Bookblogging: Failure of the EMH

Another section from the forthcoming book. Casting suggestions for the blockbuster movie will be gratefully accepted, along with more prosaic correction of errors, omissions, and of course, compliments. I’m trying to get a nice HTML version, but will see how it goes

Failure

As with the other doctrines discussed in this book, the failure of the EMH is not a sudden shock arising from the Global Financial Crisis. The evidence for the strong forms of the EMH was never particularly convincing. Rather, it was an idea that suited both the demands of the times and the intellectual tendencies that were dominant within the economics profession.

During the 1970s and 1980s, assessment of the EMH was largely confined to econometric studies. The process of financial deregulation, beginning with the breakdown of the Bretton Woods system in the early 1970s was a gradual one. It was only by the mid-1980s that most restrictions on international capital flows and exchange rate movements were completely removed. The end of domestic deregulation took even longer. So, it was not until the 1990s that failures of the global financial system could reasonably be regarded as evidence against the EMH.

That evidence was not long in coming. A number of developing countries experienced severe financial crises in the 1990s, even though their governments had done their best to follow the policy prescriptions of economic liberalism, in particular by deregulating financial markets and encouraging private investment. 

The experience of the US itself provided plenty of evidence against the EMH. The near-collapse and government-orchestrated rescue of hedge fund LTCM provided a preview of the massive bailouts of 2008 and 2009, undermining some key assumptions of the EMH in the process. Even more significantly, the boom and bust in the shares of ‘dotcom’ companies that promised to generate vast profits from the Internet showed that all the sophistication and complexity of modern financial markets only served to make possible bigger and better bubbles.

Sadly, these lessons went unlearned. Despite repeated failures to meet the test of experience, the EMH remained central to finance theory and to policy practice, right up to its final catastrophic collapse in the meltdown of 2008.

Econometric testing

Like many of the dead ideas discussed in the book, the efficient markets hypothesis was not particularly well supported by empirical evidence even in its heyday. The weak form of the hypothesis, was reasonably successful when subjected to the statistical tests applied by econometricians, but the strong and semi-strong forms much less so.

As was discussed above, the weak form of the hypothesis precludes the existence of predictable patterns in asset prices (unless predictability is so low that transactions costs exceed the profits that could be gained by trading on them). Broadly speaking, this weak version of the EMH is consistent with the evidence. At least, there are no simple and reliable trading strategies that have been shown to beat the market consistently.

On the other hand, econometric studies given little support to the semi-strong and strong forms of the hypothesis.  Most importantly, as economists such as Robert Shiller has shown, the volatility of asset prices is much greater than is predicted by the EMH. That is, where the EMH suggests that financial markets provide a way of managing economic risk, the evidence suggests that they are actually a major source of such risk.

The behavior of currency markets after the breakdown of the Bretton Woods agreement provides a good example.  A large body of economic analysis shows that, in the long run, exchange rates must be somewhere close to Purchasing Power Parity, that is, the rate at which a US dollar can buy the same goods in the US as it can if converted into euros and spent in Europe ¹ Advocates of floating exchange rates confidently expected that, whereas governments had frequently tried to maintain rates inconsistent with PPP, financial markets would bring exchange rates into line with underlying economic values, and thereby lead to greater long-term stability. In fact, the reverse has happened. In the decade since the creation of the euro, its value has been as low as 85 US cents and as high as $US1.50, even though price levels in both the EU and the US have been quite stable, and interest rates fairly comparable.Stock markets similarly display much more volatility than the EMH suggests is consistent with the observed variability of aggregate consumption. 

As with the closely related equity premium puzzle (see CH ..), it is easy enough to see that the standard theory underlying the EMH (and the closely associated capital asset pricing model). A recent survey by Stephen LeRoy of UCSB concluded 

no single convincing explanation has been provided for the volatility of equity prices. The conclusion that appears to follow from the equity premium and price volatility puzzles is that, for whatever reason, prices of financial assets do not behave as the theory of consumption-based asset pricing predicts.

¹ For a number of reasons, including the fact that many services aren’t traded on international markets, this equality isn’t expected to hold exactly. But it should be close.

