57552 POVERTY THE WORLD BANK REDUCTION AND ECONOMIC MANAGEMENT NETWORK (PREM) Economic Premise OCTOBER 2010 · Number 37 Russia 1998 Revisited: Lessons for Financial Globalization Brian Pinto and Sergei Ulatov In 1998, the Russian Federation experienced one of the most severe emerging market crises of the 1997­2001 period. It occurred less than six months after the attainment of single-digit inflation, which was supposed to launch the economy onto a sustainable growth path. This note sets out why that occurred and discusses the lessons learned. The Russian crisis of 1998 is yet another instance of financial by the crisis in East Asia and those in Argentina and Turkey. globalization contributing to an emerging market crisis in- Russia 1998 threatened to bring down the U.S. financial sys- stead of better resource allocation and faster growth. Exter- tem via the hedge fund managed by Long-Term Capital nal financial liberalization took place in the presence of weak Management. That threat prompted the New York Federal country fundamentals, with financial globalization eventu- Reserve to persuade 14 banks to pump $3.6 billion into the ally amplifying the vulnerability from the Russian Federa- fund while the Federal Reserve's Board of Governors aggres- tion's combination of a fixed exchange rate and unsustain- sively eased monetary policy by cutting interest rates thrice able government debt dynamics. In particular, external in quick succession (see Dungey et al. [2006]). portfolio investors--motivated by the expectation of a big Although the contagion effects of Russia 1998 have been official bailout--continued to finance the government's debt studied extensively (in Dungey et al. [2006], for example), build-up after mid-May 1998, even though it was obvious the country crisis itself has received scant attention--with by then that a fundamentals-based crisis ŕ la Krugman- Russia probably seen as too oil driven and geopolitical to Flood-Garber (Krugman 1979; Flood and Garber 1984) and have broad appeal. However, lessons on financial globaliza- Sargent-Wallace (1981) was unavoidable. As a result, Russia tion from Russia 1998 are of general applicability. Indeed, ended up with a much bigger external debt burden when paying more attention to Russia might have had a beneficial the crisis eventually hit that August. impact on the design of the rescue package for Argentina in By February 1998, Russia had both achieved single-digit 2001.1 inflation and substantially completed its privatization pro- gram. Yet it endured a massive exchange rate/banking/public Country Fundamentals debt crisis just six months later, in August 1998. This melt- down, which we shall refer to as "Russia 1998," was proba- Russia was going through two transitions: one from more than bly the most serious emerging market crisis witnessed over 70 years of central planning to a market economy; the other the 1997­2001 time frame, a turbulent period bookended from triple- to single-digit inflation and considerably lower fis- 1 POVERTY REDUCTION AND ECONOMIC MANAGEMENT (PREM) NETWORK www.worldbank.org/economicpremise cal deficits, a quest that began in earnest as part of a three-year ment debt was about 70 percent, and the dollar-ruble real stabilization program launched with International Monetary exchange rate appreciated by some 22 percent. That alone Fund (IMF) support in July 1995.2 But even though inflation would have reduced the debt-to-GDP ratio by some 8 per- came down rapidly, the fiscal deficit targets were significantly centage points and would have served to offset the impact exceeded. As a result, even though the nominal exchange rate of high fiscal deficits and disappointing growth. was fixed (or managed with narrow bounds) to lower infla- In addition to the inhospitable macroeconomic environ- tion, interest rates on ruble Treasury bills (GKOs) stayed at ment and weak corporate governance, a free-for-all atmos- exceptionally high levels, averaging 56 percent in real terms phere developed as widespread barter, noncash settlements, between May 1995 and July 1997. The real effective ex- and arrears of all types ("nonpayments") took hold. Ironical- change rate appreciated some 55 percent over the same peri- ly, the government itself became a prime instigator of non- od, a more or less inevitable arithmetical consequence of the payments, resorting to giving various forms of quasi-monies exchange rate­based stabilization: the exchange rate was and IOUs to its suppliers to economize on the use of cash fixed, but inflation came down only gradually (a point made because of the astronomically high real interest rates (Pinto, earlier by Dornbusch and Werner [1994]). Drebentsov, and Morozov 2000; Commander and Mumssen The persistently high real interest rates and large real ap- 1998). The private