The 1998 Russian Default and the Near-Failure of LTCM

Key Takeaways

• The trigger was small and the reaction was not. The Bank for International Settlements observed that it is difficult to reconcile the negligible fraction of global portfolios that ruble-denominated securities represented with the magnitude of the strains that followed in mature markets.
• The announcement was read as a change of regime, because official support programmes had until then consistently prevented large unilateral sovereign defaults. Every position priced on that assumption repriced at once.
• LTCM held over 125 billion dollars of assets on 31 August 1998, implying leverage of more than 25 to 1 against the 4.8 billion of capital it began the year with, before off-balance-sheet exposure. Its notional futures positions exceeded 500 billion dollars and its swaps exceeded 750 billion.
• Nobody had the aggregate number. The President's Working Group found that individual counterparties imposed bilateral limits but none of the fund's investors, creditors or counterparties provided an effective check on its overall activities.
• The rescue was private. Fourteen firms invested about 3.6 billion dollars on 23 September 1998, and Chairman Greenspan testified that no Federal Reserve funds were put at risk and no firm was pressured to take part.
• The Federal Reserve entered August 1998 with a tightening bias, cut three times by 17 November to 4.75 percent, then reversed all three cuts during 1999, returning the target to 5.50 percent by 16 November 1999.
• The recovery was unusually fast. Despite declines of 20 to 40 percent in major indices between mid-July and early October, the Federal Reserve reports the S&P 500 finished 1998 up more than 25 percent and the Nasdaq Composite up nearly 40 percent, while the Russell 2000 ended the year down 3 percent.

What Happened Between August and October 1998?

The summer of 1998 did not look like the eve of a crisis in the United States. The Federal Reserve Board's account of the year records an unemployment rate at its lowest quarterly reading in nearly thirty years, subdued inflation, most major equity indexes hitting record highs in July, and a Federal Open Market Committee that had judged at every meeting from March through July that if it acted at all, a tightening was more likely than an easing. The trouble was offshore and had been for a year, in the Asian economies that began to weaken in 1997 and in a Japanese recession that kept deepening.

Russia was where those pressures broke something. On 17 August 1998 the Russian authorities declared a moratorium on debt payments and effectively devalued the ruble. The Federal Reserve Board's annual report notes the timing with some understatement: the devaluation came the day before the Committee met on 18 August. The Committee still left policy unchanged that day, though it dropped its tightening bias to a symmetric directive.

What followed was not a slow bleed. The Federal Reserve Bank of New York's quarterly report on foreign exchange operations describes risk aversion intensifying sharply after 17 August, with losses in Russian markets and a dramatic widening of risk premiums producing successive waves of selling in emerging-market assets, and dollar-denominated emerging-market yield spreads over Treasuries reaching their highest levels since early 1995.

Chronology of the episode

Dated events with the figure each institutional source attaches to them.

One feature of that sequence stands out. The worst of the market damage happened in the six weeks between the LTCM letter surfacing and the intermeeting rate cut, which is after the sovereign event and before any policy response. That gap is where the episode actually lives, and it is why the story is not really about Russia.

Why Did a Russian Default Reach Portfolios That Owned No Russian Assets?

The Bank for International Settlements posed this question with unusual directness in its 69th Annual Report, noting the difficulty of reconciling the negligible fraction of global investment portfolios represented by ruble-denominated securities with the magnitude of the strains experienced in mature financial markets. The direct losses cannot explain the reaction.

The BIS answer is that the announcement was read as marking a shift in regime, because official support programmes had consistently prevented large unilateral sovereign defaults up to that point. Investors had not merely been holding Russian debt. They held a much larger family of positions whose pricing assumed that a sovereign in trouble gets rescued before it repudiates, and when Russia repudiated instead, the assumption failed everywhere it had been used at once.

The rest of the answer is about how the exposure was financed. The Federal Reserve Bank of New York describes emerging-market prices falling faster as leveraged investors were forced to liquidate to meet margin calls, with the speed of the declines then producing substantially illiquid trading conditions. The BIS traces where that pressure landed: higher margin calls and curtailed credit lines forced investors to raise funds by selling securities in markets that initially appeared relatively liquid, transferring the strains to the government bond markets of advanced industrial countries. Nobody sold Russia and bought Germany in an orderly rotation. They sold whatever could still be sold, so the healthiest markets absorbed the pressure from the sickest.

