I have the following model

Expenditures = a + $b_1$ Output + $b_2$ Labor + $b_3$ CapitalStock + $b_4$ D

where D is dummy variable that is Categorical. If tax is imposed, it takes 1. Or if tax is not imposed, then it takes 0.

Which regression model should I apply to estimate the equation? By the way, I have a panel dataset.

Please share your suggestions with me. I will search for them. Thank you.


2 Answers 2


I did a quick search of other questions similar to yours and I found this article and this one on CrossValidated. It may not be a direct answer but it might help lead you in the right direction and you might learn some interesting things about panel regression in the process :)

Good Luck!

  • $\begingroup$ Thank you so much :) $\endgroup$
    – studentp
    Feb 1, 2022 at 19:31
  • 1
    $\begingroup$ Currently this is link only answer. You should expand the answer to include explanation otherwise if the link breaks the answer will become useless. Also link only answers are against our rules $\endgroup$
    – 1muflon1
    Feb 1, 2022 at 20:13
  • $\begingroup$ Apologies for not expanding. I also didnt know that the links were against the rules. $\endgroup$ Feb 1, 2022 at 20:17
  • 2
    $\begingroup$ @KeithRoberts no links are fine but only having link saying go look at the link is not fine. You can keep the links but answer has to answer the question not just say you get answer once you click on the link. You should still keep the links and use citation formatting ‘>’ to avoid plagiarism if you want to copy something from the other answers or write an original answer referencing those old ones but just pasting link and saying answer is there is not fine. Good rule of thumb is to see if your answer contains sufficient information to answer Q without leaving this post $\endgroup$
    – 1muflon1
    Feb 2, 2022 at 0:23

I'm not exactly sure what your data is, but you probably will have some collinearity within your independent variables. Your output is typically determined by the combination of labor and capital within a production function, so your coefficients on output, labor, and capital will be underestimated. Whenever modelling your data, I would suggest to start by doing a DAG (Directed Acyclic Graph) because it allows you to think through and visualize how each variable might have a causal effect on another. Take a look at this great primer on DAG: [https://mixtape.scunning.com/dag.html][1]

Also, whenever you have a panel data set, you should utilize one of the most important benefits of panel data, and that is controlling for unobservable factors between each of your cross sectional groups. This is simply done with a categorical dummy for each group.

Honestly, without other information about the panel data set you are using, and what exactly each one of your variables represents, it is difficult to give any specific insights. It appears your model is very similar to a production function, but typically the dependent variable is output and not expenditures.


Your Answer

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

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