Closed mandarpriya closed 10 months ago
Should you be interested in computing time-series regressions per permno you likely want to assign permno as a grouping variable. Without having tested the code, it seems like
risk_premiums <- data_fama_macbeth |> nest(data = c(ret_excess_lead, beta, log_mktcap, bm, month)) |> mutate(estimates = map( data, ~ tidy(lm(ret_excess_lead ~ beta + log_mktcap + bm, data = .x)) )) |> unnest(estimates)
may do the job such that you receive a time-series regression for each permno. However, note that the result is not related to Fama-Macbeth regressions anymore but only yields per firm exposure of future excess returns to the different stock characteristics.
I do not see a direct to-do for the codebase of tidy finance and thus close the issue.
Sir, In this part of code risk_premiums <- data_fama_macbeth |> nest(data = c(ret_excess_lead, beta, log_mktcap, bm, permno)) |> mutate(estimates = map( data, ~ tidy(lm(ret_excess_lead ~ beta + log_mktcap + bm, data = .x)) )) |> unnest(estimates)
i see that the result obtained are general. We dont get the results for each id or permno. So if for each permno we need then what modifications are needed?