Skip to main content Blog Help (new window)
Home
Blog > Posts > Do Sectors Matter?
Back to Home
Do Sectors Matter?
As the number of exchange-traded funds (ETFs) has multiplied, the number of sector ETFs has exploded. There are now ETFs that cover every sector of the U.S. stock market imaginable (health care, telecommunications, consumer discretionary, etc.) and even international and emerging markets sector ETFs. Lost in all this is the general finding that sectors really aren’t that important, particularly over longer periods of time.
 
The long-term returns of most sectors are basically determined by:
  • The size of the companies in the sector
  • Whether the sector is on average dominated by value or growth companies
  • Its exposure to momentum

In other words, you should generally avoid concentrating your investment portfolio in a particular sector, because you’re giving up diversification and likely getting nothing in return that you couldn’t get with a better diversified portfolio of stocks.

Using 30 years of monthly returns data ending December 2012, I ran an analysis of the 12 sector portfolios on Ken French’s site to re-examine whether sectors matter. If sectors truly matter, you should see a large number of sectors that have had returns well above or well below what you would expect given the size, value and momentum characteristics of each sector. In industry lingo, these excess returns are called “alpha.” Below are the two relevant statistics from that analysis:

Of the two statistics, the second one is the most important. If that particular statistic is between –2 and +2, the sector returns were about what you would expect given the size, value and momentum characteristics of the sector. We see that 10 of the 12 sectors are within this range, basically indicating there’s nothing unique about these sectors (including utilities for all the utilities crazed investors out there). Two are outside of this range but one is above and one is below so I wouldn’t put too much emphasis on that portion of the results. The basic conclusion (reconfirming earlier findings) is that sectors don’t matter.

Are There Any Sectors That Do Matter?

The one possible exception is real estate investment trusts (REITs). REITs are stocks that provide exposure to the commercial real estate market and have historically behaved quite differently from the overall stock market. However, even REITs have been much more highly correlated with the stock market beginning at about the time of the financial crisis. There’s no way to know though whether this is just temporarily higher correlation or a permanent change.

Random Links and Commentary of the Week

Entertaining read from Bill Simmons analyzing the significance of the Miami Heat streak. 

Jared Kizer is the director of investment strategy for The BAM ALLIANCE. See our disclosures page for more information.

Comments

Which return data set are you using?

I'm assuming you're using the the value weighted returns (as opposed to the equal weighted) because our results are similar.  I ask, however, because I came up with alphas and t-stats that were of the same sign, but of a different magnitude.  For example, using 360 months of data (from Jan 1983 to Dec 2012) for Non-Durables, I calculated an annual alpha of 4.16% with a T-Stat of 2.53.
at 4/8/2013 6:16 PM

Differences

Not sure what explains the differences. Just rechecked my calcs and I did use the VW series and the same period of time you did with market, size, value and momentum as the variables. I did use Ken's research factors though. If you used his benchmark factors that could lead to differences.

JK
at 4/10/2013 3:20 PM

Add Comment

Items on this list require content approval. Your submission will not appear in public views until approved by someone with proper rights. More information on content approval.

Title


Body *


Attachments