MODERN PORTFOLIO THEORY IS HARMING YOUR PORTFOLIO

By JJ Abodeely, CFA, Value Restoration Project

Thoughtful minds are reading Scott Vincent’s recent paper, “Is Portfolio Theory Harming Your Portfolio”(HT: @jasonbremer)

In the paper Vincent argues that the flawed foundation of Modern Portfolio Theory (MPT) that risk=volatility has allowed MPT advocates to control the language of the debate and set the stage for the obvious conclusion that passive index-based investing is inherently superior. And don’t think for a second that this debate is simply theoretical, academic, or unimportant– the basic tenets of MPT shape the decisions of nearly every institutional money manager, wealth management firm, investment counselor/consultant, and financial planner in profound and often disturbing ways. YOUR money is almost certainly being managed with these ideas at the core. The traditional approach to asset allocation is built on false axioms.

While Vincent’s direct assault seems to be focused on highlighting the mistreatment of active, concentrated equity or fixed income asset managers vs. holding a passive equity or fixed income index, his arguments hold sway over the much larger and dangerous consequences of MPT on asset allocation. My assertion is that most damage to investors portfolios from the traditional approach to investing comes from the foundation of static, backward looking assumptions informing broad asset allocation decisions.

Vincent’s Case

In the piece, the author does a fantastic job of summarizing the history of Modern Portfolio Theory and it’s building block components:

  • Harry Markowitz’s work on “Portfolio Selection” (which informs the chart above) proved mathematically that diversification could reduce volatility which he equated with risk
  • William Sharpe’s Capital Asset Pricing Model (CAPM) which mathematically defines an asset’s return into two parts:
    • systemic risk or “beta” which describes how an asset has historically behaved relative to “the market” or a broad index
    • idiosyncratic (stock specific) risk which can and should be diversified away
  • Eugene Fama’s Efficient Market Hypothesis (EMH) which asserts that the market is ultimately efficient at pricing all available information

Vincent notes that most advocates for passive investing and MPT start with a compelling statement:

Active managers in general have been shown to underperform passive funds, especially when taking into account their higher management fees, taxes, sales charges, and trading costs. If you can make more money in index funds then why bother with the hassle of trying to find a good manager?

On the surface, that’s hard to argue with. I’ve seen studies that indicate after taxes, 80% of active long-only equity managers underperform their benchmarks over long time periods. Here’s the problem: Most “active” managers aren’t active at all. They are closet indexers:

While diversification has always been a selling point for actively managed mutual funds, the average number of holdings in a fund have increased dramatically since MPT made the scene. The average number of stocks held in actively managed funds is up roughly one hundred percent since 1980… the average fund holdings had risen to approximately 140 positions by 2000. The actual number of holdings in a given year could easily surpass 200 because portfolio turnover exceeds 100 percent per year on average…Investors in actively managed funds suffer – they receive quasi-active management at full active management prices.

MPT and the quantification of investing has further (mis)informed the debate by seeking a easy way to label and quantify “risk.” In 1952, Harry Markowitz chose variance or volatility of prices or returns to define risk. He did so because it was mathematically elegant and computationally simple. However, this idea has serious limitations (most of which Markowitz has since acknowledged).

On the individual stock level, Vincent notes

Risk is often in the eye of the beholder. While “quants” (who rely heavily on MPT) might view a stock that has fallen in value by 50 percent over a short period of time as quite risky (i.e. it has a high beta), others might view the investment as extremely safe, offering an almost guaranteed return. Perhaps the stock trades well below the cash on its books and the company is likely to generate cash going forward. This latter group of investors might even view volatility as a positive; not something that they need to be paid more to accept. On the other hand, a stock that has climbed slowly and steadily for years and accordingly has a relatively low beta might sell at an astronomical multiple to revenue or earnings. A risk-averse, beta-focused investor is happy to add the stock to his diversified portfolio, while demanding relatively small expected upside, because of the stock’s consistent track record and low volatility. But a fundamentally-inclined investor might consider the stock a high risk investment, even in a diversified portfolio, due to its valuation. There’s a tradeoff between risk and return, but volatility and return shouldn’t necessarily have this same relationship.

I would add to this sentiment, particularly for equities as an asset class; there is actually more risk (chance of loss) when volatility is low, and less likelihood of loss when volatility is high.

Additionally, not all volatility or standard deviation is created equal. Most investors, for example, do not think that the prospect of achieving returns that are 2 standard deviations ABOVE average is risky, where the opposite is clearly so. So MPT’s cornerstone– its definition of risk– is incomplete at best and at worst, completely misleading.

Vincent continues

Regardless of MPT’s shortcomings on both a theoretical and empirical level, its dominating influence will not easily be dislodged. MPT is deeply woven into the fabric of our financial system, its mathematical grounding and precise answers inspire confidence. Further, its application is crucial in bringing increased scale and profitability to the financial services industry. Few want to see change. As such, common sense and judgment will continue to diminish in importance as top-down, quantitative strategies and blind diversification gain investment dollars.

And this is the part the ticks me off the most. The ideas behind MPT are so self-serving for our industry that despite their shortcomings, we can’t seem to move past them. Of course fear of embracing Maverick Risk (see my recent post) help keep the status quo alive and well

There’s a feeling of safety that accompanies index investing; neither the advisor nor the investor risks losing face or losing a job over putting money to work in a broad index. We enjoy the mathematical certainty of MPT, it’s reassuring that we can fix a value to assets, and that we can quantify risk in a non-subjective manner – free from human error…When defending an entrenched system that furthers the economic interests of powerful entities, the rationale doesn’t need to be sound, it just has to be somewhat convincing.

Vincent’s Prescription

The author contends that smart investors should welcome the dumb money move to highly diversified, passive strategies, while acknowledging if you ARE the dumb money or are forced on some level into limited choices (like 401(k) plan participants or beneficiaries of trusts with bank trustees– at least don’t pay active management fees for closet indexers.

