The power of discounting

Across the industrialised world house prices appear extremely high. Much of the commentary in the UK, where house prices have become a national obsession over the years, continues to bemoan the elevated level of house prices, arguing that it is unfair on the younger generation of potential buyers who have been shut out of the market. But it is not just a UK problem: The Riksbank in Sweden has been worrying about excessively high house prices for what seems like much of the last decade; the Swiss National Bank has made references to high residential property prices and even in Germany we hear concerns at the extent of valuations.

We can value house prices in one of two main ways, depending on whether you are on owner-occupier or a renter. An owner-occupier is concerned about affordability relative to their income so the price-to-income multiple is an appropriate measure. In the UK, average house prices are worth around 5x the annual income of a first time buyer versus a multiple of 3.5x over the period 1983 to 2007. On the assumption that they are unable (or do not want) to borrow 100% of the purchase price, this implies that prospective home buyers either have to save for longer to build up a deposit, or they put down a lower deposit and borrow more (i.e. they are more highly leveraged). On the basis of the OECD's data, UK house prices relative to incomes are 28% above their long-term average. But valuations are more extreme elsewhere. The data suggest that prices relative to incomes in Sweden are 44% above their long-run average; in Australia this rises to 46% and 56% in Canada (chart 1).

Renters care about their short-term rental costs relative to income, and I have used the OECD’s data on house price-to-income ratios and price-to-rent ratios to construct indices on rental costs relative to incomes. For the most part, rents appear to have grown more slowly than incomes over the last decade. Across the OECD as a whole, they are around 12% down relative to their long-term average versus 4% below in the euro zone (chart 2). But the cost of buying houses relative to the stream of rental income has increased sharply. UK house prices relative to rentals are currently 44% above their long-term average, whilst the corresponding figures for Sweden and Canada are 72% and 97%.

One of the standard arguments used to explain the elevated level of house prices is that the great recession of a decade ago impacted severely on residential construction, chopping out at least a couple of years of construction that would otherwise have been expected to take place to cope with rising populations. In the UK, this has been compounded by the lack of public sector housing, which has collapsed in recent decades. As recently as the 1970s, more than 40% of all housing completions were in the public sector. In the period 2003 to 2010 it collapsed to 0.1%. But as a couple of neat blog posts on the BoE’s Bank Underground website make clear (here and here) a lack of supply is not the whole story.

Like any asset, housing is valued according to the stream of services it yields. If we think of a dividend discount model, in which the price is determined as the future value of housing services discounted by the interest rate, it stands to reason that as interest rates collapse to all-time lows, so the discounted future value of future services (i.e. prices) will rise sharply. Housing services are valued in terms of rental income: If you buy a house and rent it out, you can directly value the income. If you buy a house to live in, the standard statistical approach is to measure the rent that you would otherwise pay to live in that house if you were a renter (imputed rent). But whichever method we use, low interest rates have clearly boosted house prices. This simple fact explains why prices have shot up hugely whilst rental costs have not – it is the power of discounting that has distorted the price-to-rents ratio. The authors of the BoE blog posts use this insight to construct a model which concludes that the rise in UK house prices since 2000 cannot be attributed to physical shortages – it is all about the interest rate.

Since it appears that for the most part, the price-to-rents ratio is further out of line with its historical average than the price-to-income ratio across the OECD, it is not unreasonable to suppose that this has been repeated across the globe. Obviously the extent to which prices diverge in different countries will depend on a variety of local factors. However, as  one respondent to the blog post noted, quoting an economist at the Australian Banking Royal Commission, ”the price of housing is a function of the demand for and supply of credit not the demand for and supply of housing.”

As a result, no amount of housebuilding is going to reduce house prices: Only central bankers can do that. However it does raise a question mark as to what happens if and when interest rates do start to rise. Reducing the discounted value of rents is likely to result in a sharp fall in house prices, and to the extent that housing forms a substantial part of household net wealth, falling house prices could squeeze consumption hard. Indeed, the severity of the 2008 recession in the US is partly attributable to the impact of falling housing wealth. At some point, some of the air will have to be let out of the bubble, even if it is still a long way off in the future. The bottom line is that some central banks will have to think long and hard about the damage that raising rates could cause and even small changes could have bigger effects than we are used to. What goes up must come down, and what goes up a long way must also come down a long way.