On Industry Selection

Recently I received an e-mail:

Always enjoy your blog – very thought provoking, and I’ve learned a lot from reading you across a variety of topics. Assuming I haven’t missed this in an older post… one thing you mention as a key investment strategy is finding the right industry at the right time, and I’ve never seen a very good explanation of how one goes about that. In one of Jim Cramer’s recent books he offered a sort of stylized graph outlining a general playbook to that effect – I’ll send it to you if you’d like – but I’d like to get a primer on how you go about industry over/underweights.

Thanks and Best,

It made me think that I should go through my basic principles of industry selection, and explain them.  JC mentions Cramer’s “playbook” — that’s the classical guide to what industries do best in an “ordinary” business cycle.  Personally, I think Cramer’s views on industry selection are more complex than that, largely for the reason that I don’t follow the “playbook” in any strict sense: global demand is more important than US demand alone for many industries.  The old playbook is no longer valid, until we get a totally integrated world economy.  (Side note: we will never get that — some war will upset the globalization — it is the nature of mankind.)

Anyway, I have four basic tenets when looking at industries:

  • Buy strong companies in weak industries when the industry pricing outlook seems hopeless.
  • Buy moderate to strong companies in strong industries where the earnings power and duration are underestimated.
  • Underweight/Ignore/Short industries where pricing power is likely to be negative for several more years, and especially industries that are in terminal decline.
  • Avoid fad industries.  There are P/E levels that no industry can grow into.

My best example of #1 is the P&C insurance industry in early 2000.  Total gloom.  I bought a lot of The St. Paul then.  Another example: Steel in 2001-2002.  I bought Nucor.

For #2, think of the energy industry — current stock prices embed oil prices far below current levels.  Or, think of the life insurance industry — low P/Es, but the demographic trends are in their favor.

On the third one, think of newspapers, whose richest revenue sources are getting eaten up by the internet.

For the last one, think of the internet/tech bubble 1998-2000.  Very few companies that were hot then are at higher prices now.

I share the results of my industry model once a quarter at minimum.  But I don’t use my model blindly.  For example, lending financials and housing have been cheap for some time, but I have avoided them.  They are cheap for a reason.

My main model uses the Value Line ranking system, and uses the nominal rank, and how it is different from the average historical rank.  It can be used in two ways: highly rated industries can be analyzed to see where the pricing power is not reflected in the stock prices yet.  Low rated industries should be analyzed for the possibility or reversal due to undeserved hopelessness.

But you can create your own model just from a series of index prices.  The idea is to look at industries that either have strong momentum that you think is deserved, or industries with weak momentum where things seem very bad but not terminal.  You can even modify it to look at industries that have bad performance over the past 3-5 years, but have rebounded over the past 6-12 months.

Behind all of that, remember my rule: sharp movements tend to mean-revert, slow, grinding, fitful movements tend to persist.  Things that are too certain tend to disappoint, while those things that are less certain tend to surprise.

One reason I have done well over the past 7+ years is that I have been willing to let my industry selection vary considerably from where the indexes have been.  If you think that you have insight into the longer-term earnings power of industries, then take your opportunity, and deviate from market weightings.