The destination shows up in the Treasury market, where the 30-year yield had traded consistently below 5.80 percent until mid-August and fell to as low as 4.96 percent on 30 September. Even inside the safest market in the world, though, the flight was selective. The BIS records the yield differential between the benchmark 30-year Treasury issue and less recent issues of the same credit peaking at over 30 basis points in mid-October. Two bonds with identical default risk traded at different yields purely because one was easier to sell, which is a liquidity premium priced explicitly and the cleanest measurement the episode produced.

Corporate credit widened on the same impulse. The Federal Reserve's February 1999 report records spreads of private rates over Treasury rates reaching levels not seen for many years, junk bond spreads roughly doubling between mid-summer and mid-autumn before falling back, and lower-tier commercial paper spreads rising substantially. Financial conditions, credit spreads and liquidity covers how those measures move together, and bond liquidity risk covers why one instrument has two prices depending on who needs to trade it. Cross-border flows reversed just as sharply: United States residents acquired about 35 billion dollars of foreign securities in the first half of 1998, stopped buying in July, and sold about 40 billion on net over August to October. That is the portfolio version of a margin call, and a corrective to the idea that international diversification is a stable property of a portfolio.

How Did One Fund Become a Question About the Financial System?

Long-Term Capital Management had been the most admired fund in the market. The President's Working Group on Financial Markets records net returns of approximately 40 percent in 1995 and 1996 and slightly under 20 percent in 1997, which the BIS separately describes as making the fund a symbol of how profitable financial sophistication could be. Its strategy was not reckless in the ordinary sense. It was reckless in a way that required a spreadsheet to see.

One decision at the end of 1997 explains most of what followed. The fund returned approximately 2.7 billion dollars of capital to its investors, cutting its capital base about 36 percent to 4.8 billion, while total assets stood at about 129 billion dollars, or 28 to 1. Greenspan's testimony describes the logic without sympathy: as profit opportunities diminished, the fund employed more leverage and increased its exposure to risk, a strategy he called destined to fail. Returning capital while keeping the positions raises leverage by construction.

LTCM by the numbers, from the President's Working Group report and the Bank for International Settlements 69th Annual Report.

The notional figures are what changed the conversation. Swap positions above 750 billion dollars notional, sitting on a 125 billion dollar balance sheet, meant the fund's footprint bore no relationship to the equity supporting it. Notional value is not the amount at risk and should never be read as a loss estimate. What it measures is how many contracts somebody would have to replace, in a hurry, if the fund stopped honouring them.

The BIS added the detail that turns this from a story about one fund into a structural observation. In several over-the-counter and derivatives markets, it noted, market-making had to some extent been performed de facto by leveraged participants such as LTCM, by virtue of the size of their positions. A firm large enough to be the market provides liquidity to everyone else right up to the moment it needs liquidity itself, and then competes with them for it. That is a different failure mode from a bank run, and it has no depositor to insure.

The disclosure problem was the other half. The President's Working Group found that although individual counterparties imposed bilateral trading limits on their own activities with LTCM, none of its investors, creditors or counterparties provided an effective check on its overall activities. Each bank knew its own exposure and considered it prudent. The aggregate existed as a fact about the world, not as a number anyone held. For scale, the BIS estimated at least 1,200 hedge funds with own assets of over 150 billion dollars by mid-1998. That measure is not the same quantity as LTCM's 125 billion dollar balance sheet and the two should not be divided into each other, but it does establish that the industry was small relative to the single fund inside it. The concern that September was never that a hedge fund might lose money, but that this one had grown into infrastructure without anyone deciding that it should.

Why Did Convergence Trades Fail All at the Same Time?

The President's Working Group describes LTCM's relative-value strategies as taking offsetting positions in two assets whose price relationship is expected to move favourably, and summarizes the fund's overall posture as a bet that liquidity, credit and volatility spreads would narrow from historically high levels. The BIS calls the same thing directional judgements on interest rate spreads and market volatility.

That description contains the trap. A relative-value position looks low risk because the two legs are similar, so most of the price movement cancels. Because it looks low risk it is run with leverage, since the residual spread is thin. Because it is levered it must be large. And because dozens of such positions are built from one insight, that spreads are historically wide and should narrow, the similarity that makes each safe in isolation is exactly what makes them a single position in aggregate.

The August 1998 flight to quality widened every one of those spreads simultaneously. The President's Working Group is blunt about the result: the size, persistence and pervasiveness of the widening of risk spreads confounded the risk management models employed by LTCM and other participants. The BIS reached the same conclusion from the market side, describing unusual price correlations after the moratorium that caused problems for many institutions, particularly the highly leveraged ones.

