The Liquidity Effect and the Change of Financial Factors in Subprime Crisis

Since the subprime crisis in 2007, the changes of the U.S. financial market have led to a substantial movement in the financial variables of the banking system. In order to identify the variation of the liquidity effect --- the negative response of the interest rate to an exogenous money stock shock --- that comes from the subprime crisis, this thesis targets to provide evidence for the hypothesis which implicitly predict a linkage between the financial variables and the liquidity effect. The hypothesis, which is implied by a generalized version of limited-participation models, says that the transaction costs in financial markets may intensify the liquidity effects. To examine the hypothesis, the thesis employs the threshold vector autoregressive (TVAR) model to estimate from the U.S. time series data. The financial variables, which are regarded as good indicators for the transaction costs, are used as the threshold variable to separate the time series data into the high- and low-transaction-cost regimes. Based on the TVAR estimates as well as the nonlinear impulse responses, the findings support the hypothesis of the generalized version of limited-participation models. The estimation implies that there is a more prominent and instant liquidity effect under the high-transaction-cost regime than under the low-transaction-cost regime. In addition, we find that most of the U.S. time series data after 2007 subprime crisis coincide with the high-transaction-cost regime. This implies that, given other things unchanged, the liquidity effect may be more prominent after the crisis than before the crisis.