Historic Small Cap Returns May be Inflated

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
Post Reply
happytrades
Posts: 84
Joined: Tue Mar 02, 2010 6:44 pm

Historic Small Cap Returns May be Inflated

Post by happytrades » Sun Nov 17, 2013 7:04 am

Jason Zweig with WSJ made an observation at the AAII convention yesterday that I had never considered. In the Q and A, he was asked about the historic small cap premium. He made a cautionary note on the posted returns. Historical indices do not reflex the actual bid/ask spread that real investors would have paid, particularly in the old days. He said that could have ranged 2%- 4% on a trade. With much trading at all, it would wipe out the small cap advantage.

larryswedroe
Posts: 15597
Joined: Thu Feb 22, 2007 8:28 am
Location: St Louis MO

Re: Historic Small Cap Returns May be Inflated

Post by larryswedroe » Sun Nov 17, 2013 9:11 am

Well that depends. First if you have a premium say of 3% and you have spreads of 3% your turnover would have to be 100% to wipe out that premium.
Even small cap index funds which have more turnover than large caps don't have turnover anywhere near that.
Second, one doesn't have to have traded on the spreads. One could have been a "patient" trader and used market or limit orders to control costs and accept tracking error as the price.
I would add if you want to go this route you could have also improved on small cap performance by screening out the "lottery ticket" type of stocks.
Bottom line is that of course trading costs matter but Jason's comment IMO exaggerates the impact.
Hope that helps
Larry

User avatar
grayfox
Posts: 4555
Joined: Sat Sep 15, 2007 4:30 am
Location: Earth

Re: Historic Small Cap Returns May be Inflated

Post by grayfox » Sun Nov 17, 2013 9:30 am

I don't know about small-cap in particular, but historic returns, in general, probably don't reflect what an actual investor would have gotten.

There is often a significant different between the price that gets recorded in the historic record and what price someone will actually pay when buying or get when selling. For example, I recently looked at a TIPS price in WSJ and the bid/ask was 100.04/100.04

But when I went to an actual broker to buy that TIPS, the price was higher and the bid/ask spread was wider, 100.07/100.234. I actually ended up paying $100.2422

If you don't believe me, here is the page with the historic price http://online.wsj.com/mdc/public/page/2 ... stcalendar. See the 2014 Jan 15 TIPS. That price will be in the record forever and anyone backtesting will use those prices, even though an ordinary investor could not get them.

The lesson is take backtested results with a huge grain of salt.
Gott mit uns.

stlutz
Posts: 3963
Joined: Fri Jan 02, 2009 1:08 am

Re: Historic Small Cap Returns May be Inflated

Post by stlutz » Sun Nov 17, 2013 10:30 am

Zweig saying that an at AAII conference is interesting. Back when I was a member, I recall that the approach was generally to select individual smallcap (+usually value) names, based largely on the same research that gets bandied around here. In short, a nice bit of contrarianism on his part in that environment.

happytrades: You've posted a number of things in that past couple of days from the conference that seem Bogleheadish. Has that been a general conference theme or are people like Zweig sort of the odd men out?

stlutz
Posts: 3963
Joined: Fri Jan 02, 2009 1:08 am

Re: Historic Small Cap Returns May be Inflated

Post by stlutz » Sun Nov 17, 2013 10:33 am

Does anyone know: Historically, has there tended to be more or less share turnover of smallcap or largecap stocks? With such data as well as an approximation of spreads, one could arrive at an estimate of what actual investor returns in these asset sub-classes were back then.

User avatar
nedsaid
Posts: 8499
Joined: Fri Nov 23, 2012 12:33 pm

Re: Historic Small Cap Returns May be Inflated

Post by nedsaid » Sun Nov 17, 2013 11:39 am

One point that John Bogle makes in his famous "Telltale Chart" speech is that the studies regarding the outperformance of value and small cap stocks do not take into account how real life investors could have invested. If an investor wanted to capture the small cap or value premiums, he would have had to use active funds as indexes did not exist until the 1970's. Or the investor could have done old fashioned stock picking.