Emerging market financial crises

After the turmoil of the 1970s and 1980s, developed countries enjoyed a period of sustained economic expansion in the 1990s, with the United States leading the way. For a while, less developed countries (now relabelled as ‘emerging markets’ in the financial sector) enjoyed similarly smooth sailing. But from the mid-1990s onwards, there were a string of financial crises in Mexico, Argentina, Russia and most spectacularly, East and South-East Asia.

The developing-country financial crises of the 1980s had followed a pattern that supported an efficient markets hypothesis. Governments had borrowed heavily, spent the proceeds on military adventures or luxury projects, intervened to distort market prices, and attempted to restrict international capital flows. When they got into trouble, as they inevitably did, they were forced to call on the International Monetary Fund for help. 

Its standard prescriptions supported by the US Treasury and the World Bank were christened the ‘Washington Consensus’ by John Williamson of the Institute for International Economics. Williamson listed ten elements of the standard package, notably including financial deregulation and privatization of state enterprises. Although Williamson’s initial presentation included discussion of the need for appropriate prudential regulation of financial institutions, later versions of the Washington Consensus dropped this element, and incorporated more radical versions of economic liberalism, to the point that Williamson himself disavowed the phrase.

The crises of the mid-1990s hit countries that had, in general, embraced the policies of the Washington consensus. The pattern was the same in each case. Following financial deregulation, countries enjoyed strong capital inflows and booming stock markets. Some seemingly minor event produced a reversal in market sentiment and a sudden flight of capital, producing an economic crisis. Following the crisis, the IMF and world markets sought to impose the 1980s package of public expenditure cuts and economic contraction, which only exacerbated the problem. Finally, in retrospect, the victims were blamed for minor divergences from the free-market ideal which, before the crisis, had been seen as unimportant, or even praiseworthy.

 Asian economies had enjoyed decades of strong growth through policies of export-oriented industrialisation, rejecting the ‘import replacement’ policies, aimed at economic self-sufficiency that had been tried and failed elsewhere. From the early 1990s onwards, they had been engaged in a process of financial deregulation. Only a year before the crisis hit, the World Bank had produced a glowing report praising the ‘Asian economic miracle’ as an exemplar

The case of Argentina was even more striking. Rejecting the failed policies of the Peron era, Argentina had adopted the most extreme version of the Washington Consensus, privatising industry on a large scale, and even establishing a currency board to guarantee a fixed exchange rate with the US dollar. Yet when the economy ran into trouble, the financial markets were left to fend for themselves. 

The Asian financial crisis cast doubt on the idea that globalization was both inevitable and beneficent, as did the failure of Washington consensus policies in Argentina. Even more embarrassing was the success of Malaysia, which r imposed controls on the movements of foreign exchange, the cardinal sin against global financial markets. Unlike neighbors who followed the advice of the IMF, Malaysia was largely unaffected by the crisis.

Despite all this, the confidence of financial markets and policymakers in the Efficient Markets Hypothesis was unshaken. The Asian countries that had been seen, only a year earlier, as reflecting the fruits of reform were denounced as embodiments of ‘crony capitalism’. The conclusion drawn was that, only with a fully-developed, transparent and incorruptible financial system, like that of Wall Street or the City of London, could the benefits of financial markets be fully realised.

The LTCM fiasco

 The debate over the EMH gave rise to the view thatt the market is just close enough to perfect efficiency that the returns available from exploiting any inefficiency are equal to the cost of the skill and effort that goes into discovering it. This idea is central to the operations of hedge funds, which seek to discover strategies by which investors willing to take a risk can earn above average returns. However, it came spectacularly adrift in 1998, with a hedge fund called Long Term Capital Management, which employed as its expert advisers, none other than Robert Merton and Myron Scholes.

The strategy employed by LTCM was to discover small deviations from efficient market pricing, on which it could make bets that were sure to win. Instead of simply betting with its investors money, it used that money as equity for massive borrowings, which ensured that the payoff from its winning bets was multiplied many times over. Reliance on ‘leverage’ to multiply profits has been a characteristic of many financial bubbles, a point reflected in the saying “Genius is leverage in a rising market”. But it has never before operated on the scale seen in the Great Bubble, and exemplified by LTCM.

Thanks to the use of complex derivatives, LTCM turned an equity base of less than $5 billion into derivative positions with a notional value of approximately $1.25 trillion. These derivatives, such as interest rate swaps, were developed with the supposed goal of allowing firms to manage risk on exchange rates and interest rate movements. Instead, they allowed speculation on an unparalleled scale.

