Castles in hot air
May 29th 2003
From The Economist print edition
Is another bubble about to burst?
THE rapid house-price inflation in many countries over the past few years is clearly unsustainable. But will house prices just flatten off or will they slump? Alan Greenspan, who criticised the stockmarket's “irrational exuberance” long before that particular bubble was fully inflated, is more sanguine about American house prices. He accepts that there are local hot spots, but dismisses the idea of a national housing bubble that could harm the whole economy if it bursts.
In other countries where house prices have been rising, many analysts also pooh-pooh the idea of a bubble. They point out that interest rates are low, so people can afford to borrow more and are therefore willing to pay more for a home. The affordability index—mortgage-interest payments on an average-priced home divided by average income—is still quite low by historical standards in
most countries.
There is something in this. First-time buyers who previously could not afford the typical mortgage payment can now get on the first rung of the property ladder. This pushes up effective demand and could possibly lift the equilibrium level of house prices. However, the cruder—and more widely used—argument that lower interest rates make homes cheaper to buy is badly flawed, because it ignores inflation. If interest rates are low only because inflation is low, then although initial mortgage payments are smaller, the real burden of mortgage debt will be eroded more slowly over time, and payments will remain high in relation to income for much longer. Lower inflation shifts the profile of payments over the life of a loan—you pay less in the early years and more (in real terms) later on—but the total real cost is the same. Home-buyers who think that low interest rates make buying a house cheaper are suffering money illusion. It is as foolish as thinking that a car loan paid off over five years is cheaper than one paid off over only two years.
Lower real interest rates, after allowing for inflation, would indeed reduce the true cost of buying a house. But in many countries real rates are not particularly low, and tax relief on mortgage payments also becomes less generous
as nominal interest rates fall. Sabina Kalyan, an economist at Capital Economics, a London-based consultancy, calculates that real after-tax mortgage rates in Britain (where tax relief has been scrapped) are now much higher than they were in the 1960s and 1970s, and only marginally lower than in the 1980s and 1990s. And even in countries that still enjoy tax breaks, such as America, real after-tax mortgage interest rates have probably increased over time, because the size of the tax benefit shrinks with falling inflation. For instance, if interest rates are 10% and inflation 8%, and your marginal income-tax rate is 40%, then the real interest rate you pay is minus 2%. With 4%
interest rates and 2% inflation (ie, the same pre-tax real interest rate), the real after-tax rate is plus 0.4%.
A second popular reason cited for a continued rise in house prices, especially in big cities such as London or New York, is the combination of a growing number of households and a limited supply of homes. Supply, it is argued, cannot keep up with demand because of a physical shortage of land, so prices will keep rising. This certainly helps to explain why house prices in big cities have tended to outpace those in rural areas, but it is not a reason to expect prices to keep rising indefinitely. Supply constraints are already factored into current prices; they will not necessarily stop prices declining in future. Just look at Hong Kong, where the supply of land is more strictly limited than in most cities. That has not prevented house prices falling by 65% over the past five years.
What's it worth?
Neither low interest rates nor population growth can justify recent house-price booms. Moreover, the thinking behind both arguments reflects a lack of understanding about how asset prices are determined. On any given day, a house is worth whatever somebody is prepared to pay. However, the lesson from the stockmarket bubble is that an asset's fundamental value must never be neglected.
There are two ways to gauge whether house prices are sustainable: the p/e ratio and the house-price-to-income ratio.
• The p/e ratio. The price of any asset should reflect its future income stream.
Just as the price of a share should equal the discounted present value of future dividends or profits, so the price of a house should reflect the future benefits of ownership—either the rental income earned by a landlord or the implicit rent saved by an owner-occupier. During the dotcom bubble, investors behaved as if profits no longer mattered. Likewise, people today are ignoring the link between house prices and rents.