An informed investor should welcome this shift. As highly-diversified strategies gain assets, inefficiencies become more prevalent because share prices are increasingly driven by factors other than fundamentals… there is compelling empirical research that shows active managers who are truly “active,” do persistently outperform indexes. The astute individual investor can seize the opportunity that blind, passive index investing provides in the form of increased market inefficiencies by hiring active managers who have shown the ability to exploit and profit from these inefficiencies.

Take advantage of the fact that your neighbors are leaving for passive funds, as their passive investments could provide the inefficiency your manager seeks to exploit. But, by all means, avoid investing in highly diversified active funds whose returns closely match an index. If index returns are what you seek, then pull your money and invest in efficient passive index funds or ETFs (emphasis mine).

This is where many investors stop thinking about their portfolio. They say, Yes! index returns is what I seek. After all, if you buy and hold the market you can earn the long-term returns right? Unfortunately, the answer to that is no. The long-term “average” returns are rarely available. In fact, depending on where you are standing, the returns are either much higher, or much lower. Consider this chart fromCrestmont Research which shows that even for periods as long as 10 years, average rarely occurs:

MPT & Asset Allocation

Scott Vincent seems to be focused on encouraging investors to choose the correct managers (truly active, relatively concentrated, fundamentally-focused) WITHIN asset classes like stocks or bonds. And he is certainly compelling. However, as I alluded to earlier, his argument holds sway over the much larger and dangerous consequences of asset allocation.

Consider this chart which you’ve probably seen in one form or another. It shows expected risk and return of various mixes of asset classes and the typical approach to asset allocation which Modern Portfolio Theory has spawned:

So what’s wrong with this picture? Lots of things.

The first is the inputs– namely expected returns and volatilities of various asset classes– most investment programs are built on logic like this:

  • Bonds will return 5% on average over the long-term but be between 0-10% in any given year
  • Stocks will return 10% on average over the long-term but be between -10% and +20% in any given year
  • Some might include other nuance regarding different types of bonds like High Yield or different types of stocks like Emerging Markets
  • Some might include different types of assets like real estate, commodities, or “alternatives”

The problem of course is this is an incomplete description of investment returns:

  • The math contends that returns are randomly and unpredictably distributed around the average
  • This “normal distribution” of returns contends that larger market movements outside of the ranges above will be relatively rare
  • “Average” returns ignore the role of valuation and the importance of when you start investing (buy) and when you finish (sell) even over multi-decade time horizons

The traditional approach to asset allocation is built on false axioms. The phenomenal secular bull market in stocks and bonds from 1982-1999 created the perfect conditions for the nearly religious acceptance of MPT. In a recent post, Expensive Markets Mean Low (or Negative) Prospective Returns, I made the case that valuation matters greatly and currently portend disappointing returns for both stocks and bonds. Traditional asset allocation has no way of dealing with this in a way that successfully protects portfolios from experiencing meaningful and unnecessary drawdowns.

As colleague Brian McAuley penned in Sitka Pacific’s March 2010 client letter

Whether or not these themes are presented directly, they underpin the advice that most investment advisors give their clients. At its core, the message is usually something similar to this: “The markets are random and unpredictable, so the best way to invest is to properly diversify and wait for the averages to play out.”

However, what most investors seem to be unaware of is that this whole theory of random movement of market prices was proven false over 50 years ago by one of the most influential mathematicians of the 20th century, Benoit Mandelbrot. The random motion of market prices was a very nice theory, but it just doesn’t match what actually happens in the real world.

In a completely random world, a large movement in prices would be a relatively rare event. But we know from market history that large movements in prices happen far more frequently than they should if prices moved completely randomly. In fact, if we look at annual market returns, there are 50% more extreme events than there should be… It’s clear that there are other forces influencing the markets that aren’t taken into account by the statistics of purely random movements—and they have to do with human behavior.

Since the movement of market prices is not random, most investment advice given today that is based on Modern Portfolio Theory is simply wrong. Particularly, the assumption that market risk can be reduced by diversification has led to frequent catastrophic results throughout market history—most recently in 2008.

Although the theory of random market movements was proven false more than 50 years ago for those fluent in mathematics, with multiple bubbles and busts in the last decade it has become painfully obvious to everyone that there is more to market movements besides a Brownian-like random motion. Markets are subject to the rational and irrational decisions made by millions of people, and as such they are prone to cycles of extremes in sentiment and valuation—booms and busts.

Even those who bill themselves as “active” asset allocators typically only move within a small range of acceptable allocations such as 40-60%. Consider even this progressive asset allocation policy:

My Prescriptions

Obviously, I think investors can do better. Just as Scott Vincent highlights the cadre of truly active, relatively concentrated investment managers who have “beaten the market” consistently– there are asset allocation strategies which aim to improve upon the traditional approach. At Sitka Pacific, we’ve taken an “Absolute Return” approach to the markets. From our 2006 Annual Review

We have the flexibility to take more meaningful actions to preserve capital, manage risk and volatility, and invest where the best opportunities are. Our goals are simple: to preserve capital first and generate an absolute positive returns second. In order to achieve those goals we look at the market as a means to generate a return, not something that should be blindly followed. We use the market when it can be useful to us, and can look elsewhere (even to cash) when there is more risk than potential reward in the market. The flexibility to manage risk may prove critical over the next several years.

Lots of managers aim to do the same thing. I’d recommend finding one who not only has demonstrated a successful approach in various market environments, but also whose process seems repeatable and makes sense to you. Take Mebane Faber at Cambria and his GTAA ETF. It’s a pretty straightforward process that using long-term momentum indicators to decide when and when not to be invested in various markets. It’s an improvement on a static approach, but might leave a little something to be desired for folks who wish to have a fundamental approach to their portfolio. John Hussman’s approachat the Hussman Funds is worth a look as well. For Do-it-Yourselfers, I’d recommend this post by David Merkel: The Impossible Dream Project for a good look at how to build your own process. For our approach to be successful, we need 3 Key Traits:

When one is trying to be dynamic and active in their approach, flexibility is hugely important. Whether it’s the inherent inflexibility of an investment committee, the curse of success (overconfidence), ingrained cultural constraints, or simply the challenge of moving around and effectively deploying a large chunk of assets, large, well established firms often lack flexibility.

