Omitted Variable Tests and Dynamic Specification

1. Introduction.- 2. The t-statistic under dynamic misspecification.- 2.1 The model.- 2.2 Properties of the estimators ?.- 2.3 The distribution of the quasi t-statistic.- 2.4 Invariance results.- 2.5 Monte Carlo experimentation.- 3. Consumer theory and the Rotterdam model.- 3.1 Commodity space and budget set.- 3.2 Preferences, direct utility function and Marshallian demand.- 3.3 Cost function and Hicksian demand.- 3.4 The Rotterdam model.- 3.5 The Rotterdam model in matrix form.- 4. Robust estimation.- 4.1 Quasi-maximum likelihood estimation.- 4.2 Estimation of the covariance matrix of the quasi-maximum likelihood estimator.- 5. Testing for homogeneity.- 5.1 Anderson's U test if the errors are time-independent (LRU).- 5.2 Functional equivalence between the LRU and Laitinen's statistic.- 5.3 Likelihood ratio test if the errors are VAR(p).- 5.4 The distribution of the LRU statistic under dynamic misspecification.- 5.5 The robust Wald test.- 5.6 Summary.- 6. Monte Carlo experimentation.- 6.1 Data.- 6.2 The data-generating process.- 6.3 Experiments.- 6.4 Simulation results.- 7. Conclusions.- A. Proof of proposition 5.4.- B. Data.- C. Values of the population parameters.- D. Algorithm for the MA(1) parameters.- List of Figures.- List of Tables.