Raising Modern Portfolio Theory (MPT) from the Dead
Investors and their advisers should accept that there is life after modern portfolio theory (MPT), C. Thomas Howard said in his discussion on behavioral finance at the 70th CFA Institute Annual Conference.
His presentation and the Active Equity Renaissance series he co-authored with Jason Voss, CFA, call for the demise of MPT and the capital asset pricing model (CAPM).
Howard and Voss maintain that “financial markets should be viewed and analyzed using a behavioral lens.”
We have a different opinion.
While behavioral finance has earned its place in investing and asset management, we believe it operates best in conjunction with MPT, not apart from it. Perhaps the father of MPT, Harry Markowitz, said it best when he remarked, “The most important job of a financial advisor is to get their clients in the right place on the efficient frontier in their portfolios . . . But their No. 2 job, a very close second, is to create portfolios that their clients are comfortable with.” That’s where the human behavior variable comes into play.
Howard argues that diversification is unnecessary: If investors understand how behavior creates opportunity and that risk reflects human behavior, then a 100% stock portfolio is all they need. Investors can take advantage of the market mispricing caused by the bad behavior of others, while simultaneously ignoring their own emotions, to generate maximum returns.
However, he also says that the starting point for successful investment management is needs-based planning wherein a client’s portfolio is allocated into liquidity, income, and growth “buckets.”
The liquidity bucket contains such low- or no-volatility securities as bank accounts or money markets. The income bucket has high-yield stocks and other high-yield alternatives, such as master limited partnerships (MLPs), and the growth bucket is filled with equities. While the latter bucket may be concentrated in a single asset class, the overall portfolio is diversified.
This is essentially a two-asset portfolio — stocks and cash — that is, by its very nature, mean-variance optimized in any combination. Perhaps life after MPT is still MPT.
Howard makes a strong point that emotional comfort is important. In fact, he states that emotion is the most critical determinant of long-horizon wealth.
A client who can’t maintain their composure in the face of the volatility of a 100% stock portfolio will be driven to allocate to other investments. While Howard would like to call this an emotion-return trade-off and not a risk-return trade-off, they are one and the same.
The common argument against standard deviation as a measure of risk is that real risk should be defined as a permanent loss of capital. Yet emotional investors who watch the value of their 100% stock portfolios drop to an intolerable level will make an emotional decision to sell at or near the lows, thus creating a permanent loss of capital. Volatility, or standard deviation, leads to emotional responses that cause a permanent loss of capital. Risk by any other name is still risk.
So how can MPT help advisers add value? (And to clarify, our discussion considers MPT in the context of multi-asset class portfolio construction, not in an equity model or single asset class allocation.)
First, MPT creates a framework for objective decision making. All advisers should avoid subjective analysis and emotional decisions.
Behavioral finance teaches that we are all susceptible to behavioral biases. Any effort we make to avoid these biases leads to better investor outcomes. However, MPT is not the only way to remain objective.
In the case of Howard’s “bucket” strategy — reminiscent of goals-based wealth management, pioneered by Jean L.P. Brunel, CFA — advisers can stay objective by using non-economic methods to determine bucket sizes. According to Howard, the liquidity bucket can be calculated based on the specific number of months or years of cash the client requires, with the other buckets holding the remainder. This then qualifies as an objective approach.
Other advisers may use a 1/N approach in which each asset class has an equal weight. Again, this is a reasonable method of portfolio construction. Where advisers often try to add value but don’t is when they allocate to an asset class “because it feels like the right time to be over- or underweight” or when they believe they can time the market. To the extent advisers feel the need to try this, they should do so at the margin. As Clifford Asness and his associates said, “If you sin, then sin only a little.”
MPT helps evaluate the incremental benefits of adding additional asset classes to an existing portfolio and how best to allocate among those asset classes. Whether augmenting a stock and cash portfolio with commodities or mixing international equities in with US stocks, the question remains: How much to allocate to each asset class? Again, with objectivity the key, any number of approaches are reasonable.
