"Finance's Wrong Turn: A Revisionist View of Modern Finance in Practice"

Beginning in March 2018, New Frontier President and CEO, Dr. Richard Michaud, will partner with CFA Society Boston to teach a 5-week course that addresses the crisis in modern finance and quantitative asset management. 

The fundamental conundrum of modern finance is the lack of credible evidence that any investment strategy can provide risk-adjusted, cost-effective, superior performance relative to an appropriate index or investment objective over a sustainable horizon.  Is this ineffectiveness due to limitations of financial theory, investment technology, or fundamental capital market conceptual errors?

The course will argue that it is, in fact, due to all three. In five lectures it will describe the foundation of modern finance, applications to 20th century quantitative investment practice, the fundamental flaws of 20th century financial theory, a 21st century statistical update of quantitative investment technology, and a wholly new rational conceptual foundation for finance with essential implications for practice.

Log in at CFA Society Boston online to learn more and register for the series.  

by Richard O. Michaud, Ph.D.


Markowitz (1952) represents the birth of modern finance.  Prior to Markowitz, finance theory was largely security valuation.  Notable examples include Graham and Dodd (1934) and Williams (1938).  There was little attention to the portfolio as an entity or portfolio risk management as the focus of financial theory.   

The Markowitz mean-variance (MV) efficient frontier was invented as a model of institutional investment manager behavior.  Harry Markowitz was a Ph.D. student in the economics department at the University of Chicago in the early 1950s.  He was searching for a thesis on the stock market.  He studied institutional mutual fund companies and noticed that the portfolios they managed were all long only (no short selling) well diversified portfolios that varied by risk level depending on a client’s risk aversion.    Risk-averse clients prefer portfolios with less risky securities and conversely.  While sitting in Chicago’s business library, he had his famous epiphany.  He realized that a linear (equality and inequality) constrained quadratic programming (QP) mathematical optimization program will create a portfolio mathematically like that of a rational informed professional investment manager.  A MV QP framework requires estimates of return with a vague notion of security and portfolio risk conveniently defined as the variance.  In one stroke, Markowitz defined a financial theory of capital markets as well as invented a mathematical procedure to replicate institutional asset management in practice.  Later, Markowitz invented the Critical Line Algorithm (CLA) that made it possible to compute MV efficient frontier portfolios. 

Readers know that the history of financial theory is not based on the Markowitz efficient frontier.  The Capital Asset Pricing Model (CAPM) due to Sharpe (1964), Lintner (1965), Treynor (1961), and Mossin (1966) has been the dominant financial theory of choice since the 1960s.  While CAPM theory is based on Markowitz MV optimization, there is one key omission:  the no-short-selling constraint included in actual institutional fund management was omitted for analytical convenience.  This is because the Markowitz MV QP framework is not analytically solvable with only calculus. 

CAPM has been a very productive financial theory.  With some simple additional assumptions, you can show that the market portfolio is MV efficient, and that a security’s expected return in equilibrium is a linear function of “beta,” or the systematic risk of the security relative to the market.  In addition, beta enables defining “alpha” and the concept of the “active” return of a security or portfolio. 

CAPM is a very popular financial theory.  It spawned a multi-trillion dollar “quantitative” asset management industry.  Academically trained sophisticated professionals started many quant shops in Boston and around the world during the 1970s and beyond, with estimates of beta, alpha, and portfolio risk in order to outperform the market.   

CAPM is not Markowitz theory.  It is not based on an empirical observation of the behavior of informed investment professionals.  It is MV utility preference theory based on Von Neumann and Morgenstern (VM) (1944) game theory rationality axioms.  VM axioms are designed to reflect human rational decision making under uncertainty.  Simple statements such as:  if I like a more than b, and b more than c, then I like a more than c.  There is, of course, great mathematical sophistication to VM game theory.  John von Neumann is one of the greatest mathematical geniuses of the 20th century.  It is almost impossible to argue that the axioms do not reflect the reasoning of rational informed human thinking.  It was a very convenient set of principles to build a rational social science of financial behavior.  What could possibly go wrong with an expected utility theory of capital markets consistent with VM game theory?

Eventually, red flags started to appear.  The first known dissenter to game theory based finance was Maurice Allais (1953) (Nobel prize 1988).  He was interested in rational low-probability decisions.  He created examples to show that low-probability bets were often not consistent with VM rationality axioms.   Paul Samuelson said of Allais that if he had written in English instead of French the path of modern finance would likely have been very different.  However, many financial economists considered Allais’ examples tricks and the implications of his work were often ignored by American financial economists.   

A second red flag appeared in a Jobson and Korkie (1981) simulation study of unconstrained MV optimized portfolios.   They showed that the CAPM unconstrained MV optimization framework was worse than equal weighting a portfolio.  The study was not widely appreciated as a fundamental critique of the usefulness of CAPM theory for investment management.

However, the third red flag was widely noticed.  The famous Kahneman and Tversky (1979) experiments showed that investors have a consistent bias towards risk aversion that is inconsistent with VM rationality axioms.  Very simply, informed investors do not think like VM and, by implication, CAPM is not a useful framework for practice.   

