Data and Code for: Betting on the House: Subjective Expectations and Market Choices

Home price expectations play a central role in macroeconomics and finance. However, there is little direct evidence on how these expectations affect market choices. We provide the first causal evidence based on a large-scale, high-stakes, and naturally occurring field experiment in the United States. We mailed letters with information on trends in home prices to 57,910 homeowners who had listed their homes on the market. Collectively, these homes were worth $34 billion. We randomized the information contained in the mailing to create non-deceptive, exogenous variation in the subjects home price expectations. We then used rich administrative data to measure the effects of these information shocks on the subjects' market choices. We found that, consistent with economic theory, higher home price expectations caused a reduction in the probability of selling the home. These effects were highly statistically significant, economically large in magnitude, and robust to a number of sharp checks. Our results indicated that market choices were highly elastic to expectations: a 1 percentage point increase in home price expectations caused a 2.63 percentage point reduction in the probability of selling the property within 12 weeks.

住房价格预期在宏观经济学与金融学领域占据核心地位。然而，目前鲜有直接证据能够阐明此类预期如何影响市场决策。本研究基于美国境内一项大规模、高利害且自然发生的实地实验，首次提供了因果性实证证据。我们向57910名已将房屋挂牌上市的房主邮寄了包含房价趋势信息的信件，涉及房屋总市值达340亿美元。我们对信件所载信息进行随机化处理，以此在受试者的房价预期中引入无欺骗性的外生变异。随后，我们借助丰富的行政数据，量化评估此类信息冲击对受试者市场决策的影响。研究结果与经济学理论一致：更高的房价预期会降低房屋售出概率。该效应具有高度统计学显著性，经济效应规模可观，且经多项严苛检验证实其稳健性。我们的研究表明，市场决策对预期具有高度弹性：房价预期每提升1个百分点，将导致房屋在12周内售出的概率降低2.63个百分点。