Thursday, Oct 18, 2007

Twenty years ago on Friday, traders in New York arrived for the start of the working week in a worried frame of mind. Wall Street stocks had shed more than 5 per cent the previous Friday, leaving them with a weekend to ponder where the market would go from there.

Before their day's dealings began, they heard news that Hong Kong's stock exchange had fallen more than 5 per cent. More dramatically, prices on the London Stock Exchange, in part still reacting to the effects of a hurricane that had hit southern England at the end of the week before, were down almost 10 per cent. There was little other news but international tensions were rising, with some sabre-rattling between the US and Iran.

The traders assumed that business on Monday October 19 1987 was going to be difficult. Some breakfasted together and discussed the losses seen overseas. Leaving the table, Art Cashin, a broker for 25 years, borrowed a phrase from the Roman gladiators: "Those who are about to die salute you."

By the time the market closed a few hours later, many on the floor of the exchange felt as though they had been fed to the lions. The Dow Jones Industrial Average fell more than 22 per cent, by far the greatest single-day fall in its history. It was double the percentage slide suffered on the worst day of the great Wall Street crash of October 1929.

It provided an extreme example of market behaviour that academics, as well as many market participants, have been trying to explain ever since. "The first lesson was that such an event can happen," says Richard Thaler, a professor at the University of Chicago and one of the leading figures in the emerging field of behavioural finance. "If you had taken a poll of financial economists before that week - which also included two of the biggest advances in history - and asked them if it could happen, the answer would have been No."

Although how it happened remains a subject of debate, consensus has emerged on a few points. In particular, two preconditions were plainly in place.

First, the market had become by many measures overvalued, setting it up for a crash. Its price/earnings ratio, at 23, was high. Meanwhile, government bonds had suffered a sell-off during the summer and risk-free yields of more than 10 per cent were available. That made the year's huge rally in stocks, which saw them gain more than 40 per cent by mid-August, hard to justify.

Jeremy Grantham, founder of Grantham Mayo Otterloo, a Boston fund manager, was quoted on the cover of a newsletter two weeks before the crash: "Buying bonds at 10.3 per cent and selling stocks is almost a free lunch."

Second, having become too expensive, the market had run out of momentum. By the weekend before Black Monday, fear had gained the upper hand over greed. The market had peaked two months earlier and slipped. Technical analysts, who follow patterns in markets, thought the Dow looked like a "sell". Ron Daino, part of the technical analysis team at Smith Barney, had issued such a warning the week before. But he admits he had no clue that the crash would be so severe.

An overvalued market where traders have become more pessimistic is likely to fall. But these factors are not enough to explain a 20 per cent dive in one day. For this, the culprit named at the time was a hedging strategy known as portfolio insurance. For Mr Grantham, the entire crash was a "technical accident caused by portfolio insurance".

This was a crude but popular strategy for limiting losses. If a portfolio of stocks was falling, computer programs would hedge by taking a position in S&P 500 index futures (on the Chicago Mercantile Exchange) at an even lower price. If things grew worse for the stocks in their portfolio, investors would make money from the futures. This would limit their losses.

The problem was that this was very predictable. As the market fell, portfolio insurance programs would seek to buy futures, and sell stocks, at lower and lower prices.

Rob Arnott, now the head of Research Affiliates, was a global equity strategist at Salomon Brothers at the time. It was his job to monitor the likely demand for portfolio insurance. After Friday's loss, he calculated that portfolio insurers alone would start the day seeking to sell more than the market's average volume for an entire day.

Once he took this to his superiors, Salomon's response was to order its traders not to take the other side of the trade. Mr Arnott suspects the other big Wall Street firms took the same stance. So the effect was as if the portfolio insurers were selling into a vacuum, with the prices on offer to them marked down and down. Arbitrageurs then worsened the problem by placing orders to sell stocks, to take advantage of differences between prices in New York and Chicago.

But stocks had still not opened, as traders had been unable to establish a price at which they could trade, given the heavy sell orders they had already received. "When you have a crash," says Mr Arnott, "you have enough uncertainty that it paralyses the people who might otherwise take the other side of your trade."

John Phelan, then the chairman of the New York Stock Exchange, says the mechanics of portfolio insurance were explained to him in 1986. "I said, my God, what you are talking about is a potential waterfall." He told the Securities and Exchange Commission that portfolio insurance was "a recipe for disaster" but says he never anticipated it could lead to a 20 per cent-plus fall in one day.

So those were the main factors that led to Black Monday: expensive stocks, combined with the perverse effects of portfolio insurance. But in understanding these elements, can we predict whether another crash is coming?

