Positively correlated residuals often occur when models are estimated using economic time series data. The residuals for

Question:

Positively correlated residuals often occur when models are estimated using economic time series data. The residuals for the consumption function estimated in this chapter turn out to be positively correlated, in that positive prediction errors tend to be followed by positive prediction errors and negative prediction errors tend to be followed by negative prediction errors. This problem can often be solved by working with changes in the values of the variables rather than the levels:

Use the data in Table 10.1 to calculate the changes in C, Y, and W; for example, the value of C − C(−1) in 1961 is 1,821.2 − 1,784.4 = 36.8. image text in transcribed