Is the impact of Mega Caps as big as ever?

• The diversification of many leading indices is lower than ever before.
• Globally, it is even lower than before the dotcom bubble burst.
• The risk-return contribution of the top 10 stocks is also higher than ever.

Comparing diversification of different benchmark indices

Against the odds of slowing economic growth, high interest rates and stubborn inflation, equity markets performed significantly well globally, especially in the US. However, the performance is largely concentrated on a few mega caps. For example, a good 70 percent of this year's performance in the S&P 500® can be attri­buted to the top 10 stocks in the index, primarily from the tech­nology sector. A corresponding dominance of the technology sector and high concentration of the top 10 positions in the S&P 500® index is reminiscent of the initial scenario of the dotcom bubble and increasingly calls into question the diversification of the index.

The sole consideration of the top 10 weighting may be an indica­tion, but ultimately has only limited significance with regard to the diversification of an index. In addition, comparing the diversi­fication of several indices to each other is only possible to a limited extent.

For this reason, the Effective Number of Constituents (ENC) is used as a diversification measure for the analysis. Stocks are typically weighted according to their market capitalisation, such that an index ultimately has a few very large positions and hun­dreds of stocks with weights of a few basis points, which have little influence on the risk-return characteristics and thus do not represent an effective position. In the context of an equally weighted portfolio, however, the situation is different, because each position has the same influence on the risk-return profile of the index and thus represents an effective position.

The ENC can be used to determine the number of effective positions for each equity index, irrespective of its weighting scheme, and thus allows comparisons between indices with regard to their diversification range.

Tab. 1: Diversification of different indices

Table 1 shows that a high number of stocks does not necessarily result in a higher diversification. For example, although the MSCI World has over 900 more stocks than the STOXX® 600, the differ­ence in the number of effective positions is marginal.

Another example is the MSCI All Country World: the number of stocks doubles compared to the MSCI World to almost 3,000, but the number of effective positions only increases from 133 to 163. This is because more than 1,500 stocks, especially from develop­ing countries, have a weighting of less than one basis point, so they have almost no influence on the performance of the index.

Indices are less diversified than during the time of the dotcom bubble

A closer analysis of the ENC for the MSCI World in Figure 1 shows that the index is even less diversified than shortly before the dotcom bubble burst. In the early 2000s, technology stocks occupied a similarly dominant index position, but there were a number of other "conventional" index heavyweights, such as Exxon Mobil or General Electric.

The question is, what are the ultimate implications of the current high index concentration? On the one hand, it is important for an active strategy to maintain diversification, which implies a high number of effective positions. On the other hand, the value driv­ers of the benchmark must be taken into account. Although an equally weighted index represents the highest possible diversi­fication according to ENC, this naturally leads to significant factor biases (especially size) compared to market capitalisation-weighted benchmark indices.

Fig. 1: Development of the number of effective positions in the MSCI World-Index

Assenagon Equity Framework

In order to work out the pure effect of the top 10 positions of the global stock market, two portfolios are constructed for the re­spective quarterly index rebalancing whose relative performance against the global stock market is shown in Figure 2:

• Global stock market ex top 10 stocks uncontrolled, i.e. there is a size bias, in particular compared to the original index.
• Global equity market ex top 10 stocks controlled, i.e. a factor-, sector- and country-neutral allocation is maintained against the original index.

Fig. 2: Realtive development compared to the global stock market
(January 2000 – July 2023)

Interestingly, both portfolios, regardless of whether controlled or uncontrolled, perform very similarly for a long time. Only the relative risk, in terms of tracking error against the global equity market, is significantly lower in the case of the controlled version. This results from the holistic portfolio construction, which neutralises factor, sector and country effects compared to the original index. In absolute terms, both portfolios outperformed the global equity market by around 10 per cent until 2015, al­though from that point onwards both portfolios underperformed the global equity market.

From 2018 onwards, the relative performance of the two port­folios decouples, with the uncontrolled portfolio performing significantly worse, owing to the size bias. The controlled portfolio on the other hand, is neutral to the original global equity market in terms of average market capitalisation and therefore has no size bias. The controlled portfolio thus shows the pure effect of the top 10 positions of the index.

The result shows impressively that there has been a strong idio­syncratic risk-return effect from the top 10 stocks on the global equity market since 2018. The comparison between the two portfolios shows that approximately half of this effect is due to the size factor and the other half is purely idiosyncratic. Another highly interesting fact is that there was not too much difference between the two portfolios from January 2000 to 2018. In other words, the pure effect of the top 10 positions adjusted for all factor, sector and country effects was not really significant until 2018.

For the investor

The index concentration is currently stronger than ever before. In addition, the idiosyncratic effect of the mega caps, i.e. the risk-return contribution of the top 10 positions to the global equity market, is higher than ever before.

Benchmark-oriented investors are thus faced with major chal­lenges, as this idiosyncratic effect has shown high volatility, especially in recent years. The tracking error, which results purely from the missing top 10 positions, has more than doubled since 2018, as can be seen in Figure 2, and has also brought phases of significant underperformance with it.

In order to cope with this problem, it is advisable, especially for benchmark-oriented investors in the context of active strategies, to overweight or underweight equities relative to the benchmark instead of leaving them out altogether - on the one hand, this reduces the impact of the idiosyncratic mega cap effect and, on the other hand, it reduces the relative risk.

PS: Read about, whether the positive 2023 performance of the stock market was a purely idionsyncratic mega cap effet, in the next equity insight.