Why the link between volatility and growth is both positive and negative.

I revisit the relationship between growth and volatility in two different disaggregated data sets. I confirm that growth and volatility are negatively related across countries, but show that the relation reverses itself across sectors. This phenomenon, sometimes called the "Simpson's fallacy", has a natural interpretation in the present context: it is the component of aggregate volatility that is common across sectors that correlates negatively with aggregate growth. Furthermore, while investment and volatility are unrelated in the aggregate, sectoral investment is shown to be more intense in volatile activities, as if the return to capital were higher there. These results call for a distinction between macroeconomic and sectoral volatilities, not unilke that between macroeconomics, where volatility often means policydriven instability, and fice, where volatility reflects risk, and thus high returns