I am carrying out an empirical study wherein I have to judge productivity for two inter-country alliances. While the translog is a better method theoretically the coefficients are insignificant, whereas in Cobb-Douglas they are significant. The gamma stat also shows a better fit for Cobb-Douglas. Is there any reason as to why Cobb-Douglas can be preferred over translog?

  • $\begingroup$ Could you explain what a "gamma stat" is? With OLS, it is mathematically not possible to obtain a poorer fit (in terms of $R^2$ or SSR) for a more general function than Cobb-Douglas. How do you measure the fit? $\endgroup$
    – Bertrand
    Nov 20, 2021 at 13:14
  • 1
    $\begingroup$ This might help: economics.stackexchange.com/questions/18998/… $\endgroup$
    – Rumi
    Nov 21, 2021 at 3:55


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.