You would never think of Harry Markowitz as a bomb-throwing revolutionary, but 50 years ago he tossed an explosive into the investment world that is still smoldering today.
Mr. Markowitz's radical ideas turned investment theory inside out, although it took more than 20 years for the institutional investment world to start catching on. In 1990, Mr. Markowitz was awarded a Nobel Memorial Prize in Economic Sciences for his work, along with William F. Sharpe and Merton Miller.
Today, virtually everyone in institutional investment management has been touched by Mr. Markowitz's groundbreaking insights that investment returns are tied to risks, and by subsequent developments in modern financial theory, ranging from the efficient-market theory to options-pricing models. Taken together, the body of work known as Modern Portfolio Theory applied mathematical techniques to the traditional clubby world of investing.
"It is portfolio theory. The `modern' part is redundant," said Eugene F. Fama, the Robert R. McCormick Distinguished Service Professor of Finance at the University of Chicago's Graduate School of Business.
Modern Portfolio Theory "is a term I've always hated," Mr. Sharpe said. "I've always told Harry that he must be an ancient portfolio theorist ... I've heard `post-modern.' I'm waiting for deconstructionist."
Whatever you call it, MPT has brought physicists and mathematicians into the heart of investment management. It has transformed the way portfolios are managed, and how pension executives assemble various investment strategies into a coherent whole. And it has raised -- although not answered -- fundamental questions about active management, including whether anybody can pick managers who can consistently beat their benchmarks.
MPT developments
"The idea of Modern Portfolio Theory just permeates everything in the pension industry," said Grant Gardner, director of research at Frank Russell Co., Tacoma, Wash.
Observed Rob Arnott, managing partner at First Quadrant Corp., Pasadena, Calif.: "All of this spawned modern finance. It's interesting to think, even in recent history looking back 25 or 30 years ago, that people who were viewed as quants, as people on the edges of finance theory, would today be viewed as very mainstream."
Among the key developments stemming from MPT:
* Portfolio construction. Mr. Markowitz's insight that the relative riskiness of an individual stock should be looked at in context of the entire portfolio has altered how portfolios are built. Diversification, which has become the byword of portfolio construction, can lower the volatility of the entire portfolio.
* Risk controls. Almost all active managers now embrace some quantitative tools, whether top-down managers using tools to shape the broad allocations within the portfolio, bottom-up stock pickers ensuring the portfolio is adequately diversified by industry or sector, or managers ensuring that a portfolio stays within a defined risk from its benchmark.
* Pension fund asset allocation. Diversification achieved by combining different asset classes has enabled pension funds to move from the typical 1950s pension portfolio of purely domestic stocks and bonds to one that now also includes international stocks, real estate, private equities and hedge funds, as well as combining managers of different styles and risk characteristics. Newfangled risk budgeting techniques also are a direct outgrowth of MPT.
* Equity valuation. The Capital Asset Pricing Model, developed independently by Mr. Sharpe, Jan Mossin and John Lintner, said the stock market as a whole is the single biggest factor in determining the value of individual stocks, changing how stocks are valued.
* Index funds. CAPM's key assumption that the stock market is the most efficient portfolio led to Mr. Fama's development of the efficient market hypothesis, and subsequently to the development of index funds. Today, indexing accounts for 38% of U.S. institutional assets invested in domestic equities, or about $1.1 trillion, according to Greenwich Associates, Greenwich, Conn.
* Quantitative strategies. Other quantitative techniques, such as tactical asset allocation and market-neutral strategies, emerged from MPT's risk-oriented approach
* Performance measurement. Risk measurement tools, such as beta and Sharpe ratios, have changed the way managers' investment performance is measured.
"In the bad old days, people said they have 5% of the portfolio in General Electric," said Mr. Sharpe. "Now, institutional managers say they have 5% in GE vs. (GE's weighting of) 4% in the S&P 500. When managers talk about performance, they say not just they have done 12% but have done 12% (and) the S&P did 14%," he explained. Comparing performance against the consensus, as reflected by a benchmark, "all comes from efficient market theory," Mr. Sharpe explained.
Diversification is key
Modern Portfolio Theory has transformed a sleepy industry run by bank trust departments and insurance companies into a dynamically changing business.
But all that was a distant vision when Mr. Markowitz first cast about for a dissertation topic in 1950 at the University of Chicago. Waiting in the anteroom of his academic adviser, Mr. Markowitz had a chance conversation with a broker who suggested applying mathematics to the stock market.
