The housing market is in the doldrums. A survey by the RICS next week is likely to say that prices are flat in most of the country and falling in London, while the Halifax should report that although prices are recovering from their falls earlier this year they are up by only around 1 per cent after inflation in the last 12 months.
Official forecasts point to the malaise continuing. The OBR last week forecast that prices will rise only 1.1 per cent per year in real terms between now and 2022. This is because weak productivity growth means weak income growth which puts a lid on prices.
And then there are some risks. The obvious ones are that a recession would cut prices and make housing even harder to sell. But there’s also the danger that a rise in mortgage rates would reduce what little affordability there is in the market. And then there’s tax risk. If incomes grow as weakly as the OBR expects, the Treasury will have to look to other sources for revenue, which would put wealth, land and houses in the cross-hairs; buy-to-let investors have already suffered from higher taxes, and things might get worse.
Some of you might think there’s also the danger that prices will be dragged down by increased housebuilding. In fact, this is doubtful; many believe that prices are insensitive to feasible changes in building in the short run.
All this poses the question: is housing, as an investment, worth the risks? Economic theory gives us a test of this. It says the required return on any asset should be the product of four things:
- The volatility of the asset: the riskier it is, the greater should be expected returns.
- Our risk aversion. The more risk-averse we are, the higher must be returns if we are to hold the asset.
- How the asset performs in bad times. An asset that does badly when our general fortunes do badly adds to our risks, so we need higher returns to compensate us. On the other hand, assets that do well in bad times offer insurance, so we accept low returns on them in normal times.
- The likelihood of bad times. The more unstable our general finances are, the less able we are to bear risk and so the higher must be returns on risky assets.
All this is intuitive. Let’s put some numbers on it. We can measure the likelihood of bad times by the standard deviation of consumer spending growth: falls in spending are a sign of bad news about our current and future prospects. Since 1955 (when quarterly data began) the standard deviation of aggregate real annual spending growth has been 2.3 percentage points. Obviously, individuals face more risk than the macroeconomy – just as individual stocks are riskier than the All-Share index – so let’s multiply this by three.
The volatility of annual changes in real house prices since 1955 has been 8.9 percentage points. Because individual houses are riskier than the house price index, let’s also multiply this by three to give us 26.6 percentage points.
We can gauge the likelihood of house price falling in bad times by their correlation with consumer spending. For annual changes, this has been 0.69.
Risk aversion is harder to measure, partly because it varies from context to context. Let’s call it three, where one represents risk-neutrality.
Multiplying these numbers together gives us 0.069 x 0.266 x 0.69 x 3 = 0.038. Which means we need a real return on housing of 3.8 per cent per year to justify its risks. In fact, house prices have risen 2.8 per cent per year in real terms since 1955.
Of course, you can quibble with these numbers. And of course price appreciation is only part of the return on housing; there’s also rental income or just the psychological benefits of having a place of one’s own – although these are offset by the costs of mortgage finance, insurance and maintenance.
Numbers don’t have to be precise to be useful, however. This exercise tells us something. It says there is probably no house price puzzle in the way that there is an equity premium puzzle. Whereas long-term returns on shares exceeded theoretical predictions – at least until the 1990s – returns on housing probably haven’t. Or at least not by much.
Of course, house prices have risen a lot – but no more so than is necessary to compensate for their risks. Viewed as an investment, housing has offered only fair returns.
Which raises the problem. If the OBR is even roughly right, housing in the next few years won’t offer the returns necessary to compensate most people for their risks. Sadly, however, it does not follow from this that any other asset will do so.