Folks in this position, MPT devotees, and folks who don’t want to rock the boat too much would be wise to consider several emerging theoretical frameworks which aim to improve upon the status quo.

  • Post-Modern Portfolio Theory (PMPT) is compellingly described in this FPA Journal article by Pete Swisher and Gregory Kasten.
  • Allocating based on Downside Risk (versus Mean Variance) seems like a no-brainer and should be evaluated by anybody trying to “optimize” portfolio allocations. Here is a good primer.
  • Fundamental Indexing: Research Affiliates is a good source.
  • Risk Parity, which Bridgewater, PanAgora, and AQR are all pursuing in various forms, takes a different view of targeting acceptable levels of risk as opposed to returns

These approaches all have their own limitations and should be evaluated critically (like anything involving your or your client’s money). Ironically, one can make the argument that “Absolute Return” or GTAA or Fundamental Indexing can be considered as distinct asset classes, with their own projected risk and return characteristics and lower correlations to traditional benchmarks. In this case, MPT would suggest that “efficient” portfolios should have an allocation to them, just as many traditional investors have embraced alternatives to varying degrees.

Regardless of how you are currently invested, the shortcomings of MPT and its pervasiveness in portfolios is becoming increasingly clear. The last two years have given MPT-based investors a reprieve from the secular bear market underway since 2000. The next two years may not be so friendly. Investors should take the leap into truly active investing and asset allocation, even if it just means “trying it” with one or more managers or funds.

JJ Abodeely

JJ is a Director and Portfolio Manager for Sitka Pacific Capital Management, LLC, a SEC-Registered Investment Advisory firm offering Absolute Return and Global Multi-Asset Class strategies. Prior to joining Sitka Pacific, JJ spent more than 10 years in various capacities for a $1.2 Billion asset and wealth management firm based in Seattle, WA. As Portfolio Manager, Analyst, and a member of the Investment Policy Committee, he was responsible for the development and management of several new investment strategies, including Global Multi-Asset Class ETF strategies and an equity long-short strategy.

More Posts - Website

  • http://www.pragcap.com Cullen Roche

    JJ,

    Let me be the first to say well done.

    Cullen

  • http://riskandreturn.net Lance Paddock

    Let me be the second JJ. Excellent.

  • VRB II

    Mr. Abodeely-

    This is the second piece I’ve read of yours. Well done again. Not all CFAs are created equal. I think you should help the institute re-write Level III material. You nailed it once again!

  • http://www.valuerestorationproject.com JJ Abodeely

    Thanks for the comments. It will be interesting to see how professional and educational organizations like the CFA Institute, Economics departments, MBA programs, etc. change their curriculum over the next several years to keep up with practitioners. Many have started embracing Behavioral Finance which starts one down the road towards questioning the foundational framework of our industry.

  • http://neweconomicperspectives.blogspot.com Scott Fullwiler

    This will be required reading in my class this fall. Thank you.

  • Different Chris

    The is an incredible piece of work.

  • asking questions

    Nice article, but what about those with limited capital that wish to enter the market. Not everyone can afford to have their accounts managed or spend hours upon hours researching investment strategies. Index funds allow people to get a broad view of the market. Your insight was very informative. Thanks.

  • Jon Goldman

    This is a fantastic article. I’ll be referring to this for a while. Sending a link to my college’s investments professor, who as you can probably guess, teaches MPT. You would have put a capstone on your argument if you’d included a short discussion of what the Financial Instability Hypothesis means for the idea of an efficient market, which MPT is in turn based on. Furthermore, both Jesse Livermore and Bernard Baruch had great statements in rebuttal to market randomness – would have been inspiring to see a quote or three from those who prospered before the dark age of macroeconomics.

    Well done sir. I look forward to more of your articles.

  • http://www.valuerestorationproject.com JJ Abodeely

    Scott and Jon-
    I sent the article and post to my old college professor as well. Academia needs to catch up, especially liberal arts schools which might have limited resources in the field.

    “asking questions”- it is a conundrum, for sure. You either spend the requisite time or you stay away from markets. We’ve made our strategies relatively accessible, but I encourage our clients to spend time reading and evaluating our comments/strategy. Alternatively, I tend to think you could spend a few hours a month and build a pretty “smart” all weather portfolio of a few funds, etfs, and make changes when valuations are at extremes.

  • Roger Ingalls

    I enjoyed this very much! Detailed, yet accessible.

    Nice to see a few recommendations for the less financially endowed in there (if you’re going to index anyways, pick the least expensive), and a few suggestions for actively managed ETFs.

    Looking forward to your next article.

  • Different Chris

    I sent the article and your brillaint post to my undergraduate advisor too, small liberal arts school in PA.

  • http://neweconomicperspectives.blogspot.com Scott Fullwiler

    I’m always looking for good critiques of MPT. I teach it, largely because it’s standard in the field, but then always present materials debunking it. It’s also taught in the same course that I teach Minsky, so that provides an alternative framework on its own, though not necessarily regarding asset pricing and portfolios.

  • CDOSQUARED5

    I think you are conflating high expected returns and risk. You want to bundle the two together by saying stocks are cheap, therefore, not risky. For Example : “Most investors, for example, do not think that the prospect of achieving returns that are 2 standard deviations ABOVE average is risky, where the opposite is clearly so” MPT takes this into account through the expected return.

    You critique seems to be more accurately be stated as that you think current variance at market peaks and troughs is not an appropriate measure for projected variance. Just as we smooth earnings through a Schiller PE, price peak or normalized profit margin, we should do the same when projecting variance. That is reasonable to say but doesn’t lead to throwing away the variance framework. Additionally, you can argue that projected return distributions have skew and are not uniform. That is reasonable to say but doesn’t lead to throwing away the variance framework.

