Gavin R. Putland vs. Fred Harrison on when (not if) the GFC will be

BOOM BUST: House Prices, Banking and the Depression of 2010, by Fred Harrison — reviewed by Gavin R. Putland.¹

Decades ago — perhaps in the early 1970s — a dirty snowball about five kilometres in diameter was making a leisurely loop through the inner Solar System. Apparently it was following an elliptical orbit about the sun, in which case the snowball had made the same journey many times before; that is, apparently it had been moving in a cycle. In any case, on this particular trip the overgrown snowball had a close encounter with the planet Jupiter and was captured by Jupiter's gravitational field; that is, it began to orbit Jupiter. The new orbit was very elongated, so that interference from the sun's gravity caused significant variations: the motion of the comet about Jupiter was recognizable as a cycle, but this cycle was only approximately periodic. In July 1992, the snowball passed so close to Jupiter that it was torn apart by the tidal effect (the variation in the gravitational force with distance), forming more than 20 major fragments and innumerable smaller ones. In March 1993, the fragmented snowball was discovered by Eugene and Carolyn Shoemaker and David Levy and became known as Comet Shoemaker-Levy 9 (or SL9). By studying the comet's motion and extrapolating it into the past and future, astronomers pieced together the preceding story. They also determined that the comet's next encounter with Jupiter would be its last: between 16th and 22nd July 1994, the fragments slammed into Jupiter's dense atmosphere, leaving “scar” clouds about the size of the earth.

What's that got to do with a book about economic cycles? Three things:

1. Cycles are not laws of nature, but are manifestations of underlying laws. If Comet SL9 orbited the sun many times, that cycle was broken by the initial encounter with Jupiter. The “cycle” of the comet's subsequent motion about Jupiter could not be extrapolated indefinitely into the past and future; it had a beginning and an end. But the beginning and the end did not represent suspensions of the laws of physics; rather, the comet's motion was governed at all times by Newton's laws of gravity and motion. As the cyclic motion of the comet, while it lasted, was a manifestation of underlying physical laws, so a cyclic variation of a key economic indicator, while it lasts, is a manifestation of underlying economic laws. And as the same underlying physical laws eventually interrupted the cyclic motion of the comet, so the same underlying economic laws may, for all we know, interrupt the cyclic variation of an economic indicator.
2. Cycles can interfere with each other through the underlying laws. When only two bodies are involved, Newton's laws of gravity and motion give rise to perfectly periodic, synchronized orbits — that is, a perfectly regular cycle. But adding a third body can cause all sorts of complications, ranging from slightly disturbed cycles (e.g. as Jupiter's gravity slightly affects the earth's orbit about the sun) to fairly stable sub-cycles (as Jupiter's moons orbit Jupiter which in turn orbits the sun) to unstable sub-cycles (as the sun's gravity caused Comet SL9 to drift closer and closer to Jupiter) to captures (as SL9 switched from a sun-centric path to a Jupiter-centric path) to collisions (as SL9 hit Jupiter) to permanent ejections (as the Pioneer 10 spacecraft was kicked right out of the solar system by its close approach to Jupiter). In each of these examples, the cause of the disturbance is itself a cycle: the motion of Jupiter about the sun. As the orbits of celestial bodies interfere with each other through the very laws that govern those orbits, so economic cycles may, for all we know, interfere with each other through the very laws that govern those cycles.
3. Consequently, a recognizable cycle may be only approximately periodic. Had Comet SL9 been discovered earlier and observed through several orbits, any prediction of its motion based on an assumption of precise periodicity would have been noticeably wrong — spectacularly wrong after July 1994. Likewise, any prediction of an economic indicator based on an assumption of precise periodicity may also be noticeably wrong, and may, for all we know, be about to go spectacularly wrong.

Fred Harrison in his latest book, Boom Bust, indeed makes a prediction on the assumption of precise periodicity. And he is so confident of this periodicity that he writes his prediction into the book's subtitle, House Prices, Banking and the Depression of 2010. How confident should he be?

