Multiple Transactions Model: A Panel Data Approach to Estimate Housing Market Indices

In this paper, a multiple transactions model with a panel data approach is used to estimate housing market indices. The multiple transactions model keeps the same features of the repeat transactions index model (i.e., tracking the price appreciation of same houses). However, the multiple transactions model overcomes the shortcomings of the repeat transactions model by avoiding the correlated error terms. The indicative empirical analysis on a small sample of actual house transaction data demonstrates that the proposed multiple transactions model is superior to the repeat transactions model in terms of index variance, robustness of estimate, index revision volatility, and out-of-sample prediction of individual house prices.