Estimating a spatial autoregressive model with autoregressive disturbances based on the indirect inference principle

This paper proposes a new estimation procedure for the first-order spatial autoregressive (SAR) model, where disturbance term also follows autoregression and its innovations may be heteroscedastic. The is based on principle of indirect inference that matches ordinary least squares estimator two SAR coefficients (one in outcome equation other equation) with approximate analytical expectation. resulting shown to consistent, asymptotically normal robust unknown heteroscedasticity. Monte Carlo experiments are provided show finite-sample performance comparison existing estimators generalized method moments. applied empirical studies teenage pregnancy rates Airbnb accommodation prices.