One afternoon in the library, reading John Burr Williams' "Theory of Investment Value," Mr. Markowitz realized Mr. Williams had ignored the impact of risk in making investments. If an investor were not concerned about risk, an investor would pick the one stock he thought would produce the maximum return, Mr. Markowitz said.
"At this moment of discovery, when I saw these correlations pop out of the page, it seemed to me to be common knowledge," he said in an interview.
But Mr. Markowitz's major breakthrough was recognizing that risk must be measured not in terms of each individual security, but how the risk of each security related to every other security in the entire portfolio. That measure, called covariance, shows how much risk, or variance, an individual security adds to the entire portfolio.
Mr. Markowitz published these findings in the March 1952 issue of the Journal of Finance in his seminal paper, "Portfolio Selection."
While this idea may seem obvious to anybody who has been through an MBA program in the past 25 years, the fact that Mr. Markowitz provided a mathematical framework to relate the relative riskiness of an individual stock within the entire portfolio was a revolutionary concept.
"Where did he get this?" said Mark Rubinstein, the Paul Stephens Professor of Applied Investment Analysis at the University of California's Haas School of Business in Berkeley, and a co-founder of Leland O'Brien Rubinstein, the portfolio insurance firm.
Mr. Rubinstein said many subsequent financial theory developments, including CAPM, led out of Mr. Markowitz's work. But as Mr. Rubinstein recently explained when introducing Mr. Markowitz to a Berkeley School of Finance program, Mr. Markowitz's findings "seemed to come out of nowhere."
"Prior to Harry Markowitz, there was a very loose sense that diversification was useful but no real understanding of the cost and benefit of diversification," said Andrew Lo, Harris & Harris Group professor at the Massachusetts Institute of Technology, director of the MIT Laboratory for Financial Engineering, andchief scientific officer of AlphaSimplex Group LLC, Cambridge, Mass. "Since then, portfolio theory and optimization are part of every portfolio."
Barr Rosenberg, a pioneer in developing mathematical tools to enhance stock selection and optimal portfolio mix who now is chairman of AXA Rosenberg Group, Orinda, Calif., said diversification was understood intuitively but could not be quantified. Typically, a portfolio manager would be required to own the stock of a large company from each industry group, he said.
The agenda is set
While Mr. Markowitz's original 14-page paper began the modern era in finance, it was his 1959 book that truly set the stage for the rest of the century. Even today, many investment professionals and academics do not realize how visionary the book was.
In the book, also titled "Portfolio Selection," Mr. Markowitz suggested replacing variance as a measure of risk with semivariance, which measures downside risk. He compares several alternative measures of risk,according to Mr. Rubinstein: standard deviation, semivariance, expected value of loss, expected absolute deviation, probability of loss, and maximum loss -- foreshadowing current debates on the correct measure of risk.
Mr. Markowitz also presented a method for figuring how to optimize returns over an individual's lifetime -- another idea 40 years ahead of its time. This so-called multiperiod approach now is the focus of research of many economists, involving complex calculations concerning taxes and the rise and fall in investor wealth.
He further developed the notion of using a mean-variance optimizer, requiring estimation of expected returns, variances and covariances. This approach remains the basic method for creating portfolios to this day.
Unfortunately, Mr. Markowitz's method of comparing the relative riskiness of hundreds or thousands of securities was time-consuming and expensive when computers were in their infancy. Calculating the covariances of 2,000 securities would require more than 2 million calculations, noted Peter L. Bernstein in his book "Capital Ideas."
But Mr. Markowitz outlined how his complex mean-variance approach could be simplified for the 1960s by relating stocks to a single factor. Mr. Markowitz passed that idea to Mr. Sharpe, then a graduate student working under Mr. Markowitz's guidance.
The birth of CAPM
Mr. Sharpe took that idea and ran with it. In his 1963 paper, "A Simplified Model for Portfolio Analysis," the young graduate student posited that a stock's price is related to a single factor, or index. That factor is the stock market itself, accounting for about one-third of any stock's movement. In a portfolio of as few as a dozen stocks, Mr. Bernstein wrote, "more than 90% of the portfolio's variability is explained by the index."
A year later, Mr. Sharpe advanced this concept into the Capital Asset Pricing Model, or CAPM. Mr. Sharpe determined the "super-efficient" portfolio defined by the late Yale University professor James Tobin -- the one that, when combined with cash equivalents, would provide the best choice for any investor -- was the stock market itself. (Academics Jan Mossin and John Lintner made similar determinations around the same time, although Mr. Sharpe beat them into print. Some experts now refer to the Sharpe-Mossin-Lintner model.)
By relating the riskiness of a stock to the entire market instead of every other stock in the universe, CAPM's single-factor model made the computations feasible, although still expensive by today's standards.
