The MFS 'Active Advantage'... Really?

by: Mark Hebner

Summary

Analysis of MFS Investment Management mutual funds.

Active managed fund advantage?

Test for statistically significant alpha for MFS funds.

For those of you who like to peruse CNBC every once in a while, there is a very high chance that you have seen commercials from an outfit called MFS Investment Management.

They are very well scripted and highlight some of the typical selling points active managers like to make when speaking to investors about topics such as risk management, downside protection, alpha, and a globally informed perspective through a network of analysts that have boots on the ground at different locations around the world. It is a very compelling story that speaks to a lot of the fears investors have when broaching the topic of investing for the future.

The biggest selling point and overall message from these commercials is the fact that MFS Investments are being active in the sense they are not just sitting back and letting things happen. This is often one of the issues we have as advisors when talking about the difference between an active and passive approach to investing in that passive insinuates this idea of not taking action and letting things like large market swings or problems in China just happen. In other words, we are not being proactive in controlling the outcomes these events have on the wealth of our clients.

And it is completely understandable. How many times in our daily life do we just sit back idly and wait? It is not the American way or the American spirit. We are supposed to take control of our day and our future. Our outcomes at work and other personal goals such as fitness or relationships are a direct reflection of the actions we take or do not take. Therefore, the same should apply when it comes to investment outcomes, right?

This is the unfortunate paradox that traps many investors into believing that the same active approach we take in other areas of our life also applies to that of the financial markets. The reality is that an individual investor has very little control over the day-to-day fluctuations of the market. Furthermore, nobody has enough foresight to know what is going to happen to the markets that somebody else doesn't already know; at least not with a high degree of certainty. What an investor does have control over is the long-term growth of wealth, which is dependent upon their individual savings rate and amount of risk taken within their portfolio. Legendary investor and mentor to Warren Buffett, Benjamin Graham, famously stated in Security Analysis, "in the short run the stock market is a voting machine, but in the long run it is a weighing machine."

Now, even though we have laid out a line of thinking that seems reasonable, it still wouldn't be a robust argument without empirical support. Let's take a look at the individual claims that MFS Investment Management makes in their commercial and the corresponding peer-reviewed academic support for their statements. This is not intended to highlight the merits of MFS Investment Management individually since their claims are very common amongst the active investment community.

Claim #1: Active management tends to perform well during market transitions, particularly more difficult markets, providing value in risk management and reducing downside volatility.

This is a classic argument active management proponents like to use. It assumes that based on current information they know exactly where we are at in a market cycle. In terms of the academic literature, there has been no conclusive evidence that demonstrates that this claim is true. Active management tends to perform similarly during down markets and up markets. We have already written articles before on the topic of active management providing protection during down markets here. We referenced a paper from the Fall 2012 edition of the Journal of Investing called Modern Fool's Gold: Alpha in Recessions, which concluded that there is very weak persistence in a manager's ability to provide superior outperformance during subsequent recessions or expansions. So while it sounds very comforting to say we reduce downside volatility but fully embrace upside volatility, managers usually do not know when each is going to occur, and by the time they act on their premonitions, the market has already moved on.

Claim #2: Focus on security selection is such that no systematic component of risk drives performance and offsets everything we are trying to do.

Let's dissect this sentence since there is a lot of fluff. No systematic component of risk refers to either overall market risk, size risk, or relative-price risk. These are more formally known as the different dimensions of risk, which were comprehensively introduced in 1992 in Eugene Fama and Ken French's seminal paper, The Cross Section of Expected Returns. So if we don't want one single component of these risk factors driving the performance of the overall strategy, then what we are really saying is that we want to be very well diversified. We want large stocks, small stocks, growth stocks, and value stocks across all different sectors and regions. The problem with this approach is the more and more diversified you get, the more and more you look like the overall market (i.e., index fund). As we will show later, this seems to be an issue MFS Investment Management is running into with a lot of their seasoned strategies.

From a security selection standpoint, most active managers have a really tough time distinguishing the next winners from the next losers. In Fama and French's paper, Luck versus Skill in the Cross-Section of Mutual Fund Returns, they looked at the performance of 819 different mutual funds over the 22 years ending in 2006. They found that the vast majority (97%) didn't beat their respective benchmark (produce positive alpha), and the small amount that did is indistinguishable from just choosing stocks at random.

Click to enlarge

Claim #3: From a philosophical standpoint we look to take risk where we think we will be compensated and budget risk for where we have the most skill and ability to consistently deliver alpha.

This claim is very similar to that of Claim #2. We already know where investors should be expected to earn a return for risk taken; that would be in stocks that are smaller in size and value-oriented. Anything beyond these factors is assuming the pricing mechanism inherent in the market is not functioning properly. In other words, current stock prices do not properly reflect future expectations of a particular company, which is just complete nonsense. To look at a price and say it is currently mispriced means that your ability to estimate the fundamentals of a particular company or your ability to predict the future are superior to that of the other few million professionals in the world. It's their collective estimate of a company versus yours. That is quite a claim to say you know something they do not. And the law of large numbers is stacking the odds against you. We can once again cite the paper by Eugene Fama and Ken French on Luck versus Skill. The vast majority cannot consistently deliver alpha across multiple asset classes and time horizons. What is so special about MFS Investment Management that they believe they can deliver it? To say they have a global operation that believes in collaboration and sharing of information is nothing special. Most of the big asset managers have global operations and are promoting the exact same thing.

Claim #4: We like to tell clients to arbitrage the time horizon and have a long-term view.

This is just fancy speak for staying diversified and having a long-term focus, which is something we also promote. Because nobody can accurately predict what is going to happen in the future, it is best to "arbitrage the time horizon" by buying low-cost index funds and rebalancing when necessary. The market, not one particular individual, knows best.

