Beyond Greed and Fear

by Hersh Shefrin

  Indonesian business people extrapolated a trend and then speculated on it. Most were surprised when they ended up with serious losses. In other words, these investors were overconfident as well.

  Opaque Framing

  Were there warning signs about the deteriorating situation in the Asian economies that were ignored by overconfident investors? Economist Paul Krugman argues that this was in fact the case. In 1994, he wrote an article in Foreign Affairs that generated a major controversy. The article summarized work by economists Lawrence Lau of Stanford University and Alwyn Young of Boston University, who suggested that the sources of growth in the Asian economies were declining. Consequently, economic growth in these countries was likely to slow.

  These economists argued that the Asian economic miracle was not a product of an “Asian system” where farsighted Asian governments promoted specific industries and technologies. Rather, it was the result of more run-of-the-mill changes such as high savings rates, good education, and the movement of underemployed peasants into the modern sector. If these arguments are true, then growth will slow when, for example, few underemployed peasants are left to move into the modern sector. This feature is truer of an economy like Singapore’s than of one like China’s. So, in short, Lau and Young, and then Krugman, argued that investors were looking at Asian economies through an opaque frame, not a transparent one.

  In the conventional economic view, investors gradually adjust their expectations about Asian economic growth as information about changing fundamentals appears. But the adjustment in 1997 was hardly gradual. Consider some hard questions. Had investors been unduly exuberant about investing in Asia, particularly in emerging markets? Did they have available to them information that would enable them to make informed decisions? Did they underreact to the economic statistics that were available? Or did they simply extrapolate past growth rates and succumb to the optimism bias? Did undue optimism lead to widespread investments in negative net present value projects all across Asia, projects that could not produce positive returns, let alone the kinds of returns investors had imagined? These are important questions. For a sense of the answers, consider the following.

  In a later article for Fortune magazine Krugman (1997) discussed investors’ sudden realization that they had been ignoring important statistics like the trade deficit. He stated: “By last year it became clear that South Korea’s and Thailand’s torrid growth rates of the first half of the 1990s had to end—wages were rising faster than productivity; overheated domestic markets were spilling over into imports, creating massive trade deficits. … And international capital markets have suddenly noticed that countries throughout the region have been running world-class trade deficits—bigger relative to their economies than Mexico’s was just before the big peso collapse.”

  Krugman’s article appeared in August 1997. Perhaps I myself am committing hindsight bias, but let me point out that two months later, managers of most mutual and pension funds were still oblivious. Then, in October, the financial markets in Hong Kong suffered a sudden jolt. Between October 20 and 23, Hong Kong’s stock market lost nearly a quarter of its value on fears concerning interest rates and pressures on the Hong Kong dollar. In the above cited Wall Street Journal article, Bill Sutton, who monitors Asian currency markets for Merrill Lynch, stated that “the vast majority of hedge funds, real-money funds and other institutional investors were fully invested in the rupiah.” Julian Robertson Jr.’s Tiger Management Corp. was a prominent example. The Wall Street Journal reported on Tiger Management’s experience in October as follows:

  Tiger Management decided to buy, figuring the rupiah, down to around 3,800 to the dollar, was due for a rebound. Tiger says it snapped up $975 million in rupiah in early October. This wasn’t a big sum for the hedge fund, but it was for the rupiah market, and it pushed up the currency—briefly.

  “Our people were very bullish on the rupiah,” Tiger’s Mr. Robertson says. But he is the boss, and “finally I just said, ‘I don’t want to be long any of this crap.’” Tiger held its position about three weeks and sold most of it in $150 million chunks at the end of October, plus the final piece at the end of November.3

  What began as a decline in Asian currencies eventually engulfed those of Latin America and Russia, not to mention Canada, illustrating the point I made earlier that excessive speculation—driven by psychology and panic—is a major issue. In the September 7, 1998, issue of Fortune, Krugman called for the controversial action of implementing exchange controls. Here is how he put it.

  In short, Asia is stuck: Its economies are dead in the water, but trying to do anything major to get them moving risks provoking another wave of capital flight and a worse crisis. In effect, the region’s economic policy has become hostage to skittish investors. Is there any way out? Yes, there is, but it is a solution so unfashionable, so stigmatized, that hardly anyone has dared suggest it. The unsayable words are “exchange controls.”

  “Skittish investors”? Time to consider the general findings about excessive speculation.

  The General Case

  Economists Kenneth Froot and Jeffrey Frankel (1993) have documented that excessive speculation occurs in normal times just as it does in extraordinary times. They focus on the determinants of forward discounts in foreign exchange markets.

  Consider the following example. On Monday, March 27, 1995, the Japanese yen was trading at 86.4 to the U.S. dollar. Speculators who wanted yen 30 days later, on Wednesday, April 26, could pursue one of the following two alternatives. They could exchange dollars for yen on March 27 at 86.4, take delivery of the yen a month before, and invest for the proceeds for one month at the short-term Japanese interest rate. Or, they could purchase the yen forward, at the thirty-day forward rate of 86.049. In the second case, the exchange of dollars for yen would take place on April 26, not March 27.

