I recall one particularly rousing (read: fiery) debate I had on an internet forum with a gentleman named ‘sense’. His position was simple: that buying a house was always better financially than renting over the long term.

Each of us produced spreadsheets to support our respective positions. Each was able to show that for a given set of assumptions, our case was sound. The equation boils down to the assumptions and how well they reflect the reality of an unknowable future.

I don’t have those assumptions handy, but I believe they were approximately as follows:

• 8% interest rates
• 4% rental yield growing by 3% pa
• 3% inflation rate
• 4.2% net yield on invested savings.
• 10% house deposit
• 2% annual house ownership costs (rates/insurance/maintenance)

All these can be estimated with some degree of confidence. There remained just one further factor, the important one that swung the outcome either to the renter’s or the buyer’s favour: Capital appreciation.

The tipping point turned out to be (perhaps surprisingly) low. At somewhere between 4% and 5% annual appreciation, buying would surpass renting after some 20 years. At a higher rate of growth, the buyer’s financial position would exceed the renter’s earlier. At a rate slightly below 4%, the renter would be able to buy a house for cash before the buyer had finished repaying his mortgage!

The debate should have been simple to close conclusively at this point. After all, prices have grown at an average (though not continuously) of 8% per year over the last 30 years. So it is not unreasonable to expect them to grow at 5% or faster for the next 30, right?

Somehow, I don’t think it’s so simple. Incidentally, I seem to recall the debate ended at this point with an agreement that the final result depended on the assumption for capital gains. Sense’s view was that 5% growth was not far fetched (even a given?), so buying was the optimum choice. I’m still not so sure.

Efficient Markets Theory

It is a widely known fact that the stock pickers for managed funds are of little value. Over any a ten year period, somewhere around 85% (if memory serves) of actively managed funds will achieve for fund-holders a lower return than the index in which they invest. ‘Index Funds’ exploit this by emulating the gains (and losses) of the index.

How is it that a fund that simply buys a few of every stock available can most of the time outperform most of the funds that follow the expert advice of a team of highly intelligent, highly educated analysts who spend countless hours and days poring over countless numbers of company reports and accounts? Proponents of efficient market theory would unsurprisingly say that it’s because the market is efficient.

Basically, efficient market theory states (in an efficient market) that share prices always reflect all known information about the companies owned by those shares. Stock exchanges do their best to ensure this is true by mandating that listed companies provide all relevant information to all market players.

A stock analyst might determine that a company will return a bumper profit next year. Clearly, its stock are a ‘buy’. But the analyst will not be alone. Others will use the same information to reach the same conclusion and prices will be bid up. Meanwhile, the sellers will realise that their stock is underpriced and refuse to sell at such a low price. Very soon the price will hit a level where all the information has been ‘priced in’. The stock is no longer a bargain. It is fairly priced relative to alternative investments.

There is some debate about efficient markets, but this seems to be more about how efficient they are and how prone they are to behavioural anomalies of market players (see below), rather than the mechanics described above. So what has this to do with the housing market? Plenty.

Imagine a house costs $100k. If this house will be worth $200k in ten years and the holding costs to a buyer will be $50k for ten years it is a bargain. We should all buy it if we can. Yet if it goes to auction or even a competitive private sale where just two people know that: a) Their holding costs will be $50k and b) It will be worth $200k, it will not sell for $100k and will not be a bargain. Its price will be bid up until one player decides he can get a better and less risky return elsewhere. The bidding will likely stop somewhere shy of $150k.

To expect otherwise, you’d have to believe that markets are not very efficient at all. That the only rational person was the buyer. That the seller, the real estate agent and the competing buyer (in reality, numerous competing buyers) were all clueless. That the only person who knew this house would double in price in ten years and was a bargain at $100k was the single buyer!

How likely is this? I mean, ‘everybody’ ‘knows’ that ‘house prices double every 7 to 10 years’, right?

Behavioural Economics

Some economists study something called behavioural economics. Among other technical stuff they notice that people tend to act in herds, buying (something) when others are buying it, selling when others sell. That people are moody, either unduly pessimistic or unduly optimistic. That market participants overestimate their own intuition/ability/understanding/knowledge. That people expect short term trends to be sustained for long periods even when this is highly unlikely or even impossible.

By acting on these behavioural instincts, market participants do themselves great harm. They buy overpriced stocks and sell underpriced ones. They sell great stocks which dip in price and buy crappy ones that are ‘going up’. They repeat past lucky moves that provided great results only to be repeatedly disappointed. They mistake rapid irrational upward moves for a ‘sure thing’ or a ‘trend’ and buy more and more at unsustainably high prices, making small gains on the way up that are far exceeded by losses on the way down as the market recovers its senses.

Of course there’s a problem here. If markets were totally efficient it would be impossible for such a person to do themselves much harm beyond incurring unnecessary trading costs. The shares they bought and sold would rarely be over or under priced. But a market where people behaved in such a way wouldn’t be very efficient! So which is it?

Probably a bit of both. Individual shares, entire markets or even whole economies never operate at their perfectly efficient optimum. They are either too high or too low, too hot or too cold. They fluctuate around the levels proscribed by efficient market theory, veering away from fair value sometimes for a brief period, sometimes for many years, but always return to trend. This is partly because information is imperfect or unevenly distributed and partly because people are irrational.

