Rational Investing

Age 55	Investment = Initial Benefit Base	$1M
Age 65	Minimum Benefit Base at Retirement	$1.5M = $1.0M + 10 × $50,000
Age 66 +	Guaranteed Minimum Yearly Income	$75,000 = $1.5M × 5 percent

The appeal of GLWBs is that the benefit base has the potential to increase faster than the minimum guarantee if the investment portfolio performs well. Therefore investors get security through the guaranteed minimum income and participation in the market through the potential increase in the benefit base. For example, the benefit base can be periodically reset (such as every year or every three years) at a higher level if the value of the portfolio net of withdrawals is above the current benefit base. Finally, because investors are offered a guaranteed income, conservative investors are often told they could tolerate riskier portfolios to target higher expected returns. For example, they could move from a 40/60 allocation to equity and fixed income to a 70/30, the most aggressive allocation often allowed in these products.¹

It all seems very attractive. However, let’s consider a few facts. On the one hand, an investment of $1,000,000 at age fifty-five leading to a minimum yearly income of $75,000 for life starting at age sixty-six represents an annualized internal rate of return of only 2.03 percent if you live until age eighty-five and capital has been depleted. The internal rate of return would increase to 2.86 percent if you live until ninety. If you live until ninety-five, it would be 3.40 percent. On the other hand, the total annual fees on GLWB contracts can be significantly higher than 3 percent when you add the fees related to investment products, guarantees, distribution, and marketing. The combination of high fees, cash withdrawals after retirement, a reset mechanism, and market uncertainty makes it difficult to forecast how the annual income could evolve beyond the minimum guaranteed level of $75,000. The income investors can expect is very much path dependent. In other words, it depends both on the cumulative performance of the portfolio and on how this performance is achieved over time because the reset of the benefit base occurs on specific dates at specific intervals. Investors in such products are usually provided with scenarios of expected yearly income under different financial conditions, but it is hard for most potential buyers to properly grasp the distribution of expected payoffs and the long-term impact of high fees on the evolution of the benefit base. Not surprisingly, the product documentation almost never explains how the participant would have done had he invested in less expensive investment alternatives over his lifetime.

Andrew Ang ² of Columbia Business School, now at Blackrock, refers to these products as complicated, expensive, and difficult for consumers to navigate. Others use more colorful language to describe them. Hence if you thought this example was complicated, it is because the product itself is truly complex, and yet it is sold to the average investor. Of course, investors in these products do not have the option of improving their investment skills over time. Buying such products is often done once and its consequences are experienced over a lifetime.

Many risk-averse investors understandably prefer to pay high fees to ensure a guaranteed level of income during retirement. Although the investment needs are fulfilled by these products, the decision to invest in them is difficult to make. You have to form an idea about the different possible market performances over a long term and the likelihood of each scenario, as well as understand how the payoff structure will interact in each case. Finally, you must judge whether these products remain a good deal against other alternatives given the fees charged. It’s complicated.

In the spring of 2010, a friend looking to invest in a similar product asked for advice. It was necessary to read the eighty-page documentation twice to fully understand the specific terms of the product and its fee structure. The most relevant problem was not to determine how the product would behave under different scenarios of market returns. Instead it was to determine if a cheaper and more conservative product not providing any guarantee was likely to outperform the proposed GLWB under most scenarios.

First, the analysis showed that the minimum payoff of this GLWB (up to the age of ninety-five) could be matched by investing 70 percent of the portfolio in a few government coupon bonds of different maturities and in one long-term fixed income ETF. The balance of the portfolio (30 percent) would be invested in equity ETFs. Second, the analysis concluded that the income generated by this low-cost portfolio would likely outperform the GLWB unless the average yearly equity performance was either above 12 percent or less than 2 percent for several decades. Thus it was unnecessary to make predictions about market returns or managers’ skills to reach the conclusion that the payoff structure of this particular product was not attractive. High fees cannot be the solution to the well-being of investors. They also make it harder to live up to financial promises because higher fees increase the likelihood that the minimum guarantee may be needed. Higher fees also reduce the income that could be expected beyond that guarantee. These comments also apply to many hedge funds and especially funds of funds that charge an extra layer of fees.

