CODEWITHSUNDEEP
rplm
safex
safex

Reputation: 2514

R plm lag() ignores panel structure

I want to create lags of a variable. In a panel data setting, obviously lags are only considered within each panel.

How come that plm's lag() does not respect the panel structure (by default) and is there a way to change that (without manually dplyr it)? 

# Load example data
data("EmplUK", package = "plm")
Em <- pdata.frame(EmplUK, index=c('firm', 'year'))

# how I think it should have worked
Em$lwage_incorrect = lag(Em$wage)

# what actually works
Em= Em %>%  group_by(firm) %>%  mutate(lwage_correct = lag(wage))

Upvotes: 2

Views: 2209

Answers (1)

tifu
tifu

Reputation: 1411

When I run your code, I get panel-specific lags using both of your methods, so you might want to check it again. I have gotten into similar trouble before when I wasn't clear what lag function I was actually using (there is one in base R, one in plm, and one in dplyr, for example). Running Em$lwage = plm::lag(Em$wage) removes this ambiguity.

Upvotes: 5

Related Questions