In America, for example, a p/e ratio of sorts for housing can be calculated by dividing an index of average house prices by the index of rents which is included in the consumer-price index. In recent years, home prices in the United States have outpaced rents, pushing the p/e ratio to record levels, 16% above its 30-year average (see chart 8). San Francisco's p/e is almost 30% above trend. Estimates of p/e ratios for Britain, Australia and the Netherlands point to an even more pronounced bubble, suggesting that house prices in all three countries are at least 30% too high.
All this means that the return from investing in property has slumped. These p/e ratios are based on average rents, but over the past year or so rents on new lettings have fallen by 20-30% in New York, San Francisco and London. In London and Sydney gross rental yields (rents divided by price) have fallen to around 4%, from 7-8% a few years ago. In these cities it is now cheaper to rent than to
buy.
The p/e ratio helps to expose the fallacy that house prices are rising because of growing populations and fixed supply, because these factors should affect the rental as well as the owner-occupier markets. The fact that prices are rising much faster than rents suggests that homes are being bought in the expectation of capital appreciation rather than underlying fundamentals. That is the definition of a bubble.
Some analysts argue that lower interest rates justify a higher p/e ratio. The same argument was made in the late 1990s for equities. But if rates are low because inflation is low, then future rents will also rise more slowly. Future rents should therefore be discounted using real interest rates, not nominal rates. As argued above, real interest rates in America or Britain are not particularly low.
The p/e ratio is probably the best measure by which to judge whether houses are overvalued, because it tries to value housing like other assets. But it is hard to get reliable, long-run series of rents for all countries. Moreover, in many continental European countries rents are still partially regulated by the government. If the p/e ratio proves difficult to obtain, another useful measure is available:
• The house-price-to-income ratio. The ratio of average house prices to average incomes, which tracks the long-term affordability of homes, is currently flashing red in America, Britain, Australia, Ireland, the Netherlands and Spain.
In all these countries the ratio is close to or above previous peaks—ie, levels that preceded previous crashes (see chart 9).
Awkwardly, this ratio is sensitive to the exact measure of income used. In America the ratio of average home prices to average personal disposable income is only 5% above its long-run average. But Ian Morris, an economist at HSBC, a bank, reckons that the ratio of average house prices to median household income is a better measure, because it reflects the circumstances of the typical home-buyer and avoids the distortions introduced by a few very rich people. This ratio is currently at a record high, 14% above its long-run average.
Mr Greenspan insists there is no “national” housing bubble in America, and indeed houses across the country are on average less overvalued than elsewhere. But if a bubble were to burst in several cities simultaneously, that would have big national consequences. Using the ratio of house prices to income, a recent analysis by Michael Youngblood, director of research at GMAC-RFC Securities, identified potential bubbles in 55 of the 210 metropolitan areas that he examined.
For the euro area as a whole, the average house-price-to-income ratio is slightly below its long-term average, but this is mainly because in Germany houses on this measure look significantly undervalued. In France, Italy and
Belgium the ratio is close to its long-term average. In Spain, the Netherlands and Ireland, however, it is 40-50% above it.
Ireland and Spain differ from most of the other bubbly housing markets in that real mortgage-interest rates fell substantially in the run-up to euro membership. This may justify a higher ratio of house prices to incomes. But how
much higher?
All of the six countries where houses appear to be overvalued (America, Britain, Australia, Ireland, the Netherlands and Spain) also share another bubble-like
symptom: an explosion in mortgage borrowing in recent years. Australian household debt has jumped from 55% to 130% of personal disposable income over the past decade. In the Netherlands the ratio has almost doubled to 180% of disposable income over the same period. The average new mortgage there is 110% of the value of a home, because lenders are happy to finance all the purchasing costs, including stamp duty and fees. In America, 21% of mortgages last year were for more than 90% of a home's purchase price, up from 7% at the peak of the boom in the late 1980s. This means that if prices were to drop, more households would be left with debts exceeding the value of their home than were a decade ago.
A sticky situation
Comfortingly, many economists believe that house prices will level out rather than collapse. In America, national average house prices have never fallen for a full year during the past three decades. Unlike share prices, house prices tend to be sticky downwards. The usual explanation for this is that people have to live somewhere and transaction costs are high, so they will not sell even if they think their home may be worth less next year. Owners are also loth to accept a capital loss. As long as they can meet their mortgage payments, they will stay put until conditions improve. So the volume of sales will decline, but prices will not.
