In this podcast, MarketFox columnist Daniel Grioli speaks with Research Affiliates founder and chairman Rob Arnott about his research into factors and addresses the criticism that AQR founder Cliff Asness had on his factor timing paper.
He makes the case you certainly can time factors and that to sell out of value now, while it is the only factor that is cheap, seems not the best thing to do.
He also believes that the industry will undergo significant change, but that machine learning is largely useless for long horizon investing.
The three papers discussed in the podcast are:
* How Can ‘Smart Beta’ Go Horribly Wrong?
* Timing ‘Smart Beta’ Strategies? Of Course! Buy Low, Sell High!
* Is Your Alpha Big Enough to Cover its Taxes?
3:00 Considering Astrophysics
4:30 Was a quantitative approach unusual in the early days?
5:45 A lot of quant go wherever the numbers lead them
8:15 Shockingly often conventional wisdom turns out not to be true when you test it
11:55 Higher dividend is higher earnings growth
17:15 Buybacks are smoke and mirrors to disguise management remuneration
18:05 What risk premium is normal?
24:15 You will never buy a bargain if you never buy what is out of favour, what is unloved.
24:45 Get people off their fixation with past returns
26:43 The spread between growth and value is wider than historic average. This means either a new normal or that value is a bargain.
31:00 The small cap effect is driven by the 2 or 3 per cent superstar winners.
31:45 Discussing the war of words with Cliff Asness
34:00 People are pouring money into multi-factor strategies, because they are tired of waiting for value to work.
36:00 Low volatility is trading at a premium, whereas historically it has traded at a discount. And people think they have less risk…?
37:45 Four of the five factors are pushing you into an anti-value direction
39:55 When momentum is chasing these bubble stocks you are slightly more likely to have a crash in momentum
42:00 If you must invest in the US, have a defensive stance
48:00 Those who say that factor timing doesn’t work, just have not done their homework
50:00 The big failing of the quant community is that we view everything as a signal, instead of viewing it as an asset
56:30 Talking ETFs
58:45 My next paper with Cam Harvey and Harry Markowitz looks at how you can screw yourself up with quantitative methods
59:55 Any research is data mining, but not all data mining is research
1:01:15 The quant community is engaged in performance chasing without realising it, for the most part
1:01:30 Start with a principal, start with a hypothesis, then test the hypothesis. Don’t go back to the same data again and again and tweak the process.
1:02:30 Machine learning is going to be useless for long horizon investing
1:08:10 The Future of the Financial Industry and discussing zero fees