The structure of possible payoff in a game of dice is easy to compute. That of GLWBs is complex. The complex structure obscures that a similar expected payoff pattern can sometimes be achieved using a simpler portfolio approach managed in an Excel spreadsheet. It is our experience that many investors in GLWBs—and perhaps sellers of these products—did not consider or did not understand this possibility. This is a case of information asymmetry in which investors have far less relevant information than required to make an informed decision.

Complex investment processes can be intimidating, and many investors prefer to pay handsomely for advice, regardless of its usefulness. We are not saying that investing is simple or that advisory services are not worth their price. However, as will be discussed in chapter 3, only a small proportion of fund managers can outperform after fees by design and not by chance. Furthermore, investors should not have to pay high fees for investment solutions that are like target-date funds or standard balanced portfolios. There are index funds and ETFs or combinations of both that deliver effective and complete portfolio solutions for less than 0.5 percent in total expenses. They can be as balanced as dedicated portfolio solutions offered by large financial institutions and advisors.

Lussier remembers advising an entrepreneur who had $50 million in liquid assets and was earning $7 million in yearly cash flows from his business that

• He was paying at least twice as much as was reasonable in fees to his advisor. His portfolio had no distinctive advantage over what could be assembled from a few ETFs and specific securities. It already contained several ETFs, and it was under-diversified on the equity side.

• He would likely perform just as well (before fees) if he invested in a balanced portfolio of ETFs and spent two mornings each year rebalancing his portfolio.

Though the entrepreneur knew that he was overpaying for these services, it was difficult to convince him that a simple self-administered process would do as well as the unexceptional portfolio his manager had designed for him. This is far from a unique situation. Investing is intimidating even to entrepreneurs.

These examples illustrate the importance of understanding the basic structure of the market before making any decisions about the potential impact of expertise. They also illustrate that not all investments offer fair odds. If details of a specific situation seem too complicated and structurally unfavorable, it may be wise to simply stay out of it.

To Be or Not to Be Active?

Let’s make a deal. We will use one equation to simplify our discussion over the next few chapters. In exchange, we promise not to use any other equations until chapter 5. To some of you, this equation will introduce many of the concepts floating around the investment world: alpha versus beta, large bets versus diversification, risk premium versus idiosyncratic risk, etc. To other readers who are already investment experts, the equation will summarize all these concepts and structure our discussion on market structure in this chapter, on delegated management in the next chapter, on forecasting in chapter 4, and on sources of performance in chapter 5.

Let’s say you invest in an asset, say a stock, bond, mutual fund, hedge fund, commodity futures contract, etc. Whether you look at the value of your investment after one minute or after one year, your realized return R is the gain that you have realized through price appreciation and cash flows that you have received. For example, if you buy a stock for $100, sell it for $105 after one year, and receive a $5 dividend, then R = 10 percent. To better understand the sources of returns of any asset, we use the following representation for returns, which we will call our return equation:

R = R_f + β_m × F_m + β₂ × F₂ +…+ α + ε.

= Risk-Free Rate + Compensation for Risks

+ Mispricing + Noise/Luck

In this equation, the return R has several components. First is the risk-free rate of return R_f that is obtained by investing in a safe asset, for example, a short-term bond from a developed country’s government. We can always park our money in this safe asset and achieve a return equal to R_f. We are interested in the return in excess of R_f that riskier assets provide because the safe asset is the natural benchmark. Clearly, an asset that has an expected return of 5 percent is not as attractive when the risk-free rate is 8 percent than when it is 2 percent.

Second, the Fs are returns on factors that capture common—or systematic—sources of risk in the economy for which investors receive compensation, while the betas (βs) govern how this asset is exposed to them. The first of these factors, F_m, is the return in excess of the risk-free rate on the value-weighted market portfolio that contains all assets weighted by their market capitalization (or market value, hence the term value-weighted). It is called the market factor. For example, the S&P 500 is a representation of a value-weighted portfolio of large capitalization U.S. stocks. The market portfolio is simply a much broader representation of the securities market that incorporates all securities. The market obviously has a market beta of 1 with itself.

