Testing for time stochastic dominance

We propose nonparametric tests for the null hypothesis of time stochastic dominance. Time dominance makes a partial order different prospects over based on net present value criteria general utility and discount function classes. For example, can be used ranking investment strategies or environmental policies expected future benefits. consider an Lp-type test statistic derive its large sample distribution under standard panel data sampling scheme with fixed dimension. suggest path-wise (or cluster) bootstrap procedure that allows individual series dependence horizon. describe two approaches, contact-set approach numerical delta method, may lead to enhanced power compared conventional least-favorable-case approach. prove asymptotic validity our testing procedures. investigate finite performance in simulation studies. As illustration, we apply proposed evaluate Million Baht Village Fund Program Thailand carbon emission trading China.