Remember though, explaining your process, rationale, and the supporting academic research is extremely powerful in any client-adviser relationship. Advisers earn credibility when times are good so that clients will trust them when times are bad.
MPT provides tools for assessing a client’s emotional (risk) tolerance. If we’re honest, this is where advisers really earn their fees. As Markowitz said, we should build portfolios that clients are comfortable with so they stay in them through all different market environments. We can talk about losses being temporary and mean reversion, but none of that matters if the variability of a client’s portfolio exceeds their emotional tolerance.
Howard and Voss repeat the claim that MPT’s measure of risk — standard deviation — does not reflect true risk, that only a permanent loss of capital is true risk. But understanding the standard deviation of a portfolio can help a client endure challenging market environments and avoid permanent losses.
Whether a portfolio is constructed using an optimization or not, having metrics showing the range of potential outcomes is incredibly valuable. Of course, these metrics are imperfect. For example, forward-looking capital market assumptions around risk and return for asset classes are preferable to historical data. Capital market assumptions are readily available from many reputable organizations, such as J.P. Morgan, and are relatively consistent across the industry. But historical data is readily available from many reputable organizations and is fairly consistent. The goal is not to be exact, but to have a context for discussion. We have found these conversations are far more fruitful in assessing a client’s emotional (risk) tolerance than any questionnaire.
Objectively analyzing diversifying asset classes for the value they bring to a portfolio is better than adding an asset class because the timing seems right or because everyone else is doing it. Discussing a range of possible performance outcomes with a client to determine their risk tolerance is better than investing 100% in stocks and hoping we can talk the client off the ledge when the market goes south. If we educate on the front end, far fewer counseling sessions and tissue boxes will be needed during the tough moments.
Despite harsh critiques in Enterprising Investor and elsewhere, MPT continues to be widely used by pension plans, foundations, endowments, and institutional consulting firms across the globe. If there was a better way to build a portfolio, these institutions would be the first to use it.
The adviser’s mission is to give clients the investing experience they deserve. The greatest danger to future wealth is failing to stay the course and not remaining invested long enough to reap the rewards. Diversification across asset classes and markets deflects emotional risk and still generates competitive returns. MPT provides a solid foundation on which to build that successful investment experience, while behavioral finance helps us counsel our clients when their emotions overtake their reason.
We don’t advocate that everyone use MPT, or even agree with it. The criticism of MPT, however, is inappropriate and off-target. No one ever promised that MPT was a way to avoid losses in all markets, nor that it was the sole methodology for portfolio construction.
MPT is a tool in an adviser’s toolbox — one they can choose to use or not. Behavioral finance and MPT are not at odds with each other or mutually exclusive. In fact, they work well together.
Investment paradigms benefit from the mosaic of research and ideas that continue to advance our field.
If you liked this post, don’t forget to subscribe to the Enterprising Investor.
All posts are the opinion of the author. As such, they should not be construed as investment advice, nor do the opinions expressed necessarily reflect the views of CFA Institute or the author’s employer.
Image credit: ©Getty Images/Craig Smallish
Thanks for the post! What are your thoughts on the limitations of mean-variance optimization techniques, a cornerstone of MPT, when looking at asset classes that have a non-normal return distribution? Is variance, or standard deviation, still then the most applicable measure of “risk” when history has shown us time and again that risky asset return distribution have significant negative skewness and excess kurtosis? One alternative may be mean-VaR optimization, but I’ve yet to see its implementation discussed by practitioners involved in the portfolio construction process.
Peter, I’ll refer you to the following link in which Markowitz discusses your question in detail with Antii Ilmanen of AQR. In short, he has tested alternative measures of risk and “determined that mean-variance was best in approximating expected utility versus proposed alternative measures of risk.”
https://www.aqr.com/~/media/files/words-from-the-wise/words-from-the-wise-harry-markowitz-on-portfolio-theory-and-practice.pdf
More semi-good tools are better than one great one, which, despite Howard’s assertion, BF isn’t. True Believers like him give me the willies. Like CAPM, BF semi-good.
