The volatility of housing markets
Housing markets exhibit excess volatility; house prices are more volatile relative to both rental prices and aggregate income. Further, prices and transaction volumes move in tandem, and are negatively correlated with the average time on market for sellers.
Several explanations have been proposed for these patterns, including financial constraints and down-payment effects (Stein 1995, Ortalo-Magné and Rady 2006), and departures from rational expectations (Shiller 2005). A number of authors (e.g. Wheaton 1990, and Krainer 2001, among others) have observed that shocks in the housing market induce movements not only in prices and quantities, but also in the seller time on market. This liquidity channel points to the importance of search frictions – the fact that it takes time and resources for buyers and sellers to meet and trade – for understanding housing market volatility.
The effect of moving homeowners’ behaviour on the housing market volatility
In a recent paper (Moen et al. 2015), we argued that the search behaviour of moving homeowners can be an important source of housing market volatility. By choosing whether to buy or sell first, households who move house affect the ratio of buyers to sellers and market liquidity, with important consequences for time on market, transaction volume, and prices.
Moving owner-occupiers have to buy a new property and sell their old residence. Since housing is an illiquid asset, there may be substantial delays between the two transactions. This is evident in Figure 1, which plots the distribution of the time difference (in days) between the ‘buy’ and ‘sell’ transactions for homeowners moving house in Copenhagen between 1993 and 2008.
Figure 1. Distribution of the time difference between ‘buy’ and ‘sell’ transactions for moving owner-occupiers, Copenhagen 1993-2008
Source: Own calculations based on registry data from Statistics Denmark
These transaction delays can be very costly, for example, because of the necessity to service two mortgages at the same time, or because the household if forced into temporary renting. Homeowners can choose the sequence of their transactions to minimise these costs. Specifically, they can condition the sequence on the state of the housing market. Indeed, as Figure 2 shows for the case of Copenhagen, the fraction of moving homeowners that buy first correlates strongly with the state of the housing market. In particular, in the ‘hot’ housing market period from 2004 to 2006, when house prices were on the rise, sales were high and seller time on market was low, a substantial share of homeowners (up to 85%) bought first. This share quickly dropped to less than 40% with the cooling of the housing market in 2007-2008.
Figure 2. Housing market dynamics, Copenhagen 2004-2008
Source: Danish Mortgage Banks’ Federation, Stats Denmark and own calculations based on Stats Denmark registry data.
We argue that this pattern can be understood as an equilibrium outcome resulting from the feedback between housing market conditions and homeowners’ transaction sequence decisions.
- We show that it is optimal for moving owner-occupiers to buy first when there are more buyers than sellers in the market, and to sell first when there are more sellers than buyers.
However, having owner-occupiers buy first crowds the buyer side of the market, resulting in a housing market with more buyers relative to sellers.
- Conversely, having owner-occupiers sell first results in a market with more seller than buyers.
The interplay between the behaviour of individual homeowners and the buyer-seller composition of the housing market leads to multiple equilibria: One with a high ratio of buyers to sellers and a short seller time on market, and one with a low buyer-seller ratio and long seller time on market. Switches between these two equilibria are associated with an increase in the seller time on market and the stock of houses for sale, as well as with a drop in transaction volume and prices, consistent with the housing market turnaround shown in Figure 2.
Why would a moving homeowner want to buy first when there are many buyers relative to sellers? The reason lies in the costly delay between the two transactions. Suppose it is more costly to extend the time between two transactions (for example, due to mortgage costs associated with double ownership or due to short-term renting at a premium) compared to extending the time spent in one’s old home prior to the completion of the first transaction. In that case, homeowners would want to minimise the delay between transactions. A market with more buyers relative to sellers exhibits short time on market for sellers and longer time on market for buyers. Minimising the delay between transactions in this case entails becoming a seller after the first transaction. In this situation, a moving homeowner would buy first. Following the same logic, a homeowner would want to sell first whenever the ratio of buyers to sellers is low.
We also show that if households expect housing prices to rise (fall), their incentives to buy first (sell first) are stronger. The reason for this is that conducting a sequence of transactions in an illiquid asset market like the housing market can expose a trader to price movements that he cannot hedge against. For example, buying first implies that a household has a long position on housing in between the two transactions. Any anticipated rise (fall) in house prices thus make it less (more) costly to buy first irrespective of the buyer-seller ratio in the market. This behaviour tends to destabilise the housing market and lead to self-fulfilling fluctuations in prices and the buyer-seller ratio.
Overall, our research points to the importance of taking liquidity effects in the housing market seriously. Liquidity considerations can have important amplification effects and can improve our understanding of housing market volatility.
This article is published in collaboration with Vox EU. Publication does not imply endorsement of views by the World Economic Forum.
To keep up with the Agenda subscribe to our weekly newsletter.
Author: Espen R. Moen is a Professor of Economics, BI Norwegian Business School. Plamen Nenov is an Assistant Professor, Norwegian Business School. Florian Sniekers is a PhD candidate, University of Amsterdam, VU University Amsterdam, and the Tinbergen Institute
Image: Property sale and rental signs are seen next to a street sign in London. REUTERS/Toby Melville