Almost every day there are news stories on the rise of passive management or the relative under-performance of active managers. Articles asking questions such as “Is MiFID II the final Nail in Active Management’s Coffin” or “Active managers exposed as most US equity funds lag behind market” have become commonplace. Meanwhile the data on active manager performance is sobering indeed. S&P Dow Jones Indices released their latest SPIVA report which said, “Over the 15-year period ending Dec. 2016, 92.15% of large-cap, 95.4% of mid-cap, and 93.21% of small-cap managers trailed their respective benchmarks.”
So, the question is why? In this note, I will attempt to list the most egregious flaws in the active management industry, but I do need to point out that not all managers have these flaws and there are some that have avoided making any of these mistakes. That said, those managers are almost certainly in the top 4.6% to 8.85% of managers that did not underperform their benchmarks.
The largest active managers are trapped in an “innovators dilemma” where, despite losing ground to passive funds, and being out innovated by more technology and quant savvy hedge funds, they are still quite profitable. Thus, rather than risk the upheaval of innovation, they continue as they have. Changes are limited to incremental improvements which their marketing departments tell them are important for gathering or retaining assets. Unfortunately, major changes are needed, and I think that the “Seven Deadly Sins” is an apt metaphor to describe the problem:
PRIDE – Star system approach that indulges the ego of “proven” managers at the expense of process improvements
GLUTTONY – Focus on asset gathering, even when it is too much for the fund to handle
SLOTH – Leave the process “as is” rather than improve it
GREED – Make as much short-term money as possible, regardless of value provided
LUST – Chase past returns, rather than focus on process or innovation
ENVY – Arguments that competitive trends are contrary to the public interest
WRATH – Direct anger towards disruptive technologies instead of embracing them
The impact of all of these sins has been sustained under-performance, and their manifestation has been that the industry has spent the last 30 years paying “lip service” to advances in quantitative and financial technology. Instead of embracing technological change, managers, and the pension consultants that are the “gatekeepers” for the investment management industry, use processes that are more or less the same since last century. These consultants focus on recent manager performance, tracking error to their favorite indices and their ability to uncover “alpha”, while areas such as portfolio construction and trading are typically afterthoughts. This is a shame, as there are many active managers with a proven ability to identify alpha, but many fail to construct optimal portfolios or trade efficiently enough to capture the alpha they expect.
To get back to the metaphor, lets dive a bit deeper into each “sin”
Pride – The tendency to hire and market individual “star” portfolio managers, runs counter to the establishment of an optimal process. My eyes were opened to this when I had a conversation with a senior quant at a large asset manager about the ideas in my “quantamental” article. He told me that, while the notion of using quantitative technology to improve returns is slowly gaining traction in his firm, to promote those ideas, he needs to use the word “fundatative” instead. This firm’s portfolio managers believed it was so important to emphasize their primacy that they would not even entertain a conversation about new methods of portfolio construction or trading without doing so…
To understand the potential performance benefit of quantitative processes vis a vis the traditional method, lets contrast the two:
- Traditional – The portfolio manager (PM) makes predictions about both individual stocks and sectors and uses those to generate orders, which are sent to the trading desk for execution. (The orders rarely contain details about the predictions) Separately, risk managers also send orders to the trading desk, to adjust exposures in the portfolio so that factor risks or tracking error are kept within fund guidelines.
- Quantitative — The PM makes predictions about either individual stock or sector outperformance and expresses them arithmetically, by assigning a magnitude and expected timeframe for the prediction to occur. These are fed into a portfolio optimizer, which also has access to a trading cost model and risk factors, which generates orders for the trading desk. These orders contain the predicted alpha (if any) and timeframes for the desk to utilize when choosing a trading strategy.
It is simple to understand how the quantitative approach would be superior, even though the source of alpha is the PM in both cases. Using a separate risk process is clearly sub-optimal since it will cause increased turnover, plus a lack of integration and communication with the trading function can be very costly as, quite often, risk reduction trades are crowded. (Many commercial risk models update at discrete times and many managers might be implementing similar trades. This means that these should be traded quite aggressively.) When orders are generated to capture alpha, however, it is impossible for a trader to properly assess the level of aggression to use or the potential value of liquidity without knowing the magnitude and time-period which the PM expects that alpha to manifest. It is also very hard to believe that individual PMs are capable of sizing positions optimally, using relative alpha, benchmark weights, risk factors and the predicted trading costs for each potential order/position size of every stock in the portfolio.
I often wonder if the reticence of stock-pickers to providing details of their predictions is due to simple vanity and misunderstanding. It is hard to make accurate predictions consistently, and star PMs care about their image. The reality, however, is that a quantitative signal which is 10-20% correlated to future returns is considered quite good, but PMs might think that such a low “batting average” would make them look bad. Of course, IF they did so, another benefit is that it will differentiate between true acumen in stock picking (alpha) from both random chance and the ability to predict industry, sector and other factor returns (beta)…
Gluttony – Managers often accept too much money into funds which can drive portfolio managers to build positions that are too large relative to their own predictions of outperformance. The result is that trading costs, which increase with the size of the position being accumulated, can reduce or overwhelm the alpha which was predicted. Some hedge funds, in recognition of this, cap the size of their funds, but that seems to be rare in the “long only” market. If, however, long only managers adopted the quantitative approach described above, the portfolio optimization process would do this automatically, as it would force incremental dollars towards the benchmark weights once the marginal alpha was insufficient to justify overweighting a stock in larger size.
