There are not many things as divisive in the investing world as the term diversification. The opinions range from concentrating in a few best ideas to owning the entire market, and everything in between. There seems to be an unwritten consensus that the best value investors hold concentrated portfolios, and to do otherwise would be return destroying. I've also seen a second trend where people talk about a guru investor's holdings and how they purchased a "basket" of stocks. Monish Pabrai bought a "basket" of Japanese net-net stocks, and other gurus have purchased other baskets. Calling a dog a duck doesn't make it a duck, it's just a dog with a funny name. In this post I want to discuss a few myths on diversification, so the next time an investor wants to add one more stock to their portfolio they don't feel ashamed. Maybe we can eradicate the investing world of baskets..
Warren Buffett says to concentrate
It is apparently well known that Buffett believes in a concentrated portfolio. Googling for this returns 3.59m results, but no direct quote. If there is a direct quote maybe a Buffett groupie who reads the blog can leave a comment.
Warren Buffett also recommends that most investors should put their money in index funds. There is actually a quote for this one "A very low-cost index is going to beat a majority of the amateur-managed money or professionally-managed money" (source) My guess is most readers aren't heading off to their brokerages to dump their stocks and buy index funds. So why pick and choose what Buffett says to do when creating your own portfolio?
As David Merkel notes Warren Buffett is different than the rest of us. I would submit to readers that Buffett's idea of concentration is a bit different than is commonly accepted. Berkshire Hathaway's website lists 50 independent subsidiaries. The annual report shows 14 major equity positions. It seems to me that owning 64 different companies in a variety of different industries is the very definition of diversification.
Only invest in your best ideas
The saying goes "Why invest in your 11th best idea when you can put money in your top ten?" This sounds great, but let's just take the question to the logical conclusion. Why invest in your second best idea when you can put all your money in your best idea? If as an investor you really have the ability to know your absolute best investment idea why waste money in anything but the top idea? The problem is we think we might know our best idea, but we really don't. My two best investments where companies that I thought at most could double, I just held and they became 10-baggers. I waffled on buying one of them, and I almost sold the other the day it doubled.
Diversification is protection against ignorance
This is another Buffett quote where he states that "Diversification is nothing more than protection against ignorance." Now I recognize in comparison to Buffett I am ignorant, I'd venture to guess most investors are. Not all of my investments go straight up, and I make plenty of mistakes. I would rather be ignorant, and understand my weakness than suffer from over confidence. I look for a large margin of safety when I invest because I know that it's easy to make a mistake. Buying with a bigger discount allows me to make more mistakes and avoid losing money. I'd prefer to never make a mistake, but that's just a dream, and nothing more.
My returns will be ruined if I spread my bets
This is my favorite myth, and one that I feel is misunderstood, just like how average companies are good investments. When I invest in a company at the point of purchase I look for a 25% discount to tangible assets (for a 50% return), or a 10-15% return hurdle. Why the 10-15% range and not something higher? Some companies growing at 6% might also be selling at 35% of book value, I'm happy to take the lower ongoing return for a bigger asset discount. Likewise some companies are selling close to book or above book, so I look for a higher earnings yield.
Every stock I add to my portfolio has the same return characteristics, either a 50% upside or 10-15% ongoing growth, or both, or both and more. Some companies have higher earning yields, but they might be less stable than a lower yielding company. As long as every company I add to my portfolio meets my minimum return potential how does adding one more company dilute my returns? It doesn't.
At this point some readers will point out that searching for stocks that match strict criteria such as profitable net-nets lowers my investable universe. This is true, but there are still plenty of very cheap stocks. As an example this weekend in Barrons there was an article that mentioned there are 183 small cap Japanese stocks selling for less than net cash. Putting a profitability filter on those stocks might reduce the set to 100; why not just buy them all? It would be hard to imagine an investor having trouble building a portfolio from 100 companies all selling for less than net cash and not earning a positive return.
What it all means
I think most investors would agree that extreme diversification is pointless unless it's the goal, such as in an index fund. Paying a manager to mimic an index is akin to throwing away returns. As a value investor the money is made at the time of purchase. As long as I stick to my process and look for a solid margin of safety, a tangible asset discount, and a discount to earnings I don't see how adding one more company that fits my criteria can do anything but help my portfolio. I doubt this post will change anyone's mind, but it might make people think a bit deeper about their style and assumptions.
Talk to Nate about diversification
Disclosure: none
Good post Nate. Agree as long as you find good ideas that fit your return view they are worth buying.
ReplyDeleteI also looked at my portfolio concentration a while ago and came across a great presentation by Zeke Ashton.
