Men In Black II, Ocean’s Twelve, & The Hangover, Part II... all of these movies have one thing in common: they all were abysmal sequels to blockbuster movies. We long to regenerate scenarios when everything comes together perfectly and the stars align, but that kind of success is rarely duplicated. In the world of money managers, success means blockbuster performance… every year! Fund managers who are successful in the short term are considered the current financial heroes, despite the fact that every reputable study of mutual fund performance over the past 30 years has found there is no reliable way to know if managers with recent winning performance will win in the future. This is why some variation of the disclaimer "past performance is no guarantee of future results" must appear in all mutual fund advertisements and prospectuses. Even still, unwitting investors chase recent performance, and the dangerous practice of manager picking ensues.
Sometimes managers can duplicate their success a few years in a row, but it just doesn’t last. As hard as it is to duplicate success in the film world, it is even more difficult for these all-star money managers to duplicate their past success.
"Most investors follow the crowd down the path to comfortable mediocrity," says David Swensen in Pioneering Portfolio Management.1 Anxious to capture the gains that come with a winning mutual fund manager, manager pickers blindly chase a hot performing fund manager's track record, failing to realize their odds for future success have vastly diminished.
Figure 5-1 shows the results of a study using Morningstar data reflecting the performance of active fund managers for the 15 years from 1998 to 2013. The chart depicts how an average of only 8 funds of the top 100 fund managers repeated their performance the following year. In the years 1999 to 2000, 2007 to 2008, and 2008 to 2009, none of them repeated their previous year's top 100 performance.
Pick Your Manager
Variations in manager performance are a function of luck and the random rotation of the style of their fund. When a particular manager's investment style is rewarded by the market, that manager is often credited with skill. As market conditions change, however, so does the performance of fund managers. Figures 5-2 and 5-3 track the rankings of the top 10 mutual fund managers in a given year and subsequent time periods. These charts reveal how quickly a "top" fund manager can slide to the bottom. For example, Figure 5-3 shows that Dynamic Gold & Precious Metals I had the highest performance out of 6,446 mutual funds in 2010. In 2011, however, the fund slipped to 6,389th place, landed in 6,648th place in 2012, and ominously, no data is available for 2013. The data contained in these two figures reveal many other examples of fund performance that sharply declined.
Top-performing funds have failed to maintain their position throughout a meaningful subsequent period. As Bob Dylan said, "the first ones now will later be last, for the times they are a changin."2
An analysis of the Morningstar database of 231 mutual funds with 10 years of returns is shown in Figure 5-4. The top graph shows the performance rankings of these 231 funds from best to worst (left to right) for the first 5-year period from 2004 to 2008. Then the same order of fund rankings is maintained in the bottom graph in order to see if fund performance was repeated in the years 2009 to 2013. Based on the above studies, it should come as no surprise that many of the managers who outperformed their peers in the first 5-year period did not do so in the second 5-year period, and vice versa.
Another tracking mechanism that can cause confusion is the reporting of mutual fund returns, often inflated when compared to actual long-term returns. The discrepancy arises from neglecting to account for funds that have closed or merged, resulting in the higher average returns of only surviving funds included in calculations. When funds go under, their records are stricken from databases, creating a survivorship bias. This bias inflates the remaining funds' average returns by 21%, according to CRSP data cited by John Bogle3. A recent Morningstar study found that for the ten years ending March 31, 2014, only 53% of actively managed funds across all categories survived the period4.
Even large institutions and pension plans chase performance, much to their detriment. A study conducted by Amit Goyal of Emory University and Sunil Wahal of Arizona State University found that manager hiring and firing decisions made by consultants, board members and trustees were a waste of time and money.
The study, "The Selection and Termination of Investment Management Firms by Plan Sponsors,"5 reveals the negative impact of manager picking. Goyal and Wahal analyzed hiring and firing decisions made by approximately 3,700 plan sponsors, representing public and corporate pension plans, unions, foundations, and endowments. Figure 5-5 shows the results of hiring 8,755 managers over a 10-year period from 1994 through 2003. Note that investment manager performance is measured by average annualized excess returns over a benchmark. The chart illustrates that managers that were hired had outperformed their benchmarks by 2.91% over the three years before being hired. However, over the following three years the managers on average underperformed their benchmarks by 0.47% per year when adjusted for management fees and transition costs. Plan sponsors often proceeded to fire managers who had underperformed in favor of other recent top performers, only to repeat the cycle again. The study concluded, "In light of such large transaction costs and positive opportunity costs, our results suggest that the termination and selection of investment managers is an exercise that is costly to plan beneficiaries."
