Help with xtabond2 and xtdpdsys

[Anna Per]

Dear Statalist,

I already searched for this topic and found similar threads, but I still have problems with making xtdpdsys command equivalent to xtabond2. The reason why I need xtabond2 is because of the "collapse" function.

This is the xtdpdsys command:

xtdpdsys deposits_ratio hhdebtgdp  nomir_depo  log_incomecap  m1_income y_*, lags(1) maxldep(1) endog(hhdebtgdp  nomir_depo  log_incomecap  m1_income) artests(2)

This is my unsuccessful try of writing a xtabond2 command to yield the same results:

xtabond2 L(0/1).deposits_ratio hhdebtgdp nomir_depo log_incomecap m1_income y_*, gmm(deposits_ratio hhdebtgdp nomir_depo log_incomecap m1_income, lag(2 .) collapse) iv(y_*, eq(diff))

I would be extremely grateful if someone could help me with translating teh xtdpdsys command into the xtabond2 command.

Thank you in advance!

[Sebastian Kripfganz]

Please see the following Statalist topic:
how to make xtdpdsys and xtabond2 equivalent?

More information on GMM estimation of linear dynamic panel data models in Stata:
XTDPDGMM: new Stata command for efficient GMM estimation of linear (dynamic) panel models with nonlinear moment conditions

[Anna Per]

Thank you Sebastian for your reply! I already saw that topic before posting my question but I still haven't found a way to make it equal so that I get the same results. I am aware that it is very much the same as the topic but I am still struggling with it.

[Sebastian Kripfganz]

First of all, notice that in the xtdpdsys syntax (in contrast to xtabond2 or xtdpdgmm), you should not specify endogenous variables in the list of (exogenous) independent variables but only as an option:

xtdpdsys deposits_ratio y_*, lags(1) maxldep(1) endog(hhdebtgdp nomir_depo log_incomecap m1_income) artests(2)

This can be replicated with xtdpdgmm as follows:

xtdpdgmm L(0/1).deposits_ratio hhdebtgdp nomir_depo log_incomecap m1_income y_*, gmmiv(L.deposits_ratio, lagrange(1 1) model(diff)) gmmiv(hhdebtgdp nomir_depo log_incomecap m1_income, l(2 .) model(diff)) iv(y_*, diff model(diff)) gmmiv(L.deposits_ratio, diff lagrange(0 0) model(level)) gmmiv(hhdebtgdp nomir_depo log_incomecap m1_income, diff lagrange(1 1) model(level)) wmatrix(independent)

[Anna Per]

Thank you so much Sebastian!