Two fund managers walk into a scorecard
A note on my own position first, since that turns out to matter for this topic. I'm a licensed financial adviser. A meaningful part of my livelihood depends on people continuing to find low-cost index funds a sensible place to put their money. You should weigh what follows accordingly. I'd suggest applying the same test to everything else you've read on this subject in the last fortnight.
In the past week, two people employed by New Zealand active fund managers have published pieces arguing that the SPIVA scorecard (the research S&P Dow Jones Indices has produced since 2002 comparing active funds against market indices) makes active management look worse than it is. Anthony Edmonds of Aurellon and Greg Smith of Generate. You can find the Aurellan article here and the NZ Herald article here.
They are not entirely wrong. Parts of both arguments are good and the passive case doesn't need to pretend otherwise.
Where they have a point
The benchmark is a real problem. SPIVA NZ measures our global equity funds against the S&P World Index. That index covers 24 developed markets. Many New Zealand managers run mandates with emerging markets exposure, ESG exclusions, or a different geographic split. Measuring a fund against an index it was never built to track produces a gap that has nothing to do with skill.
Worse, the S&P World Index (NZD) and its hedged sibling were only launched on 25 October 2024. Every figure before that date is back-tested. S&P discloses this in a footnote, to their credit, but it means the long-run history is a reconstruction, not a record.
The categories are too broad. With only four reported categories covering the whole country, SPIVA NZ has to sweep infrastructure funds, clean energy funds, thematic strategies and broad global managers into the same bucket and measure them all against one index. Some of those funds were never trying to track a broad developed market benchmark. In a market our size, you can't build clean cohorts without ending up with samples too small to publish.
No one can buy the index. SPIVA compares active funds to a costless mathematical abstraction. Real index funds charge fees and have to trade. Morningstar's Active/Passive Barometer has used real investable passive funds as the comparison for years, and it's the right approach. It's also a small adjustment. Single digit basis points in the US. Typically a bit more here.
Some dead funds aren't failures. SPIVA treats every fund that closes during a measurement period as an underperformer. In a market as consolidated as ours, plenty of funds have closed because a business was sold or a platform migrated, not because the manager was bad. That's a genuine flaw and a specifically New Zealand one.
If someone gets a proper study done on this, and there is talk of exactly that, it would be worth having.
Where it falls apart
The 86% figure is not evidence of anything.
You may have seen a claim in the referenced articles that 86% of New Zealand passive global share funds underperformed the same index, offered as proof that the methodology is broken. It appeared in the Herald as "a separate analysis found."
Here is what that analysis was. Fourteen funds. A three year window ending 31 July 2026, chosen by the author. A benchmark deliberately selected because no New Zealand fund tracks it. Two of the fourteen were Kernel sector funds that the author openly says shouldn't have been included. And the original author has stated, repeatedly and in writing, that he does not believe the result and published it only to demonstrate that the method is flawed.
That's a legitimate rhetorical device in its original context. It is not a finding. By the time it reached a national newspaper it had lost the author's name, the sample size, the disclaimer and the construction, and had become a fact sitting alongside a recommendation to consider active managers.
86% of 14 funds is 12 funds. It is not a statistic. It is a demonstration, and the person who built it says so. Not to mention the carefully crafted starting point and short timeframe.
The horizon argument is made and then abandoned.
The Herald piece argues, correctly, that longer periods tell a more honest story. That a decade spanning different market environments is a better test than one favourable stretch, and that rolling ten year returns are the demanding test.
It then cites Morningstar's Mid-Year 2026 Active/Passive Barometer for the claim that just over 40% of active funds beat their passive peers, and 27% in US large-cap. Both of those are one-year figures, for the twelve months to June 2026.
The Barometer publishes ten year success rates in the same report. On the same page even, which the NZ Herald article conveniently left out. 10 year results are considerably worse for active management. Over the decade to June 2026, 25% of active strategies survived and beat their passive counterparts. In US large-cap, the category the 27% figure came from, the ten-year rate drops significantly to 13%.
So an article arguing that a decade spanning different market environments is the honest test, did the opposite and reached for 40% and 27% (one year data) when 25% and 13% (10 year data) were printed right beside them.
An article whose central argument is that short windows mislead should not be reaching for the shortest window available.
"Five of the six most consistent performers are active managers."
This is the piece's only New Zealand evidence, and it arrives with no source, no fund universe, and no definition of "consistent."
It also has a problem that ought to be obvious to anyone who has watched this market. Most New Zealand passive options are younger than ten years! Simplicity launched in 2016. Kernel in 2019. Foundation Series only in 2022. A ten-year New Zealand ranking is dominated by active managers in large part because passive funds weren't available to be ranked.
Fund managers have spent a week telling us that SPIVA's survivorship treatment distorts the picture. This is survivorship bias of a purer kind, deployed in the opposite direction, in the same article.