What Mr. Sharpe also did was introduce a measure of "systematic risk" -- that is, the risk of the market as a whole. This risk measure became known as beta. Every stock's beta related its risk level relative to the market, which was given a beta of one. Thus, a stock with a beta of 1.1 would be 10% more variable than the market as a whole; a stock with a beta of 0.9 would be 10% less variable. "The concept that a higher beta stock should offer a higher return was a radical insight," Mr. Rosenberg said.
Efficient market theory
As an economist, Mr. Sharpe placed his capital asset pricing model in the context of a market equilibrium. That is, he said that prices of riskier assets would go down as demand declined, and prices of less risky assets would go up, or vice versa, depending on investors' appetite for risk. In such an equilibrium, all stocks would be correctly priced.
Mr. Fama took the next step. In his dissertation, which was published in January 1965, Mr. Fama asserted the market is efficient: it reflects new information so quickly the market "knows" more than any individual can. The bottom line is it is very hard for the average investor, even with the best available information, to beat the market.
Performance studies have borne out this contention. In a groundbreaking study of mutual-fund performance published in 1969, Michael Jensen -- then a graduate student studying with Mr. Fama -- found the average mutual fund had drastically underperformed the market from 1955 to 1964.
Mr. Fama later backed off his contention that stock prices were unpredictable. "The initial idea was, if markets were efficient, then prices would be a random walk. That turns out not to be true," Mr. Fama said in an interview. He noted that if an equilibrium model included expected returns that vary with the business cycle, stock prices become predictable.
The question remains, however, is anybody savvy enough to outsmart the market?
Mr. Fama says the answer is yes. "My hero is Warren Buffett. He's the king of investing. He says he can pick a (good) company every few years... If the best can pick a company only every few years, what should the rest of them be doing?"
Buffett's blasts
However, Mr. Buffett, the legendary chairman of Berkshire Hathaway Inc., Omaha, Neb., does not have kind words for the academics. In Berkshire Hathaway's 1988 annual report, Mr. Buffett roundly condemned efficient-market theorists, noting his own track record had doubled the markets' average 10% return from 1956 through 1988. "That strikes us as a statistically-significant differential that might, conceivably, arouse one's curiosity," he wrote in the report.
He added: "Naturally the disservice done students and gullible investment professionals who have swallowed EMT has been an extraordinary service to us and other followers of (Benjamin) Graham. In any sort of a contest -- financial, mental, or physical -- it's an enormous advantage to have opponents who have been taught that it's useless to even try. From a selfish point of view, Grahamites should probably endow chairs to ensure the perpetual teaching of EMT."
Nor does Mr. Buffett spare the lash for followers of Messrs. Markowitz and Sharpe. In Berkshire Hathaway's 1993 annual report, he wrote: "The strategy we've adopted precludes our following standard diversification dogma. ... We believe that a policy of portfolio concentration may well decrease risk if it raises, as it should, both the intensity with which an investor thinks about a business and the comfort-level he must feel with its economic characteristics before buying into it."
He further lambastes academic use of beta as a measure of a stock's volatility. "In their hunger for a single statistic to measure risk, however, they (academics) forget a fundamental principle: It is better to be approximately right than precisely wrong."
And Mr. Buffett is not alone. Here's a view on risk control from Oak Value Capital Management Inc., Durham, N.C.: "We think that the attempt to graft higher order mathematics onto investments simply because we can crunch the numbers is, frankly, rubbish, and we spend zero time considering it."
From theory to practice
For years, Modern Portfolio Theory languished on the library shelf. But a number of events brought the theory to the attention of institutional investors.
For one thing, pension executives had become aware that equities offered superior returns to bonds. Using data from the University of Chicago's Center for Research in Security Prices, professors James Lorie and Lawrence Fisher revealed in January 1964 that stocks returned 9% annually from Jan. 30, 1926, through 1960 -- despite the effect of the Great Depression.
Another issue was that early stabs at performance measurement helped pension executives realize how much their funds were lagging mutual fund performance.
In 1965, A.G. Becker Co., a Chicago brokerage firm, started building its institutional database that swiftly became the dominant industry source of performance data.
But the first real move into quantitatively managed investment strategies required two more influences to start capturing pension executives' attention: the 1973-'74 bear market and the development of the computer.
Reassessment period
In the 1960s, money managers ignored the new academic theory. By the mid-1970s, they were in peril if they did so.
