Further packages for microeconometrics

[maebruck]

AER pg. 145

• gam: generalised additive models: logLik posterior assigned, no real "log-likelihood contributions"
• lme4: nonlinear random-effects models: counts, binary dependent variables, etc. - similar to above, seems to be closer to a "real" log-likelihood, but still does not seem to make sense.
• mgcv: generalised additive (mixed) models: as gam
• micEcon: Demand systems, cost, and production functions: probably possible to get log likelihood functions for
• mlogit: Multinomial logit models with choice-specific variables: seems feasible: https://stats.idre.ucla.edu/stata/dae/multinomiallogistic-regression/ http://qed.econ.queensu.ca/ETM/corrections/Fourth-pdf/pg468.pdf
• robustbase: Robust/resistant regression for GLMs: no, we are doing robust regression too.
• sampleSelection: Selection models: generalised tobit, heckit: Try heckit, and maybe generalised tobit after implementation of the AER::tobit