Help comparing factor loads on VTMSX (Tax-Managed Small Cap)

Discuss all general (i.e. non-personal) investing questions and issues, investing news, and theory.
Post Reply
Topic Author
cb474
Posts: 905
Joined: Tue Jan 19, 2010 5:32 am

Help comparing factor loads on VTMSX (Tax-Managed Small Cap)

Post by cb474 »

The question came up in another thread about how Vanguard's Tax-Managed Small Cap fund (VTMSX) rarely gets discussed in the context of tilting portfolios.

I was interested in comparing VTMSX (based on the S&P 600) to Vanguard's Small Cap Value fund VBR/VISVX/VSIAX (now based on the CRSP) and to the popular iShares IJS fund (based on the S&P 600 Value). It seems like depending on the time period, VTMSX has almost as much value load as VBR.

Using the PortfolioVisualizer website, since inception in 1999, VTMSX has a HmL of .38, despite being a blend fund, which compares pretty favorably to the HmL for CRSP small value .48 (which I get from this post: http://www.bogleheads.org/forum/viewtop ... 0#p1511454). Given how much smaller the VTMSX (i.e. the S&P 600) is than CRSP small, perhaps this could be a good substitute for VBR, which also allows one the option of having a S&P 600 based mutual fund, rather than having to use ETFs.

On the other hand, I know factor loads vary over time. Looking at more recent time periods VTMSX does not have such a high HmL. It is also hard to compare funds that have different inceptions dates. And of course, VBR has changed indices more than once adding to the complications. And I even notice, for reasons I cannot explain, that over identical time periods different versions of the same fund VBR, VISVX, and VSIAX, don't have exactly the same factor loads (as rendered by the Fund Factors site).

So my question is what is the right way to compare factor loads of VTMSX to VBR and IJS? Thanks.
berntson
Posts: 1366
Joined: Mon Oct 29, 2012 12:10 pm

Re: Help comparing factor loads on VTMSX (Tax-Managed Small

Post by berntson »

Hi cb474! I'm not sure what to make of VTMSX. At first, I thought that maybe the tax management was creating a value loading. On further inspection, VTMSX has the same loading as other S&P 600 funds over the same time period. Other S&P 600 funds also show a high value loading around the turn of the century.

This is a guess. The value premium can show up in different places at different times. It need not show up uniformly across large value, medium value, and small value. When it shows up uniformly across all three value boxes, the small blend funds will have low value loadings as one would expect. But now suppose that the entire value premium is concentrated in the small value box for a period of time. That is, value stocks beat growth stocks entirely because small value is having a period of enormous outperformance. A small blend fund will then temporarily own more or less all the companies experiencing the value effect and those companies will be driving the returns of that fund. As such, the small blend fund will "look" like a value fund for a period of time---it will have high value loadings.

This sort of phenomenon is one reason, I think, to diversify one's value tilt across size. The value premium does not need to show up in the small value box. For a period of time, it could be large value or midcap value stocks that are the reason that value beats growth. Investors with only a small value fund will miss out on the value premium in those sorts of years.
Topic Author
cb474
Posts: 905
Joined: Tue Jan 19, 2010 5:32 am

Re: Help comparing factor loads on VTMSX (Tax-Managed Small

Post by cb474 »

Thanks for the thoughts berntson! That's very helpful.

Yes, I also noticed that the same anomalies to VTMSX's value load can be seen in other S&P 600 funds over the same timer period, like IJR. So I suppose my question is more generally about how to compare S&P 600 funds to MSCI and (now) CRSP small funds and what to expect going forward. Curiously, I've even notice that over certain time periods, the former MSCI Vanguard small funds had higher SmB than the S&P 600 funds (e.g. NAESX vs IJR), even though S&P 600 is generally seen here as more concentrated in small.

In any case, I take your point about not concentrating value loadings in small. I suppose that could be an advantage to the more midcap leaning CRSP version of small. It does seem hard to make judgements about funds when factor loads change over time.
berntson
Posts: 1366
Joined: Mon Oct 29, 2012 12:10 pm

Re: Help comparing factor loads on VTMSX (Tax-Managed Small

Post by berntson »

That's interesting about the shifting small loadings. Figuring out what to expect from funds is surprisingly hard I've found. Something I've started doing is paying more attention to actual valuations--price to book, price to cashflow, price to sales, and maybe dividend yields.* I look at the valuations first, then use rolling regressions to revise my initial estimate of how funds will perform. I think this sort of process is likely why many posters think that the S&P 600 is smaller than the CSRP funds. It owns smaller companies, so over time, one would expect it to have a higher sml loading.

One of the problems is that in many cases, the only valuations I can find are the ones on Morningstar. Those are forward looking and generated by some sort of opaque process that they don't explain. So they can't be completed trusted when doing this sort of thing.

*Hml tracks a value premium associated with low book to value stocks. But as things turn out, if you build a portfolio with low price to sales or price to sales or high dividend yield, you get similar results. Arnott even tried weighting stocks using price to number of employees. Turns out that works. All of these strategies seem to be strategies for doing more or less the same thing. Maybe it's buying companies with more risk. Maybe it's increasing the odds of buying companies that are underpriced.
Topic Author
cb474
Posts: 905
Joined: Tue Jan 19, 2010 5:32 am

Re: Help comparing factor loads on VTMSX (Tax-Managed Small

Post by cb474 »

Thanks for the further thoughts, berntson. It does seem to get more and more complicated, the more one learns about it.

People who are skeptical about the small and value premiums often say that it's one thing to contruct hypothetical possibilities, based on academic research, using indices devised for this purpose. It's another thing to construct a fund that actually captures these possibilities. I usually think that argument is overstated to just dismiss the idea of small and value tilting altogether.

But I am appreciating more that it's not so easy to capture these premiums, with the funds available (perhaps especially without access to DFA or the advistor only Bridgeway funds). And it seems like with something like VBR or the S&P 600 funds, one would have to hold them almost exclusively, to get a significant amount of tilt.

Anyway, I appreciate the help thinking through some of the things I noticed with the factor loads. I guess (like the Morningstar style charts) factor loads are not an end all and be all to evaluating one's portfolio. I have tried a little bit recently, like you, to take a look at other measures (something I also need to learn more about), but I agree the numbers can be hard to find and Morningstar's forward looking measures seem dubious. And of course, those are also numbers that change over time.
Post Reply