![Sample our Built Environment journals, sign in here to start your access, Latest two full volumes FREE to you for 14 days]({"type":"image" , "src" : "/sda/5926/TOC-Subject-Sample-2021-Built-Environment.jpg"})
Abstract
This study emphasizes the importance of asset-based – as opposed to agent-based – housing liquidity measures, a house’s time on the market, and the time to buy for its buyer rather than its seller. It uses unique survey data to analyze spatial differences in the imbalance between these two measures in boom and bust markets of the 2000s. We find evidence of higher overall imbalance and its inter-ZIP code area dispersion in the booming years. Buyer search intensity diminishes the imbalance in boom and intensifies it in bust. Other revealed significant covariates have different effects under the two market conditions. Our model explains the liquidity imbalance variations on the ZIP code area level well, but not on the individual property level, highlighting the importance of regional idiosyncratic transaction dynamics.
ACKNOWLEDGEMENTS
We are grateful to Bill Watkins and the Center for Economic Research and Forecasting at the California Lutheran University and The First American Title Company in Santa Barbara, California, for assistance with our surveys. We also thank the University of California Santa Barbara and California State Polytechnic University Pomona for the financial support. We are very grateful to the two anonymous referees for their valuable comments. A previous version of this article won an award for the best article in real estate cycles category sponsored by Pyhrr/Born Trust for Real Estate Cycle Research presented at the 2018 annual conference of the American Real Estate Society.
Notes
1 This should hold in theory if the population size and homeownership rate are stable for owner-occupants and if the market is equally represented by first-time buyers and last-time sellers.
2 The National Association of Realtors’ (NAR) Home Buyer and Seller Survey appears to be the most widely used data source for house buyer duration, as shown in our review of rather limited relevant literature on this topic. The NAR survey was conducted biannually until 2003, and annually afterwards. According to the NAR website, it was first administered in 1981 with just 59 questions. The version from 2017 contained 131 questions. Today the survey can be taken on paper or online, and in English or Spanish. Over three decades of data collection has been revealing trends in “the demographic characteristics of home buyers and sellers, buyers and sellers’ experience in the home transaction process, as well as market characteristics including the use of real estate agents.” The survey targets a national sample of recent homebuyers who purchased a primary residence in the 12-month period between July and June. The recent homebuyers' names are provided by Experian, which has a database with such information derived from county records.
3 It may be natural to question the accuracy with which time to buy (TTB) is measured. While its endpoint is easily measurable with the exact date a buying contract is signed, the starting point of the duration may be largely subjective. Different buyers may perceive differently the time when they started their search, whether it is checking an online source, such as Zillow, for the very first time, or contacting a real estate agent, with the two being potentially weeks or months apart. The average search duration should reflect a point somewhere in the middle between these two possible starting points. In our study we find a very striking and statistically significant difference between the market average TTB in the boom and bust time periods of the housing bubble. The same can be said about the accuracy of time on the market (TOM). It may be largely understated if a house is delisted and relisted but TOM only reflects the time that elapsed since the most recent listing date.
4 The addresses were provided by the First American Title Company in Santa Barbara, California.
5 The questionnaire was mailed to recent house buyers in the tri-county asked: “When did you first start actively searching for a house? When did you finish searching and signed the house-buying contract for the house?” It also clarified that “active” search means, for example, looking for a house “via the Internet or journals, through friends or realtors, by contacting house seller or by visiting houses for sale on your own.” The two questions then asked to provide either the month and year of the beginning and the end of the search, or to directly write the number of months and/or weeks that the buyer spend on search.
6 Out of all returned questionnaires, 62 percent in the boom market and 61 percent in the bust market had the information on the time on the market for the purchased house. The corresponding question in the survey asked the following: “Please indicate, if possible, the total time the house was for sale by its previous owner, before you purchased it: __ months, __ weeks.”
7 Because our mailed questionnaires cover a random sample of all recent house buyers, we assume no bias in responses coming from certain areas. In other words, ZIP code areas with higher number of responses should reflect a higher actual buying-and-selling activity in that market.
8 An important factor of the bust market is the high number of houses that were sold due to foreclosure. According to a study done by Mayer, Pence, and Sherlund (Citation2008) at the Federal Reserve, California and Florida represent top two states that account for about 40% of the foreclosure filings during 2008. As of middle of 2008, these two states together with Arizona and Nevada ranked highest in nation by the percentage of mortgage loan defaults: 26% of subprime loans originated in 2006, and 18% of all subprime loans originated in 2007, as compared to only 13% and 9%, respectively, for the nation as a whole. In our questionnaire we inquired house buyers whether the reason for the house sale was known. Foreclosure was among most common reasons provided by the respondents. We tested the inclusion of Foreclosure (constructed as a dummy variable on property level, which equals 1 if the purchased house was sold due to foreclosure, and 0 otherwise) in our regression model for the bust time period. Out of all returned questionnaires, 15.25% indicated foreclosure as the reason for sale. The bust regression models that include Foreclosure show a positive, but statistically insignificant, coefficient. The insignificance may be due to insufficient number of respondents that indicated a known reason for sale. For more active ZIP code areas this potential issue implies understatement of the percentage of homes sold due to their owners’ financial difficulties. For less active ZIP code areas this implies severe under- or over-statement of the presence of such sales. The empirical regression results that include Foreclosure are available upon request.