PsyFi Search

Saturday 8 May 2010

Why Markets Crash

An Unsteady Aim

Oddly there’s little agreement amongst the experts about why markets crash. Although given that experts in fields without objective measures of success are generally less accurate than a drunk in a ship’s urinal during a storm that’s not really surprising. Still, if the best that the world of investment has to offer doesn’t know when stuff’s overvalued then how can we possibly hope for an end to boom and bust?

There’s no getting away from the reality that the inability of analysts to know whether markets are overvalued or not leads to serious problems. Pundits, who have a record of prediction that makes amateur astrologers look like geniuses, are delighted to proclaim the inadequacies of regulators and analysts but, frankly, have nothing better to offer. Sadly, history doesn’t offer much in the way of solace: it’s only hindsight that gives us superior knowledge.

Irving Fisher’s Great Error

The market bust exemplar that’s scored into folk memory is the crash of ’29. History, written in retrospect, records how the over-valued nature of the market made the collapse inevitable. Indeed Irving Fisher, who stated that “stock prices have reached what looks like a permanently high plateau” with the uncanny precision of the general whose last words were “don’t be stupid, they’ll never hit me from over th …” is largely remembered for that most infelicitous of predictions. Yet Fisher is one of the greatest economists of the twentieth century and a pioneer of behavioural finance.

Fisher largely lost his reputation as a result of this prediction when, really, we should probably draw a couple of more general conclusions – either that if a great economic theorist can’t see a market crash in front of his eyes then probably all economists are hopeless at prediction or that no-one can tell whether a market is overvalued or not. Either way it doesn’t offer much hope that anyone else can do any better.

1929, Not a Bubble?

Harold Bierman has looked at the evidence for obvious signs of overheating ahead of the 1929 Wall Street Crash and has come up with some interesting data. For instance:
“The financial fundamentals of the markets were also strong. During 1928, the price-earnings ratio for 45 industrial stocks increased from approximately 12 to approximately 14. It was over 15 in 1929 for industrials and then decreased to approximately 10 by the end of 1929. While not low, these price-earnings (P/E) ratios were by no means out of line historically.”
We’d hardly classify a P/E ratio of 15 as overvalued these days which suggests, as a minimum, that attempts to draw long-term conclusions from simple analysis of market valuations are potentially seriously flawed. If the norm of what constitutes a sensible valuation can change significantly over time we need to look for another way of assessing whether markets are over or under valued.

Government Interference

Bierman’s analysis shows that the narrative of speculation and overvaluation that’s been created around the crash of ’29 is largely a classic case of historical revisionism, as commentators use their 20/20 hindsight. He suggests a number of other factors as more relevant to the market collapse. In particular there was a bubble in public utility stocks which was pricked by regulatory concerns which, in turn, forced highly leveraged investors to sell a wide range of stocks to cover their losses. This triggered a panic. This all sounds eerily familiar.

There’s some agreement that the actions of the US government were at least partly responsible for the problems as incipient fears of overvaluation and speculation were fuelled by the incoming Hoover administration. There’s another argument that the relatively timid approach of government and regulators in more recent times is coloured by these historical events. Perhaps those commentators that argue that central banks should have deflated the last bubble need to be careful what they wish for.

War Stories

If it was difficult, if not impossible, to foresee the Wall Street Crash what about something a bit more obvious? Like, say, the two World Wars? We all know the history, about how they were more or less inevitable. Only the data says differently.

Niall Ferguson’s analysis of government bond markets on the lead into the First World War suggests that either markets completely failed to predict the “obvious” slide into conflict or that the war itself was far from certain until the last possible moment. The subsequent devastation to bond prices makes it hard to believe that investors wouldn’t have reacted if they’d realised that war was inevitable.

Meanwhile Frey and Waldenström suggest that Second World War markets didn’t predict anything very much either. In particular they note that the effective outbreak of the war led to large breaks in government bond prices in Sweden and Switzerland suggesting that the “obvious” nature of the war wasn’t so darn obvious at the time.

The Dangers of Hindsight

It’s a bit shocking that three of the seminal events of the last century – the outbreak of World War I, the Wall Street Crash and the start of World War II – weren’t obvious to people, as judged by the evidence of the markets, at the time. Yet, rationally, this must have been the case otherwise people would have been panicking well in advance of these events. It’s a critical lesson – history is only obvious with hindsight, and the commentators picking over the bones of more recent issues are simply blind to their own hindsight biases.

More practically, it might be of use to try and figure out whether more recent events were predictable. Warren Buffett’s (who else?) analysis of the prospects of markets in 2001 is instructive.

Buffett’s Crystal Ball

Back in the Sun Valley conference of 1999 Buffett predicted lousy returns for years to come and was pretty much dismissed as a busted flush by the dotcom hoards. By 2001 the story was rather different. Asked to explain his earlier prescience he pointed to a couple of things: the effects of interest rates on investment returns and the relationships of stockmarkets to GDP.

