There are three main limitations of the DW test as a test for serial correlation. First, the dis- tribution of the DW statistic under the null hypothesis depends on the data matrix . The usual approach to handling this problem is to place bounds on the critical region, creating a region where the test results are inconclusive. Second, if there are lagged dependent vari- ables on the right-hand side of the regression, the DW test is no longer valid. Lastly, you may only test the null hypothesis of no serial correlation against the alternative hypothesis of first-order serial correlation.