This article is republished in the S&P Dow Jones Indices Index blog.
After the dominant performance of large-cap technology stocks in 2020, concentration problems naturally come to mind. The HHI, or Herfindahl-Hirschman index, is a widely used measure of concentration; is defined as the sum of the percentage weights of the components of the square index (usually taken as integers). For example, the HHI for an equally weighted portfolio of 50 shares is 200 (50 x 22); the HHI of the S&P 500 Weight Index, which comprises 500 shares, is 20 (500 x 0.22).
Other things like that, a higher HHI indicates an increased level of concentration, but, as the simple illustration above shows, even for equally totally unconcentrated weight portfolios, the HHI level is inversely related to the number of names. Therefore, as we use HHI to make comparisons within the technology sector over time, we need to use a tight metric. The adjusted HHI is the sector HHI divided by the HHI of an equal weighted portfolio with the same number of shares. A tighter HHI means that a sector is increasingly concentrated, regardless of the number of stocks it contains.
Exhibit 1 shows an adjusted HHI chart of the S&P 500 information technology, using monthly observations from January 1990. We observe an average value of 4.3, an interquartile range of 3.9 to 5 , 4, a minimum of 2.7 and eight large atypical values. Interestingly, the current adjusted HHI level of 7.2 is at the 95th percentile, indicating a historically high level of concentration for the technology sector.
Sample 2 illustrates the relationship between the adjusted HHI of the technology sector and the relative performance of the S&P 500 equal weight information technology compared to its weighted counterpart. After the peaks of concentration (as during 1990, 1999 and 2002), technology with equal weight seems to be outdated.
Another way to understand the relationship between concentration and relative performance is to plot the change in the adjusted Tech HHI compared to the relative performance of Tech with equal weight (see Figure 3). We observe a positive relationship between concentration and subsequent performance of the same weighting within Tech, with an R2 of 0.4.
The strong performance of large-cap technology names last year led to an increase in sector concentration levels. History tells us that after the peaks of concentration, technology with equal weight has tended to surpass. We can look at history to provide a perspective on these trends.
The posts on this blog are opinions, not advice. Read our legal notices.