Willingness to pay estimates from lclogit2 when price is interacted with another variable

[hankyul yoo]

Hello,

My question concerns gettting standard errors of willingness to pay (WTP) estimates, in a latent class analysis, when price is included in an interaction with another variable.

I am using the lclogit2 command to conduct a latent class analysis of my choice experiment data.
My assumed utility function has interaction terms, and one involves the price.
So the formula to get the willingness to pay for an attribute is not simply the ratio of coefficients, but a slightly adjusted formula.
As lclogit2 does not allow interactions in the code, I multiplied the variables and coded them in lclogit as normal variables (not interactions like var1##var2)
This is the code:

lclogit2 decision, group(resp_task) nclasses(2) membership(educ sig_status uniqueness env_belief energy_id hp_int age) rand(vis des_bas des_top des_none att3 att4 att5 att6 visxbas visxtop visxadp visxems visxprc) seed(1234)

As interactions are not allowed in lclogit2, I tried using nlcom instead of lclogitwtp2, but I don't know how to get the WTP estimates for each of the latent classes. Althougth nlcom was conducted after a latent class estimation, which assumes multiple classes, it only provides one set of estimates.

I am able to calcuate the WTP manually, but would like to avoid doing the delta method by hand, if possible.

Does anyone have experience getting standard errors for WTP values for each latent class, when the price is interacted with other variables?

Thanks in advance for any insight you might have!