Health & Fitness

Does IMF conditionality benefit lenders?

Does IMF conditionality benefit lenders?
of 12
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
  Does IMF Conditionality Benefit Lenders? Stephen T. Easton*Duane W. Rockerbie**July 1997JEL Code: F34  Abstract This paper assesses the benefits to private lenders from IMF participation and conditionality inrescheduling sovereign debt contracts through the Paris Club. A benefit is defined as an increase in the expectedvalue of rescheduled or new loan contracts via a reduction in the weighted probability of default. Default is defined asthe accumulation of positive arrears in any period. Using a sample of 84 countries, of which 56 rescheduled debtsthrough the Paris Club, over the sample period 1977-87, we find that the immediate provision of IMF liquiditythrough various loan facilities does lower the probability of default by a significant amount. Using a simple model of interest spreads, this translates into an expected decrease in the average spread of between 155 and 179 basis pointsfor a representative sovereign borrower. Borrowers who are in significant liquidity and insolvency trouble willexperience larger reductions. The presence of IMF conditionality does not appear to significantly lower theprobability of default based on loan facilities with differing degrees of conditionality attached. Hence the papersuggests that IMF liquidity is important to private lenders but IMF conditionality programs are not.* Department of Economics, Simon Fraser University, Burnaby B.C. Canada, V5A 1S6** Department of Economics, University of Lethbridge, 4401 University Drive, Lethbridge AB. Canada, T1K3M4  1. Introduction Since the abandonment of the Bretton Woods agreement and the subsequent adoption of flexible exchangerates among most of the developed economies since 1971, many economists and others have questioned the role of the International Monetary Fund (IMF) in international lending. The IMF provides immediate assistance to LDCsthat are unable to meet immediate debt service payments by negotiation through the Paris Club. In fact, the IMF’sprovisions were amended to include this function in 1952 with the formal establishment of stand-by arrangements.Stand-by arrangements are typically short-term loans of not more than twelve to eighteen months (Polak (1991))which are agreed upon by the IMF and the sovereign borrower before formal Paris Club negotiations can proceed toreschedule more sizeable official debts. It is common practice for the more recent London Club of private lenders toinsist on an IMF stand-by arrangement before private debts can be renegotiated. This activity provides a beneficialrole for the IMF since it is likely than many rescheduling agreements could not have been made without IMFparticipation to provide immediate liquidity. 1 In the cases of LDCs with extreme balance of payments problems leading to imminent default, the IMFencourages LDC borrowing from several of it’s other facilities in concert with a stand-by arrangement. TheExtended Fund Facility (EFF) was established in 1974, the Structural Adjustment Facility (SAF) in 1986, and theExtended Structural Adjustment Facility (ESAF) in 1988, all with the purpose of providing liquidity with asomewhat longer maturity period and differing degrees of conditionality than stand-by arrangements. Typically athree to four year maturity period is stipulated for all three facilities. Only low income developing countries canreceive SAF and ESAF loans which carry a nominal interest rate of 0.05% annually. Stand-by arrangements andEFF loans carried market interest rates determined by the riskiness of the sovereign borrower.Lending through IMF facilities requires an annual letter of intent provided by the sovereign borroweroutlining the necessary adjustment policies to insure that future debt service payments to the IMF and othercreditors are met. Discussions of the nature of these conditions is provided in Polak (1991), Mosley (1987) andSpraos (1986). These discussions usually include promoting a current account position consistent with expectedand sustainable capital inflows, a targeted real growth rate and a target rate of inflation. The extent to which theseconditions are enforced differs greatly depending on the loan facility accessed. The least stringent is the stand-byarrangement due to its short term to maturity and the ability of the sovereign borrower to withdraw funds wheneverdesired over the course of the loan. This provides the IMF with little time to monitor the status of the imposed  1 1  Bhattacharya and Detragiache (1984) show formally that joint contracts with private and official creditors effectivelyallows creditor country governments to commit credibly not to bail out problem debtors and to restore the effectiveness of the enforcement mechanism of potential credit rationing. Multilateral agencies, such as the IMF and the World Bank, willoptimally provide subsidies to sovereign debtors who face default, hence providing a role for their participation. Eckaus(1982) suggests that the IMF provides a role as an information collector in Paris Club reschedulings which no otherprivate or official agency could achieve profitably.  