Please help with this question.

2. Suppose you want to investigate the effect of union membership on wages. You want to estimate Yit = lUMit + at + \"at where Y5, is the hourly wage of individual 2' in year t and [Ill/1,, is a dummy variable taking the value 1 if individual i2 is a union member in year t and 0 otherwise. You have a balanced panel with T = 2 time consecutive time periods. For the purposes of this question, you may assume that individuals can only change their union membership status at the beginning of each year. (a) Name at least one variable that is effectively being controlled for by the inclusion of individual xed effects. (b) Deseribe what regression you would run to estimate the model using rst differences, and whether this is equivalent to using the within approach. (c) Suppose you estimate the model from (b) using the full sample. You then estimate the model again using only the subset of switchers (individuals who changed their union status from t = 1 to t = 2). Would your answer be numerically equivalent? Explain. ((1) Suppose the number of switchers in the sample was very small, say only 10 observations, but the overall sample size n was very large. Would your panel data approach yield precise estimates of 51