for any errors and omissions.

3

The ECCU has one of only two FPCU in the world and hence can be analyzed relative to the other exchange rate regimes in the Caribbean. Besides the ECCU, the only other currently operating currency union with a fixed exchange rate regime is the CFA franc zone—comprising the West African Economic and Monetary Union (WAEMU) and the Central African Economic and Monetary Union (CEMAC)—time series data on which is sparse, and hence is not included in our empirical study.

5

In other words, economic policy is characterized by fiscal dominance, a feature more characteristic in developing countries. Conversely, policy making in developed countries is generally characterized by a systematic tendency for fiscal and monetary policy to counterbalance one another (see Debrun, 2000).

6

See Sun (2003), however, for an analysis of fiscal policies in a context of “fragmented policymaking,” that is, many fiscal authorities operating in a single country with a fixed exchange rate, which can be adapted to a multi-country setup. The author shows that if the punishment for abandonment of the peg is high enough, fixed exchange rates might induce more fiscal discipline and inflation may not occur.

7

Note that the only type of free-riding behavior under consideration is with respect to the burden of the inflation tax. Other forms of free-riding, e.g., higher future taxes or lower future social expenditure, are not considered here.

8

The model abstracts from other structural characteristics and institutional factors that also might be affecting fiscal performance under different exchange rate regimes.

9

Under this setup, while monetary policy is not able to precommit to stable prices, fiscal authorities are able to precommit to repay debt. The abstraction from sovereign default risk allows the model to focus on its objective of analyzing the differential consequences on fiscal behavior of different exchange rate regimes. However, even with sovereign default risk in the model, the results of the paper would hold as long as the inflation tax is part of the ultimate policy response to fiscal solvency.

10

Note that πt=(PtPt1)Pt and by the law of one price, devaluation is the same as inflation.

11

The following transversality condition is imposed: fi,2=mi,2=0.

12

Note that i = r +π.

13

This assumption guarantees the economy operates in the upward-sloping side of the Laffer curve.

14

To simplify, the inability of the monetary authority to precommit is introduced mechanically, abstracting from explicit time inconsistency considerations. An important result derived from the latter strategy is in Sun (2003), who shows that if punishment associated with the abandonment of the peg is big enough, fixed exchange rates might induce more fiscal discipline.

15

The beneficial effects of inflation on public accounts are twofold. Tornell and Velasco (2000) stress seignorage revenues deriving from the devaluation, while Chari and Kehoe (2004) stress the deflation of debt in domestic currency.

16

In the context of high foreign currency debt, the threshold is likely to kick in after the foreign currency debt has been defaulted, as Rocha et al. (2002) show.

17

Reinhart (2002) shows that 85 percent of debt crisis are accompanied by currency crisis (and hence inflation), even though her definition of debt crisis does not include bailouts by international financial institutions (as in Manasse et al., 2003). Thus, the correlation between debt crisis and currency crisis could be even higher.

18

μt=(MtMt1)Pt

19

See Woodford (1998) and Bergin (2000) for a justification of why the relevant solvency condition in a currency union is that of the aggregate government.

20

In the sample, the ECCU countries (Antigua and Barbuda, Dominica, Grenada, St. Kitts and Nevis, St. Lucia, and St. Vincent and the Grenadines) maintained their FPCU regime through out the sample period. The Bahamas, Barbados, Belize and Suriname maintained conventional fixed peg regimes through out the sample period. Guyana maintained a fixed peg until 1989, Haiti until 1991, Jamaica until 1990 and Trinidad and Tobago until 1993. These countries abandoned their fixed peg regimes in favor of a variety of more flexible exchange rate regimes, including floats. The Dominican Republic maintained a crawling peg until 2002, when it adopted a floating exchange rate regime.

21

This result also has direct implications for the regression analysis presented in the next sub-section. In particular, even if there were any causality from fiscal performance to exchange rate regime choice, one would expect fixers to be associated with stronger fiscal balances, which is contrary to what is shown in Figure 1.

22

See Kaminsky, Reinhart, and Vegh (2004), and Rasmussen and Tolosa (2005). The higher influx of net capital inflows since the mid-1990s was unrelated to changes in capital account policies, as the region had eliminated most capital controls in the early 1980s (see IMF Annual Report on Exchange Arrangements and Exchange Restrictions, various issues).

23

Reinhart et al. (2003) find evidence that borrowing capacity is significantly related to default histories and the nature of macroeconomic volatilities.

24

Kufa and others (2004) show that fiscal policies in the ECCU have consistently worsened over time, increasing the risk of unsustainability of the public sector debt. Duttagupta and Tolosa (2005) show that the growth in fiscal spending in the ECCU during 1983–2004 generally surpassed GDP growth irrespective of the nature of the business cycle, implying that fiscal stances were influenced by other factors besides growth.

25

The country-specific, time-invariant factors also help proxy for “institutions” data for which is very poor in the Caribbean (e.g., fiscal transparency, characteristics of the budget process, independence of the Ministry of Finance over the Cabinet, the degree of expenditure control by the budget authority). See von Hagen and Harden (1996).

26

The data sources of all the indicators are documented in Appendix II.

27

The reserve pooling agreement in the ECCB implies that no individual country reserves are allocated, but each member has unrestricted access to the common pool of reserves, as long as it has the domestic currency to make it effective (see Williams et al., 2005). Thus, the use of foreign reserves to determine the bailout capacity of the central bank from each member’s perspective appears reasonable.

28

See Wildasin (1997) for a similar argument.

29

To avoid endogeneity between some of the right-hand side explanatory variables (real GDP growth, trade openness, and foreign reserves) with the primary balance, one-year lagged values of the explanatory variables are used.

30

The data on natural disasters during the sample period was obtained from the “EM-DAT” database and comprised disasters including hurricanes, floods, drought, earthquakes, slides, famine, volcano, and mudslides (also see Rasmussen, 2004).

31

This result could have two implications: either, the increase in fiscal spending in response to a natural disaster was accompanied with at least the same increase in grant-financed fiscal revenue; or contrary to expectations, primary expenditure did not increase or was actually compressed during a disaster episode (column (b) provides some support for this argument).

32

Note that the primary spending only summarizes the expenditure side of fiscal stance and hence, is a relatively poor proxy of fiscal policy compared to the primary balance. In other words, a deterioration of fiscal stance is associated with an increase in primary expenditure, only if the former overshoots any increase in fiscal revenue.

33

These regression results are not presented here, but are available upon request.

34

For a model of discipline and bargaining power, see Berger et al. (2004) for the case of eastern European countries.

Fiscal Discipline and Exchange Rate Regimes: Evidence From the Caribbean
Author: Rupa Duttagupta and Mr. Guillermo Tolosa
  • View in gallery

    Fiscal Stances of Caribbean Countries, Average 1990–2004

    (primary balance in percent of GDP)

  • View in gallery

    Eastern Caribbean Currency Union: Nature of Fiscal Stance, 1990–2004

  • View in gallery

    Private Net Capital Inflows (less FDI)

    (In percent of GDP)