Bernd Scherer tells Martin Steward that asset managers should take a good look at - and possibly hedge - the market risk embedded in their fees. How might that change the relationship with clients?
Many asset management companies went bust thanks to the financial crisis. No wonder many more are now taking a closer look at their key business risks.
“Sometimes it takes a catalyst to create awareness - and P&L is an excellent teacher,” says Bernd Scherer, professor of finance at the EDHEC Business School and former Morgan Stanley managing director.
Scherer’s attention is particularly focused on the pure market exposure risk that asset managers run via the asset-based fees that generate the bulk of their revenues. He feels this is massively overlooked by the industry, which assumes that its key risks are operational. But if an asset manager charges fees as a percentage of assets under management, and if those assets fall in line with the market, it stands to reason that revenues will fall in line with the market. A manager running €10bn and charging a fee of 0.50% immediately has a directional long market exposure of €50m before he even starts - a hugely variable income stream in a business with very high fixed costs. “What has been called scaleability is really just operating leverage,” Scherer observes.
It took 2008 and the need to “fight post-bonus depression” to spur him into writing a series of papers that are now appearing in academic and practitioner journals. His most recent empirical work - ‘Market Risks in Asset Management Companies’ - has been developed with a bigger and higher-frequency dataset with the help of Frankfurt-based consultancy Klimek Advisors. “The subtitle might have been, ‘Was this Necessary?’” he says. “In my opinion it was not - it was due to a failure across the entire industry to expand risk management sufficiently. Asset managers have been focused on their clients’ risks, while ignoring their own.”
The questions are important for all asset management entities, but especially so for those, like investment banks, that maintain these businesses for ‘annuity’-like revenue streams to smooth out the cyclicality of their core business. The crisis showed asset management to be anything but ‘annuity’-like. Its revenues have been falling just as external funding costs have been rising.
To test the hypothesis that there is a direct relationship between asset management revenues and market returns (which should therefore be incorporated into risk management), Scherer studied the quarterly revenue growth of 17 publicly-traded US asset management firms between Q2 2004 and Q1 2010, regressed against the quarterly percentage changes in the average price of the Russell 1000 Index relative to prices at the start of each quarter.
There is very little annuity-like in these revenue numbers (see table). Volatility ranges from Federated Investors’ sedate 4.98% all the way to Invesco’s hair-raising 94.90%. When Scherer regressed average revenue growth against revenue volatility he found an R-Squared of 0.76, suggesting that different firms exhibited different levels of leverage to a common factor, while the high levels of non-normality (skew and kurtosis) look suspiciously like market risk.
After subjecting the revenues and market returns to a rigorous statistical workout, Scherer finds that market risk was indeed the dominant source of revenue volatility for most firms, but particularly for the larger, diversified firms that had experienced no M&A activity. T Rowe Price stood out as the most extreme example. Its 0.93 revenue R-squared against the market suggests that just 7% of revenue volatility could be attributed to the traditional operational risk factors. The conclusion: market risk matters, but no-one seems to be doing anything about it.
“Any risk that is incidental to your core business is risk you don’t want to take,” Scherer observes. “The core business of asset management is alpha generation, distribution, product development, which has nothing to do with this market risk that is deeply and mechanically ingrained into the fee contract. Your beta benchmark is imposed upon you by your client. They want to take that risk, and it may well be a hedge against their liabilities - but for you it is an outright risk.”
It is also a risk for your shareholders, who presumably own you for your asset-management expertise rather than as a leveraged beta play. Even if that shareholder hedges that risk in its portfolio, this cannot undo the frictional costs associated with taxes, agency costs and liquidity that these exposures create within the asset management firm, says Scherer. Firms may not want to go the whole hog and make revenues ‘beta-zero’ by hedging market risk out of their balance sheets (the Klimek white paper suggests Asian forwards for this), but Scherer says they should at least be aware of this risk. They need to rethink their risk management function.
If they do decide to hedge that exposure, how might that change their relationship with clients? When Scherer says they spend most of their risk management resource on clients and not enough on themselves, do clients need to worry that this is a zero-sum game?
It is not in the client’s interest to see its asset manager go out of business, or holding cash reserves against P&L risks that could be upgrading IT or hiring and retaining alpha-generating staff. There is no misalignment there. Nor would any client want to see its manager swing wildly for the fence with super-high beta risk to recoup revenue losses suffered because of a market downturn. “You don’t want an asset manager to manage his own exposure with your money,” as Scherer puts it.
But he goes further, arguing that hedging market risk should provide clients with “a clearer signal of whether its managers are succeeding in the firm’s mission of generating alpha”, because fee levels would then be determined purely by manager performance.
But there is some complexity associated with this. Paradoxically, perhaps, the model works best with passive management. Here, the client pays you to minimise tracking error against its benchmark. Hedging that benchmark risk out has little effect on your incentives to limit tracking error - if anything, it increases the incentive.
At the other extreme, for an unconstrained absolute return mandate with a cash benchmark where returns should be generated by skill alone, the model does not apply - you cannot hedge your core business. The biggest revenue risk to firms running these mandates ought to be straightforward redemption, outflow risk, which Scherer and Cambridge University’s Steve Satchell are exploring for a forthcoming paper.
But what happens as we move along that spectrum, with the client maintaining a market benchmark but allowing greater tracking error (implying desire for higher alpha management)? Unhedged, if the market tanks, your revenues tank with it. If you hedge, because the revenue lost from the market effect is replaced with profits from your Asian forwards, you massively increase the relative impact of your alpha on your final P&L. If zero tracking error means zero revenue risk, and you are confident in your alpha-generating abilities, it is clearly in your interest to maximise alpha.
However, the manager who is not so confident arguably has a greater incentive to hug its benchmark. (These points might be relevant in the context of the EU’s recently-passed AIFM Directive, which requires AIFM “staff members” to undertake “not to use personal hedging strategies or remuneration- and liability-related insurance to undermine the risk alignment effects embedded in their remuneration arrangements”.)
“If you decide it’s safer not to take the alpha risk to your fee revenue, your client is going to say, ‘This guy didn’t produce any alpha, let’s fire him’,” says Scherer.
But how much information would the decision to hedge give, ex-ante, to prospective clients? There seems to be an incentive for both the confident manager (who gets the advantage of higher alpha impact on revenues) and the not-so-confident manager (who can hug the benchmark and be rewarded with less revenue volatility).
As Scherer implies, the light would now be shone directly on the manager’s alpha-generating skill, cutting through the market noise. If a track record shows benchmark-hugging despite generous tracking error risk budgets from clients, it may now be safe to assume that it is because the manager recognises that this is the best way to minimise his revenue volatility in the absence of alpha-generating ability. When his revenues were volatile anyway, he may well have run high tracking error risk despite his inability to deliver a high information ratio.
So there should be a broad incentive to adopt Scherer’s suggestion to be cognizant of, and possibly hedge, revenue market exposure, and it should certainly align interests with clients in terms of business robustness and the ability to deploy resources efficiently. As important, perhaps, it could deliver a valuable piece of extra information about a manager’s confidence in his own ability to generate alpha, too.