Highlights from AQR’s interview with Ed Thorp

Below is a fascinating interview with Ed Thorp, who discovered/invented many of quantitative methods used to beat casinos and markets. Full interview can be found here (PDF).

A few of my favorite quotes (and, yes, my confirmation bias is in full effect, since my firm focuses on passive index investing for equities and active investing for fixed-income and private equity/debt/real estate):

On the challenges of finding true alpha in liquid markets:

For liquid asset classes like US bonds and stocks, for instance, this means that everybody who is active, or not indexing, are
collectively a big index fund, on average. That big actively-traded “index fund” is being managed, so it’s also paying costs. So, a couple of percent is being drained out of that pool, compared with the guys who are paying very low amounts for passive indexing. So, these active investors collectively have a couple percent disadvantage. So, all the institutions that are battling for an edge in those liquid asset classes aren’t going to get alpha collectively. They should just index those parts of the portfolio, in my opinion.

On the opportunities in illiquid and opaque markets:

Sullivan: Where do you see additional opportunities for institutional investors?
Thorp: What I see for institutional investors is access to the more illiquid asset classes like private equity. That’s something ordinary investors don’t get a shot at, and it requires active management because there’s a lot of work in evaluating and
hiring managers.

On the importance of approach, strategy, and risk-management, even when the odds are with you.

Brown: A recent paper by Haghani and Dewey (2016) indicated that students in finance often lack the basic quantitative skills to properly think about risk.
Thorp: Yes. The authors conducted a live experiment with college-aged students and young professionals at asset management firms who were knowledgeable about investing. The experiment went like this: each participant gets 30 minutes and $25 to start with. Each has a computer terminal and is informed that they will flip a computerized coin that comes up heads 60% of the time and tails 40% of the time. Then they can bet as much as they want on each coin flip. After 30 minutes, time stops and they get to keep their gains up to a certain dollar limit (otherwise the experimenters might go broke!). If they reach the limit sooner, then betting stops, because they’ve won as much as they can. The others go on betting. So, the question is, what betting policy should you follow? Many of the participants had no idea what to do. Quite a few of them went broke and a rather large portion of them didn’t make any money. Another rather large section made some money but not a lot. The average amount of winnings was around $70 for those not going broke. Aaron wrote a nice piece (2016) which analyzed all this in detail. Winnings should be something like $240 if they follow optimal policy.
Brown: Yes. Very high probability that they’d win about $240 if they used the Kelly method, which as you already know says to bet 20% of their bankroll each flip on heads, calculated as 2(.6)-1=20%. Basically, there seemed to be two types of bettors,
risk-takers who went broke and non-risk takers who bet small amounts like $1 each time, so average winnings of even those not going bankrupt were quite low.
Thorp: Every year in Las Vegas they have something called the Blackjack Ball, where about 50 of the best gamblers in the world gather. If you were to ask any of these professional blackjack players what to do, they would have said, well, I’ll just use the Kelly Criterion because it’s a close approximation to an optimal solution. So, the professional blackjack players would know the answer, but the finance people did not.

Full interview can be found here: https://www.aqr.com/-/media/files/papers/aqr-words-from-the-wise-ed-thorp.pdf


Housel: What Other Industries Teach Us About Investing

I was amped for the release of Blue Planet II today, but I think this was the best video release of the day.

My favorite thing about this talk is that Housel covers some of the most important investing issues without even talking about investing! Nobody is immune from the challenges and biases that Housel illustrates, so its a great watch for both professional and non-professional investors.


Two More Ways to View the Curve

Following up on my previous post about the 2s10s spread, I should note that the 2s10s is just one of many spreads. Its a benchmark. But you could just as easily look at the 2s5s, 5s30s, 10s30s, and so on. Below are a couple of alternative ways to keep tabs on the shape of the yield curve.

I like the below chart because it allows one to see how various constant maturities move in relation to one another over time. As you can see, the curve flattened during the last two major tightening cycles and so I remain skeptical of those that are calling for long-rate rises to outpace those of short-rate rises.

Another common way to view the shape and history of the yield curve is to simply view it next to previous iterations of itself. See below for an example:

Happy Friday!

2s10s Spread

[Update: This article has been updated with live data at my new site here: https://mattshibata.com/2s10s]

A challenge (not unique) to investing is separating the signal from the noise. I find that the highly-stochastic nature of most data makes them practically useless (and thus noise). One of the few data points that I find useful to keep tabs on is the 2s10s spread.

The above chart shows the 2s10s spread, which is the difference between the 2-year Treasury yield and the 10-year Treasury yield.

Above is a graph of the underlying 2-year and 10-year yields. The 2s10s spread (the first chart) is simply the red line minus the blue line.

The reason that I and many other investors (and economists too) reference the 2s10s spread is that it is a quick and simple indication of the slope of the yield curve, which is used to measure and estimate all sorts of things. Generally, the economy and markets tend to do well when the yield curve is steep and not so great when it is flattish or inverted (meaning the short-end of the yield curve is higher than the long-end).

The first chart shows the yield curve is rapidly flattening, which is of particular interest these days since the yield curve typically flattens and then inverts just before recessions (as indicated by the gray bars). This usually occurs because the 2-year yield rises much faster than the 10-year yield (as is happening now) and eventually surpasses it. The yield curve is not inverted yet and rapid flattening often coincides with tremendous economic and market performance; it does not appear that the curve will invert for at least 6-12 months (if it does at all), so no need to panic yet.

It is worth mentioning that some of the smartest asset managers out there think that the yield curve conveys less information now than in the past, due to a variety of reasons that I won’t get into here. Those managers may very well be right, but the 2s10s spread has a much better recession-calling record than anyone I know of. Further, past episodes of yield curve flattenings and inversions were “explained” as benign signs by the top minds during those respective times. Not sayin’, just sayin’.

The 2s10s spread is just one of many data points and I’m neither supporting nor denying its significance. It is something that many people watch and I believe it bears watching as well.

***A six-month update can be found here: http://thoughtfulfinance.com/2018/07/18/yield-curve-inversions/