Modern Portfolio Theory: Break free dude!

CAPM / Alpha Theory 12 Aug 2010

There were a series of government-sponsored television commercials in Canada in the mid-1980s that in simple yet effective terms sought to encourage young people to “break free” from smoking tobacco.

Showing teenagers at school playgrounds, on the streets and in other social settings shouting “break free!” (“Fumer, c’est fini!” en Francais) the ads (click here for a humorous glimpse at one of them on YouTube [Ed: Yikes.  Holy 80’s hair!]) quickly caught on as a catchphrase among the Canuck young, with the at-the-time-prerequisite “dude” tossed in at the end for good measure.

State Street might easily have drawn more attention to its recent report on re-thinking asset allocation by using the slogan as its title, given its message of encouraging institutions to “break free” from the more traditional and in their view somewhat outdated confines of modern portfolio theory, or MPT.

While noting both its usefulness and its ability to mitigate damage during market extremes, particularly during the 2007-2009 financial crisis, the report (click here to download) indeed makes the case that perhaps it’s time institutional investors wean themselves away from relying solely on MPT and at least consider applying it in conjunction with other more modern and effective variables.

As most in the allocation business know, MPT is a theory of investment which tries to maximize portfolio expected return for a given amount of portfolio risk, or equivalently minimize risk for a given level of expected return, by carefully choosing the proportions of various assets.

More recently, however, the effectiveness of MPT, which as the report notes dates back to the 1950s, has been challenged by fields such as behavioral economics, particularly in the wake of the financial crisis and the multitude of black swans that flocked to Wall Street and every other financial center.

State Street makes the case that there are now better and more effective tools centered on market volatility, portfolio construction and trading liquidity that in combination with MPT can help asset allocators much more effectively than straight MPT alone. What’s more, the emergence of quantitative approaches aimed specifically at tackling turbulent market activity and ensuing “non-normal” investment returns that can be harnessed to help construct more robust risk models.

The chart below from the report illustrates the distribution of differences between hedged and unhedged on a five-year rolling basis.

“Investors should think about whether two basic assumptions of MPT apply to their circumstances,” the report says. “Are their returns normally distributed, and is there a smooth trade-off between wealth and satisfaction (i.e. quadratic utility?) If these assumptions do not apply, they should consider an alternative optimization methodology called full-scale optimization.”

In other words, augment the traditional MPT route by throwing some cash at hedge funds and other types of alternative strategies that can diversify the portfolio, help keep returns decent, mitigate risk in the event hell freezes over again and all assets correlate to one, and at the same time ensure there’s an option to pull the chute, if absolutely necessary, according to State Street.

Case in point: According to the report, when US and non-US equity markets produce returns greater than one standard deviation above the mean, the correlation stands at 17%. However, when the return drops to more than one standard deviation below the mean, correlation sharply rises to 76%, meaning investments are likely to suffer with the market.

Indeed, the report notes that one of the greatest failures of risk management in recent years was using average risk numbers rather than regime-specific risk analysis to gauge risk – something MPT doesn’t incorporate or address in the new era of extremes, in State Street’s view (see chart below).

The bottom line, according to State Street, is that MPT still remains deeply relevant in terms of portfolio construction, diversification and risk management. But given new and updated techniques in contemporary financial markets that take advantage of computational and information aggregation capabilities, it makes more sense to cut back a little rather than kick the MPT habit outright.

Be Sociable, Share!

7 Comments

  1. Daniel Plainview
    August 13, 2010 at 8:54 am

    Markowitz, says the same thing as he talks about MVO “I never ever assumed that probability distributions were normal. I never justified Mean Variance analysis in terms of probability distribution being normal.” “My basic assumption is that you act under uncertainty to maximize expected utility.” It all most sounds like he expects people to act as rational free thinking beings, and as we know this is not all ways the case, people tend to anchor to ideas and seek out answers supporting our existing beliefs. Markowitz also talks about using a “Student’s t-distributions with 4.5 degrees of freedom plus or minus a half degree as a good model of stock return”. My long winded point is I agree with State Street, MPT and MVO are not perfect in the real world. We need to act as rational investors when applying academic theory to real world asset management.


  2. Peter Urbani
    August 13, 2010 at 5:37 pm

    Nice to see institutional inertia being overcome a mere 20 years after the ‘discovery’ of Post Modern Portfolio Theory methods. Markowitz bashing although fashionable is also a little unfair given that his seminal work was done in the 1950’s BC (Before Computers) and he did articulate the assumptions and the fact that it was a ‘normative model’ or how investors aught to behave and not necessarily how they did behave given the assumptions of i.i.d, quadratic utility, elliptical distributions etc etc. The primary problem of diversification is its reliance on the concept of correlation which is only a limited and linear measure of association and also on the assumption that it remains constant. In the real world of time varying correlations, volatility clustering and asymmetric preferences and effects it is not that surprising that the ‘classical’ theory doesn’t work that well. Still both the largest provider of Risk Measurement softaware, also being a little unfairly


  3. Peter Urbani
    August 13, 2010 at 5:40 pm

    Nice to see institutional inertia being overcome a mere 20 years after the ‘discovery’ of Post Modern Portfolio Theory methods. Markowitz bashing although fashionable is also a little unfair given that his seminal work was done in the 1950’s BC (Before Computers) and he did articulate the assumptions and the fact that it was a ‘normative model’ or how investors aught to behave and not necessarily how they did behave given the assumptions of i.i.d, quadratic utility, elliptical distributions etc etc. The primary problem of diversification is its reliance on the concept of correlation which is only a limited and linear measure of association and also on the assumption that it remains constant. In the real world of time varying correlations, volatility clustering and asymmetric preferences and effects it is not that surprising that the ‘classical’ theory doesn’t work that well. Still both the largest provider of Risk Measurement software, The largest Ratings Agency, almost all 3rd party Optimisation software and most institutions continue to use the ‘normal’ assumption.


  4. Mark H. Melin
    August 17, 2010 at 4:19 pm

    The obvious yet often overlooked truth about Modern Portfolio Theory is that it works best when uncorrelated assets are utilized. Most asset allocation models do not use truly uncorrelated assets, such as managed futures.

    Those who utilize MPT or other asset allocation models would do well to consider perhaps what can be argued is the most uncorrelated asset class in history, managed futures. How the managed futures asset class remains generally ignored and mis-understood by Wall Street financial advisors remains a mystery.


  5. Patrick Burns
    September 20, 2010 at 4:25 am

    The blog post “Ancient portfolio theory” contains my comments on this:
    http://www.portfolioprobe.com/2010/09/20/ancient-portfolio-theory/


Leave A Reply

← Research shows that hedge fund investors may want to join others by "selling in May and going away" Founder of the Journal of Environmental Investing: "It's all about alpha!" →