A higher β_m indicates that a security moves more in tandem with the market than another security with a lower β_m. For example, Walmart Stores (a safer stock) has a β_m of about 0.5, while Amazon (a riskier stock) has a β_m of about 1.5. Assuming, for example, that investors require a risk premium of 4 percent to be exposed to the market portfolio, the portion of return attributed to the market factor would be 0.5 × 4 percent = 2 percent for Walmart and 1.5 × 4 percent = 6 percent for Amazon.

However, a security with a lower β_m does not necessarily indicate a lower risk overall. A security with a low β_m could be sensitive to risks other than the overall market. Hence our representation for returns allows for the possibility that there may exist other common risk factors: F₂, F₃, etc. As will be discussed in chapter 5, it is now recognized that the compensation investors get for investing in an asset is linked to exposure to other risk factors than simply the market. Among the best-known empirical factors are size (a measure of the relative performance of small versus large stocks), value (the performance of stocks with a low price-to-accounting value ratio relative to stocks with a high ratio), and momentum (the relative performance of stocks that have performed well recently to stocks that have not).

Third, α + ε is the idiosyncratic or unsystematic part of the return on the asset, the portion of realized return that cannot be explained by the risk-free rate and exposure to risk factors. On the one hand, the alpha (α) is a constant, for example, a 2 percent excess return per year. A positive α implies a higher return every period, regardless of other factors, like a company or a hedge fund manager outperforming the market on average because his unique skills allow him to better identify mispriced securities. On the other hand, ε is a random shock that affects only this asset and no others. This shock can be favorable or unfavorable. As an example, think of a company announcing either disappointing or surprising earnings, not of broad movements in the market. Another example is the price impact of an independent study indicating the main drug sold by a pharmaceutical firm to manage sleep apnea either increases the probability of cancer or reduces it. The impact of ε on return (scale and sign) is simply unexpected. ε could also be called good or bad luck.

To obtain this equation, we are not making any unrealistic assumptions and we are not relying on an abstract economic model either. Instead it is a simple statistical way to describe returns. All the equation does is state that realized returns among different assets will differ because they have different exposures to risk factors, because they are not similarly mispriced (α), and because luck (ε) is a random and asset-specific phenomenon. Unfortunately, mispricings and noise hinder our ability to estimate the expected return required by investors in Amazon or Walmart.

We will use this return equation in various ways in this book. We now turn to the first use, which concerns one of the most enduring debates in finance: Should we be passive or active investors?

What Does It Mean to Be an Active Investor?

Should we try to pick stocks that will outperform the market? Or should we simply buy a low-cost index fund that invests in all stocks according to their relative importance in the market? Evidence shows that active investors do not outperform the market, and investors may be better off investing in passive portfolios. Yet passive investing remains an unappealing approach for many of us. Buying the market portfolio implies that we buy poor quality companies as well as good ones. Surely we can do better? Therefore let’s begin by examining the issues around being active and successful.

If we sum the returns R for all assets by weighting each of them by their market capitalization (e.g., the total value of all outstanding shares of a company), we get the return on the market portfolio R_m. At this time, Apple Inc. has a market capitalization of about 600 billion U.S. dollars and therefore contributes more to the return on the market portfolio than a company with a smaller market capitalization. By taking this weighted sum over all assets, we get our return equation for the entire market portfolio. But the market return in excess of the risk-free rate F_m is already on the right-hand side of the return equation. Therefore our return equation implies that for the entire market portfolio we necessarily have: α = 0, β_m = 1, and β₂ = 0, β₃ = 0, etc. More specifically,

R_m = R_f + F_m

This result may seem obvious, but it leads to a very important conclusion. It means that the value-weighted market portfolio of all assets has no alpha and no net exposure to risk factors other than itself, the market portfolio. We collectively own the market, and we obviously cannot collectively outperform ourselves.