conditions and to react to failure to achieve them. ESAF arrangements are enforced using semi-annual performancereviews and installments are paid out semi-annually after the review. EFF loans use the same reviews but conductthem quarterly while SAF loans use annual reviews and installments. SAF loans are considered the most stringentdue to the length of time available to the IMF between reviews and installments. ESAF loans are the next moststringent followed by EFF loans and finally stand-by arrangements. The conditions attached to the loans areusually only as stringent as the ability of the IMF to evaluate their performance hence SAF loans typically carrydetailed conditions while stand-by arrangements carry only vague conditions.If private lenders are rational then they must form an expectation about how IMF liquidity and an attachedconditionality program will improve the ability of the sovereign borrower to meet future debt service payments andconsequently increase the expected value of any existing, new or rescheduled private loan. This is the argumentsuggested by Sachs (1989). The conditions attached to officially endorsed loans makes the loans economicallyviable in the sense of maximizing expected private bank profits. The degree to which private lenders perceive areduction in the default probability will depend on the amount of liquidity provided by the IMF, the conditionsattached to the loan, and the credibility and enforceability of the conditions. Private lenders can “free ride” on IMFarrangements worked out through the Paris Club, either through rescheduling with the same borrowers in theLondon Club, providing fresh loans, or simply benefitting from the improved prospects for existing loans. ThusIMF participation benefits sovereign borrowers through increased liquidity and lower future debt service (Claessensand Diwan (1990)) and provides a positive externality to all creditors.Although the focus of this paper is whether IMF conditions benefit lenders, many attempts have been madeto determine whether IMF conditions benefit sovereign borrowers. Good reviews are Polak (1991), Frenkel andKhan (1990) and Khan (1990a, 1990b). There is a consensus view that IMF conditionality programs have notbenefitted LDC borrowers significantly, and the borrowers should be allowed to submit their own conditions to theParis Club. The statistical analysis usually involves (i) comparing the economic conditions of the borrower beforeand then after the imposition of IMF conditions, taking into account the economic conditions which would haveprevailed had the IMF not intervened; or (ii) determining whether the targets outlined in the conditions wereachieved or not. Both approaches suffer serious methodological flaws. In the first approach, external factors whichaffect the borrower’s position: the terms of trade, world demand, world interest rates, etc., should be held constant.This is not usually done in the studies to date... In the second approach, the borrower’s welfare is measured onlyby the target variables: typically the inflation rate, real economic growth and the current account. Whether thesetargets reflect the welfare of the borrower is unknown. Sometimes the method in which a target is achieved is asimportant as achieving the target. For example, by using import compression via trade restrictions to improve thecurrent account rather than using an exchange rate adjustment has ramifications for both short and long run realincome and welfare..  This paper shall use statistical methods to determine if the imposition of IMF conditions, in concert withthe provision of IMF liquidity, serves to lower the perceived probability of default of a sovereign LDC borrower. Adiscussion of how this could translate into a fall in interest spreads over LIBOR follows. The paper ends with abrief and conclusion section.  2. Modeling the Effects    of Conditionality The relationship between the probability of default and loan pricing has been modeled in a number of papers (Eaton and Gersovitz (1980), Edwards (1984, 1986) and Easton and Rockerbie (1996)). Surveys of earlierattempts can be found in McDonald (1982), Saini and Bates (1984) and Eaton and Taylor (1986). A more recentsurvey is Rockerbie (1993). The approach depends largely on whether a dynamic model is used and on thedefinition of default. Here a simple two-period model will be used to explain the basics and will be fully exploredin the next section. Default is defined as the accumulation of arrears on existing debt. This is not the standarddefinition where default is defined as the occurrence of a debt rescheduling either through the Paris Club or theLondon Club. Here rescheduling is viewed as the avoidance of default which is the non-payment of owed monies.Often sovereign borrowers reschedule debts due to an improvement in the economic outlook of the country whichallows them to seek more favorable terms (Easton and Rockerbie (1996)).Since the probability of default is in the zero to one range, a probit or logit model is required to estimate anexplicit relationship between the default probability and the set of risk indicators X. In our framework , theprobability of default (the presence of positive arrears) is weighted by the arrears ratio a t  = A t  /L t  to represent thescale of default 2 , resulting in a Tobit model of the form y t   = X t β + e t   whereThe set of risk indicators X was chosen on the basis of performance in previous studies. 