Bubbles and balloons

Quite a few people criticised the title of my last post — There was no British housing bubble — on the basis that, even if there was no overconstruction of housing (and thus no Austrian-style distortion in the structure of production), there was a bubble in the sense that prices rose rapidly, and so on.

But is this right? I suppose it depends on what you mean by a 'bubble'. As far as I can tell, there are at least three different meanings of the word 'bubble':

1. A speculative bubble, like the Beanie Baby craze. As Arnold Kling put it recently, "If investors who are buying the asset have estimates of the discounted present value of the income from that asset that imply a negative real return, then it is a bubble."
2. An Austrian-style bubble that distorts the real economy by incentivising production in an area where much of the demand is illusory (typically created by credit expansion, according to the Austrians).
3. A government-created rise in price above 'real' (or endogenous) factors.

Take the third kind of bubble, which I think is what we are currently seeing in the British housing market. A ban on the construction of new houses would cause the price of housing to rise significantly, for instance (and this isn't a million miles away from current government policy). Though the government policy is probably very harmful, given that it exists it is perfectly rational for markets to drive the price up, and that price should stay up for as long as the political factors dictate. The policy might be crazy, but the market's reaction isn't.

Let's take a look at historical UK house prices (in real terms).

Clearly, prices were above trend in the 2000s and then fell after 2008, but compared to the early 1990s prices are still extremely high. I'm willing to believe that quite a bit of that rise was a type-1 or type-2 bubble, but unless you think we're still in the midst of that kind of bubble (which could pop at any time), it's not the whole story and doesn't even seem to be most of the story. (As some commenters have pointed out, some aspects of this price increase were likely attributable to foolish financial wizardry, probably driven by regulation.)

More likely, that rise in house prices since the 1990s, since it is still high, is a type-3 bubble — a sensible reaction by markets to foolish government policies constraining the construction of new homes. I can't explain why this rise only took place in the 1990s (population growth and decreases in household sizes may explain this, but I don't know), but unless you're saying that right now markets are wrong and you know better, that rise doesn't seem like the sort of unsustainable bubble that leads to sudden crashes.

Type-3 bubbles are different to type-1 and -2 bubbles in that they do not run the risk of sudden crashes. A type-3 bubble is created by government fiat and it can only be undone by government fiat. This difference is sufficiently great that I suggest a new term for type-3 bubbles: "balloons". A term like that might communicate the fact that prices have been blown up by human agency and, unlike bubbles, require an active popping or disinflating before they go away.

There was no British housing bubble

Marcus Nunes graphed the Housing Stock to Population ratio in the US recently, showing that housing reached something like a steady-state in the US from the mid-1980s onwards. As Philip Stephens says, “The constructing in US housing was exactly what was needed to maintain the housing-population ratio in the face of increased population growth. You cannot have an “unsustainable boom” without oversupply.”

This is what the UK looked like over this period (my thanks to Daniel Knowles for the data):

The long-term trend (black dotted line) is attributable to the tendency in recent decades to smaller households, but what’s interesting is that, not only was there no spike in the run-up to 2008, the growth of dwellings over population actually fell below the trend. This is not what we would expect to see if there had been a bubble in housing production.

Dwellings data is a little bit unreliable, though – splitting a house into two flats creates an extra dwelling – so it might be better to look at the amount of new houses that were actually built. Here’s a chart showing the number of residential construction permits granted over the past forty years:

And here’s the ratio of new residential construction permits over population across the same period:

These charts show that housing construction was actually well below historical levels in the 1990s and 2000s, both in absolute terms and relative to population. It is difficult to see how someone could claim that the 2008 bust was caused by too many resources flowing toward housing and subsequently needing time to reallocate if there was no bubble in housing to begin with.

What this suggests is that the Austrian story about the crisis may be wrong in the UK (and, if Nunes’s graphs are right, the US as well). The Hayek-Mises story of boom and bust is not just about rises in the price of housing: it is about malinvestments, or distortions to the structure of production, that come about when relative prices are distorted by credit expansion.

What did cause the crisis? Jeffrey Friedman has shown that bank regulation (most notably, the Basel accords) was one of the major factors that led to the financial crisis, and Robert Hetzel has outlined a convincing theory that central bankers’ tightening of monetary policy in early-to-mid 2008 was the overriding cause of the world’s economic collapse. There is also the possibility that financial investment in the housing market was a simple error.

I was once convinced that the Mises-Hayek story about the boom and bust was true, but the evidence does not seem to bear this out.

Update: A lot of people seem to be implying that Austrian Business Cycle Theory (ABCT) means: Easy money -> high prices ("bubble") -> bubble burst, people lose money. This is incorrect. ABCT relies on distortions to the structure of production (that is, the "real" economy) which have to be liquidated over a period of time following the point at which it becomes clear that they are not good investments. If a 'bubble' just meant that people had lost money it would not cause a long-running recession, it would just mean that overnight a lot of people had lost money (like a stock market crash). The reason the recession takes time according to the ABCT is that resources have been invested in a sector where price signals take a considerable amount of time to adjust after a credit-induced malinvestment bubble and so it takes a while for people to determine which investments are 'mal'.

In short: There may have been a price bubble in British housing market, but there was not the production bubble that ABCT predicts.

PS: I am interested in seeing these data for countries like Ireland and Spain, where the Austrian story may be more valid. It is also possible, as Anton Howes has pointed out, that a regional breakdown would show that there were bubble-like expansions in housing supply in certain parts of the UK, which the country-wide figures hide. If you have these data please let me know, either in the comments or by emailing me at sam@old.adamsmith.org.