Hi,
I am estimating the impact of remittances on labor disincentive effect where both these variables are binary with an instrument z being a continuous variable. I am using the approach discussed in Econometric Analysis of Cross Section and Panel Data by Je¤rey M. Wooldridge section 19.6.2: Endogenous Explanatory Variables as I have self selection and endogeneity issues.
In Step 1) I estimate a probit model with a binary dependent variable being people not working (i.e unemployed) on independent variables remittances x2 x3 - where remittances is 1 "individual receives remittances" and 0 "individual doesn't receives remittances"
in Step 2) I compute the Inverse mills ratio
in Step 3) I estimate a 2sls with binary dependent variable labor disincentive (people not working and not seeking work) with independent variable x2 x3 and instrument z using the ivreg2 command
Commands:
probit unemployed remittances x2 x3 -- where unemployed is 1 "unemployed" 0 "employed" , remittances is 1 "individual receives remittances" and 0 "individual doesn't receives remittances"
ivreg2 labor_disincentive x2 x3 lambda (remittances = z) -- where labor_disincentive is 1"Unemployed and not seeking work" and 0 "Unemployed but seeking work", lambda is the inverse mills ratio and Z instrument is a continuous variable.
Questions:
1) I am wondering if this is correct?
2) Do i need to check for linearity when using this approach and 2sls due to the binary nature of my variables? if so how do i check it?
Many thanks in advance
Salman
I am estimating the impact of remittances on labor disincentive effect where both these variables are binary with an instrument z being a continuous variable. I am using the approach discussed in Econometric Analysis of Cross Section and Panel Data by Je¤rey M. Wooldridge section 19.6.2: Endogenous Explanatory Variables as I have self selection and endogeneity issues.
In Step 1) I estimate a probit model with a binary dependent variable being people not working (i.e unemployed) on independent variables remittances x2 x3 - where remittances is 1 "individual receives remittances" and 0 "individual doesn't receives remittances"
in Step 2) I compute the Inverse mills ratio
in Step 3) I estimate a 2sls with binary dependent variable labor disincentive (people not working and not seeking work) with independent variable x2 x3 and instrument z using the ivreg2 command
Commands:
probit unemployed remittances x2 x3 -- where unemployed is 1 "unemployed" 0 "employed" , remittances is 1 "individual receives remittances" and 0 "individual doesn't receives remittances"
ivreg2 labor_disincentive x2 x3 lambda (remittances = z) -- where labor_disincentive is 1"Unemployed and not seeking work" and 0 "Unemployed but seeking work", lambda is the inverse mills ratio and Z instrument is a continuous variable.
Questions:
1) I am wondering if this is correct?
2) Do i need to check for linearity when using this approach and 2sls due to the binary nature of my variables? if so how do i check it?
Many thanks in advance
Salman
Comment