Principles for Navigating Big Debt Crises

I spent my career watching things happen that seemed unprecedented, only to discover they had happened many times before. In my fifty years as a global macro investor I encountered financial events that felt shocking and new—the 1971 dollar crisis, the 1982 Latin American debt blow-up, the 1987 stock market crash, the 1990 Japanese deflation, the 2008 financial collapse. Each time, if I hadn’t studied history, I would have been completely unprepared.

Rather than seeing lots of individual things happening, I saw fewer things happening over and over again, like an experienced doctor who sees each case of a certain type of disease unfolding as “another one of those.” That shift in perception—from chaos to pattern—is the entire point of this study. The template that follows is built on 48 big debt cycles, defined as all cases that produced real GDP declines of more than 3 percent in large countries.

To me, watching the economy and markets on a day-to-day basis is like being in an evolving snowstorm with millions of bits and pieces of information coming at me. To understand what I’m really seeing requires stepping back to the perspective of 100 years rather than 100 days. From above, the snowstorm becomes a pattern. The same forces, the same sequences, the same types of leaders making the same types of mistakes at the same stages of the cycle.

I am sharing this template in the hope of reducing the likelihood of future debt crises and helping them be better managed. The two biggest impediments to managing a debt crisis well are not the debt itself. They are ignorance and a lack of authority. Policy makers who know how these things work, and who have the legal powers to act, can handle even severe debt crises without catastrophe. The cases that went badly wrong were almost always failures of understanding first, and debt second.

Credit is the giving of buying power. This buying power is granted in exchange for a promise to pay it back, which is debt. Giving the ability to make purchases by providing credit is, in and of itself, a good thing. Not providing the power to buy and do good things can be a bad thing. Too little credit growth can create as bad or worse economic problems as too much, with the costs coming in the form of foregone opportunities.

Whether more credit is desirable depends on whether the borrowed money is used productively enough to generate sufficient income to service the debt. If that occurs, the resources will have been well allocated and both the lender and the borrower will benefit. If it doesn’t, both will suffer. The failure is not in borrowing. The failure is in misjudging what the borrowed money will produce. That misjudgment, repeated widely enough, produces a bubble.

You create a cycle virtually anytime you borrow money. Buying something you can’t afford means spending more than you make. You are not just borrowing from your lender; you are borrowing from your future self. Essentially, you are creating a time in the future in which you will need to spend less than you make so that you can pay it back. That future arrives. When it arrives for an entire economy at once, it arrives as a depression.

Lending naturally creates self-reinforcing upward movements that eventually reverse to create self-reinforcing downward movements that must reverse in turn. During the upswings, lending supports spending and investment, which supports incomes and asset prices; increased incomes and asset prices support further borrowing and spending. During the downswings, reduced lending leads to reduced spending, which reduces incomes and asset prices, which further reduces the willingness to lend.

Throughout history, only a few well-disciplined countries have avoided debt crises. That is because lending is never done perfectly and is often done badly due to how the cycle affects people’s psychology. More often than not, policy makers err on the side of being too loose with credit because the near-term rewards—faster growth—seem to justify it. It is also politically easier to allow easy credit than to tighten it. The consequences arrive later, for someone else to manage.

In the early part of the cycle, debt is not growing faster than incomes, even though debt growth is strong. That is because debt growth is being used to finance activities that produce fast income growth—borrowed money going toward expanding a business and making it more productive, supporting growth in revenues. Debt burdens are low and balance sheets are healthy, so there is plenty of room for the private sector, government, and banks to lever up.

Suppose you earn $50,000 a year and have a net worth of $50,000. You have the capacity to borrow $10,000 per year, so you could spend $60,000 per year for a number of years. For an economy as a whole, increased borrowing and spending can lead to higher incomes and rising asset values, giving people more collateral to borrow against. They borrow more and more. As long as the borrowing drives growth, it is affordable. This phase feels wonderful to live through.

Debt burdens can’t rise faster than the money and income needed to service them forever. When the limits of debt growth relative to income growth are reached, the process works in reverse. Asset prices fall, debtors have problems servicing their debts, investors get scared and pull back. Liquidity problems develop. People cut back on their spending, which reduces incomes further. This self-reinforcing process typically happens suddenly, after years of quiet buildup.

