"All the market return comes from just 4% of shares"

[Duke of Marmalade]

Yep

 

[Gordon Gekko]

Is the issue here that people are looking at it the wrong way? It’s not about identifying the winners per se, which is bloody tough. It’s about avoiding the losers which, although still difficult, is easier than identifying the winners. The more that I read up on this, the more my own view strengthens; basically, it’s mad to own less than (say) 30 shares.

 

[Brendan Burgess]

Hi Gordon

But the stars give most of the gain, so you have to have one or two of them in your selection.

You can lose only 100% on a dog, but the stars are unlimited.

I thought that 10 blue chips were enough - based on experience. I would like to see the data for other markets.

Brendan

 

[Duke of Marmalade]

Positive skewness is when a few stars pull the average of the index up above its median. In positive skewness the fewer stocks you hold the more likely that you will underperform the index by missing the stars.
Negative skewness is when a few dogs pull the average of the index down below its median. In negative skewness the fewer stocks you hold the more likely you will outperform the index.

The empirical evidence is that there is positive skewness, thus you should hold more stocks. So Gordon is right but I humbly suggest for the opposite reason.

 

[Gordon Gekko]

Which is easier Duke, to pick a winner or to avoid a loser? I dare say the latter.

I broadly agree with your point; in my view lone wolves such as ourselves cannot effectively run concentrated equity portfolios and succeed.

 

[cremeegg]

The article states that the average growth of the S&P was 1.5% higher than the median growth.

If an investor correctly buys the market, ie buys each share in proportion to its market cap, then he or she receives the average return, not the median.

 

[Duke of Marmalade]

If an investor correctly buys the market, ie buys each share in proportion to its market cap, then he or she receives the average return, not the median.

yep and that's why there is such a trend toward passive investment. An interesting thought experiment is if everybody is a passive investor, who sets the price?

 

[Sarenco]

The more that I read up on this, the more my own view strengthens; basically, it’s mad to own less than (say) 30 shares.

It's not mad if your objective is to beat the market return.

Which is why I don't understand Brendan's question...

The main point for investors is whether 10 shares is enough.

Enough for what exactly? Enough to achieve something akin to the market return? Enough to materially reduce the idiosyncratic risk of individual stocks? Enough to diversify accross sectors, geography, company size or something else.

Without knowing an investor's objective, it's an impossible question to answer.

 

[john luc]

just being reading here and curious but what criteria do ye apply when picking a stock. Do you use market cap, industry leader, management team,

 

[Colm Fagan]

Whilst I don't have any statistics to prove the conjecture, I am confident that a portfolio of 10 shares selected randomly from the top 100 shares in the world index, with weightings broadly consistent with their respective market valuations, would deliver the same return as the index with a margin of error of 1% in any 12 month period. In the shorter term, more significant variations from the performance of the index are likely, but the longer the measurement period, the smaller the variation in annual percentage terms. The experience of the "sloth fund", which I mentioned in posting #15 on this thread, supports this contention.

A portfolio constructed on these lines has a number of advantages. Buying a small number of shares reduces the costs (assuming you're dealing with a broker that has a minimum fee irrespective of deal size). You also get to understand the underlying businesses, which is vitally important from my perspective. I cannot stand the idea of buying into businesses without having some idea what I'm buying into. I love to be able to identify with the companies and the major players in those companies. I think 10 is about the most that an average private investor can get their heads around. Of vital importance too is the fact that you save on investment management fees. The saving would broadly equal the margin of error.

Further advantages accrue over time. As you get more familiar with the businesses, you can vary your weightings (gradually) to favour the companies that you have more confidence in. I would not advise anyone to get to the levels of concentration I have now reached with some of the companies in my portfolio. I recognise that I am definitely an outlier and I would not advise anyone to follow the approach I’ve taken, but it’s possible to achieve superior performance over time by shifting weightings on a gradual basis.

 

[Sarenco]

The experience of the "sloth fund", which I mentioned in posting #15 on this thread, supports this contention.

Not really Colm. The sloth fund was initially constituted with 30 (not 10) US "mega-caps". If you established a portfolio with 30 of the largest cap companies from the S&P500 then, yes, the correlation with the S&P500 as a whole would be very high.

However, these days ETFs that track the S&P500 have TERs of as low as 5bps so nobody bothers trying to establish a portfolio that acts as a proxy for the market anymore - they just buy the market!

