Fixed-Effects Vector Decomposition

Warning! Avoid the so-called "Fixed-Effects Vector Decomposition" (FEVD) estimator, introduced by Plümper and Troeger in a 2007 issue of Political Analysis.

A recent "Symposium on Fixed-Effects Vector Decomposition" in the 2011 volume of that journal, which included critiques by William Greene and by Trevor Breusch et al.,  reveals just what this estimator is.... and isn't!

Integrated & Cointegrated Data

Last week I had a post titled More About Spurious Regressions. Implicitly, in that post, I assumed that readers would be familiar with terms such as "integrated data", "cointegration", "differencing", and "error correction model".

It tuns out that my assumption was wrong, as was apparent from the comment/request left  on that post by one of my favourite readers (Anonymous), who wrote:

"The headlined subject of this post is of great interest to me -- a non-specialist. But this communication suffers greatly from the absence of a single real-world example of, e.g. "integrated" or "co-integrated" data, "differencing" (?), "error-correction model," etc. etc. 

I'm not trying to be querulous. It's just that not all your interested readers are specialists. And the extra intellectual effort required to provide examples would help us..."

Monte Carlo Experiments With gretl

I keep saying that I must make more use of the gretl econometrics package. It's great software, and it's free! So, shame on me for not putting my effort where my mouth is.

Fortunately, Riccardo (Jack) Lucchetti keeps a bit of an eye on me in this regard!

Panel Unit Root Tests

Testing for unit roots in panel data is pretty standard stuff these days. Any decent econometrics package has everything set up to make life easy for practitioners who want to apply such tests. As usual, though, ease of application doesn't guarantee correct application of the tests, or the interpretation of the associated results.

Yet Another Reason for Avoiding the Linear Probability Model

Oh dear, here we go again. Hard on the heels of this post, as well an earlier one here, I'm moved to share even more misgivings about the Linear Probability Model (LPM). That's just a fancy name for a linear regression model in which the dependent variable is a binary dummy variable, and the coefficients are estimated by OLS.

Another Gripe About the Linear Probability Model

NOTE: This post was revised significantly on 15 February, 2019, as a result of correcting an error in my original EViews code. The code file and the Eviews workfile that are available elsewhere on separate pages of this blog were also revised. I would like to thank Frederico Belotti for drawing my attention to the earlier coding error.

So you're still thinking of using a Linear Probability Model (LPM) - also known in the business as good old OLS - to estimate a binary dependent variable model?

Well, I'm stunned!

Yes, yes, I've heard all of the "justifications" (excuses) for using the LPM, as opposed to using a Logit or Probit model. Here are a few of them:

Econometrics Beat on Twitter

Last week I finally caved in and joined Twitter!

My intention is to use it in tandem with this blog, but we'll see how that works out. I might get totally carried away.

I'll be tweeting @DEAGiles.

© 2012, David E. Giles

More About Spurious Regressions

Students of econometrics are familiar with the "spurious regression" problem that can arise with (non-stationary) time-series data. 

As was pointed out by Granger and Newbold (1974), the “levels” of many economic time-series are integrated (or nearly so), and if these data are used in a regression model then a high value for the coefficient of determination (R²) is likely to arise, even when the series are actually independent of each other. 

They also demonstrated that the associated regression residuals are likely to be positively autocorrelated, resulting a very low value for the Durbin-Watson (DW) statistic. There was a time when we tended to describe a “spurious regression” as one in which R² > DW. 

"Disappearing" Historical Data

It's the bane of my life - and probably that of every other economist who (tries to) work(s) with time-series data. Historical data that simply disappear.

Shazam! Now you see, it; now you don't!

You know the sort of thing I mean, I'm sure. Just when you think you have a nice long, consistent, series of economic data, the statistical agency in question suddenly stops recording it. They change the basis on which the data are gathered (usually for perfectly good reasons), and leave you with a big red DISCONTINUED descriptor.

Gee - thanks a lot!

Sometimes they even "pull" the historical series that you were really counting on, and just provide you with a pretty new series that started yesterday, and won't be of any use to anyone for ages.

Forecasting: Principles and Practice

Forecasting: Principles and Practice is the title of a new book by Rob Hyndman and George Athanasopoulos.

As Rob says on his webpage:

"The book is dif­fer­ent from other fore­cast­ing text­books in sev­eral ways.

• It is free and online, mak­ing it acces­si­ble to a wide audience.
• It is based around the fore­cast pack­age for R.
• It is con­tin­u­ously updated. You don’t have to wait until the next edi­tion for errors to be removed or new meth­ods to be dis­cussed. We will update the book frequently.
• There are dozens of real data exam­ples taken from our own con­sult­ing prac­tice. We have worked with hun­dreds of busi­nesses and orga­ni­za­tions help­ing them with fore­cast­ing issues, and this expe­ri­ence has con­tributed directly to many of the exam­ples given here, as well as guid­ing our gen­eral phi­los­o­phy of forecasting.
• We empha­sise graph­i­cal meth­ods more than most fore­cast­ers. We use graphs to explore the data, analyse the valid­ity of the mod­els fit­ted and present the fore­cast­ing results."

This looks really good!

© 2012, David E. Giles