You guys nailed it. Thanks.
Please keep preaching MPT to the choir. It’s a zero sum game and the returns of those of us who know MPT is demonstrably bunk have to come from somewhere.
MPT is still used because it uses SD , and thats the only way we can quantify risk whether we define it as permanent loss of capital or a downward fluctuation in portfolio value. If we can quantify it then we can build models and attempt to predit the future or at least sell it as the best educated guess.
As an investor , my test on allocation for an equity advisor is performance after fees for at least one full market cycle ( now by coincidence sbout 10 years )
and better still two full cycles .
Annualized return,sortino alpha, beta , max drawdown , up/down capture metrics .Very few advisors do very well risk adjusted long term.
Frankly MPT has its uses if not too rigidly applied as does BF ..and risk reduction based on extreme valuations .
all these theories are fine , but long term full cycle net risk adjusted returns as above are what finally cut the mustard
Have read many authors on and off this site that sound good but cannot show metrics as good as a handful of mutual funds with same manager we screened on these metrics metrics defined above and outperform handily .. and persist at least 5-10 years .
Happy to hear from any RIAs /CFAs who agree and have such track records .
Rare birds indeed .
thanks , marvin
Thanks Marvin, I appreciate the comment. We wouldn’t consider ourselves “equity advisors” but instead, asset allocators. We believe that to be the single most important factor in determining portfolio outcomes. If we can get that right our client will have a high probability of success. In allocating to equities, we try to find managers with a sound, defensible approach and a track record of consistent execution.
Nathan and Richard-
Your focus is on mean-variance optimization, which is one of the three MPT pillars. Based on your comments, it appears you have given up on the other two pillars, CAPM and EMH.
I agree that keeping clients in their seats is a critical challenge faced by advisers. I just don’t think the efficient frontier is of any help in this regard. Using the bucket model for making asset allocation decisions leads to increased client control, confidence, and comfort. It is irrelevant whether or not the resulting portfolio lies on the efficient frontier.
The advantage of the bucket model is that it reassures clients that liquidity/emergency and income needs have been met. This means the final bucket can be invested for long-term growth.
The growth bucket is where keeping clients in their seats is most challenging. The goal is to maximize long horizon wealth by investing in the highest expected return market, most likely the stock market. Emotional coaching by the adviser plays a critical role by helping clients avoid costly investment errors.
But of course there is only so much that can be done to allay client fears. When making the requested emotion-return portfolio decisions, standard deviation is a poor metric to use. At my firm Athena, when asked by our adviser clients to build less than optimal growth portfolios, we use max draw down as our emotion metric. It is not the ups but the downs that strike fear in the heart of clients. We take pains to frame the resulting asset allocation decisions as emotion-return rather than a risk-return trade-offs.
For 10 years now, we at Athena have used objective behavioral measures and methodologies for successfully managing fund, stock, and tactical portfolios, as well as for client communications. I see no compelling reason to bring MPT back from the dead.
Tom,
As an investor ..not an advisor, I think your bucket idea is useful.
It helps with emotions . I also like your use of max drawdown rather than SD
to gage risk tolerance .
For investors who have enough assets that safe buckets are already full, there remains the problem of equity allocation and management over the longer term . No one enjoys watching their equity bucket fall 50 pct .
In complex problems like investing in public markets there is seldom one answer . While MPT has its flaws in practice , it has some benefits if not rigidly applied..
I am not a theorist . I am looking for managers and systems that outperform
over full market cycles on a risk adjusted basis ..
I recently ran across one that reduces equity exposure gradually as valuations rise over adjusted CAPE averages.. Not perfect .,but very worthwhile .
As an investor , I welcome hearing from any manager ..author or reader .. with a demonstrable equity bucket performance over a full market cycle ( now 10 years ) that outperforms ..risk adjusted the handful of MF we have screened
for said performance ..maximum drawdowns is one of the criteria .