Back to houses. Supposing somebody did buy that house for $100k ten years back and now it is worth $200k. What does that tell us? Does it mean that house prices double every ten years? Heck no! At $200k that house is either fairly priced, overpriced or underpriced (the balance of probabilities is not with the former).

Suppose it is fairly priced today. That means that a decade ago it was underpriced at $100k and should have sold for almost $150k in a competitive market. In that case it should be implied that house prices rise 33% on a fair value basis over ten years. NOT 100%! If it is overpriced today, it should be expected to gain less than 100% over the coming decade. Could it have been underpriced at $100k and still be underpriced at $200k today?

Well… yes. But you’d have to be mad to bet on it, unless you had good reason. One such reason would be if its rental yield had risen so that holding costs had fallen relative to the expected gain. I’ll ignore this because it most certainly does not apply to the recent housing market!!!

So how can we apply this information to the recent housing market? Let’s look at rational behaviour. Consider the expectation that house prices will double every seven to ten years. This was a commonly held belief in Perth in 2005 (as it still is across much of the country). Fair enough, house prices had doubled over the last ten years and stood at $300k. Over the following year, house prices rose 50% to $450k. And guess what? Many (most?) people still expect house prices to double every seven to ten years! Irrational.

That means people expect house prices to hit $900k whereas a year earlier they were expected to reach just $600k. If the rule is true, and people are rational, they should conclude either that prices were 33% underpriced in 2005 or they are 50% overpriced today.

By and large, they conclude neither. They instead concoct a convenient, self-serving and frankly deluded explanation. House prices move in predictable cycles. Several years of slow growth followed by two or three of explosive growth. Over the long term this averages out to 7% to 10% per year. This provides comfort to those who bought in 2006 only to have prices flatten or fall. I say phooey to this.

For starters, efficient market theory tells us that rational players would take advantage of such an observable cycle. They would buy after several years of sub 7% growth and sell after a couple of years of 10% plus growth. Doing so would far exceed the returns offered by buying and holding. But if people did this, the ‘cycle’ would soon disappear. Prices would rise at a steady rate somewhere between 7% and 10% per year. But they don’t.

What if the fair value of a house does rise by 7% to 10% per year and ‘the cycle’ is simply a reflection of behavioural economics theory? That implies that most people are not rational and there is scope for a rational person to take great advantage of them. How? By buying a house at the cyclical low (+/- a few years), in other words, NOT NOW!

But what evidence do we have of this ‘cycle’? It’s true that in most parts of the county house prices have grown at a compound annual rate between 7% and 10% over the last 37 years. It is also true that they’ve sometimes been above and sometimes below this range. This does not prove the existence of a cycle.

The first half of this period was dominated by high rates of inflation, which spent as much time above ten percent as below. Naturally, house prices grew rapidly in nominal terms but really didn’t go far in real (inflation adjusted) terms or anywhere compared to incomes. The first real peak came after the boom of the late 1980s. This was followed by a lull through most of the 1990s. Prices during this time (by most accounts) remained elevated in real terms relative to the 70s, 60s and beyond. Prices have been growing very rapidly since the late 1990s, though 2004 and 2005 were muted relative to the earlier and more recent boom years.

So where’s this ‘cycle’? We’ve had a peak, a trough, then one more peak. This is hardly evidence of a recurring cycle! Take a piece of paper. Draw three dots in random locations. Okay, you’ve drawn a triangle. What? You haven’t? If you draw lines connecting these dots we’ll have a triangle. Clearly your dots are in a triangular arrangement.

This is silly, right? Not as silly as looking at three dots and seeing a cyclical pattern, projecting it as self-replicating into the unknowable future and borrowing more money than you could ever hope to repay to gamble on this imaginary line’s existence!

Back to the original question of renting or buying. Remember how it only takes 5% annual capital gains for buying to beat renting (all else being equal, predictable, and static)? Surely that’s much more of a sure bet than prices doubling every 7 to 10 years?

Yes it is. In fact, a study by some clever university types in 2004 (Abelson and Chung) found that real house prices had increased just 100% over the 1970 to 2003 period, or an average of 2.1% per annum. So with the RBA mandated to keep inflation below 3%, 5% for houses sounds like a reasonable figure.

There are a couple of problems here though. The study authors noted the recent rise in house prices which was as much as 18% above annual inflation. They suggested that these rises might turn out to be temporary, in which case real returns would be lower. The precise warning was that the 100% figure would “give an exaggerated view of real price increases if, as we expect, there is a real house price downturn post 2003”. In fact, if we wind the bubble back to 1998 house prices had struggled to keep up with inflation for much of the preceding 28 years.

This poses a problem. We know that house prices today are either undervalued, fair value or overvalued. If we assume that inflation plus 2% is fair value growth over the long term (knowing this is a generous assumption) then when inflation is 3%, is the present fair value:

• Whatever price is offered today
• 5% per annum above the 1995 ‘trough’ price?
• 5% per annum above the 1990 ‘peak’ price?

Either of the latter options implies (due to recent growth well in excess of 5%) that prices are currently above fair value and future returns will be well below 5%. Renting would be better. If we take the former, we are saying that house prices have been grossly undervalued (that is, that every seller and almost every buyer has behaved irrationally) for every single year for the last 150 years or more.

Expectations are funny things. Manias* are just weird.

* Psst: You're standing in it.