Harrison's thesis is that the land market follows an 18-year cycle, with a short recession at the mid-point of each cycle and a longer recession at the end-point (as summarized in his diagram on p.87). To support this claim, he starts with the slump of 1992, which he treats as a “primary” or “end-cycle” recession, and counts backwards to establish hypothetical dates of all the mid-cycle and end-cycle recessions since 1776 (p.101). Then he cites historical evidence in support of each date.

The scientifically literate reader will immediately notice two possible sources of confirmation bias in this procedure. First, attention is drawn to particular years and away from other years during which equally interesting things might have happened. Second, attention is drawn to the ways in which the events of the proposed years are consistent with recessions, and away from other possible interpretations of the same events. Even then, Harrison admits that the end-cycle recessions of 1920, 1938 and 1956 didn't happen, and cites the two world wars as the reason. At the end of this historical survey he remarks (p.115):

We do not claim that the trends that may be traced in the historical record worked with the precision that would impress a Swiss clockmaker. But the deviation by six or even 12 months on either side of the end of an 18-year period, or its mid-way point, does not discredit our theory.

Indeed it doesn't; it makes the data look too perfect, raising the suspicion that there has been some inadvertent methodological bias, in which case the theory is neither discredited nor confirmed.

But on the same page, Harrison then quotes historian Llewellyn Woodward as referring to commercial crises in 1825, 1836–9, 1847, 1857, and 1866. If we try to fit a 9/18-year cycle to those dates, the best we can do is 1829, 1838, 1847, 1856, and 1865, which means that Woodward's last crisis occurred a year late and the first one four years early (or 5 years according to Harrison's hypothetical dates). If Harrison's “depression of 2010” comes four years early, it will happen in 2006 (which is roughly what this reviewer has predicted). In Harrison's defence, one might point out that the crisis of 1825 was apparently related to shares rather than land. But if a recession can be caused by something other than the land market, a regular land cycle does not guarantee a regular cycle in the wider economy. Moreover, even the land market can behave in “unscheduled” ways. The present global property bubble — described by The Economist as the biggest asset bubble in history — inflated right after what should have been Harrison's mid-cycle recession; but according to Harrison, such huge bubbles are not supposed to appear until the “winner's curse” phase at the end of the cycle.

Let's have a closer look at that scheduled mid-cycle recession. According to Harrison, the U.K. was due to go into recession in 2001 (pp.1,13). It didn't, he says, because when Gordon Brown became Chancellor of the Exchequer in 1997, he directed the Bank of England to conduct monetary policy so as to maintain inflation at 2.5% per annum, based on a price index that excluded mortgage interest (p.8). In other words, inflation in the residential land market was deemed not to count. So buyers were allowed to bid up prices far beyond the levels that, under previous policies, would have provoked remedial action. The housing bubble, instead of popping and precipitating a recession, merely expanded at a reduced rate into 2001, allowing home owners to borrow against their rising land values and spend the country out of the trough. So the British economy as a whole did not suffer a technical recession (two consecutive quarters of negative growth), although its manufacturing sector did (p.196).

That's all eminently plausible. But, having explained the missing recession in terms of a radical change in policy, why does Harrison not entertain the idea that the same change in policy could produce a change in the period between recessions — or at least a change in the length of the current cycle? Why does he assume that the mid-cycle recession has been averted and not merely delayed, especially when the housing bubble that should have caused the recession has been allowed to keep on growing? Shouldn't he rather say that the mid-cycle recession is overdue? And if it happens late, might it not have some characteristics of an end-cycle recession, so that one could just as well say that the end-cycle recession has come early?

Let's see if we can make sense of a combined “late mid-cycle recession” and “early end-cycle recession” in terms of the underlying dynamics.