The particular focus of LTCM was what is called, in the picturesque jargon of the financial markets, the “yen carry trade”. The basis for the trade is the fact that, ever since Japan’s own bubble economy exploded in 1990, interest rates in Japan, and therefore any interest rates for debts denominated in Japanese yen, have been very low. Hence, money can be made by borrowing money in yen, “carrying” it to the US market, and lending it in US dollars. The catch is that if the dollar depreciates against the yen,

LTCM did not focus on the yen carry trade itself, but on the interest rate margins that drove the trade. It looked for divergences between the margins generated by the markets and the values predicted by its computer models, then bet that the market would ‘correct itself’ over time. These bets paid off for a number of years, making big profits for LTCM owners and investors. But, in 1997 with Asian financial crisis, all its bets failed at once.

The unregulated status of hedge funds had been justified on the basis that the investors were sophisticated and wealthy individuals, and that only their own money was at risk. But it soon turned out that the leveraged investments made by LTCM had been financed by huge loans from major Wall Street and international banks, and that a failure by LTCM ran the risk of generating a systemic collapse. The US Federal Reserve, under Chairman Alan Greenspan, orchestrated a rescue package, with major banks being pressured to contribute (among the Wall Street investment banks, Bear Stearns was the only one to refuse). The crisis was staved off, and the LTCM principals and investors escaped with much of the wealth gained from their earlier successful bets intact.

The LTCM episode had numerous lessons, many of which pointed out at the time but few of which were taken to heart by policymakers.

In retrospect, the strategy pursued by LTCM can be seen as a variant on the ancient ‘martingale’ betting strategy. As Slate writer Jordan Ellenberg explained in an excellent article last year, the strategy can be illustrated by betting on a coin 

Bet 100 bucks on heads. If you win, you walk away $100 richer. If you lose, no problem; on the next flip, bet $200 on heads, and if you win this time, take your $100 profit and quit. If you lose, you’re down $300 on the day; so you double down again and bet $400. The coin can’t come up tails forever! Eventually, you’ve got to win your $100 back.

The problem with the martingale is that you are trading off a steadily diminishing probability of losing against a steadily increasing loss if you do. At some point, there will be a run of tails long enough to bankrupt you.

What is true of the martingale is true of more sophisticated variants, like the strategies of LTCM. This point ought to have been evident from the LTCM failure. Instead, the conclusion drawn by both financial markets and regulators was that the problem could be solved by designing ever more complicated derivatives. By the time the whole thing blew up in 2007, the complex trades that had brought LTCM to grief looked like child’s play.

A second crucial point is that diversification of risks only works to a certain extent, and can be undermined by attempts to exploit it.Once the limits of diversification have been reached rearranging the set of claims involved isn’t going to reduce risk any further, so if all parties appear to be making risk-free profits, the risk must have been shifted to some low-probability, high-consequence event. 

LTCM relied in part on the assumption that currency markets were largely independent, so that losing bets made on one currency would, in general, be cancelled out by wins on others.

There were also lessons for regulators. The first was that no system of financial regulation can survive if some firms are guaranteed and regulated, but are allowed to deal on a large scale with others that are not regulated. The second is the old one of ‘moral hazard’: if people are protected by insurance from the bad consequences of risks, they will tend to take more risk as a result.

Financial market players ignored a ll these lessons, but they did learn one big one, which was the opposite of the moral hazard lesson ignored by regulators, namely, the existence of ‘Greenspan put’. A put is a kind of option allowing you to sell a stock at an agreed price on a given date. In effect, the holder of a put has a one-way bet on the stock they own. If it goes up, they sell the stock on the market and collect the profits. If it goes down, they exercise the put option and collect the agreed price.

Precisely because they are so attractive, put options are valuable ( the Black-Scholes rule shows how to value them). What was special about the Greenspan put was that it was free. The LTCM precedent showed that, if financial markets ever got into really serious trouble, the Federal Reserve would bail them out. So, any kind of risk taking behavior became a one-way bet, as long as sufficiently many of the big financial institutions were making the same bet. 

The first exercise of the Greenspan put came in the wake of the dotcom crash, discussed in the next section. The second, and much bigger one came in October 2008, at the hands of Greenspan’s successor, Ben Bernanke. This time, though, not even a trillion dollar bailout was enough to save all the big financial institutions from the consequences of their reckless speculation.