Two consequences follow and they are separable. A valuation loss is survivable on its own, because a convergence trade that widens will usually converge if it is held. A funding loss is not, because widening spreads generate collateral calls payable in cash today regardless of what the position is worth in two years. The fund was probably right about many of its spreads and it did not matter. Being right later is only a strategy if the financing lasts that long.

The general version is not a hedge fund problem. Correlations measured in calm conditions systematically overstate the diversification available in stressed ones, treated directly in how correlations change across regimes. The BIS made the same point about equities in 1998, recording exceptionally high cross-country correlations of stock price changes and drawing the parallel to October 1987 explicitly. The habit that follows is the one in liquidity-adjusted position sizing: size against what exiting costs under stress, not what entering costs today. LTCM sized against normal-market liquidity in instruments where the fund itself was a large part of that liquidity.

What Was Visible Before August 1998, and What Was Not?

Retrospectives on 1998 tend to flatten two very different kinds of information into one word, foresight. Separating them is the exercise.

Signals classified by whether an outside investor could have acted on them before 17 August 1998.

The distinction that matters is between the first two rows and the fifth. An investor in July 1998 could reasonably have concluded that risk was underpriced and that spread compression had gone a long way, and that conclusion supported carrying less leverage and more cash. It supported no view at all about which institution would fail, or in which week, because the concentration data did not exist.

The fourth row is the one that fools people. LTCM's track record was fully public and it was excellent. A returns series cannot distinguish a strategy that earns a real premium from one that collects a thin spread with large borrowings until the spread moves; three years of smooth returns are consistent with both, and the difference shows up only in the tail. That is why performance history is a weak input to risk assessment, and why risk of ruin and capital depletion is worth reading against any record that looks too even.

Hindsight check. The 1998 version of this test has an unusually clean answer, because the decisive fact had never been written down anywhere. For crowded relative-value trades, no amount of diligence would have helped: the crowding sat in the risk systems of dozens of counterparties, each seeing a prudent bilateral exposure, none summing them. For compressed credit spreads, the answer is a qualified yes with no timing at all. Treating the two as equally available is how a case study manufactures false confidence, and cognitive biases in trading explains why the effect survives being pointed out.

How Did the Federal Reserve and LTCM's Counterparties Respond?

Two responses ran in parallel that autumn and they are constantly conflated. One was a private recapitalization organized in a New York Fed conference room. The other was monetary policy.

The private recapitalization

The President's Working Group dates the beginning to Tuesday 22 September 1998, when a core group of four of the most concerned counterparties began discussions. The following day, fourteen firms agreed to invest about 3.6 billion dollars of new equity. The report describes the New York Fed's role as providing the facilities and encouraging the firms to seek the least disruptive solution in their own collective self-interest. Greenspan's testimony on 1 October is explicit that no Federal Reserve funds were put at risk, no promises were made and no firms were pressured, and gives as his reason that a forced liquidation would have distorted prices and produced severe losses for participants with no connection to LTCM at all.

Notice what the fourteen firms were choosing between. Not helping versus not helping, but an orderly wind-down they controlled versus a default that would have forced each of them to replace its own contracts simultaneously, in the same direction, in illiquid markets. Framed that way the decision is neither charity nor coercion. It is a group of creditors recognizing that they were collectively the market they would have to sell into.

Monetary policy

Federal funds target rate changes, September 1998 to November 1999, from the Federal Reserve Board's record of open market operations.

The dates matter more than the size. The first cut came on 29 September, six days after the recapitalization was already agreed, and the Federal Open Market Committee minutes for that meeting give the reasoning as turmoil spreading from the Russian crisis into United States markets, strong demands for safety and liquidity driving Treasury yields down while private spreads gapped higher, and banks tightening credit standards. LTCM is not the stated reason.

The Federal Reserve Board's annual report then records something easy to miss. In the days after the September cut, market disturbances got worse, not better: price movements were exacerbated by deteriorating liquidity as some securities dealers cut back their market-making, and by increased anticipation of an unwinding of positions by hedge funds and other leveraged investors. In early October Treasury yields briefly tumbled to their lowest levels in many years. Easing did not stop the deleveraging, because the deleveraging was not about the level of interest rates.

