This post was originally published and is credit to this site
Active fund selection is a part beta, part alpha decision.For traditional asset classes, it is a fairly straightforward process. Investors seek an alpha/beta blend. In selecting an active large-cap manager, for example, investors desire a skilled manager who outperforms by investing in large-cap stocks. But alternative investments tend to blur the alpha/beta line, and too often, investors incorrectly base their fund analysis on a belief that alternatives lean closer to the manager skill/alpha side of the spectrum. This thinking can lead to a few key problems when selecting alternatives managers, because unlike the typical large-cap fund, a pure-alpha alts fund has few substitutes, and it is generally more expensive. Conversely, beta is straightforward exposure to an asset class. With beta’s lower cost and ease of comparability, it’s no wonder that fund investors are talking about beta again. But herein lies a problem for alternatives: How do we go about selecting alts funds when the line between what is considered alpha and what is considered beta is so tenuous?
We believe the focus-on-beta approach is highly liberating, but it hinges on first determining asset allocation. Simply put, this is a matter of where alternatives allocations are funded from.As we will detail later, a long-short credit fund, for example, can offer more credit exposure and much less duration than a traditional fixed-income fund. Allocating assets to one of these funds from traditional fixed-income offerings should be viewed as an asset-allocation decision. But if the goal is to exchange a duration allocation for credit, we have a whole host of available options. Sure, a long-short credit fund may be one option, but we might also achieve similar results with a traditional low-duration high-yield bond fund or a bank-loan strategy. We refer to these alternatives to alternatives as “next-best options.” Next-best options are not only helpful because they can be thought of as lower-cost substitutes, but like alternatives, they also can be diversifying. Alternatives can be great diversifiers, but they don’t have a monopoly on diversification. While imperfect substitutes, next-best options can help investors consider low correlations in the appropriate context, and we can use this concept to quantify if an alternative fund is worth its added cost.Finally, traditional assets can be used to build best-fit portfolios of stocks and bonds. These portfolios can be used as benchmarks to represent a fund’s asset-class exposures (beta); any outperformance, or underperformance, relative to the benchmark may signify a manager’s skill (alpha). To perform this analysis, we built simplified factor models that represent an investor’s choice between an alternatives fund and a custom combination of traditional asset classes. While we aren’t the first to perform this type of work, our application varies slightly. Instead of using these models to select individual funds, we aggregate our data to make a holistic determination on each category—we believe certain categories simply offer more attractive opportunities than others for investors choosing between alternatives and traditional funds.The Bits and Pieces of Alts
Many of the fundamental building blocks of alternatives funds are traditional asset classes. Long-short equity funds, which bet both for and against stocks but generally are only around 50% to 60% of assets net long, are an easy example because they can be benchmarked to a simple portfolio of stocks and cash, with the appropriate mix depending on the fund’s beta.A bitter pill to swallow, however, is that many long-short funds fail to beat that benchmark (they don’t generate alpha), even before the massive fee hurdles these managers would have to overcome to beat a low-cost next-best option.The average fee for funds in the Morningstar Category of long-short equity funds was 1.86%, as of year-end 2017. The average fees2 for large-blend and ultrashort fixed-income funds were 1.02% and 0.52%, respectively, for 2017. This comparison shows it can be far cheaper to lower beta using cash. A long-short equity fund would need to outperform a combined fee hurdle by approximately 1 percentage point to justify the greater cost. While we will detail later which alternative categories we do believe have the potential to generate alpha, for now it’s important to understand what exactly constitutes alpha and what might be mistaken as alpha.Let’s move to a slightly more complicated Morningstar Category: long-short credit funds. These funds are the fixed-income variant to long-short equity funds and can go long and short various bonds. The products have a very low (nearly zero) duration profile. Being long one bond and short another with a similar duration would equate to having no duration. A cursory analysis may reveal these funds to be lowly correlated to bonds. And, indeed, they are—sort of. Their correlation to the Bloomberg Barclays U.S. Aggregate Bond Index averages around 0.24.There’s a problem with this comparison, however: It isn’t valid. Long-short credit funds take on little duration risk, and interest-rate movements explain approximately two thirds of the volatility in the U.S. aggregate bond index. So, if we used this index as our