Forum: help

RE: Inverse Mills Ratio in MicEconAids [ Reply ]
By: Arne Henningsen on 2014-05-07 20:05

Dear Manoj

It is a long time ago that I looked at the issue of censoring in demand systems. I am not sure if you get consistent results if you add the inverse Mills Ratios as shift variables (using argument "shifterNames") in an Almost Ideal Demand System. Please check the extensive literature on censoring in demand systems, e,g,:

Yen, ST (2005): A Multivariate Sample-Selection Model: Estimating Cigarette and Alcohol Demands with Zero Observations, American Journal of Agricultural Economics 87, p. 453-466.

and

http://purl.umn.edu/123946

Best regards,
Arne

Inverse Mills Ratio in MicEconAids [ Reply ]
By: Manoj Rajakaruna Mudalige on 2014-05-06 06:39

[forum:41004]

I am trying to do a demand estimation in AIDS using 'MicEconAids' package in R. I have categorized my demand system into 20 food groups. Still I find the problem of missing consumption in food groups which I am trying to correct by incorporating Inverse Mills Ratio (IMR) in to the demand model. I can use 'systemfit' package and incorporate IMR for each OLS as follows.

olsq1 = dw1~dp1+dp2+dp3+dp4+dp5+dp6+dp7+dp8+dp9+dp10+dp11+dp12+dp13+dp14+dp15+dp16+dp17+dp18+dp19+dp20+dlnx_p+IMR1

But, it is quite hectic business to specify the restrictions when you have 20 food groups. In that sense, MicEconAids is much easier to use. But, I am just wondering if it has provisions for including IMR. Is it correct to use all 20 IMRs with the other shifter variables?

priceNames = c("p1",....,"p20")
shareNames = c("w1",...,"w20")
shifterNames=c("z1",....,"z5")

estResult <- aidsEst( priceNames, shareNames, shifterNames,"x",data = ap, method = "LA" ,priceIndex="T")

If not, Is there any other method to incorporate IMR in "MicEconAids" or is there any other easier way to specify such a large number of restrictions in "systemfit"?

Cheers,
Manoj