Cold, Hard Facts

Let's look at the cold, hard facts. We examined the performance of all 87 different mutual funds that MFS Investment manages to see if their integrated research platform around the world that promotes the sharing of information, protecting investors from downside volatility, taking risk where they expect to be compensated, and budgeting risk for where they think they can provide alpha has worked out well for them.

We first looked at all of their US-based strategies that had at least 10 years of performance history to see if they were able to deliver outperformance over a medium time horizon. Of the 14 funds that were US-based strategies and had at least 10 years of data, not a single one produced a positive alpha that was statistically significant to a high degree of certainty (95% confidence level) once we adjusted for the known dimensions of expected return using the Fama/French 3 Factor Model. See chart below.

Click to enlarge

As you can see, most of the funds hug the dashed line that represents zero annualized alpha. This is what we would expect to see not only in an efficient market, but also for someone who is tracking the overall market, which is the point we made under Claim #2.

The second thing we looked at was the performance of all 87 mutual funds against their Morningstar assigned benchmark, since inception, to see if there were any funds that produced a statistically significant alpha.

Here is what we found:

  • 53 (61% of all funds) displayed a NEGATIVE alpha compared to their Morningstar assigned benchmark since inception
  • Only 1 fund (1.15% of all funds) displayed a POSITIVE alpha that was statistically significant at the 95% confidence level compared to its Morningstar assigned benchmark
  • 27 (31%) displayed a POSITIVE alpha compared to their Morningstar assigned benchmark
  • 7 (8%) of the funds had no alpha to report since they had been around for less than 1 year.

It is important to note that these performance figures do not take into account the front-load fees that are associated with these funds. The average maximum front load-fee as of 10/31/2015 across all MFS Investment strategies is 5.16%. These fees may or may not be paid by investors based on broker recommendation or custodial and/or broker platforms. The table below lists the fees for all MFS funds. Readers can find a more in-depth analysis and illustration of the costs associated with strategies from MFS Global Management here.

So the vast majority (61%) of their funds displayed negative alpha and only 1 strategy had a positive alpha that was statistically significant at the 95% confidence level. Which was the only fund whose performance may have been attributable to skill? It is the MFS International New Discovery A (MUTF:MIDAX) coming in with an annualized alpha of 4.96% and a t-statistic of 2.02. Below is its alpha chart showing the performance comparison against the MSCI All Country World ex US Index.

Click to enlarge

Does this necessarily mean that we would concede that investors should reliably expect to be better off investing in MIDAX versus a comparable index fund or mix of index funds? Absolutely not! There are key reasons why, although MIDAX has historically produced a statistically significant alpha, that investors should still stick with a passively managed index fund instead:

  • First has to do with the Morningstar assigned benchmark, which happens to be the MSCI All Country World ex US Index. The Index has a 99% developed and 1% emerging market makeup, while MIDAX has consistently had 10-15% allocated to emerging markets, which we know has experienced a higher historical return than that of its developed counterpart.
  • We cannot control for the overall size and value tilts in this particular strategy like we can with US-based companies since we do not have independent size and value factors for a global ex US portfolio of stocks. As we showed with the 14 US-based strategies, the alpha is diminished down to nothing once we have controlled for known dimensions of expected return.
  • A lot of the alpha seems to have happened during the first 4 years (very tall green bars) of the fund's history, but since then, results have been mixed. This means the statistical significance might be subject to in-sample bias. As we have mentioned before in this article, the best way to control for in-sample bias is to look at two independent time periods to see if the statistically significant alpha persists.
  • Although the alpha for the strategy is statistically significant at the 95% confidence level, there is still a 5% chance that the outperformance was due to random luck. Our opinion that it's random luck is bolstered by the very weak performance for the remaining 79 funds in the MFS Investment lineup. The policies, procedures, and systems in place in terms of hiring professional staff, gathering and analyzing information, and implementing investment strategies would seem consistent across the business. Unfortunately, this has not translated into an overwhelming success story across all of their strategies. Unless they applied a process, which was unique to MIDAX, it may seem reasonable to believe that just by random chance 1 out of their 87 strategies would have success. There is nothing distinguishing about MFS Investment Management's process that would give us confidence they could deliver future outperformance.

As we have said before in other articles that examine the performance of major mutual fund lineups (see Fidelity Funds Part 1, and Part 2, American Funds, Lord Abbett Funds, JP Morgan Funds, Hedge Funds, Olstein All-Cap), this in no way is supposed to critique the level of intelligence and competence of the individuals in these organizations. They are the very brightest and most knowledgeable in our industry. Their reason for their lackluster performance has to do with the environment in which they operate. The free market that allows for the continuous sharing of information and exchanges that have the ability to aggregate that information are the ultimate culprits for the existence of the manager who can consistently deliver "alpha."

No single individual is more powerful than the overall market and no single individual will ever have all the information at their finger-tips about a particular company and its future earnings potential. All individuals are subject to estimation error when it comes to analyzing financial information and future growth prospects, and the aggregating of these estimation errors by the markets allows the price to be the single best estimation at any given time. Some active managers may be right at times at the expense of other active managers, but that is what we would expect. Unfortunately, the persistence of any manager's outperformance comes down to either better estimation faculties, quicker access to information, or illegal insider information. In today's overwhelmingly competitive markets, the latter or blind luck seems to be the biggest reason why most managers have experienced long-term success. Below are a few examples of the alpha charts we do in our analysis based on Morningstar assigned benchmarks.

Click to enlarge

Click to enlarge

Click to enlarge

You can find alpha comparisons across all 87 MFS strategies in the original article here.

Disclosure: I/we have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours.

I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it. I have no business relationship with any company whose stock is mentioned in this article.