Good morning,
I am creating a model for my thesis that relates a company's share price to its EBITDA. I have the companies in the S&P 500 for my cross section, and a time span of 5 years (N = 503, T = 5). Pooling is not an option, and the Hausman test rejects the null, so I opted for a fixed effects model (within). However, I would like to interact EBITDA with Capex to show that it is not significant for companies with elevated capital expenditure. Since I cannot include time-invariant dummies in a within estimation, what can I do? Perhaps I can use the Mundlak approach and distinguish within and between effects?
Thank you,
Lorenzo Garbarino
I am creating a model for my thesis that relates a company's share price to its EBITDA. I have the companies in the S&P 500 for my cross section, and a time span of 5 years (N = 503, T = 5). Pooling is not an option, and the Hausman test rejects the null, so I opted for a fixed effects model (within). However, I would like to interact EBITDA with Capex to show that it is not significant for companies with elevated capital expenditure. Since I cannot include time-invariant dummies in a within estimation, what can I do? Perhaps I can use the Mundlak approach and distinguish within and between effects?
Thank you,
Lorenzo Garbarino
Comment