A new approach to calculating welfare measures in Kuhn-Tucker demand models

I develop a new approach to calculating welfare measures in Kuhn-Tucker consumer demand models that uses the analytical properties of the Multiple Discrete-Continuous Extreme Value (MDCEV) specification. I adapt Pinjari and Bhat's (2011) Marshallian demand forecasting routine to calculate Hicksian demands that are useful for computing welfare measures. Simulations demonstrate that this new approach substantially reduces computational time relative to the existing approach using a numerical bisection routine. The new approach performs best relative to the numerical bisection routine if i) a γ-profile utility function is specified, ii) the number of choice alternatives available is high, or iii) the average number of chosen alternatives is low.