How many shares should you have in your portfolio?

It’s not often you get important investment lessons from popular TV programmes.

So it was nice to see that ITV managed to squeeze a warning to investors into one of the plot lines on Downton Abbey. The well-heeled Grantham family learned just how dangerous it is to put all your eggs in one basket, when they lost all their money betting on a single Canadian railway company.

However, while diversification seems to be simple common sense, just how much of it do you need? And can you have too much?

How diversification works

We’ve written about diversification and asset allocation before. But here’s a quick recap.

Investing is risky. No investment is truly ‘risk-free’. But risk comes in different forms.

For example, if you buy a share in Tesco, you are betting on the quality of the company itself. If the company runs into trouble, the shares will probably fall. So you are taking ‘company’ risk.

But you are also betting on the supermarket business. If supermarkets as a sector start to fall out of favour – if the internet destroys their margins, say – then Tesco is likely to suffer too. This is ‘sector’ or ‘industry’ risk.

Taking an even wider view, there’s ‘market’ risk. Maybe investors just don’t want to buy British-listed stocks, or shun equities altogether.

Diversification can help combat all of these risks. By buying more than one supermarket company, you could reduce company risk. By buying a business in a different industry too – technology maybe – you could reduce your sector risk.

And you could reduce market risk by investing some of your money in stocks listed in another country, or even in a different asset class, such as bonds.

How well does it work?

This is the theory behind diversification, and – unlike many clever-sounding ideas – in practice it has been shown to work. A 1977 study by academics Edwin J Elton and Martin J Gruber showed how even a little bit of diversification could cut your risk.

It found that the average standard deviation of a US share over a year was 49.3%. That sounds technical. What it basically means is that roughly a third of the time, you could expect the average US share to be 50% higher or 50% lower than the point at which you bought it. Very nice if you’re up 50% – not so nice if you’re down that much.

But if you bought another share, the standard deviation of the portfolio would fall to 37.4%. Having ten shares would cut this to 24%. Diversifying across asset classes, and across countries also helps.

How much is too much?

But you can have too much diversification. If you have 100 shares instead of ten, for example, that’s 100 sets of dealing costs.

Unless you have a very large amount of money invested, that’s going to be very expensive as a percentage of your portfolio. In other words, holding that many stocks is effectively impossible for a retail investor.

So what’s the optimum amount? The good news is that Elton and Gruber found that moving from a 20 to a 1,000-share portfolio would only cut risk from 21.7% to 19.2%. So you don’t need to have a ridiculously large portfolio to be properly diversified.

Many of the best investors run fairly tight portfolios. Hedge fund manager Seth Klarman holds only 22 shares in his US portfolio. He thinks the ideal number is even lower at just 15. An Australian study in 2006 also found that households with the most focused portfolios tend to do better than those who spread themselves thinner.

This makes sense. The fact is, investing is hard work. No one with a 100-stock portfolio can possibly know every one of those companies well enough to be making properly informed decisions about them.

The best level of diversification does depend partly on the sector you plan to invest in. Someone buying into a specialist, high-risk sector where lots of companies go bust, such as biotech or small miners, will want to buy enough shares to ensure their winners more than offset their many losers. This is where a fund can come in very useful.

But for a more straightforward portfolio – such as large, income stocks, say – then an investor should be able to build their own diversified portfolio easily and cheaply enough, with around 16 to 20 stocks. My colleague Stephen Bland shows you how to do this in The Dividend Letter newsletter – you can find out more about his methods here.

The Dividend Letter is a regulated product issued by Fleet Street Publications Ltd.