That is why the 15 October cut happened between scheduled meetings, at the Chairman's initiative and after a conference call, with a matching quarter-point discount rate reduction. Afterwards strains diminished considerably, safe-haven demand ebbed, Treasury yields trended higher and volatility eased, though risk spreads remained very wide and liquidity limited. A coordinated support package for Brazil in mid-November added to the same easing.

The full arc is the useful part. The Committee went from a tightening bias in July to 75 basis points of cuts by November, then reversed all of it during 1999, ending back at 5.50 percent on 16 November. Read against Federal Reserve policy rates and forward guidance, the 1998 sequence is the awkward case: three cuts delivered into an economy that was growing above trend the whole time, and withdrawn once the market plumbing they were aimed at had been repaired. An easing cycle can be round-tripped inside twelve months when the problem it addressed was liquidity rather than the economy.

How Long Did the 1998 Recovery Take?

Weeks, which makes 1998 the outlier of this library in the direction almost nobody expects from a crisis with a sovereign default and a systemic hedge fund in it.

The BIS records that stock markets suffered their largest setbacks since 1987, with major market indices declining by 20 to 40 percent between mid-July and the first week of October, the fall being especially severe in continental Europe, possibly reflecting closer economic ties to Russia. It then records that equity markets in all major economies recovered during the fourth quarter, that prices in many cases climbed back to their early summer highs, and that the rebound was particularly strong in the United States, where equity prices closed the year at record highs. The pace of it, the BIS notes, surprised many observers.

The Federal Reserve's February 1999 report gives the full-year outcome: the Nasdaq Composite up nearly 40 percent, the S&P 500 up more than 25 percent, the Dow Jones Industrial Average and the NYSE Composite up more than 15 percent, and the Russell 2000 down 3 percent.

That last figure is the one worth sitting with. In the same calendar year the largest United States companies produced a strong gain and the smallest ones produced a loss. A saver in a broad large-cap index fund experienced 1998 as an unusually good year with a scary autumn. A saver holding small-cap or emerging-market exposure experienced something else entirely. Index-level recovery statistics conceal exactly this dispersion, which is why the drawdown and recovery calculator is more useful applied to what you hold than to a headline index.

Three qualifications belong with any 1998 recovery claim. It was a recovery in prices: the BIS is careful to say that corporate and emerging market bond spreads fell back from their peaks without returning to pre-crisis levels, so credit did not recover with equities. It was not universal by geography, since Japan's ongoing recession kept its equity prices depressed throughout. And the speed is a caution rather than a comfort. The BIS observed that a reader armed only with annual observations would see the period as an uneventful continuation of previous trends, with nothing to betray that the defining events happened in the two months after mid-August. An annual chart of 1998 is a lie of omission.

What Was Specifically Different About 1998?

Every episode in this library has features that do not transfer. For 1998 there are four, and three of them are unusually favourable.

The losses were mostly not credit losses. Compare this with the 2008 financial crisis, where the underlying asset genuinely defaulted at scale and the write-downs were real. In 1998 the mature-market damage came overwhelmingly from repricing and forced liquidation, not from borrowers failing to pay. That is why the recovery could be so fast: there was much less permanent economic loss to work through.

The United States economy was strong the entire time. The Federal Reserve was describing an unemployment rate at a thirty-year low and rapid output growth while it was cutting. There was no recession to recover from. A financial crisis that leaves the real economy intact is a different animal from one that does not, and treating 1998 as evidence that markets bounce back quickly imports an assumption that was specific to that year.

The transmission ran through a funding currency. The Federal Reserve's H.10 noon buying rates show the dollar at 135.60 yen on 2 October 1998 and 117.00 on 9 October, a decline of roughly 13.7 percent in five business days, with two of those sessions taking it from 131.15 on 6 October to 118.85 on 8 October. The New York Fed attributes the move to participants unwinding short yen positions in an environment of increasing risk aversion. Nothing about Japanese fundamentals improved in that week. This is what forced deleveraging looks like when the leverage was raised in a foreign currency, and it has no analogue in the 2020 COVID crash or the 2022 rate shock.

The systemic institution was a fund, not a bank. There was no deposit insurance question, no lender of last resort with an obvious claim on the problem, and no regulator with the authority to see the whole position. The resolution had to be assembled out of the private interests of fourteen firms in twenty-four hours. That is a structurally weaker safety net than a banking crisis has, and it worked in 1998 partly because the counterparty list was short enough to fit in one room.

