At the end of the year, when people inevitably start comparing their performance with others one common theme usually emerges: every great stock picker runs a concentrated portfolio. When you “only” generate a 15% positive return while others are making 30%, 50% or even 100%+ it’s easy to envy those mind-boggling returns and start wondering what you should do to emulate those results. When you learn that those results were without exception generated by investors who had the balls to concentrate on just a few good ideas the implication seems obvious: time to bet big!
And as a matter of fact the undeniable truth is that the only way, to achieve those mind-boggling returns, is by having a concentrated portfolio. Intuitively concentration makes sense, but unfortunately everybody’s intuition sucks at statistics. What is ignored here is the base rate of people trying a concentrated strategy. How many shoot for the moon only to end up with an evaporated portfolio at the end of the year? Because that is the flipside of the story; the investors with the mind-boggling losses are also the concentrated investors! But that is, of course, a story that most people prefer not to tell (kudos to the few that do!).
Let’s believe for one second in an efficient market and imagine that 50% of investors are concentrated and 50% are diversified. Since the market is efficient no-one can generate any alpha, yet at the same time a large number of investors is expected to outperform the average. The diversified investors should generate a return that is close to the average, but approximately 50% of the concentrated investors – 25% of the total population – is expected to outperform the average while the other half is expected to underperform. Depending on how concentrated they are some could outperform spectacularly. A naive interpretation of the phenomenon would be that concentration would be the way to go while the opposite is true since it only increases risk in an efficient market.
If you are reading this you probably don’t believe in an efficient market. I know I don’t. But the reality is that most outcomes in investing are the result of random noise. You can’t know what will happen. No-one expected oil to drop from $80/barrel to less than $50/barrel in two months time. No-one expected that the Swiss franc would gain 30% in minutes yesterday. You can only expect the unexpected. And the only realistic way, to protect yourself against the unexpected, is to be diversified across industries, countries, and asset classes.
Another problem with the large amount of luck that is involved in investment results is that it is easy to look like a genius and think that you are while you aren’t. Even when you are a pretty good investor it is incredibly hard to be right consistently, and you are probably delusional when you think that you can outperform the benchmark by a large margin. But the bigger problem is – even when you are realistic about your skill – that there is a large amount of uncertainty about your skill level. There is always the possibility that you don’t add alpha.
If you concentrate and you are wrong about your skills it is almost inevitable that you will blow up at some point in time. Betting too big relative to the expected value of the bet will result in a suboptimal geometric growth rate: that’s what Kelly teaches us. Investing isn’t about hitting a great return in a single year. It’s not even about generating on average a great return: any series of returns, that includes a zero, has a zero as the final portfolio value! The geometric growth rate is the only thing that matters and if you keep everything else equal you maximize it by reducing variance.
To be fair: at the same time Kelly also provides a strong argument for concentration. If you find a great idea with limited downside and/or a small probability of a large loss and a large upside the formula is probably going to suggest that you put the majority of your portfolio in a single stock. But you have to account for the uncertainty in your estimates, including the unknowable, making a strict application of the formula unwise.
It is also only true when it’s the only good bet available. There are more than 100,000 listed companies worldwide. If you aren’t able to find a decent amount of reasonably comparable good bets your are probably lazy or the markets are incredibly efficient. But if they are so efficient is it reasonable to assume that there is somehow one overlooked huge opportunity? That’s not something I would want to bet on, and being lazy is certainly no excuse.
I consider myself to be concentrated and diversified at the same time. The fact that I’m writing this post with 50% of my portfolio in my top six positions is kinda ironic…
Very good post, thanks!
Of those top six stocks, one is Berkshire Hathaway, which I wouldn’t count as the same level of concentration as e.g. Conduril.
Absolutely, and I actually included cash as one of those positions which is obviously even lower risk. But note that I’m not saying that you never should concentrate and always diversify across a huge number of positions. It’s not what I’m doing! But when you read the posts of some people you start to think that a three stock portfolio is already over diversified.
I dont think concentration is the root cause of blowups. The main problem is overconfidence, leading to concentrated portfolios / the use of leverage / buying ‘story stocks’ without a margin of safety. And the problem with investing is that it appeals the most to those who are overconfident in the first place. And that the variance of your risky strategies will only increase your overconfidence until you go broke.
The hard thing about investing is that pretty much every human bias (as described by Kahneman for example) works against you.
That’s a good point. Overconfidence is a huge factor in excessive risk taking, but I don’t think it is necessarily the root cause. If the flawed idea of the requirement of (extreme) concentration would be less widespread people could still be overconfident while being diversified at the same time.
Hi, great post, i will look into this geometric series things.
My take on it is that one should be humble with regards to how much correct information one is able to gather about a company and the market. Its not possible to know it all.
Assume for a company there is a total of N things to know, where each of those N things is significant enough to affect the outcome of the company and therefore the share price in the future. Then you should ask yourself how big % of N is it possible for a human and yourself to know? If you know all the N things you will not lose and there is no reason to diversity. But its of course impossible. E.g the CEO will die in 1 week and had tremendous importance for the business which will from then on deterorientate. Or e.g you bought a cheap stock but it will take 4 years for people to notice and in the mean time new technology or a disaster ruins it for you. Can you know 40, 50 or even 60%? A financial report is limited.