sector retaliated by becoming delin- preciation made it all but impossible for enterprise man- quent on tax payments; and that culminated in elaborate agers to operate profitably. At the same time, privatization offset schemes whereby tax arrears were settled, in effect, did not yield the improvements in corporate governance at a hefty discount. Nonpayments thus turned into a web of and productivity one might have expected. The process was hidden subsidies. Added to explicit budgetary subsidies, to- marred, resulting either in control going to insiders as part tal subsidies were estimated at 15­20 percent of GDP in of mass privatization; or valuable companies in oil, metals, 1996 and 1997, leading to ingenious forms of asset stripping and telecommunications going to the powerful Moscow instead of enterprise restructuring, and contributing directly banks as part of the opaque "loans-for-shares" auctions car- (if opaquely) to the fiscal crisis via the need for increased ried out in late-1995. debt issuance as taxes fell short. Why did the adverse economic outcomes not prompt a mid-course correction? For two reasons, in our assessment: Financial Globalization First, there was a deeply entrenched belief that lowering in- flation would lead to fast, sustainable growth--and inflation The liberalization of foreign portfolio investment in early was coming down quickly.3 Second, the debt dynamics of the 1997 swelled reserves to record levels of $25 billion by government seemed under control. Table 1 shows that despite midyear and pushed the stock market to new highs, while significant primary fiscal deficits, high interest payments, and interest rates on GKOs came down. That situation coincided small or negative growth rates, the debt-to-GDP ratio re- with falling inflation, which ignited hopes that growth was mained more or less constant over the period 1995­97. Based about to take off. In addition, political certainty was firming, on standard debt dynamics, one would have expected the with Boris Yeltsin's reelection and the appointment of an debt-to-GDP ratio to have been on an explosive path.4 economic "dream team" consisting of high-profile reformers. The explanation for the constancy of the debt-to-GDP But, by that time, nonpayments had become deeply en- ratio is to be found in the real appreciation of the ruble: for trenched, with cash tax shortfalls a chronic problem. The example, in 1996, the dollar-denominated share of govern- East Asian crisis spilled over in October 1997, leading to the first of a series of speculative attacks on the ruble. That No- vember, the central bank spent $6 billion (out of its reserves Table 1. Public Finances and Economic Growth, 1995­98 of $23 billion) defending the ruble's peg to the dollar; but, eventually, it was forced to raise interest rates. Primary Real GDP deficit Interest payments Government debta growth The final attack on the ruble began in mid-May 1998. By (percent Percent Percent of US$ Percent (precent then, with single-digit inflation having been attained and Year of GDP) of GDP revenues billions of GDP annually) the real exchange rate holding steady, the government's 1995 2.2 3.6 28 170 50 ­4.0 debt dynamics were visibly unsustainable. Instead of aban- 1996 2.5 5.9 47 201 48 ­3.4 doning the peg, however, Russia decided to mount a last de- 1997 2.4 4.6 38 218 50 0.9 fense with the help of a $22.6 billion international rescue 1998b 1.3 4.6 43 242 75 ­4.9 package, led by the IMF and including the World Bank and the government of Japan, that was announced in mid-July. Source: Pinto, Gurvich, and Ulatov 2005, table 9.1. a. Domestic plus foreign, end of period. Advised by Wall Street, the government also dabbled in b. Does not capture the subsequent debt renegotiation. financial engineering, deciding to swap its short-term, costly 2 POVERTY REDUCTION AND ECONOMIC MANAGEMENT (PREM) NETWORK www.worldbank.org/economicpremise GKOs for long-term, Eurobonds yielding 12­15 percent a Lessons year. The logic was seemingly unimpeachable: the swap would start a virtuous cycle by lowering interest payments The behavior of the private investors shows that moral haz- and rollover risk. That, plus the rescue package, would raise ard is alive and well, and sheds light on the "allocation puz- market confidence and buy time for the government to zle." Gourinchas and Jeanne (2007) argue that not only does raise primary fiscal surpluses and at last resolve the non- capital tend to flow from developing to developed countries, payments problem that was finally recognized as a crip- in line with the Lucas paradox; but within developing coun- pling impediment to revenue mobilization and faster tries, poorer performers (in terms of growth and productiv- growth. ity) receive the bulk of the capital flows--contrary to the Ironically, the