This paper analyzes how the combination of borrowing constraints and idiosyncratic risk affects the equity premium in an overlapping generations economy. I find that introducing a zero-borrowing constraint in an economy without idiosyncratic risk increases the equity premium by 70 percent, which means that the mechanism described in Constantinides, Donaldson, and Mehra (2002) is dampened because of the large number of generations and production. With social security the effect of the zero-borrowing constraint is a lot weaker. More surprisingly, when I introduce idiosyncratic labor income risk in an economy without a zero-borrowing constraint, the equity premium increases by 50 percent, even though the income shocks are independent of aggregate risk and are not permanent. The reason is that idiosyncratic risk makes the endogenous natural borrowing limits much tighter, so that they have a similar effect to an exogenously imposed zero-borrowing constraint. This intuition is confirmed when I add idiosyncratic risk in an economy with a zero-borrowing constraint: neither the equity premium nor the Sharpe ratio change, because the zero-borrowing constraint is already tighter than the natural borrowing limits that result when idiosyncratic risk is added.
In this Section: