Over the weekend, I was looking at some research from JP Morgan that showed the percentage weight of the top ten stocks in the S&P 500 over time, and I admit that I was surprised.

When we started Acropolis, 20-years ago in August, the top ten stocks made up about 24 percent of the index. I thought that was pretty high back then and was one of the reasons that I thought that our global diversification made sense.

That percentage dropped fairly steadily until about 2016 when it hit about 16 percent. Then, for reasons that aren’t exactly clear to me, the weight of the top ten stocks rose steadily. According to JP Morgan, the top ten stocks in the S&P 500 as of Jan 31, was 30.1 percent – almost double what it was in 2016 and 25 percent more than when we started in 2002.

All of the stocks are household names, such as Amazon, Google, and Tesla. Speaking of Google (or, more technically, Alphabet, the parent company), the weighting scheme counts each of Alphabet’s share classes as its own stock, so really, we’re looking at the top nine stocks in the index.

The JP Morgan research then goes on to illustrate how much earnings the top ten contributes. In 2002, it was about 24 percent, matching the market weight. Outside of the 2008 global financial crisis when everything was out of whack, the top 10 companies contributed about as much as their market weights would imply.

Until recently, that is. As of January 31st, the top 10 companies in the index contributed just 24 percent of the earnings but were still 30 percent of the value – suggesting that the value is off compared to the earnings.

In fact, the price-earnings (PE) ratio of the top 10 stocks as of Jan 31st according to JP Morgan was 30.6. Since the inception of their data in 1996, the average PE for the top ten stocks was 19.8. The other 490 stocks have a PE ratio today of 18.5. The average PE for the remaining 490 since 1996 was 15.7.

A few things jump out. First, the top ten stocks are always more expensive. On average, they have a PE of 19.8 versus 15.7, a premium of about 25 percent. That makes sense to me because the big stocks are big because everyone loves them and is willing to pay high prices for them.

Second, the premium is about 65, almost three times more than average. I don’t have access to the underlying data, so I can’t do a detailed analysis, but I don’t think you need the data to believe that the premium is probably unsustainable.

Third, the other 490 stocks are not that expensive compared to the average, which is somewhat reassuring. Yes, they are more than 20 percent more expensive than the average, but some of that should be explained by much lower than average interest rates.

As noted above, I was a believer 20-years ago that the S&P 500 was too concentrated, and it’s much more concentrated today than it was then, so I feel even better about our identification strategy.

Sure, over the last five years, I haven’t been so happy with it since the S&P 500 beat everything else and the diversification represented a drag. But seeing how much of the results came from just nine or ten stocks helps me feel better about missing some of those returns and better still about where we stand going forward.

To be sure, we still have exposure to these stocks – four of them are on our Approved List and we own them through our index and quantitative funds.

But, we have less exposure. For one thing, even for our clients that are 100 percent stocks (the distinct minority), only about 50 percent are in large-cap S&P 500 stocks. To a less extent, many of the top ten stocks are growth-oriented tech stocks, where we are generally underweight.

As always, we don’t know where stocks will go from here – especially to the top ten or the remaining 490 stocks in the S&P 500. But, I’m glad to have a less concentrated approach than the broad index.