I wrote an article about it which you can find here:
http://www.eurosharelab.com/newsletter-archive/255-how-concentrated-should-you-be
I hope it helps.
Here's a quote from Buffett's 1961 partnership letter with a brief comment on diversification of the "generals" bucket in his portfolio:
ReplyDelete“The first section consists of generally undervalued securities (hereinafter called “generals”) where we have nothing to say about corporate policies and no time table as to when the undervaluation may correct itself…Sometimes these work out very fast; many times they takes years. It is difficult at the time of purchase to know any specific reason why they should appreciate in price. However, because of this lack of glamour or anything pending which might create immediate favorable market action, they are available at very cheap prices. A lot of value can be obtained for the price paid…This individual margin of safety, coupled with a diversity of commitments creates a most attractive package of safety and appreciation potential.”
Here's another quote from his 1958 letter. He was certainly willing to stake a large portion of partnership assets in single ideas. But he wasn't always concentrated. Remember one of his letters noted the total number of positions in the 40 range. But if the opportunity came up, he'd concentrate, essentially when the odds were in his favor. But my guess is having some measure of control over the destiny probably played into his thinking as well.
“This new situation is somewhat larger than Commonwealth and represents about 25% of the assets of the various partnerships. While the degree of undervaluation is no greater than in many other securities we own…we are the largest stockholder and this has substantial advantages many times in determining the length of time required to correct the undervaluation.”
I agree too Nate, good post.
ReplyDelete"The problem is we think we might know our best idea, but we really don't."
ReplyDeleteIndeed. There's a PDF being linked (sorry, don't know what it is) on Twitter about catching falling knives. It noted that the median return was quite poor, but the mean was performed above average. I seem to recall Geoff Gannon saying the same thing about net-nets.
So it could be that there's a significant "long tail" to value investing - in other words, it's the few outliers that hit the ball out of the park, pay off the losers, with performance left to spare. That requires diversification, not concentration.
I’ll pipe up on this.
ReplyDeleteFirst, there are two separate objectives in diversification which I’ll call (i) balance and (ii) sheer numbers. Balance refers to the avoidance of (or to the seeking out of) heavy exposure to a particular identified risk. If you own 20 stocks and they’re all in construction, you’re diversified in the second sense but not diversified in the first sense. When people talk of this or that supposed guru holding a basket of stocks they are talking about diversification via numbers (the second sense) but concentration in the first sense. E.g. Templeton held very many Japanese stocks in the 1960s: he was diversified but not diversified. Two different concepts. The sense in which you’re applying “diversification” in this post doesn’t apply to, or coincide with, “basket” as used in the conventional sense.
Second, diversification (in the sheer numbers sense) is not a binary concept. I know of only one person who would put all their money in one stock. It’s a foolish thing to do, not because one cannot know a stock that well – one can – but because shit happens: acts of god that can hit one stock and wipe it from the face of the earth in the blink of an eye. I know HA very well, for example, but if a nuclear N. Korean (or Israeli) nuclear missile hits Maui, HA is done for. And if all my money’s in that one stock, I don’t get to fight again; I, too, am done for.
That’s why I don’t put all my money in my best idea, not because I think Halstead or GEA are just as attractive on a risk/reward basis. I know they’re not. I know that, barring an act of god event, every dollar I devote to my 2nd and 3rd best ideas will reduce the marginal return on my investment dollar.
The question then, is how many stocks does one need to own to practically eliminate non-market risk.
And the answer depends on what you’re trying to do and what form the playground takes.
If you’re trying to mimic the market, 1 stock will do it.
If you’re trying to outperform the S&P 500, it could be 500 stocks weighted equally or 500 minus the airlines, whatever that comes out to, say 488.
If you’re investing in risky and/or low quality assets – distressed debt, highly leveraged equities, net-nets – you’d be crazy not to diversify massively. With these stocks you’re hoping that something changes, that the future will be different from the present or the past. For 80% or 90% of them, nothing will change – they’ll die. But the Lazarus group will do spiffingly well and will more than make up for the failures. That’s a strategy or pattern that relies on (numeric) diversification, that lives or dies by it. 20 to 30 stocks, equally weighted. That’s what Graham, Kahn, Shloss, etc did and that’s why they did it.
If, on the other hand, as “Buffett prime” was (as we know,there is no Buffett, there have been at least three Buffetts), you’re a stock-picker focused on quality issues, 4 to 7 stocks will do it; more will kill your upside. I will venture this: there is no way to earn 30+% annual returns over a full business cycle by holding more than 7 stocks at a time. And that is the context in which most intelligent investment commentary cautions against diversification. In this particular context , diversification is most definitely a bad thing.
