TSLS IV Estimation Exercise

[seer00]

I would like to test the validity of the permanent income hypothesis in a small open economy using panel data on income deciles. I have real consumption and real income data on households divided in income deciles. The deciles are my cross-sections. I use interaction terms in the spirit of Shea 1995 and my specification is as follows:

DLOG(consumption_expenditure_sa) = c(1)+c(2)*DLOG(income_sa)*income_neg_dummy + C(3)*DLOG(income_sa)*income_pos*income_pos_dummy + C(4)*LOG(level_of_interest_rate_household)
_lending

where income_neg_dummy=1 a dummy for periods when DLOG(income_sa)<0 and income_pos_dummy=1 when DLOG(income_sa)>0. level_of_interest_rate_household is interest rate for household consumption lending. All variables are seasonally adjusted.

Now I have reasons to suspect there might be an endogeneity issue with consumption and income and I want to use TSLS to do an IV estimation. I want to use lags of income for instruments on income. Given I have interaction terms is there something different about the way I input instruments in the instruments pane that I have to know to do this right?
The way I do it I simply put in DLOG(income_sa_(-1) to (-6)) and the interest rate as my instruments, but have some reasons to suspect this is not correct. I Do I have to put in the interaction terms the way they are in the specification or just the exact variables I think are the right instruments?

Thanks in advance!