Common Myths About LTCM and the Russian Default

"The Fed bailed out LTCM." No public money was involved. Fourteen private firms provided about 3.6 billion dollars, and Greenspan testified that no Federal Reserve funds were at risk, no promises were made and no firm was pressured. What the New York Fed supplied was a room and a deadline. Whether convening the creditors of a private fund is itself a form of official support is a fair question about precedent, but it is a different question from whether taxpayers paid.

"LTCM caused the 1998 crisis." The sequence rules this out. Emerging-market spreads had been widening for a year, Russia defaulted on 17 August, and the fund's difficulties became public only after 2 September. LTCM was the most alarming casualty of the flight to quality and it amplified the illiquidity that followed, but it was downstream of the trigger.

"The models were wrong." The President's Working Group says something more specific and more uncomfortable: the size, persistence and pervasiveness of the widening of risk spreads confounded the risk management models employed by LTCM and other participants. The models were not wrong about the distribution of ordinary days. They were calibrated on a history in which the fund itself had not yet grown large enough to change what a bad day looks like. A better model does not fix that.

"A hedged position is a safe position." Every LTCM trade was hedged in the sense of having two legs. The hedge removed the direction and left the spread, and the spread was then levered until it mattered. Hedging changes which risk you own; it does not reduce the total unless size falls with it.

"Diversification failed in 1998." Diversification across spreads that all express one view failed, and international equity diversification failed too, since the BIS recorded exceptionally high cross-country correlations of stock price changes. Diversification into on-the-run Treasuries did not fail; it worked so well that off-the-run Treasuries of identical credit lagged by over 30 basis points. The question is not whether to diversify but whether the things combined are genuinely different when it counts, which is the subject of stress testing and scenario analysis.

"It ended when the Fed cut." Conditions deteriorated after the 29 September cut, not before, which is why an intermeeting cut followed on 15 October, and even then risk spreads remained very wide and liquidity limited. Compare the same pattern in Black Monday 1987, where the policy response and the price bottom were also separate events.

What a Reader Can Actually Carry Forward

1998 is the episode in this library with the least useful outcome and the most useful mechanism. A full recovery inside one calendar year, with the index finishing at a record, is not something to plan around. The mechanism recurs constantly and at every scale.

What generalizes

• Positions that look diversified because they are numerous can be one position. The test is not how many holdings you have but how many distinct things have to be true for them to work. If one sentence describes the thesis behind all of them, that is your position count.
• Leverage converts being early into being wrong. Many of LTCM's spreads did eventually converge and the fund could not wait, because collateral calls arrive on a schedule set by the market rather than by the thesis. Any position whose financing can be withdrawn faster than the thesis can resolve is a timing bet however it was analyzed. Risk management is where that becomes a limit.
• Size against exit liquidity, not entry liquidity. If your position is a meaningful share of the normal trading in an instrument, you are part of the liquidity you are counting on. Position sizing and risk per trade is the retail-scale version of the same constraint.
• A liquidity premium is a real, priced risk. The 30-plus basis point gap between on-the-run and off-the-run Treasuries in October 1998 is the cleanest demonstration available that identical credit trades at different yields on tradeability alone. Yield pickups from less liquid versions of the same exposure are compensation for exactly that, collected in calm markets and repaid in stressed ones.
• The catalyst can be far smaller than the reaction. A market that reprices violently on an event too small to explain the loss is telling you the event invalidated an assumption, not that it destroyed capital. Ask which assumption, then ask what else was priced on it.

What does not generalize

• The speed of the recovery. A rebound to record highs within a quarter happened because the mature-market losses were largely repricing rather than defaults, and because the United States economy never contracted. The 1929 crash and the dot-com bubble are the counterexamples here, and they are measured in decades.
• The private rescue. It worked because the exposure sat in a small enough group of firms to convene in a day. A more distributed version of the same problem has no such room.
• The availability of 75 basis points of cuts. The Federal Reserve had a 5.50 percent starting point and no inflation constraint, and reversed the whole easing within a year.
• The specific instruments. Nothing important about 1998 depends on ruble bonds or fixed-income arbitrage. Looking for the next LTCM in the hedge fund industry is the reliable way to miss it somewhere else.

The one question worth asking now

Take the positions you hold and write down the single sentence that would have to be false for most of them to lose money together. If you cannot find such a sentence, the portfolio is genuinely diversified. If you can write it easily, that sentence is your actual position and its size is your actual exposure, regardless of how many line items sit underneath it. The exercise takes ten minutes, requires no forecast, and is the part of 1998 that transfers to a portfolio of any size.