Bubbles are driven by extrapolation. When things have been good for a long time, most people believe they will continue being good. Prices are discounting further rapid appreciation from already high levels. There is broad bullish sentiment. Purchases are being financed by high leverage. In markets, when there is a consensus, it gets priced in. Because humans tend to move in crowds and weigh recent experience too heavily, extrapolation becomes the dominant force.

One classic warning sign that a bubble is coming is when an increasing amount of money is being borrowed to make debt service payments, which of course compounds the borrowers’ indebtedness. Another sign is the proliferation of new financial intermediaries and new financial instruments that develop outside the supervised and protected banking system. Fast-growing lending markets operating outside regulation are classically symptomatic of the bubble phase.

Bubbles emerge in one or a couple of markets and are often hidden beneath the averages. Central banks focus on inflation and growth, and when neither is a problem they don’t adequately worry about debt-financed purchases of investment assets. This is one of the biggest problems with most central bank policies—because they don’t target the management of bubbles, the debt growth they enable can go to finance the creation of bubbles while the overall numbers look fine.

When things are so good that they can’t get better—yet everyone believes they will get better—tops of markets are being made. Most tops are triggered when the central bank starts to tighten and interest rates rise. Sometimes the tightening is brought about by the bubble itself, because growth and inflation rise while capacity constraints begin to pinch. The inversion of the yield curve—long-term rates at their lowest relative to short-term rates—is the classic signal.

Early on in the top, some parts of the credit system suffer but others remain robust, so it isn’t clear that the economy is weakening. The central bank is still raising interest rates and tightening credit while parts of the financial system are quietly failing. This is the cruelest aspect of the top: the bad news accumulates before it shows up in the headline numbers. By the time policymakers see it clearly, the reversal is already irreversible without aggressive action.

The squeeze manifests as: debt service payments rising faster than income; asset prices falling, creating a negative wealth effect; lenders growing worried about whether they will get their money back. Borrowers are squeezed as an increasing share of their new borrowing goes to pay debt service or isn’t rolled over. This is the transition point—when the positive self-reinforcing upswing begins to operate in reverse. The top is not an event. It is a process.

Severe debt crises—depressions—occur when the usual remedy no longer works. Typically, debt crises can be alleviated by lowering real and nominal interest rates. But when rates reach about 0 percent, that lever is no longer effective. At that point, debt restructuring and austerity dominate, which are deflationary and depressive. The debts are too heavy to roll over or refinance, and the losses must be absorbed by someone.

In a deflationary depression, debtors have more debt than they can service. Banks are undercapitalized and no longer function as lenders. The wealth effect of falling asset prices reduces spending. There are not enough buyers for the debt assets that need to be sold. Everyone wants to deleverage simultaneously. A classic death spiral: falling asset prices hurt borrowers, hurt lenders, reduce spending, reduce incomes, which further reduces asset prices.

One person’s financial assets are another’s financial liabilities—promises to deliver money. When the claims on financial assets are too high relative to the money available to meet them, a big deleveraging must occur. The free-market credit system that finances spending ceases to work well and typically works in reverse, necessitating the government to intervene in a big way. The central bank becomes a big buyer of debt; the government becomes the lender of last resort.

Policy makers typically fail to recognize the magnitude of the problem initially, instead enacting a number of one-off policies that are insufficient to move the needle. It is only after what is usually a couple of years—and a lot of unnecessary economic pain—that they finally act decisively. On average, that decisive shift comes two to three years into the depression, after stocks have fallen more than 50 percent, economic activity has fallen about 10 percent, and unemployment has risen to around 10–15 percent.

Policy makers have four levers: 1) austerity—spending less; 2) debt defaults and restructurings; 3) the central bank printing money and making purchases; 4) transfers of money and credit from those who have more than they need to those who have less. The first two are deflationary and reduce debt burdens. The third is inflationary and stimulates growth. The key is to get the mix right so that deflationary and depressive forces are balanced with inflationary and stimulative ones.

When the right balance is achieved, a “beautiful deleveraging” occurs. Income needs to grow faster than debt. If a country has a debt-to-income ratio of 100 percent, and incomes can be made to grow at a rate 2 percent faster than debt grows, then the debt burden will fall by roughly 2 percent per year. The central bank needs to print enough money to get nominal growth above nominal interest rates—but not so much that inflation accelerates, the currency devalues, and a new bubble arises.

