One reason for this structure was that there always would be someone on an investment committee who remembered an unpleasant experience in some industry that might under present-day conditions offer profit. The best way to circumvent this impediment in the late stages of a bull market was "to rent a kid"; that is, hire someone who had no preconceptions about particular investment alternatives.

It was in this setting that I assumed management of over $300 million of aggressive funds. I had an almost complete mandate to do whatever was necessary to achieve the desired result—to attain and maintain a position at the top of the list of competitive growth funds.

The new manager tends to concentrate quite heavily in what may have been the consensus before he assumed responsibility for the portfolio. In my own case, I have rarely felt that the institutional consensus in investment management is rewarding (consensus may often be right, but often it has been fully discounted in prices). I was attracted to unusually large positions in heavily institutionalized, low-multiple securities; the securities whose earnings had gone up much faster than price and those cyclical stocks not necessarily tied directly to the business cycle.

The largest position I instituted was in the airlines, approximately 17 percent of the total portfolio. Airlines are very volatile securities: not all of them fly in "friendly skies." They move in and out of institutional favor on wide swings in earnings. I was biased in favor of the group anyway as I had done some pioneering research on airline securities several years earlier. It was theoretically possible to forecast their earnings, hence their price, on the basis of equipment purchases. In the late sixties, however, equipment purchases slowed down and the principal issue was guessing at traffic growth. In this industry small increments in traffic produce large earnings gains. At this time the projection of traffic on its past trend of growth of twice the rate of the economy would have produced earnings gains in excess of 30 percent. For slightly different reasons, I also undertook excessively large nonconsensus positions in autos and aerospace. They each had a different reason for earnings gains but were also selling at an earnings multiple well below the market.

These groups were diversified according to economic forces but were undiversified relative to the market. In the bear down-trend of 1969-1970, each of these three groups substantially underperformed the market, producing painful results for any investor who was heavily committed.

Diversification

Diversification is a technique for minimizing the risks that accompany any securities venture. If a portfolio manager could look ahead and pinpoint the precise timing of a single event, he would naturally follow the right course. But since this is virtually impossible to do, he seeks to minimize the ever-present possibility of error by hedging his position through diversification. Diversification precludes putting all your investment eggs in "one basket." It should be pointed out, however, that too much diversification is worse than none at all. Just as a general of the army may invite defeat by spreading his available forces too thin, so may a portfolio manager become susceptible to error by overdiversifying the number of securities in an account.

Diversification is a frequently elusive factor for which academia has coined a new term. The investor should attempt to minimize the covariance (like the common forces on each security) among securities in his portfolio. Industry diversification isn’t enough; quality diversification is not helpful. In fact, the investor may have to rely upon the historical relationships between securities and assume that such relationships, or covariants, will remain in force.

One of the better forms of diversification may be geographic. A number of interesting studies suggest that diversification in different stock markets is useful and that there is a very low correlation or covariance among foreign markets. This work suggests that portfolios in the future will be highly internationalized and will achieve better diversification without reducing returns by cutting across national borders.

In the future, most investment portfolios in the United States will be internationalized. In many respects they already are, through the activities of large American corporations. Export trade and overseas earnings are approximately a third of most large U.S. companies. They happen to be the most profitable part of their activities, although that may be due to the method by which internal accounts are kept.

Traditionally, Americans have concentrated their portfolios away from the various international capital markets. This tendency will change. Now that the U.S. currency and the economy are more closely aligned with those of the Western world, we can expect that capital will be allocated around the world and portfolios will be balanced in different currency areas.

American portfolio managers have not been good international investors. They have not understood the currency problems and have not been aware of the government forces in each of the countries where they might invest. These forces are more important than in our own country in determining investment returns. We have always expected that funds placed overseas would be returned for ultimate reinvestment. A balancing of different stock markets provides diversification that may be difficult to achieve in any one area. Our portfolios will increasingly assume the aspects of European portfolios as they were constructed ten years ago. Trips through the banking fraternities in overseas countries are good preparation for the next decade.

