Can beta be an exponent within a function prior to taking the log of that function to arrive at a simple, log-linear estimation equation?

I am curious because I found an equation that had beta as an exponent on page 7 of this article by Lee Branstetter(2004): http://bit.ly/1OGSFx9 titled: Is Foreign Direct Investment a Channel of Knowledge Spillovers? Evidence from Japan’s FDI in the United States

I believe there this is not out-of-order. I am however, seeking an explanation to the equation.

I have tried to check different sources and the closest explanation is that the function is a linear-log function.


1 Answer 1


This is not a structural paper. The authors are not "deriving" a log-linear relationship from a model, which they then proceed to estimate. They rather start by postulat the relationship:

Let $C^J_{it}$ be the number of citations made by the patent applications Japanese firm i filed in year t to the cumulated stock of bindigenousQ U.S.-invented patents granted as of year $t$. I can then write the log of $C^J_{it}$ as a simple log-linear function of several other observables

$$ c^J_{it} = \beta_0 + \beta_1 p_{it} + \beta_2 FDI_{it} + \gamma_i + > \alpha_t + \epsilon_{it}$$

which is their equation (1).

  • $\begingroup$ Thanks @Foobar but: (1) how is postulating different from estimating a model and; (2) under what circumstances would each be more suitable. I ask because I believe both requires making assumptions and testing those assumptions with an equation. Thanks. $\endgroup$
    – od320
    Sep 11, 2015 at 19:05

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.