Conference Agenda

This study explores the possible impacts of short-term house trades, often called “house flips”, on local housing markets. Using nationally representative data across multiple U.S. housing markets, we test the Spillover Hypothesis and the alternative Competition Hypothesis for the influences of flips on the neighborhood non-flip house transactions. We find that the local flip tendency is negatively associated with the subsequent neighborhood non-flip transactions’ house holding time, and positively associated with those transactions’ performances, after we control for market conditions that affect both flip and non-flip transactions and other possible determinants for holding time and performance. The effects are generally more prominent when the local flip tendency is measured for a longer run, or for a larger neighborhood area, supporting the Spillover Hypothesis. The results are persistent regardless of whether the local flip tendency is defined as the proportion of flips in the local resales by transaction number or by transaction volume. The performance spillover remains after we incorporate transaction costs and tax effects, and is likely to be related to the market timing skill spillover at the property listing/resale time. These findings demonstrate the substantial influences of flip activities to the local housing market turnover and price changes.

In this paper, we study the causal effect of information friction on housing market outcomes using an administrative micro-level housing resale transaction dataset in Beijing, China. We apply a boundary discontinuity design and difference-in-differences model to explore the prohibition of school district marks in the online listing information, which increases the information friction during housing search. Our results show that the prohibition leads to a reduction of 2.55% in transaction prices and a 22.99% increase in the seller's time on the market for the houses corresponding to the key primary schools relative to the other houses. The longer time on the market is mainly due to the increased difficulty for potential buyers in finding desirable dwellings. Moreover, the changes in buyers' viewing behavior and sellers' time on the market prompt the sellers to lower the listing prices to attract potential buyers and lead to lower transaction prices. Overall, the empirical evidence highlights the importance of information friction during the housing search process.

Digital finance has witnessed a rapid explosion in China and plays a significant role in individuals' behavior. Using the China Household Financial Survey (CHFS), this study employs two-stage least squares (2SLS) regressions to estimate how the housing price affects entrepreneurship, further exploring how digital inclusive finance alters such an effect. The result shows that higher housing price discourages entrepreneurship and more developed digital inclusive finance encourages entrepreneurship. In addition, digital inclusive finance alleviates the negative impact of housing prices on entrepreneurship by mediating the crowding-out effect mechanisms. This study yields implications on how to utilize the housing market development and financial market to spur entrepreneurship.

House price forecasting has received worldwide attention due to its significance for residents, development companies and the government. However, there are some limitations (such as sparsity and complexity) in data, which brings huge challenge for modeling. This paper proposes a novel hybrid model integrating Baidu Search Index for predicting housing prices index. The proposed hybrid model includes four steps: first, we obtained web search data and extracted core indicators from the Baidu search engine to build the Internet Concerns (IC). Secondly, house price index is decomposed by variational modal decomposition (VMD), and the Envelope Entropy (EE) was introduced to determine the number of mode decompositions. Thirdly, we use least squares support vector machine (LSSVR) to predict each input integrated by each mode and IC, Enhanced levy bat algorithm (ELBA) was used for parameter selection. Fourthly, generating the final result by adding the forecasting results of each input. The experiment was carried out on the Beijing’s new commercial housing sales price index. The experiment results show that our model can effectively capture the characteristics of nonlinear and nonstationary time series data and achieve higher prediction accuracy. Thus, it acts as an effective mechanism for house price prediction research.