Consider the regression model y=a+x+u IN(0, 1) = 1, 2,..., T Suppose that the model refers to semi-annual data, but the data available are either: 1. Annual data, that is 51=31+92 52=y3+34 93ys+y6 and so on with 1, 2, 3,... defined analogously (assuming that T is even), or: 2. Moving average data, that is 31+y2 32+33 y+ 33+34 ++ and s so o on with x, x, x.... defined analogously.

(a) What are the properties of the error term in the regression model with each set of data: (y, x), (5, &), and (y, x)?

(b) How would you estimate in the case of annual data?

(c) How would you estimate 8 in the case of moving average data?

(d) Would you rather have the annual data or moving average data? Why?