I know of one around-the-world trip, conducted in ten days by a brokerage firm, designed to circumnavigate the globe with briefings held at the airports of major world capitals. Unfortunately, it was a far better study of first-class international air travel than of economic conditions in the countries involved.

Americans have taken a chauvinistic attitude toward investment markets other than their own. Now that we recognize that our problems are at least as great as those of the rest of the world, the America-first policy applied to portfolios should not remain a great barrier. In fact, just as American portfolios may place some of their capital outside this country, the investments in dollars held outside the U.S. by the Europeans, Japanese, and Arabs may well be converted to longer-term investment in the United States, thus balancing the dollar outflow from American institutional portfolios.

Another form of diversification is by investment type. There are three alternatives for most publicly traded portfolios: cash, fixed income, and equity. Markets for warrants, options, convertibles, venture capital, and so forth are either amply specialized or sufficiently narrow and are understood by investors for whom such possibilities are available.

Cash has emerged as a principal diversification tool. It formerly was used as a market hedge, in conjunction with forecasts of declining prices. But its application was limited, because the returns on cash or its equivalent, principally Treasury bills or Certificates of Deposit, were usually less than the long-term rising trend rate of common stock prices. In 1974, however, cash returns were on the order of 10 percent. At the same time a major bear market was in process. Thus, cash as a diversification tool seems to carry a lower price tag, particularly if an investor cannot conveniently dampen the volatility of a portfolio due to the instability of short-term equity risks. Cash may be the only way in which a portfolio can be conveniently dampened. Fixed-income alternatives, principally bonds, used to be the main offset to equities in a typically balanced portfolio. In fact, as bonds appreciated, equities declined and vice versa.

A number of market timing systems were developed in conjunction with the opposite moving forces between bonds and equities. Essentially, lower interest rates would be associated with a reduced demand for funds in a recessionary condition, or the government’s stimulus of business, by making money easier. Under such circumstances, bonds would appreciate in price while equities declined, reflecting more difficult business conditions or reduced earnings, accompanied by dividend cuts or omissions.

Historical studies have suggested that the true rate of return on bonds, adjusted for inflation, was something on the order of 3 to 3½ percent (some have computed figures as low as 1 percent), and that the actual rate might be more nearly the real rate, plus the rate of imbedded inflation. It could be assumed, in a period of declining business, that bond rates could come down, because of a slight adjustment in the expected real rate of return, perhaps from 3½ to 2½ percent, and that prices would moderate, reflecting the newer competitive environment of both labor and goods under recession conditions. Declining interest rates would raise bond prices, producing better results for the balanced portfolio which had a substantial portion of fixed-income obligations.

Diversification through fixed income became quite expensive during the twenty-five-year bull market which ended in 1966, especially if one of the market cycles was missed. It became increasingly apparent that long-term oriented funds, like pension and endowment funds, should have a much higher proportion of equities and not attempt to call a market turn. Failure caused by missing a cycle would produce an opportunity loss of 20 to 30 percent. This would seriously penalize the long-term result. Diversification turned out to be too expensive to utilize for a typical portfolio, which might have had bonds during the latter part of the sixties with low interest rates and high expectations for equity prices.

A recent diversification technique being used by institutional investors is to split fund managements. Ten years ago accounts were split to create a competitive position. It was assumed that managers would respond in line with their own commercial interests for competitive accounts. A large fund might be split into several parts and entrusted to several managers. Operating costs were higher, since most of the managers had progressively lower charges on the incrementally higher dollars. Motivation was expected to offset this operational limitation.

One of the principal purposes investment management clients have in splitting a fund apart is to obtain information on different investment styles. The problem of performance measurement is to determine the time over which it is relevant. Not only do performance measurements, like stock classifications, have to be adjusted for volatility levels, but one has to determine whether there is a minimum amount of time in which it is relevant, or is it indeed, as some say, one full market cycle.

Furthermore, is every market cycle like every other market cycle, so that perhaps one market cycle alone is not enough? There is a tendency to use performance measurements over far too short a period of time and even one full market cycle may be too short.

