Hi,

Thanks Sven. 

I have ended up doing just that. I have convinced the student not to compare the regions and do panel regression.

Thanks again

Alison 

Sent from Outlook for Android

From: Sven Schreiber <sven.schreiber@fu-berlin.de>
Sent: Saturday, August 26, 2023 12:16:03 PM
To: gretl-users@gretlml.univpm.it <gretl-users@gretlml.univpm.it>
Subject: [EXTERNAL] [Gretl-users] Re: Econometric problem
 

*** CAUTION - EXTERNAL EMAIL. Think before replying, opening attachments, or clicking any links. If in doubt, please forward to phishing@northampton.ac.uk or click ‘Report Phishing’ button. ***

Am 23.08.2023 um 14:05 schrieb Alison Loddick:
> I’m wondering if you can help me. Firstly, I am not an > econometrician but a statistician, so I know I have lots to learn. I > generally use Gretl with students to teach them how to do panel > regression and correlation. ... > The student has 17 years of data and wants to compare two variables > (pay and productivity) over 17 regions to understand how the regions > differ over time and whether the two measures have a relationship. > I’m wondering if it is some sort of multivariate time series model. > > > > Can anyone help me help the student?

Your question is quite broad I think, but I will try to answer as gretl-specific as I can. You said you're teaching panel regressions, and that's what I think is the natural approach here. Why not just run a fixed-effects regression between the two variables? For example, using the grunfeld sample dataset in gretl to regress 'value' on 'invest' with the standard fixed effects, gretl reports a LSDV-R2 of 0.96 and a within-R2 of 0.37. This is just to calculate a panel-context correlation between the two variables, not as a model per se.

Of course then you have the usual panel-related questions on whether you really want fixed effects and so on, but that's a different (and not directly gretl-related topic).

I don't see this as a time series model in the usual sense, because with 17 regions and 2 panel variables you have 34 time series, but apparently only 17 periods (years) of data. But if you want to have dynamic effects in your panel regression is of course yet another question. Finally, wanting to understand "how regions differ over time" strikes me as an ambitious question in general, because then you would need some interaction effects or maybe even time-varying parameters. Or maybe you just want to calculate some statistics over rolling windows, that's of course feasible (in principle, not saying there's a built-in function for that).

cheers

sven




University of Northampton: Transforming Lives and Inspiring Change www.northampton.ac.uk This e-mail is private and may be confidential and is for the intended recipient only. If you are not the intended recipient you are strictly prohibited from using, printing, copying, distributing or disseminating this e-mail or any information contained in it. We virus scan all E-mails leaving The University of Northampton but no warranty is given that this E-mail and any attachments are virus free. You should undertake your own virus checking. The right to monitor E-mail communications through our networks is reserved by us.


Disclaimer

The information contained in this communication from the sender is confidential. It is intended solely for use by the recipient and others authorized to receive it. If you are not the recipient, you are hereby notified that any disclosure, copying, distribution or taking action in relation of the contents of this information is strictly prohibited and may be unlawful.

This email has been scanned for viruses and malware, and may have been automatically archived by Mimecast, a leader in email security and cyber resilience. Mimecast integrates email defenses with brand protection, security awareness training, web security, compliance and other essential capabilities. Mimecast helps protect large and small organizations from malicious activity, human error and technology failure; and to lead the movement toward building a more resilient world. To find out more, visit our website.