    The problem with saying variance shouldn’t be used is that I don’t believe your paper addresses what replaces it.. You have no method to determine how to allocate capital across various asset classes. The logic seems to be put 100% of your eggs in the stock with the highest expected return. But is the prescience and discipline of stock analysts to continually pick winners a good assumption?

    Risk Parity, Portable Alpha, fundamental indexation, etc. are firmly within the MPT framework.

  • VRB II

    Mr. Abodeely-

    My firm has 2 CPAs, 3 CFPs and a doctor.
    I have some nice letters behind my name also. When I did my CFP my hair stood up studying the investment module. I manage our firms money and really struggle with an older CFP, E.A who always wants to be fully invested and is a die hard MPT diversification disciple.
    We’ve gotten into some heated arguments. Last week I threw my monthly letter to my clients at him and had to leave for the day.
    I really appreciate your ability to explain your thoughts.
    But I really can’t thank you enough for offering your approach, being transparent in your process, and the intellectual capital of other professionals.
    I wish you all the best and thank you for your service to our proffesion.

  • Uncle Tupelo

    Everyone has limted capital in some sense. I have given a fair amount of thought to this question and I respectfully submit that if one does not have the passion, aptitude and time to study the markets to a meaningful degree, then one should not be in the markets. This is related to one of the author’s points in that the ability to judge true risk (which is not an absolute but may/will differ from investor to investor) is one of the key essentials for investment survival and hopefully success, and judging risk takes some time and attention.

  • Octavio Richetta

    A+! But for the average guy, or even the clever ones, it is not so easy to act on the insights provided. The way to outperform the market is to take significant, no diversified positions, that prove to be right. You can try to do it with a margin of safety but it is still very hard.

    If one assembles a mosaic of the above vehicle ideas one probably end up with kind of an “index” asset allocation fund.

  • El Viejo

    “. . . Markowitz, now at the University of California, San Diego, followed a hunch in 1952 when he split paycheck contributions to his retirement account equally between stocks and bonds.
    Economists call this approach ‘1 over N,’ distributing money evenly among the number of available investment options, the Ns. The 1/N strategy is also called ‘naive diversification,’ . . . ”

    “People can make better economic decisions by using simple rules of thumb.”

    “. . . inspired by a 2009 study led by economists Victor DeMiguel of the London Business School and Lorenzo Garlappi of the University of Texas at Austin. Using 40 years of data from the U.S. stock market, DeMiguel and Garlappi found that 1/N portfolios consisting of either 25 or 50 stocks usually generated greater returns than 14 complex models for investing in the same stocks.”

    From Science News: June 4, 2011 (www.sciencenews.org)

    So us amateurs have an equal chance of doing well.

  • El Viejo

    Title of above article:

    “Simple Heresy” by Bruce bower

  • Inge Berqvam

    Is it correct to claim, as in the article that the fluctuations in the market is not random because it’s disrbution is not normal? I mean, there are a lot of other random distributions than the normal. A t-distribution would take care of the fat tail. However, it is harder to work with matematically.

    That aside, I personally do not believe in MPT.

  • hkm

    Wow. Is there no one who questions any of this?

    Personally I agree with everything that is said about MPT, having long felt that the industry has gone head over heels for a theory based ultimately upon the utopian idea that, averaged across some mix of assets, the markets will always magically go up. Kind of like what everyone said about Real Estate in 2006 (and a thousand other situations before). Moronic.

    So there is a lot to admire here.

    To me, though, it all goes off the rails a bit when I get to the mention of John Hussman, who presumably would be the kind of manager that one would need.

    I accept that Mr Hussman is widely respected. By all accounts he really is a decent sort, too, so I am not looking to pick on him.

    But actually, to me his fund performance has been almost counter-productive. Yes, when he started the ‘Growth’ fund in 2000, it was a shining stand-out, zigging upward for a couple years when everything else was really crumbling. But after that, for years it did very little. Then in 2008-2009, when one would expect that it might come back to shining as a short fund again, all it did was lose less than the rest. Which still would have been highly admirable, if only it had been gaining at least a bit in all those positive years. But, basically, if you had money in that fund since 2004 you might as well have just held cash for what you got out of it.

    Of course, a better bet over that period would have been his newer Total Return fund. Over its lifetime it has left HSGFX in the relative shade, and has been comfortably positive every year. It has a great history. But basically that has been a bond fund in one of the great bond markets of all time. How’s that going to work out if the bond world goes bad, maybe for ten years or more? Going forward, there must be a decent chance that Strategic Return could end up just like Growth: little different than holding cash.

    Which brings me to what I think is the salient point.

    Yes, I agree that good active managers are better than MPT, but 1) how is that measured, and, just to be really practical, 2) how do we identify those Magical people?

    To me, the only answer to both questions is that we look at their….past performance. What they have done in the past. Which, of course, is no predictor of the future performance. Which, by the way, may also be far more random than we might ever plan for it to be. What if they just got lucky? What if they had the right style for (that) right time? What if they just lose it (as several notable managers have)?

    As an active investor for over 25 years, I have run into this little wall more times than I care to remember.

    To me, sadly, that makes this theory just as utopian and misleading as MPT. It’s a great sales pitch for besieged active managers, but sadly nothing that one can actually activate with any kind of certainty.

    In fact, it probably involves massively more risk (through entrustment to the wrong manager or methodology) than fluffy old MPT.

    If someone can tell me how you identify the Magic Manager with any degree of certainty, I will have my assets at his or her doorstep first thing tomorrow morning. (The Manager’s doorstep).

  • http://www.valuerestorationproject.com JJ Abodeely

    I would agree and add to this comment that putting money away pre-tax, in a short-term bond fund or savings account, and letting that compound over time, you’d do just fine.

  • http://www.valuerestorationproject.com JJ Abodeely

    “Risk Parity, Portable Alpha, fundamental indexation, etc. are firmly within the MPT framework.”

