Despite their “high-flyer” status, FT.com reported today that several high profile UK asset managers have recently begun bailing out of long-only shops. Reasons ranged from retirement to “spending more time with family”. But as this article points out, many have been wooed by hedge fund companies. Says FT.com:
“Immediately after retiring from CSAM, (Bill) Mott (a star Credit Suisse fixed income manager) announced that he could not resist the lure of running money on a day-to-day basis in an environment unencumbered by bureaucracy. He is joining former colleagues to set up Psolve, an investment boutique, which he describes as a ‘dynamic and free-thinking environment where the fund management team’s interests can be aligned with those of both the group and our investors’.
“The lure of boutiques of specialist and hedge fund managers, which promise managers more freedom and higher pay linked to performance, is a sensitive issue for big investment groups.”
It comes as no surprise that “freedom” and “free-thinking” are important cultural attributes to asset managers – as they are to all participants in industries reliant on human and intellectual capital. But the quest for flexibility also has a more mercantile rationale. Managers who leave the constrained world of long-only investing in search of wide open spaces subscribe to the fundamental law of active management. They are “active management fundamentalists”.
They are aware of their own ability to identify investment opportunities (“skill”). And apparently, they see plenty of investment opportunities around them. But their problem seems to be that they are unable, due to the strict investment guidelines and mandates of traditional long-only investing, to bake these opportunities into their portfolios.
This problem was first studied back in 1989 when Richard Grinold identified skill and opportunity as the drivers of success in the fundamental law of active management…
He said a manager’s information ratio (their risk-adjusted value-add) was a factor of the manager’s true skill (called the information coefficient) times the amount of opportunity he had to ply his trade. No opportunity means no value-add, regardless of the level of skill; and no skill means no value-add, regardless of the number of opportunities facing a manager.
The information coefficient is just the correlation between the manager’s forecasts and what actually transpires. It’s 1.0 if the manager has a crystal ball and is named “the Amazing Kreskin” and it’s 0.0 if the manager is a chimpanzee taking target practice. (Do yourself a favour and don’t ask why the number of discrete opportunities to make a bet (i.e. stocks x time periods) is square-rooted).
More recently (in 2001), Roger Clarke, Harindra de Silva, and Steven Thorley added a third variable to measure the extent to which a manager is able to bake his predictions into the portfolio. They argued that constraints prevented a manager’s IC from being as high as it could. In the presence of these common investing constraints, the IR is actually lower than what Grinold’s original theory would suggest. To account for this, they grossed-down the formula by a factor they called the “transfer coefficient”.
The transfer coefficienct ranges between 0.0 (no ability to bake-in ideas) to 1.0 (total freedom in baking-in ideas). Together, the IC and TC are mashed together by the authors and referred to simply as the “performance coefficient” or “PC”. Their seminal paper contains the following chart that unifies a manager’s forecasts, her weights and her returns:
Today, the fundamental law of active management is often used as an intellectual framework to justify 1X0/X0 investing. Strategies involving long positions equaling 120% of NAV and short positions amounting to -20% of NAV allow managers to bet against a name more than simply avoiding it altogether in their portfolio. As a result of dropping this natural floor (i.e. 0%) on position size, a fund’s information ratio should rise.
Clarke, de Silva & Thorley show that the higher the transfer coefficient, the more likely excess returns will be the result of skill, while the lower the transfer coefficient, the more likely excess returns were actually a normally distributed fluke – destined to right itself in the future.
As the following chart (by Clarke & de Silva’s Analytic Investors) illustrates, a lower TC (i.e. more constraints) doesn’t necessarily mean that returns will be lower – just that the proportion of returns from skill will be lower. Lady luck might intervene to goose returns at any time.
This is a very powerful observation that makes a lot of intuitive sense. In the presence of absolutely no constraints, a manager has no excuses, returns are entirely the result of her skill or buffoonery. In the presence of severe constraints, a manager can always argue that losses weren’t their fault. (Although the reverse – when managers claim great returns were just luck – seems to be more rare for some reason.) Efficient markets theorists will note that alpha is assumed to be positive in the presence of zero constraints. This is based on the assumption that “skill” is non-zero – an assumption you will surely question.
Managers that bail out of long-only funds at cruising altitude aren’t just looking for a rush. They know that the removal of constraints is critical to their success (assuming, of course, that their “skill” is non-zero). This is a line of argument that should be used, if it’s not already, by the headhunters that are poaching talent from long-only shops.
At the end of the day, this return to asset management fundamentalism is driving alpha-beta bifurcation and alpha-centric investing. Says FT.com:
“But some think they are fighting a losing battle. Investors continue to replace portfolios of mainstream equities with a more diverse range of passively managed funds and actively managed funds of alternative assets, such as property and hedge funds, promising risk-constrained or market-beating returns.”