It has been a tough five years or so for active equity managers in the US. In the main, correlation between different stocks has been high and the dispersion low.
This has meant, even if you have successfully picked the right stocks there has been less reward on offer because the best performing stocks haven’t been that much better on average. Dispersion is often discussed in terms of the active versus passives debate, but it can also tell us useful things about the nature of market drivers.
Why has dispersion been low in recent years?
A common argument is that stocks tend to perform similarly when big macro issues are driving markets. This seems intuitive; investors will be less prepared to focus on the prospects for a single company if the only issue that matters is whether the US banking sector will survive.
You can see this in the chart below, which shows a moving average of the dispersion of weekly returns of US (S&P) sectors. Dispersion declined materially in the years following the financial crisis, and had stayed relatively low until recently.
This mirrors what has happened at a stock level in the US, but also the trend around the world. Figure 2 shows dispersion measures in Europe and in Emerging Markets (looking at variation at a country level). You can see the same decline post crisis, and the same recent pick up.
These trends echo Ritu’s observations on correlation over at the Equities Forum last month. Stocks are showing more differential behaviour now than they were in the middle of last year and more than they have for much of the period since the financial crisis.
Why has dispersion picked up?
This pick up in dispersion (and fall in correlation) has confused many. Using the rule of thumb about macro risk above, surely we might expect the reverse? After all, it feels as if macro uncertainty is very high, as I discussed in December. Donald Trump must have raised macro uncertainty in the US?
Similar confusion has greeted the fact that implied volatility has been at or close to its recent lows in US equities, but also in many other assets around the world.
Last week I also discussed whether asset price moves were indicative of complacency in Europe. Ultimately it seems that the confusion is being driven by two elements:
1. We forget where we have come from. In January, I discussed the dangers of “efficient markets thinking,” that is, the tendency to believe that all price moves must have a reason behind them and ignoring the fact that assets might have been priced incorrectly to start with. As we have mentioned many times, the first half of 2016 was characterised by deep pessimism. You can see the lows in dispersion reached in February in figures one and two above. The simple unwinding of the sharp moves of H1 2016 accounts for much of the “game of two halves” nature of price action in the year as a whole.
2. Macro uncertainty doesn’t always affect all companies in the same way. While some of the pick-up in dispersion recently is simply the unwinding of the forces that had driven markets in the preceding period, this does not explain why dispersion in the US would reach a high For that, we have to look to the specifics of the current environment.
3. Trump has indeed increased uncertainty, but he has also made very specific policy pronouncements which would create clear winners and losers. At the same time – and more importantly – the concurrent rising interest rate environment has been a clear boost to US banks and a challenge for consumer staples. This is shown in the chart below (though the effect will also be exaggerated through the effect of rebasing).
A similar pattern will be evident in the UK post-Brexit. Yes, the Brexit vote has created macro uncertainty but stocks won’t all move together because of the role of Sterling on firms with overseas earnings versus those without.
Many commentators are concerned about the nature of recent equity market gains and confused that this has taken place against a background of an apparent increase in macro uncertainty. This may be a valid argument, but stock market dispersion should not be used as evidence in its favour.