RDP 2014-06: Is Housing Overvalued? 1. Introduction

This paper examines whether it is more expensive to own a house or to rent. We assess houses as ‘overvalued’ if home buyers pay too much, in the sense that they would be better off renting than buying. This involves comparing the financial cost of renting a home with the cost of owning a similar dwelling, where the latter depends on the purchase price, interest rates, repairs, council rates and so on. We briefly also examine non-financial costs but find these are small, on average.

We decompose housing values into contributions from rents, interest rates, expected appreciation and other factors, which we hope will be directly useful to potential buyers. The decomposition may also be useful to market participants, policymakers and others who need to understand the reasons for house price movements. For example, we find that the boom in house prices in 2002–2003 can largely be attributed to expectations of further capital appreciation.^[1] That has implications for lending and prudential standards. Interest rates and rents have been more important determinants of house prices at other times, with a different set of policy implications. Our estimates can be readily updated, which may assist in the early detection of bubbles.^[2]

Given that the supply of housing is fixed in the short run, prices are determined by how much buyers are willing to pay. Hence a comparison of the costs of home ownership with the costs of the nearest alternative seems central to a measure of overvaluation. In contrast, other popular measures of overvaluation, such as the price-to-income ratio, are not obviously a part of any individual's decision-making process. We compare various measures of overvaluation in the next section.

As we discuss in Section 2, our paper contributes to a large literature that compares house prices to rents and the user cost of housing (a term we define precisely in Section 3). Our paper is unusual, though not unique, in two important respects. First, we focus on conditions in Australia. Second, we use a new dataset that matches prices with rents for a large representative sample of properties. In contrast, most previous comparisons of the cost of owning and renting have used different and inconsistent data sources for house prices and rents. Because houses that are bought differ from those that are rented, in both observable and unobservable ways, it has been difficult to discern whether differences in cost reflect differences in quality. Accordingly, researchers could only compare changes in prices with changes in rents. Even then, they have needed to assume that quality changes are controlled for similarly in the two series. This assumption becomes increasingly doubtful over longer periods. In contrast, our matched data enables comparisons of the level of prices with the level of rents. Hence, we can estimate the level of overvaluation. It also facilitates an analysis of other interesting properties of dwelling prices, such as their implications for expected capital appreciation.

To summarise our results, we find that assessments of house prices are sensitive to assumptions about expected capital gains. If real house prices were to continue to grow at the average rate of the past six decades, then buying a house now would be about as costly as renting. To put this another way, the expectations of future capital gains implied by current house prices are in line with historical norms. That allays some concerns about a housing ‘bubble’. If house price growth were to be slower than the historical average, as some forecasters predict, then the average home buyer would be financially better off renting.^[3]

These findings relate to average housing conditions, around which individual circumstances will differ. For example, a household expecting historically average capital appreciation will be better off owning than renting if it values home ownership for non-financial reasons, if it expects to remain in the house for longer than average, or if it has substantial financial savings that it cannot profitably invest elsewhere. Given that individual circumstances vary, no-one should base personal investment decisions solely on our estimates. However, we do hope that our approach provides a guide to how these decisions could be made. We also hope that our detailed estimates, which are based on average conditions, are useful when information on individual conditions is unavailable.

Several limitations of the paper (shared by other research on the user cost) are worth noting. First, our analysis is partial equilibrium. We focus on the home-buying decision; this involves comparing prices to rents and expected appreciation, which we take as given. Comparisons of prices to other benchmarks would be relevant to other decisions. For example, a comparison of prices to construction costs would be relevant to builders. Comparing current prices to future prices would be relevant to deciding when to buy or sell. A general equilibrium analysis would explore how all these decisions might be made consistent. But these comparisons are beyond the scope of this paper. Put slightly differently, we examine whether house prices are in line with rents. A broader study could examine whether housing prices and rents are jointly over or undervalued relative to other consumer prices.

Second, we only examine purchases by owner-occupiers, who account for two-thirds of all dwellings. Investors make similar decisions, but these are complicated by taxes.

Third, we focus on whether households are financially better off buying or renting their house and by how much. The decision to buy rather than rent also reflects subjective factors that are difficult to measure such as security of tenure, freedom to renovate, access to finance, pride of ownership, the risk of capital losses and the flexibility of moving. However, although these considerations are important at an individual level, at an aggregate level they seem to cancel out. As we discuss in Section 5.2, non-financial considerations do not seem to have a substantial effect on the average price level. Even if they did, it seems neither useful nor feasible to tell households what their subjective preferences are. In our view, it is informative to calculate average financial costs – about which potential buyers are presumably interested – and let individuals decide for themselves whether these are worth incurring.

Footnotes

We follow common usage in using the term ‘house prices’ to refer to both detached houses and units except when the distinction is material. [1]

Stiglitz (1990, p 13) defines a bubble: ‘if the reason that the price is high today is only because investors believe that the selling price will be high tomorrow—when “fundamental” factors do not seem to justify such a price—then a bubble exists’. [2]

Full disclosure: during the preparation of this paper one of the authors, Peter Tulip, bought a house. The other author, Ryan Fox, continues to rent. [3]