Have We Been Too Hard on Active Funds?

Could it be that some of us have gotten active funds wrong? That they’ve topped their indexes more often than we’ve given them credit for? That we’ve undersold their value?

Those are the main contentions of a recent study published by Martijn Cremers, Jon Fulkerson, and Timothy Riley, in which they dissect the popular “S&P Index Versus Active” (SPIVA) scorecard’s methodology, arguing that it makes active funds look worse than they actually are. It’s an interesting paper, and well worth reading, but ultimately I didn’t find it convincing.

Before I get to the key pillars of the authors’ argument, a word on how the SPIVA report works: It tallies up the number of active funds that existed at the start of a time horizon and then counts how many made it to the end and beat their style-appropriate index. So if there were 100 active funds that began a 10-year period, 40 were subsequently merged or liquidated away and only 20 of the 60 surviving funds topped their index, it would mean 80% of the funds failed. And so on.

(Note: We have our own version of SPIVA called the “Morningstar Active/Passive Barometer.” It’s an excellent report. But since the authors focused on SPIVA, that explains why I address that study and not our own barometer report here.)

Here are the three main critiques that constitute the authors’ argument, in order of persuasiveness.

1. Comparison with a costless benchmark as SPIVA does isn’t valid, as investors have to pay at least something for an index fund in the real world;
2. Equal-weighting active funds like SPIVA does, instead of weighting by assets, presents a skewed picture of how likely it was for the average dollar to outperform;
3. Contrary to how the SPIVA study treats them, dead funds aren’t necessarily all failures.

The authors test the effects of modifying the SPIVA study—by using actual passive funds in lieu of indexes, asset-weighting instead of equal-weighting, and crediting obsolete funds if they outperform before dying, respectively—and show the differences compared with what was reported in the SPIVA study. They do so for each change individually and for all three in combination.

Here’s how they described what they found for active US equity funds:

Active management, in each asset class, fares significantly better after our changes. Averaged across the US equity categories, the Scorecard indicates that 92% of active funds underperformed over the last 20 years. In contrast, we find that 55% of assets underperformed, with half of the investment categories showing a majority of their active fund assets outperforming. Put another way, rather than overwhelming underperformance, our results suggest that the probability of outperformance for a given dollar invested in the US equity class over the last 20 years approximated a coin flip.

This chart, which is drawn from the dataset in the paper, compares the percentage of active US equity funds that underperformed their benchmark per the SPIVA Scorecard and the corresponding figures the authors came up with after making these adjustments.

For instance, over the 10 years ended Dec. 31, 2024, SPIVA reported that almost 90% of active US equity funds failed to beat their index, but the authors found only 63% lagged after the adjustments. The differences tended to be wider over shorter horizons than longer horizons.

Critiquing the Critique

As mentioned, I think the paper is worthwhile. Methodological choices matter, and the authors have rightly scrutinized some of the choices S&P has made in implementing the SPIVA Scorecard (we are no strangers to such scrutiny ourselves). Nonetheless, I find their overall argument unconvincing for a few reasons.

The Picture Doesn’t Change That Drastically

Even when the authors asset-weighted, utilized actual passive funds, and credited any dead funds for outperforming while they were alive, they still found that nearly two-thirds of active US stock fund assets failed to outperform over the decade ended Dec. 31, 2024.

It was a similar story for active foreign equity funds—the fail rates fell 15 to 30 percentage points over the 10-year horizon, but even in the best-case scenario (international small-cap), more than half the assets still fell short by the end.

There were some categories where the underperformance rate fell significantly. For instance, SPIVA reported that more than 61% of active “investment-grade intermediate” bond funds underperformed over the trailing 10 years ended Dec. 31, 2024, but that figure fell to around 24% after the authors made their adjustments.

Yet when you dig into the details, you find there’s quite a bit of noise. For instance, nearly one-fifth of the 37-percentage-point improvement of active investment-grade intermediate bond funds stemmed from replication errors—that is, the authors not being able to exactly reproduce SPIVA’s findings. (The authors couldn’t get closer than a 54% failure rate, meaning the funds the authors examined—before adjustments—were slightly higher performing than the group SPIVA used.)

