In an Uncertain World (49 page)

   

ANYONE WHO PARTICIPATES in financial markets—whether as an individual investor, a Wall Street bond trader, or a company CEO—has to make fundamental decisions about risk. At an individual level, such choices affect one's financial well-being and peace of mind. At the corporate level, they affect profitability levels. And at the broadest collective level, decisions about risk have potentially enormous economic consequences. In a world of immense global trading markets, the success of institutions, the liquidity and efficiency of markets, and the safety of our financial system depend to a great extent on how well the complex process of risk determination is carried out by the vast numbers of participants—individuals and businesses—involved in the financial markets.

Many people arrive at accepting some level of market or trading risk through instinct, aversion, or feel. My approach, by contrast, has always been to try to make risk decisions on an analytic basis—even if they involve judgments about such intangibles as how much one can handle emotionally and the less-than-rational actions of others in the markets.

In making any decision about risk, the logical first step is to try to determine at what point additional risk no longer carries potential rewards that exceed the potential losses, given the respective probabilities of the good and the bad. The actual measurement of these probabilities, of course, is immensely complicated. But this calculation, which can be organized on an expected-value table, remains the most fundamental basis for decisions about risk.

However, the problem quickly grows more complicated. As an illustration, consider the choice a major financial institution such as Citigroup faces in choosing an optimal level of trading risk—that is, the level of risk incurred by the firm's own traders trying to maximize return on some portion of the firm's own capital. That is a far more complex question than it might seem to be at first, with many dimensions that can't be quantified on an expected-value table.

To begin with, a public company such as Citigroup is very different from the private firm Goldman Sachs was when I was there. A public company's stock price is a function of its earnings and the multiple the market decides to put on those earnings. And financial markets value stable earnings growth more highly than a volatile earnings path, even if the total profits at some projected end point are the same. So a public company has to determine both what will maximize its earnings over time—adjusting that calculation for risk—and also the effect of greater volatility on the multiple. Only after making these judgments can one make an educated guess about what level of risk will maximize a company's share price over time.

At a large financial company today, none of this is easy. In part, the difficulty arises because of the immense size and complexity of a firm's positions, which typically include stocks, bonds, derivatives, and currencies. In part, it arises because of the intrinsic difficulty of deciding which risks correlate with one another and which are likely to offset one another. And in part, it is due to the inherent uncertainty in trying to estimate rewards, risks, and probabilities. It is also extremely difficult to get managers to be ruthlessly analytic and to put their emotions and opinions aside.

Nor is that the end of the road with respect to risk management. Maximizing the expected value of results over time is a critical but not sufficient guide to decision making. Even when the expected value for additional risk is positive, every trader or investor—from the smallest individual to the largest commercial or investment bank—must decide, as a separate matter, how much loss he or it can tolerate, financially and psychologically. As an additional complication, there is one type of financial risk, the risk of remote contingencies—which, if they occur, can be devastating—that market participants of all kinds almost always systematically underestimate. The list of firms and individuals who have gone broke by failing to focus on remote risks is a long one. Even people who think probabilistically, and are highly analytical and systematic, often dismiss remote contingencies as irrelevant.

In this regard, I often think of an example from Goldman Sachs when Jon Corzine, then the firm's exceedingly successful head of fixed-income activities, wanted to take a large position in farm credit bonds. The expectation was of a high return, and because the bonds were backed by the “moral obligation” of the U.S. government through a new agency known as Farmer Mac, the probability of their defaulting seemed close to zero. But what Steve Friedman and I asked Jon was “What if a problem develops in farm credit and as extremely unlikely as it might be, the government declines to stand by its so-called moral obligation?”

“That's silly,” Corzine replied. It was inconceivable to him that the government would not honor its moral obligation, and in a sense he was right.

But Steve and I didn't want Goldman Sachs to cease to exist after 130 years because something that we agreed was virtually inconceivable actually happened. In theory, you don't ever want to be in a position where even a remote risk can hurt you beyond a certain point—and you have to decide what that point is. Too often, risks that seem remote are treated as essentially nonexistent. In this case, the remote contingency never occurred, but the decision to limit the risk was right.

As a practical matter, if you want to be in the business of trading or want to invest, all sorts of remote contingencies have to be set aside, from systemic financial collapse to catastrophic meteorites to nuclear war. We once explored this idea at Goldman and concluded that if we really wanted to take into account every catastrophic contingency, we couldn't be in business. But even if you can't avoid all distant risks in practice, it's sensible to think explicitly about which ones you're choosing to take and which should be diminished or avoided.

   

WHEN I ARRIVED at Citigroup, it had of course done much risk analysis, and Sandy Weill and his team had reached judgments suited to the institution. But from time to time, senior executives revisited the issue. After becoming Citigroup's president in 2002, Bob Willumstad raised a risk-related question not just about trading but about our allocation of capital overall. We had more than $80 billion in equity capital employed in our various businesses. Were our assets allocated in the most effective way possible? Perhaps a bit more should be in Brazil and a bit less in credit cards, or vice versa. Willumstad pointed out that at Citi, as at any company, all kinds of operating assumptions had built up over time, none of which was necessarily right. Some additional analytical work on risks and returns could challenge a lot of the conventional thinking, perhaps causing us to reduce some activities and increase others.

This kicked off an even broader discussion of how to evaluate conventional wisdom inside a company. I remember one meeting in my office that included Hamid Biglari, an Iranian-born Ph.D. in nuclear physics who was then our head of corporate strategy; Chuck Prince, then Citi's chief administrative officer; Kim Schoenholtz, our chief economist; and Lewis Alexander, our chief emerging-market economist. My response to Willumstad was that any institution will always have immense resistance to conclusions that differ from the conventional wisdom. Any conclusion that pointed toward significant change would inevitably be attacked. People with a vested interest in the status quo would find flaws in the study's methodology, in the way risks were calculated, in indirect effects and factors that hadn't been considered. Under that kind of attack, a program for change can easily dissolve into an endless, irresolvable debate.

