Overidentifying restrictions test
Summary
- Setup: more instruments than explanatory variables
- Random sample \(\left( y_i,x_i,z_i \right)\) for \(i=1, \ldots ,n\) where \(x_i\) is \(k×1\) and \(z_i\) is \(r×1\) with \(r≥k\)
- A linear regression model, \(y_i=x'_iβ+ε_i\)
- Exogeneity fails, \(E\left( ε_i \mid x_i \right)≠0\)
- z-variables are instruments, \(E\left( ε_i|z_i \right)=0\) and \(E\left( z_ix'_i \right)=Σ_{zx'}\) is \(r×k\) with rank \(k\) .
- \(Var\left( ε_i|z_i \right)=σ^2\) .
- We have \(r\) equations \(E\left( {z_iε}_i \right)=0\) but only \(k\) unknowns. If \(r>k\) then we say that we have overidentification .
- We can test if all instruments are exogenous using an overidentifying restrictions test (also called a Sargan test). The null hypothesis is that all instruments are exogenous.
- To perform the test in Stata, do “estat overid” after estimating the model using ivregress (with strictly more instruments than endogenous variables)