Investment Management: What Really Matters
Many prospective clients ask us about how we think about investments and portfolio management. There are three broad approaches an investor might consider. The first approach, which we’ll call the Conventional Approach, usually involves a smart person (the portfolio manager) with a crystal ball in one hand and stack of analysis in the other. The underlying assumption is that with enough brainpower and a clear enough crystal ball, the manager can outsmart the market, making more money for investors in the process. Intuitively, this has a lot of appeal and would seem to make sense; unfortunately, the evidence doesn’t support this. Over time, fewer than 20% of these conventional (or “active”) portfolio managers outperform their respective benchmarks because the effort to beat the market costs more than it yields in extra returns. So, why not just look back at the last 5 or 10 years and see who did well and then use only those managers? While intuitively one would expect this to be helpful, again, the evidence disappoints: how well a manager has done over the last 5 or 10 years tells us next-to-nothing about how he or she is likely to do over the next 5 or 10 years.
The second approach, an Index-based Approach, is almost always a better choice than the Conventional Approach. With indexing, portfolios are constructed to be consistent with (or representative of) some established list of investment holdings, that list usually being maintained by an independent organization. Examples of such indices include the S&P 500, a list of 500 stocks maintained by Standard & Poor’s that is representative of the broad US large company stock universe, or the Barclay’s Aggregate Bond Index, a list of government and corporate bonds representative of the US bond universe. The primary benefit of this approach is its low cost: many index funds have expenses that are a fraction of conventionally managed funds, and as stated above, over time, index funds outperform similar style actively managed funds about 80% of the time.
So, at opposite ends of the spectrum, we have two broad approaches to investing: conventional (or active) at the one extreme, and Indexing (or passive) at the other. But, is there a better alternative? Let’s start with a short analogy to look at these two approaches. Imagine a community basketball league where you’re going to recruit teams from a pool of 1,200 eligible 13 to 15-year olds. In the conventional (active) approach, the coach might try to scout all 1,200 players individually. This might be very effective if you had unlimited time and were the only coach doing this, but this effort would take A LOT of time. And, if each of the other coaches in the league were doing the same thing, all competing for the best players, it might not be that much of an advantage in the end due to the competition for the best players (e.g., one coach would get some of the best players, but so would the other coaches). So, the process of scouting 1200 individual players would be expensive and competitive, probably yielding so-so relative results if all of the coaches were doing the same thing (as they are in the investment world). That’s active management in a nutshell: because it’s so competitive, it costs a fair amount – in both time and money – to get a competitive result and these costs end up driving down the absolute result (i.e., the returns an investor ultimately receives after expenses).
Sticking with our basketball analogy, how would an index-based approach work? In this case, we might choose a team (or teams) by applying some simple rule to select 12 players who are, on average, representative of the community – for example, every 100th eligible player from the all-players list – or perhaps we might choose one player from each of the 12 neighborhoods in the community. This approach has the advantage of building a team that will be, on average, fairly representative of the community AND it’s very inexpensive from a time and effort stand point…you just look at the list and select every 100th player. This will get you an average team in terms of absolute skill, given the players in the community, but do so at a very low cost in terms of time and effort. That’s index investing in a nutshell, and despite the use of the word “average”, it’s worth keeping in mind that this result still beats 80% of actively managed funds.
But is there a better way to produce a team that is still low cost / low effort, but also likely to be more effective than just selecting every 100th player (i.e., “indexing” the team)? Let’s assume we know several details about each potential player in our league: age, address, height, speed in the 100-yard dash, how much time they’ve spent on the court, and so on. What if we knew that, on average, taller players will be better on the court than the shorter players? And that the faster players are more likely to be better than the slower ones? Not that every taller player will be better than every shorter player, or that every faster player will be better than every slower player, but that on average, all other things being equal. So, instead of selecting 12 generally representative players for our team, we now randomly select from, say, the tallest 20% of candidates, or the fastest 20% of candidates – or even the intersection of the two – the tallest and the fastest? Wouldn’t this be likely to produce a better outcome than just randomly selecting 12 players – AND with much less effort (i.e., cost) than trying to scout all 1200 candidates individually?