Recently diversification has been following more specialized paths. The equity man of the sixties was expected to be the Renaissance Man. He supposedly was able to operate in bonds, stocks, and in different types of market environments—"the compleat investor." Now it is assumed that the best investor is the narrow, specialized, and disciplined individual who has his own operating trademark. Perhaps at all times, even when a particular style is out of favor, investors will be able to differentiate between them and pick the principal proponent from each style category. It is assumed, then, that each will rotate eventually into favor at some point in the market cycle: This is known as diversification within equity styles. It may involve income stocks, venture capital, special situations, and even institutional consensus investing.

The emerging trend in the quest for diversification without penalizing results is the stress on mechanical strategies. Increasingly, no investor has a proprietary information edge over other investors. In fact, there are substantial legal prohibitions that suggest that buyer and seller must together have identical information about the security they propose to deal in. They may differ in judgment, but even that is rather difficult if they were similarly trained and similarly motivated and receive the information at approximately the same time. Accepting a notion of this perfect information world leads investors to operate with a series of mechanical rules. A set of rules allows him to operate against a pricing inefficiency that may exist for a very short time, so that he does not have to spend the conventional time of making an economic analysis or an on-the-spot judgment as to the motivation of management and its future plans.

Quantifiable rules (sometimes called "screens") provide for the operation of portfolios containing larger numbers of securities than have been true in the past. These rules achieve more controlled diversification and a better examination of the causal factors that produce results. Mechanical strategies suit the age of computers in investment decision-making and perhaps could only have evolved through the frustration of dealing with more conventional techniques.

I am quite convinced that, while many investors think they are adopting the conventional techniques of economic and corporate analysis, they really are employing a limited number of quantifiable rules that worked in the most recent market period. The intelligent investor would tend to downgrade these rules from the most recent market period, since that is probably the condition most firmly planted in an investor’s mind and discounted in the marketplace.

Diversification can be achieved without lowering performance. It may be a tool to achieve higher performance, adjusting for uncertainty, and may become adjunctive to the use of mechanical strategies when constructing portfolios over the next decade.

Security classification systems have received a continuous amount of attention. At Keystone Funds in Boston at least one person is kept busy working out new classification schemes, with which to compare the various elements of a portfolio against acceptable targets.

A classification system is designed to provide information about the way a portfolio is balanced. The normal portfolio breakdown and balance is directed more towards sources of net income than towards schemes that relate the holdings to industries or to economic effects like business cycles. If a portfolio is designed to be responsive to market forces, then the quantitative terms that relate to its breakdown and volatility characteristics should be most predictive of expectations under a predetermined set of future circumstances.

Portfolio breakdown techniques are often used in connection with different performance measures. Brokers in bull markets publicize an unweighted index of the market results derived from their research recommendations. Rarely do they publish such figures in a bear market. Mutual funds keep close track of how the competition is doing and whether they stand above or below average. Most mutual-fund organizations attempt to spread their funds sufficiently so that they will have at least one entrant in the upper categories irrespective of general market conditions. Most investment managers submit their performance data to measurement services like Becker Securities and Frank Russell & Co., so that they can use performance figures to support their case if they are favorable or ignore them if they are unfavorable.

Some investors use regular indices, like Standard & Poor’s industry breakdowns. Others devise their own in terms like defensive, cyclical, or growth. Still others use quantitative measures, such as separating securities according to volatility characteristics or price objectives under certain market hypotheses. Whichever classification is used, the investor must weigh the advantages and disadvantages of each.

The one very strong advantage of using a market index, such as provided by Standard & Poor’s or Dow Jones, for performance measurements is that their investment selections are available at low cost. For most investors they most nearly approximate a randomization of the market and can readily be duplicated with some sampling techniques. Once one elects to duplicate a market index, the proprietary work of the firm elected would be to minimize client expenses.