10 stocks randomly selected from the ~10,000 publicly traded companies across the globe is not going to be anywhere near as closely correlated.

 

[Colm Fagan]

10 stocks randomly selected from the ~10,000 publicly traded companies across the globe is not going to be anywhere near as closely correlated.

I accept that the performance of the 10 stocks will definitely not be as closely correlated to the index as an ETF. Sometimes it will perform slightly better than the index, other times slightly worse than it, but blind adherence to an index is not the primary goal as far as I'm concerned. Far from it. I cannot bear the thought of investing in companies where I cannot identify with the companies concerned. That to me is an absolute no-no. There’s no fun in it. Like most other things in life, we should try to get some fun out of investing, to enjoy it. I hope that my “Private Investor” postings show that investing can be exciting. Also, as I said in my final paragraph, following those 10 companies over a number of years can help you identify which ones are superior performers and possibly more importantly, as Gordon Gekko noted earlier, which are potential laggards. Those insights come, not from detailed technical evaluation of financials, but from the consistency of the message from the Board and senior management and from the language used. You don’t have to be a financial genius. Over the longer term, I am pretty confident that this strategy will deliver above-average performance, without costing you a cent in investment management fees. Try it!!!

 

[jpd]

An interesting thought experiment is if everybody is a passive investor, who sets the price?

I think about this too. Presumably there will always be enough active investors to move the price to it's "proper" value, whatever that is!

 

[Duke of Marmalade]

Whilst I don't have any statistics to prove the conjecture, I am confident that a portfolio of 10 shares selected randomly from the top 100 shares in the world index, with weightings broadly consistent with their respective market valuations, would deliver the same return as the index with a margin of error of 1% in any 12 month period.

Colm, that seems a very strong conjecture. I know picking randomly from FTSE 100 is not directly comparable with your algorithm but it was the most available to a Google search. I have uploaded my findings (I love uploading). This is the current 12 month performance of the individual components. I was really startled by the variance between companies. No less than a staggering 29% standard deviation In words, using say 2 standard deviations as a measure of outlier a single stock picked at random might be 60% off the average. For 10 stocks chosen at random the standard deviation would be 9% with the outlier being a possible c.20% off the benchmark.

 

[Colm Fagan]

Duke, I guessed that you would be on the job!! Yes, I'm surprised too by the differences. Did you allow for weightings relative to their respective market capitalisations? It would also be interesting to know if my related conjecture that the variation reduces over time holds true. Can you look at those same stocks over (say) a three year period? For completeness, we should look at total returns, not just price.

 

[Duke of Marmalade]

Duke, I guessed that you would be on the job!! Yes, I'm surprised too by the differences. Did you allow for weightings relative to their respective market capitalisations? It would also be interesting to know if my related conjecture that the variation reduces over time holds true. Can you look at those same stocks over (say) a three year period? For completeness, we should look at total returns, not just price.

Nothing as fancy as market weightings, and I think am using the FTSE price only index, but it gives a flavour of the concepts involved. I upload the spreadsheet with 5 year figures included. Yes the standard deviation falls from 29.2% to 11.8% but we would expect that order of reduction fro purely statistical considerations (divide by SQRT(5)).

 

[Duke of Marmalade]

forgot to upload

 

[Colm Fagan]

Duke
My conjecture isn't looking good at this stage, but hasn't been disproved - yet.
I still hold with my other (subjective) contentions re being able to identify with the companies in which my money is invested, getting some craic from it, and being able to modify and improve the strategy as I become more familiar with the companies. Remember that the investment horizon is at least 5 to 10 years, so there's plenty of time to get to know the companies and the people running them.

 

[Duke of Marmalade]

I'm almost in a state of shock over the variation between stocks in the FTSE 100. The 12 month performance varies from -42% to +160% (ok +66% if we ignore the outlier Evraz). I think of the index as being a relatively tame beast but underneath it is a sea of turmoil.

 

[Sarenco]

I'm almost in a state of shock over the variation between stocks in the FTSE 100.

And that's just large caps - imagine the variation in returns across the full FTSE All-Share Index.

Unecessarily taking on the obvious risks of a highly concentrated 10-stock portfolio wouldn't be my idea of fun. But each to his own I guess.