This article proposes a test for the cost-based explanation of nonparticipation, by estimating a lower bound to the forgone gains of incomplete portfolios; these are in turn a lower bound to the costs that could rationalize nonparticipation in financial markets: high bounds would imply implausibly high costs. Assuming isoelastic utility and a relative risk aversion of three or less, for the stock market I estimate an average lower bound of between 0.7 and 3.3 percent of consumption. Since total annual (observable plus unobservable) participation costs are likely to exceed these bounds, the cost-based explanation is not rejected by this test.