Expense control would insure a closer relationship between theoretical returns and actual returns than might otherwise be the case. What has been called index-matching is really the first of a series of mechanical strategies. Index-matching, which has substantial advantages for very large portfolios, is like stripping the investment business of its mystique and recognizing that the emperor is walking down the street naked.

Index-matching offers consistency by definition, since it will closely approximate the performance measure established for it by the market index. It is economical since most costs can be predetermined. It should, over a time, compare favorably for most large funds against an actively managed account. More directly, it will encourage managers who have specialized skills to engage in those skills other than merely duplicating the work that can be done by mechanical strategy. It will return the artists to his art, rather than have him spend his time on making frames.

A number of portfolios have been strikingly similar to an index portfolio without admitting it. The large mutual fund, Massachusetts Investors Trust (MIT)., has a volatility of approximately 1.0 and nearly a zero alpha, suggesting that the fund approximates the market. There is every reason why it should. It is so large it virtually has to be at market risk at all times, but one can easily raise the question, why a large staff of portfolio managers and analysts are required to duplicate this task. One could borrow from the collective judgment and wisdom of the investment community by accepting the notion that the securities in which it must deal are efficiently priced.

If MIT’s objective is to do as well as the market, minus its expenses, which are lower than many funds, then it comes very close to meeting those objectives. It is an example of a modest objective with an investment strategy designed to have a high probability of meeting that objective. Since satisfactions of this nature are unique within the investment community, index-matching may be merely an extension of existing practice, rather than a new revolution.

Identifying Market Phases

Next to diversification techniques, the nearly impossible task of identifying the nature of market forces is critical in portfolio construction. The error rate will be high but not nearly high enough to make the effort worthless. I doubt if any investor can afford to disregard this second key consideration, especially at major turning points.

Many investors can do this systematically, but the present consensus is that one should not even try. Investors may take the attitude that they will participate in all markets or none at all, or they will attempt to find a system that suits almost all conditions. Unfortunately, the worst thing that one can do in a poor market is to employ technical tools that are designed for good markets. Relative strength, for example, is perhaps one of the more compellingly poor guides to use in a bear market. Many investors, particularly institutions, tend in a bad market to sell securities that have been behaving the worst and buy those that have been the best performers. This is a distinctly awkward tactic that oftentimes causes one to switch into a security that is on the verge of a new decline. The same tactic may be another expression of the commonly followed activity that involves moving up the quality scale as the market decline continues.

Most bad markets cascade, with the lowest quality issues dropping first. Sometimes this is reflected by the American Stock Exchange price index declining before the New York Stock Exchange Composite Index. Low-quality securities seem especially depressed relative to earnings at the peak of the market, whereas their earnings are probably increasing at a greater rate than the top-grade securities. They generally rank second or third in an industry, but have a greater rate of increase in new sales. The operating leverage for such companies may be slightly higher than for the companies of better quality.

Nearly everything imaginable tends to force one into poor quality companies at, or just after, a bull market tops out and the move in a southerly direction begins. To be loaded up with low-quality securities at the start of a bear market and to upgrade gradually as the decline progresses is one of the surest ways to invite a burial by subscription later on.

The correct method, assuming that one is not going to employ cash, is to shift the low-quality securities gradually to higher quality, and the more volatile to more defensive securities as the market progresses upward. As the market recedes, the quality of the securities can be gradually lowered until the lowest quality is achieved at what might be a bottom level.

There is virtually no way to pinpoint an exact market bottom, of which there are classical types. One is the emotional, high-volume release usually associated with climactic events. Under these circumstances, prices behave completely irrationally. Securities are sold regardless of price into thin and unreceptive markets. Margin calls are common. Purchases made under these forced conditions have a high probability of success.

The other type of classical bottom is called an "exhaustion" bottom. Here the market just gets tired from going down. There are neither buyers nor sellers at these low prices that are often not associated with any specific event. These two bottom characteristics are the most common, but not necessarily the only conditions under which a market turn may be made. I suspect that market valuations, like individual securities, are nearly always an intelligent assessment of current conditions. It may be only during occasional extreme circumstances that inefficiency sneaks in and only for short periods of less than a year. The half-life principle seems as appropriate for the general markets as for individual stocks.