    I completely agree– however they contain incremental improvements from a theoretical and or practical standpoint (from my perspective). I will also add (and perhaps did not make this clear enough) that one of the major problems with MPT in action, is the degree to which people explicitly or implicitly rely on historical inputs for returns and volatility. Even those attempting to use forecasts are still “anchored” to (mostly recent) past experience in ways that limit its effectiveness.

    The variance framework could be improved, for example, by focusing on semi-variance or downside deviation. That is what PMPT promotes (see the link from my piece). Markowitz himself thought it was an improvement, however he lacked the computational power to use it in the 1950s. But even semi-variance has limitations– so it is obvious what the answer is:

    “The perfect investment, as everyone knows, is positively skewed, leptokurtic, and has low semi-variance. But these moments about the mean of an investment’s return probability distribution are at least partially incompatible since investment returns are non-Gaussian, and variance/semi-variance obviously loses its utility in non-mesokurtic (that is, non-Gaussian) skewed distributions. Which is exactly the point.”

    Translation: Ideal investment portfolios make above average amounts of money frequently and when they do (inevitably) perform poorly it looses small amounts and not too often.

    There are different ways of trying to accomplish this, both within asset classes and allocating amongst them. There is no one right way, but there are lots of wrong ways. Please see my comment below for how we do it.

  • http://www.valuerestorationproject.com JJ Abodeely

    I don’t think concentration (non-diversification) is required. We’ve generated absolute returns with a highly diversified portfolio across multiple asset classes. The key is that it is active, not static.

  • http://www.valuerestorationproject.com JJ Abodeely

    I would agree the 1/N is a fantastic approach for those not wishing to be active. I recently heard a presentation by Michael Falk (msfassetconsulting.com) that tore into MPT and then drew the same conclusion about the power of 1/N.

  • Spellcheck

    Financial professionals should know the difference between “Risk of principle loss” and “Risk of principal loss”.

    The first makes you unprincipled, while the second makes you poor.

  • Andrea Malagoli

    Great post. I am fully supportive of the arguments presented and I have done some work along the same lines myself here:
    http://www.scribd.com/doc/55836743/Articles-Malagoli-Young-PI-Published-0211

    The one further step I’d go is to say that most variations of MPT based on quantitative methods are likely to go nowhere as well. For example, using downside risk as an alternative measure is only a minor change. It turns out that the risk and size of a drawdown is mathematically related to volatility, so the two risks are actually essentially the same.

    The other important concept to keep in mind is that the randomicity of returns is a more important factor than skewness oand or fat tails in the distribution of the period returns. I.e. what matters for risk is whether or not sequences of returns are truly independent or become autocorrelated, as it happens during market crashes. This is a much stronger effect than the existence of fat tails in the single-period distribution.

    One of the big limitations of MPT and other modern finance methods is that they only look at the statistics of single-period returns, and not at the dynamics of how returns compound. This has lead to numerous misunderstandings about risk and returns, like the famed notion that stocks become safer in the long run.

    I agree that MPT is hard to eradicate, but I think that all the marginally improved solutions, like risk parity portfolios, are equally dangerous ideas.

    When it comes to managing portfolios, I think the time spent thinking about how market work (micro and macro) is much better spent than on quantitative analysis. Quantitative analysis is inevitably backward looking, and no amount of fancy math will ever change that.

    This does not mean that portfolio construction should be a purely qualitative endeavor. One just uses data differently. I compare it to the process of “pattern recognition”, where one looks at the current data and looks back in history to see where similar patterns may have occurred. If these patterns exist, then one can make a reasoned judgment about how the current and past situation are similar and how they differ. It is not the clean and rigorous looking process that makes MPT so attractive, but it is much more effective by far, at least the way I have used it.

    What worries me about theories like MPT is that they give people a major excuse to suppress critical thinking and plain old common sense.

    MPT essentially tells people that it is ok to buy assets at any price, because they’ll make money eventually if they diversify enough. Who would ever shop for anything like that?

  • http://www.valuerestorationproject.com JJ Abodeely

    MKM- I hear what your saying and at the risk of sounding like a jerk, I think your expectations might be too high.

    I wrote in an above comment that ideal investment portfolios make above average amounts of money frequently and when they do (inevitably) perform poorly, they lose small amounts and not too often.

    A relative return manager might define “average” as the benchmark while an absolute return manager defines average as 0 or maybe short-term bond yields.

    I can’t comment on Hussman’s funds specifically as I have not analyzed his results, only his process. Every manager makes mistakes, even Hussman, GMO, Seth Klarman, Bridgewater, and Warren Buffett. The key, IMO, is finding managers that when they make mistakes, your capital doesn’t get killed. Most of the time when these guys are wrong, it usually means that you are not making as much money as you would have in “the market.”

    Also, size plays a role. Even for these guys, the bigger they got, perhaps the harder it was to earn returns. I believe Hussman’s technique for hedging, for example, has changed over the years by necessity because of his large AUM.

    Now these are all managers who “have done it” through various market cycles and shifts in style. Each would perhaps say that luck has played a role as well.

    Identifying the Magic Manager is as you suggest impossible, and every person might have a slightly different list of characteristics they seek. Here is what I’d look for and what we strive for:

    1. A focus on preserving capital first, earning a positive absolute return second. An comprehensive view of risk and the proper temperament (not computer models) to manage it.

    2. A focus on anticipating fundamental changes (whether macro, company specific, etc.) and the ability to assess value relative to price.

    Overlaid on all of this, is a fiduciary framework whereby manager and investor interests are aligned and the manager can be deemed to be genuinely passionate and insatiably focused on improving his process, and ultimately, results.

    Finally, evaluation over the long-term will help you determine when or if an manager has indeed “lost it” and you can move on.

  • http://www.valuerestorationproject.com JJ Abodeely

    I apologize if I have errors, sometimes I move to fast. I can’t find the typo anywhere, though. Can you point me to it?

