Journal Issue

Countries’ Repayment Performance Vis-àVis the IMF

International Monetary Fund. Research Dept.
Published Date:
January 1999
  • ShareShare
Show Summary Details

Aylward and Thorne (1998) investigate sovereign debtors’ repayment performance vis-à-vis the International Monetary Fund. They distinguish through logit analysis between countries that make timely repayments and those that run into arrears. Their conclusion is, inter alia, that the inclusion of IMF-specific financial variables and a small number of macroeconomic variables yields a highly significant econometric model of the probability of a country incurring IMF arrears. Aylward and Thorne also compare the countries’ repayment behavior vis-à-vis the IMF and other creditors, respectively. They find that the number of countries that have incurred protracted arrears to the IMF is much smaller than the number of countries that have failed to service their debt to other creditors and/or had to enter into debt-rescheduling arrangements. In addition, political instability seems to play a more prominent role among the most protracted cases of IMF arrears than it does among countries with non-IMF-specific repayment problems.

Although Aylward and Thorne point out the importance of political and sociopolitical factors for debt servicing, they rely mainly on financial and macroeconomic indicators in their empirical analysis. Past experience of threatened and deliberate debt-service interruptions, especially since the debt crisis in the 1980s, has led to an understandable focus in the literature on debtors’ willingness to honor their debt-service obligations (see the overview in Cataquet, 1985, and Saunders, 1986). Most approaches in that strand of the literature take as their starting point a “social planner” (not specified in detail) who maximizes the country’s expected utility by deciding for or against honoring its debt contracts as a benevolent dictator. In the empirical country-risk literature, payment interruptions are explained using multivariate risk indicators. Their choice does not, in general, reflect a theoretically sound explanation model, but rather plausibility considerations. This may, of course, lead to a misspecification of the econometric model, which is especially problematic if the model is used for credit evaluations.

Financial indicators are the result of past economic conditions and policies as well as having an impact on current conditions and policies. But it is the general policy stance of the government that determines the direction of financial policies. These might, in addition, be influenced by short-term factors like external crises or internal political pressures. A pure ex post analysis of financial and macroeconomic indicators can therefore be misleading if used for credit decisions, even if the identified parameters linking them to the likelihood of arrears are highly significant. Consequently, the sociopolitical background must be included in the picture to check against policy changes, etc., especially if the political structures embodied in time series (past repayment history) are used for ex ante or out-of-sample forecasts (time-specific effects model). Aylward and Thorne (p. 612) take up a similar point with their notion of a “state-dependent design for the model.” But for ex ante forecasts of arrears, a dummy variable for past arrears is only a beginning. A change of government, or even changing moods within a government, can suffice per se to change its debt-servicing stance and should therefore be taken into account. Therefore, credit institutions, insurance companies, regulatory authorities, and rating agencies, which have high value at risk, explicitly include sociopolitical elements in their sovereign-risk models (see overview in Haque, Mathieson, and Park, 1997, and in Bäcker, 1998a, ch. 2). This is despite the fact that standardized economic data are generally more detailed, more timely, and less expensive to obtain than the demographic, social, and political data of some emerging countries.

A Politicoeconomic Approach

The academic country-risk literature has so far been influenced very little by the findings of the political or new political economics departing from Nordhaus (1975). Therefore, the question arises as to how internal political pressure from lobby or interest groups can lead to debt-service problems, for example, opposition to adjustment measures recommended by the IMF or public protest against foreign banks. (See Haggard and Kaufmann, 1989, p. 248, and Nunnenkamp and Picht, 1989, pp. 696–97). Bäcker (1998a and 1998b) introduces a stylized politicoeconomic model that integrates these considerations. It departs from the fact that citizens are affected in many ways by their government’s external obligations and the adjustment measures these imply. In the framework of the balance of payments and excluding asset transactions, the external debt service ultimately has to be raised by primary current account surpluses. That implies a corresponding reduction of internal absorption. Looked at the other way round, the current account surpluses of an open economy entail a surplus of savings over investment. This means that, other things being equal, higher external debt-service payments lead to heavier burdens through taxation, reductions in public investment, social welfare spending, and public services. General political and psychological factors such as status thinking etc. also play a role (Lissakers, 1991, pp. 199–200).