We had a recent opportunity to observe an important market cycle reverse. I visited the financial district in New York on May 25, 1970—one day before the panic low of 628 on the Dow Industrials. At that time, the corner of Broad and Wall was guarded by 500 mounted police. Their function was unrelated to the financial community; there were riots between construction workers and pedestrians. But I came away with a sense of chagrin at seeing the nation’s financial center under police guard.

I felt this was a close approximation to the Rothschild dictum about "buying ‘rentes’ when there was blood in the streets." As it happened, 1970 was also the year of the Kent State incident, in which national-guard troops fired on rioting students. One could not pinpoint a reason to be bullish for a single event, but there were several emotional signals that usually characterize in important trend reversal.

In this particular case, there was also the convenience of large volume*—a further classical sign of a major market bottom. The market declined again and successfully tested its May 26 low in June on sharply reduced volume This further confirmed the classical definition of a major turnaround.

* N.Y.S.E. volume in May 1970 totaled 258,220,000 shares—averaging about 12.3 million shares daily in twenty-one trading sessions. The turnover in June averaged 10.3 million and 10.4 million in July. August was low-volume month for the year with 218.7 million shares traded.

My firm, Batterymarch, at that time was investing in high-quality, high-yield industries. We had especially large positions in oil and tobacco stocks. Our reason for holding them was not that they would produce maximum gain. Like others, we were mainly trying to find ways to hedge our investments in the event that the market continued to go down, in which case we would want to be in less volatile securities. These industries became market leaders of the late 1970 period. We were able to take quick profits and reinvest the proceeds in a more dynamic selection of companies—principally life, fire, and casualty insurance and banks.

In a normal sense, an investment in high-yield securities at the bottom would have represented rather neutral strategy. It just happened that this particular market did not behave in an entirely normal manner on the recovery. But it did lead us to conclude that high-yielding, relatively stable securities are quite likely to perform better than their more volatile counterparts in the early phase of a market rise. Investors may want to reduce their cash position and participate in a market turn, but they are not wholly confident about the situation. Therefore, they pick these hedged securities, which was our practice as well.

There is a tendency in a rising market to become price unconscious, to let one’s gains move on further than one thinks is reasonable and to buy quite aggressively before prices move out of sight. In the early stages of a bull market this is quite a satisfactory procedure. However, in the late stages, one runs a high risk of being swept along with the crowd. It is very difficult, with foresight, to identify the various stages of market development.

The classical form has changed many times, and just as greed tends to overtake rational behavior at market tops and fear sets in at the bottom, most investors adjust for the market condition that has just passed. It is my view that that is the one condition, because of this technique of discounting prices, which can be ignored.

It is difficult to imagine that anyone could not invest successfully during a bull market. The trouble is that some people credit their own brilliance to a stock’s ability to appreciate in price, and this makes them overconfident. They should realize that bull markets have their pitfalls, as do bear markets. Bull markets are but a prelude to the next inevitable declining phase, and investors’ switch over from a bearish to a bullish psychology should be accomplished gracefully and with the utmost care. There is never anything "easy" about the market, irrespective of the prevailing trend.

It seems natural now to pinpoint the May 1970 low as a clear signal of a market reversal but it was difficult at the time. Volume was high and prices fell emotionally for almost four weeks. Except for certain periods in 1973-1974, it was the longest panic decline since I entered the investment business. After the first week of heavy selling, it seemed clear to me that reversal was imminent, although the precise price level and timing were impossible to forecast. The main thing was that the decline that had started the previous year was over. The reservoirs of cash that remained uncommitted on the market sidelines should be reinvested immediately. But in what?

