How Many Holdings Makes Sense?

My post yesterday about Marty Whitman drew a lot of comments at Seeking Alpha with one in particular worth exploring a little more.

So, to the author of this column I say: If you want to invest in your 35th, 36th, and 37th best ideas, go right ahead, but I don't know anyone (including myself) who thinks that even his *sixth* best idea is as good as his top five.

A lot of people assume that in a diversified portfolio there should be a best idea and then some sort of general priority or ranking of holdings where something must be the best idea and something must be the weakest (not sure whether weakest is the best word). This certainly can be a valid mindset and I have zero doubt that people have success allocating the largest weighting of their portfolio into their "best idea."

Lost in my article (it is not possible to recreate every thread I've written about for each post) is the way I prefer to construct a portfolio and how the framing of the comment quoted above is not how I choose to build and then manage a portfolio. While I believe my method is valid so too are many other ways including that of the commenter.

I build portfolios from the top down benchmarking against the S&P 500 total return index (price plus dividends). My opinion about how to do this is to consider each of the ten big sectors in the index and decide whether to underweight, overweight or equalweight versus the index. This is done by combining what I know about market history with what I believe is going on now which hopefully yields a correct forward looking analysis.

The easiest example is the financial sector from a few years ago. I had been concerned about that sector's weight exceeding 20% of the index (typically a bad omen) and so had been a little underweight years before the crisis. When the fundamentals began to deteriorate coinciding with the inverted yield curve, I was not willing to give the sector the benefit of the doubt and so got aggressively underweight. I did not go zeroweight as I view zero to be a big bet which I tend to avoid.

As there are ten sectors that gets me to at least ten holdings. As a matter of philosophy 10% in one stock is more than I am willing to go. In addition to sector decisions I am a big believer in foreign investing at the country level, meaning there are some countries I want to own and some I do not. In choosing countries I am looking for countries with different fundamental attributes than the US (better chance at diversification than Western Europe ex-scandies), on firmer economic footing than the US (rules out Western Europe ex-scandies and Japan), have some reasonable prospects for doing well and that have something the world needs (mostly resources or labor).

Countries we own now (excluding narrow client mandates) include Brazil, Chile, Canada, Australia, Israel, Switzerland, Denmark, Norway, Sweden, UK and a little bit to Asia via a couple of narrow ETFs. In the next couple of years I can see going back into China, adding more Asian countries, going back to South Africa, more in Latin America and further down the road countries that are not easily accessed now depending on developments in those places. Forgetting the sectors for a moment exposure to all those places could require 15 or 16 different holdings.

Also embedded into my idea of a diversified portfolio is exposure to themes which can include water, infrastructure or food. There of course can be and is overlap; I've disclosed many times owning Vale (VALE) for clients which covers part of the materials sector, Brazil and an ascending middle class where there used to be no middle class.

Additionally I try to manage things like volatility, average cap size and yield of the portfolio. As an example with two stocks I have never owned selling AT&T (T) and its $170 billion market cap and replacing it with Frontier Communication (FTR) and its $8.5 billion market cap can reduce the cap size of the entire portfolio by a noticeable amount which at times makes sense to do (usually earlier cycle).

One of the building blocks of my process is that a company can add value for many years but not always be a great performer--most stocks that turn out to be great over long periods of time can't out perform 100% of the time. As an example I've owned Teva (TEVA) for clients for about six years. Looking at a chart going back exactly six years the stock is up 97% versus up 7.75% for the S&P 500 (price only). However in that six years there have been short stretches where the stock has lagged the index noticeably. Looking forward there is no convincing me the name won't continue to add value to the portfolio but there will be periods of underperformance, YTD it has lagged the market.

A final strategic point is that any stock can be the top performer in a portfolio. I've mentioned a few times being surprised in the past at what name ended up being the best performer for a given period. A tie in here is that not every stock chosen can work as expected. I mentioned this yesterday. If I know a couple (or more) will be wrong, then yes having more holdings minimizes the consequence of those incorrect picks and while not everyone believes in that I do.

The above is neither universally right or universally wrong, it is right for me. Whatever you do should be right for you, which is a point I have made many times before. Another point made often is that no strategy can be the best for all times and that includes Whitman's. I believe it is crucial to understand the weaknesses in your strategy so as not to have any sort of panicked reaction to adverse market conditions. And we didn't even get to my thoughts on taking defensive action in the portfolio; no post can cover every past thread.