  • Majorajam

    Where to start? Firstly, indexing does not require a belief in normally distributed returns or MPT (neither implies the other, fyi). It’s simply a concession to the existence of both opportunity cost and the fact that, after costs, the sum total of active managers will always underperform the sum total of index managers. The latter seemingly controversial exposition is actually the indisputable consequence of mathematical identity.

    Secondly, please show me any shred of empirical evidence that the semivariance or downside deviation, or any technically based indicator for that matter, is any better at forecasting risk than simple variance. It is not, your dissembling on ‘skewed non-mesokurtic distributions’ notwithstanding. If you’re looking for where the action on improving risk forecasting is, look to correlations (and even here, the upside is highly limited).

    Thirdly, please define absolute return and explain how it both a) applies to your strategy and b) does not apply to passive investing. This word, I do not think it means what you think it means.

    Lastly, I’ll just note the curious synergy here in someone who’s strategy depends on a preponderance of ‘greater fools’ willing to bid their asks and ask their bids encouraging unsophisticated retail investors to tilt at exceptional return windmills.

    For my own part, I think MPT is an exceptionally poor analytical framework for predicting, or even understanding, high-level market dynamics. It is even less suited, as if that were possible, for informing macroeconomic policy or otherwise providing insight into structural parameters, notwithstanding its continued use by freshwater dead enders toward those ends. But despite all that, and only somewhat paradoxically given its partial validity, it can be put to highly effective use for managing money in the hands of smart practitioners, and in any case is far less dangerous in the hands of dumb ones than the loose cannon approach you seemingly advocate, largely by virtue of what I have written above.

    Certainly, given my first point above, people could do much worse than indexing. And do, all the time. Ergo successful active managers. Ergo the bloated financial industry.

  • Octavio Richetta

    Hussman is the best example that comes to mind to demonstrate that market time does not work/is hard has hell. If this guy cannot time the market who can?:-) All his fancy measures of market action/risk etc. are just of fancy way of trying to time the market.

  • Andrea Malagoli

    Where to start?

    I am with you on that. You have clearly neither read my comment carefully or my article, and therefore have totally misunderstood the points I was trying to make.You are also incorrect in saying that indexing does not imply MPT. MPT and indexing were both born out of the Chicago school of “efficient markets”.

    I am not going to comment on the many other incorrect statements, but I want to take issue on this:

    “Lastly, I’ll just note the curious synergy here in someone who’s strategy depends on a preponderance of ‘greater fools’ willing to bid their asks and ask their bids encouraging unsophisticated retail investors to tilt at exceptional return windmills”

    these types of comments are really inappropriate and are a clear mis-representation of my comments. You should not feel that hiding your identity should give you permission to make derogatory and arbitrary statements that you cannot support with either data or competent financial analysis.

  • alex p

    Cullen,

    Many, many thanks for posting this. Don’t always agree with your reasoning on some things, but you are always respectful and infromative. That says a lot.

  • http://riskandreturn.net Lance Paddock

    Actually, Hussman is a great illustration of how understanding a manager requires looking past short-term issues of performance. During the crisis Hussman performed brilliantly, and his set of techniques applied faithfully would have had his clients showing a large absolute return over any reasonable time frame. The issue is not his performance prior to the crisis, or his performance during. Essentially he went away from his normal discipline for part of one year do to a concern that extraordinary circumstances would have exposed his clients to potentially disastrous results.

    He could have been right about that and looked stunning, but we all got lucky and he ended up not going back to his standard methodology until markets were once again overvalued. That cost him the upside that would have made his caution prior to March 2009 a huge plus.

    As an analyst of managers you now have the ability to move beyond complaints about his returns over a single period, and a framework for judging him prospectively. Assuming that a similar event does not occur, one can see that if he behaves as he has in all other time periods, that his returns should be quite good. From studying him you also know that he has come up with a framework for dealing with those extreme events should they (or other types) occur in the future which should allow him to take advantage of bounces as he did in 2003.

    More importantly, given the levels of overvaluation in todays markets we should expect his prospective returns to be superior to market benchmarks until they have declined a good bit. He made a mistake (and his mistake was not unreasonable and could have paid off and a product of extreme events) and as JJ said it was one which didn’t cause massive losses. So, when we invest with a manager we should invest not based on an outlier (whether up or down) but on what is most likely to occur. The most likely expectation is that Hussman wil be relatively to extremely cautious until markets get significantly cheaper, whereupon he will increase his exposure. That is a time tested method for outperforming and achieving high absolute returns over time.

    Throwing him out because of March 2009 through December of 2009 would be a big mistake unless you do not believe that the method will work in the future as well as it has in the past or that he will depart from his typical methodology in the future. He didn’t do it in 2002-3 when events shook many and he tok advantage of the upside, he didn’t do it in 2006-7 when his caution was wrongly derided. I assume that his willingness to endure short term uderperformance is thus intact and that we should see him faithfully carry out his method. If his behavior and communication suggested he didn’t see the issues with what he did, or radically altered his approach to try and “make up” for his trailing, then I would be concerned even though it would have made his numbers look better.

    Looking under the hood, his emphasis on higher quality companies as opposed to the riskier fare that has become extremely overvalued, should give him a large alpha boost over the next few years. For that reason I would suggest allocating more money to him now, not less. Using our own high quality holdings as a guide we have seen a huge shift with quality radically outperforming since Mid April, and we see the same effect helping Hussman as well (almost 500 basis points above the S&P 500 over the last 30 days.) That worm is turning, so not only would I expect that his general approach will add value over time, but his specific positioning is in an area which I expect to beat the benchmarks by six to seven hundred basis pints over the next 5-7 years.

    Anyway, that is not a full analysis, but it shows how one should begin thinking about analyzing managers. Trailing returns can be misleading, both positively and negatively without a thorough understanding of what a manager does and only then can it help you look at the likely value they can add in the future, which may be less or more than in the past. Periods that fall out of what you expected need to be looked at carefully to determine if they are likely outliers, or does it show a deeper issue. In Hussman’s case I would expect his returns to very good on both a relative and absolute basis going forward.