The dotcom bubble

Stock markets in the US and elsewhere rose strongly in the 1980s and 1990s, interrupted only briefly by the crash of October 1987 (which, in retrospect, fostered the illusion that any decline in stock prices would be quickly reversed). By 1996, the boom had reached the point where, with the Dow Jones index at 8000, Alan Greenspan warned of the dangers of ‘irrational exuberance’ in asset markets. Greenspan never repeated the warning, and soon returned to his customary role a cheerleader for speculative markets. However, the catchphrase was adopted by economist Robert Shiller as the title of a penetrating analysis of the role of self-deception and collective over-optimism in stock market bubbles

The bubbles had raised stock prices in general, but it was propelled to new heights by the arrival of the ‘dotcom’ sectors. The Internet, developed as a public service by the US government research agency DARPA and by the university sector worldwide, was opened to commercial use in the 1990s, just as its most popular manifestation, the Worldwide Web was coming online. 

In 1995, the Mosaic web browser, created at the publicly-funded National Center for Supercomputing Applications, was converted into a commercial product named Netscape, which formed the basis of a spectacularly successful Initial Public Offering (IPO). The stock was set to be offered at $14 per share. But, a last-minute decision doubled the initial offering to $28 per share. The stock’s value soared to $75 on the first day of trading, nearly a record for first-day gain. 

Never profitable on an annual basis, Netscape was acquired by America Online (AOL) in a stock-swap valued at US$4.2 billion in 1998. A couple of years later, in the biggest merger in history, AOL merged with Time Warner. The deal gave AOL a market value of more than $100 billion; it is now valued at around $4 billion.

The Netscape IPO and AOL takeover set the pattern for a string of ever more dubious “dotcom” ventures, producing huge gains for investors despite the absence of significant profits, and in many cases, even revenues or products. The history of Netscape and AOL was mirrored by thousands of firms which attached the dotcom suffix to businesses as mundane as selling dogfood and garden supplies, or as spurious as that of the pioneering entrepreneur of the 1713 South Sea Bubble who sold shares in “a company for carrying out an undertaking of great advantage, but nobody to know what it is”. Indeed whereas Netscape and AOL had substantial revenues, and AOL had a profitable business as an Internet service provider, the typical dotcom never made a genuine sale, let alone a profit.

Speculation on dotcoms centred on the NASDAQ stock exchange. ¹ The NASDAQ index rose from 800 in the mid-1990s, to over 5000 at its peak in March 2000 when it collapsed suddenly, falling below 2000 (as of May 2009, the index stood at 1700). Hundreds of dotcom companies failed or were taken over at prices far below those of the late 1990s. 

Even more than the complex global crisis now underway, the NASDAQ bubble and bust provided a sharp test of the efficient markets hypothesis, which failed egregiously.  It was obvious, and pointed out by many observers, that the prices being paid for dotcom investments could not be justified on the basis of standard principles of valuation. Even if some turned out to be the spectacular successes promised in their business plans, it was impossible that the sector as a whole could do so. (In reality, the only company created in this period that is earning large and growing profits as of 2009 is Google, and its shares were not traded until 2004).`

Previous bubbles might have been dismissed on the basis that the markets concerned weren’t fully informed and transparent, or that speculators were prevented from betting against the bubble assets and thereby bringing prices back to earth. The dotcom bubble showed that none of these defences worked.

As regards transparency, no market in history has been subject to such intense scrutiny and obsessive coverage as the NASDAQ of the late 1990s. Stocks and the companies that issued them were assessed by investment banks, stockbrokers and the financial press. The dubious projections on which they relied were set out in prospectuses which warned (in a pro forma fashion) that they might not be fulfilled.

Speculators did attempt to burst the bubble. Julian Robertson of Tiger Investments, short-sold grossly overvalued tech stocks in the late 1990s and lost billions when the stocks rose even further in 1999. He quit managing other people’s money, telling clients that he no longer understood the markets.

Although the dotcom bubble and bust was spectacular, the 2000-01 crash was at least equally significant for the exposure of corporate fraud on a scale unparallelled (at the time) since the 1920s.  The two biggest frauds, Enron and Worldcom offered a sharp contrast. The Enron frauds relied on a complex network of trading schemes, special purpose vehicles and elaborate accounting devices. By contrast, the managers of Worldcom simply invented revenue numbers that made the company look massively profitable when it was actually losing money hand over fist.

