Sovereign Debt Structure for Crisis Prevention IMF OP

Sovereign Debt Structure for Crisis Prevention (IMF OP 237) E. Borensztein, M. Chamon, O. Jeanne, P. Mauro, J. Zettelmeyer (IMF Research Department) • Existing debt structures are crisis-prone. Much crisis prevention can be done with existing instruments. • The case for growth-indexed bonds • Pricing growth-indexed bonds

• Motivation: – Existing debt structures are crisis-prone, especially in emerging markets (short term, foreign currency). – Debt servicing more costly in recessions. Procyclical fiscal policies. • Crisis Prevention with Existing Debt Instruments: – A few countries have improved their debt structures within a few years – How? Credible inflation stabilization, institutional reforms committing to stable policies, lengthening debt structure first with inflation-indexed.

The Case for Growth-Indexed Bonds • • An old idea Advantages Obstacles (theory, practice, survey) Precedents (how past innovations emerged) • Resolve one obstacle: pricing

Benefits of Growth-Indexed Bonds For issuing countries: • Fewer defaults and collateral damage • Less pro-cyclical fiscal policy • Greater international risk sharing For investors: • New assets whose return has low correlation with GDP of investors’ country • Fewer defaults

Growth-Indexed Bonds—an example Consider a floating-rate bond with a coupon rate equal to: Coupon = r*+(g-g*) with a minimum of zero. Suppose that in 1990: • r* = 7 percent • g*= average real GDP growth rate over previous 20 years • 50 percent of government debt is growth-indexed Define Fiscal saving = reduction in annual interest payments resulting from indexation

Interest Savings over the Economic Cycle: Argentina

Table 2. Correlation between the Primary Balance and Real GDP Growth ________________________________ Emerging Markets Advanced Countries Without With Indexation _________________________________ Mean 0. 30 0. 77 0. 40 0. 64 Median 0. 37 0. 80 0. 45 0. 74 _________________________________ Sources: OECD Analytical Database; and International Financial Statistics, Government Finance Statistics, and Country Reports, International Monetary Fund. Note: This analysis includes 25 emerging market countries and 20 advanced countries.

But is it going to be too expensive? How much of an insurance premium should countries pay over and above what they pay on standard bonds? Not much: cross-country GDP risk is largely diversifiable

Table 4 : Comovement of Individual Country Real GDP Growth with World Market Portfolio Real GDP Growth World Real Stock Returns United States World United States R 2 coeff std. error 0. 266 0. 277 1. 011 0. 988 0. 431 0. 307 0. 430 0. 076 0. 135 0. 289 0. 345 0. 696 0. 183 0. 245 0. 159 0. 138 0. 125 0. 875 0. 060 0. 083 0. 042 0. 077 0. 086 0. 095 0. 905 0. 062 0. 088 Emerging Markets Median Average of Abs (1 -coeff) 0. 056 0. 105 0. 715 0. 810 0. 783 0. 587 0. 668 0. 016 0. 028 0. 236 0. 203 0. 797 0. 291 0. 310 0. 019 0. 038 0. 023 0. 032 0. 968 0. 119 0. 151 0. 013 0. 033 -0. 048 -0. 055 1. 055 0. 108 0. 137 Developing Economies Median Average of Abs (1 -coeff) 0. 127 0. 125 1. 370 1. 117 0. 963 0. 869 0. 995 0. 022 0. 054 0. 322 0. 447 0. 841 0. 372 0. 500 0. 037 0. 068 0. 087 0. 070 0. 930 0. 184 0. 190 0. 012 0. 041 -0. 020 0. 002 1. 019 0. 120 0. 161 Advanced Countries Median Average of Abs (1 -coeff) Notes: Regressions of individual country real GDP growth rates on world (or U. S. ) real GDP growth or world (or U. S. ) real stock returns (one-year lead). The data refer to 27 advanced countries, 39 emerging markets (S&P/IFC definition), and 26 relatively large developing countries. Sources: International Monetary Fund and Morgan Stanley Capital International.

Obstacles to Financial Innovation • Financial Innovation is always difficult: externalities and coordination problems; product uncertainty; need for critical mass; standards; community of professionals. • Innovation by corporates vs. sovereigns: scale; control over policies and statistics.

Precedents • • Brady Value Recovery Rights (VRRs) GDP VRRs (Costa Rica, Bulgaria, Bosnia) Ciudad de Buenos Aires Options on US economic statistics (Longitude-DB-GS) • GDP warrants in Argentine restructuring • Inflation-indexed bonds

Challenges • Too risky? Already exposed to GDP. Less default risk. Diversifiable across countries. • Too complicated? Difficult to price? Growth is well-understood and followed. Inflationindexed. Pricing model at the end. • Misreporting of GDP data? Political incentives. Could be audited • Moral Hazard? • Political resistance to pay insurance?