In the end, policy makers always print. That is because austerity causes more pain than benefit, big restructurings wipe out too much wealth too fast, and transfers of wealth from haves to have-nots don’t happen in sufficient size without revolutions. Also, printing money is not inflationary if the size of the money creation is simply offsetting the size of the credit contraction. What matters is the total of money and credit, not money alone.

An ugly deleveraging becomes beautiful when the mix shifts from deflationary to balanced. The ugly phase—falling asset prices, unemployment, austerity, contracting credit—transitions to the beautiful phase when the money printing is large enough to neutralize deflation and lift nominal growth above nominal interest rates. The shift happens not because the debts disappear but because the economy grows faster than the burden of debt service. The debt ratio falls gradually, over years.

Even during a beautiful deleveraging, the recovery is slow. It typically takes roughly 5 to 10 years for real economic activity to reach its former peak—hence the term “lost decade.” It takes longer still, often more than a decade, for stock prices to reach former highs, because it takes a very long time for investors to become comfortable taking the risk of holding equities again. The emotional scars outlast the financial ones.

The deflationary template applies when a country’s debts are largely in its own currency—it can print. The inflationary template applies when debts are in a foreign currency or when the country lacks a reserve currency that the world wants to hold. In that case, capital flows out, the currency weakens, inflation rises, and the central bank cannot lower interest rates enough to offset the weakness without accelerating the flight. This is the harder path.

Countries most vulnerable to severe inflationary deleveragings or hyperinflations share common features: no reserve currency; low foreign-exchange reserves; large foreign debt; a history of high inflation; a large trade deficit. When a country meets these conditions, a debt crisis translates directly into a currency crisis. The sequence is: capital flight → currency collapse → imported inflation → political pressure to print → more inflation.

The most important characteristic of cases that spiral into hyperinflation is that policy makers don’t close the imbalance between income and spending and debt service; instead, they fund and keep funding spending over sustained periods of time by printing large amounts of money. Once an inflationary depression reaches the hyperinflationary stage, it becomes extremely difficult to stop printing. Stopping creates an extreme tightness of liquidity. The longer the crisis goes on, the harder it becomes.

Germany entered the postwar period meeting every condition for an inflationary disaster. Losing the war meant the mark would not be the reserve currency of the postwar era. A large stock of external debts had been acquired. The Allies would impose reparations in hard currency—gold marks—that Germany had to earn but could not print. Foreign exchange reserves were thin. The confidence of German citizens in their own currency was already eroding.

The final terms of the Treaty of Versailles came as a huge shock. Germany was to lose 12 percent of its territory, 10 percent of its population, 43 percent of its pig-iron capacity, and 38 percent of its steel capacity. The reparations were enormous and structured so that debt service burdens would get bigger if economic conditions improved—the perverse design of a victor who wanted the punishment to scale with recovery. Germany signed because the alternative was total occupation.

By September 1922, Germany was trapped in a classic hyperinflationary spiral. Extreme capital withdrawals and rapidly rising prices were forcing the Reichsbank to choose between extreme illiquidity and printing more money. Reckless money printing was less the cause of the hyperinflation than what was required to prevent massive deflationary defaults by banks and a deflationary economic collapse. The institution understood the trap perfectly and could not escape it.

From July 1922 until November 1923, the mark depreciated by 99.99999997 percent versus the dollar—the cost of dollars increased 1,570 billion percent—and prices rose by 387 billion percent. In 1913, six billion marks had circulated in the entire German economy. By late 1923 the Reichsbank was printing 400 quintillion marks per month. As one economist noted at the time: stopping the printing press would mean that in a very short time all public and private commerce would cease entirely.

The stabilization came through abolishing the old currency entirely and issuing a new one—the rentenmark—backed by a fixed claim on German land and industrial assets. The new currency required extreme limits on monetization, credit creation, and government spending. Germany needed a comprehensive and aggressive policy shift, not incremental adjustments. Accompanied by the Dawes Plan renegotiating reparations and extending foreign exchange loans, the hyperinflation ended almost overnight.

The late 1920s bull market had every feature of a classic bubble. Stocks sold at multiples as high as 30 times earnings, financed by margin debt at accelerating rates. The call loan market—a new innovation—let investors access short-term debt to fund long-term risky holdings, a classic asset/liability mismatch. Investment trusts, another innovation, drew new speculators into the market. The more prices rose, the more aggressively speculators bet they would rise still more.

