The Phillips curve in economics is a theory due to Bill Phillips, a New Zealand economist, about the trade-off between inflation and unemployment rate. According to this, lower than normal unemployment rate goes hand in hand with higher than normal inflation, and vice versa. The Wooldridge dataset "Phillips" includes data on the US annual inflation (inf) and unemployment (unem) rates from 1948 to 2003. We want to see if this theory is supported by the data or not.  

Consider the regression of inflation on a constant and unemployment rate. Suppose that the errors are generated by an AR(1) process. Amend the model and estimate it. Do the estimates support the theory?