¹ A competitor to the New York Stock Exchange that had been established by brokers including Bernard Madoff, who confessed in late 2008 to having operated the biggest Ponzi scheme in history

The crisis of 2008

The bursting of the dotcom bubble spelt, or should have spelt, the end of belief the strong forms of the efficient market hypothesis. On the other hand, by exposing weaknesses in the systems that were supposed to keep financial markets operating properly, it gave regulators and financial institutions a chance to clean up, so that future outcomes could be more like those predicted by theory.

Neither of these things happened. Advocates of the efficient markets hypothesis, in general, simply ignored the dotcom fiasco, and went on as if nothing had happened. The accounting scandals at Enron and other companies produced the Sarbanes-Oxley Act, which sought to reform corporate governance. But the Act was limited and largely ineffectual. Within a year or two, the conventional wisdom of the financial markets was that Sarbanes-Oxley was an over-reaction to isolated cases of fraud, and that a new push for deregulation was needed.

Financial institutions could disregard the failures of the dotcom bubble because of the (seemingly successful) operation of the Greenspan put. Rather than let the financial sector suffer the consequences of the bursting bubble, Greenspan relaxed monetary policy and inflated a whole new bubble, this time in housing.

The housing boom in the US was not spectacular by global standards. Its crucial characteristic was that both the boom and the subsequent bust took place in all major markets simultaneously.  As with the LTCM disaster a decade earlier, the models used by financial instruments to rate the riskiness of mortgages and assets derived from those mortgages incorporated the assumption that separate housing markets in the US were largely independent of each other. Hence, a diversified portfolio of US mortgages was highly unlikely to suffer losses on all or most of its holdings at once.

But the very transactions justified by the models undermined the assumptions on which they were based. The demand for diversified portfolios meant that lenders lowered their standards in all markets at once. Whereas previous US real estate booms had been based on local factors leading to optimism about the prospects for particular markets, the boom of the early 2000s was based on a general believe that real estate, as an asset, was bound to go up in value.

This assumption was embodied in the construction and pricing of an ever more complex range of financial derivatives. The process began with the observation that, if house prices kept on rising, the absence of a downpayment was not a problem, since the borrower’s equity would rise with the price of the house. That in turn meant that it would be possible to refinance a loan on more favorable terms. 

Hence, on this assumption, it made sense to offer ‘negative-amortisation’ loans, in which, for an initial period of two or three years, the borrower did not pay down the debt at all, but added to it. After the initial ‘honeymoon’ period, these loans were set to revert to much more stringent terms, but it was convenient for everyone to assume that, when the time came, the loan could be refinanced.

Based on these assumptions, investment banks were prepared to buy securities based on loans made by mortgage lenders such as Countrywide. The resulting loss of market share by Fannie Mae and Freddie Mac led these institution to lower their standards. Beginning in 2004, they entered the subprime market on a large scale, relying on their implicit guarantee to hold down borrowing costs. Increasingly competitive securitisation also reduced the incentive of the original lenders to monitor the creditworthiness of borrowers; once they had packaged the mortgages into securities they were no longer exposed to the risk of default, and the demand for securities was so strong that quality was not a major problem.

The growth in demand for mortgage-backed securities reflected a range of innovations, such as the rise of bond guarantors, and the development of collateralized debt obligations (CDOs) under which a portfolio of mortgage-backed securities was transformed into a set of assets some of which were supposed to pay off even in the event of a downturn in local housing markets (the possibility of a national downturn was excluded from consideration in the models used to rate these securities). 

These and other devices, combined with optimistic assumptions about default and repayment rate, made it appear that the risks associated with lending could be made to vanish. With the blessing of ratings agencies such as Moody’s and Standard & Poors, loans to people who might have neither a regular income, nor a job, nor any  asset except the house itself (the acryonym NINJA (no income, no job or assets) was used to describe them) were transformed into ‘super-senior’ bonds given the same AAA credit rating accorded to the US government itself.

