In a recent article, I asked the question: Rather than build a globally diversified portfolio, why not just buy stock in Berkshire Hathaway (BRK.A )? An analysis of the performance of BRK.A relative to the DFA U.S. Large Cap Value Fund (DFLVX) and the firm’s U.S. Small Cap Value Fund (DFSVX), as well as to similar Vanguard index funds (VIVAX and VISVX), showed live results for value investors who instead implemented their allocations through well-diversified, passively managed funds.

• Using data from Morningstar, for the 15-year period ending September 23, 2016, BRK.A had annualized returns of 8.67%. DFLVX returned 9.17% and DFSVX returned 11.37%. BRK.A underperformed both, and also underperformed an average of the two funds by 1.6 percentage points a year.

• Over the same time period, BRK.A outperformed the large-cap Vanguard Value Index Fund (VIVAX), which had annualized returns of 7.18%, but it underperformed Vanguard’s Small Cap Value Index Fund (VISVX), which returned 10.5%. It also underperformed the average return of the two funds by 0.39 percentage points a year.

After reading the article, one skeptical reader questioned why I chose a 15-year lookback. I chose it simply because it’s convenient, as it’s the longest period Morningstar shows on its site. The reader guessed, however, that it was because BRK.A was shown to outperform over longer periods, and I selected the timeframe in which my “desired outcome” was displayed.

To address that question, I decided to go back and review the results over longer periods (to include the inception dates of the comparable funds). Thanks to my colleague, Dan Campbell, we can see the data over various longer periods. Using data from Bloomberg and Lipper, he was able to update the figures to show annualized returns through September 30, 2016. Returns are net of fund expenses. (Full disclosure, my firm, Buckingham, recommends DFA funds in constructing client portfolios.)

The data shows that, at least compared to the average of the two DFA funds, BRK.A underperformed whether we look at 15, 16, 17 or even 18 years. It outperformed slightly over the 19-year period, but underperformed slightly over the 20-year and 21-year periods. We have to go back to the 22-year and 23-year periods to see some real outperformance. If we consider only the DFA large value fund, DFLVX outperformed as far back as 18 years.

When examining the data for the two Vanguard funds, it’s important to understand that the biggest difference between their index funds and the DFA funds is that by design the DFA funds have significantly more exposure to the size and value factors, which have historically provided premiums.

I presented this data so that you can draw your own conclusions. Here are some, however, you should at least consider. And they are similar to the ones I drew in my previous article. The first is that perhaps Warren Buffett has been burdened by the huge size of assets under management. Or, as my co-author, Andrew Berkin, and I demonstrate in our book, The Incredible Shrinking Alpha, perhaps markets have become more efficient over time, making it harder for Buffett to identify undervalued companies. Or, it might also be that DFA has more recently incorporated research showing that Buffett not only bought value stocks, but higher quality (or more profitable) stocks as well. Maybe it’s a combination of the three.

At any rate, whether we are looking at 15, 16, 17, 18, 19, 20 or 21 years of data, it seems hard to conclude that investors in BRK.A were well rewarded for taking the risk of concentrating their assets in just one stock rather than in a broad portfolio of stocks. How many investors in an actively managed fund would be willing to accept such long periods of underperformance and still cling to the belief that they are likely to be rewarded for taking concentration risk going forward?