On valuation, one of the models for showing when bull markets are about to turn looks at share price multiples compared with cyclically adjusted corporate earnings. Profits are cyclical. Therefore, when profits are high, earnings multiples should be lower. Robert Shiller, the Yale University economist, developed a system of cyclically adjusted price/earnings ratios for his book Irrational Exuberance, which famously foretold the bursting of the technology bubble.

Results produced by that method (see charts) show the market was wildly overvalued in 1929 and 2000. More worryingly, 1987 does not look so extreme - and valuations today have crept above the levels associated with the top of a bull market.

Such models cannot predict one-day falls but they do show very clearly when markets are at risk of turning from bull cycles - steadily trending upwards - into bear markets. It looks as if those conditions are currently in place.

Another element for predicting a crash is technical analysis. Markets do follow predictable patterns, as is obvious from the remarkable similarity between charts of the markets in 1929 and 1987. The danger signals for technical analysts flash when a period of euphoria or momentum runs out and doubts creep in. Barring external news, a period of euphoria rarely moves immediately to panic without some intermediate signals that sentiment has run out of steam.

But even its leading practitioners admit that technical analysis cannot predict a crash. "If I knew, I wouldn't have called half a dozen crashes that didn't happen!" jokes Robert Prechter of Elliott Wave International, a widely followed group of technical analysts.

While trends and turns in the market can be predicted with some reliability, he says, crashes are extreme and harder to see.

That leads to another question: when is a crash a buying opportunity? In retrospect, October 1987 was a great time to buy. The Dow was back to its pre-crash level in 16 months. Tokyo's Nikkei got there by April 1988 and in London the FTSE had regained its losses by August 1989.

But October 1929 was emphatically not a good time to buy. Anyone who had bought after the second of the two days of that great crash (and many did) would have lost 84 per cent of their money within two years.

What explains the speedy recovery in 1987? First, the selling was overdone - it moved the market from being overvalued to undervalued in one day. "It went into the day normally expensive and came out normally cheap," says Mr Grantham.

With valuations cheap, the conditions were already in place for a recovery. This was not true in 1929 or in 2000, when the market had entered the crash at much more extreme valuations.

Second, the Federal Reserve played its part, along with other financial authorities. Interest rates were cut. Troubled savings and loans providers were bailed out. The underlying US economy remained sound. In 1929, by contrast, a series of errors by the Fed helped turn the financial crisis into an economic disaster.

So crashes can be great buying opportunities, but only if markets reach favourable valuations and if there is reason to be confident in the underlying economy.

Can they be prevented? Markets are different now. Average daily volumes on the New York Stock Exchange alone have increased tenfold since 1987, the venues on which stocks can be traded have proliferated and the speed of trading is unimaginably faster. The range of hedging opportunities using derivatives is much larger than it was.

The view from the floor is that the improvements in technology will not prevent crashes. "But it will help them happen that much faster," says Mr Cashin, the veteran broker.

Andrew Lo, an economist at the Massachusetts Institute of Technology, argues that these developments have also made crashes more likely. "We have a much more well-integrated and well-connected set of financial markets. Disruption in one market can very easily spill over into other financial markets."

He draws comparisons with recent events, which saw a number of quantitatively managed hedge funds - which use complex mathematical algorithms to trade swiftly among different stocks and asset classes - suffer huge losses in August amid the fall-out from the credit crisis. Portfolio insurance, Mr Lo suggests, was "a microcosm of what we see today, writ large".

Further, a growing field known as neurofinance attempts to apply insights from cognitive psychology to decisions made by investors. These insights suggest that the human mind is "hard-wired" to make the kind of collectively irrational decisions that can lead to crashes such as Black Monday. No amount of technology will change this.

According to Richard Peterson, author of Inside the Investor's Brain, when markets are rising, investors show primitive "chasing" behaviour, expecting gains. Once gains fail to fulfil their expectations, the loss-avoidance part of the brain is engaged. This is when herd instincts come into play and sentiment grows more negative. Then comes panic - and the crash.

These factors are constant. Mr Peterson suggests that investors can try to be more emotionally self-aware, but this is difficult and requires behaviour that many people find unnatural.

Mr Lo puts it differently. "Our brains are hard-wired and optimised for decision-making on the African savannah 100,000 years ago," he says. "If we are being chased by a sabre-toothed tiger, it's perfectly appropriate for the brain to get a shot of adrenaline and run like hell. Unfortunately, that won't help you on the floor of the NYSE when the S&P is down 20 per cent."

Twenty autumns on, the response needs to be different

Christopher Cox, chairman of the US Securities and Exchange Commission, makes a habit of pointing out that regulators must recognise that regulation "is not the fuel that drives our markets". Instead, it is the "oil that greases the gears" write Jeremy Grant and Gillian Tett.

