HOME Recent Events RMI in the News ALUMNI
  Issue 6 | Archive  

February 2011

Why Do Sellers Hold Out in the Down Market? An Option-based Explanation

A paper by Wenlan Qian (National University of Singapore).

In the residential housing market, sellers tend to hold onto their homes when prices fall. RMI affiliated researcher Dr. Wenlan Qian's working paper offers one rational explanation of the "hold-out" phenomenon in down times in the housing market. Motivated by the concept of real options, the paper models the owner's selling decision as an endogenous optimal stopping time. It provides a theoretical basis for understanding the market determinants of home owners' listing and selling decisions. Evidence from U.S. aggregate residential housing data is consistent with the theoretical model. Cross sectional variation in transaction volume are strongly associated with the asset expected growth rate and volatility; and their effects are amplified in areas with low supply elasticities and in times with low market prices.

In the theoretical model, in response to some random shocks, people change their demand for owner-occupied housing. An existing home owner becomes a potential seller whose reservation price, in the model, is equal to a fixed transaction cost. However, buyer's reservation price for the home consists of two components. One is the equivalent rental cost she needs to pay during her duration of stay if she chooses to live in the rental housing instead. Since home ownership, compared to renting, grants owners resale benefits, buyer's reservation price also includes the resale option component in the event she needs to move and sell the house. As a result, gain from trade is a time-varying endogenous function of economic fundamentals. Each seller and buyer pair's encounter is modelled as an alternating bargaining game (Rubinstein and Wolinsky, 1985) in which the time-varying trade gain is split based on the population size of the buyer and seller group.

Given the pricing rule and the underlying rent process, the decision on the sale timing becomes an optimal stopping problem for the seller to exercise her resale option. Given a nonzero constant trading cost, there will be a critical value such that sellers will exercise their resale option as long as the market rent exceeds this threshold. When the rent process is below the threshold, however, sellers will be better off waiting since the gain from trade is not sufficient enough to offset the value associated with further waiting. As a result, in post-boom markets when economic fundamentals are weak, sellers hold on to their resale option and wait until the market conditions get better. This leads to a decrease in the intentions to sell resulting in a decrease in trading volume. The paper also studies the role of supply conditions on the exercise of the resale option. Specifically, when land supply can be competitively added, the exercise threshold is smaller and sellers wait less than the case where no new construction is allowed.

In the empirical section, the evidence is broadly consistent with the predictions implied by the theoretical model. First, the model predicts that the threshold (volume) is decreasing (increasing) in the growth rate of the rent process and is increasing (decreasing) in the rent volatility. She finds a one percentage increase in the rent cap rate - a measure that is inversely related to the expected rent growth rate - is associated with a decrease in transaction volume by roughly 3-4%. Second, transaction activities decrease (increase) in the rent growth rate (volatility) at a faster rate if supply is more inelastic. Given the same cap rate, being in the supply inelastic state (i.e. in the top 10% of the population density distribution) would further decrease on average transaction volume by 7%. If the market is in inelastic supply state, a one percentage increase in the rent volatility is associated with an 8% decrease in average transaction volume. However, if the market is in elastic supply state, the rental volatility is not associated with variation in transaction volume. Third, the model implies that the option effect on the trading volume should be stronger in the down market, where the no-trade threshold is more binding. Consistently, the paper documents a lower transaction volume associated with a lower rent expected growth rate or a higher rent volatility especially when the U.S. housing market was experiencing a trough period.

Overall, this paper provides a rational option-based explanation for delayed trading decisions in the post-boom housing markets. Decrease in mobility during down times could be the result of home owners' optimal response rather than the suboptimal outcome due to financial constraints or behavioral limitations.

Back To Newsletter

Published quarterly by Risk Management Institute, NUS
Editor: Ivy Wang (rmiwy@nus.edu.sg)