英文摘要 |
The simultaneity problem of independent variables in a production function is caused by the correlation of the error term that contains unobserved productivities with input demands, which then generates an endogenous problem. When estimating a production function, Olley and Pakes (1996) suggest jointly solving two problems - i.e., selection and simultaneity - to obtain consistent parameter estimates. Our research ignores the selection problem, because there are few tourist hotels leaving the market during the sample period, and merely takes the simultaneity problem into account. We thus propose to jointly estimate two production frontiers for Taiwan's tourist hotels, corresponding to accommodation and restaurant divisions, using the copula methods. After considering the simultaneity problem, we find that the average technical efficiency of both divisions increases, the management ability of the restaurant division exceeds that of the accommodation division, chained hotels outperform independent hotels, the average efficiency of the sample hotels before and after 2008 (the subprime crisis) remains the same, and the data do not support the quiet life hypothesis. |