Using the Ownership Lens to Select Funds
Practical considerations for using the new statistic.
The previous two columns have discussed the Morningstar Ownership Lens, a new measure born from a 2005 Journal of Finance paper. (“Judging Fund Managers by the Company They Keep,” Randolph Cohen, Joshua Coval, Lubos Pastor.) The Ownership Lens evaluates equity funds not by their own past performances but instead by the past performances of funds that hold similar portfolios.
This seems to me a strange and indirect method for assessing funds, but the Ownership Lens’ proponents counter that criticism with two claims: 1) By evaluating a group of funds, rather than a single fund, the measure reduces statistical “noise”; and 2) The calculation works. That is, funds with strong Ownership Lens scores post higher future returns than those with low scores.
The latter argument was accompanied by 21 years’ worth of history for diversified U.S. equity funds in the Journal of Finance article and an additional 17 years’ worth, using the much larger universe of all global mutual funds, in Morningstar’s follow-up. In short, while the hypothesis might be questioned, the evidence could not. The Ownership Lens measure has shown strong predictive ability.
(There’s much more to be said about the statistic’s calculation, but not here. Those seeking additional detail will find a high-level discussion in the first of my column links, considerably more depth in Morningstar’s paper, and the full treatment in the Journal of Finance article.)
Regrettably, those results are strictly theoretical. Academics study factors, not investment strategies. The portfolios used for their research are both huge and frequently traded, to most accurately capture a factor’s effect. That approach suits their purpose--but it cannot be implemented. Nobody will buy equal stakes in all equity funds that have high Ownership Lens rankings, rebalance monthly, and then reconstitute quarterly, when funds file their latest portfolio positions.
And taking more practical approaches might reduce, or even eliminate, the statistic’s power.
For one, holding only a few funds that have high Ownership Lens scores, rather than a large pool, creates the very statistical “noise” that the measure attempts to avoid. That is, investors following the Ownership Lens strategy might fail even if the measure generally succeeds, should those buyers have purchased duds. Professors who research entire universes need not worry about such flukes, but shareholders must.
For another, what prospers for three months may not do so for three years. The professors (and Morningstar’s sequel) recalculated the Ownership Lens statistic every quarter as new data arrived. The portfolios that they created were always closely aligned with the measure. In contrast, the funds owned by actual investors, over time periods of several years, will drift. Inevitably, the funds that had the highest Ownership Lens scores will regress toward the mean.
A Trial Run
Conducting a thorough test of how the Ownership Lens statistic might be employed in practice would be a very large effort--far beyond the scope of this column. However, Morningstar’s researchers were generous enough to provide me with the results for a single time period. It ran from September 2016 through August 2019, covering diversified U.S. equity funds. The conclusion was disappointing. By and large, when implemented over years rather than months, the usefulness of the Ownership Lens disappeared.
Not entirely, it should be said. On average, the U.S. equity funds that placed in the lowest two quintiles for Ownership Lens scores as of September 2016 subsequently posted three-year total returns that were (very) slightly below their category norms. Meanwhile, the top-quintile funds outgained their category averages by about 40 basis points per year. Thus, favoring funds that had high Ownership Lens scores did prove to be beneficial.
However, the benefit was weak, with no guarantee that it would pay off for investors who formed manageable portfolios. As it turned out, buying the five funds that had the highest Ownership Lens scores when the time period began would have been profitable, as four of the five subsequently posted good relative returns. But selecting the top 10 or top 20 funds wouldn’t have been particularly successful. The performance of those portfolios roughly matched the category averages.
While much more work must be done before a conclusion can be reached, indications are that the Ownership Lens cannot stand alone as a signal for intermediate- to long-term equity-fund performance. The single experiment that was run to test the measure’s practicality supported the suspicion that its effects diminish as the investment holding period lengthens. Be intrigued by the academic findings--but wary of how they translate into investment advice.
That is the bad news for using the Ownership Lens. But there is also good news.
Two Is Better Than One
In their original Journal of Finance article, the authors had the clever idea of re-introducing a fund’s performance history into the equation. As previously stated, they viewed returns alone as an untrustworthy indicator. However, they reasoned, once the Ownership Lens scores were determined, past performance might be helpful in supplementing that measure. Good if a fund had a top Ownership Lens score; better yet if that fund also boasted excellent returns.
Yet another winning hypothesis! Adding performance data to the Ownership Lens does indeed improve predictability. When the authors sorted funds into 5v5 grids, the highest future returns for the 25 resulting groups came from funds that had placed in the top quintile for both Ownership Lens and trailing risk-adjusted returns, while the lowest returns came from funds that had landed in each measure’s bottom quartile.
The same pattern held for my single test. Adding past performance rescued the usefulness of the Ownership Lens measure by tripling the size of the performance gap. Whereas half a percentage point separated the future returns of the top- and bottom-quintile funds, when the Ownership Lens statistic operated on its own, the difference rose to almost 150 basis points when using the double sort.
In summary, when combined with past performance (and presumably also expense ratio), the Ownership Lens appears to be a useful screen for selecting mutual funds. Operational difficulties prevent the measure from showing its full power, but it remains strong enough to be meaningful.
John Rekenthaler has been researching the fund industry since 1988. He is now a columnist for Morningstar.com and a member of Morningstar's investment research department. John is quick to point out that while Morningstar typically agrees with the views of the Rekenthaler Report, his views are his own.