Using data from the same study by Goyal and Wahal, Figure 5-6 conveys the tendency for investment committees or plan sponsors to hire investment managers with a history of above-benchmark returns and fire managers with lower performance. The chart shows that after managers were hired, their post-hiring excess returns were indistinguishable from zero, and the managers that were fired performed better than the hired managers. The plan sponsors should have just bought index funds and forgotten about manager picking in the first place.
In the 2009 edition of Pioneering Portfolio Management: An Unconventional Approach to Institutional Investment,6 Yale Endowment Chief Financial Officer David Swensen states, "Active management strategies, whether in public markets or private, generally fail to meet investor expectations... In spite of the daunting obstacles to active management success, the overwhelming majority of market participants choose to play the loser's game."
Despite Swensen's admonition, active manager selection and termination remains a common practice among public pension plans. Plan sponsors hire investment fund managers whom they expect to deliver above benchmark returns.
An investigative journalist for St. Petersburg Times, approached Index Fund Advisors (IFA) and a handful of other investment experts to collect some in-depth analysis of the risks and returns of the Florida State Pension Plan for various periods of time relative to various index portfolio strategies. The research results were revealed in a July 31, 2011 article titled, "Easy investments beat state's expert pension planners,"76 which concluded that a simple index portfolio would have outperformed the Florida state pension plan's investment performance over the last ten years.
"The professionally managed SBA [State Board of Administration] performed worse — by more than a percentage point — than seven index-fund portfolios for the decade ending Dec. 31, 2010," the article reports. "On average, a $100 investment in an index portfolio grew to $184, while Florida's pension delivered just $157," the reporter concluded.
The findings prompted further query for IFA. If Florida's $124 billion pension plan fared so poorly against the index portfolios, what about the other states? IFA has attempted to analyze the employee retirement systems in all 50 states. Data on more than 40 state pension plans have been received to date, yielding similar results with varying degrees of underperformance relative to the index portfolios.
Figures 5-7 through 5-10 show the annual risk and return of various state pension plans, net of fees, compared to passively managed index portfolios comprised of a blend of diversified asset allocations. A best effort was made to estimate fees in states that report returns before fees are deducted. States were analyzed for both 13-year periods and 26-year periods and were charted based on either a June 30th or December 31st year-end date. The data shows that in the 26-year study, only 2 states (South Dakota and Delaware) matched the index portfolios, and not one of them outperformed in any of the time periods analyzed. For data sources, go to pension-gate.com.
Figure 5-7 7(permalink)
Figure 5-8 8(permalink)
Figure 5-9 9(permalink)
Figure 5-10 10(permalink)
Directors of these pension plans have access to so-called "top" money managers, which would lead one to believe that these plans fired their very best shots at earning above-benchmark returns, only to fall short. This analysis reveals that the widely implemented and costly process of hiring and firing of investment managers for state pension plans has delivered a negative payout relative to a risk-appropriate set of index benchmarks.
Next: Step 6: Style Drifters
As discussed briefly in Step 3, one solution to determine manager skill is to identify if there are enough years of performance data to be statistically significant by measuing a manager’s t-stat. If the t-stat is 2 or greater, then the investor has at least a 95% confidence level that the manager’s above-benchmark returns were due to skill, with up to a 5% chance that they were due to luck.
The four "alpha charts" (Figures 5-11 through 5-14) show the year-by-year difference between the fund return and the benchmark return for four funds with at least ten years of data whose managers who were recognized by Morningstar as "Manager of the Year" for 2013. While all of these funds had a positive average alpha, none of them had a t-stat of 2 or greater, which means that their alpha was not consistent enough to be deemed a function of skill rather than luck. Among the four managers shown to have positive alpha, the average minimum number of years to determine luck vs. skill was 47.5, with the most consistent fund having a minimum track record of 20 years (Figure 5-13) and the least consistent having a minimum track record of 118 years (Figure 5-14).
The Alpha Myth
When managers are subjected to the scrutiny of a simple t-test, the idea of manager skill to produce consistent alpha quickly becomes relegated to the realm of fantasy, taking its rightful place alongside unicorns, Bigfoot, and the Loch Ness Monster - as seen depicted in The Alpha Myth painting.