房價對台灣消費的影響

Abstract

根據107年主計處國富統計報告，房地產占家戶部門總資產的比例高達四成，故房價變動會對家戶財富造成很大影響。為了探討房市對總體經濟的影響，本文建構兩個包含房屋市場的動態隨機一般均衡模型，其中一個模型允許經濟個體用房屋作為擔保品，另一個模型則否。我們利用貝氏法估計兩個模型的參數，並使用上述模型所隱含的限制式來建立貝氏向量自我迴歸模型的先驗機率分配 (DSGE-VAR)，進而比較兩個模型與資料的配適度。以邊際概似值為衡量基準，我們發現房屋不能作為擔保品的模型比房屋可做為擔保品的模型佳。我們認為上述結果顯示以房屋為擔保品的借貸管道，不是影響台灣經濟活動的重要管道。同時反映資料中房價變動不能影響消費的實證發現。

According to National Wealth Statistics in 2018, real estate accounts for about 40% of the households assets. Therefore, house price fluctuations will have great influences to households wealth. In order to explore the connection between the housing market and the macroeconomy in Taiwan, this paper constructs two dynamic stochastic general equilibrium models with housing market. One model allows economic individuals to use houses as the collateral for mortgage loans, but the other model does the exact opposite. We apply the Bayesian method to estimate the parameters of the two models. Next, in order to compare the goodness of fit of the two models, we use the cross-coefficient restrictions implied by the above models to construct the prior distribution of the Bayesian vector autoregression (DSGE-VAR). Using the marginal likelihood as the benchmark, we discover that the model without house collateral is better than the model with house collateral. We believe this result shows the credit channel is not a key factor in affecting economic activities in Taiwan. The result also reflects the empirical finding: fluctuations in house prices do not affect consumption.