Credit cycles revisited

Credit and business cycles play an important role in economic research, especially for central banks and supervisors. We reexamine a very useful dynamic model proposed by Kiyotaki and Moore (1997) of an economy with an endogenous credit limit. They claim that a small temporary shock generates large and persistent deviations from the steady state due to a positive feedback loop and the endogenous credit constraint. We mathematically show that contrary to common belief the model does not show amplification and persistence is visible only for a few parameter settings. Kiyotaki and Moore have linearized the model despite higher order terms being more important, rendering the Taylor expansion invalid. Further, we show that spillover effects in an economy with two distinct sectors are small. The strong amplification present in the original results, which supposedly is due to the large inter-temporal or dynamic multiplier effect, is spurious. The dynamic multiplier effect is of similar size than the static effect and in all cases numerically small.