economic and financial situation unraveled predictions of the neoclassical growth model. But Russia soon after the swap was completed. Less than four weeks 1998 shows that if investors care only about short-term later, Russia abandoned the ruble peg and defaulted on a gains and are driven by moral hazard, one is likely to see cap- substantial portion of its ruble-denominated debt on August ital flows in line with what might appear to be an allocation 17, 1998, plunging its banks (which were heavily exposed puzzle. That is the only reason one can offer to explain why to sovereign debt) into a crisis as well. Including the GKO- Russia was able to increase its external debt so significantly Eurobond swap, the government's dollar-denominated debt after mid-May 1998, when it became crystal clear that the increased by more than $16 billion between June 1 and the fiscal situation was unsustainable. Investors clearly wanted meltdown (in excess of 8 percent of postcrisis GDP). Para- to have their cake (charge interest rates reflecting high de- doxically, the amounts lent to Russia grew as the fundamen- fault and devaluation risks) and eat it (exit with 100 percent tals worsened and even as market investors themselves were of ruble proceeds at the precrisis exchange rate when a large signaling exceptionally high levels of default and devalua- official bailout package arrived). tion risk. Table 2 shows how the risk premia demanded by Instead of averting a crisis when fiscal fundamentals are the market evolved after mid-May. weak, financial engineering actually may trigger a crisis, as Table 2. GKO Yield and Sovereign and Devaluation Risk Premia for Key Dates, 1998 Percent annually Sovereign (default) Devaluation GKO risk risk Date and event yield premium premium May 15: Fiscal sustainability clearly in question. 39.3 4.8 23.0 July 13: Announcement is made of IMF-led $22.6 billion rescue package and GKO-Eurobond swap. 102.3 8.5 82.3 July 14: (day following rescue announcement) 58.2 8.1 38.6 July 20: IMF board approves package. Disbursement is reduced from $5.6 billion to $4.8 billion as 51.6 7.8 32.3 parliament stalls on key expenditure control and tax measures. July 23 (day preceding GKO-Eurobond swap completion) 54.2 8.2 34.4 July 24: GKO-Eurobond swap completed. GKO yields return to "crisis"` levels of mid-June, 66.4 10.0 44.9 sovereign risk premium jumps. August 6: World Bank board approves crisis package­related loan, and disburses $300 million. 77.7 12.0 54.1 August 10 (Monday, one week before the meltdown) 99.0 20.0 67.5 August 14 (Friday preceding meltdown): $1.7 billion in reserves is lost as portfolio investors exit, 144.9 23.8 109.5 bringing total loss from July 10 to August 14 to $4.5 billion. Russia's central bank bails out SBS-Agro bank with a $100 million loan. August 17 (meltdown Monday): Devaluation and default. Source: Pinto and Ulatov forthcoming, table 4. Note: If the yield on the one-year U.S. Treasury bill is 5 percent and the ruble/dollar target rate of depreciation is 6 percent, then the one-year GKO yield should be 11 percent (based on interest parity). If the yield is higher, the excess is the sum of the sovereign (default) risk premium and the devaluation risk premium (or the compensation for depreciation greater than the target of 6 percent). The sovereign risk premium was proxied by the spread on a short- maturity Russian government dollar borrowing, relative to the U.S. government; and the devaluation risk premium then was obtained as a residual. See Frankel and MacArthur (1988), who first used such a decomposition; and Pinto and Ulatov (forthcoming). 3 POVERTY REDUCTION AND ECONOMIC MANAGEMENT (PREM) NETWORK www.worldbank.org/economicpremise shown by the GKO-Eurobond swap. By its nature, a market- achieved macroeconomic stabilization. . . . The IMF is virtu- based, voluntary swap cannot be expected to lower the pres- ally certain, he declared, that real growth is underestimated ent value of the government's debt obligations, along the and will soon show up in official figures. . ." (p. 3). lines of the Modigliani-Miller theorem. But a swap may 4. Neither privatization proceeds nor seigniorage explains backfire and precipitate a crisis when fiscal solvency prob- the constancy of debt-to-GDP over the 1995­97 time frame. lems are present, as in Russia 1998. If the present value of As shown in table 1, there was a sizable increase in nominal future primary surpluses is less than outstanding debt, one debt--some $50 billion--between 1995 and 1997. way for the government to restore intertemporal budget bal- 5. A formal analytical statement of this argument can be ance is to let the nominal exchange rate depreciate--thereby found in Aizenman, Kletzer, and Pinto (2005). For more on lowering the real value of its domestic currency obligations. the Russian swap and why it failed, see Pinto and Ulatov When a swap out of local currency debt into dollar-denom- (forthcoming). inated debt occurs, however, it lowers the outstanding stock of domestic currency obligations (the tax base), calling for References an even larger depreciation (the tax rate).5 This recognition Aizenman, Joshua, Kenneth M. Kletzer, and Brian Pinto. 