Financial Crises are Predictable

Macroeconomists have debated whether financial crises are predictable. New research shows that indeed they are – and are caused by the rapid expansions of credit accompanied by asset-price booms.

Robin Greenwood, Samuel Hanson and Jacob Sorensen contribute to the macroeconomic literature with their March 2021 study, “Predictable Financial Crises,” published in the January 2022 issue of The Journal of Finance, in which they analyzed the probability of financial crises as a function of past credit and asset-price growth. They combined historical data on the growth of outstanding credit to nonfinancial businesses and households with data on the growth of equity and home prices to estimate the future probability of a financial crisis in a panel of 42 countries over the period 1950-2016.

Greenwood, Hanson and Sorensen constructed an indicator variable they called the red-zone (R-zone) that identified periods of potential credit market overheating. A country was in the business R-zone if nonfinancial business credit growth over the past three years was in the top quintile of the full sample distribution, and stock market returns over the same window were in the top tercile. Following is a summary of their findings:

  • Crises can be predicted using past credit growth in simple linear forecasting regressions – both nonfinancial business and household credit growth forecast the onset of a future crisis.
  • The degree of predictability, however, is modest, even at horizons of up to five years – a one standard deviation increase in real one-year credit growth led to about a 3-percentage point increase in the probability of a crisis within the next five years. However, the degree of predictability increases when large credit expansions are accompanied by asset price booms.
  • When nonfinancial business credit growth is high and stock market valuations have risen sharply, or when household credit growth is high and home prices have risen sharply, the probability of a subsequent crisis is substantially elevated.
  • The probability of a crisis at a one-year horizon was 13% if a country was in the business R-zone, a more than threefold increase over the unconditional probability of 4%. The comparable one-year probability was 14% if a country was in the household R-zone – if household credit growth and home price growth were jointly elevated.