And I agree with WEB that most small investors should avoid stock-picking altogether and stick with an index (or, even better, a value-weighted index) because stock-picking makes heavy demands of one’s time and one’s character and, unless one enjoys that sort of thing, one will take short cuts – naïve extrapolation of past trends into the future, interpreting market share as economies of scale, etc – that will leave one exposed as naked when the tide goes out.
So yes, diversification is very bad -- in some contexts and for some purposes. And yes, diversification is very good -- in some cases and for some purposes.
The important thing, in my views. is to match the diversification strategy with the investment strategy.
I think it is worth noting that WEB's universe is much smaller than ours.
ReplyDeleteI was at an annual meeting where he said that if he had only a million dollars to manage, he believes he could achieve a 50% annual return.
He did not say how... but I am guessing he would be much more concentrated than BRKB is.
Also, during his initial purchase of KO stock back in the 80's, I think he had over 25% of the company tied up in this stock. American Express (during the Salad Oil think).... similar I believe.
Tim... thanks for the link. Good article.
There's a video on neatvalue.com where someone asked him this question. He said net-nets, special situations and smaller stocks. I'm not Buffett, I never will be so I'm not expecting to do 50% a year, 15% is fine with me.
DeleteGood article, but let's take your proposition (adding one more is okay) to its logical conclusion. Where does it stop? The 100th idea? The 1000th idea?
ReplyDeleteIn answer to Anon above, if all stocks meet the investment criteria then I guess it would stop based on trading costs. If you have £1000 in an account and you pick 1000 stocks that's not going to work out well, i.e. the trading costs are more than 100% on each stock.
ReplyDeleteAs for diversification, I like to think of it in all senses - pure numbers, geography, industry, operational, plus anything else I can think off. It's not just for safety either as some geographies are booming while others are in bust mode, and the same is sometimes true of industries. When one industry get popular you can sell and buy something that is having a hard time, for example.
So that means if you have a $10m portfolio, you will go to 1000 stocks?
ReplyDeleteI am disappointed that the blog writer sees it unfit to comment on an issue that goes to the whole basis of his article.
The whole argument, as I see it, against concentrating and for diversification is that one cannot reasonably rank ideas (dubious proposition IMO) and the concentration idea, taken logically, means it is possible to reason for a one stock portfolio.
I think the article has missed the whole point.
Firstly, 8 to 20 positions uncorrelated positioins will confer close to 95% of the benefits of diversification. Anything more confers very little additional benefit for the additional work involved.
Secondly, the idea against overconcentration is to protect against risk of ruin. We need to refer to Kelly's criterion, and the logicality of why a conservative half-Kelly is always better than betting over Kelly. The arguments are muddied somewhat due to the difficulties in determining exact probabilities of risks, which is why positions are usually only a fraction of Kelly's criterion.
Thirdly, it is quite...ambitious...to believe that one could have heaps and heaps of great ideas. This goes back to the heart of risk management in value investing.
If there were 1000 stocks that fit my criteria I guess I could.
DeleteI find comments fascinating because often readers such as yourself read things I never wrote but you inferred. If I am looking at 10 ideas and only have enough capital for 5 I'd take the five "better" ideas. What does that mean? I'd probably be some combination of cheapness and prospects. There's no rule that I have to buy all 10 because they fit my criteria for a good investment.
As I mentioned in the post, there are 183 companies in Japan trading below net cash. That's a lot of ideas. Expand the investable universe a bit and I'm willing to bet there are easily 500 or more companies that have similar return characteristics. Maybe it's not worth buying them all, but that's a LOT of companies to dig through.
Nets nets ala Graham, and Buffett concentration of holdings in great business are two entirely separate things. Which one are you talking about? Buffetts idea of concentration do not apply to net net investing.
ReplyDeleteNeither specifically, I don't think anyone just invests in net-nets, although some might just buy good companies. I try to buy net-nets when they're attractive, special situations, and the proverbial good company at a great price.
DeleteAs a reference 35% of my portfolio is in net-nets, another 15% are low P/B or low EV/EBIT stocks, about 20% are special situations, and the rest are growth companies. When I say growth I mean moat companies that have growth that I bought when they were cheap and have since recovered.
I agree concentrating in net-nets is crazy, but some people do it.
Hi Nate,
ReplyDeleteNice blog!
If you are still looking for a direct quote, take a look at this video. The topic is brought up at about 1:30 into the video.
To find the video, search on youtube for "Warren Buffett MBA Talk - part 8".
-Igor
Hi
ReplyDeleteHow many compagny in our portfolio ?