Ambiguity aversion and trade

What is the effect of ambiguity aversion on trade? Although in a Bewley's model ambiguity aversion always lead to less trade, in other models this is not always true. However, we show that if the endowments are unambiguous then more ambiguity aversion implies less trade, for a very general class of preferences. The reduction in trade caused by ambiguity aversion can be as severe as to lead to no-trade. In an economy with MEU decision makers, we show that if the aggregate endowment is unanimously unambiguous then every Pareto optima allocation is also unambiguous. We also characterize the situation in which every unanimously unambiguous allocation is Pareto optimal. Finally, we show how our results can be used to explain the home-bias effect. As a useful result for our methods, we also obtain an additivity theorem for CEU and MEU decision makers that does not require comonotonicity.