It was the tightening that popped the bubble. The first signs of trouble appeared in March 1929, when rumors of a Federal Reserve clampdown on speculative credit emerged. After the New York Fed raised its discount rate to 6 percent in August, the blow-off phase was over. On Black Tuesday, October 29, large blocks of shares flooded the market at the open. When a rumor spread that the Bankers’ Pool had shifted from buying to selling, panic accelerated. It collapsed the way all bubbles do.

The banking system held up briefly after the crash. In 1930, bank net earnings declined about 40 percent compared to the prior year—but they remained on sound footing. Many analysts believed they would remain strong. That confidence was the error. As the economy weakened and those banks had lent to suffered losses, the banks began to fail. The most important early failure was the Bank of the United States—400,000 depositors, triggered by a false rumor. It set off a national panic.

It is classic in a big debt crisis: policy makers play around with deflationary levers to bring down debt for a couple of years but eventually wake up to the fact that the depressing effects of debt reduction and austerity are both too painful and inadequate. Hoover’s attempt to balance the budget through austerity was a rookie move that is classic in depressions. The debate about what to do became antagonistically political, with strong populist overtones. Roosevelt came on the scene as the necessary pivot.

Roosevelt broke the gold peg in April 1933. Gold rose overnight. Within two weeks, the Federal Reserve was able to decrease its liquidity injections. Short-term rates dropped. The money supply increased. As explained in the archetypal template, a beautiful deleveraging requires enough stimulation to offset deflationary forces and bring nominal growth above nominal interest rates. That is what breaking the gold peg accomplished. The economy roared to life over the following three months.

The events of the 1930s have happened many times before for the exact same reasons. Changing laws in ways that would have made the last crisis less bad is typical at the end of big debt crises. The Glass-Steagall separation of commercial and investment banking, deposit insurance, new securities regulation—all were efforts to redesign the system so that the same dynamics could not repeat. They delayed the next crisis by several decades. That is the best outcome policy can achieve.

Between 1987 and 1989, Japan experienced a bubble driven by a self-reinforcing cycle of rising debt, strong growth, and strong asset returns. Debts rose by 24 percent of GDP during the bubble to a pre-crisis peak of 307 percent of GDP. The debt was in Japan’s domestic currency and the majority was owned domestically. Japan was a net creditor. These facts meant Japan could not have an inflationary collapse—it would have a deflationary one, and a long one.

The bubble was allowed to inflate because central banks focus on inflation and growth—and in Japan in the late 1980s, neither was a problem in the conventional measures. That is exactly what allowed the debt growth to finance the creation of the bubble. When it became clear that assets were overvalued, the Bank of Japan tightened. Land and stock prices collapsed. Japanese companies and banks held assets on their books at peak valuations while the market moved against them.

Japan’s resolution was very slow. Monetary policy was not sufficiently easy to push nominal GDP growth above nominal interest rates for quite some time. The country did not print aggressively or devalue until decades later. It allowed the debt problems to sit on bank balance sheets rather than forcing restructuring and recapitalization. The result was the “lost decade”—or rather, two lost decades. Japan is the canonical example of the ugly deleveraging that never quite becomes beautiful.

Between 2004 and 2007, the United States experienced a bubble driven by a self-reinforcing cycle of rising debt, strong growth, and strong asset returns. Debts rose by 38 percent of GDP during the bubble to a pre-crisis peak of 349 percent of GDP. The US housing market was showing every sign of a classic bubble: prices high relative to traditional measures, discounting further rapid appreciation, broad bullish sentiment, purchases financed by high leverage, buyers speculating in forward inventory.

Shadow banking—financial institutions building new channels that got around the more established and better-regulated banking system—is a common feature of bubble periods. Borrowers and lenders had severe asset/liability mismatches: borrowing short-term and lending long-term; holding illiquid assets while financing with liquid debt. The US financial regulatory system did not keep pace with these developments. It did not have adequate visibility into the shadow banks, nor the powers to curb their excesses.

The tightening popped the bubble. As interest rates rose, home prices declined. Many subprime borrowers had adjustable-rate mortgages, so their payments rose immediately with rates. Declining asset prices created negative wealth effects, which fed back into the economy through declining spending and incomes. This was exactly the self-reinforcing downward spiral the template predicts—a system behaving as it always behaves when the debt load exceeds the income that can service it.

