There are well-known theoretical concerns regarding the use of price correlations to determine antitrust markets. However, this has not deterred their use or the application of Granger causality, stationarity, and cointegration tests in the determination of antitrust markets. In this paper, we explore the empirical performance of these various tests. In particular, we want to know whether these tests are capable of generating the correct inference both when two products are in the same relevant market and when they are not. Our results imply that, in the absence of common shocks, simple price correlations may be capable of providing reliable evidence on market delineation. However, in samples sizes similar to those currently available, the performance of other commonly employed price-based tests suggests that they provide little economically meaningful information to antitrust practitioners.