Using a dataset of $17 trillion of assets under management, researchers document that actively managed institutional accounts outperformed strategy benchmarks by 86 (42) basis points gross (net) during 2000 to 2012.

In return, asset managers collected $162 billion in fees per year for managing 29% of worldwide capital. Estimates from a Sharpe model imply that their outperformance comes from greater nuance in factor exposures (smart beta).

Joseph Gerakos, from the Tuck School of Business at Dartmouth College; Adair Morse, from the University of California, Berkeley; and Juhani T. Linnainmaa, from the University of Southern California Marshall School of Business, obtained data for the 2000 to 2012 period from a global consultant that advises pension funds, endowments, and other institutional investors on the allocation of capital to asset managers.

In asset manager language, they explored the importance of smart beta or tactical factor allocations. They explain that the words smart and tactical refer to tilting portfolios toward toward better-performing factors. Their estimates tie positive performance directly to smart beta investing. They document that institutional asset managers outperformed strategy benchmarks by 42 basis points net of $162 billion in annual fees and that smart beta investing entirely explained this outperformance.

Researcher suggest that because the unit of observation in institution-level studies includes both delegated and non-delegated capital, an implication of the results is that non-delegated institutional capital likely underperforms delegated institutional capital. Furthermore, there are differences in asset classes covered. Most institution-level studies focus on the U.S. public equity asset class. In these results, U.S. public equities have the lowest positive alpha relative to strategy benchmarks.

The results from the Sharpe analysis raise the question as to whether delegation to an asset manager was worth $162 billion per year. Could institutions have performed as well over the sample period by instead managing their assets in-house, assuming that they had the knowledge and ability to implement a factor portfolio?

The researchers consider the investment opportunity set of tradable indices that was available to institutions during the sample period, and find that if institutions had implemented dynamic, long-only mean-variance portfolios to obtain their within-asset class exposures, they would have obtained a similar Sharpe ratio as asset manager funds, taking into account trading and administrative costs.

This finding suggests that asset managers earned their fees at the margin. The researchers’ estimates also imply that the introduction of liquid, low-cost factor exchange-traded funds (ETFs) is likely eroding the comparative advantage of asset manager funds.

The data cover $18 trillion of annual assets on average over 2000 to 2012. The data include quarterly assets and client counts, monthly returns, and fee structures for 22,289 asset manager funds marketed by 3,272 asset manager firms. The median fund pools six clients and has $285 million in capital invested in a strategy. The analysis focuses on four asset classes: U.S. fixed income (21% of delegated institutional assets), global fixed income (27%), U.S. public equity (21%) and global public equities (31%). These asset classes represent the lion’s share of global invested capital. In these asset classes, researchers have close to the universe of institutional asset managers that were open to new investors during this period.

The report, “Asset Managers: Institutional Performance and Smart Betas” may be purchased or downloaded for free from http://www.nber.org/papers/w22982.