This study examines the asset pricing implications of preferences over the higher moments of returns’ distributions. We show that in a market populated by risk-averse, prudent and temperate investors, firms whose returns exhibit negative coskewness or positive cokurtosis should yield higher premia relative to counterpart firms with positive coskewness and negative cokurtosis respectively. These theoretical predictions are empirically tested using a comprehensive dataset of shares listed on the London Stock Exchange during the period 1986–2008. Our empirical results confirm that coskewness and cokurtosis premia are genuinely priced in the UK market, over and above what covariance risk, size, value and momentum factors can explain. We also show that a theoretically motivated, higher co-moment asset pricing model has significant explanatory ability over the cross-section of coskewness and cokurtosis portfolio returns.

We run a field experiment offering graduated microcredit clients the opportunity to finance a business asset worth four times their usual borrowing limit. We implement this using a hire-purchase contract; our control group is offered a zero interest loan at the usual borrowing limit. We find large, significant, and persistent effects: treated microenterprise owners run larger businesses with higher profits; consequently, household consumption increases, particularly on food and children’s education. A dynamic structural model with nonconvex capital adjustment costs rationalizes our results and allows counterfactual analysis; this highlights the potential for welfare improvements through large capital injections that are financially sustainable.