Question on Bernstein's "Skating"

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
Post Reply
Topic Author
Taggerung
Posts: 52
Joined: Sat Oct 22, 2011 1:59 pm

Question on Bernstein's "Skating"

Post by Taggerung »

For anyone who as read, or has a handy copy of, "Skating Where the Puck Was" I have a question about the point being made on pp 22-23.

While examining asset correlations over the short-term vs. long-term, Bernstein cites Large-cap Growth and Small-Cap Value as being, "the two least-related available asset classes." That quote I believe is referencing a M* type style box, that is, these two asset classes are in opposite corners so to speak.

On page 23 he then goes on to demonstrate that the correlation between these two classes does indeed decrease over time, such that over 5 year periods between 1931-2010 the correlation is only 0.43.

Is it a reasonable interpretation then that we should consider including what may be a lesser yielding asset class like Large-Cap Growth over Large-Cap Value if it is to be held in conjunction with Small-Cap Value for a very long time horizon? Since a main point of the booklet is that finding any asset classes with low correlation is extremely difficult yet extremely valuable, is this a reasonable take away?
User avatar
Kevin M
Posts: 15789
Joined: Mon Jun 29, 2009 3:24 pm
Contact:

Re: Question on Bernstein's "Skating"

Post by Kevin M »

It seems like a rational conclusion, but in practice most folks seem to favor holding large-cap blend or total market, and then tilting to small and/or value with additional mutual funds. One argument I've heard is that large-cap blend are actually tilted enough to growth already. Another might be that a rational starting point is "the market", and then one tilts away from the market based on individual preference.

Most of W. Bernstein's sample portfolios shown in his books follow this approach, using either S&P 500 or Total Market for the large-cap core position.

Kevin
If I make a calculation error, #Cruncher probably will let me know.
User avatar
nisiprius
Advisory Board
Posts: 52216
Joined: Thu Jul 26, 2007 9:33 am
Location: The terrestrial, globular, planetary hunk of matter, flattened at the poles, is my abode.--O. Henry

Re: Question on Bernstein's "Skating"

Post by nisiprius »

That's a weird table. "The" correlation, monthly 1927-2010, is 0.77. The next set of figures is weirdly overlapping. 1927-2006, 5-year periods is 0.65, which isn't that different from 0.77. The correlations get lower and lower as the times get slid along, 1927-2007, 1929-2008, 1930-2009, and then increase slightly 1931-2010.

Notice that what this means is that he is assuming rebalancing at 5-year intervals on the schedules shown.

What I take from this is not that correlations are lower if we use 5-year periods, but that correlations are lower for periods that don't include 1929.

What I also take from this is that in the real world, the chances of how your personal investment moves happen to line up with financial crises are more important than what "the" long-term statistics are.
Annual income twenty pounds, annual expenditure nineteen nineteen and six, result happiness; Annual income twenty pounds, annual expenditure twenty pounds ought and six, result misery.
Costanza
Posts: 52
Joined: Sun Dec 08, 2013 7:08 pm

Re: Question on Bernstein's "Skating"

Post by Costanza »

Taggerung wrote:For anyone who as read, or has a handy copy of, "Skating Where the Puck Was" I have a question about the point being made on pp 22-23.

While examining asset correlations over the short-term vs. long-term, Bernstein cites Large-cap Growth and Small-Cap Value as being, "the two least-related available asset classes." That quote I believe is referencing a M* type style box, that is, these two asset classes are in opposite corners so to speak.

On page 23 he then goes on to demonstrate that the correlation between these two classes does indeed decrease over time, such that over 5 year periods between 1931-2010 the correlation is only 0.43.

Is it a reasonable interpretation then that we should consider including what may be a lesser yielding asset class like Large-Cap Growth over Large-Cap Value if it is to be held in conjunction with Small-Cap Value for a very long time horizon? Since a main point of the booklet is that finding any asset classes with low correlation is extremely difficult yet extremely valuable, is this a reasonable take away?
Clever inference. But any portfolio that holds a substantial amount of a total market index will have a lot of large growth stocks. On face, then, it's hardly clear that adding additional large growth exposure will provide a net benefit, much less that whatever diversification benefit there is will outweigh the value premium obtained by tilting toward large-cap value, not growth. But if the asset correlations truly are as modest as 0.43, someone more statistically savvy than I should run the numbers.
terrabiped
Posts: 160
Joined: Mon Dec 30, 2013 1:29 pm

Re: Question on Bernstein's "Skating"

Post by terrabiped »

Hmmm, I had a very different takeaway from that book that can be expressed in the following quotes:
In short, you have to move on when you get too much company; the first people to invest in an asset class with high expected returns and low correlations enjoy sirloin, while the Johnny-come-latelys get hamburger.

Kindle Locations 161-163


Remember, investing early in true alternative asset classes is, by definition, never easy, and once they do become securitized, liquid, and popular, all bets are off. In the digital age, any risky asset class you can buy with a keystroke can, and most likely, will bite you when things head south.

Kindle Locations 301-303
This point cannot be made strongly enough: when a risky asset class becomes too popular, the fact that it is over-owned by “weak hands” means that it will simultaneously have both low expected returns and high correlations; in other words, low expected returns and high correlations go together, a fact demonstrated by most alternative asset classes over the past decade.

Kindle Locations 362-365
How easy was it to buy a diversified basket of small cap value stocks 50 years ago? How easy is it today? How might that change their characteristics? I suspect it changes everything.
lazyday
Posts: 3849
Joined: Wed Mar 14, 2007 10:27 pm

Re: Question on Bernstein's "Skating"

Post by lazyday »

Growth hasn't had its own return premium, while Value has.

So has Momentum. I don't know if Growth and Momentum are highly correlated, but Value and Momentum have had low correlation. Been suggested to use both, google:
growth and momentum correlation
or try:
value and momentum

Robert T posted about Mom and small value, I think in a discussion on the "Larry Portfolio" also called "Swedroe Portfolio". Idea is using extreme small value, no large or market funds, and lots of bonds.
Topic Author
Taggerung
Posts: 52
Joined: Sat Oct 22, 2011 1:59 pm

Re: Question on Bernstein's "Skating"

Post by Taggerung »

terrabiped wrote:Hmmm, I had a very different takeaway from that book that can be expressed in the following quotes:
In short, you have to move on when you get too much company; the first people to invest in an asset class with high expected returns and low correlations enjoy sirloin, while the Johnny-come-latelys get hamburger.

Kindle Locations 161-163


Remember, investing early in true alternative asset classes is, by definition, never easy, and once they do become securitized, liquid, and popular, all bets are off. In the digital age, any risky asset class you can buy with a keystroke can, and most likely, will bite you when things head south.

Kindle Locations 301-303
This point cannot be made strongly enough: when a risky asset class becomes too popular, the fact that it is over-owned by “weak hands” means that it will simultaneously have both low expected returns and high correlations; in other words, low expected returns and high correlations go together, a fact demonstrated by most alternative asset classes over the past decade.

Kindle Locations 362-365
How easy was it to buy a diversified basket of small cap value stocks 50 years ago? How easy is it today? How might that change their characteristics? I suspect it changes everything.

All great quotes and I certainly wouldn't contest or disagree with any of your takeaways, I just thought that my point was more in keeping with these than maybe it truly is. Maybe it has to do with my mushy definition of "alternative assets." I'm not sure that any style boxes are risky in the way that he is using that term above (I think he had commodities or international REITs in mind?). You are certainly right though about the ease with which we can all buy into previously obscure sectors, styles, regions, or various other baskets of securities.

I love it when this forum acts like the guardrails on my highway to retirement!
Post Reply