How Will It Happen? • Financial innovation in history evolves randomly, e. g. inflation-indexed bonds • Externalities and coordination problems, official intervention

Table 5. Introduction of Inflation Indexed Securities by Sovereigns Indexed Public Debt Outstanding in 1999 Argentina Australia Brazil Canada Chile Colombia Czech Republic Finland France Greece Hungary Iceland India Ireland Israel Italy Mexico New Zealand Norway Poland Sweden Turkey UK UK USA Period of Issue 1972 -1989 1985 -1988 19931964199119561967199719451998199719951955 n. a. 198319551983 19891977 -1984 199419821992 -2000 1952 1994197519811997 - Average CPI Inflation Rate in 3 Years Prior to Introduction (in percent) 18. 6 8. 4 3. 8 n. a. 4. 6 39. 6 13. 7 9. 3 n. a. 1. 7 7. 6 23. 2 4. 3 n. a. 18. 6 32. 7 18. 5 110. 7 14. 2 1. 4 9. 8 292. 2 8. 2 5. 4 80. 8 10. 7 13. 2 2. 8 in million US$ 0 --27, 860 45, 291 6, 636 14, 960 4, 949 150 0. 7 3, 994 197 394 494 166 260 79, 037 0 2, 528 n. a. 361 30 0 --15, 475 8, 561 --55, 288 57, 014 in Percent of Total Govt Debt 0. 0 --29. 5 19. 6 1. 5 62. 0 13. 2 1. 7 0. 0 0. 6 0. 2 3. 0 11. 5 0. 2 1. 1 80. 2 0. 0 8. 4 n. a. 2. 3 0. 1 0. 0 --12. 5 24. 3 --12. 0 0. 8

How to Price Growth-Indexed Bonds • How do people price standard bonds? p(no default)*return on sovereign bond+ p(default)*recovery value = return on U. S. treasuries • Extract from there reasonable combinations of p(default) and recovery values • Estimate standard deviations and correlations of growth, primary balance, and real exchange rate, based on past data • Extract debt/GDP default trigger that yields the probability of default implicit in the spreads • Simulate 100, 000 paths (Monte. Carlo) and compute expected payoffs

Distribution of NPDV pay-offs (baseline where all debt is plain-vanilla)

Pricing: Key Messages • Growth-indexed bonds are not much more difficult to price than are plain vanilla bonds. • Externality from growth-indexed bonds to standard bonds: large issuance of growthindexed bonds reduces the probability of default, thereby reducing the required coupon on standard bonds as well.

Additional Slides Survey of Investor Attitudes

Background and Assumptions Assume an emerging market sovereign borrower that has successfully tapped international financial markets for a number of years, currently not experiencing major problems, but whose bonds trade at substantial spreads above US Treasuries. This country will be called Emerging. Land (“EL”). EL has experienced real gross domestic product (GDP) growth of 3 percent on average over the past 15 years, with a maximum of 7 percent, and a minimum of negative 8 percent. Most analysts expect average growth and volatility of GDP to be similar in the next decade. We will ask you to price a new financial instrument--and variations on it--to be issued by this country. Assume that ten-year eurobonds (U. S. dollar-denominated) issued by EL with a coupon of 7 percent currently trade at a spread of 400 basis points above U. S. Treasuries. We’ll call these bonds the “Plain Vanilla Bonds” (“PV Bonds”). EL is contemplating issuance of a growth-indexed bond (“Growth Bond”) with a ten-year maturity and with a coupon of 7 percent plus the difference between real GDP growth during that year and 3 percent. However, coupon payments can never be negative: Coupon = 7 percent + (real GDP growth – 3 percent), with a minimum of zero

1. At what spreads would you be willing to purchase the Growth Bonds? A) Spreads more than 50 basis points lower than PV Bonds B) Spreads 10 -50 basis points lower than PV Bonds C) Same spreads as PV Bonds D) Spreads 10 -50 basis points higher than PV Bonds E) Spreads 50 -100 basis points higher than PV Bonds F) Spreads 100 -200 basis points higher than PV Bonds G) Spreads 200 -300 basis points higher than PV Bonds H) Spreads more than 300 basis points higher than PV Bonds I) Unwilling to purchase regardless of the spreads

2. Consider this variation. The yearly coupon now has a minimum of 3. 5 percent, and an extra payoff in years of positive growth. The formula is the following: Coupon = 3. 5 percent + real GDP growth, with a minimum of 3. 5 percent Assume that, given expectations of the growth rate of GDP in EL, the average expected payoff of this bond is the same as for the bonds described in question 1.