And only two of those adjustments—crediting dead funds and asset-weighting—actually improved the results. The others—the benchmark adjustment and the interaction effect of applying all three adjustments together—made the results worse.

That seems to underscore the extent to which the improved results hinged on modifications that strike me as either dubious (that is, the treatment of dead funds; more on that shortly) or less than robust (asset weighting).

Dead Funds Are Failures, Period

Maybe I’m too hardheaded, but I don’t see a case for treating a dead fund as anything other than a failure. If we define successful active funds as those that a) make it to the end of the time horizon in question and b) earn a higher return than whatever they’re being benchmarked against, then there’s no way a dead fund can be anything other than a failure, as it flunks condition a.

My concern isn’t purely philosophical, it’s also practical: When a fund dies, it imposes the decision on investors of what to do next. In the case of a liquidation, the proceeds have to be redeployed elsewhere, with all the associated selection risk. With respect to a fund merger, shareholders of the acquired fund must take the acquiring fund’s measure and determine whether it’s worth sticking with. More choices, more problems.

With respect to the paper itself, the authors’ approach to crediting dead funds also seems questionable. They appear to have treated the truncated performance of dead funds, covering the portion of the period they lived, as equivalent to the performance of funds that survived the entire period. And this adjustment had a big effect in some categories like high-yield bonds, of which the authors wrote:

The impact, though, is largest within the fixed-income class. To illustrate, the underperformance rate over a 20-year horizon in the high-yield category decreases from 78% to 52%. Put another way, with this change alone, the typical active high-yield fund becomes about as likely to outperform as underperform.

I don’t have access to the data behind the SPIVA scorecard, but I can proxy for it using our data, focusing on the 150 funds (largest share class of each fund) that were assigned to the high-yield bond Morningstar Category as of Dec. 31, 2004. Seventy-two of those funds survived the ensuing 20 years. Of the 78 funds that died, the average lifespan was 8.3 years, or less than half the 20-year horizon.

Here’s the distribution of those dead funds based on how long they lived and whether they outperformed the Markit iBoxx Liquid High Yield TR USD Index from Jan. 1, 2005, to the month-end immediately preceding the date they died.

Thirty-two of the 78 dead funds outperformed the index between Jan. 1, 2005, and the last month-end preceding the date they died. But more than half the outperformers failed to survive even half of the full 20-year period. All told, the dead funds that outperformed survived 2,844 months in aggregate, while those that lagged lived nearly twice as long: 4,518 months in total.

This explains why, when you track the performance of the average obsolete fund from Jan. 1, 2005, through Oct. 31, 2024 (the last month-end the longest-living dead fund survived through), it badly lags the index.

Take it all together, and the argument for crediting for dead funds doesn’t hold much water, in my opinion.

Asset Weighting Is Capturing Something Else

By contrast, I have less of an issue with asset weighting. My colleagues who publish the Active/Passive Barometer report don’t asset-weight when tallying success rates like the authors did. But they do include asset-weighted average returns for active and passive funds in the report. It’s a useful measure, and the authors are right to cite it.

However, it doesn’t necessarily tell you how deft active funds were in beating their indexes. Or at least that’s not all it tells you. For when you asset-weight, you essentially bake in the choices that investors at large have made about which funds to own. That is, it’s as much a referendum on investors as it is on the managers running those active funds. A cynic might say that it’s a measure of investors’ success despite the fund industry.

What have investors succeeded with? The low-cost funds they’ve favored. You can see this by measuring the average expense ratios of the largest active stock and bond funds by net assets, which I’ve shown below using net assets and expense ratios as of Dec. 31, 2014.

The top 1% of funds in their peer group by asset size—which accounted for 34% of overall assets—were about half the cost of the smallest 25% of funds. So when you asset-weight, you are de facto overweighting the cheapest funds which, in turn, explains why asset-weighted average returns usually exceed equal-weighted average returns.

