Nov 072011
 

Dylan Grice, the ever-interesting strategist at Société Générale, put out a good note recently arguing that quality investing outperforms. In other words, buying safer stocks leads to higher returns on average.

This is not what conventional investment thinking says. The standard rule of thumb is that risk and return are positively correlated. Getting higher returns means taking on more risk.

Taken over the long run, that holds true across assets. The data shows that higher risk asset classes have delivered higher returns over time. The chart below shows long-term returns for US asset classes (chart is via Grice’s note, data taken from Expected Returns by Antti Ilmanen).

This pattern isn’t just true of the US. Take the data that Elroy Dimson, Paul Marsh and Mike Staunton have compiled on stock and bond returns from 1900 to 2010, covering 19 countries. In each case, the very long-term equity risk premium has been positive – in other words, stocks have beaten bonds over long periods (and usually by a good margin) .

(Chart taken from the Credit Suisse Investment Yearbook 2011. Obviously the last ten years have been bad for equities in many countries, showing that the long term can sometimes need to be quite long.)

So far, no surprises. However, this relationship does not seem to hold within an asset class.

For example, the riskiest CCC-rated bonds would in theory be expected to outperform the safest AAA ones. But Ilmanen’s data on average returns in the bond market since 1952 shows the two extremes actually perform about the same. Mid-ranked bonds around BBB and BB (just above and below the cut-off for investment grade) have done the best.

So historically, buying the riskiest bonds did not get you the highest returns. With stocks, the relationship is even more inverted. Grice’s chart below shows that higher-quality shares have given higher returns than low quality ones over time – the exact opposite of the usual assumption on the risk/return trade-off.

Quality here is measured using the Piotroski score, which is essentially a nine-point checklist of firm’s financial health – higher scores are better. Crucially, there are no valuation measures in the Piotroski score, so this selects stocks purely because the business appears sound, rather than cheap.

And in fact, it turns out that stocks with higher Piotroski scores have typically traded on higher valuations – yet despite this have delivered higher average returns. In other words, investors have paid more for quality companies (as you might expect), but they haven’t paid as much of a premium as the stocks are worth.

This is by no means the first study to show this. The argument that quality investing outperforms has been around for a long time and there is a fair amount of data to support it. Benjamin Graham’s work back in the 1930s distinguished between good quality and low quality stocks – something that many value investors who aim to copy his approach don’t always take on board.

The reason why good quality good businesses might outperform is easy to see: they are usually in better financial shape, in stronger positions in their markets and better managed. Hence they can survive difficult times and take advantage of new opportunities better than weak ones.

However, this should be obvious. Exactly why the stocks of these quality firms should be persistently undervalued by investors – and so able to deliver better returns – is less obvious.

It may be psychological. Good quality stocks often look dull and predictable. Lower quality ones seem more exciting and have more potential upside.

Investors may simply be drawn to the riskier long shot because they focus too much on the upside and now enough on the downside. Sometimes that long shot will come off and deliver better returns than any quality stock. But they’ll offset that by paying too much for the ones that don’t.

There’s a fair amount of evidence from horse-racing that bettors overvalue longshots and undervalue favourites. While professional investors might like to think they’re more disciplined than gamblers, it wouldn’t be surprising to discover that they’re not.

 Leave a Reply

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>

(required)

(required)