  Notice that the March 27 forward rate was less favorable than the corresponding spot rate, because a dollar bought fewer April 26 yen than it did March 27 yen. On a per yen basis, the forward dollar value of the yen was more than the dollar spot value. In other words, the yen was trading at a premium on March 27.

  What does this mean? For one thing, Japanese interest rates were lower than U.S. rates. On March 27, 1995, the thirty-day Japanese interest rate was only 1.72 percent, whereas the corresponding U.S. rate was 6.16 percent. If The yen did not trade at a premium, then investors could make large arbitrage profits very easily. They could borrow yen at the low Japanese rate, purchase dollars at the spot rate, and invest the dollars at the higher U.S. rate for thirty days. Of course, they would have to repay their yen loan on April 26. But if the yen were not trading at a premium on March 27, they could lock in a favorable rate by converting their dollars back into yen in April. Most important, the rate they locked in would enable them to earn a guaranteed profit by borrowing at a low interest rate and investing the proceeds of that loan at a higher interest rate.

  When the yen trades at a discount, does that mean that investors expect the yen to depreciate over the next thirty days? Perhaps, but not necessarily. Money Market Services (MMS) is one of the major firms that track investor expectations about currency rates. It reported that on March 27, 1995, investors expected that in thirty days time, the yen would move to 90.4 So, although the yen was trading at a premium, investors expected that it would depreciate over the next thirty days. In fact, the yen appreciated to 84 over that period.

  How frequently do investors’ expectations about exchange-rate movements run in the opposite direction from the forward discount? It happened 42 percent of the time during the period March 27, 1995, through January 5, 1998, in the thirty-day forward yen market.

  It definitely appears that investors are willing to bet against the direction implied by the forward discount. Is this sensible? It depends on how well the forward discount does as a predictor of future movements in the spot rate. If, over time, the spot rate tends to move by about the amount predicted by the forward discount, then
it would be foolish to bet against the forward discount. However, the truth is that during the period August 8, 1994, through January 5, 1998, the yen depreciated against the dollar despite the yen’s trading at a premium throughout this entire period. Figure 21-1 shows the behavior of the yen (per dollar).

  How would traders have made money speculating against the forward discount? When should they buy low, and when should they sell

  Figure 21-1 Time Series of Yen per U.S. Dollar

  How has the dollar value of the yen fluctuated between 1994 and 1998? high? If they expect the yen to depreciate over the next thirty days, then selling high means selling the yen forward today, since the forward yen trades at a premium. They could then plan to buy low by buying yen in thirty days time at the April 26 spot rate, whatever that turns out to be.

  How would it have turned out? Speculators could have sold yen at 86 on March 27, agreeing to deliver 86 yen in exchange for one dollar. Then on April 26, they would have expected to purchase yen for immediate delivery at a rate of 90, which would have been good news. Traders would expect only to have to spend $0.96 to purchase the 86 yen that they promised to deliver, and would receive $1.00 in return. So, they would expect to make a 4-cent profit per yen in completing their forward transaction. But the news on April 26 was bad, not good. The yen had fallen to 84. Hence, speculators would have had to spend $1.02, more than a dollar, to purchase the 86 yen they committed to deliver a month earlier, thereby losing 2 cents per yen instead of making 4 cents.

  Is betting against the forward discount speculative? If so, is it excessively speculative? To answer that question, Froot and Frankel point out that investors form their expectations about exchange-rate movements by considering the contemporaneous spot rate and a host of other variables. They ask whether we can tell if investors should place more or less weight on the spot rate.

  To answer the question, they suggest comparing the accuracy of investors’ predictions to their forecasts. They measure accuracy as the difference between the forecasted change in the exchange rate and the actual change. The point is that if investors make rational use of information, then accuracy at any date should be a matter of luck. On the other hand, suppose that investors tend to be too extreme in their predictions. Then their forecasts will not be rational. When they predict appreciation, they predict too much appreciation; when they predict depreciation, they predict too much depreciation. In short, investors overreact. In this case, we could improve their forecasts by toning down the predicted size of the changes being forecast.

  Froot and Frankel find that as a general matter, investors’ predictions about exchange-rate changes are too extreme. How about their yen forecasts during the period we have been discussing? The MMS forecasts fit the general pattern of excessive speculation.

  Moreover, forecast errors tend to be highly predictable. You can make a good guess about how far off investors’ predicted change in the exchange rate would be. This is because forecast errors tend to be

  Figure 21-2 Forecast Errors in Yen Appreciation

  The time pattern of MMS forecast errors concerning the value of the Japanese yen. An efficient forecast is not excessively volatile, and it fluctuates randomly around zero. But once positive, the yen forecasts tend to stay positive, and once negative they tend to stay negative. highly persistent. You can see this feature in figure 21-2. Notice that the prediction error tends to move in cycles. When the error is positive, it tends to stay positive for a while before turning negative. Once negative, it stays negative for a while.

