生產(chǎn)率與效率分析lecture8-隨機(jī)前沿效率1-資料下載頁
2025-01-07 12:34本頁面
【正文】 CALC 。list 。1chi(,1)$ Result = .99720998294368760D01 Area in rejection region = 37 Basic Model ?TEi terms ui . halfnormal or exponential ?CobbDouglas form ?production frontier ?crosssectional data 38 Extensions of Basic Model ?1. ui . truncatednormal ?halfnormal exponential arbitrary choices ?both have mode at zero ?implies inefficiency effects near zero ?implies high TE ?truncation at zero of N(?, ?2) ?depending on sign and magnitude of ? ?avoids problem of halfnormal exponential 39 Ui ? halfN has Mode at Zero Ui having mode at zero means highest probability at ui = 0. Highest probability result is estimating low/no inefficiency. This result is purely a function of the distributional assumption about U. ui f(u) 0 40 Extensions (cont.) ?2. other functional forms ?linear ?loglinear ?translog ?ZellnerRevankar generalized function ?etc. 41 Extensions (cont.) ?3. costminimizing cases ?cost frontiers ?u ? 0 ?LIMDEP can estimate this model ?Default case in LIMDEP ?production = maximizing behavior ?Command for cost function = minimizing behavior ? 。cost 42 Extensions (cont.) ?4. panel data models ?random effects model ?LIMDEP can estimate this model 43 Extensions (cont.) ?5. applications beyond production cost ?earnings ?Hofler Polacheck (1982) ?reservation wages ?minimum unemployment rates ?extent of labor market effort ?Hofler Sen (2023) ?Can you think of a situation in which a stochastic frontier would be appropriate? 44 Summary ? yi = xi ? + vi ui ? ui ? 0 maximizing behavior ? ui ? 0 minimizing behavior ? u measures efficiency ? Can estimate ui for all i ? ui . truncatednormal preferred 45 Summary (cont.) ? Data types ?single cross section or time series ?panel ? Contexts ?production ?cost ?others 46 Summary (cont.) ? Two key questions ?Are the observational units either maximizing or minimizing? ?Is the variable being studied bounded either above or below? 47
教學(xué)課件相關(guān)推薦
鄂ICP備17016276號-1