Diversification benefits of thematic equities

It has been a volatile start to the year for global financial markets – the kind of environment that amply demonstrates the value of having a well-diversified portfolio.

On the face of it, the MSCI All Countries World index (ACWI) seems to tick that box. It features over 2,600 large and mid-cap companies based in 50 of the world's advanced and emerging economies. 

However, our analysis shows that, much like US stock indices, the ACWI index has also become increasingly reliant on the returns generated by a small handful of stocks, mainly in the tech sector. ACWI’s skew is reflected in its performance. Last year, the so-called Magnificent Seven grouping of tech stocks – Alphabet, Amazon, Apple, Meta, Microsoft, NVIDIA and Tesla – were responsible for 45% of the index’s return. That share is even greater than might be expected given their cumulative average weight in the index of 19%.

That concentration is becoming a problem. In our view, the turbulence we have seen in these stocks in recent weeks (such as a 17% one-day drop in NVIDIA shares in January) and their collective year-to-date underperformance of global equity markets by over 10% as of mid-March serve as reminders that their dominance is unlikely to last forever. The equity market is likely to rebalance to better reflect economic fundamentals, secular trends, earnings prospects and company dynamics.

When that happens, index-tracking investors may find their portfolios aren’t as diversified as they thought – while actively-managed portfolios might be better placed to win out. 

To see how diversified (or not) ACWI really is, we have compared it with our Global Megatrends Selection (GMS) strategy, which offers broad exposure to thematic equities. In first approximation, ACWI, with 2,645 constituents would appear to be more diversified than GMS with just 418 holdings.As of 31.01.2025 However, this neglects a key component of diversification, namely the distribution of  stock weightings across constituents. From this vantage point, GMS in fact has a much more even distribution of weights than ACWI, where mega caps dominate (see Figure 1).

Another way to gauge the diversification of either a portfolio or an index is to use a methodology known as the effective number of constituents (ENC). This metric gives an indication of index concentration ranging from one (signalling the dominance of a single stock) to the total number of constituents (suggesting an equally-weighted portfolio). ENC essentially reveals for a given portfolio the number of constituents of a hypothetical equal-weighted portfolio that would provide the same amount of diversification as the original one. It is calculated using the inverse of the Herfindahl index, which is itself the sum of squared portfolio shares. The ENC is smaller than the actual number of constituents, except for an equally-weighted portfolio.

Our analysis shows that the effective number of constituents in ACWI has been steadily declining over the past decade and a half, and now stands at just 125. Despite having significantly fewer stocks, GMS actually has a higher ENC than ACWI, suggesting that the smaller portfolio is more diversified than the index.

Now, some investors might argue that only looking at portfolio weights merely scratches the surface. Diversification, they would say, is primarily related to how the stocks in a portfolio behave, whether that’s their correlations to the index and one another or their sensitivity to common factors (such as market, sector or interest rates).

It’s a valid observation. One way to assess the diversity of a portfolio’s holdings is to conduct a principal component analysis (PCA) – a statistical technique that seeks to identify the main factors (known as “principal components”) that explain the variation of stock returns. 

In another analysis, we applied PCA to both ACWI and GMS using the last 12 months of daily stock returns and weighing the stocks by their share in the portfolio or the index. The results showed that for GMS, only a relatively small proportion of performance (34%) was driven by the top three PCs, suggesting high levels of diversification and greater resilience to market shocks. For ACWI, the share was 75%. Data covering period 31.12.2023-31.01.2025.

Moreover, beyond GMS, Pictet Asset Management’s 16 other thematic equity portfolios all ranked better than ACWI based on a PCA, suggesting that, on this metric, they are more diversified than the reference index.

ACWI’s skew is reflected in its performance. While the overall index returned 17.5% in the period under review, taking a simple average of the performance of all the stocks in it was just 5.5%, and the median was 1.6%. Furthermore, nearly half of the stocks – 47% – actually finished 2024 in the red, even as the overall index posted one of its best years on record.

Interestingly, investor behaviour is beginning to reveal some concern about the overly concentrated nature of market-cap-weighted indices, with equal-weighted passive strategies seeing record inflows recently.

Our study shows that, after a three-year outperformance streak, the odds of the sector underperforming in the subsequent few years are about even. After a five-year winning streak, the probability of a long period of underperformance rises to 60%. These results augur badly for parts of IT and communication services sectors that have outperformed in five of the past six years (with 2022 being the exception). This highlights the need for stock selection in those areas.

Of course, the probability of underperformance is only one aspect of the risk stock investors face. The magnitude of underperformance in cases when it occurs is another. In our analysis, a reversal could see the former stock market stars lose 10-15% in one year, and over 30% if the underperformance persists over three years.

Clearly, when a reversal in market leadership occurs, it can be significant. We saw some evidence of this in January, when the launch of China’s DeepSeek artificial intelligence assistant sent shockwaves through the tech sector. US President Donald Trump’s tariffs brought further pain to some previously popular stocks, with investors choosing to take profits.

While these market moves may, of course, prove short-lived, we believe a broader rebalancing could well be on the cards. The Magnificent Seven’s dominance was in large part down to their strong earnings growth but also reflected a macroeconomic environment that was negative for other industry sectors – interest rates staying higher for longer, and global weakness in manufacturing. While our strategy team sees continued investment opportunities around the widespread adoption of AI, they do not expect the theme to be the dominant factor in stock markets over the medium term – and for four key reasons:

1. It’s no longer all about ‘growth’. Our strategy team believes that growth as a style is unlikely to outperform strongly, with country, sector and stock fundamentals mattering more than style rotations driven by central bank policy. Indeed, for the first time in three years, the secular argument is balanced between growth stocks and value stocks. The latter will be boosted by government policy through near-shoring, reindustrialisation of developed economies and selective energy and banking deregulation.
2. Good news already priced in.  The future growth of Magnificent Seven companies is already baked into their share prices. The group is close to 33% of the US market but only 23% of its earnings, so current valuations already reflect a period of continued earnings outperformance. The hurdle to positive surprises is thus high; in contrast our strategy team expect the gap between the tech giants and the rest of the market to narrow – something we already seeing during the current reporting season.
3. Macroeconomic instability will matter. While they are well-run companies with solid free cash generation, moats around their business models and structural growth prospects, they are not immune to economic growth scares – and we expect to see more of these over the medium and longer term as US growth slows towards trend.
4. Selection is key. While the market has so largely treated the seven as one entity, they are in fact quite different companies with different industry dynamics and won’t always perform in sync with each other. Furthermore, the DeepSeek news showed that AI leadership is not just about a capex ‘arms race’ but also about who can put available technology to best use – in other words, the incumbents do not have a monopoly on innovation.

As the global equity performance becomes more balanced, we are likely to see the broad market-cap weighted indices such as ACWI begin to underperform the equity market at large. That will present an opportunity for actively-managed strategies to take the lead and show off their diversification credentials.