He showed that throughout the twentieth century we’d seen a consistent decoupling of movements in GDP and the stockmarket. All things being equal if a country’s per capita GDP is going up so should the value of the nation’s stocks. However, that simply wasn’t the case as US stockmarkets remained rooted in gloom while the economy was booming and then took off on a rip just as the pace of growth slowed. What market movements in the late twentieth century were closely correlated with were changes in interest rates. It’s these, allied to psychological effects, that really move markets.

Basically, lowering interest rates causes stock prices start to rise, all things being equal, driving down yields. People notice that shares are rising and start to buy, increasingly because other people are buying. Eventually this pushes stocks to unsustainable levels and, on the grounds that when something can’t continue it won’t, markets drop, sometimes vertiginously.

The Madness of Markets

As Buffett more than adequately demonstrated at Sun Valley, even the professionals get trapped by hindsight into assuming that the future will look the same as the past.
"By 1982 … [the pension funds] had moved up their assumptions a little bit, most to around 7%. But now you could buy long-term governments at 10.4%. You could in fact have locked in that yield for decades by buying so-called strips that guaranteed you a 10.4% reinvestment rate. In effect, your idiot nephew could have managed the fund and achieved returns far higher than the investment assumptions corporations were using."
Why in the world would a company be assuming 7.5% when it could get nearly 10.5% on government bonds? The answer is that rear-view mirror again: investors who'd been through the collapse of the Nifty Fifty in the early 1970s were still feeling the pain of the period and were out of date in their thinking about returns. They couldn't make the necessary mental adjustment.

Hindsight hurts investors both ways. We look back and think that everyone foresaw the future and assume we can as well, yet simultaneously we make predictions from our own little bubble of past experience. Meanwhile herd effects drive people in and out of assets far more effectively than the numbers ever do. History tells us how to invest; not by counting the followers of a trend but by following the numbers that count. Patience does the rest.

Be prepared.

Related Articles: Hindsight Bias, Investing in the Rear View Mirror, It's Not Different This Time


  1. It is one of the small ironies of the market that is it easier to predict long-term returns than short-term returns.

    That is why someone sensible like Buffett will occasionally make long-term predictions (1969 bearish, Oct 1974 bullish, July 1999 bearish, October 2008 bullish), but will never make a prediction of where the market will be in a month or a year.

  2. From 2/24 Interview
    Robert Prechter of Elliott Wave International says that the huge and inflated bond market is a dangerous bubble. “The individual investor has been more or less abandoning stocks” and buying bond funds, Prechter (says). “I think that is going from the frying pan into the fire. The bond market is the biggest bubble in the history of the world. ”
    On 2/24 yield on 10 year Treasury = 3.69
    Today yield on 10 year Treasury = 3.43.
    Prechter not alone in his prediction. How many people took short positions because they bought into this obvious bubble?
    What people can't get is that if we identified a bubble there wouldn't be a bubble! This is called market efficiency.
    Of course if the yield on the 10 year was 5% today there would be a lot "I told you sos".
    I thought we had a bubble when the DJIA moved from 1,000 to 5,000 so don't read me as thinking that I got the answers.

  3. It is one of the small ironies of the market that is it easier to predict long-term returns than short-term returns.

    I'm with Parker.

    Lots of people have a hard time understanding why short-term predictions of stock returns never work and yet long-term predictions of stock returns always work. I believe that the key is focusing on what it is that makes stock-return prediction possible.

    What makes prediction possible is that stock returns must ultimately reflect the economic realities. Now -- Do economic realities always assert themselves immediately or do they sometimes only do this over time?

    Consider the house in a gambling casino. The house always wins in the long run, right? But the house does not always win in the short run. A gambler who pulls a slot machine lever 10 times stands a good chance of coming out ahead. A gambler who pulls a slot machine lever 10,000 times stands no chance of coming out ahead. Statistical probabilities often take time to assert themselves.

    So it is with stock investing. Those who take valuations into consideration always end up ahead eventually. But they can be behind for three years or five years or ten years. There is a law that says that those with the odds on their side are always going to over-perform. But there is no law that says that those with the odds on their side are always going to out-perform quickly.

    Rational Investing is patient investing.


  4. From the looks of your numbers Tim, this is another example showing the fallacy of low P/E investing as a foolproof tool.

    Just as high P/Es can presage big rebounds with growing earnings, so can low P/Es be rampant when cyclicals are at their peak.

    The UK banks and miners were on fairly low P/Es in 2007, and look what happened to them!

    I strongly favour behavioural aspects now when it comes to trying to judge (for my sins!) when a market is over-valued.

    When Rob Bennett thinks the market has reached a permanently high plateau, I'll sell and buy gold. ;)