This is the middle ground between conventional, active management and purely passive index investing: giving up the costly effort of trying to outthink the market and pick which companies will be the big winners, while also not settling for an indexed list of holdings that ignores strong academic research about the characteristics which historically have added to returns. These investment characteristics, or dimensions as we’ll call them, offer important information about expected returns – and offer that information at a relatively low cost. This evidence-based approach, which focuses on the dimensions of expected investment returns, is grounded in strong academic research and is the low-cost middle ground between the expensive conventional approach and a purely passive approach. This is the approach we use to manage client portfolios.
So, what are these dimensions of expected return? 2013 Nobel prize winning economist, Eugene Fama, from the University of Chicago and Kenneth French, Professor of Finance at Dartmouth, have shown that most long-term investment returns come from just three factors. These factors are equity market exposure (or “beta”), and then – within the equity portion of the portfolio, small size exposure and value exposure.
Equity Market Exposure: Imagine a risk-free, very conservative investment – let’s say the one-month US Treasury bill. As an investor moves out of that conservative or risk-free portfolio into the more volatile equity markets at-large, he or she could expect – over a long enough period of time – to see both portfolio risk and expected investment returns increase accordingly. So, the first dimension of investment return is broad equity market exposure. This is probably not a terribly big surprise to many investors but it is, without question, the most significant dimension in the construction of any portfolio.
Small Size Exposure: Once an investor has exposure to the equity markets at large, the next dimension of returns to consider is company size exposure. In a 1981 paper published in The Journal of Financial Economics, University of Chicago professor Rolf Banz showed there is a higher expected return for owning smaller companies as a class relative to the broader market at-large. To illustrate, let’s assume our hypothetical investor were to annually sort all available US stocks from largest to smallest (as measured by the total value of each company’s outstanding stock or its market capitalization), and then each year from 1926 onward, have owned only the smallest 20% of those stocks. Our hypothetical investor would again have seen an increase volatility – but also a significant increase in investment returns for having held the smaller stocks. This size exposure is our second dimension of expected returns.
Value Exposure: The next dimension of expected returns is value exposure. In 1992, the Journal of Finance published a paper by Kenneth French of Dartmouth and Nobel laureate Eugene Fama titled “The Cross Section of Expected Stock Returns”. Like the 1981 Banz size paper, the 1992 Fama-French paper demonstrated higher expected returns for owning lower-priced, out of favor “value” type companies relative to the broader market at-large. Using a similar annual re-sorting and re-selecting process to that Banz had used – but this time with low price-to-book value companies, Fama and French illustrated significantly higher long term returns for having owned value-oriented holdings. This is the third dimension of expected returns.
The common theme for each of these three dimensions is that by increasing exposure to one of these factors (e.g., equity market exposure, small size exposure, or value exposure), an investor would have increased expected returns over the long run. Another way of saying this would be to say that the investor was paying for higher potential returns with higher risk.
Since the 1992 publication of their value paper, other such dimensions have been vetted, some of which seem to be risk-based and others which may be behavioral in nature. Two additional factors worth mentioning are momentum and profitability. Momentum (Jagadeesh and Titman, 1993; Carhart, 1997) is the effect whereby holdings (e.g., stocks, markets, asset classes, even entire economies) that have done well in the recent past seem to be a bit more likely than not to continue to do well in the near future. Profitability (Novy-Marx, 2012) captures the effect where stocks of companies with higher-than-average past profitability (operating profitability less interest expense) have outperformed stocks of companies with lower profitability. The unified framework based on these first three dimensions (market, size and value) came to be known in academic parlance as the Fama-French 3-Factor Model, recently updated to be a 4-factor model to reflect the inclusion of Profitability as a fourth primary factor.
Why Should I Diversify?
An often-asked question with the factor models model is this: ‘If small, value-y, stocks with high profitability have the highest historical and potential returns over time, why not just put all of my money into those categories?’
The answer is two-fold. First, while these factors have been shown to have increased returns over long periods of time, there can be – and have been – very long stretches lasting 5 years, 10 years, or longer when a single asset class or dimension significantly underperforms its long-term historical average. Experience has shown that very few investors are able to patiently sit through a 10-year stretch of significant underperformance of any investment approach when all of their money was invested in that one approach – even if we think that approach may offer the highest potential return over the long haul. The second reason for not making huge single-category or single-factor bets is that in any one category or dimension, there can be significant volatility in both directions – up and down. By blending together several types of investments that have low correlation to each other, an investor can help minimize the overall volatility of a portfolio without sacrificing potential returns. So, for these two reasons, diversification should be part of the overall strategy.