Another factor is that there used to be "stock funds" and "bond funds". In recent years, you have seen more and more specialization by size and valuation. It used to be if you wanted to own a diversified portfolio of stocks, you just bought a stock mutual fund.

Bogle shows a flaw in the methodology though I don't think it is a fatal flaw. My understanding is that academic research shows that the small and value premiums show up in almost all the stock markets of the world. The data is very persistent.

I do think that if DFA and the indexes been around in 1926, that the value and small premiums would still exist though probably somewhat less than what the research does today. That these premiums are persistent shows me that these are rooted in human behavior and psychology and are not easily arbitraged away.
A fool and his money are good for business.

steve_14
Posts: 1507
Joined: Wed Jun 20, 2012 12:05 am

Re: Historic Small Cap Returns May be Inflated

Post by steve_14 » Sun Nov 17, 2013 2:15 pm

Personally I'm mildly skeptical of return data more than about 40 years old for large caps. For small caps, I'm highly skeptical.

Random Walker
Posts: 1959
Joined: Fri Feb 23, 2007 8:21 pm

Re: Historic Small Cap Returns May be Inflated

Post by Random Walker » Sun Nov 17, 2013 2:58 pm

History is important. But we can only earn future returns. Looking forward, it intuitively makes sense that smaller companies are more risky. And they do have higher costs of capital. Thus they have higher expected returns. It's important to have an historical record to support decisions regarding investing in an asset class. But also very important to have plausible risk or behavioral story that adds to investor conviction.

Dave

User avatar
3CT_Paddler
Posts: 3111
Joined: Wed Feb 04, 2009 5:28 pm
Location: Marietta, GA

Re: Historic Small Cap Returns May be Inflated

Post by 3CT_Paddler » Sun Nov 17, 2013 3:07 pm

nedsaid wrote:Bogle shows a flaw in the methodology though I don't think it is a fatal flaw. My understanding is that academic research shows that the small and value premiums show up in almost all the stock markets of the world. The data is very persistent.


I read a study that looked at five or so of the bigger European economies (can't find it through Google at the moment). One of the countries showed no premium (UK) and the others had a mixed record.

Rodc
Posts: 13601
Joined: Tue Jun 26, 2007 9:46 am

Re: Historic Small Cap Returns May be Inflated

Post by Rodc » Sun Nov 17, 2013 3:15 pm

Historical returns also don't back out taxes which have changed over the years.

Or prevailing bond rates at the time.

Or the cost of trading which includes the rate of turn over not just the cost of trading. My sense and someone posted on this once (I think quoting Jack Bogle) is that trading rate was much lower when the per trade cost was much higher, so this might be a wash.

Lots of trouble with past data.
We live a world with knowledge of the future markets has less than one significant figure. And people will still and always demand answers to three significant digits.

afan
Posts: 2905
Joined: Sun Jul 25, 2010 4:01 pm

Re: Historic Small Cap Returns May be Inflated

Post by afan » Sun Nov 17, 2013 5:20 pm

That transaction cost caveat has been studied rigorously. As you might expect, it reduces the premiums. As best I recall, at least the value premium survived as positive after making reasonable adjustments for trading costs. I have not seen anyone attempt to aacount for taxes on a top bracket taxpayer each year. As indicated, that would not only require digging up the tax laws for each year, but also making a detailed set of assumptions about other income, deductions, and whatever may have counted at any time in the past. Certainly the tax effect would reduce the returns to the strategy. But you need not guess about this. There have been small, value, and small value funds for years. You can look up their factor loadings, the predicted performance of a portfolio with that loading, and calculate the difference. That difference will include all the costs as well as the impact of tracking errors & positive or negative and securities blending. It could be hard to determine exactly how much of the difference was due to each factor, but you would have an estimate of the returns to the strategy.