3  They included thetotal external debt to GNP ratio (DBGNP), the ratio of total external debt service to exports (DSEX), the grossdomestic investment to GDP ratio (IGDP), the real growth rate of GDP (G), the inflation rate computed as theannual percentage change in the GDP deflator (INF), a terms of trade index (1987 = 100, TT), the ratio of foreignreserves to imports (RESIMP) and an index of trade openness computed as exports plus imports divided by GDP(OPEN). The one period lagged arrears to total external debt ratio a t-1  was also included to account for laggedadjustment. Thus the final Tobit specification is  2 2 Easton and Rockerbie (1996) model and explain statistically why this indicator of the expected value of a default is abetter characterization of the economic significance of “arrears” than the simple probability of default. 3 3   A useful survey is contained in Rockerbie (1993), Table 1. More recent studies include Lee (1993), Hernandez-Trillo(1995) and Cantor and Packer (1996). y t ===  001 if if ππ t tt  a  π t a t  = f( π t-1 a t-1 , DBGNP t , DSEX t , IGDP t , TT t  ,G t , INF t , RESIMP t , OPEN t )(1)All data were annual covering the sample period 1977-87 and were obtained from the World DevelopmentIndicators CD-ROM, International Financial Statistics and the World Debt Tables. Post-1987 data were not useddue to the establishment of concessional terms on official lending in 1988 (Toronto terms). These concessionalterms were further enhanced in 1992 (London terms) and again in 1994 (Naples terms). 4 Two dummy variables were also included in the set of risk indicators to account for IMF conditionalarrangements. The dummy variable SB took on the value one whenever an IMF stand-by arrangement was in placefor the particular sovereign borrower, while the dummy variable EFF took on the value one whenever an IMF loanfrom the EFF or SAF facilities was in place. These data were obtained from IMF Annual Reports. In some casesboth dummy variables took on the value one at the same time, but there were enough non-overlapping cases toallow for estimation of each effect.A total of 84 LDCs were included in the sample. They were chosen using the following method. First the65 countries which rescheduled debts through the Paris Club were found in Butte (1995), Table A25. Non-rescheduling countries were chosen by including those countries whose real per capita income fell within the rangefound for the rescheduling countries. This increased the sample size to 112 countries. Data limitations reduced thesample size to 84 of which 56 were rescheduling countries. Of the total pooled sample size of 840 observations,470 observations took on the limiting value of zero for the arrears ratio.  3. Statistical Results To account for fixed effects across countries, the pooled time-series cross-section data were converted todeviations from means for all variables except the dummy variables SB and EFF. This was viewed as a morepractical way to account for fixed effects than to include 83 country-specific dummy variables. The transformeddependent variable was thus π π t t t t  a a −  which no longer has a lower limit value of zero, hence the transformeddependent variable was assumed to have no upper or lower limit allowing the use of ordinary least squares.White’s (1980) heteroscedasticity consistent covariance matrix was estimated to adjust the standard errors of theslope coefficients for an unknown form of heteroscedasticity. The lagged dependent variable, π t-1 a t-1 , will becorrelated with the error term even if the error term itself is not serially correlated (Greene (1997), 640-641), hencean instrumental variable was used in it’s place using π t-2 a t-2  and one period lagged values of all the risk indicatorsin (1) as instruments, except for the dummy variables SB and EFF. The estimation results appear in Table 1.  4 4   The Naples terms allow up to a two-thirds reduction in the amount of debt service owing in present value terms. Fordetails see Boote (1995).
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks

We need your sign to support Project to invent "SMART AND CONTROLLABLE REFLECTIVE BALLOONS" to cover the Sun and Save Our Earth.

More details...

Sign Now!

We are very appreciated for your Prompt Action!