Patterns of financial crises


In light of the recent fears of recession I decided to summarize one of the best books on crises, “This time is different” by C. Reinhart & K. Rogoff. The book describes financial and economic crises since 1300 up to early 2000s including 2007–2008 Great Financial Crisis. The authors divide crises into 5 categories (inflation crises, currency crashes, external debt default, domestic debt default, banking crises), and seek patterns in them.

Below is a thread on the historical patterns that the book describes.

1. Historically, a frequent occurrence of global banking crises have been followed with high rate of sovereign defaults on external debt.

There are many ways global financial crises can lead to high incidence of sovereign debt crises worldwide, but especially in emerging economies.

- Banking crises in developed countries has tended to result in decreasing global growth which causes the commodity prices to drop. This fall in commodity prices affects negatively export revenues of commodity. As a result, their ability to service external debt is diminished.

- Herd behavior also plays a role in the channels through which financial crises lead to sovereign debt crises. Banking failures in one country may lead to credibility problems in the neighbor countries. Investors tend to overgeneralize the failure of one economy to other countries. They also avoid risky investments which is followed by capital outflows from emerging markets. That investors decrease their exposure when their wealth reduces is detrimental to emerging countries’ ability to service their sovereign debt.

- Banking crises in advanced economies have historically been associated with credit crunches. When money is scarce in these countries, it is difficult for emerging markets to borrow which results in weakening economic activity in these countries and increases debt burden.

2. There has been a clear historical pattern between inflation and sovereign external debt default. Countries that defaulted on external debt have limited or no access at all to global financial markets. And if prior to this, they hadn’t managed to decrease government spendings, they resort to inflation tax which sometimes even may result in hyperinflations. Therefore, inflation tends to worsen after sovereign default on external debt.

3. Since most emerging markets are commodity exporters, it isn’t surprising that commodity price cycles may lead the capital flow cycle. Troughs in the former may be followed by a high incidence of sovereign defaults.

Several researches have shown that there is a high procyclicality in emerging markets borrowing. The uptrend in prices of commodities tend to increased borrowing for emerging markets. When the cycle ends and commodity prices go down, capital inflow to these countries dries which leads to sovereign defaults.

The authors note that there’s a pattern found in the research on financial crises that countries experiencing “capital bonanza” (sudden capital inflow) are more likely to have a debt crisis.

4. Several words on domestic debt defaults. It may seem surprising but defaults on domestic debt are not that rare in economic history. Since the government can theoretically always inflate away its domestic debt, one may wonder why it chooses not to service the debt. But inflation can do more harm to financial sector. That’s why the government sometimes can regard the repudiation as a better choice.

Some conditions may render inflation much less desirable. If the debt is short-term or indexed, the government should achieve a more aggressive inflation to be able to drastically decrease the real payments of its domestic debt.

5. The authors compare the aftermath and the run-up of domestic and external debt defaults. Their conclusion is that in most aspects default on domestic debt is worse. For example, GDP has historically decreased 8% during the 3-year period before the default. If we look at the crisis year only, average output declines were 4 and 1.2 percent for domestic and external debt crises respectively.

According to Reinhart and Rogoff, the situation is much worse if we compare inflation during and after external and domestic debt crises. Historically, the year when external default occurred, the inflation rate was 33 percent. But it pales in comparison to the inflation during domestic debt crises. In domestic default the inflation rate has historically been 170 percent. That is 5 times higher than the inflation rate during external debt default! Even after the domestic default, the rate doesn’t fall to normal levels and remains above 100 percent after the crisis.

6. One type of crises that the authors devote many pages to is banking crises. An interesting feature of banking crises is that even developed economies failed to avoid them even though these countries were able to handle sovereign debt crises many years ago. Authors call this “graduation”. In the case of sovereign debt defaults, we can say that advanced economies seem to achieve graduation. However, graduation from banking crises remains elusive even for the most advanced economies as the 2007–2008 crisis demonstrated so spectacularly. That banking systems are vulnerable to bank runs exacerbates the situation with banking crises.

