In a recent Wall Street Journal ‘Intelligent Investor’ column, Jason Zweig showed how index funds that are supposed to be near-identical twins are performing like strangers. The iShares Russell 1000 Growth ETF was up about 4.6% this year, while the Invesco S&P 500 Pure Growth ETF had gained 24.3%. Among the leading large-value ETFs, the spread ran to 11 percentage points. Much of the difference came from classification. Micron Technology and Sandisk, two chipmakers riding the AI boom, counted as value stocks under Russell’s price-to-book math because factories and equipment inflate book value. FTSE Russell says the pair drove roughly a fifth of the Russell 1000 Value Index’s return before the June reshuffle moved them to growth. Zweig warned that, “the finer you slice-and-dice the market, the more concentrated it becomes.”
I’ve spent decades designing index methodologies and performing due diligence on the indexes that underpin investment products. At S&P Dow Jones Indices, I ran U.S. equities, including the pure style family behind that Invesco fund, so I know firsthand how gaps like these open up. The methodology determines what a fund’s label means, yet most investors never get past the name.
The word “pure” in that fund’s name describes a construction most buyers never look into. A traditional style index like the S&P 500 Growth classifies the whole parent index, holds every stock with growth characteristics even when it also scores as value, and weights the holdings by market cap, so it behaves like a mildly tilted version of the market. The pure version keeps only the stocks with the strongest growth scores, removes the middle where names score as both growth and value, and weights what remains by growth score. That puts the most growth-like stocks at the top rather than the biggest companies. The result is a more style-pure portfolio that excludes blended stocks and gives the greatest weight to companies with the strongest growth scores. Anyone who reads the two methodologies should expect it to move differently from the Russell 1000 Growth. Investors often miss that construction difference, and also often miss the case where a prospectus swaps one benchmark for another while the fund’s name stays put.
That gap between a fund’s label and its methodology becomes more consequential in thematic investing, where differences in eligibility, weighting, and reconstitution can produce entirely different exposures under the same broad label.
I lived through one kind of label drift in 2018, when S&P Dow Jones Indices and MSCI redrew the sector map – an effort I helped lead at S&P. Telecom expanded from a handful of sleepy dividend payers, about 2% of the S&P 500, to a communication services sector that held Alphabet, Facebook, Netflix, and Disney, and made up roughly 10% of the market. Anyone holding a telecom fund for income woke up in a growth sector. Another kind occurs when the definition stays fixed while the companies change underneath it, as happened with Micron. Its factory-heavy balance sheet kept it scored as value while the business became an AI trade. Either way, the fund’s name stays familiar while the exposure changes.
Digital assets were an extreme version of this challenge. Before I built the first classification standard for them at CoinDesk, there was no established framework – only thousands of assets and no consistent way to distinguish a smart contract platform from a payment token or a meme coin. You can’t perform due diligence on an exposure that hasn’t been defined. Building that standard reinforced the need to define an exposure precisely before deciding which constituents belong in an index.
That same discipline matters when an index turns a theme into an investable universe. For a theme like space, revenue provides a clear test of whether the theme is central to a company’s business or merely incidental. In a recent Syntax Data paper, I analyzed the Syntax Space Industry Index, which requires a company to generate at least half of its revenue from space activities and to pass liquidity and tradability screens. Its 16 U.S.-listed constituents average roughly 95% of revenue generated from space-related activities, and 12 have 100% space exposure. By comparison, the S&P Aerospace & Defense Select Industry Index averages about 15% space revenue, while the S&P Kensho Final Frontiers Index averages about 22%. Three indexes with similar titles can therefore produce very different compositions, with different return drivers and risk profiles. Investors need to compare the eligibility rules and weighting methods behind them before choosing one, because those decisions determine what they actually own and how it is likely to behave.
Zweig’s warning still applies, and anyone marketing a thematic index should be clear about the concentration that follows. A strict revenue screen leaves only 16 companies. Equal weighting limits the effect of any one name or launch failure, but in its modeled history this index still ran annualized volatility near 70% and fell almost 49% from its high – more than twice the volatility of the broader benchmarks. For an investor using it, that level of risk may argue for a modest satellite allocation sized to withstand a severe drawdown.
When a stock enters an index is a methodology decision too, and in a fast-moving theme it can matter a great deal. Invesco’s pure-growth fund caught the chip rally partly because its December rebalance swept in the right names at the right moment. SpaceX’s June listing made the same point on a larger scale. The space index I analyzed added it effective June 15 at its first-day close, under a provision for IPOs with a market cap of at least $250 billion and at least half their revenue from space. Another space index added it a day later, FTSE Russell waited for its late-June reconstitution, the Nasdaq-100 added it after 15 trading days, and the S&P 500 requires a year of trading history before it will consider a new listing. The same stock entered those indexes on different dates and at different prices, while it traded anywhere from about $150 to $225 across those weeks, and every rule that decided it was published in advance for anyone who looked.
Before buying an index fund, read enough of the methodology to understand what qualifies, how constituents are weighted, what prevents one stock from dominating, when the index reconstitutes and whether the performance history is live or backtested. For a thematic fund, pay particular attention to whether eligibility rests on a broad label or a measurable threshold such as revenue. If the document can’t answer those questions in a page or two, investors are being asked to rely too heavily on the fund’s name.
Zweig ends by asking whether investors need to divide the market into growth and value at all, since holding broad growth and value funds together can largely recreate the market. Pure style produces a different portfolio. Pure growth and pure value exclude the blended middle and together cover only about half of the parent index. They also weight stocks by the strength of their style scores, so holding both still emphasizes companies with the strongest growth and value characteristics. The same relationship between purity and concentration applies to thematic indexes. The methodology determines which companies qualify, how much of their business must come from the theme, how they are weighted and when they enter or leave the index. As those rules define a narrower slice of the market, the portfolio becomes more concentrated. That concentration changes which companies drive returns, how the fund behaves across market conditions and the risks investors take. Owning a thematic fund therefore means owning both the intended exposure and the distinct performance and risk profile created by that narrower portfolio.
Jodie Gunzberg, CFA is Managing Partner of InFi Strategies, an independent advisory practice serving asset managers, index providers, and wealth advisors on investment strategy, product commercialization, and index design. She is also a consultant with Syntax LLC, a New York-based fintech company pioneering economic data intelligence and next-generation index technologies.
