Back to Portfolio for the Future™

Alpha Generation and Real-Time Data

October 10, 2017

In a recent paper, Michael W. Brandt and two colleagues look into whether hedge funds “adjust the conditional market exposure in response to real-time changes in macroeconomic conditions, and whether doing so improves their performance.”

The paper is a short one but it may punch above its weight in several important academic discussions, as we shall see.

Brandt is a professor of business administration at Duke’s Fuqua School of Business. He’s also with the National Bureau of Economic Research. The overlap of macro and micro is a characteristic concern of his: his other recent articles of his discuss how macroeconomic news predicts stock returns and the economic fundamentals behind Eurozone sovereign yield spreads. His co-authors on the paper concerning hedge fund market timing are Federico Nucera and Giorgio Valente, of LUISS Guido Carli University and the Hong Kong Monetary Authority, respectively.

On one level this is a contribution to a growing body of literature on “nowcasting,” that is, on the ability of institutions to process and employ real time economic information. These authors use a nowcasting economic growth index developed by Brandt and two other collaborators (Alessandro Beber and Maurizio Luisa) and published two years ago in the Journal of Financial Economics.

A Macroeconomic News Flow Index

The “macroeconomic news flow” index uses only raw figures (in order to remain real-time, it must forego any use of revised figures) and it serves as a proxy for the information that a sophisticated data collector and analyst might extract from publicly available releases. The first part of the Brandt Nucera Valente project then is to determine whether hedge funds (equity hedge funds in particular) are in fact reacting to changes in the sort of information the MNF index reflects. The short answer is: some are; some aren’t. The good ones are.

Economic nowcasting has long (since around 1989 or so) been split between state-space approaches on the one hand and a balanced panel regression on the other. The Beber Brandt Luisa contribution was to combine those two approaches. They abandoned the effort to estimate a real-time series of gross domestic product, and focused on economic news releases that fell into one or more of four categories: inflation, output, employment, and sentiment. They excluded from their database any information based on the financial markets, as the aim was to get behind the interpretation that those markets are putting on the same underlying data. So (to describe the result in terms of a couple of prevailing journalistic clichés that may fit) this was an effort to measure “Main Street,” and second-guess “Wall Street,” which is presumably also looking to the health or otherwise of Main Street.

But let’s get back to Brandt’s work with Nucera and Valente.

In the paper on hedge fund timing, the authors find that there are three groupings: counter-cyclical responders, pro-cyclical responders, and unresponsive hedge funds. Among these groups, the pro-cyclical responders have the best performance, even controlling for standard hedge fund risk factors.

The point, if it holds up under scholarly scrutiny (it certainly seems likely to receive some, given the stakes involved) would amount to a blasted hole in the hull of the efficient markets hypothesis. It suggests that the best hedge funds, presumably on the basis of managerial ability, gain alpha by doing precisely what the EMH says they cannot do – reading publicly available information better or quicker than the markets themselves. Quantitatively, “the most pro-cyclical market timers outperform their less active and counter-cyclical peers by over four percent annualized with a risk-adjusted alpha of 5.5 percent.”

As the authors observe, this research contributes to an ongoing debate over the question whether hedge funds earn their fees by measures of risk-adjusted performance. It answers, some of them do! and it goes some distance at least within the realm of equity, to say something about who does and why.

Finally, this paper also helps clear those alpha generators of the charge that they thrive only because of inside information. The index generated in the 2015 Beber, Brandt & Luisa paper, and employed as the basis of the assessment by Brandt, Nucera & Valente, is based entirely on public information.