benchmark, we would be led to believe that long-short credit managers had a high degree of alpha. In fact, we believe a better comparison would be against the Bloomberg Barclays U.S. Corporate High Yield Bond Index, because long-short credit funds take on a fair amount of credit risk. Here, the correlation with the index nearly doubles to 0.48. As we can see, understanding the inner workings of a fund is key, because sometimes we might mistake beta as alpha. In this case, longshort credit funds offer a beta play on credit.3 But even the high-yield index comparison isn’t perfect, because the high-yield index does have a few years of duration. So, what are long-short credit funds’ closest substitutes or, more specifically, their next-best options?Low-duration credit funds such as high-yield and bank-loan funds would be a great place to start. Nontraditional bond funds are similar; they tend to have less duration. But they are more unconstrained. They can bet on a wide array of assets, from currencies to sovereign emergingmarkets credits. That diversification shows; the funds are only 0.12 correlated to the U.S. aggregate bond index. However, against the high-yield index, they post a correlation of 0.67.Alternatives to Alternatives
By using correlations as a guide, and understanding a bit about what’s going on under the hood of funds, we can find many examples of assets that we’d consider comparable to alternatives, or next-best options. Once we realize where the core risks (the beta exposures) of alternatives truly lie, examples tend to crop up all over.While we don’t think there are prefect substitutes lying in plain sight, we do think that reasonable substitutes abound. As mentioned, for low-duration, credit-sensitive funds like long-short credit and nontraditional bond funds, a low-duration high-yield or bank-loan fund might fit the bill, as bank loans also offer credit exposure without duration. Categories such as option-based strategies conceptionally may be a bit trickier; strategies such as covered calls offer both income and equity risk. However, we find that many of these strategies still have a beta of around 0.5 to the S&P 500. While there are no perfect substitutes, it may help to look at these strategies against the broader backdrop of other income-producing strategies that bear equity risks, such as high-dividend stocks, REITs and master limited partnerships, or possibly preferred stock.To identify other next-best options for alternatives, we look at the betas, correlations, and return expectations. For example, multialternative categories are filled with an amalgamation of various strategies, but they are generally funds of alternative funds. So, while no single next-best option is available, it can be helpful to benchmark this group to a low-risk, low-return category, such as a conservative multiasset category. Given the average multialternative fund provides a 0.25 beta to equities, a category that has between 15% and 30% equity exposure isn’t a bad comparison in our opinion.The Value of Low Correlations
To calculate if a lowly correlated alternatives fund’s higher fees are justifiable, we offer a simple, back-of-the-envelope technique. To make this work, we need to compare an alts fund to its next-best options.The conventional approach would be to add all these asset classes to an efficient frontier. But an efficient frontier leaves no leeway for intuition. If alternatives are even slightly less correlated than our next-best option, all else equal, the frontier will cast aside our cheaper, passive option. Our goal instead is to calculate the low correlation benefit of alternatives, or essentially the maximum amount an investor should be willing to pay.For example, we could start by building two very simple portfolios, both with a 90% allocation to a 60/40 stock-bond mix (i.e., the stock allocation would be 54% of portfolio assets, bonds 36%) and 10% allocated to either an alternative fund or our next-best option. In this example, we’ll use a long-short credit fund and a bank-loan fund, respectively. To keep everything very simple, we assume that both the long-short credit fund and the bank-loan fund will return 3.5% and have a standard deviation of 3%. (The process still works if we relax these assumptions.) After looking at historical data, we conclude the bank-loan fund will have a 0.55 correlation to the 60/40 allocation versus 0.3 for the alternative fund. Everything is the same between these two portfolios except the alternative fund has a 0.25 lower correlation to the 60/40 allocation.Bank-loan funds have an average fee of 0.85%. What should an investor be willing to pay for the lower correlation benefit of the long-short credit fund? To find out, we calculated the Sharpe ratio on both portfolios. We mathematically back out the fee for the alternatives fund as if the Sharpe ratios of the two portfolios were the same. This would be the maximum anyone would want to pay for the long-short credit fund. However, we’d argue that investors should pay much less to have an adequate margin of safety. We found that the most an investor should pay for the 0.25 lower correlation was 1.4%, given our assumptions. The average long-short credit fund costs 1.35%. But 20% of the funds in the category charge less than 1%, so lower-cost options certainly exist.Improving the Odds