Bar chart comparing active fund success rates over one year and ten years. All active funds: 40% over one year, 25% over ten years. US large-cap: 27% over one year, 13% over ten years. Source: Morningstar US Active/Passive Barometer, mid-year 2026.
Persistence is never mentioned.
This is the hole in the middle of the whole argument. The case being made is not "the market is inefficient." It's "find the managers who can consistently convert opportunity into results."
That claim requires evidence that past outperformance predicts future outperformance. S&P publishes a separate Persistence Scorecard for precisely this question, and its findings are consistently unkind. Top quartile managers rarely stay top quartile. You can grant every methodological criticism above and still be left with the problem that you have to pick the winner in advance.
Softening 80% underperformance to 60% doesn't help an investor who has no reliable way to identify which 40% to buy.
Of the active US large growth funds that existed two decades ago, 66% (two thirds) have closed, and fewer than 1% both survived and outperformed their average indexed peer.
The Schroders numbers are net of fees. But whose fees?
The article cites Schroders research finding the median global large-cap manager added 0.5–1% a year over 10–20 years. My first assumption was that this was gross of fees, as manager composite data often is. It isn't. Schroders state that they measure active performance against indices net of fees. Credit where it's due. That's the harder test and it makes the finding more interesting than I first allowed.
There are two things to keep in mind however.
The first is whose fees. Composite data of this kind reflects institutional mandates, where pricing is a fraction of retail. A median manager clearing 0.5–1% after an institutional fee tells you very little about a New Zealander paying 1.09% in a KiwiSaver fund. The distance between those two fee levels is wider than the entire claimed outperformance.
The second is what's in the sample. Manager composites are self reported, and a firm chooses which composites to submit and when. That's a different bias from the one SPIVA is accused of, but it isn't nothing.
And it's worth seeing what Schroders themselves do with the argument. Applying a 0.30% passive fee to UK equities over 26 years, they note the index fund would have trailed its benchmark. So, in their framing, 100% of passive funds underperformed.
It's the 86% again. A passive fund trailing its index by roughly its fee is the product doing exactly what it says on the tin. An active fund trailing by considerably more, having charged considerably more for the attempt, is not the same event. Collapsing the two into one underperformance count is the active fund manager move, and it keeps turning up.
The paper everyone is quoting
Both pieces lean on a 2026 working paper by Cremers, Fulkerson and Riley arguing that SPIVA understates active performance.
It was sponsored by the Investment Adviser Association's Active Managers Council, the active management industry's lobby group. That doesn't make it wrong. Cremers in particular is a serious academic. But it is a disclosure that has been missing from every New Zealand retelling I've seen, in articles whose central complaint is about undisclosed bias.
And Morningstar's Jeffrey Ptak, who actually reviewed it, wasn't persuaded. He doesn't accept crediting dead funds. A failure is a failure, in his view. He thinks asset-weighting captures something other than manager skill. His conclusion was that all three adjustments applied together "didn't change the picture that dramatically," and that even on the paper's own generous treatment, roughly two-thirds of active US stock fund assets still failed over the decade to the end of 2024.
There's a detail in his review worth knowing too. When the authors substituted real passive funds for indices in fixed income, underperformance rose in three of the four largest peer groups. The adjustments don't uniformly favour active management. They're just reported that way.
I'd stop short of Ptak's position, though. A fund that closed because a business was sold or a platform was retired isn't always a failure of management, and treating it as one is wrong. Ptak's rule is too absolute.
But Cremers' fix runs to the opposite extreme, and it's the more distorting of the two. Crediting a dead fund with its performance up to the day it closed assumes the investor's experience ended there. It didn't. The money had to go somewhere, often at a moment not of the investor's choosing. And the ordinary reason a fund closes is that assets bled away after a stretch of poor performance. Corporate tidy ups are the exception, not the rule.
So the honest answer sits between the two, and closer to Ptak than to Cremers. The treatment nobody in this debate uses would be to chain a dead fund's return to its category average, or to the fund that absorbed it, from the closure date onward. That tracks what actually happened to the money. It's also considerably more work than either side has done, which is probably why neither has done it.
What none of this touches
Strip out every disputed number and two things remain.
The arithmetic. Before costs, the aggregate of active investors must earn the market return, because collectively they are the market. After costs, they must earn less. This isn't a study anyone can re-specify. It's arithmetic, and it doesn't depend on SPIVA existing.
The selection problem. Even in a world where 45% of active managers beat their benchmark, an investor still has to identify them beforehand, pay for the attempt, and be right often enough to cover the cost of being wrong. Nothing published in the last month makes that easier.
In Morningstar’s active/passive barometer report, they quantifiy the second half of that. In US large-cap, the distribution of ten-year excess returns among surviving active funds skewed negative. Meaning the penalty for picking a losing manager outstripped the reward for finding a winning one. Low odds are one problem. An unfavourable payout on those odds is a separate one, and it compounds.
