Consumer Demand—Empirical Analysis II

Abstract

This chapter focuses on empirical application of the system of demand functions approach using the Rotterdam model (RM) and Almost Ideal Demand System (AIDS). Attention is given to the nature of the data sets typically available for use in time series analysis. I discuss issues involved in implementation and interpretation using data sets created for estimation of US demand for meats. In the process, an evaluation is made of the relative performance of the RM and AIDS models. Both conditional and unconditional demand elasticities are estimated using results from two-stage budgeting in Chapter 3. Extensions of the analysis to imposing negative definiteness are indicated. Other data sets besides USDA disappearance are discussed.

Problems

• 5.1 Given the elasticities and sample mean shares for the compensated conditional demand elasticities in Table 5.5, calculate the matrix of uncompensated conditional demand elasticities. Why are the uncompensated conditional elasticities so different from the compensated conditional elasticities, yet the uncompensated unconditional elasticities are very similar to the compensated unconditional elasticities?

• 5.2 Verify that the eigenvalues of the matrix $\left[ {c_{ij} } \right]$ from Table 5.5 are all non-positive.

• 5.3 Derive the compensated cross-price elasticities for the four meat goods with respect to non-meat food and non-food goods using the parameter estimates from Table 5.12. Sample mean estimates of the relevant budget shares are shown in Table 5.14. The relevant formulas can be found in Chapter 3.

• 5.4 Explain why the money flexibility coefficient $\phi$ must be estimated jointly with the other parameters of the first-stage equations.

• 5.5 Use the data in Tables 5.1, 5.2 and 5.3 to estimate the Indirect Translog demand system. Compare the compensated price and expenditure elasticities with those obtained from the RM and AIDS in this chapter.