The basic cause of boom-bust cycles is clear enough. As land is in fixed supply, land prices increase with economic growth. That creates a speculative demand for land, which accelerates the price rise, and so on, until “the bubble bursts”. The peak in land prices is accompanied by a peak in building activity as investors try to justify the exorbitant prices paid for sites. This activity, plus consumption financed by borrowing against rising equity in land, plus the knock-on effects, induce a general economic boom.

In general, the bursting of a bubble in a particular asset market has two counteracting effects. On the one hand, it drives investors away from that asset class and, by default, towards some other asset class that may also be susceptible to bubbles. On the other hand, those who have invested heavily in the collapsed market have to reduce their expenditure, and some become insolvent. As one agent's expenditure is another's income, and as one agent's debt is another's asset, a chain reaction ensues, reducing the funds available for investment in other asset markets, possibly causing them to collapse, and so on; these are the ingredients of a recession. In the late 1980s, the stock-market burst led into a land bubble, which then popped to cause a recession. In the mid-late 1920s it was the other way around. But in all cases, a bursting bubble in one asset market interferes with other asset markets. These are the ways in which economic cycles, like celestial orbits, “interfere with each other through the underlying laws”.

Now let's focus on land. One possible explanation for the mid-cycle and end-cycle effects, which Harrison doesn't seem to consider, is that there are actually two land cycles: a commercial land cycle of roughly 18 years, superimposed on a residential land cycle of roughly 9 years. (According to the work of Bryan Kavanagh at the Land Values Research Group, this model would be consistent with experience since 1970.) In that case, a “mid-cycle” recession is triggered by a residential burst alone, while an “end-cycle” recession is triggered by a combined residential-commercial burst.² A normal residential crash squeezes home owners and small investors, causing a fall in consumption and hence a minor recession, but drives bigger investors towards other assets, including commercial land. But what if a new and irresponsible monetary or fiscal policy allows a residential bubble to grow much longer and bigger than usual? When it bursts, might not the ensuing recession be severe enough to bring down the commercial land market as well? Would that not be a combined “late mid-cycle recession” and “early end-cycle recession”? In failing to consider this possibility, Harrison seems to pay too much attention to schedules and too little attention to the underlying dynamics — in particular, the unusual size and timing of the latest residential land bubble.

Predictions aside, there is much to recommend in Harrison's 266 pages of analysis. With appropriate irony he exposes the injustice of privatized economic rent and explains the virtues of taxing it (such as the encouragement of development and the suppression of cycles). He gives hints as to what factors, other than speculation, might have influenced the periodicity of land markets at various times in history; these include the average adult lifespan, the planned lifetimes of terminating societies (ancestors of building societies), anti-usury laws (affecting the time taken to repay loans), and the frequency with which people change addresses. Policymakers' lack of interest in the land market is a recurring theme. A priceless quote from Alan Greenspan (p.65) debunks the excuse that recessions are caused by oil shocks. The learned pronouncements of economic commentators about the “new economy” and the “new prudence”, with the assurances that “this time it's different” and that a few “bad apples” don't spoil the whole barrel, are duly quoted while Harrison, like a latter-day Ecclesiastes, reminds us that we've heard it all before.

Of course Harrison is right in his basic assertion that the land market is cyclic, and that bubbles and bursts in the land market are the unrecognized pointers to what are called economic booms and busts. But by claiming that the boom-bust cycle has been almost perfectly regular for more than two centuries, he has, in this reviewers' opinion, overplayed his hand and cast doubt on his methodology. By further assuming that the claimed regularity will continue through the present cycle, and announcing the next depression on that basis, he has gambled much of the credibility that he has laboriously earned over more than 20 years.

That said, let us give him credit for taking a stand. Is he right? We'll find out when the hard times begin.

Fred Harrison's BOOM BUST: House Prices, Banking and the Depression of 2010 is published by Shepheard-Walwyn (London, 2005).

Notes (added Nov.24, 2009)