Moreover, the people who raised these objections to analytic conclusions pointing to change wouldn't necessarily be wrong. Analytic rigor was a critical starting point for the discussion. But financial analysis could never encompass all of the relevant considerations. There were factors in some of our businesses that couldn't be quantified but that could have great economic impact, including synergies between the businesses, impact on our image, and so on.

Not long after I became a partner at Goldman in the early 1970s, the late Hyman Weinberg, our capable CFO, did a business-by-business profit and loss statement. That would now seem rudimentary, but we'd never done one because some people felt it might cause discord in the partnership. Hy found that investment banking—the business Goldman was best known for—wasn't actually profitable for the period he was looking at. John Whitehead, the highly respected head of investment banking at the time, responded that he knew that investment banking was profitable. If Hy's methodology didn't show that result, the methodology must be flawed.

Oddly enough, both Hy and John may have been correct. Investment banking is people-intensive, and the people tend to be extremely well paid. Thus at times good profit margins can be hard to achieve, although, as in Goldman Sachs's case, the business can be highly profitable over long periods, when conditions are at least moderately favorable. But even if Hy was right about investment banking not being profitable for the period he was looking at, John may also have been right in the sense that investment banking probably made a substantial nonquantifiable contribution to profitability by enhancing the prestige and standing of the firm, thereby attracting clients in other areas.

The right conclusion, I think, is that ongoing change is critical to success, and even to survival, but that ultimately challenging conventional wisdom is a matter of judgment. For example, measuring risk-adjusted returns on capital, as Bob Willumstad originally suggested, can be very useful as a tool to inform judgments. But a great danger for any institution is that such tools can take on a bureaucratic life of their own—especially as they come to be viewed as more sophisticated and “scientific”—and begin to drive decisions formulaically rather than contributing as inputs to broader judgments.

   

EVEN AFTER THE MARKET had started to fall, the tendency of otherwise intelligent and thoughtful people not to think clearly about stocks and investing continued to strike me. I remember when a shrewd venture capitalist came to my office and gave me a book. “These people have a system that works,” he said.

“Well, you know, I'm pretty skeptical about systems,” I responded.

And my friend said, “You read this, and you'll see. These people have done it.”

The book covered a period from about 1995 to 1999, during which the stock market had essentially gone straight up except for a severe downdraft in 1998. Once I saw the four-year track record, I didn't bother reading to see what the theory was. During those years, all you had to do was buy stocks to do extremely well. I threw the book out because I was afraid someone might see it on my desk.

During the strong market of the 1990s, most investors who rode the wave ignored traditional ideas about valuation. Some money managers remained invested on the basis of a practical calculation: “If the market continues to rise and I'm not participating, I'll lose my job. But if it falls dramatically, I'll be in the same situation as everyone else.” Others were conscious market cynics who thought they could successfully exploit the foolishness of others. Momentum investors didn't need an opinion about valuation. They were consciously saying, “The market may be overvalued—we don't know and we don't care. All we know is, it's been going up, and we're going to invest as long as it does—and get off the train before everyone else.” The problem lies in executing the greater-fool theory. If you get off every time the market ticks down and then reestablish your position when the market starts to go up again, you're going to get killed, because even rising markets fluctuate on the way up. And if you wait, you risk going down with everyone else.

With the stock market decline after 2000, many of the new-paradigm theories lost credibility. That should have brought renewed attention to the failure of most “active” fund managers—people who pick individual stocks rather than passively investing in an index—to outperform the market over the long run. Over long periods, the S&P 500 index has done better than the preponderance of active managers. Few mutual funds, hedge funds, or money managers consistently beat the indexes over a decade, let alone several decades.

When I express these kinds of views in speeches or interviews, I am often asked, “How should people with limited resources think about investing?” I'm not in the business of giving financial advice, but I have learned something about how to think about these matters. My guidance is fairly conventional and hardly exciting. But however commonplace it may seem, people tend not to follow it, especially when the adrenaline starts rushing.

The most important part of my answer is that investors should recognize the risks they're taking. Stocks outperformed bonds for every decade of the twentieth century, except for the 1930s and a roughly equal performance in the 1970s. But there is no guarantee that the future will replicate the past or, at the least, that there won't be long periods of poor or mediocre stock market performance. And even if stocks as a whole continue to outperform bonds over the long run, many individual stocks will still perform badly. You might simply pick the wrong stocks and as a result significantly underperform the market—just as many professional money managers do.

And even if the historical outperformance of equities holds true and you do not pick the wrong stock or group of stocks, there may still be extended periods of adverse market performance. You, as an investor, simply may not be able to stand a wait of five to ten years or even longer, either financially or psychologically. Stocks that outperform bonds in the long run may also underperform them badly for lengthy periods within that “long run.” Over the seventeen-year stretch from the last day of 1964 to the last day of 1981, the Dow closed at the same level—874 in 1964 versus 875 in 1981. While the S&P 500 did rise by 45 percent during the same period, that is still only roughly 2½ percent per annum (plus the average dividend yield of roughly 4 percent). And after several years of difficulty, you have no way of knowing if the markets will improve or continue to deteriorate. Investors who want to retire or send their children to college may discover that they can't continue to wait for the hoped-for benefits of the long term, while other investors may simply not be able to live with the stress of large loans and the uncertainty of what might happen next. In reality, people often can't tolerate an extended downturn, so they get out and take their losses, only to reinvest when the market rises. When it falls again, they get whipsawed.