In the immediate aftermath of this summer's credit squeeze, regulators are struggling with a depressing fact: the crisis has yet again exposed shortcomings in the regulatory regime. The gears, in effect, have rusted up - irrespective of all the lessons in the past few decades about what makes markets run smoothly.

The Federal Reserve, Treasury, SEC and others have been in constant motion since the problem of subprime mortgages erupted. But an immediately clear regulatory response did not emerge.

Barney Frank, Democratic chairman of the House of Representatives financial services committee, said this week: "We have been here too often - financial market innovation leading to, in [former Federal Reserve chairman] Alan Greenspan's words, 'irrational exuberance', while market participants and regulators found it increasingly difficult to measure and account for risk."

There is political pressure to ensure that something is done in response to the credit squeeze, particularly in the US. But central bankers and regulators agree it would be foolish to rush policy responses.

As one senior policymaker observes, it is still far from clear that the dust has fully settled from the events of the past two months. There is also a common desire to avoid kneejerk overreactions. Painful memories exist of the overkill many perceive in the Sarbanes-Oxley law that followed the Enron accounting scandal.

"In the US, there is real political pressure to do something - but nobody wants to create even more mistakes," says one senior policymaker.

This means that in the short term, at least, policymakers are likely to focus on two potentially easy scapegoats: the mortgage brokerage industry in the US and the credit rating agencies.

Evidence of action came this week when Hank Paulson, Treasury secretary, called for the "patchwork" of mortgage regulation to be streamlined. Calling the behaviour of some mortgage originators and brokers "shameful", he urged a nationwide system of regulation for mortgage brokers that would enhance the state-by-state system that is seen as having allowed shoddy underwriting standards to proliferate.

On ratings agencies there is less consensus. While some Europeans want to see much tighter scrutiny, their US counterparts appear more wary.

Meanwhile, efforts are under way to restart some markets that have become paralysed. Citigroup, JPMorgan Chase and Bank of America, the three top US banks, this week unveiled Treasury-blessed plans for a fund that would buy up to $100bn (70bn, GBP49bn) of complex assets from structured investment vehicles.

Nonetheless, policymakers argue that in the longer term, bigger shifts are probably needed in the way that regulators and central bankers approach the financial world. That is because the issue that has hurt the markets this summer is strikingly different from the troubles of 1987: the difficulty of valuing assets at a time when risk has been widely dispersed through complex and opaque financial instruments.

This means, some policymakers say, that what is needed now is more regulation - but a different kind of regulation that addresses the developments that have sprung from financial market creativity. "It's really hard to figure out what to do," says Kathleen Hagerty, professor of finance at the Kellogg School of Management at Chicago's Northwestern University. "So much of this is how you as a regulator are going to see what they [market participants] have. You have all these instruments and it's really hard to value those things."

That is why Mr Frank, while welcoming the effort by the three big banks, says banks must still "increase the scrutiny they give to the opaque and difficult-to-value securities that, beginning with the mortgage market, underlie the financial turmoil of the last two months".

The problem confronting policymakers is that the financial system is so integrated that banks and other investors can easily straddle borders - picking and choosing regimes or playing one off against another.

Most banks have stopped retaining credit risk on their own books, instead slicing and dicing some of it into securities that can be sold to other, largely unregulated, investors such as hedge funds that operate globally. This has made it hard for regulators to work out where this risk has gone or how investors might behave under stress. Some observers conclude that regulators will need to collect far more data from non-bank players - or perhaps demand more transparency in opaque corners of the market. But there is little appetite in the US for a wider clampdown on hedge funds.

One key issue for reform is how banks handle their liquidity needs. Another is the accounting treatment of investment vehicles that partly lie off banks' balance sheets - such as structured investment vehicles and conduits. Prof Hagerty says: "It's hard to value them accurately, because they are not exchange-traded, so there is no ready price. I can't see what good comes from keeping things off-balance sheet in terms of the clean running of the markets."

Others argue that banks should be forced to post capital reserves against loans they have sold on, if they have retained a reputational risk or other links. "This is an area where there needs to be some more debate," says David Dodge, Canadian central bank governor.

The Financial Stability Forum, where central bankers and regulators around the world meet to debate, is expected to produce recommendations on these issues early next year. Separately, the US Treasury is preparing to unveil a "blueprint" for regulatory reform early next year.

Mr Paulson set in motion a top-to-bottom examination of the structure of US financial regulation last November - before the current crisis. This week, he said his department would review the accounting treatment of off-balance sheet vehicles in the mortgage asset market.

David Nason, the Treasury's assistant secretary for financial institutions, said in a speech this week that regulations must adapt to changing times: "Markets are constantly evolving. We should analyse and understand the rationale or justification for our current regulatory structure as well as the inefficiencies it can breed."

John Authers