In 1997, residential property in Hong Kong was more expensive than almost anywhere else in the world.
While visitors were sometimes amazed at the sky-high prices paid in Hong Kong, residents were not. Hong Kong was intensely business-oriented, and had pushed itself within a generation from third world poverty, to glittering first world status. It seemed the most dynamic city in the world.
This is the place to see how the world will do business in the future, opined Lord William Rees-Mogg, former editor of The Times. Seven years later, that same apartment was worth 30% of its price at the peak.
A similar story unfolded in Tokyo eight years earlier. In 1989, the land on which the Imperial Palace stood was valued at more than the whole of New York.
Then the Bank of Japan jacked up interest rates and prices crashed.
Moral: What seems reasonable at the time, often looks very foolish later!
Bubbles characterize all asset markets. A bubble is an irrational over-investment in one particular sector or economy. But what is particularly interesting about property markets is that its cycles, its bubbles and depressions are (unlike the stock market) so smoothly predictable.
Since the start of the industrial revolution the world has experienced long waves of growth and contraction, with severe economic depressions occuring at intervals of 45 to 65 years (the Kondratieff cycle). There is also a regular shorter term cycle of around six to 12 years, which came to be called the business cycle. In the latter part of the 19th century a large body of literature built up to explain the business cycle, which came to be seen as the central problem of economics.
Yet in the 1930s there was a new perspective. Under the influence of Maynard Keynes most economists abandoned the business-cycle framework. Instead, depressions were seen as caused by policy mistakes, and economists hoped the business cycle could be smoothed out. This was a period of tremendous optimism in the economics profession.
Alas, the cycles did not disappear. The business cycle remained in place.
In the 1970s, another theory became fashionable. Based on studies of the stock market behaviour, the 'efficient market hypothesis' argued that asset prices tend to reflect the best current information about the value of an asset.
When a new piece of information about an individual company enters the market, the stock price adjusts upwards very rapidly, within a day or two. Prices do not drift up slowly over weeks or months.
The academics drew the conclusion that the stock market is a very efficient mechanism for absorbing new, objective information. It is a 'perfect market'. A company's stock price chart has a pattern which came to be called a 'random walk' (random, because it can't be forecast). The stock price jitters about like an angry mosquito, reflecting today's news. The market absorbs new information so well that it is almost impossible to take advantage of news to make profitable trades on a company's stock - that's how perfect the market is.
Therefore there are no predictable long-term cycles in the stock market. Because if cycles were predictable, the information (that they were predictable) would already have been acted on, smoothing out any cycles (because the market is perfect).
Thus, predictable cycles were believed impossible.
Property markets are different
The perfect market hypothesis achieved an extraordinary hegemony within academic finance during the 1970s and 1980s. Even practitioners who tended to be 'ho-hum' about the theory couldn't put their doubts into words.
Cracks in the theory appeared from around 1985 onwards, with the rise of the new school of 'behavioral finance'. But these cracks tended to be hairline cracks, minor faults, not major corrections. Stock markets do better in January (the January effect), which shouldn't be happening. Individual stock prices have a slight tendency to 'mean revert', which would not be present if the market was truly perfect. But none of this is very major.
One of the leading lights of the behavioral finance school, Robert Shiller, began his research in the real estate markets. In property markets the 'perfect market' theory simply doesn't work. The normal pattern of real estate price movements is a smooth curve, upward or downward-sloping (see summary 1 Pyhrr, Roulac and Born (1999)). Graphs of property prices look very different from stock market graphs - they are smooth curves, not angry trails left by buzzing mosquitoes.
Though property markets behave completely contrary to perfect market theory, "as recently as the late 1980s," note Pyhrr, Roulac and Born, "it was not uncommon to hear a finance professor dismiss the concept of real estate cycles... and suggest that research on the subject was misguided."
Perfect market theory goes bust
A pioneering series of papers published in 1986-1988 began fatally undermining the existing consensus. Case and Shiller studied the psychology of the price movements during the US 1983-87 house price boom (Case and Shiller, NBER 1988). House prices doubled in many parts of the US over four years.
"Housing prices are booming," buyers polled by Case and Shiller tended to worry: "Unless I buy now, I won't be able to afford a home later."
Buyers never cited any concrete evidence about the economy, or housing supply, etc. Rather than being based on objective economic facts, the optimism of buyers was based on mutually reinforcing conversations.
"An especially striking feature of the...answers," noted Case and Shiller, "is that not a single respondent referred to explicit quantitative evidence relevant to future supply of or demand for housing."
Economic impacts on house-prices
'Objective' factors only explain a portion of house price rises and falls. Much of the rest is cyclical, and is explained by the market's own history.
This blew a hole in the perfect markets theory. Property markets move in something like a smooth curve - which is what you get when you follow a rule such as: 'buy in period T, if the price has risen in T-1'.
Changes in macro-economic variables do (also) tend to impact house prices. The most important factors are:
- Increases in national income (GNP). In a recent BIS six-country study (US, Canada, UK, Australia, Netherlands and Ireland) a 1% increase in GNP growth was associated with a 1-4% rise in real house prices after three years. The effect was greatest in Ireland, least strong in the U.S.
- Reductions in interest rates. A 1% reduction in the short-term interest rates is associated with ½ to 1 1/2% increases in house prices within a year. The effect is strongest in Canada and the Netherlands, weakest in the U.S. and the UK.
- Equities prices increases. In the UK, variations in stock prices explain 35% of all house price growth over a three-year time horizon, while variations in GNP explain around 20%. In the US, Canada, and Ireland, a 10% equities price increase causes about a 1% house price increase over three years.
Yet above all, buyers pay attention to what other buyers are prepared to pay - not some exogenous 'objective' information. So there is no perfect market driven by external news - but rather a predictable market, feeding on itself.
Real estate returns
The moral is simple: do not always believe that the current market high is due to 'fundamental' factors.
Real estate markets tend to fluctuate between undervaluation and overvaluation. What is 'high' for a market, and what is 'low' is largely a matter of history, and of comparison with other assets.
Over-valued property markets are extremely seductive. There are often good reasons to believe that a particular market deserves higher valuations than it did in the past, or than other markets deserve. And sometimes (to make matters more complicated) these arguments are right.
To get an indication of when prices are reasonable, there are yardsticks.
We review the leading yardsticks in the article How to avoid buying into a bubble.