My last post concluded with the remark: „actively adapting asset allocation to address changing market conditions is the best practical method in the real world.“
This conclusion stems from analyzing the behavior gap evident in the Dalbar study, which leaves the average fund investor with only about 50% of the returns of a passive index in his pocket. The easy solution of simply becoming a passive investor unfortunately doesn’t fly, as many of these underperforming investors actually are passive indexers – at least until they face a major drawdown and panic. They sell their holdings near the bottom and reinvest too late, which causes the gap. From my research and personal experience it becomes clear, that the greatest challenge in investing lies in finding an investment approach that you can stick to and that prevents you from falling into behavioral traps – especially during bad and also heady times. Everyone seems to fall prey to performance chasing and liquidating in a panic – at least to some degree.
Two approaches to avoid that come up again and again:
- Hire a financial planner to act as a behavioral coach, consciously settle for average returns and shell out for the fees. Trying to find an outperforming manager most likely would result in performance chasing and lead to a similar outcome as detailed above – I think you would be better of managing your money yourself as described below. Similarily a low- fee Robo-Advisor looks great on paper, but lacks the behavioral coaching that, in my mind, is the main function of an advisor. For a low maintenance approach I would still recommend that route, because of the low fees. Combined with a written investment statement which states the course you want to stick to through thick and thin to serve as your personal behavior coach, this can be a good solution.
- Be active and employ a dynamic asset allocation that concentrates on reducing drawdowns and volatility to your tolerance level in any market environment.
The key question is: how is successfully adapting an asset allocation to market conditions possible and can it yield higher risk- adjusted returns than a static portfolio?
Intelligent students of financial markets agree that market regimes (like most things) move through cycles, but there is a strong division between people that believe it is impossible to time those cycles and those who think there are ways to do just that – both on the academic and on the practitioners side.
I think, elements of the right answer can be found in both camps´ point of view and, I believe, in the middle ground lies opportunity to successfully adjust your portfolio to different market regimes, if you keep an open mind.
I agree with the idea, that you cannot usefully (as in “make money with it”) predict the future in any consistent fashion – you would simply have to be exactly right at the right time way too often. The vast amount of predictions and explanations for market movements touted daily by pundits everywhere are nothing but noise – useless, but extremely difficult to ignore. As Jason Zweig says: „our brains are wired to force us into forecasting; it is a biological imperative. In fact, humans are born with what I’ve come to call ‘the prediction addiction.‘“
Rather than trying to forecast market movements, my philosophy is grounded in the idea, that market regimes are persistent enough to draw profitable conclusions from analyzing which state the market is in at the moment. Several market anomalies point to that conclusion: value and relative momentum effects, time series momentum (trend) and volatility clustering. In the second part of this article I will analyze the possibilities to successfully use these anomalies in a dynamic portfolio allocation.
As I think that it is imperative to build a clearly defined investment philosophy that you can really trust and believe in (the exact concepts it embraces are secondary as long as they are solidly evidence-based), I first want to point to and comment on several valuable resources from practitioners that are founded in academic research. These approaches form a major part of today´s investment landscape.
1 Passive buy and hold- index investors for example Index Fund Advisors.
Passive investors think it is impossible to actively manage your portfolio allocation with success, apart from periodic rebalancing. This is based on Eugene Fama´s Efficient Markets Hypothesis (EMH) and related theories from the 1950´s to 70´s which they adhere to religiously. They often do employ factors from the original Fama five- factor model – which in practice does mean using a form of active market timing to me (for example using a value factor means buying stocks at a time when they are cheap). As my own ideas embrace contrary views to the buy and hold approach as, I see this investment philosophy as quite restricting and narrow minded, but you have to be aware that this is still a prevalent conviction in the investment industry. Valuable insights from this investment philosophy are:
- Emotion- based investing is to be avoided; disciplined, evidence- based, quantitative investing to be embraced. The most significant problem for active investors is not relying on empirical evidence in selecting investments.
- Qualitative stock picking is a fools game.
- Predicting the future doesn’t work as numerous studies on expert forecasting show.
