Sunday, October 31, 2010

Low Volatility Conference


Last Friday I spoke at a Robeco seminar on ‘low volatility’ (note the calm low-vol investor within a panicking buffalo herd). Why did I speak? Well, I have corresponded with Pim van Vliet for a while, as there aren't a lot of us 'low vol' researchers out there, so we tend to appreciate each other's work. He and Arlette van Ditshuizen run some of their low volatility portfolios.

The strategy is based on the well-documented fact that there is no obvious risk premium within equities. This implies that you can generate a dominant Sharpe ratio by simply focusing on only those stocks with low volatility. Reduce volatility by 33% relative to the indices (very doable), and you increase your Sharpe 50%!

Now, a slightly less obvious benefit of the low volatility focus is that the return is higher too, which was most prominently demonstrated in a series of papers by Ang et al in the Journal of Finance. You can download relevant data from my betaarbitrage.com website here, where I have beta portfolios since 1962, and minimum variance portfolios since 1998. Beta, total volatility, idiosyncratic volatility, it doesn’t matter, the highest of that stuff will have a lower return than average. Thus, removing them raises the portfolio return.

My talk focused on the idea that risk and return are not correlated in practice. This was a theme of my book Finding Alpha, and so I went over my empirical proof, which is to present the large scope of data that contradict a positive risk-return relationship, about 20 different asset classes. If risk and return are empirically uncorrelated, low volatility investing is a rather straightforward normative implication of what a rational investor should do, so clearly my invite served to help their argument and I was only too happy to shill for the truth.

The ultimate test of an economic theory is its out of sample performance, and this is perhaps mostit is most successful. I documented it in my 1994 dissertation, using data from 1926-1992 (you merely had to control for price or size, and it shows up). I then set up a C-corp that invested successfully on this principle from 1996-2001. I then worked at a hedge fund, and while I can’t say what or how I did, my old employer Telluride legally tried to prevent me from using ‘volatility’ ever again, because they thought it was valuable (and, creatively, their 'property’). So, I’ll just let my rational reader infer what I did and how well it worked for that period. Now, Robeco has shown it has worked from 2007-present, as the Robeco European Conservative Equity fund recently garnered the coveted ‘5 star’ Morningstar rating. That collection of real-time success is more compelling than any GMM test, because it is too easy to fit a specific historical sample using specific conception of volatility or really anything.

Vanguard became very powerful because they were the first big institution to really believe in the value of index funds relative to active managers. A key was, John Bogle , the CEO, really believed it, having written a senior thesis at Princeton on the subject decades earlier. Robeco has several researchers who all believe in the dominance of low volatility investing, and being right on a basic principle like this makes it a lot easier to be tactically proficient. I hope they can keep the imitators at bay, because I've sent innumerable letters and made presentations to funds that have all been veriy dismissive of this volatility finding. Thus, I would prefer they pay for their late-coming to he facts, because their crowd has been singularly unhelpful to those of us trying to sell the world, in academia or in the private sector, on this idea.

Given your average manager lags the indices by 1-2%, and has higher volatility than the indices, index investing does not offer wildly better performance than active investing, but it is clearly better. The case for low volatility portfolios is even more compelling in a Sharpe ratio sense, because there you increase returns by 1-4%, and lower volatility by 30%. One fund manager at the seminar discussed how they were now allocating 30% of their portfolio to low volatility investing. I suspect more and more institutions will find this a good idea, and it’s going to be a big trend.

7 comments:

Anonymous said...

Is there a link to the slides?

Anonymous said...

when everyone is on board, what does the new equilibrium look like?

Anonymous said...

Academics are insane. Why publicize something that works? When this notion picks up enough investors it will force the cycles to change and the strategy itself will become less reliable.

Unknown said...

Not to be nit picky but Ned Johnson was the Fidelity CEO, John Bogle is the Vanguard guy

Eric Falkenstein said...

thanks Eric!

Mebane Faber said...

Nice article Eric. Seems like some of the big boys are rushing into the space...Russell/Axioma launching some ETFs including one on low beta (LBTA). Hard to find much info or a fact sheet on the indexes...:

http://www.indexuniverse.com/sections/news/8171-russell-adds-5-large-cap-etfs-to-pipeline.html

Anonymous said...

"Why publicize something that works?"

Why not publicize? If you keep things secret it only ensures that nobody gives you money to manage, or if they do, they will drop you when the inevitable patch of bad luck hits, because they can only assume that your bad performance is due to incompetence.