However, countries other than the United States have seen falls in average house prices over past decades. And even within America there have been regional busts, for example in Boston, Manhattan and San Francisco in the early 1990s, where prices dropped by 15-20%.
The optimists still maintain that house prices do not have to fall to restore the relationship between house prices and rents or incomes. All it needs, they argue, is for house prices to stay constant as incomes and rents rise. This is how most of the adjustment after previous house-price booms has been achieved. But now inflation is so low that gradual adjustment would take too long, so nominal prices are much more likely to drop.
Take Britain, where house prices appear to be more than 30% overvalued relative to their long-term trend. If house prices stay constant and wages grow by 3% a year, it will take ten years to bring the ratio of prices to income back to its fair value. Moreover, history shows that after a boom prices usually undershoot their fair value. If the ratio of house prices to income were to plunge to its
trough of the last cycle, then even with rising incomes average nominal house prices would slump by almost 40% over four years. That would be overdoing it, but a price fall of 20-25% looks likely.
Table 10 takes a guess at how far prices might fall in the six countries with suspected housing bubbles. The first column estimates the amount by which houses are overvalued (relative to the average of the past three decades), based on both the price-to-income ratio and the price-to-rent ratio where available. Our forecast allows for a 10% rise in the equilibrium level of house prices in Ireland and Spain, as a result of lower real interest rates. It also assumes that over the next four years wages (and rents) will grow by 3% a year in America, Britain, Australia and the Netherlands and by 4% in Ireland and Spain, helping to restore the link between house prices and incomes; and that prices will undershoot their fair value by 5-10%. The end result of these calculations is that average house prices in America could fall by 10% in money terms over
the next four years and those in the other countries listed by 20-30%. If America and Europe were to suffer a spell of deflation, as some economists fear, prices would plunge even more steeply.
These are averages across the whole country, but some regions are more overvalued than others. A 20-25% average price decline in Britain could imply a fall of over one-third in London house prices but a much smaller drop in the north. In America prices in New York, San Francisco and Boston might drop by 20% or more.
A bigger drop in nominal prices is also that much more likely this time because so many more people have been buying houses as an investment. In Australia, investors accounted for over 40% of all housing loans approved last year, up from 18% in the early 1990s. In London, one in eight housing transactions in recent years was for investment. If house prices start falling, owner-occupiers
who can afford to meet their interest payments will stay put, but investors are more likely to sell—especially if rents do not cover their interest payments—putting further downward pressure on prices. The abundance of cheap
rented property is also affecting the market in another way. Anecdotal evidence in London and Sydney suggests that some owner-occupiers are selling and renting to lock in a profit before prices fall.
The optimists still maintain that previous house-price booms turned to bust only because central banks sharply raised interest rates, so buyers could no longer afford to service their mortgages. This time, with inflation so low, interest rates are unlikely to rise significantly, they say. Perhaps; but Japan and Germany show that low interest rates are no protection against a decline in house prices.
All that is required to burst the bubble is a change in sentiment. Just as investors in the stockmarket eventually decided that enough was enough, at some stage would-be home-buyers will refuse to pay ridiculous prices. In many
countries, first-time buyers are now finding it impossible to get on the bottom rung of the property ladder because they cannot scrape together the deposit. In Britain, mortgage lending to first-time buyers has fallen sharply this year. As demand at the lower end dries up, the whole market could quickly turn. And if prices start to dip, one good reason to buy—the expectation of capital gains—will disappear, pushing prices even lower.
Mr Greenspan says that any analogy between house prices and the stockmarket bubble is “a rather large stretch”. He needs to exercise his imagination. House prices in America and other economies are unlikely to fall as precipitously as share prices have done over the past three years, but even a much smaller drop could do more serious economic harm.
Copyright © 2003 The Economist Newspaper and The Economist Group. All rights reserved. |