If you really wanted to chase it, you could create a portfolio and deduct each year what you taxes would have been.

Or you could go TSM.
We don't know how to beat the market on a risk-adjusted basis, and we don't know anyone that does know either | --Swedroe | We assume that markets are efficient, that prices are right | --Fama

larryswedroe
Posts: 15597
Joined: Thu Feb 22, 2007 8:28 am
Location: St Louis MO

Re: Historic Small Cap Returns May be Inflated

Post by larryswedroe » Sun Nov 17, 2013 6:47 pm

there is definitely far more turnover among small than large

What most don't understand is that the outperformance of an asset class is mostly related to the migration of a small number of stocks OUT of the asset class due to superior performance. For example, the entire small cap premium on average is explained by 1% of the stocks each month, on average (another reason need to basically own them all)


Larry

DetroitRed
Posts: 175
Joined: Tue May 17, 2011 4:16 pm

Re: Historic Small Cap Returns May be Inflated

Post by DetroitRed » Sun Nov 17, 2013 7:39 pm

The main database for historical US stock returns is CRSP which has NYSE stocks back to 12/31/25.

The American Stock Exchange (AMEX) was not covered by CRSP until 07/31/1962. It used to be difficult to get an NYSE listing, so many small caps were listed on AMEX. Therefore the historical small cap returns between 1925 and 1962 are calced from an incomplete universe of US stocks.

User avatar
stevewolfe
Posts: 1523
Joined: Fri Oct 10, 2008 7:07 pm

Re: Historic Small Cap Returns May be Inflated

Post by stevewolfe » Sun Nov 17, 2013 8:19 pm

DetroitRed wrote:The main database for historical US stock returns is CRSP which has NYSE stocks back to 12/31/25.

The American Stock Exchange (AMEX) was not covered by CRSP until 07/31/1962. It used to be difficult to get an NYSE listing, so many small caps were listed on AMEX. Therefore the historical small cap returns between 1925 and 1962 are calced from an incomplete universe of US stocks.


Now that ^^^^ is very interesting bit of information that in the last 5 years I've been here I do not recall ever hearing before. Thanks for sharing that!

DetroitRed
Posts: 175
Joined: Tue May 17, 2011 4:16 pm

Re: Historic Small Cap Returns May be Inflated

Post by DetroitRed » Mon Nov 18, 2013 11:32 am

stevewolfe wrote:
DetroitRed wrote:The main database for historical US stock returns is CRSP which has NYSE stocks back to 12/31/25.

The American Stock Exchange (AMEX) was not covered by CRSP until 07/31/1962. It used to be difficult to get an NYSE listing, so many small caps were listed on AMEX. Therefore the historical small cap returns between 1925 and 1962 are calced from an incomplete universe of US stocks.


Now that ^^^^ is very interesting bit of information that in the last 5 years I've been here I do not recall ever hearing before. Thanks for sharing that!


I wish that fact would be more widely reported.

happytrades
Posts: 84
Joined: Tue Mar 02, 2010 6:44 pm

Re: Historic Small Cap Returns May be Inflated

Post by happytrades » Mon Nov 18, 2013 4:23 pm

stlutz wrote:Zweig saying that an at AAII conference is interesting. Back when I was a member, I recall that the approach was generally to select individual smallcap (+usually value) names, based largely on the same research that gets bandied around here. In short, a nice bit of contrarianism on his part in that environment.

happytrades: You've posted a number of things in that past couple of days from the conference that seem Bogleheadish. Has that been a general conference theme or are people like Zweig sort of the odd men out?


This was my first time to attend the conference. Being a Boglehead vulnerable to cognitive bias, I gravitate to speakers who confirm my existing biases toward passive investing. I would say most of the attendees were stock pickers and the presentations were geared toward them. They generally seem to believe that there is a large universe of small stocks that Wall Street does not follow, and that you can find some good buys there.

However, there were several presentations on how truly difficult it is for people to overcome their emotions in being successful investors.

Post Reply