Consider the first differences of quarterly log US industrial production, over the period 1963.1–2002.4. Create 4 quarterly dummies D s , t and estimate the model y t = α + μ 1 D 1 , t + μ 2 D 2 , t + μ 3 D 3 , t + μ 4 D 4 , t + ε t . What happens? Why? What happens if α is restricted to 0? What happens if μ 1 is restricted to 0? How can the parameters be interpreted in these two cases?