A word on celebrating 40%
Just over 40% of active funds beat their passive peers. That is being reported as encouraging news. In what other field is a 40% success rate a result you'd put in a press release? Not surgery. Not aviation. Not construction. You would not board a plane on those odds and you would not sign off a building on them.
The reference point is what makes it work. After a decade of 80–90% failure figures, 40% success reads as a recovery. The Herald piece even notes it's up seven points on the year before, which quietly converts a losing rate into a trend line.
And the 40% is doing less work than it looks. It's a single year. It's measured against a passive peer average, which roughly half the field clears by construction. And it sits before the investor's actual problem, which is selection. You don't get to buy a 40% chance. You buy the specific fund you chose, having paid extra for the attempt.
The reason this survives in funds management, and wouldn't survive anywhere else, is that the counterfactual is invisible. Nobody sees the index fund they didn't buy. If your active fund returns 7% you feel fine about it, even if the market returned 9% and you paid 1% a year for the privilege of missing it. There's no moment where the loss announces itself. It just quietly isn't there at the end.
The argument that eats itself
There's one more thing in Ptak's review worth pulling out, because it undoes a good deal of what's been built on top of it.
In conceding what he could to the paper's authors, Ptak wrote that they aren't wrong about active management having some merit, especially when it's delivered at low cost, and that this comes through loud and clear when they asset-weight the results, because asset-weighting places heavier emphasis on the cheaper active funds that have been popular with investors.
Asset-weighting is the single biggest driver of the 92% to 55% shift both New Zealand articles have been leaning on. And what asset-weighting actually surfaces is that large, cheap active funds have done relatively well. The headline number being used to argue for active management is, at its foundation, a finding about low fees.
Which is what the passive case has been saying the whole time. Cost is the one variable in this debate that behaves predictably. You can re-specify a benchmark, argue about dead funds, and swap fund count for asset-weight. You cannot re-specify what you paid.
And this isn't an inference. The Barometer measures it directly. Over the ten years to June 2026, 33% of active funds in the cheapest quintile of their category beat their average passive peer, against 20% in the priciest quintile. In US large-cap the gap is wider: 22% for the cheapest, 9% for the most expensive. Morningstar's own methodology section puts it plainly. Fees are one of the best predictors of future fund performance.
In the aggressive category, Generate's Focused Growth fund charges 1.25% a year, against a Morningstar category average near 0.95%. Kernel's High Growth fund, same category, charges 0.25%. Simplicity charges 0.23% across its entire range from 1 October. In the growth category, Generate sits around 1.09% against a category average of 0.97%. There is also a membership fee on top of Generate's.
So an above average-cost manager (Generate) is citing research whose favourable finding is driven by emphasising low cost funds. I don't think that's deliberate. I think it's what happens when a result gets quoted at headline level without the sentence underneath it.
In Generate's own words, their KiwiSaver supplementary brochure (page 6) states that over 90% of their members joined through an adviser, and that over 80% of members' funds sit in growth funds against a market average of 46%.
The second figure matters for the arithmetic. The fees most Generate members actually pay are the 1.09% and the 1.25%.
The fact that over 90% of Generate’s members joined through an advisor (aka commission) matters for what the fee buys. In an adviser distributed product, the annual charge covers distribution/commission as well as investment management. That's not a criticism of the model, and I'm hardly placed to make one. But it is part of what a fee comparison is actually comparing, and it's a reason a like-for-like number against a self-service provider isn't quite like-for-like.
Final thoughts
SPIVA NZ is noisier than the confidence with which it gets quoted. Two editions, four categories, a benchmark barely two years old, and a fund universe of perhaps seventy. The headline number moves a long way between readings: 83% of global equity funds underperforming in the calendar year 2024, 74% in calendar year 2025, and just 40% over the first half of 2025. The half year figure isn't like-for-like with the annual ones, but that's rather the point. The answer depends heavily on which window you pick up. Anyone citing it as settled proof of anything is over reading it, and that includes people on my side of this.
But "this measurement is imprecise" is not the same claim as "therefore choose active management." The first is a methodology argument, and a decent one. The second is a sales conclusion, and it requires evidence about manager persistence and net-of-fee returns that neither article provides.
If the proposed New Zealand research gets done properly with pre-registered methodology and sensitivity analysis on the big specification choices, I'll write about it here and I'll take whatever it says. That would be a real contribution.
The Barometer's clearest finding isn't about active versus passive at all. It's that cheap funds beat expensive ones, inside active management as much as outside it. If someone wants to go active, the evidence says go cheap. Low turnover, low fee, long horizon. That's a defensible position and I'd have no argument with an adviser who took it. What it isn't is a case for paying a percentage point more than you need to.
Until then, what we have is two fund managers, a 14 fund demonstration nobody stands behind, a one-year figure in an article about 10 year horizons, and a ranking of New Zealand's most consistent performers that quietly excludes most of the competition for not having existed yet.