Financial markets and the monetary transmission mechanism.

摘要

Does monetary policy affect the real economy? If it does, what role does the financial system play in the monetary policy transmission mechanism? This series of three essays addresses these questions using a mix of econometric and institutional analysis. The essays share in common the conclusion that the classical dichotomy does not hold and that monetary policy influences real economic activity in important ways.;Essay Two looks more deeply into the distributional impact of the credit channel and how the development of the Japanese financial market has affected the nature of the transmission mechanism. This essay uses a panel data set that includes 115 manufacturing firms from 1975 to 1992. The investment behavior of these firms is analyzed during the tight money episodes of 1979-80 and 1989-91. The firms are divided into two sets: the first set has close ties to the banks that serve as their primary source of external funds and are likely to be well informed about the firm; the second set has weaker links to the banks and, therefore, is assumed to have more difficulty obtaining external financing. During the tight monetary policy episodes, investment is more sensitive to liquidity for the second set of firms than the first. This result holds even for the 1989-91 episode that occurred after significant steps had been taken to liberalize the Japanese financial markets.;Essay Three looks more broadly at the correlation between movements in the money supply and movements in the real economy. If monetary policy actions work through either the lending or money channels, then movements in the money supply should help predict future movements in real activity. This essay concentrates on the impact on the money-output correlation of a univariate specification that allows a time series to be characterized as stationary around a broken trend function. For data on the United States and Japan, the measure of real output can be characterized as broken-trend stationary and its positive Granger-causal relationship with detrended money supply is more robust than that found by using log first-differenced output. For the US data before 1985, money Granger causes broken-detrended output even when different short-term interest rates are included as regressors. (Abstract shortened by UMI.).;Essay One begins by using a simple model to make concrete the distinction between the money and lending channels. Then, a number of primary documents from the Bank of Japan are used to identify shifts to tight monetary policy that were motivated by either inflation or balance-of-payments concerns. Focusing on these dates as the beginning of tight money episodes, it is shown that bank lending volume falls by more than would be predicted using forecasting equations. Additionally, during these episodes, the composition of firm borrowing shifts away from bank loans and toward other sources of external financing. Impulse-response functions indicate that bank lending falls in response to innovations in a variable used to proxy for the stance of monetary policy; this decline occurs before movements in real output are observed. Finally, reduced-form evidence suggests that bank lending has important predictive power for future output movements. The results from all of these tests are consistent with the lending channel hypothesis.