This paper studies the idiosyncratic risk component of individual house capital gains using data on resales and intermediate capital investments. The idiosyncratic component is large; its dynamics do not follow a random walk; and its magnitude is associated with proxies of information quality and market liquidity at the level of individual properties. Accounting for idiosyncratic risk substantially changes the assessment of the risk-return trade-off for housing: it reduces Sharpe ratios and makes them holding period dependent. I use a simple quantitative portfolio model to show that homeowners may be willing to make significant payments to insure against idiosyncratic housing risk.

We study the effects of underquoting, the practice of setting listing prices below sellers’ reservation values, on housing auctions. Laws introduced in Australia to deter underquoting lead to higher listing prices, but also to declines in sales prices and sales probabilities. We develop a quantitative model to formalize predictions under different assumptions about bidders’ information and rationality. The effects of the laws are matched by a version of the model in which participating bidders overbid. While both behavioral biases arising during the auction and switching costs can explain overbidding, empirical and survey evidence points to behavioral biases as the main mechanism.

Homeowners who originally bought when marketwide price levels were high (low) fetch high (low) sales prices and rents, even decades later. We study the propagation of reference-dependence to neighboring listings. The “spillover” reference point effect is about one-half as large as the “own” reference point effect. Neither house quality nor location appears capable of explaining the result. Using a simple model to provide empirical predictions, we find support for a competition-based mechanism. We quantify the aggregate effect of own and spillover reference point effects on aggregate prices and/or rents at the ZIP code level.