In August 2007, we wrote to our clients: “This is the Big One—the financial market unraveling we have been expecting, the unwinding of widely held, irresponsibly created positions that occurred as a result of financial middlemen pressing to invest for high returns the immense amount of liquidity flooding the financial system.” We had studied the template. What looked unprecedented to most observers looked to us like another one of those.

AIG alone had $1 trillion in assets at peak and had issued hundreds of billions of dollars of insurance contracts on bonds. When those bonds faced losses, AIG had to pay out. On September 16, 2008—the day after Lehman filed for bankruptcy—the Fed announced $85 billion in emergency funding to AIG. The deal was drafted in a rush on a Tuesday afternoon. That kind of quick thinking and creativity—working around regulatory and political constraints—was exactly what the crisis required.

In 2008, the US had a team of policy makers that understood what it would take to manage a debt crisis about as well as one could expect given that debt crises of this magnitude happen about once in a lifetime. Treasury Secretary Paulson had 30-plus years of financial market experience. Fed Chairman Bernanke had spent his academic career studying the Great Depression specifically. Bernanke knew that what killed the 1930s economy was premature tightening; he would not repeat it.

The quantitative easing broadly worked in providing the needed additional stimulus. Despite continued debt problems in Europe, the US economy and markets finished 2010 on a high note. Growth picked up. Frequent concerns over inflation stemming from the fast pace of central bank printing didn’t materialize, laying to rest the incorrect belief that printing a lot of money would always cause inflation to accelerate. The critical variable is not money alone—it is money plus credit.

The compendium covers 48 big debt cycles—all the cases that produced real GDP declines of more than 3 percent in large countries. They divide into two groups: primarily domestic currency crises (deflationary deleveragings, like the US in 1929 and 2008 and Japan in 1990) and foreign currency crises (inflationary deleveragings, like Weimar Germany 1920s, Argentina in the 1980s, and the Asian crisis countries in 1997). The two paths diverge at the moment the currency comes under pressure.

In virtually every deflationary case where the deleveraging became beautiful—where policy makers pulled enough of the right levers—the pattern was the same: after an ugly phase of 2–3 years, aggressive stimulation arrived. M0 increased substantially, interest rates were pushed toward zero, and real exchange rates were depreciated. The speed and aggressiveness of the pivot determined whether the lost decade lasted 5 years or 25. Sweden in 1993 did it right. Japan in 1990 did not.

The case studies make a further point visible only in aggregate: the recovery in economic activity and capital formation is always slow, even in the beautiful cases. Stock prices take longer to recover than GDP. GDP takes longer to recover than employment. Employment takes longer to recover than sentiment. The emotional memory of the crash—the reluctance to take risk again—is the last thing to heal, and it is what makes the next bubble possible once it finally does.

The best-managed cases are those in which policy makers swiftly recognize the magnitude of the credit problems; don’t save every institution that is expendable, but do save those that are systemically important; ensure that the essential institutions have enough capital and liquidity to function; and work to restructure the debts so that the pain is distributed over the population and over time so that the debt does not impose an intolerable burden.

Can most debt crises be managed so there aren’t big problems? Based on my examination of 48 cases and the ways the levers available to policy makers work, I believe it is possible for policy makers to manage debt crises so that they are not catastrophic. But this requires knowing how they work. The two biggest impediments are not debt levels—they are ignorance and lack of authority. Policy makers who understand the template and have the powers to act can navigate even severe crises.

When debts are denominated in foreign currencies rather than one’s own currency, it is much harder for a country’s policy makers to do the sorts of things that spread out debt problems over time. The reserve currency privilege—the ability to print the currency the world wants to hold—is the most powerful single advantage a country can have in a debt crisis. It is the difference between Weimar and Washington. That privilege must be earned, and it can be lost.

The template did not give me the ability to predict the precise timing or trigger of any crisis. What it gave me was the ability to see the conditions that make a crisis inevitable—the accumulation of debt relative to income, the asset/liability mismatches, the overconfidence of investors extrapolating recent trends. When I saw those conditions, I knew another one of those was coming. The blizzard looked different from above than from inside it. That is the value of the study of history.