A passive investor is allocated exactly as the market portfolio is, hence the term indexing. What does it mean to be an active investor? Being active means that we adopt a portfolio allocation that differs from the market capitalization-weighted portfolio. We can either take an active position on a stock because we believe it is mispriced (its alpha is different from zero), take an active position on a risk factor because we believe it will perform well, or both. In all cases, other investors need to take the opposite bets to ensure our collective position is that of the market portfolio. It cannot be otherwise because investors collectively are the market. Therefore if we ignore all fees (management, transaction, custody, etc.), the aggregate performance of all active investors will be equal to that of the market and to that of each passive investor. It also means that the aggregate portfolio of all active investors has the same sensitivity to risk factors as the market (β_m of 1 and no exposure to other factors) and no alpha (α = 0). In other words, if an investor has a greater value bias than the market (a positive beta on the value factor), another investor must have a growth bias (a negative beta on the value factor). These statements are independent of the investment horizon. Before fees, active management is what economists call a zero-sum game.

For example, let’s assume that the size of the U.S. large capitalization stock universe is $10 trillion and that 20 percent of the market ($2 trillion) is owned by indexed investors. If the entire market is up by 5 percent, the 20 percent invested by index managers must also be up by 5 percent. It then follows that the 80 percent ($8 trillion) invested by all active investors must, in aggregate, be up by 5 percent for the total market to be up by 5 percent. Whether the market is up by 5 percent in one week or in one year does not change this reality. Again, this arithmetic is independent of the horizon.

When an active investor decides to take a specific bet on a stock, he can overweight it in its portfolio compared to its weight in the market portfolio. If he beats the markets, then necessarily another investor who had underweighted the stock underperforms the market. Therefore if some active investors beat the market by $1, then some active investors underperform the market by the same amount because all investors as a group must perform as the market does. This is called the arithmetic of active management or equilibrium accounting; it is a direct consequence of simple arithmetic. It is structural, not a forecast.

Why should we care about the economics of active management? Because we need to be fully aware of the strong headwind we are facing if we decide to be active investors. To outperform the market portfolio, we need to find investments with non-zero alpha, determine if we should be over- or underexposed to risk factors other than the market portfolio, or both. But we need to do it at the expense of someone else. We all have good stories about how we managed to spot a good stock before it shot up, but whether we can do it on a consistent basis is what is important. Competition among investment professionals means that this is difficult. The fact that positive investment stories are disclosed more frequently than negative ones does not change this reality.

The economics of active management are also important in properly assessing a new investment strategy. Any strategy or portfolio solution that adopts weights that are not the market capitalization weights has to be understood through this perspective. To be profitable, somebody else has to take the other side of these deviations from the market portfolio and lose on average, either by being fully aware of it or not. We will come back to this point in chapter 5.

The Impact of Costs

Active investing is even more difficult than the previous section implies. Finding positive alpha and predicting risk factors is not a costless endeavor. There is a cost in acquiring information about different companies, predicting the economic landscape, etc. Alternatively, you can delegate the management of your portfolio, that is, hire a professional fund manager who will hunt for alpha for you. But this manager will face the same challenges as you. A professionally managed fund is itself an asset with a positive or negative alpha and exposure to different risk factors. A fund’s performance can also be analyzed through the lens provided by our return equation above.

Therefore when you consider the costs and fees involved in active management, our equilibrium accounting relation from the previous section becomes a negative-sum game. When you take a specific bet on one or many stocks, outperforming the market means that another investor underperforms. But it also means that both of your over/underperformances are reduced by the cost incurred. To be a successful investor, it is not enough to have expertise, a good understanding of investment and of financial markets, and good information. Your expertise, understanding, and information needs to be good enough to consistently beat other investors and cover the cost of acquiring this expertise and information. Similarly, if you hire an active manager to do it on your behalf, he needs to be good enough to consistently beat the market and cover its fees. Bottom line: If you want to be active and successful in the long run, you better be really good at it. Winning at the investment game in the long run is not about how much expertise you have, but how much more expertise and staying power you have than everyone else.

