THE INS AND OUTS OF SELLING HOUSES: UNDERSTANDING HOUSING‐MARKET VOLATILITY

• Published date: 01 August 2024
• Author: L. Rachel Ngai,Kevin D. Sheedy
• Date: 01 August 2024
• DOI: http://doi.org/10.1111/iere.12693

INTERNATIONALECONOMIC REVIEW

Vol. 65, No. 3, August 2024 DOI: 10.1111/iere.12693

THE INS AND OUTS OF SELLING HOUSES: UNDERSTANDING

HOUSING-MARKET VOLATILITY∗

By L. Rachel Ngai and Kevin D. Sheedy

Imperial College London, London School of Economics, CEPR, and CfM, U.K.; London

School of Economics and CfM, U.K.

This article documents the role of inflows (new listings) and outflows (sales) in explaining the volatility and

comovement of housing-market variables. An “ins versus outs” decomposition shows that both flows are quan-

titatively important for housing-market volatility. Thecorrelations between sales, prices, new listings, and time-

to-sell are stable over time, whereas the signs of their correlations with houses for sale are found to be time-

varying. A calibrated search-and-matching model with endogenous inflows and outflows and shocks to housing

demand matches many of the stable correlations and predicts that the correlations with houses for sale depend

on the source and persistence of shocks.

1. introduction

The importance of search frictions in buying and selling houses is widely acknowledged,

with buyers and sellers spending considerable amounts of time searching. The essence of the

search approach to markets is to understand how the stocks of buyers and sellers evolve

through inflows and outflows. Applied to the labor market, this has been the subject of an ex-

tensive literature. However, for the housing market, there has been little work that aims to

understand inflows and outflows jointly, especially with regard to cyclical fluctuations.

This article assembles a collection of stylized facts about the cyclical properties of a broad

set of U.S. housing-market variables over the last three decades, including house prices and

the key stocks and flows, comprising houses for sale, sales transactions, new listings, and the

average time taken for houses to sell. A calibrated search-and-matching model with both en-

dogenous inflows (new listings) and outflows (sales) is used to explain the empirical findings.

One contribution of the article is to document two novel facts. First, both inflows and out-

flows are quantitatively important in understanding housing-market volatility. This is shown

using an “ins versus outs” decomposition of the type that has been applied to the labor mar-

ket. Here, the stock of houses for sale is the equivalent of unemployment, the evolution of

which depends on the difference between new listings and sales. The second novel fact is that

houses for sale does not have a stable correlation with house prices, sales, or new listings,

whereas correlations among all other pairs of variables remain stable. The correlations among

prices, sales, and new listings are all positive, whereas the correlations of these with time-to-

sell are all robustly negative (with the possible exception of prices). In contrast, though the

∗Manuscript received July 2020; revised November 2023.

We are grateful to the editor and four anonymous referees for their comments. We thank Adam Guren, Morris

Davis, Mike Elsby, Martin Gervais, Lu Han, Chris Pissarides, and especially Allen Head for helpful discussions, and

Thomas Doyle and Christopher Jenkins for assistance with the data. Wealso thank seminar participants at Essex Uni-

versity, the REDg Dynamic General Equilibrium Macroeconomics conference, the Search-and-Matching Research

Group conference, the Spring Housing-Urban-Labor-Macro conference, and the Society for Economic Dynamics An-

nual Conferences for their comments. Rachel Ngai acknowledges support from the British Academy Mid-Career Fel-

lowship. Please address correspondence to: L. Rachel Ngai, Department of Economics, London School of Economics,

Houghton Street, WC2A 2AE, UK. E-mail: l.ngai@lse.ac.uk.

1415

© 2024 The Authors. International Economic Review published by Wiley Periodicals LLC on behalf of the Economics

Department of the University of Pennsylvania and the Osaka University Institute of Social and Economic Research

Association.

This is an open access article under the terms of the Creative Commons Attribution-NonCommercial-NoDerivs Li-

cense, which permits use and distribution in any medium, provided the original work is properly cited, the use is non-

commercial and no modifications or adaptations are made.

1416 ngai and sheedy

correlation of houses for sales with time-to-sell has been positive throughout the period stud-

ied, the correlations of houses for sale with prices, sales, and new listings have changed from

positive to negative in recent times.

A second contribution of this article is to demonstrate two new quantitative results us-

ing a stochastic search-and-matching model with endogenous inflows and outflows. Central

to the model is the idea of idiosyncratic match quality between a house and its owner, and

the dynamics of the distribution of ongoing match quality. Decisions to buy houses are de-

scribed by a cutoff rule whereby a sale occurs when a draw of new match quality is above

a certain threshold. Individual match quality is a persistent variable, but is subject to occa-

sional idiosyncratic shocks that degrade it. After such shocks, homeowners decide whether

to move house, and the moving decision is also described by a cutoff rule for match quality.

These decision processes give rise to an endogenous distribution of match quality across all

homeowners.

The first novel quantitative result is that housing-demand shocks coming from changes in

interest rates and expenditures complementary to housing can explain most of the patterns of

comovement among housing-market variables. In the model, since moving house represents

an investment in match quality, interest rates affect the incentive to invest in better match

quality by changing the relative importance of future payoffs compared to current costs. A

fall in the real interest rate increases the total surplus from a transaction and raises the price

paid by buyers. Hence, a lower interest rate has a positive effect on house prices and new

listings. A positive expenditure shock, associated with an increase in the flow utility received

from occupying a house, raises the total surplus from a transaction and thus increases house

prices. This shock increases the rate at which transactions occur, lowering time-to-sell. The

positive expenditure shock also boosts homeowners’ incentives to invest in better match qual-

ity by moving house, which leads to a rise in new listings, and these listings ultimately result in

more sales.

Match quality plays a crucial role in the workings of the model and its ability to explain

the stylized facts. The presence of a distribution of new match quality is central to generating

a positive correlation between sales and prices. Given the equilibrium distribution of match

quality among existing homeowners, a persistent housing-demand shock increases the incen-

tive to invest in better match quality, leading to more listings. This explains the positive corre-

lation between new listings and sales and prices.

The second quantitative result is that the model predicts different correlations between

houses for sale and other variables when there is a change to the source or persistence of

housing-market shocks. By simulating the model for two subsample periods, the lower mea-

sured persistence of the empirical proxy for the housing-demand shock can explain the switch

from positive to negative in the correlations of houses for sale with sales, prices, and listings,

as is seen empirically in recent times. Therefore, the model can offer an explanation of why

the signs of the correlations between houses for sale and other variables have not been stable

over time, while also being consistent with most of the empirically stable correlations among

other housing-market variables.

The relative importance of interest-rate and expenditure shocks also matters because pos-

itive housing-demand shocks from these two sources have opposite effects on time-to-sell.

Lower interest rates raise the return to searching and thus increase time-to-sell, leaving

more houses on the market. In contrast, positive expenditure shocks increase the desire

to complete transactions and hence reduce time-to-sell, depleting the stock of houses for

sale.

The main reason for the switch in the sign of the correlations between houses for sale and

sales, prices, and new listings is a reduction in the measured persistence of the expenditure

shock in the second subsample. The key point is that new listings rise by more than sales with

a more persistent shock, which increases the stock of houses for sale. On the contrary, the less

persistent shock fails to induce enough moving to replenish the stock of houses for sale. This

explanation comes from understanding moving decisions as investments in match quality: a