Jonathan Berk

Jonathan Berk speaking at the fiduciary investor think tank at Stanford University in the US – an event staged by, a sister publication of Professional Planner.

One of the enduring areas of asset management academic study, and practitioner query, is whether or not fund managers have skill. In his work, professor of finance at Stanford, Jonathan Berk, shows the answer doesn’t lie in returns at all, but in manager compensation.

Determining whether active skill exists is not the same as looking at whether investors get a net return from managers. It’s a difficult concept for institutional investors to grasp; their focus for so long (rightly) is on how they generate a return for their beneficiaries. But in an academic sense, and ultimately in choosing managers which generate those returns, skill is not found by looking at returns.

Berk’s research shows that return does not measure skill: rather, manager compensation is a better proxy for sorting out the top managers from the mediocre ones.

Berk, who teaches a course in critical analytical thinking in addition to his money management course, argues that net alpha measures the competitiveness of markets and the rationality of investors, giving information about investors but not about managers.

Compensation: a better proxy

Instead, he argues that compensation is a better proxy for ranking good managers.

“You can look at compensation out of sample,” Berk says.

“The top-decile managers make more money than the lower; the only way variation can arise is through size, which means investors are compensating them. Investors are predicting good managers,” he says, and in this way the current compensation of managers can predict future performance.

“This analysis does a much better job of sorting out the mediocre managers.”

Berk demonstrated this argument to delegates at the investment think tank, made up of chief investment officers of public and corporate funds as well as foundations and endowments, which falls back on one of the foundations of economics: that people with skill are in short supply and earn economic rents.

In the paper co-authored with Jules H. van Binsbergen, Measuring Skill in the Mutual Fund Industry, published in November last yearBerk looks at whether mutual funds earn economic rents without possessing skill. It’s a different look at an old problem.

The paper, which was the basis of Berk’s presentation at the investment think tank, finds that the average mutual fund has added value by extracting about $2 million a year from financial markets. And, the authors say, most importantly the cross-sectional differences in value added are persistent for as long as 10 years.

They conclude that it is hard to reconcile the findings with anything other than the existence of money management skill.

Investors identify talent

The argument follows that higher skilled managers manage larger funds and reap higher rewards. And that, surprisingly, investors appear to identify talent and compensate it, so current compensation predicts future performance.

“Look at the most successful mutual fund manager of all time, Peter Lynch,” Berk says. “For the first five years he returned 2 per cent gross on $40 million a value added of $1 million per month, in the last five years he returned 20 basis points gross alpha on assets that grew to $10 billion, a value added of $20 million per month,” Berk says. “If you looked at alpha instead of value added, you have concluded that Peter Lynch was lucky.”

Berk argues that net alpha measures the competitiveness of markets and the rationality of investors, but tells you little about managers, or whether they possess skill.

“Net alpha doesn’t tell you about managers, it tells you about investors,” he says. “Using net alpha has led to people concluding that mutual fund managers are unskilled when in fact they are actually highly skilled.”

The Michael Jordan analogy

Berk uses a basketball analogy to demonstrate why net alpha, or the return to the investor, is not useful in measuring skill.

“Is field goal percentage a good measure of a basketballer’s performance? No, because the defence adjusts. For example Michael Jordon was double teamed, and this affected not only Michael Jordan’s field goal percentage but also Scott Pippen’s,” he says.

“If you were a coach (without integer constraints) and were optimising, then the field goal percentage of each player would be the same. Similarly investors all want to invest with managers that can deliver positive alpha. In equilibrium this inflow of capital drives alphas down to zero. Like basketball players, all managers make the same alpha and so alpha tells you nothing about managerial quality.”

Similarly gross alpha, which is just net alpha plus the fee, is also not useful to measure skill.

“The manager picks his fee so how can that be used to measure his skill?”

Berk presented his paper which was followed by a roundtable discussion with the investors and sponsors.

The presentation incited many questions and discussion points from the investors around the table, including:
Are benchmarks properly diversified?
David Cooper, chief investment officer, Indiana Retirement System

Will skill get competed away?
Farouki Majeed, chief investment officer, Ohio Public Schools Retirement System

Do inflows cost performance or the other way?
David Long, co-chief investment officer, HOOPP

Are their liquidity and capacity constraints?
Eric Baggessen, senior investment officer, head of asset allocation, CalPERS

As assets grow do managers need to change their strategy because the AUM is restrictive?
Hershel Harper, chief investment officer, South Carolina Retirement System

As a result of the discussion, investors want to see further academic study answering the questions:
Is skill is with the individual or the firm?
Do investors need skill to pick managers?




Join the discussion