So, what is “the right” balance?
Let’s start with this acknowledgement: there is no perfect balance. The structure of a reasonable and diversified portfolio of asset classes is as much an art as a science. However, there are portfolios that offer higher return potential depending on their exposure to the different investment dimensions discussed above. There are also geographic factors to consider (i.e., international diversification), which by themselves may not offer a potentially higher return, but which do help to diversify a portfolio and thus reduce its expected volatility. One study supporting the inclusion of international assets in a diversified portfolio was published in the Journal of Investing (David Lester, 1988). This study concluded that adding international holdings of up to 40% of an overall portfolio actually decreased total portfolio volatility, despite the international holdings being somewhat more volatile by themselves.
Considering the above, our general construct rules start with a healthy, but not over-weight, exposure to international and emerging markets, as well as small company holdings (both US and international) and a decided tilt toward value-oriented investments. This approach seeks a balanced weighting toward the important size and value factors, while incorporating meaningful international diversification.
How to Implement?
If the critical decisions of portfolio construction are the four dimensions of market exposure, size, value and profitability, then how do we actually implement such a portfolio strategy? Going back to the beginning of this discussion, the preferred answer would clearly NOT be to use conventional and actively-managed funds, as doing so has been shown to be unreliable at best and almost always comes with higher costs, both in fees and in taxes.
The remaining options are true index funds (open-end mutual funds or exchange traded funds), which are very low cost but also strictly tied to established indices, or quantitatively structured funds that seek better exposure to the evidence-based dimensions of stock returns (especially the size, value and profitability dimensions). Many of these options are very low cost, although generally not quite as low cost as pure index funds.
We manage portfolios which make use of both true index funds (including Vanguard, iShares, and other low-cost providers) as well as factor funds, especially those of Dimensional Fund Advisors (DFA), a pioneer in evidence-based investing strategies. All of these companies offer excellent choices. Well-diversified, low cost portfolios can be built using products from any of these companies. At the margins, DFA has the strong appeal of an unparalleled commitment to implementing evidence-based strategies, and in many cases, they tend to have better, more focused exposure to the dimensions of expected stock returns than the alternatives.
A good illustration of this is DFA’s Small Cap Value fund (DFSVX) vs. Vanguard’s Small Cap Value Index (VISVX), which is based on the value component of the Russell 2000 index. Both would be respectable choices for a US small cap value type holding, but DFA’s fund targets much smaller companies (an average company size of $1.5 billion vs. Vanguard’s $3.3 billion) and has much more of a value tilt, with a lower price-to-book (P/B) value ratio of 1.2 vs. Vanguard’s 1.8). If there are, in fact, small cap and value premiums, as the evidence suggests, the DFA fund will likely capture more of those premiums, and it’s likely to provide better diversification relative to large company holdings as it’s further away from them on both the size and value spectrums. DFA also has strong affiliations with some of the leading minds in finance-related academia. Its director of research is 2013 Nobel Prize winning economist Eugene Fama, and three of its current or former board members (Myron Scholes, Merton Miller, and Robert Merton) are also Nobel laureates in the field of economics. Because of the company’s strong ties to academia, their funds tend to pursue a pure implementation of the most current and actionable research in the field of investment finance.
That all said, we do manage portfolios for some clients which are built on non-DFA holdings (usually Vanguard and /or iShare Exchange Traded funds). As noted above, all offer excellent choices and we’re not philosophically (or economically) bound to any particular investment company.
In closing, it’s worth noting this: while how an investment portfolio is structured is very important to long term results, perhaps of even greater importance is how a portfolio owner reacts (or doesn’t) to market volatility. While academic evidence suggests that up to 95% of portfolio returns are driven by allocation decisions like asset class and factor exposure, in the real world, 95% of ultimate investor returns are probably driven by investor behavior. Therefore, what is really important in the end is to select an appropriate allocation, one that you can live with, in both good times and bad, and then do just that – live with it! This is part of what we help clients do.