关键词: 协整, 时间序列, 长短期因果效应, 房产税

Abstract: In order to avoid the bias of the heterogeneity of control samples in the causality inference method based on the classic panel data model and the possible " pseudo-regression" of interruption time series (ITS) method, and to relax the restriction on the result variable " second-moment stability" , the time series can be used to estimate and infer the policy causality. Using the co-integration theory of time series, this paper proposes a causal effect inference method for co-integration time series based on the error correction model and tests of structural breakpoint. This method expands ITS from the traditional trend-stationary process to a stochastic trend process, can not only infer and evaluate the policy treatment effect, but also solve the identification problem of temporary and continuous policy effect, and the method is more suitable for the analysis of policy causal effect in the absence of effective control samples. In addition, the empirical analysis of Shanghai and Chongqing pilots shows that the property tax pilot policy has a long-term inhibitive effect on the price growth of commercial housing, and the policy effect of the two cities shows a significant difference—the effect on Shanghai is strong and quick, while the effect of Chongqing’s policy has a longer time lag.

Key words: Co-integration, Time Series, Short and Long Term Causality, Property Tax