| IN THE ARCHIVES: November 2011 September 2011 July 2011 May 2011 March 2011 January 2011 More... |
 January 2012 The Search for Forward-Looking Factors The Past is All We Have
This isn't to say investors don't need to think about current and future conditions when deciding where and if to invest their hard-earned money. That's really the essence of investing. What investors can't do, however, is see into the future. Short of a reliable crystal ball, there's no way to do that. Rather than search for such magic or rely on pundits who claim to possess it, investors would be better off taking a survey of current market and economic conditions, and then examining how their prospective investments have performed in similar past situations. This is essentially a quantitative approach based on the assumption assets will behave similarly in like conditions. This doesn't mean conditions are ever completely identical to those in the past or that relations between alternative investments don't change over time. Nothing's ever completely identical and adjustments have to be made. Investors who master this are the ones that make better choices or "predictions", it's not because of some sort of magical looking glass into the future. Interestingly, quantitative investors are often faulted for focusing on the past when making decisions for the future. Ever since the first lump of gold was used as currency, there have always been those who sought "forward looking" measures rather than those derived from the past. Initially that makes sense because if there are certain factors that lead to future success, they should work regardless of what happened in the past. Unfortunately, what often sounds reasonable in theory isn't possible to apply in practice.
A Recent Example Morningstar is the leading provider of mutual fund data and information. You're undoubtedly familiar with their "star ranking' in which their analysts rate funds from 1 (worst) to 5 (best). The stars have become so ingrained with the investing public, many advisors won't even suggest funds with fewer than three 3-4 stars. Retail customers especially have come to rely on them as a key, if not the key, reason to purchase a mutual fund. Despite the overwhelming success of the star ratings, Morningstar sees them as only quantitative and backward-looking: The Morningstar Rating, most commonly referred to as the “star rating,” is a purely quantitative, backward-looking measure of past performance. It is based on a fund’s risk- and cost-adjusted performance over three-, five-, and 10-year periods and helps investors to quickly and easily assess a fund’s track record relative to its peers. [This and all subsequent references come from the Morningstar FAQ's released in November 2011]. It’s always interesting to see Morningstar deny the predictive ability of the star rating. They often do when questioned about the generally mediocre return of four- and five-star funds. Nevertheless, there’s never been a concerted effort to correct investors’ mistaken belief that star ratings are predictive of future performance. That is, of course, until now as Morningstar rolls new "Analyst Ratings". Unlike the star ratings, the Analyst Rating is a qualitative, forward-looking measure, based on analyst research...considering both numeric as well as analyst-driven factors. This approach notably puts only partial weight on past performance and backward-looking risk measures and does not dismiss funds that have underperformed or have limited track records. It will also be more responsive to significant changes at a fund or parent organization. Notice there are two important differences between the Analyst Ratings and the old stars. First, as already pointed out, the Analyst Ratings are expected to be forward-looking. The fact that "past performance is no guarantee of future results" won't hinder them as it would, presumably, affect the stars. Secondly, the Analyst ratings aren't strictly quantitative like the stars. Because of this, they can look beyond the numbers at qualitative factors as well. Morningstar's analysts do this by evaluating each fund on each of five key "pillars":
The first thing that jumps out is the fact that this list isn't purely qualitative. Indeed, Pillars 5 and 6 are purely quantitative. In fact, they're considered in the star ratings so these two pillars aren't the distinguishing forward-looking qualitative factors. The secret must lie in the other three.
Forward vs. Backward Looking Maybe this pillar is based on the assumption that teams that work well together tend to stay together and provide a consistent approach. But doesn't the quantitative measure known as “manager tenure” also address this? Not only that, doesn't it have the benefit of being quantifiable? What about the “Alignment of interests with fund shareholders”? This is usually evaluated by reviewing compensation schemes and ultimately the fund’s expense ratio – both highly quantitative factors. It's starting to look like that qualitative insight must be in the remaining two pillars. Pillar 2 is Process. Is there a clearly defined repeatable approach to security selection and portfolio construction? Again, this is often emphasized by analysts seeking qualitative information from investment managers. Essentially it’s an effort to determine if performance up to this point has been the result of skill or luck. Generally the best way to evaluate this is to compare what the manager says he or she does to what he or she actually did.
Most investment managers can provide a written security selection and portfolio construction policy. The best ones also have a documented "sell" policy, too. Speaking from experience in creating these documents, the goal is to make them sound as specific as possible while not limiting the managers’ options. This results in general statements that do little to cast much light on the actual investment process – particularly in turbulent markets. Arguably, the best way to evaluate this is to looking back at how the fund’s holdings and portfolio actually changed during periods of market stress. This, by the way, also helps assess “whether the portfolio is consistent with the stated process” and is – you guessed it – a quantitative factor. That leaves the final pillar, Parent. This one’s a real head-scratcher for quants. It’s hardly arguable that you don’t want to place money in a fund with a parent on the verge of going under, but does this really take a major investigation? Presumably the “capacity and risk management, recruitment and retention of talent” refers to administrative staff because these factors were already considered in the other pillars for the investment team. How’s that helpful? Incentive pay should also be covered in Price and People pillars. Finally, the “culture of stewardship”: How does that possibly help the fund investor? Does the highly rated parent firm send excess profits to the fund as a charitable contribution? At this point it’s hard to see what benefits the qualitative factors in the Analyst Ratings add over purely quantitative approaches such as the traditional star rating. Perhaps more puzzling is why the former is expected to be more predictive than the latter? On the contrary, one might reasonably conclude one is not more predictive than the other because neither really is.
The Real Issue This is a philosophical issue that dates back to Plato and Aristotle. It's the difference between deduction and induction, and it's the principle that defines the scientific method. There are few things we can know with complete certainty, the rest we simply simply observe and infer connections such as cause and effect. We think a high level of Factor X means an investment will provide outstanding returns because we've noticed that's what's happened in the past. Put differently, if we hadn't noticed it in the past, why would we expect Factor X to be predictive at all? Unless you're truly an old fashioned stock picker -- wiling to go with your gut feeling without knowing or caring why your gut feeling is what it is -- then you're relying on past observations. As you've probably guessed by now, this blurs the line between qualitative and quantitative measures. Even those that seem most qualitative (such as the "feel" for the investment team in the example above) ultimately rely on past observation. If you're still not convinced, take a few minutes to come up with a "predictive" factor that is not based at all on past observation. Good luck with that. So although past performance may is not a guarantee of future results, it's still one of the best measures we have to go on. Quants may not be so misguided after all. Search this site! Just enter you key word or words:
Get current quotes or follow your own custom portfolio,
courtesy of E-Line Financials:
|
||||||||||||||||
HIS IS THAT TIME OF YEAR when market pundits flood the media with their predictions/stock picks/market projections/guesses/whatever for the upcoming year. Some will be right, most will be wrong. Those that are right are often more a result of luck than skill. The vast majority aren't worth the time spent reading them. Most will be forgotten by year end which tells you how unimportant they are. It's also why the pundits who were so wrong last year will be writing new predictions next year.