In Bäcker’s model (1998a and 1998b), the population is divided into distinct groups, which assess the advantages and disadvantages stemming from debt service differently. Their subjective reactions determine the individual support or opposition of the group members for or against the government’s (debt-service) policy. This results in an individual net support, which can be translated into a probability of that individual voting for the government’s policy. The government, therefore, chooses a debt-service level that maximizes the overall support for its policy among the population (measured as a weighted sum of support probabilities). Aylward and Thome’s observation of a relatively low probability of countries defaulting on IMF debt versus other debt fits in well with this. This is no doubt owing to the central role that the IMF plays in the international financial system. The IMF’s status means that a default on IMF loans will probably prove far costlier to a debtor country, in terms of direct or indirect sanctions, than will a default on commercial debt; therefore, it will normally be considered only as a last resort.

Debt-servicing capacity can consequently be interpreted as the result of a politicoeconomic evaluation process that integrates the general economic and sociopolitical situation, the individual preferences of each citizen as well as the political optimization calculus of the government. One of this model’s advantages lies, inter alia, in its ability to allow for an analysis of internal policy conflicts. In an extreme case, the politicopsychological state of a society could be so strained that a debt reduction could make an increase in net debt service possible by dampening the opposition. This result implies the existence of a politicoeconomic debt-service Laffer curve, analogous to the original debt relief Laffer curve, which was founded on pure economic terms.

Determinants that have so far been rather neglected in the literature are now highlighted: the costs and benefits of contractual debt obligations as subjectively perceived by different population groups; their individual information and time preference; their evaluation, depending on the opportunity cost of the debt service, which translates the perceived costs and benefits into a measure of government policy support; features that transform the resulting net support into a likelihood of voting for the government’s current (debt service) policy; and the different group sizes or the political influence of population groups with diverging opinions on foreign debt service. This is in line with Aylward and Thome’s finding (p. 603) that a large part of the variation in countries’ repayment behavior is country specific rather than attributable to macroeconomic variables.


    AylwardLynn and RupertThorne1998“Countries’ Repayment Performance vis-à-vis the IMF: An Empirical Analysis,”Staff Papers International Monetary FundVol. 45 (December) pp. 595618.

    BäckerArno1998aPolitische und ökonomische Länderrisikeneine theoretische und empirische Analyse (Berlin: Overseas Publishers).

    BäckerArno1998b“A Politico-Economic View of the Debt-Servicing Capacity of Emerging Economies,”IntereconomicsVol. 33 (iSeptemberOctober) pp. 23037.

    CataquetHarold1985“Country Risk Analysis: Art, Science, Sorcery?”Beihefte zu Kredit und KapitalVol. 8 pp. 77117.

    HaggardStephan and RobertKaufmann1989“The Politics of Stabilization and Structural Adjustment,”in Developing Country Debt and Economic PerformanceVol. 1The International Financial Systemed. byJeffreySachs (Chicago and London: University of Chicago Press) pp. 20954.

    HaqueNadeem UlDonaldMathieson and NelsonPark1997“Rating the Raters of Country Creditworthiness,”Finance & DevelopmentVol. 34 (March) pp. 1013.

    LissakersKarin1991Banks Borrowers and the Establishment: A Revisionist Account of the International Debt Crisis (New York: Basic Books).

    NordhausWilliam D.1975“The Political Business Cycle,”Review of Economic StudiesVol. 42 pp. 16990.

    NunnenkampPeter and HartmutPicht1989“Willful Default by Developing Countries in the 1980s: A Cross-Country Analysis of Major Determinants,”Weltwirtschaftliches ArchivVol. 125No. 4 pp. 680701.

    SaundersAnthony1986“The Determinants of Country Risk: A Selective Survey of the Literature,”Studies in Banking and FinanceVol. 3 pp. 138.

Arno Bäcker is an Economist in the Deutsche Bundesbank’s Department of Economics. This comment is, however, exclusively of a private nature. The author is indebted to Martin Klein for valuable discussions.

Other Resources Citing This Publication