R: How to stop update.formula from wrapping RHS in extra brackets

Question

I'm writing a function that, among other things, calls pgmm from the PLM package based on an updated formula. However when I use update.formula from stats the right-hand side (RHS) of the formula becomes inexplicably wrapped in parentheses. I wouldn't care except that the formula= argument in pgmm does not accept this syntax.

My baseline formula:

 model.AR1.1X = y ~ lag(y,1) + lag(x,1)  

I use this for some preliminary estimations and then update via:

gmm.form = update.formula(model.AR1.1X, . ~ . | lag(y, 2:6) 
# calling this formula shows:
gmm.form
y ~ (lag(yi, 1) + lag(x, 1) | lag(y, 2:6))

Notice the bracket after the tilde ~, ( lag(y ... etc.. and the double-bracket at the end 2:6))

When I then call pgmm I get the error:

 Error in terms.default(formula) : no terms component nor attribute

There is no problem when I write out the formula myself (without extra brackets), but this would complicate my code.

Is there a way to prevent this behaviour in update.formula? Alternatively, is there a gsub equivalent that can be used on a formula object to remove these brackets?

Answer 1

You can either modify the call or use paste as in the comments. Here's the modify method.

gmm.form[[3]] <- gmm.form[[3]][[2]]
#y ~ lag(y, 1) + lag(x, 1) | lag(y, 2:6)