2005. "Sargent- could spur a speculative attack on foreign exchange reserves, Wallace Meets Krugman-Flood-Garber, or: Why Sovereign Debt Swaps triggering a crisis. Financial engineering definitely is not a Don't Avert Macroeconomic Crises." Economic Journal 115 (503): 343­ free lunch, and the hidden tab may be surprisingly high! 67. Similarly, implementing a successful official bailout is ex- Commander, Simon J., and Christian Mumssen. 1998. "Understanding tremely hard when fiscal solvency problems are present. In Barter in Russia." Working Paper 37. European Bank for Reconstruction Russia's case, the liquidity injection to reserves financed by and Development, London, U.K. Dornbusch, Rudiger, and Alejandro Werner. 1994. "Mexico: Stabilization, implicitly senior debt from the rescue package would demote Reform, and No Growth." Brookings Papers on Economic Activity 1994 the claims of GKO holders and become the perfect time to (1): 253­315. exit, with the liquidity injection providing the means of es- Dungey, Mardi, Renee Fry, Brenda Gonzalez-Hermosillo, and Vance Martin. cape. Together with the GKO-Eurobond swap, the interna- 2006. "Contagion in International Bond Markets during the Russian and tional rescue package itself helped trigger the 1998 crisis. the LTCM Crises." Journal of Financial Stability 2: 1­27. Flood, Robert P., and Peter M. Garber. 1984. "Collapsing Exchange-Rate Regimes: Some Linear Examples." Journal of International Economic 17 About the Authors (1-2): 1­13. Frankel, Jeffrey A., and Alan T. MacArthur. 1988. "Political vs. Currency Pre- Brian Pinto is senior adviser in the Poverty Reduction and Eco- mia in International Real Interest Rate Differentials: A Study of Forward nomic Management Anchor, Washington, DC. Sergei Ulatov is Rates for 24 Countries." European Economic Review 32 (5): 1083­114. an economist in the Moscow office of the World Bank. This note Gourinchas, Pierre-Olivier, and Olivier Jeanne. 2007. "Capital Flows to De- is based on Pinto and Ulatov's chapter in the forthcoming Else- veloping Countries: The Allocation Puzzle." Working Paper 13602. Na- tional Bureau of Economic Research, Cambridge, MA. vier Encyclopedia of Financial Globalization. Kharas, Homi, Brian Pinto, and Sergei Ulatov. 2001. "An Analysis of Russia's 1998 Meltdown: Fundamentals and Market Signals." Brookings Papers Notes on Economic Activity 37 (1): 1­68. Krugman, Paul R. 1979. "A Model of Balance-of-Payments Crises." Journal 1. More generally, Russia 1998 demonstrates the perils in of Money, Credit and Banking 11 (3): 311­25. official bailouts when fiscal fundamentals are weak (illustrat- Pinto, Brian, Vladimir Drebentsov, and Alexander Morozov. 2000. "Give Macroeconomic Stability and Growth in Russia a Chance: Harden ed in 2010 by Greece). On June 8, 2010, when the European Budgets by Eliminating Non-payments." Economics of Transition 8 (2): Union put finishing touches on a 440 billion package to 297­324. help Greece and curb contagion, the 10-year Greek/German Pinto, Brian, Evsey Gurvich, and Sergei Ulatov. 2005. "Lessons from the sovereign bond yield spread was 560 basis points. On Octo- Russian Crisis of 1998 and Recovery." In Managing Economic Volatility ber 11, 2010, it was much higher: 700 basis points! and Crises: A Practitioner's Guide, ed. Joshua Aizenman and Brian Pinto. 2. This section is based on Kharas, Pinto, and Ulatov Cambridge, U.K.: Cambridge University Press. Pinto, Brian, and Sergei Ulatov. Forthcoming. "Financial Globalization and (2001); and Pinto and Ulatov (forthcoming). the Russian Crisis of 1998." In Elsevier Encyclopedia of Financial Global- 3. The then first deputy managing director of the IMF, ization. Amsterdam, The Netherlands. Stanley Fischer, was quoted as follows in the final report of Sargent, Thomas J., and Neil Wallace. 1981. "Some Unpleasant Monetarist the January 9­12, 1997, U.S.­Russian Investment Sympo- Arithmetic." Federal Reserve Bank of Minneapolis Quarterly Review Fall: sium held at Harvard University: "Russia, he said, has 1­17. The Economic Premise note series is intended to summarize good practices and key policy findings on topics related to economic policy. It is produced by the Poverty Reduction and Economic Management (PREM) Network Vice-Presidency of the World Bank. The views expressed here are those of the authors and do not necessarily reflect those of the World Bank. The notes are available at www.worldbank.org/economicpremise.