3
$\begingroup$

I have a regression model of unemployment on vacancy over 20 years. The model also includes a time trend

time trend = 0, 1, 2 , 3 ,....

What's the reason for including a time trend?

Thank you

$\endgroup$

2 Answers 2

3
$\begingroup$

If both unemployment and vacancy have a trend in the long-run, then regressing one against the other would be very misleading. (Every time series with strong positive trend could be well regressed with practically anything that also has a strong positive trend!). That is a classical example of spurious regression; more generally speaking you should be cautious to think over whether time series in the regression are stationary.

Including time trend is essentially a form of de-trending. (If you include linear time trend, it means fitting and subtracting a linear trend.) Makes sense if the time series become stationary by de-trending.

Chapter 10 of the Wooldridge book discusses the basics of this question. Actually, Section 10.5 is dedicated to your very question.

$\endgroup$
2
$\begingroup$

Time is important in terms of inflation etc. Therefore, when dealing with economic situations (also many other situations) its effect should be controlled for. Good responses to your question are available from Time trend or time dummies in a panel

$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.