    The bond fund is a different case, but if you e-mail me I will be glad to give you an analysis about why his fund may have more merit than the bond bull market comment implies, though we do not use it at this point.

  • Geoff

    Mr. Roche, it would to have a “printer friendly” feature on your site so that readers could print excellent posts like this one.

  • Majorajam

    Andrea,

    You clearly did not read my comment and are incorrect in assuming I was responding to you, or that I give two figs about your comment or your paper.

    You are also incorrect in claiming that index funds ‘imply MPT’ (not that my original remarks could even be fairly described as the converse). While it’s true that strong form market efficiency dictates that beta-adjusted active management underperforms in every period, stock, etc., and that even weaker expressions of market efficiency have strong implications for the attractiveness of active management, that bit of fancy from notoriously fanciful freshwater economists has nothing to do with index funds.

    Here on the mainland we look for a bit more than historical happenstance to draw conclusions. Here’s a bit you might consider relevant from the Wikipedia article on what led to Vanguard’s setting up the first index mutual fund (on the S&P500):

    For his undergraduate thesis at Princeton, John C. Bogle conducted a study in which he found that around three quarters of mutual funds did not earn any more money than if they invested in the largest 500 companies simultaneously, using the S&P 500 stock market index as a benchmark.[3] In other words, three out of four of the managers could not pick better specific “winners” than someone passively holding a basket of the 500 largest public U.S. companies. The managers could pick specific stocks which would do as well as picking the 500 largest stocks (essentially doing as well as random chance would dictate), but the cost to pay their expenses, as well as the high taxes incurred through active trading, resulted in underperforming the index.[4][5]

    Please note that empirical performance studies != theoretical nonsense peddled by the Chicago school.

    These types of comments are really inappropriate clearly evidence a stunted intellect. You really should consider posting anonymously if you’re going to make such inane statements as cannot be supported by a modicum of understanding, data or relevant analysis.

  • http://www.valuerestorationproject.com JJ Abodeely

    Majorajam & Andrea-

    First let me say that I have no intention of playing tit for tat or responding to every criticism. I put my thoughts out there and engage others in discussion because I believe it can only help raise the bar for all involved, especially myself. Civility should be required and were it not for a well-placed Princess Bride quote, I’d probably not wish to respond to Majorajam’s criticisms. But he makes a few good points and raises some good questions that I will seek to answer.

    I agree that indexing does not equal strict adherence to MPT and for many its a simple antidote to the simply stated fact that “active managers on average underperform.” HOWEVER, that simply stated fact (as Vincent’s original article points out) is incomplete for those willing to use a bit of critical thought and apply a bit of their own judgement. Namely, if one thinks that some of the principals of MPT are flawed (various forms of EMH and the idea that risk=volatility)then it is not only possible, but quite easy to find managers that are likely to “outperform” over a reasonably long-time period.

    What does “outperform” mean? For relative return, asset-class specific investors or funds it means beating an appropriate and specific benchmark. I argued in my previous post on Maverick Risk that most widely used benchmarks (S&P 500) are artificial constructs that have little to do with the natural liabilities of investors (funding retirement, etc.).

    Absolute Return minded investors like ourselves, “out-performance” means beating a natural benchmark such as zero or inflation. In other words, protecting and growing the future purchasing power of current assets. We do this by attempting to make above average (0 or inflation) amounts of money as frequently and consistently as possible and positioning the portfolio such that when we do (inevitably) perform poorly it happens not too often and in small amounts. We believe that is the best way to grow assets over time.

    Passive Investors can be Absolute Return oriented as well (as Majorajam rightly points out) however in my experience, MOST ARE NOT. In order for a passive investor to be able to expect positive real returns, they must have a time horizon of many decades, not years. For example, at the lows in 1982 stocks reached the same level adjusted for inflation as 1905 (77 years earlier!). How long will investors passively buying the market portfolio in 2000 have to wait for their returns to become positive on an inflation-adjusted basis (as a point of reference, I believe it took the 1929 buyer of stocks nearly 20 years)? Why do most passive investors not really focus on absolute returns?

    1. They lack sufficient perspective on the market. An absolute return oriented passive investor would not subject his portfolio to 16 years of negative real returns in “the market portfolio” of stocks AND bonds such as those that occurred when starting from expensive markets in 1965.

    2. Most humans lack the wherewithal and discipline to stick with it and tend to buy the highs and sell the lows or change the “market portfolio” with the fads/overvaluations of the time. The passive investor held significant amounts of Tech stocks at high valuations in 1999, Japanese stocks in 1989, etc. How long before those positions become positive, real, absolute returns?

    I will also add that when you start and finish matters greatly. If we were looking at a stock market with a P/E multiple of 7 and interest rates at 15% (such as in 1982) I would be a much greater advocate for a passive, efficient, inexpensive approach.

    People can (and often do) much worse than indexing, I completely agree. The money management business is bloated and corrupt. I agree. People can do better. A focus on absolute returns as I’ve described them is how I approach it. I am talking my own book as that is what we do. I wouldn’t want it any other way.

  • http://www.valuerestorationproject.com JJ Abodeely

    Thanks. Sometimes it’s a bit like debating religion. Which is kind of pointless. People have to experience something to learn it for themselves (it takes 3 negative experiences for people to acknowledge something is hurting them). I will say that investment philosophy and values are so critical to success that you need to have alignment with your colleagues (whether that is to embracing indexing or embrace some other approach). That way your whole organization can live and breathe the values under which you operate. Good Luck.

  • Fred

    Abstract:
    Minimizing the classical definition of risk should be a counterintuitive venture as the explanatory nature of historical metrics’ construction challenges their ability to serve a predictive purpose on a non-stationary process. We uncover an ill conceived bias in these metrics and discover they provide a contrary indication to an investment’s survivability. The breakdown in the explanatory-predictive link is troubling and we aim to correct this via a better derived explanatory metric. The predictive variant of our metric will directly question the notion of optimization in order to serve the first priority of any continuous system, survival.
    http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1550519

    “I would add to this sentiment, particularly for equities as an asset class; there is actually more risk (chance of loss) when volatility is low, and less likelihood of loss when volatility is high.”
    This is due to the influence of Brownian Motion, and is also quite relevant for manager analysis.