Marvin
Tom,
Perhaps we can discuss CAPM and EMH offline, as it’s outside the scope of the article. In my opinion they all have merit and imperfections, as each of their creators would readily admit. We have to use the tools in our toolbox as best we can. A bucket method of two asset classes (cash for liquidity / emergency and equities for growth) gets you to the same place as an efficient frontier of two asset classes. One challenge with the bucket approach is that the amount of assets required in the liquidity bucket to keep clients in their seats when they look at their growth bucket can decrease their long-term success, with neither the advisor nor the client knowing it. For example, if a 50 year old client with $5M tells you they need $2M in cash to “stay in their seats”, and that cash has a return of 0%, you’re at best getting a 6% return on the $5M over the long-term (and that’s assuming a generous 10% for equities in today’s world). At 50 with any normal level of spending, that money will not last throughout their lifetime. MPT can help me find ways to have the same level of volatility (or max drawdown), keeping the client in their seat, but also incorporate other asset classes with a greater than 0% return. It can also help me frame a discussion around a portfolio’s potential return path using standard deviation, max drawdown, or a number of other measures.
With MPT I have more information, a better discussion, and if I want, an objective methodology to diversify beyond stocks and cash.
As we stated in the article, we believe objective decision making in asset allocation is the key and MPT is one way to do it. You don’t have to bring MPT back from the dead if you don’t want to, but I would argue it never died in the first place.
I believe standard deviation is an incomplete metric used to measure risk. So any risk adjusted return that uses standard deviation in its calculation is irrelevant to me.
Ok Chuck , so how do you calculate a risk adjusted return excluding SD ?
Also how does your 10 year Annualized returns on equity bucket compare to BM on absolute and your own risk adjusted basis ?
maybe you can help my search .
thx marvin
We cant identify all the risks of an investment or portfolio. There are risks that we can identify but there are also unknown risks that we cant. So the total risk is unknown. Lets assume that SD is one of the known risks, which I believe it is. We will never know SD’s percentage of total risk and that wouldnt be a static metric even if we did. So, to answer your question “what is the risk adjusted return”, I have no idea neither do you. You can figure a SD adjusted return and call it risk and pitch that to clients because the average investor doesnt even know what SD is. Or better yet, put it into a model and really confuse them.
Cheers
Right on Chuck!
Chuck
I dont pitch to clients– have none- i am an investor for my family .
Re risk adjment calcs – im no theorist but one partial tool is to back out cash from equity exposure.
eg if i decide that instead of 60% equity because of high valuations I will cut exposure (dirty word– raise cash) to 40%, the risk i bear is on only 2/3 of my equity bucket — one third is risk free.
similarly I use Xray to back out cash and bonds from my MF holdings to get
the net equity exposure rather then the nominal ,
lastly i use full cycle history to check max drawdowns and up vs down captures to see risk beyond sortino whic is better than SD .
I am not CFA- an amateur — best I can do — can CFAs show me a better way?
Can any CFAs readers or authors meet my challenge to show me a better folio for equities than the few MF passing a screen –full cycle (now 10 years) for say 2-3 pct better net AR than BM and max drawdowns 2/3 of BM??
If so I’m ready to invest with you!
Marvin
Chuck, I think every advisor would admit that investing requires making decisions with incomplete information. If it didn’t, we shouldn’t earn a greater than cash return because there would be no “risk”. The best investors create a strategy and stick to it, regardless of what happens in the short-term. Unfortunately very, very, few people are able to do that. We get in our own ways and allow emotions to affect our convictions. If standard deviation can help me educate a client before those emotions appear about the potential short term performance of their portfolio and therefore help them remain rational in the moment because they’re not completely caught off guard, then I see it as a valuable tool. It’s not a sales pitch, it’s just information.
Agreed. That is a very useful application of SD. When too much attention is given to SD, thats where I have problem.
Probabily will be the subject of my monography.