By late 2006, loans to borrowers with weak or non-existent credit formed the basis of an inverted pyramid amounting to trillions of dollars of spurious assets created by banks and hedge funds around the world. Some of these institutions were explicitly backed by national governments. Many others were ‘too big to fail’; given the complex and fragile web of financial transactions built up since the 1970s, the breakdown of even a medium-sized player could bring the whole system to a halt.

The stage was set for a global economic meltdown. The crisis built up slowly over the course of 2007, as the growth in house prices slowed, and ‘subprime’ borrowers faced foreclosure. By mid 2007, the problems had spread more widely, to classes of borrowers seen as less risky. CDOs and other derivatives, originally rated as AAA, were downgraded on a large scale and some went into default.

Throughout all this, the dominant view, informed by the EMH, was that nothing could or would go badly wrong. It was not until investment bank Bear Stearns was rescued from imminent bankruptcy in March 2008, that confidence started to crack. By this time, as the National Bureau of Economic Research subsequently determined, the US economy had been in recession for several months. But as late as August 2008, the most common response from financial markets was that of denial.

The meltdown began with the sudden nationalisation of the main US mortgage agencies, Fannie Mae and Freddie Mac in early September 2008. Two months later, the investment banking industry had collapsed, with Lehman Brothers bankrupt, Merrill Lynch swallowed by Bank of America, and Goldman Sachs and JP Morgan forced to seek the safety of government regulation. A year later, the list of casualties included banks around the world, whole countries such as Iceland, and the archetypal embodiment of corporate capitalism, General Motors. While a gradual recovery looks likely at the time of writing (June 2009), many more failures and defaults are inevitable.

In some ways, the boom and bust of  Every bubble in history has come with a story to show why “this time it’s different”. But the current crisis has two features that should spell the end of the efficient markets hypothesis once and for all. The first is that, in scale and scope, it is larger than any financial failure since the Great Depression. The estimated losses from financial failures amount to $4 trillion or about 10 per cent of the world’s annual income. Losses in output from the global recession are also likely to be in the trillions before the world economy recovers.

And, unlike the Great Depression, this crisis was entirely the product of financial markets. There was nothing like the postwar turmoil of the 1920s, the struggles over gold convertibility and reparations, or the Smoot-Hawley tariff, all of which have shared the blame for the Great Depression. Financial markets and major banks were lightly regulated by governments under systems that relied, in large measure, on risk assessments undertaken by the banks themselves, and based, in large measure on the ratings issued by agencies such as Standard & Poors and Moodys. 

All of the checks and balances in the system failed comprehensively. The ratings agencies offered AAA ratings to assets that turned out to be worthless, on the basis of models that assumed … Clearly, this was not simple incompetence. The entire ratings agency model, in which issuers pay for ratings, proved to be fundamentally unsound. But, these very ratings were embedded in official systems of regulation. Thanks to the EMH, crucial public policy decisions were, in effect, outsourced to for-profit firms that had a strong incentive to get the answers wrong. 

To these systemic failures was added the exposure of long-running fraud on a massive scale. The Ponzi scheme operated by Bernie Madoff, former head of the NASDAQ exchange and leading light of the New York financial sector, took place on a scale that matched the gargantuan growth of the financial sector itself. The original Ponzi scheme, promoted by Charles Ponzi on bogus investments in postal coupons 1920, brought in an amount equal to $5 million in today’s value. Madoff estimated the proceeds of his racket at $50 billion – ten thousand times Ponzi’s taking. And while Madoff put others in the shade, the collapse of the bubble brought to light a string of frauds involving tens or hundreds of millions of dollars.

The cases of Madoff and other frauds brings to mind JK Galbraith’s idea of the bezzle. The bezzle is the amount of undetected corporate fraud. As a boom continues, and everyone does well, people realise they can siphon off money and use it to make even more money. If they are threatened with detection, the original amount stolen can be returned to the till, and thye are still ahead. But, in a crisis, this can’t be done and, in any case, outside accountants are all over the books. So, embezzlers are caught and the bezzle shrinks. It stays small in the early stages of recovery when most decisions are being made by the cautious types who survived the crisis. But as the boom continues, hungrier and less-risk averse types come to the fore and the bezzle begins to grow again.

Under a system where the financial sector grows out of proportion to the real economy, and where, by virtue of the EMH, values recorded in financial markets are taken to be real, however absurd they may seem, the bezzle grew to unprecedented magnitudes.