What had happened was that many bank trust departments, realizing that mutual fund managers were beating the pants off them, started investing in hot Nifty Fifty stocks, such as Avon and Polaroid, also known as "single-decision" stocks. The rationale was these stocks were so good one could buy and hold them forever. When the market started crashing in 1973, these portfolios were hard hit. All in all, stocks lost 50% of their value in real terms.
"First, the private pension industry was embarrassed against the private mutual funds, and then embarrassed against the benchmarks," said William L. Fouse, chairman emeritus of Mellon Capital Management Corp., San Francisco.
"After '73-'74, there was a period of self-examination on the part of practitioners," said Dean LeBaron, founder of Batterymarch Financial Management, Boston, who now splits his time between Switzerland and New Hampshire.
Birth of the index fund
The main way quantitative techniques penetrated investment management was through index funds, a direct outgrowth of Mr. Sharpe's assumption of market efficiency and Mr. Fama's efficient market hypothesis.
The first index fund was created by Wells Fargo Bank, San Francisco, in 1970 for Samsonite Corp.'s pension fund. The problem was the index fund was invested on an equal-weighted basis in New York Stock Exchange stocks, racking up huge transaction fees from constant rebalancing.
The Stagecoach Fund, which invested in low-beta stocks and then was levered up to market risk, was launched the next year but was dropped in 1973 because of lack of client interest. "I threw ice water on it because I said the problem is beta doesn't capture everything," said Mr. Fouse, who had joined Wells Fargo in 1970 from Mellon Bank after his cutting-edge ideas were ignored by the Pittsburgh bank.
The Stagecoach Fund was replaced by the first fund that matched the Standard & Poor's 500 index; the Illinois Bell Telephone Co. pension fund was the first client.
Other managers were on the heels of Wells Fargo, including American National Bank and Trust Co., Chicago, and Batterymarch, which later exited the indexing business. Wells Fargo's management sciences unit subsequently was sold and renamed Barclays Global Investors, the world's biggest institutional money manager, and has remained at the forefront of quantitative investing.
Meanwhile, Mr. Rosenberg refined the notion of beta. By analyzing many financial variables for each company, such as size and financial condition, he was able to figure which factors contributed to beta. He was able to construct better risk measures -- "Barr's better betas" or "bionic betas." The fact that the betas could be explained through fundamental analysis made them more acceptable to Wall Street, noted Michael J. Clowes, editorial director of Pensions & Investments, in "The Money Flood."
Later, Mr. Rosenberg developed a product that looked at combining managers with different risk characteristics into the most efficient portfolio. He also developed tailored benchmarks for equity managers and performance attribution of managers.
The index fund revolution and quantifiable betas helped spread Modern Portfolio Theory to pension executives and money managers.
Optimize, young man
The use of optimizers in money management did not catch on until the late 1970s. "In 1977, when I started my career, I built an optimizer for The Boston Co.," said First Quadrant's Mr. Arnott. Using an IBM 470/145 mainframe computer, "It took 45 minutes to do a 45-asset optimization. Today, I can do a 1,000-asset optimization on my laptop computer in five to 10 seconds."
For many pension funds, it took a bit longer. Robert Shultz, who joined International Business Machines Corp.'s pension fund in 1980 after running the New York Telephone pension fund, said he knew Harry Markowitz was working at IBM's Yorktown Heights, N.Y., research facility. When various topics came up, he kept asking his boss, the assistant treasurer, "why don't we ask Harry Markowitz about that?" Mr. Shultz related.
"Someone got pissed off and said: `If you want to know what Harry Markowitz thinks, why don't you invite him in for lunch some time?"' Mr. Shultz explained. So IBM brought in Mr. Markowitz to discuss capital market theory. After the discussion was over, Mr. Markowitz "thanked us, came back in, and said `now, don't give up on active management."'
But that was no contradiction. When the radical notions of the new thinking became more widely disseminated, they came out in a jumble. Mr. Markowitz's efficient frontier became intertwined with efficient market theory. In reality, Mr. Markowitz never bought the efficient-market line.
The efficient-market theory "was not my hypothesis," Mr. Markowitz said. "Portfolio theory, as I presented it in my 1959 book, had to do with one investor acting rationally to achieve certain objectives." There is no assumption, he explained, "that anyone else is acting rationally."
Efficient-market theory aside, Mr. Markowitz's crowning achievement of developing the theory behind the means-variance optimizer remains undiminished. Observed Mark Kritzman, managing partner of Windham Capital Management Boston LLC, and research director of the Research Foundation of the Association for Investment Management and Research, Charlottesville, Va.: "The testament to Markowitz (is) here it is 50 years later, it's still the dominant method for solving most portfolio information problems."
