Key Takeaways
- The Global Industry Classification Standard (GICS) was created in 1999 and assigns every public company to a single sector based primarily on its largest revenue source.
- Company misclassifications distort sector-level analysis and the ETFs built on top of it.
- A purpose-built taxonomy — one that reflects how companies actually generate value — produces a fundamentally different, and more accurate, view of a portfolio's true exposures.
When MSCI and Standard & Poor's launched the Global Industry Classification Standard in 1999, the timing made sense. The dot-com boom had created genuine confusion about how to categorize a new generation of technology companies. Investors needed a framework. GICS provided one.
That framework has not fundamentally changed since then. The same 11 sectors, 25 industry groups, 74 industries, and 163 sub-industries structure that organized markets in the early 2000s is the same structure that organizes them today. The world's businesses, however, are not the same.
The gap between how GICS describes modern companies and how those companies actually operate has become large enough to matter.
One Company, One Box — Even When It Doesn't Fit
The defining rule of GICS is deceptively simple: a company is assigned to a single sub-industry based on what its principal activity is that drives revenues.
While that rule worked reasonably well when companies had reasonably coherent business models, it works less well when a single company is simultaneously one of the world's largest retailers, the dominant force in cloud infrastructure, a major streaming entertainment provider, an advertising platform, and a logistics operator.
For instance, Amazon is classified as Consumer Discretionary, and it has remained there despite Amazon Web Services (AWS) becoming the company's primary profit driver — generating more operating income than the entire retail segment. An investor holding a Consumer Discretionary ETF that includes Amazon believes they have consumer spending exposure. What they actually have, in a meaningful portion, is cloud computing exposure, which is a fundamentally different bet.
Amazon is the most visible example, but not the only one. Alphabet's advertising revenue earns it a place in Communication Services; Microsoft is Technology — but Azure, Office 365, LinkedIn, and Xbox make it simultaneously a cloud business, a productivity software company, a professional network, and a gaming platform.
The labels capture only a portion of the whole picture.
When the System Updates, It Creates New Problems
To its credit, the GICS system does update. But the updates themselves illustrate the problem. In March 2023, MSCI and S&P reclassified 14 firms across five sectors. Visa, Mastercard, PayPal, Fiserv, and others — previously classified as Technology — moved to Financials. The Technology sector, which had been anchored by these companies, became even more concentrated in Apple and Microsoft.
For investors, that reclassification was not a minor administrative update. An advisor holding a Technology ETF before March 2023 woke up after the change holding a fundamentally different set of exposures in a fundamentally different sector — not because any of the businesses changed, but because the taxonomy did.
A classification system that requires periodic large-scale corrections is not a stable foundation for portfolio construction.
What It Means for ETFs Built on GICS
The implications run directly into the ETF market. The majority of sector ETFs — the funds that let investors express precise views on specific parts of the economy — are structured around GICS. The Technology Select Sector SPDR, the Consumer Discretionary Select Sector SPDR, and their equivalents track GICS-defined sectors. An investor who buys a Consumer Discretionary ETF because they want exposure to consumer spending is, in some portion, buying Amazon Web Services.
This affects investors at the portfolio level as well. Advisors using GICS believe they are building a diversified client portfolio with sector ETFs, but if the underlying classification system creates artificial concentrations — piling cloud computing exposure into three different sectors, for instance — the portfolio's actual risk profile looks quite different from its stated one.
In other words: the fund is fine; the index is fine; the taxonomy underneath them is not fine — not for an economy where the most valuable companies have become genuinely multi-sector businesses.
What a Better Taxonomy Looks Like
A classification system built for today's market would allow for several things that we cannot do with GICS:
- Multi-dimensional classification — acknowledging that a company can generate meaningful revenue from cloud computing and e-commerce simultaneously, and representing both in its profile.
- Update continuously, not in annual or semi-annual cycles.
- Reflect actual revenue composition rather than historical identity.
It would also be designed specifically for the question investors are actually trying to answer: what am I exposed to, and how much? Not: which box does this company belong in?
While GICS was the right tool for 1999, it does not work for today. Portfolio construction, risk management, and sector analysis all change when the underlying taxonomy is precise enough to reflect the actual structure of modern businesses. We need to re-evaluate our tools and ask whether the portfolios built on top of it are actually representing what their owners believe they represent.
The answer, for most advisors who look closely enough, is no.
Langar Technology, Inc. — Redefine Investing. For more information on our proprietary classification system, email hello@langartech.com.