Let us reiterate one important point: This book is not about passive over active management. The goal is to understand the difficulties faced when you decide to be an active investor or hire an active fund manager. The poor performance of the average active individual investor and of the average active fund manager generated the popular advice that investors should invest in low-cost index funds instead. Yet not all of us can follow this advice. We need some investors to be active, do research, work hard to find information, and form opinions about different companies such that asset prices can form in the market. That low-cost index funds are a good deal for many of us investors is a consequence of market prices incorporating a lot of information about an investment. Surprisingly, this important point is often overlooked in the popular press, but it has been well known in the academic literature for more than three decades.³ If we all behaved as though the market portfolio was unbeatable and none of us spent any resources (time and money) on trying to beat the market, the market prices would not reflect any information and it would be easy to beat the market. Hence the market has to be inefficient enough to be attractive for talented investors to gather information and make good allocation decisions. In return, this search for outperformance makes the market relatively efficient and the passive index portfolio a good investment for those of us with no special information or investment expertise. It is a delicate balance.

The issue at stake is whether you have or can identify true investment expertise, and most importantly, how much you are paying for it. There is no doubt that a brand new entry-level Mercedes is a better car than a twenty-year-old Chevrolet. But it is not necessarily a good deal if you need to pay a million dollars for the Mercedes while the Chevrolet sells for a few hundred dollars. The debate on passive versus active management is about what the investor gets in the end, not about whether individual investors or fund managers have expertise or not. There is no doubt that some managers are truly skilled at picking good investments. The relevant question is whether a positive alpha mutual fund turns into a negative alpha fund after all fees and expenses are accounted for. Another question is how the value added is split between the skilled manager and his clients once competition among funds and market frictions are added to the mix. But let’s not get ahead of ourselves.

How Do Different Groups of Investors Fare?

The market is owned by different groups of investors (individuals, mutual funds, insurance companies, pension funds, hedge funds, endowment funds, and others). The distribution of equity ownership has significantly changed over the last sixty years or so. For example, individual investors used to own 92 percent of all equity in 1950. The current percentage is closer to 30 percent currently. In 1980, the share of equity owned by mutual funds was less than 3 percent, while it is closer to 23 percent today.⁴

It is possible that one group of investors is sufficiently talented to constantly beat and extract value from another group of investors. Think of mutual fund or hedge fund managers versus retail investors. Yet very little evidence exists to support such a case, especially when fees are considered. We will examine the decision to delegate and hire a professional fund manager in more detail in chapter 3, but here we look at their collective performance as a group to understand the arithmetic of active management.

Professors Eugene Fama of the Chicago Booth School of Business and Kenneth French of Dartmouth College recently conducted a study of the performance of all equity mutual fund active managers in the United States over more than two decades.⁵ Whether you look at the performance of all of their mutual funds as one big portfolio or the average performance of these managers, the same conclusion is reached. They collectively perform the same as the market index before fees and underperform the market index after fees. What this means is that they as a group do not consistently extract value from another group of investors. Obviously, we never invest in the average mutual fund manager and some managers in the lot certainly add value. Chapter 3 further explores this aspect.

Other studies provide similar evidence for different groups of professional managers. Taken as a group, there is no evidence of aggregate outperformance by institutional funds sold to public and private retirement plans, endowments, foundations, and multi-employer unions⁶ or by retail and institutional mutual funds that invest in international equities.⁷ On this issue, the consensus is strong.

The Subtleties in the Economics of Active Management

When thinking about the odds of being successful at the active management game, it is important to keep in mind the following issues.

The arithmetic of active management does not say that active investors don’t create value. Active investors, by gathering information about different assets to make investment decisions, choose how capital is allocated in an economy. Efficient capital allocation is crucial to economic development because it determines which projects are funded and at what cost, which contributes to improving our standards of living. The arithmetic of active management only says that active managers as a group cannot outperform the market and that for winners there are also losers as a result.

The arithmetic of active management does not say that half of active managers will outperform and half will underperform the market (even before fees). Actively managed funds vary widely in size and operate under various business models. Although there are approximately 8,750 open-end mutual funds in the United States and slightly more than 775 fund complexes (e.g., Fidelity and Franklin Templeton), twenty-five of these complexes control three-quarters of the assets, and the top 50 percent of funds, measured by size, control nearly 99 percent of all assets.⁸ Thus it would be a mistake to assume that necessarily half the funds should outperform or underperform because they have such different sizes and investment styles.

