Stationarity in dated data sets.

[Swarles_Barkley]

My friend and I have this assignment for EViews and we just cannot solve the last question. Just to paint the picture, I will quote that question here:

In sheet3 you find annual data of consumption and disposable income of a particular OECD-country. Estimate a reliable consumption function expressing the relationship between consumption and disposable income. Calculate the marginal propensity to consume. If necessary, distinguish between the short-run and the long-run.

The problem is that these data are non-stationary (we found this using an augmented Dickey-Fuller test) so we can't make reliable predictions with them.

So, the question that's been killing us for the past couple of days: how do you make non-stationary data stationary? If it matters, we use EViews 6.


Thanks

[trubador]

The answer is really simple and you should better figure it out on your own in order to make the most out of your assignment. You can either refer to standard econometrics textbooks or do a simple search in the forum, since there have been numerous discussions and posts on this subject.

[hum2004]

or you simply take differences. But i would add, that you make a mistake if you estimate in differences (omitted variable). you should rather test for cointegration