Calibrating macroprudential policy

This paper contributes in a quantitative manner to the debate on macroprudential policy in three ways. First, it illustrates how macroprudential surveillance could have been better undertaken with a highly stable crisisprediction model for 14 OECD countries estimated over 19801997. Second, in terms of macroprudential regulation, it uses the related estimates to calibrate the increase in capital and liquidity needed to reduce average crisis probabilities to a desired level, as is being discussed in Basel. We show that an international consensus on regulatory changes will generate “winners” and “losers” in terms of capital and liquidity adjustments, and we suggest that raising capital and liquidity standards by 3.7 percentage points across the board will reduce the annual average probability of a financial crisis to around 1%. Finally, we show that countercyclical macroprudential policy is best calibrated on house prices and current accounts rather than GDP and credit, as are widely recommended at present. Credit growth, for example, has neither a direct nor an indirect statistical impact on crisis risk. Our results have important implications for the next generation of international banking regulations.