Now, we turn to alpha. Some alts funds will justify their fees. To improve our chances of finding one, we can average each alternative category’s alpha. Our study used a Fama-French threefactor model for equitylike categories, a two-factor (credit and duration) model for bondlike categories, and a mix for multialternative funds. We only looked at funds with at least a threeyear track record and included dead funds to avoid survivorship biases. We ran our calculations from inception.Here are the results of our study:
The results jibe with many well-established investment truisms, such as that equity investing tends to be a zero-sum game. We stated that long-short equity managers as a whole don’t add much alpha in the period we studied. Indeed, the average long-short equity fund generated nearly zero alpha. Markets are extremely competitive, and adding the ability to sell stocks short doesn’t appear to give managers a new competitive edge; shorting is just as difficult, if not more difficult, as long bets.Another investment mantra for many is that there are more active opportunities in fixed income. We also found this to be the case with alternatives. But we were a bit surprised by the results. We expected to see some modest alpha for the nontraditional bond category, as certain funds take on international exposure that isn’t fully captured by our simple models. However, we were surprised to see long-short credit funds generating approximately 1.8% alpha. That group is more U.S.-centric, so the results are compelling; however, the category is small, so sample size may be a factor.Our results also showed that market-neutral funds generated the highest alpha at around 2.4% per year. A few caveats are necessary, however. We report all these numbers gross of fees4 because we believe it’s important to first establish which categories might have an inherent advantage to offer manager skill, and which do not. Including fees can cloud that picture. But fees are still an important consideration. For market-neutral funds, fees tend to be high, averaging around 1.76%. However, it’s important to stress that lower-cost options exist and paying over two thirds the category’s historical alpha in fees wouldn’t likely be advisable. Finally, returns in this space tend to be mostly alpha, meaning that while in other categories investors can still generate returns absent of alpha, in market-neutral land, alpha generation tends to drive returns.Multialternative funds take on an array of exposures and international positions, as well. While we wouldn’t have been surprised to see the category generate zero alpha, perhaps our models’ negative 1.2%-per-annum assessment is too bearish. These funds can have momentum and other features baked in that we left out for simplicity (and momentum has performed poorly recently). Still, the negative alpha isn’t ideal, especially once fees are accounted for.Finally, option-based funds in our study generated about 1% alpha, better than long-short equity, but worse than fixed-income categories that generated lower returns overall (so alpha per unit of returns is higher for fixed income). With fees hovering around 1.55%, these funds appear to be charging a lot more than they’re worth.Alpha as a Guide
Investors can use these category average alpha figures to help guide their investment decision-making process. Categories with less-than-stellar track records of generating alpha should be viewed more cautiously, while those with higher potential could have more opportunities. As mentioned, however, alpha is only part of the story. The full spectrum of opportunity should be thoroughly scrutinized when analyzing alternatives. For example, the diversification power of other assets should also play a key role in decision-making. That is why the ability to separate alpha potential from passive beta exposures is so important; it helps to broaden the opportunity set. Similarly, categories such as market-neutral can offer higher alpha potential but have lower-return profiles overall, so while alpha may be a good way to measure risk-adjusted outperformance, alpha alone can’t tell us how to make asset-allocation determinations.
1 The free cash flow yields/sales ratio is cash from operations minus capital expenditure divided by sales.2 Fee data throughout come from Morningstar Direct as of Dec. 31, 2017.3 Beta here really only tells us half the story. The Greek letter is meant to signify an asset class’ sensitivity to the market. So, a beta of 0.5 to high yield can be interpreted to mean that these funds will gain or lose at half of the rate of high yield. Missing in beta, however, is how strong this relationship truly is, or its R-squared. In this case, that relationship is quite strong, as average R-squared values clock in at 0.72 for the category, relative to high yield.4 We chose to run our study gross of fees because fees can muddy the performance appraisal waters. By casting fees aside, it is easier to discern the categories where managers have added alpha. Armed with that information, investors can better gauge what fee levels are more appropriate.
This article originally appeared in the February/March 2018 issue of Morningstar magazine. To learn more about Morningstar magazine, please visit our corporate website.