There can also be unintended consequences when you asset-weight as the authors have, especially in broader peer groups. For instance, the authors found that once you asset-weight, the underperformance rate of active “international” stock funds fell to 66.8% from 85.3% over the 10 years ended Dec. 31, 2024, a nearly 19-percentage-point improvement. (It was actually closer to a 15-percentage-point bump once you factor in replication errors.)

That looks impressive at first glance until you consider that “international” encompasses all manner of foreign large-cap stock funds, from value to growth. Those stylistic differences can be significant at times, and that was true of the decade ended Dec. 31, 2024. Again, I don’t have SPIVA Scorecard data, so I’ll have to proxy using Morningstar data, but here are the annual returns of the average active fund in the three primary foreign-large cap Morningstar peer groups.

The average growth-leaning foreign large-cap fund handily outperformed the average value-oriented fund over this period. Moreover, active foreign large-blend and foreign large-growth funds were far likelier to survive and outperform than foreign large-value offerings.

No doubt, these differences express themselves in the equal-weighted success rate calculation, but they’re potentially further magnified when asset-weighted. Why? Active foreign-large blend and foreign large-growth funds were much larger than foreign large-value funds by assets than by sheer number of funds as of Dec. 31, 2014.

And in that way, asset weighting could further torque a measure—the equal-weighted success rate—that’s already arguably being distorted by stylistic differences among funds.

Changing the Benchmark Is Fine, But It Doesn’t Move the Needle Much

The one adjustment I concurred with was the authors’ modification of the benchmark using the asset-weighted average net return of passive funds in lieu of a costless index. This is a practice my colleagues employ as part of the Active/Passive Barometer.

But I’m not sure it’s necessarily a game changer, and it seems to involve its own nuance. To illustrate, here are the improvements the authors attributed to the use of actual passive bond funds instead of the indexes assigned to those S&P categories.

It’s apparent that substituting in this fashion pushed the underperformance rates lower in most of the categories. But the results were mixed in three of the four biggest peer groups by number of funds—investment-grade intermediate funds, investment-grade short and intermediate, and global income—where the underperformance rate rose, not fell, over the one-, three-, five-, and 10-year periods ended Dec. 31, 2024.

Also, the degree to which these improvements varied over the trailing periods seems to reinforce the sensitivity of this input to the mix of assets across passive funds—which may or may not correspond to the cross-section of styles among active funds in the same peer group—and how it might change depending on the period being examined.

Conclusion

I found this to be a thought-provoking paper, and I appreciate its focus on how methodological choices can affect our assessment of active-fund success. It’s good to have research like this that’s trained on practical questions that investors, advisors, and allocators grapple with in real life.

Also, the authors aren’t wrong about active management having some merit, especially when it’s delivered at low cost, something that comes through loud and clear when they asset-weight the results, thereby placing heavier emphasis on the cheaper active funds that have been popular with investors. For our part, our analysts assign Morningstar Medalist Ratings of Gold, Silver, and Bronze to scores of active funds the world over, reflecting their conviction in these managers to outperform a benchmark after fees.

Having said that, I just didn’t find the argument persuasive. Even when the authors made all three adjustments—crediting dead funds, asset-weighting, and using actual passive funds—it didn’t change the picture that dramatically. I don’t buy the argument for crediting dead funds—failures in my book—if they outperformed over their lives, and while there’s a case for asset weighting, it seems to capture something besides manager skill while also courting its own issues.

In all, a worthwhile study, even if I wasn’t convinced by the arguments and conclusion.

Switched On

Here are other things I’m reading:

• The Long View talks to Bill Bengen (he of the “4% Rule”)
• “Investors saved an estimated $6.8 billion in fund expenses last year”

Don’t Be a Stranger

I love hearing from you. Have some feedback? An angle for an article? Email me at jeffrey.ptak@morningstar.com. If you’re so inclined, you can also follow me on Twitter/X at @syouth1, and I do some odds-and-ends writing on a Substack called Basis Pointing.