Hello everyone,

My regression looks in the following way:

yit=a0 + a1xit + a2Dit +a3Lit + a4mit + u

y - is the native employment rate in the region i and year t. (nativeemp)

x - the migrant/native ration in region i and year t. (mig_nat)

D - is a vector of variables for the demographics of individuals in region i and year t. These include the ratio of individuals aged 25-49 against those aged 16-25 (mid_young), 50-64/16-25 ratio (old_young)

L - a dummy variable for year t

m - a dummy variable for region i

In stata, I am running the following code:

reg nativeemp mig_nat mid_yound old_young i.year i.region, cluster(region)

I am also running the first difference:

reg D.(nativeemp mig_nat mid_yound old_young) i.year i.region, cluster(region)

Finally, I want to combine estimation in differences with the use of instrumental variables calculated by GMM imposing the moment restrictions that d.u (error term) are uncorrelated with the chosen instrument, which is in this case 3 and 4-period lags of the endogenous regressor (nativeemp). I have denoted 3-4 period lags with lmignat3 lmignat4.

I would really appreciate it if you could tell me the code which applies the Arellano-Bond method to my regression.

Thank you for your help