AdaemmerP / lpirfs

40 stars 21 forks source link

Cutoff for longer horizons in the lpirfs package #44

Open nper12 opened 1 month ago

nper12 commented 1 month ago

Grettings Dr. Adämmer,

I have a question regarding the lpirfs package, specifically lp_lin_iv function. I am wondering why do I encounter the following error:

                          task 1 failed - "Mat::submat(): indices out of bounds or incorrectly used"

when trying to estimate responses at higher horizons (say, I have data for 55 periods, but I want the impulse response to show the response for 48 horizons). Why is there a cutoff at 27 horizons in this case (the function only produces results until this horizon). Note that my specification is really rudimentary (without additional lags): results_lp <- lp_lin_iv( endog_data = endog_data, shock = shock, lags_endog_lin = 0, trend = 0, confint = 1.68, hor = 48). Does the function artificially stop solving for really long horizons relative to the data?

Thank you.