Consider this model for the health of an individual:

healthi = Alpha + Beta1(age)i + Beta2(weight)i + Beta3(height)i + Beta4(work)i + Beta5(exercise)i + Epsiloni

where health is a measure of health, age, weight, and height are self explanatory, work is weekly hours worked, and exercise is the hours of exercise per week. We are interested in examining the effect of exercise on health, therefore the rest of the variables are "control" variables.

(a) Why might you be concerned about exercise being correlated with the error term ui? Tell a story to show that the OLS estimator may be invalid.

(b) What other variable(s) might be missing in the above model? Since the "missing variable" is excluded from the model, how does that affect the above estimated coefficients?