Also one must be aware that of all the knowledge you (a human) has gathered over the years a significant percentage of it is just going to be plain wrong. This is just because you will not live infinitely long and cant use your energy to check and recheck every aspect of the information you gather. The longer you live the more sense it makes to invest time in getting rid of wrong information since the consequence would be to be wrong about something infinitely many times. But humans die after a short life so we dont have time for that stuff and are better off taking in alot of low-filtered information every day and being wrong a decent amount doesnt follow you to infinity.
At the same time, yesterday i read a really solid blog post ( the last) by the red corner and it seems that he has 20p cash and 4-6 stocks at any time, having the opinion that having 8-12 stocks would be to stretch his abilities.
I think that you have to recognize that uncertainty can exist on a more fundamental level. You don’t even know how big the number N of important things is, and many of these things are unknowable no matter how much experience you have or how much research you do. You don’t even really know how important the stuff is that you don’t know.
That is sort of my point. N is very big and unknown to any human, therefore one should not concentrate.
Thank you for all your posts! They are very helpful.
There is one additional benefit of diversification that I have never seen mentioned. If you are an investor who, while holding a security long, is willing to buy on down trends and sell on up trends I believe that additional money can be made as compared to a strict buy and hold strategy. Sure, sometimes you buy on a down trend and the stock never recovers; other times you sell and the stock keeps going higher and you never buy back to your original holding amount. But I find that most stocks move up and down, up and down at the micro month-to-month or year-to-year level. A diversified portfolio that you are watching daily and keeping up with in regards to news affords you many more “opportunities” to exercise this type of investing because at any one time only a few stocks in the portfolio seems like a really good buy or are overpriced. Only a few stocks have good news or bad news and you are able to act quickly. One could argue that simply following a lot of stocks without being invested in them would give you the same opportunities to be a short/intermediate term trader but I find that being an owner keeps me more engaged and more likely to actually act. A concentrated portfolio of only 6 stocks would simply give you less “opportunities” to trade in and out based on aberrant price swings.
I don’t think that is a different benefit: it’s a different side of the same coin. When you are diversified you have more trading opportunities, but when you are concentrated you can bet bigger and make more money on a few opportunities (unless there are for example liquidity constraints).
I agree with you we suck instinctively suck at statistics, but I run my portfolio with 10 to 20 stocks. Here are 3 reasons why:
1) In their book Investment Analysis and Portfolio Management, Frank Reilly and Keith Brown report that in one set of studies for randomly selected stocks, “…about 90% of the maximum benefit of diversification was derived from portfolios of 12 to 18 stocks.” So to eliminate the unsystematic risk, you don’t need 100 stocks.
2) Part of the reason why UCITS have so many stocks in their portfolio is regulation (the 10/40 rule). You cannot have more than 40% of your portofolio on stock which represent more than 5% of your portfolio (max 8) and you cannot invest more than 10% into a single stock. But regulation is not a good enough reason to diversify.
3) I cannot seriously focus on 100 stocks at any point in time. And I challenge a team 2 portfolio managers to do so… Beyond financial theory, the main reason why I run a fairly concentrated portfolio is that stock sreening / selection / follow up is time consuming activity. And there are not so many great opportunities out there, the market is in fact fairly efficient.
I understand diversification, but this is not what stock picking is about, and you have so much time to spend on analysing stocks, so I am confortable in a 10 to 20 stock concentration (plus cash)! If I had a portfolio management team, that number would maybe increase to 30, but I would put a cap at that level.
I don’t know the book you are talking about, but you might be taking your decision based on a flawed interpretation of outdated research. A couple of random links that I found in one minute of using Google that all suggest higher numbers:
Also, note that this research is comparing diversification against the S&P 500 index, in itself not a diversified portfolio since it misses international exposure, small caps and exposure to other asset classes such as bonds. To replicate the diversification benefits of a more broadly diversified portfolio you would need more positions.
Agree with you on point 2 and 3 :).
“The diversified investors should generate a return that is close to the average, but approximately 50% of the concentrated investors – 25% of the total population – is expected to outperform the average while the other half is expected to underperform”
I don’t know why people do this all the time — they seem to conflate the median with the mean. To the extent you’re not confusing the two, what should be said here is that you are assuming that returns are symmetric (or that you’re going to use a coin flipping model for the sake of the example or… )
It is certainly possible to have an efficient market with skewed returns. This can occur many ways. The most intuitively obvious way probably occurs with certain investors being overweight a given sector (say financial companies in early 2008 or oil & gas companies in early 2014).
I was implicitly assuming a normal distribution of returns. You are right that you could have an efficient market with skewed returns.
Hi, I totally agree
we tend to be overconfident and underestimate the possibility of business developing in unexpected ways. I am very conscious of my own limits and like to have a diversified portfolio.
It is reasonable to split a portfolio in 10-15 or more positions (if we take some speculative bets such as commodities or emerging markets).
The only way I could see a possibility of a very concentrated portfolio would be to find non cyclical blue chips trading at very low valuations. It is not happening in these markets.