I really like the Horizon Kinetics letters. This can be interpreted in the way that Thorp qualified his ‘favorite books’ list on the Masters in Business podcast I summarized by saying

“I don’t necessarily agree with the books, I just believe that great books should provoke thoughts”. – Ed Thorp

On the NASDAQ 100 ETF

The top 5 holdings of the NASDAQ 100 ETF ‘QQQ’ account for 41% of the total AUM (FAANG + Microsoft). If this fund were active, this concentration would be forbidden in European UCITS. Yet, it is available in the passive vehicle iShares NASDAQ 100 UCITS ETF.

QQQ P/E calculation

The P/E ratio advertised for the QQQ ETF is 22.2x, by taking many assumptions:

  • trailing earnings
  • excluding loss-making companies
  • for the positive earning companies, excessive P/E companies are weighted less by using the ‘weighted harmonic mean’

TC comment

Electrical engineers are familiar to the harmonic mean, as the equivalent single resistance of two parallel electrical resistances is calculated by taking the harmonic mean of those two parallel resistances. What happens is that the harmonic mean for the equivalent resistance is closer to the smallest value (in case of QQQ, the smallest P/E companies) as electrical flow takes the path of least resistance. Another way to put it is that the smallest resistances (P/E companies) will be weighted the most in the average.

It is evident that real-world portfolios cannot return the harmonic mean of position returns.

Active manager YTD S&P500 tracking error from not owning FANG stocks

Bregman writes that by merely excluding Facebook, Apple, Amazon, Netflix, Google, Microsoft from the index, one would trail the YTD S&P500 performance by 243 basis points, or more than a quarter of the S&P500 returns!

In fact, the situation is even more dire when looked at over the last years:

  • 2015: owning 5% of the S&P500 accounted for 50% of the index return
  • 2016: the 10 best performing stocks, including FANG, accounted for more than 100% of the index return
  • 2017YTD: FANG accounted for 25% of the index return

TC comments

The key takeaway is that I think that the monstrous performance of FANG creates an unbearable fear of not owning FANG for many active managers as well. The huge upside volatility of the last years creates career risk of not owning FANG. Moreover, because of the great performance the weight increases every year so the buying pressure mounts, thus sucking in more capital each period.

It is interesting from a historical perspective to see how the fear of missing out reappears in different forms.

Google and Facebook versus AOL

The fund extrapolates 25% annualized ad spend growth on Facebook and Google, while assuming a 4% annualized growth of worldwide ad spending. Using these numbers, by 2020 the two internet giants would account for 40% of worldwide ad spending (offline + online). Today this number is at 23%.

It is funny how Horizon Kinetics voices exactly the same concern as I have written about just weeks ago on this blog:

Of course, as Google’s and Facebook’s share of worldwide advertising expenses increases, they must eventually reflect the cyclical attributes of the industry that clearly everyone expects they will dominate.

The appendix contains an Horizon piece from 1999, reasoning from first principles, how AOL was overvalued. This reasoning proved correct. For an introduction to reasoning from first principles, see the book Elon Musk and Superforecasting (and The Fermi Technique in my Superforecasting summary).

Til next time,

TC