RECOVERING OUTPUT-SPECIFIC INPUTS FROM AGGREGATE INPUT DATA - A GENERALIZED CROSS-ENTROPY APPROACH

For multiproduct firms, data on aggregate input usage are typically av ailable but data on activity-specific inputs are not. The present stud y proposes a generalized cross-entropy approach to estimate activity-s pecific input allocations that are consistent with the aggregate infor mation. The proposed method does not require behavioral assumptions (e .g., profit maximization) but does accommodate behavioral restrictions as well as nonsample information about the plausible factor shares ac ross enterprises. Monte Carlo experiments using simulated data for mul tifactor-multiproduct firms are used to evaluate the performance of th e proposed method.