Volatile stocks are the most profitable to own in the very low foothills of a new bull market. Keystone Funds has a unique separation of fully invested equity funds in which the manager cannot build up cash positions and is, therefore, ready at the bottom in a fully invested portfolio of a defined quality type. The Keystone system, devised by a noninvestment man in the thirties, separates the common-stock universe into four groups—the so-called S-l through S-4. Each higher grade is roughly more volatile, or riskier, than the grade that precedes it. The most famous fund, S-4, is a selection of highly speculative securities which, although very painful to own in a declining market, responds very quickly during the early stages of a rise.

These funds provide many useful clues for determining the character of a market because the managers can affect only the selection of stocks and not the quality of the portfolio. Internally, Keystone has an excellent tool called the "contominant variable," a ratio of actual results against the results of a universe from which the manager can pick his stocks. This index is far more than a comparison with a market index; it keeps the organization up to date on the manager’s activities. There have been notable occasions when the character of the market favored one particular style of investing, such as one Keystone fund or another when the selectivity was down ("poor CV"), but was considered to be an excellent performance by the investment fraternity.

Although the usual type of gain from the bottom is in volatile securities, other types may do better, or at least better than might be expected. If yield is most desirable, high-income stocks might be the leaders, especially if they have other characteristics such as a low dividend payout with high probability of an increase. In the fifties asset-rich stocks were the leaders. It is far more important to identify individual stocks that are likely to pace the market upward than to know the industry to which they belong. Ideally, the single security most likely to show the greatest profit is one on the brink of bankruptcy when the market first begins to turn. With high leverage—assuming it is reorganized in time—it soon regains financial health and prosperity. No investor, to my knowledge, has the skill to accurately time such an event. If he happens to do so, such good fortune should be ascribed more to luck than to his own financial acumen.

Impartiality

The probabilities considered in portfolio optimization and game theory are used to determine a future course of action. The outcome of these decisions determines whether the result will bring rewards or penalties. The portfolio manager usually focuses attention on the one most likely outcome. He has a single favorite security, or industry, or market hypothesis, and one economic forecast. His tendency is to incorporate this set of outcomes into his investment thinking and to exclude all else from mind.

A good example of this procedure is that for many years there has been some possibility, however remote, of a worldwide depression and economic holocaust. Few investment managers consciously consider this very seriously. They are concerned that by doing so they would adopt such a dismal frame of mind that they would be unable to operate at all. However, when such a long-odds forecast becomes more probable, it looms as an enormous surprise and causes the portfolio manager, who should have been continually adjusting his portfolio for this small but perhaps increasing possibility, to be shocked and dismayed.

The place one goes to reach major portfolio construction decisions is especially important. Few good investment decisions are reached in the office due to the bias of former thought patterns. In mid-1968 the market was recovering in the expectation that we would move into a very exciting period in the seventies. We called them "The Soaring Seventies."

Considering that we had a very speculative market in 1967, investment opinions returned quickly to the idea that prosperous conditions were returning. I was running a $350 million growth fund at the time and was troubled by this universal acceptance of optimism. Finding that everyone held the same idea, I initiated a practice that I have followed many times since.

I left the office for several days to go away by myself to think. I sought people who were not directly involved in the day-to-day investment business. At that time, I went to New York to talk to businessmen, bankers, and just a few investment people. I returned to Boston convinced that investment managers were attempting to project their own optimism upon an already inflated market

I wrote a memo and circulated it throughout the organization, itemizing in points 1 through 8 my concern about the investment outlook and the economy. These thoughts were diametrically opposed to our own investment policy and the views of our economic department. An economist was assigned to provide a critique of my memo, which he did by expounding a strong positive case for the consensus.

Knowing that my fund was expected to be an appreciation fund, I decided to buy low-multiple, nonconsensus industries as protection against what I saw coming for the market rather than assuming a cash position. This hedged position helped somewhat in the first stage of the decline. After six months, however, I began looking for signs that the decline was over. Now others in the organization began to feel that the decline would continue perpetually, if only because the trend was well established. For my part, I knew that market declines had rarely lasted more than nine months since World War II, and that my fund, having taken a defensive position initially, should become more aggressive. My technique in doing so was to increase the weighting in those defensive areas of low-multiple issues that had held up so well.