References

Every figure on this page was verified against the following sources, each retrieved on 26 August 2026:

• President's Working Group on Financial Markets: Hedge Funds, Leverage, and the Lessons of Long-Term Capital Management: the 17 August 1998 devaluation and moratorium; every LTCM figure in the table above; the 22 September core group and 23 September recapitalization by fourteen firms; the relative-value strategy description; and the findings on bilateral limits and aggregate leverage.
• Federal Reserve Board: Testimony of Chairman Alan Greenspan on the Private-Sector Refinancing of Long-Term Capital Management, 1 October 1998: that no Federal Reserve funds were put at risk, no promises were made and no firms were pressured, and why a forced liquidation would have harmed unrelated participants.
• Federal Reserve Bulletin: Treasury and Federal Reserve Foreign Exchange Operations, July to September 1998: emerging-market spreads at their highest since early 1995; the 30-year Treasury yield below 5.80 percent until mid-August and 4.96 percent on 30 September; the 24.2 percent DAX decline; and the 11.6 percent Dow fall between 26 and 31 August with the 7539.07 low and 7842.62 close.
• Board of Governors of the Federal Reserve System: 85th Annual Report, 1998: the March to July tightening bias; the unemployment rate at its lowest quarterly reading in nearly thirty years; most major indexes at record highs in July; the devaluation the day before the 18 August FOMC meeting; indexes back near start-of-year levels by early September; the post-September worsening as dealers cut market-making; the 15 October intermeeting and discount rate cuts; and the reversal from about 35 billion dollars of foreign securities bought to about 40 billion sold on net. This volume also reprints the Monetary Policy Report of 21 July 1998, the source for the Federal Reserve already citing raised risk spreads on external debts in Asia, Russia and Latin America before the moratorium.
• Federal Reserve Board: Monetary Policy Report to the Congress, February 1999, Section 2: Economic and Financial Developments in 1998 and Early 1999: the ruble depreciating more than 70 percent; junk bond spreads roughly doubling; private and lower-tier commercial paper spreads; the Treasury on-the-run premium; and the full-year 1998 index outcomes.
• Bank for International Settlements: 69th Annual Report, Chapter V, Turmoil in Asset Markets: the 20 to 40 percent index declines and the 1987 comparison; ruble securities as a negligible fraction of global portfolios; the regime-shift reading of the moratorium; the margin-call mechanism transferring strain to advanced-country bond markets; the over 30 basis point on-the-run premium; the 52 percent loss in the partners' letter; the 1,200 hedge fund estimate; bond yield volatility at or near a trough in July outside Japan; de facto market-making by leveraged participants; and the fourth-quarter recovery.
• Federal Reserve Board: Open Market Operations Archive: every federal funds target change and resulting level in the 1998 and 1999 table above.
• Federal Reserve Board: H.10 Foreign Exchange Rates, Historical Country Data for Japan: the yen per dollar noon buying rates of 135.60 on 2 October 1998, 131.15 on 6 October, 118.85 on 8 October and 117.00 on 9 October.
• Federal Reserve Board: Minutes of the Federal Open Market Committee, 29 September 1998: turmoil spreading from the Russian crisis into United States markets, Treasury yields falling while private spreads gapped higher, tightened bank credit standards, and the 25 basis point reduction.

Figures deliberately not stated. This page gives no S&P 500 peak or trough close and no exact peak-to-trough percentage for 1998, because no primary or institutional source verified in this session supplied one; the BIS range of 20 to 40 percent across major indices is used instead. For the same reason it gives no Russian GKO yield, no size or foreign-ownership share for the Russian domestic debt market, no Emerging Markets Bond Index spread level, and no loss figures for individual banks. Those omissions cost the page something real: the GKO market is where the default happened, and this account reaches it only through the reaction it caused elsewhere. An unverified yield for it would have read better and asserted more than the record here supports.

Method note: the percentage changes in the dollar against the yen were computed by Swoopr Investment from the Federal Reserve H.10 noon buying rates listed above, with no adjustment for interest differentials. Every other figure is stated as its source states it, and where a source gives a range or an approximation the range is reproduced rather than converted into a point estimate.

This is educational content about a historical episode. It is not investment advice, it is not a forecast, and nothing here should be read as a claim about how any future market disruption will behave.

Frequently Asked Questions