Essays in dynamic adjustment, structural change and data analysis: applications to the demand for meat in the US

A number of reasons are cited in the literature to explain the over rejection of theoretical demand properties in empirical studies. This dissertation examines two of the most common explanations; the Dynamic Adjustment Hypothesis (DAH) and the Structural Change Hypothesis (SCH), as it is applied to meat demand in the U.S. That is, failure to account for dynamics and variation of parameters over time may be the root cause of the over rejection. Despite the inclusion of dynamics and parameter variation in the model, however, the theoretical demand properties still can not be accepted;The first paper clearly shows that identification of break-points in a time varying consumption model may be spurious with the presence of serial correlation. In the autoregressive specification, a break in the coefficient of the beef model is observed in 1974.4, and 1978.4 for chicken. There is no evidence for a break in the pork model. Usually ignored in earlier studies, the stability of the variance was also tested and was found stable. The break-points identified can therefore be validly attributed to the coefficient vector;The second paper shows that the data support both the dynamic adjustment and structural change hypotheses. However, both in the actual analysis and in a Monte Carlo experiment the structural change appears to be the more dominant influence in the data. Also, the result of the test for structural change is not too sensitive to the specification of the break-points as long as they are not too far away from each other;The assumptions underlying standard demand analysis mostly remain unquestioned in may empirical studies. The third paper examines these untested assumptions as testable propositions to compare how closely they approximate the data, and give indications as to the cost associated with these assumptions;The stationarity assumption is clearly violated. Only the pork quantity and price, and chicken price are stationary. Cointegration tests show that there exist a single long-run equilibrium relationship that governs the co-movements of the demand variables for the three meats over time. This implies that the three meats really belong to a single market, giving justification to model the three as a system. The short-run impulse response displays a demand-like adjustment. Shock in their own-price has permanent effect in beef, but transitory in pork and chicken. The variance decomposition suggests that the assumption of exogeneity of prices and expenditure do not closely correspond to the actual behavior of the data. A strong simultaneity is observed in pork and chicken. It is also shown that supply effects are stronger in pork and chicken than in beef.