This paper presents two computationally convenient specification tests for the multinomial logit (MNL) model. The first test is based on the Hausman (1978) specification test procedure, which tests the reverse implication of the independence from irrelevant alternatives (IIA) property. The second set of tests are based on classical test procedures, including the Wald, likelihood ratio (LR), and Lagrange multiplier (LM) tests, applied to a nested logit model. The nested logit model is a generalization of the MNL model that allows for more flexible patterns of similarities among alternatives. The Hausman test involves estimating the unknown parameters from both the unrestricted and restricted choice sets. If the parameter estimates are approximately the same, then the MNL specification is not rejected. The test statistic is easy to compute as it only requires the computation of a quadratic form involving the difference of the parameter estimates and the differences of the estimated covariance matrices. The classical tests are applied to the same data as the Hausman test. The results show that the Wald test has significantly greater power than the other two classical tests. The LM test, which is based on inconsistent estimates, is distinctly inferior to the Wald test. The Hausman test, which is based on asymptotically efficient estimates, also performs well. The paper compares the two sets of specification tests for an example. The results show that the Wald test has significantly greater power than the other two classical tests. The LM test, which is based on inconsistent estimates, is distinctly inferior to the Wald test. The Hausman test, which is based on asymptotically efficient estimates, also performs well. The paper concludes that the Hausman test is a more powerful test than the classical tests. The results also show that the classical tests are not necessarily consistent against all members of the family of alternatives defined by a given specification of explanatory variables and any distribution of these variables. The paper also notes that the results are not necessarily consistent against all members of the family of alternatives defined by a given specification of explanatory variables and any distribution of these variables.This paper presents two computationally convenient specification tests for the multinomial logit (MNL) model. The first test is based on the Hausman (1978) specification test procedure, which tests the reverse implication of the independence from irrelevant alternatives (IIA) property. The second set of tests are based on classical test procedures, including the Wald, likelihood ratio (LR), and Lagrange multiplier (LM) tests, applied to a nested logit model. The nested logit model is a generalization of the MNL model that allows for more flexible patterns of similarities among alternatives. The Hausman test involves estimating the unknown parameters from both the unrestricted and restricted choice sets. If the parameter estimates are approximately the same, then the MNL specification is not rejected. The test statistic is easy to compute as it only requires the computation of a quadratic form involving the difference of the parameter estimates and the differences of the estimated covariance matrices. The classical tests are applied to the same data as the Hausman test. The results show that the Wald test has significantly greater power than the other two classical tests. The LM test, which is based on inconsistent estimates, is distinctly inferior to the Wald test. The Hausman test, which is based on asymptotically efficient estimates, also performs well. The paper compares the two sets of specification tests for an example. The results show that the Wald test has significantly greater power than the other two classical tests. The LM test, which is based on inconsistent estimates, is distinctly inferior to the Wald test. The Hausman test, which is based on asymptotically efficient estimates, also performs well. The paper concludes that the Hausman test is a more powerful test than the classical tests. The results also show that the classical tests are not necessarily consistent against all members of the family of alternatives defined by a given specification of explanatory variables and any distribution of these variables. The paper also notes that the results are not necessarily consistent against all members of the family of alternatives defined by a given specification of explanatory variables and any distribution of these variables.