Calculating marginal effects in a nested logit model (dicrete choice)

[Laura Meier]

Hey Guys,

i have a Problem: I would like to predict the heat market with a discrete choice model, combined with a microeconomic model until 2030.

I have a dataset where the relations between the heat technology, the costs, the building type and so on are viewed. So its easy to caclulate marginal effects.

But i would like to add some qualitative data (motivational factors influencing the choice, like environment). Some authors of journals made surveys about motivational factors influencing the choice and assessed them with values from 1-5. ( 5 important ; 1 less important)

My problem: They just realesed the mean and the std. for the technologys. For example homeowners who bought a gas-heating have a preference for environmental protection of 2,3 and a std. of 1.2. homeowners with a heatpump 4(mean) and 1 (std) and so on. Is it possible to calculate useful marginal effects without the total dataset and how do i integrate the mean and std in the dataset?


The formula is:
Unj = Vnj + enj
Vnj = aj + bnj cnj

U= Utility
V= Observable Utility
a = preference parameter of any technology ( motivational factor)
c = costs
b = cost parameter - determines the influence of perceived cost differences

Thanks in advance!

[Dyah Pritadrajati]

Hi everyone,

I am experiencing the same problem, did anyone figure out how to get the marginal effects (AME and MEAM) after nlogit command in Stata. Maybe someone can provide an alternative command "by hand" to compute marginal effect after nlogit. I would really appreciate any help.

Thank you,
Dyah

[Nieves Valdes]

Hi everyone! Same question here...someone knows how to compute marginal effects after nlogit? I've read about the nlogit software by Greene as an alternative solution. Someone has used Greene's software? Thanks in advance! Nieves