Sloth – There are many examples that fit in this category, but the best one is under-investment. Funds, who have no expertise in market timing, sometimes keep large cash positions instead of being fully invested. It is extremely easy and low cost to “equitize” the fund using ETFs when allowed, or by trading passively constructed portfolio slices. Consider the math of a $100 million small cap equity fund that had an average of 7% in cash throughout a year when it’s equity benchmark (Russell 2000) rose 10%. Let’s assume that the fund would need to fully liquidate the ETF or equity portfolio slice 4 times per year to fund individual stock purchases. Assuming this small cap equity fund would use roughly 50,000 shares of IWM for equitization (Russell 2000 IShares ETF that trades at roughly 139 to equitize the $7 million in cash), it might cost them 4 cents round trip (including electronic commissions) to equitize and liquidate, but we will assume double that amount (8 cents per share). In that example, the fund would suffer 70 basis points of underperformance for being in cash and roughly 1.6 basis points of transaction costs if they equitized and liquidated with IWM four times throughout the year. This is a very significant difference, and it is worth considering that the cost estimates are very conservative for a fund of this size. While it might be slightly more expensive to use baskets of stocks (instead of ETFs), it would not change the fact that the cost would be insignificant compared to the performance risk by not being fully invested.
Greed – Managers often charge fees that are too large relative to their value added. Many managers create portfolios which have some active “bets” and other positions established explicitly to better track their benchmark index. This behavior, often called “closet indexing”, should result in lower fees, as the index related positions do not require either research or intra-day trading. Of course, active managers who are aware that it is hard to justify their fees when holding substantial percentages of their portfolio at index weights, often do the opposite and either hold cash or allow substantial tracking error to be present in their portfolio, which directly lead to performance issues.
Lust – This is the behavior where the lure of PAST returns drives consultants and marketing departments to perpetuate a “musical chairs” game of manager selection, instead of focusing on manager processes and the sustainability of returns net of fees. According to IPE “A study sponsored by Rotman International Centre for Pension Management and conducted by academics at Columbia Business School, Universite de Lausanne and AQR Capital Management finds anecdotal and statistical evidence that pension fund asset allocations mirror past performance in asset classes at the expense of returns.” My suggestion is that, along with a focus on fees and value added, consultants and funds should revamp their management selection processes to include more detailed attribution. This should include both process attribution as well as performance attribution. Process, in this context, means developing an understanding of how accurately alpha is predicted vis a vis how much alpha is captured/lost when trading. Trading costs are likely to be relatively consistent, which means that firms with high costs are much more likely to underperform in the future and vice versa. Performance attribution, however, means understanding the relative contributions of individual stock alpha vis a vis beta “bets” on sector, industry, geography or other factors. This is particularly important when evaluating fund expenditures on single company versus industry and macroeconomic research.
Envy – There have been many articles and commentaries from the active management community complaining about passive strategies, or casting aspersions against quantitative funds with better performance. It would be much better for their investors, however, if these funds decided to embrace relevant components of these strategies instead. One example is that passive funds have done a very good job implementing trading strategies to handle fund inflows and outflows. Active managers, asked to outperform an equity index, could mirror those techniques, as part of a hybrid strategy that incorporates both passive and active components. In practice, this requires funds to keep track of both the desired portfolio and existing portfolio to know what to buy or sell into the closing auction for inflows and outflows of fund assets. It is worth noting that as fund families start to offer passive products, many have adopted these techniques.
Wrath – I have written extensively on this topic, with a focus on the outcry over high frequency trading, and technology in general. My belief is that funds should embrace these changes and make it work for themselves by investing in better analysis of both their own trading process and those of the brokers they use. Much of the infrastructure used by HF firms is available for “rent” by brokers and there is an evolving selection of firms to provide better contextual analysis of a fund manager’s trading. It is particularly troubling to hear about funds that ask brokers to use specific venues and order types without analysis to support those decisions and for venues to market services based on nothing more than popular narratives. As I have previously written, IEX is the most obvious offender of this category, with funds directing brokers to use them based on an anti-HFT narrative. The point of this commentary is to focus on data and process analysis, and not to pick on IEX, so I will leave those details to an upcoming commentary.
In conclusion, whatever metaphor one uses to describe the failings of active managers and the lack of worthwhile alternatives to passive funds in the “long only” management sphere, there is clearly an issue. My thesis is that, for “traditional” managers to survive and thrive they must adopt scientific rigor and use of quantitative technologies throughout their entire investment process. This goes beyond “big data” for informing the discovery of alpha and requires the full integration of asset price prediction, risk models, portfolio optimization and trading to succeed. Those firms that make this leap, and price their products based on the value they create, will grow rapidly and those that do not, will eventually follow most video rental stores and travel agents into the dustbin of history.