Do long-distance moves discourage homeownership? Evidence from England

Highlights

• Longer-distance moves decrease the reliability of information on the destination housing market and increase the risk of ill-informed housing purchase decisions

• Long-distance movers have a 5.5 percentage point lower probability of owning their next home compared to shorter-distance movers

• Long-distance movers are aware that they possess less reliable information

• Long-distance movers, especially renters, are more likely to move again quickly after accruing better information on the local area

Abstract

We hypothesize that as the distance of a residential move increases, the amount and quality of information collected on the destination housing market fall. This in turn increases the chances of making an ill-informed housing purchase decision, thus reducing the likelihood of such a purchase. Using data from the Survey of English Housing from 1993 to 2008, we document that, consistent with our prior, households moving over long distances – defined as 50 miles or more – have, on average, a 5.5 percentage point lower probability of owning their next home compared to shorter-distance movers. We also provide evidence consistent with the views that long-distance movers (i) are aware that they possess less and/or lower quality information and (ii) are more likely, especially if they are renters, to move again quickly after presumably having accrued better information on the property and local area.

Introduction

The decisions ‘whether to own or rent a home’, ‘where to buy’, and ‘what property to buy’ are all risky. In this paper, we hypothesize that as households move farther away from their original residence, search in the destination housing market becomes more difficult and more costly.¹ This difficulty of search largely arises from the heterogeneity of housing and neighborhood in terms of their characteristics and their location. As moving distance increases, the amount and quality of information collected on the destination housing market and individual housing units decrease, and, therefore, purchase decisions become riskier.

A prospective buyer of a property arguably faces risks in multiple dimensions. Risks can be property or neighborhood specific. They can also be idiosyncratic (leading to mismatch) or general/systemic. And there are interactions in these dimensions. For example, some risks are idiosyncratic in nature and related to the property (i.e., I hate the house that I just bought), idiosyncratic and neighborhood specific (i.e., I can't get to my work place as easily as anticipated), systemic and property specific (i.e., a leaking roof) or general and neighborhood specific (i.e., crime is higher than expected or rising after moving in).²

Owner-occupiers are more exposed to these risks than renters because ill-informed purchases cannot be easily reversed. Renters may also choose the ‘wrong’ neighborhood in remote destination markets, but this has less grave consequences as, in contrast to owner-occupiers, renters do not have to bear the capital loss associated with the sale of a home plus they face much lower housing transaction costs and can therefore move more easily.³ Moreover, if say a neighborhood turns bad, this should, at least in the longer-run, also be reflected in lower subsequent rents or smaller rent increases, compensating the renter for the bad event occurring. This line of reasoning is supported by Simonsohn and Loewenstein (2006) who find that renters move repeatedly to adjust their housing expenditure.⁴

A reasonable strategy for a mover to a remote destination market would therefore seem to be to rent a property first and delay a home purchase until more (reliable) knowledge can be accumulated about the new neighborhood and property stock. As a consequence, we would expect that, all else held constant, households that move farther away from their original residence are less likely to own their next property than households that move nearby. This is the main hypothesis that we put to the data.

Our empirical analysis employs data from the Survey of English Housing (SEH). The SEH is a rich dataset that provides essential information such as the housing tenure status of households (the dependent variable), the distance moved (used to compute our key explanatory variable) and various demographic and socioeconomic characteristics of households and household heads (the control variables). Controlling for demographic and socioeconomic characteristics of mover households helps mitigate concerns of spurious correlation and omitted-variable bias. A crucial additional advantage of the SEH is that it provides information on the pre-move conditions of households. In particular, the tenure status at the previous accommodation helps control for unobservable preferences and the ability to own of households.

Our empirical analysis reveals three novel insights. Firstly, we provide evidence in favor of our main hypothesis that the probability to own is adversely affected by a lower quantity and quality of information as proxied by the long-distance move dummy variable, all else equal. Our analysis reveals that the negative effect is not only highly significant in a statistical sense but also reasonably meaningful in an economic sense. Based on our most rigorous specification, the average marginal effect of a long-distance move (longer than 50 miles) as opposed to a shorter-distance one is to decrease the probability for a household to own their next home by 5.5 percentage points.

Secondly, we demonstrate that long-distance movers have less negative subjective assessments of specific problems – crime, vandalism, litter and graffiti in the neighborhood – than short-distance movers, conditional on the objective neighborhood quality and the resulting equilibrium house prices. We verify by contradiction that this is consistent with the view that long-distance movers are aware of the fact that they possess less reliable information on the destination housing market.

Thirdly, we test the hypothesis that the adverse effect of a lower quantity and quality of information, as proxied by the long-distance move indicator, on the probability to own is of a temporary rather than permanent nature. We find that the length of stay at the destination market is adversely affected by long-distance moves and this negative effect is stronger for private renters, consistent with the view that the optimal strategy for longer distance movers is to rent first and accumulate information on the destination housing market prior to making a momentous investment decision.

Our paper is structured as follows. In Section 2, we discuss the findings of previous related studies, clarify the contribution of our paper to the literature and derive empirically testable hypotheses. Section 3 describes the data, outlines our empirical strategy and presents our findings. The last section concludes.