| Summary: | 碩士 === 國立屏東商業技術學院 === 不動產經營系 === 97 === Forecasting the trends of real estate cycle’s volatility and changes, it has been the most important issue in the area of macroeconomic analysis and decision making. In the past, most of study’s results showed that it did have the business cycle in real estate’s markets, and recur constantly. And once the events of business cycle fluctuations in real estate market occur, immediately it will be driven synchronous fluctuations of other industries. The impact of business cycle fluctuations in real estate market could be very important issues.
Using the leading indicator composite index of Taiwan’s real estate cycle, this paper attempts to extend the Markov switching model to employ the discrete hidden Markov model for capturing the information of Markov switching model’s inner states set not directly-observed, and pre-detect the real estate cycle’s volatility. The empirical results show that HMM model can capture the asymmetry in duration of states. Compared with the real estate leading indicator announced by Taiwan Real Estate Research Center, HMM model has the same results on forecasting the trends of cycle fluctuations. On the other way, compared with Markov switching model, the explanatory power of HMM model in 4-steps out-of-sample forecasting is supported both conceptually and methodologically.
|
|---|