    We peel back the layers of this argument even further and offer practical solutions in generating ex ante efficiency.
    http://amzn.com/1460933788

  • Anonymous

    load of crap. Bias, misrepresentation, false “common truths,” framed in terms of “trader mentality” (for example, the quote “…for periods AS LONG AS 10 years” wheras true long term investment lasts 30 to 50 years of accumulation, and then another 30-50 retirement drawdown.

    If everyone indexed, then there would be a significant reduction in both tails. That would be good, because the goal is not to make as much as possible. It is to guarantee that you will make enough to fund your retirement and not run out before you die.

  • Anonymous

    And 2008 was not a catastophe. Anyone who followed a standard rebalancing plan would be in significant profits today.

  • K50

    There is just so much here. First, modern portfolio theory in my mind does not include CAPM or ECM. It just simply says that we can use math programming to reduce the variability of portfolio returns. (I wrote my first mean-variance optimization program 40 years ago.) We’ve known since Richard Roll in 1977 that CAPM is not empirically testable. So how anyone computes an alpha strategy etc. is beyond me. This also questions all of those studies that says managers do not outperform some benchmark because they used a CAPM measure that we can’t compute in the first place because we don’t know the appropriate benchmark and then we are making linearity and normality assumptions. (In addition, most market index calculations involve continuous rebalancing without transaction costs. Continous rebalancing is very tough to beat when you are paying transaction costs.) BTW, Markowitz is on record as not believing in either CAPM or ECM so it is hard to say that MPT came out of ECM other than Markowitz graduated from Chicago.

    An alpha strategy to me is just simply trying to outperform a benchmark but when you cannot compute the alpha accurately you can’t say you are adding alpha, portable or otherwise. Don’t get me wrong – alpha strategies are pretty neat and you might be happy with the results, just don’t ask anyone to measure the results and conclude it worked or didn’t work.

    As Markowitz argues, portfolio theory is a normative theory and represents a recommendation of how to approach the decision. It is not an explanatory theory of financial markets or investment management. Having said that, optimization itself is explanatory (explains history) and rarely provides a very good prediction. It is actually too complex of an explanation. When we use portfolio theory we need to use predictive models not explanatory models and full covariance optimization does not have a very good forecasting record.

    As we have seen empirically, rarely do we see higher volatility – higher returns. (Put a monthly volatility chart and a monthly return chart together) In fact, empirically we see very little relationship between our risk measures and our return measures. I would even argue that it might be negative – so much for the CAPM-ECMH argument.

    Markowitz also recently provided a utility theory argument for geometric mean – semivariance as the appropriate portfolio theory model and I agree as I helped supply a lot of the empirical evidence over the past 3 decades in favor of using semivariance. (I published an article on using reward to semivariability ratio about a decade before Brian Rom/Frank Sortino decided to call it the Sortino ratio.) However, the first rule of investment is to not lose your investment. You can measure downside risk statistically and diversify against downside risk as much as you want but if you haven’t done your fundamental and macroeconomic analysis, semivariance is not going to protect your backside.

    Markowitz developed portfolio theory for individual stocks. Diversification works when correlations are low (0.30 and lower). I don’t know how much of a bang for the buck you get when correlations are over 0.60. Most asset classes are highly correlated in the first place and then whenever we have a crisis, correlations go to one. In the 2008 crisis, all asset classes dropped dramatically with the exception of gold and fixed income. Fixed income hung in there because Fed Funds went from over 5% to zero. (What happens when interest rates drop?) The problem is that the fixed income markets in the US froze up and without the massive Treasury/Fed bailout you might be using your fixed income securities as quaint bathroom wallpaper. Gold worked but for roughly 25 years before that it was a sucker bet. (The Brinson et al. asset diversification FAJ article was probably the worse finance article ever written as it talked people into doing only asset allocation. If you want to argue, tell me what their control group was. Hint: there wasn’t one.)

    A good old-fashioned top-down approach is best in my mind. First, do a solid business cycle analysis and decide which sectors will do best under current conditions. Sector rotate without regard to the sector weights in the S&P 500 – you are just herding if you watch the S&P weights. Next, given the business cycle, do some good fundamental security analysis to pick good solid stocks. If you want to do deep value, then don’t worry about the business cycle as we will go through the cycle several times before you derive the deep value. If you are staying within the current business cycle, then you have to look for relative value for the next 12-24 months. At this point, I would still want to use portfolio theory to build portfolios of individual investments. Asset class diversification to me means hold cash when things are going bad. Manage the portfolio to not lose your investment and know your investments. Investment management is hard work. (To me, that is what deep and relative value guys do.) Warren Buffett talks about his 19% annual return as “we beat the market on the downside and hang in there on the upside.” If you look at his record, he rarely beats the market on the upside.

    Don’t trust a rating agency that puts a AAA rating on a structured product that contains 100% subprime instruments. Diversification is not going to turn a sow’s ear into a silk purse. I can’t believe informed professional investors were buying those things from 2005 on. I saw a Business Week article in October 2006 that talked about explosive loans that took 100% junk rated loans and turned them into a CDO with 90% of the tranches investment grade (75% were AAA???). We didn’t buy any financials after that even though they were over 30% of the S&P 500.

    Doing business cycle analysis, security analysis, portfolio analysis, and actively managing your downside provides a lot of checks and balances that are required in a changing and complex environment. I think of the Level III materials in the CFA program as akin to going into Baskin-Robbins and finding all the flavors are plain vanilla.

    To put my views into context, I’m from the theory of “self-organized criticality” which tells us, that in complex and tightly coupled systems even tiny events can have massive and unpredictable effects. This means manage your investments as though tomorrow starts the next crisis.