Homework 1

You need to type and email me your answers.

  1. (Due 1/22) What happens if we relax the assumption in the equation (1) on page 278 in Arellano and Bond (1991)? Suppose $\alpha =1$ (or close to one) and this is a random walk with an individual specific drift given by
    MATH
    Why do we care this issue? How is $\alpha =1$ related to the issues of weak instruments?

  2. (Due 1/29) How would you test the null of $\alpha =1?$ Note in this literature the time dimension $T$ is small or fixed. This implies that large T asymptotics often used in panel time series econometrics or panel-time econometrics cannot be used, though we do assume the cross-sectional dimension, $n$, is large.

This document created by Scientific WorkPlace 4.1.