  There is a long-standing question about whether the presence of a risk premium inhibits the forward discount from predicting future spot-currency rates. The risk premium is defined as the difference between the forward discount and the predicted change in the spot rate. In theory, the risk premium should explain some portion of the discount. Of course, when the yen trades at a premium, and investors predict that the currency will depreciate, as was the case on March 27, 1995, the risk premium exceeds the value of the discount.

  Froot and Frankel indicate that there may well be a nonzero risk premium. But they argue that the behavior of the risk premium is not the major impediment preventing the forward discount from serving as an efficient predictor of future spot rates. In fact, for the period under discussion, the correlation between the risk premium and the forward discount prediction error is very close to zero.

  Summary

  Foreign-exchange traders engage in excessive speculation, stemming from heuristic-driven bias. In particular, foreign-exchange traders seem to overreact, bet on trends, and be overconfident. Therefore, the foreign-exchange market shares similar features with the pick-a-number game described in chapter 1. Foreign-exchange rates are affected by both fundamental economic factors and by sentiment. This is yet one more instance where heuristic-driven bias leads to market inefficiency.

  Final Remarks

  Behavioral finance is everywhere that people make financial decisions. Psychology is hard to escape; it touches every corner of the financial landscape, and it’s important. Financial practitioners need to understand the impact that psychology has on them and on those around them. Practitioners ignore psychology at their peril.

  I began this book with an excerpt from a quotation about a “change in psychology” that reads: “Well, I was worried about the change in the attitude of investors, particularly portfolio managers. … So, it was a decided change in psychology. There’s no change in the fundamentals. I’m still bullish long term.”1

  You will have read the same point again, in the last chapter. There I included a quotation, invoking psychology, that read: “It isn’t just economics at work now; it’s a psychology that at times borders on panic.”2

  The two quotes pertain to different financial settings, one the market for U.S. equities and the other the market for foreign exchange. The quotations, coming from the first and last chapter, symbolically make the point that psychology is everywhere. Moreover, both quotations clearly indicate that psychology matters. What the authors of these and similar quotations are less clear about is what psychology means.

  The goal of this book has been to explain the current state of thinking about what psychology means for finance. To accomplish this task, I organized the subject matter of behavioral finance into three broad themes: heuristic-driven bias, frame dependence, and inefficient markets. For the moment, let’s focus on the first two. Heuristic-driven bias concerns the biases to which people are prone because they rely on heuristic rules of thumb. Frame dependence concerns issues of form and substance: People are very sensitive to issues of form, and so form can become substantive.

  I wrote this book about practitioner psychology for a practitioner audience. Practitioners who read the book should learn to recognize heuristic-driven bias and frame dependence in themselves and in others. Both heuristic-driven bias and frame dependence predispose practitioners to commit specific errors. Recognizing behavior is the first step towards modifying behavior.

  A place for final remarks is no place to recapitulate details about gambler’s fallacy, loss aversion, or self-control. However, when it comes to the third theme, inefficient markets, I do want to reiterate a word about overconfidence.

  An inefficient market means that price differs from fundamental value. The evidence that markets are inefficient is, by now, compelling. But behavioral finance emphasizes that smart-money investors, who are free from overconfidence, do not try to exploit every mispricing opportunity that they see. The smart-money investor:

  • distinguishes luck from skill;

  • recognizes that the mistakes of other traders produce an extra source of risk as well as a potential profit opportunity; and

  • knows that only some risks are worth taking.

  Overconfident investors who know only a little of behavioral finance can do themselves great harm trying to exploit market inefficiencies.

  So, to the overconfident, and that means most of us:

  Forewa
rned is forearmed.

  Notes

  Chapter 1

  1. This occurred on the August 18, 1998, program Wall $treet Week with Louis Rukeyser.

  2. A synonym for traditional finance is “standard finance.” Meir Statman (1995a) wrote an intriguing 1995 article that he titled “Behavioral Finance vs. Standard Finance.”

  3. Proponents of traditional finance accept minor violations of error-free beliefs and frame independence, but argue that errors cancel out at the level of the market. For example, the errors by investors who are overly bullish get offset by the errors of investors who are excessively bearish. So the market finds just the right middle ground.

  4. “Financial Planning: Win a Flight to the U.S.” Financial Times, May 10, 1997.

  5. Zero is also an equilibrium choice in this model.

  6. There were 1,468 entries in this contest, and the average response was 18.91. There were 31 entries with the winning guess of 13.

  7. Scholes has since left LTCM.

  8. LTCM’s returns were 19.9 percent in 1994, 42.8 percent in 1995, 40.8 percent in 1996, and 17.1 percent in 1997.

  9. Royal Dutch shares are included in the S&P 500, and Shell is included in the FTSE (Financial Times Allshare Index). Notably, Royal Dutch share prices appear to be driven by the factors that drive the S&P 500, while Shell prices are driven by the factors that drive the FTSE 100.

  10. Robert C. Merton, the Nobel laureate, is the son of Robert K. Merton, the well-known behaviorist whose 1948 work constitutes an important contribution to the understanding of self-fulfilling prophecies.