CAPM is “Finance’s Wrong Turn.” It is what the course given at CFA Society, Boston in March is all about.  It is about returning to the empirically based Markowitz MV efficient frontier framework and resetting financial theory on a more reliable path for practice.  The landscape needs to clear many artifacts of 20th century theory and practice that have accumulated for more than 50 years.  There is much to talk about and much to think about.  The effort to begin again is solvable. 

I hope to see you all in class.




Allais, M.  1953. “Le Comportement de l’Homme Rationel devant le Risque:  Critique des Postulats et Axiomes de l’Ecole Americaine.”  Econometrica 21(4):503-546. 

Graham, Benjamin, 1934.  Security Analysis.  McGraw-Hill, New York. 

Jobson. D. and B. Korkie 1981.  Putting Markowitz Theory to Work.”  Journal of Portfolio Management 7(4):70-74. 

Kahneman, D. and A. Tversky 1979.  Prospect Theory:  An Analysis of Decision under Risk.” Econometrica 47(2):263-269. 

Lintner, J. 1965.  “The Valuation of Risk Assets and the Selection of Risky Investments in Stock Portfolios and Capital Budgets.”  The Review of Economics and Statistics 47(1):13-37. 

Markowitz, H. 1952. “Portfolio Selection.” Journal of Finance 7(1): 77-91.

Mossin, J.  1966.  “Equilibrium in a Capital Market.” Econometrica 34(4):768-783. 

Sharpe, William F. (1964). Capital asset prices: A theory of market equilibrium under conditions of risk, Journal of Finance, 19 (3), 425–442

Treynor, Jack L. (1961). “Market Value, Time, and Risk,” Unpublished manuscript.

Williams, J.B., The Theory of Investment Value.  Harvard University Press, Boston.

1. This course centers on three issues that explain the crisis in modern finance: finance theory is in error, institutional investment technology is out of date, and capital markets are not well understood.

2. CAPM is based on theoretical assumptions, Markowitz is based on informed investor behavior.

3. From a Markowitz theory perspective, finance is a lot like sociology where a back test is not likely reliable.

4. Unconstrained MV optimization is not useful for practice.

5. Constraints are generally necessary for producing useful results.

1. 20th century quantitative active equity management is a mathematically inconsistent hybrid of CAPM theory and Markowitz MV optimization

2. Commercial equity risk models do not measure risk but provide a description of portfolio characteristics in familiar terms convenient for practice.

3. Many long-term positive relationships of factors with return may not be helpful in actual practice.

4. Brinson and Hensel studies of asset allocation are based on similar long-term institutional portfolio historical data yet reach very different conclusions.

5. Roll (1992) shows that index-relative mean-variance optimization relative to the market portfolio often leads to suboptimal portfolios.

1. Two very respected financial economists (Andrew Lo and Meir Statman) have independently expressed concerns about the crisis in economics.

2. Neoclassical finance posits “rational agents” in capital markets, however investor behavior is often inconsistent with Von Neumann-Morgenstern game theory utility. 

3. Based on Markowitz observed behavior axioms, CAPM is not a realizable theory for asset management practice.

4. The Efficient Market Hypothesis may be useful for understanding simple paradoxes but may have limited relevance as a framework for understanding capital markets. 

1. Markowitz mean-variance (MV) optimization, the institutional standard for more than 60 years, fails to work in practice due to estimation error insensitivity.

2. Black-Litterman (1992), which set out to fix the instability in MV optimization, was in fact nothing new and inherited all of the MV optimization limitations.

3. The Michaud Efficient Frontier (1998) achieves greater diversification and portfolio composition stability by averaging properly associated portfolios from MV.

4. Forecast Confidence (FC) allows Michaud optimization to be customizable to investor styles, market outlook, and client/institution investment horizon by adjusting the number of simulated periods.

5. A statistical similarity test to determine rebalancing (as opposed to a calendar-based or asset-weight-based method) can help to avoid trading in noise and improve the effectiveness of trades.

1. Frank Knight (1921) proposed three types of uncertainty, notably the uncertainty of non-repeatable random events which characterizes many business decisions. In Knightian Uncertainty, the investor has no other decision making instrument than their judgement.

2. Orlean (2014) shares the Knight-Keynes view of radical uncertainty and proposes two capital market frameworks: negative and positive feedback markets. In a positive feedback market, liquidity is the ultimate source of price.

3. The Mimetism Hypothesis states that the market is a sociological, self-referential social psychology system geared by mimetism. The conventions that emerge represent stability points in investor belief. Calm markets followed by random events such as a volatility spike are associated with mimetic convention.

4. Young (1998) uses a Coordination Game (CG) framework as a contrast to neoclassical finance and economics. In a CG, equilibrium exists within a dynamic framework and evolves over time as prices and markets coordinate economic activity. Agents gather information, adapt, and act sensibly but are not VM rational in this framework, and often iterate to dominant strategies.

5. Goals-based investing in mimetic markets requires continuously diversified, actively monitored strategies across time.