- Behavioral mistakes are the downfall of the individual investor and the role of a passive advisor is to curb those.
Unfortunately many academic insights after 1990 seem to be completely ignored as they challenge a set of assumptions EMH is built on (the random walk and standard distribution of security prices; static expected returns for assets; rational investors etc.). Even the founders of these theories acknowledge this – a path that is described succinctly in the beginning of this whitepaper.
A good example of omission is the conspicuous absence of any mention of momentum effects in the passive approach, an anomaly that is the subject of hundreds of academic papers and that has many implications. Even Eugene Fama states: „Extreme losers continue to lose and extreme winners continue to win. There’s a lot of speculation about why that happens……..But in my view that’s the biggest challenge to market efficiency. Momentum is the premier anomaly.”
Another strong and very common argument, that concludes that market timing is virtually impossible, is “missing the best and worst days“. Market timing would have to be incredibly accurate as less than the best 1% of days account for all the market´s gains. Missing both the best and worst days on the other hand would improve returns. The simple solution lies in analyzing market regimes: 60%-80% of the best and worst days can be pinpointed (and eliminated) in markets that are already downtrending which causes volatility to rise with abnormaly large daily losses and gains. (Table from Meb Faber´s paper)
2 Taking the best ideas of both the efficient market proponents and critics. The 2013 Nobel prize to both Eugene Fama and Robert Shiller symbolizes this and is analyzed and commented in an article by hedge fund AQR´s founder Cliff Asness. AQR´s success is founded largely on the market- neutral implementation of the classic Fama- French factors (primarily value) plus momentum as an additional factor.
The great value in their approach is utilizing the pure factors, that are quite uncorrelated to the market as a whole. This is done by building leveraged long- short portfolios of different factors, which makes it hard to implement for an individual investor as it requires frequent rebalancing and shorting a large number of individual securities and high leverage (around 8x). I am researching ways to do this and for the moment actively employ factor tilts in my long- only portfolio. The closest approximation to a pure momentum factor is the systematic structure of my short options portfolio.
The powerful idea of combining value and momentum, studying them together as a system, stems from AQR´s research. A well-constructed value strategy diversifies momentum – a combination strategy of the two is far better than either alone. They are more than the sum of their parts, because their negative correlation to each other yields abnormally high risk- adjusted returns. An explanation of this blends a fundamental (value) and a technical (momentum) approach as two sides of the same coin: a medium term technical overreaction in the market is followed by a long term mean reversion towards fundamental value.
This has become a core component of my investment strategy.
AQR states that factor timing is hard, if not impossible to do. Arguing about this with Rob Arnott of Research Affiliates has become a hot topic with factor investors recently. I think that factors can be successfully integrated in a market regime framework and that it is a good idea to do so when following a long- only approach.
One of Arnott´s most interesting idea is the concept of „overrebalancing“ , which can be applied to asset allocation as well as factors. Instead of periodically rebalancing a portfolio to a set allocation by selling the good performers and buying the bad ones, one overshoots the target allocation somewhat – this uses the powerful effect of mean- reversion and functions as a behavioral hack against performance chasing.
3 Dynamic asset allocation models in practice. The concept of these models is to add downside protection to achieve less volatility and lower drawdowns. Even though this idea has the power to significantly enhance risk- adjusted returns, its real value lies in protecting the investor from his own behavioral mistakes – making it possible to actually earn those theoretical returns. Tolerable levels of drawdown are achieved by reducing portfolio exposure in bad times in the market. A major tool used by different practitioners is time- series momentum or trend- following. Similar models with detailed background and performance histories are provided by Meb Faber, Alpha Architects, Blue Sky Asset Management, ReSolve, Gary Antonacci and others – these are among the most useful practical papers to me, because they blend the best ideas from different investment ideologies.
Dynamic allocation models depend on identifying different market regimes to work – the subject of part II of this article.
To recap, these building blocks form the blueprint of my portfolio model:
- 1 risk premia of different asset classes
- 2 additional returns through factors
- 3 systematic downside protection enhances risk- adjusted returns