However, the evidence does show that once fees are taken into account, fewer than 30 percent of managers do outperform their benchmark over a horizon of ten years. For example, for the ten-year period ending in December 2014, 76.5 percent of U.S. domestic equity funds were outperformed by their benchmark. The percentage for international equity funds was 79.2 percent. Fixed income funds did not do any better.⁹

When evaluating whether an active investor or a fund manager is successful, we assume that they play a fair game. In a fair game, engagement rules are coherent and respected by all. In hockey, for example, each team has six players on the ice, assuming no penalty. No team would allow their opponent to have a seventh player while they have only six. Similarly, if the objective were to understand the structure of excess returns of active managers in the large capitalization universe, a fair game would require managers to select all of the securities in their portfolio from the same population of stocks and to use a benchmark whose structure is representative of their style (such as a value benchmark for a value manager). If managers define their investment universe differently, this would not be a fair game and it would be more difficult to draw inferences about managers’ expertise from their performance (such as a manager using the Russell 3000 as his universe of securities but the S&P 500 as his benchmark).

Are managers using appropriate benchmarks to measure their performance? Morningstar says its classification is based on the underlying securities in the portfolios and on specific statistics, but some funds may not fit perfectly in any specific category. Furthermore, the great multiplication of funds in the United States from about 450 in the mid-1970s to more than 8,750 today has certainly made classification more difficult. There have been several studies whose objective is to determine if mutual funds have attributes (characteristics, investment styles, risk and return measures) that are consistent with their stated mandate.

Studies found that between one-third and one-half of funds do not have attributes that are consistent with their stated mandates.¹⁰ This observation was confirmed by our own experience.¹¹ In the mid-2000s, we were involved in a process to select a fairly large number of external fund managers in the equity space to create multimanager portfolios covering four geographic zones. The initial database incorporated more than 1,500 managers. Using an internal process, this number was narrowed to less than 120 managers who were subjected to a greater due diligence over a period of nearly eight months. In the end, about twenty managers were selected.

Our process for the initial screening was unusual. We did not allow the individuals involved in this project to use the managers’ performances as a criterion. Recent historical performances can consciously or unconsciously impact the opinions of decision makers. However, our teams were allowed to use the performance data to determine if the actual styles of managers were consistent with their stated mandates. If it was not, they were eliminated. We found that more than one-third of all managers in our database had performance attributes that were inconsistent with their stated mandate, such as a manager in the large market capitalization segment with a significant sensitivity to small market capitalization firms.

Let’s consider another real example of two large North American pension funds. In recent years, these two organizations have internally structured global equity portfolios that emphasize value and quality tilts. One institution chose to evaluate the performance of their managers against a benchmark that is adjusted for several documented risk factors Fs. The other institution made the assumption that their process leads to a more risk-efficient portfolio than the global market, that this increased efficiency is not explained by other risk premiums, and has built a benchmark around a standard global equity index and cash. The first institution adjusted to the portfolio orientation by creating a more complete benchmark that incorporates several risk premiums, while the second simplistically assumed that there is a single risk premium and created a benchmark that managers are very likely to outperform in the long run. Thus inappropriate benchmarks may create the wrong impression that the zero-sum game argument is not valid. It can also lead to bigger managers’ bonuses than what is deserved. Finally, it may artificially inflate the percentage of managers that are reported to have outperformed their benchmarks.

These observations do not change our prior statement. If most investors pursue the objective of outperforming the market, active investing still remains a negative-sum game after fees. However, some empirical observations could theoretically appear to contradict this statement if managers were using inappropriate benchmarks.

Concluding Remarks

To outperform the market portfolio, you need to be an active investor, which implies that you need to look for non-zero alpha assets and expose yourself to risk factors other than the market portfolio. To be consistently successful, you need to be better than many other investors that are playing the same game and also be good enough to cover your costs. You may also need to avoid bad luck, especially when the investment horizon is short. Intense competition among investors says that achieving such a feat is undoubtedly hard.

Why don’t we simply delegate this task to specialized investment professionals? The economics of active management does not preclude one group of investors from consistently beating another group, although this possibility is not clearly supported by the evidence. While the same evidence also indicates that investment professionals as a group do not consistently outperform and that less than 30 percent of active managers outperform in the long run, surely it is possible to identify some of the most successful managers among them. The next chapter deals with this question.