At what spreads would you be willing to purchase this alternative Growth Bond? A) Spreads more than 50 basis points lower than PV Bonds B) Spreads 10 -50 basis points lower than PV Bonds C) Same spreads as PV Bonds D) Spreads 10 -50 basis points higher than PV Bonds E) Spreads 50 -100 basis points higher than PV Bonds F) Spreads 100 -200 basis points higher than PV Bonds G) Spreads 200 -300 basis points higher than PV Bonds H) Spreads more than 300 basis points higher than PV Bonds I) Unwilling to purchase regardless of the spreads

3. Please tell us which of the following would lead you to reduce the spreads you require to hold Growth Bonds of the type described in question 1 issued by EL. (1 is very important and would lead you to reduce the spreads by 50 basis points or more; 2 would lead you to reduce the spreads by 20 -50 basis points; 3 would lead you to reduce the spreads by 10 -20 basis points; 4 is irrelevant. ) As a reminder, EL faces spreads of 400 basis points on its plain vanilla bonds. 1 2 3 4 You find out that the United States is planning to issue Growth Bonds at about the same time. 1 2 3 4 You find out that five other major emerging market sovereigns are planning to issue Growth Bonds at about the same time. 1 2 3 4 A reliable economic consultancy firm announces it will provide a free software with a formula for pricing growth bonds. 1 2 3 4 A well-respected international consortium reports a study showing that the GDP data provided by the country are reliable, and announces it will monitor GDP data quality annually. 1 2 3 4 Growth Bonds covering at least 50 percent of the country’s debt are issued in the context of a negotiated restructuring of EL ’s debt.

4. Please tell us which of the following factors make you reluctant to hold Growth Bonds regardless of their specific design. (1 is very important, 4 is not important). 1 2 3 4 Uncertainty about future liquidity of Growth Bonds 1 2 3 4 Complexity/difficulty in pricing 1 2 3 4 Uncertainty about integrity of GDP data reported by EL 1 2 3 4 Concern that EL will have less incentive to promote economic growth 1 2 3 4 Variable coupon instead of fixed coupon What other factors make you reluctant to hold growth bonds

Commodity-Indexed Bonds 5. Assume that EL’s economy is heavily dependent on exports of a single commodity, and it decides to issue commodity-indexed bonds, that is, bonds whose return is indexed to the price of the commodity. Please indicate which factors would make you reluctant to invest in commodity-indexed bonds (1 is very important, 4 is not important). 1 2 3 4 It is too difficult to forecast commodity prices beyond 3 to 5 years 1 2 3 4 You invest in many countries and only few countries are heavily dependent on a single commodity. It is not worth your time to learn about commodity prices 1 2 3 4 You are not interested in direct exposure to commodity price fluctuations, even if many of the countries you invest in are heavily dependent on commodities 1 2 3 4 You are interested in exposure to commodity price fluctuations, but prefer to obtain it directly through forwards or futures linked to commodity prices What other factors are important to you?

Domestic Currency Bonds. 6. EL is now considering issuing a domestic-currency denominated bond. Please indicate which factors would make you reluctant to invest in bonds denominated in EL’s currency (1 is very important, 4 is not important). 1 2 3 4 You are not interested in exposure to currency fluctuations in EL 1 2 3 4 You are concerned about the possibility of a rise in inflation in EL 1 2 3 4 You are concerned that EL’s central bank could intervene in foreign exchange markets and pursue an unfavorable exchange rate close to the time when bond payments are due. 1 2 3 4 You are concerned about your ability to hedge exposure to EL’s currency because of an illiquid NDF market. What other factors are important to you?

Additional Slides miscellaneous

Interest Savings over the Economic Cycle: Mexico

Figure 2. France: Primary Balance with and without Indexed Debt, 1979 -2002 6. 00 4. 00 Real GDP Growth Rate 2. 00 0. 00 -2. 00 -4. 00 3 percent limit -6. 00 Total Balance/GDP -8. 00 1979 1984 1989 1994 1999 2002 2. 50 1. 50 Primary Balance/GDP under SGP 0. 50 -1. 50 Primary Balance/GDP under SGP with Indexed Debt -2. 50 Primary Balance/GDP -3. 50 1979 1984 1989 1994 1999 2002