This paper uses the wild bootstrap to compute empirically relevant critical values for the test statistics proposed by Emerson and Kao (2001). Monte Carlo simulations were then performed to evaluate the size and power properties of the bootstrapped tests.