Many banking crises were related to real estate prices. These price cycles don’t differ significantly between advanced and emerging markets. However, most macroeconomic variables, such as government spending, interest rates and time series on income and consumption etc, show higher volatility in emerging markets. So the finding that real estate prices cycles around financial crises exhibit the similar pattern in two categories of countries (advanced and emerging) is surprising.

Another not so conspicuous feature of banking crises is their relationship with financial liberalization. Kaminsky and Reinhart have demonstrated that in most banking crises cases (eighteen out of twenty-six) after 1970, the financial sector had been liberalized in the five-year run-up to the crisis. In the 1980s and 1990, many liberalization cases resulted in financial crises. This doesn’t mean that no country managed to arrange the financial liberalization to proceed easily. Very few countries (e.g, Canada) could escape the crisis.

There’s a mathematical evidence to this claim too. The same authors, Kaminsky and Reinhart, provide evidence that the probability of a financial crisis conditional on liberalization is higher than the unconditional probability of a banking crisis. In another research, Demirguc-Kunt and Detragiache confirm this by using a fifty-three country sample and presenting evidence that the effect of financial liberalization on the banking sector is negative. In other words, financial liberalization is a destabilizing factor for the financial sector.

Another feature of the banking crises is related “capital flow bonanzas” mentioned in the paragraph on debt crises. It turns out that in the run-up to a crisis capital inflow to the country persistently increases. One research also provided evidence that a high capital inflow (equaling several percent of GDP) makes the economy more crisis prone.

Capital flow bonanzas are related not only to banking crises but also to credit cycles. Two researchers, Mendoza and Terrones, looked at credit cycles in developed and emerging markets. They found that increase in capital inflows tend to be followed by credit booms in emerging markets. Though credit booms are not a necessary condition for a banking crisis, there were credit booms in the run-up to most financial crises in emerging market economies.

Another regularity of the banking crises is their link to housing prices as previously mentioned. Two economists (Bordo and Jeanne) looking at advanced economies during the 1970–2001 period found that banking crises usually occur when real housing prices reach the peak or immediately after the bust.

Are there any markers that may predict a banking crisis? The research on financial crises answers yes to this question. Some forerunners of these crises include significantly increasing asset prices and big current account deficits. Ramping up of public or private debt can also predict a financial crisis. The other two precursor factors to banking crises have already been mentioned. Capital inflow bonanzas and financial liberalization more often resulted in financial crises than not.

Other than precursors, most severe banking crises also are followed by some common results which the authors categorize into three groups.

The first aftermath of the financial crises is that asset prices markedly fall which may last more than several years. Average decrease in real housing prices is 35 percent, while stock market crashes average more than 55 percent. The average duration for real estate and equity markets collapses is 6 and 3.5 years respectively. One episode is especially remarkable in the case of housing prices declines. In Japan real housing prices collapsed during seventeen consecutive years. But even if we exclude Japan’s experience, the average duration of housing markets falls remains more than five years.

Second, unsurprisingly banking crises negatively affect the output and employment. During the decline phase of the crises, which can last more than more than four years, the average unemployment rate jumps to 7 percent. The magnitude of the output decrease is more than that of the unemployment rate — output falls on average 9.3 percent. However, the duration is less prolonged averaging two years.

Though the decline in real per capita GDP has averaged more than 9 percent as noted, there is a considerable difference between the experiences of advanced and emerging market economies. During the post-World War II period, the former group of economies experienced less decline in real per capita GDP than the latter group. The authors suggest an explanation for more severe contractions of emerging market economies related to credit availability. When foreign credit decreases, the economy recedes which is translated to the decline in real GDP.

Third, as a result of a banking crisis the government debt drastically increases in value. In the post-World War II cases the real value of the sovereign debt almost doubled in comparison to precrisis levels (actually it jumped to an average of 86 percent but this doesn’t change the fact banking crises are followed by the explosion in government debt). The literature on financial crises usually points to bailout costs as the main driver of debt increase. However, the authors note that there is another factor which dwarfs the effect of bailout costs. They mention that since output reduces there’s a marked reduction in tax revenues. And this “inevitable collapse in tax revenues” is the most significant driver of debt explosion. Another factor contributing to the debt buildup is the increase in interest expenses since interest rates increase during the crises.