Testing Up, or Testing Down?

Students are told that if you're going to go in for sequential testing, when determining the specification of a model, then the sequence that you follow should be "from the general to the specific". That is, you should start off with a "large" model, and then simplify it - not vice versa.

At least, I hope this is what they're told!

But are they told why they should "test down", rather than "test up"? Judging by some of the things I read and hear, I think the answer to the last question is "no"!

The "general-to-specific" modelling strategy is usually attributed to David Hendry, and an accessible overview of the associated literature is provided by Campos et al. (2005).

Let's take a look at just one aspect of this important topic. 

Rob Hyndman on Forecasting

If you have an interest in forecasting, especially economic forecasting, the Rob Hyndman's name will be familiar to you. Hailing from my old stamping ground - Monash University - Rob is one of the world's top forecasting experts. 

Without going into all of the details, Rob is very widely published, and also has a great blog, Hyndsight. He's author of the well-known  "forecast" package for R (version 5 just released); and the co-author of several important books.

Last year, Rob taught an on-line forecasting course, titled, "Time Series Forecasting Using R". It comprised 12 one-hour lectures, on the following topics (with exercises):

• Introduction to forecasting 
• The forecaster's toolbox 
• Autocorrelation and seasonality 
• White noise and time series decomposition 
• Exponential smoothing methods 
• ETS models 
• Transformations and adjustments 
• Stationarity and differencing 
• Non-seasonal ARIMA models 
• Seasonal ARIMA models 
• Dynamic regression 
• Advanced methods

The really good news? You can access these presentations right here!

© 2014, David E. Giles

An ARDL Add-in for EViews

My posts on ARDL models and bounds testing (here and here) have certainly been popular. So, I was really pleased to see that Yashar Tarverdi has produced an "Add-In" for the EViews package that makes this type of econometric analysis somewhat easier.

You can download the the add-in program and its installer here. The add-in is called "ARDLbound", and it largely automates the key steps associated with bounds testing using an ARDL model.

Jim Hamilton on R-Squared and Economic Prediction

I always tell my students that, when it comes to regression results, the value of the coefficient of determination (R²), is pretty much the last thing that I look at. And I'm serious! I've blogged about this before (see here, for example), but it's worth reiterating, and I was reminded of this when I saw Jim Hamilton's post on this topic today.

Read it, and enjoy!

© 2014, David E. Giles

Six Word Peer Review

A "Six Word Peer Review" competition has been running on Twitter (#sixwordpeereview).

Here are a few gems that might be a little too close to home for comfort:

• You didn't cite my paper: Reject!
• Taking my time. Love, your competitor.
• Bayes would turn in his grave.
• Sorry for the huge delay. Reject!
• Author made all required revisions. Reject!
• Your conclusions contradict your actual results.
• Has author considered another direction entirely?

Not among the tweets, but the punch-line to a report I was handling as a member of the editorial board for Journal of Econometrics some years ago:

                        "This dog should be put down".

It's true - I swear!

© 2014, David E. Giles

Thanks a Milllion!

So,.......... by reading this post you'll assist in pushing the total number of page-views for this blog, since its inception in 2011, above the 1 Million mark. Thanks for your interest, support, and questions.

It's been a blast!

© 2014, David E. Giles

An Interesting New Book

Here's a new book that looks as if it will be interesting, and I'm looking forward to reading it myself: Panel Data Analysis Using Eviews, writen by I Gusti Ngurah Agung. Two other related books by this author have been published perviously - see here.

I'll give my opinion in more detail at a later date.

© 2014, David E. Giles

Estimating the Generalized Pareto Distribution

The generalized Pareto distribution (GPD) arises in the modelling of "extremes", especially if the "peaks-over-threshold" approach is being used. Estimating the parameters of the GPD by the method of maximum likelihood is especially challenging. The challenges arise because the likelihood function doesn't satisfy the usual regularity conditions for all possible values of the parameters.

I've discussed some of these issue in earlier posts, here and here.

When my colleagues, Helen Feng and Ryan Godwin, and I started looking at analytic bias reduction techniques for maximum likelihood estimators that can't be expressed in closed form, we first tackled the case of the GPD. It was well motivated, because you usually start with a very large sample size, the number of extreme data-points that lie above a given threshold, and which form the sample for estimation purposes, is generally very small. So, small-sample bias is a real issue.

Well, we bit off a lot more than we realized at the time, and bias reduction when estimating the GPD's parameters turned into a bit of a nightmare! We published several papers (including ones with Jacob Schwartz) dealing with bias reduction for other distributions, but the GPD problem was always there in the background.

I'm pleased to be able to report that our paper on this problem is now accepted for publication in Communications in Statistics - Theory & Methods. You can access a pre-print here.

© 2014, David E. Giles

Thanks for the Honour!

Recently, I became an Honorary Professor in the Department of Economics in the Management School at the University of Waikato, in New Zealand. The appointment took effect on 1 January.

The economics group at Waikato is very energetic and productive, so I'm really grateful for this honour and for the opportunity to interact with them .

I'm still on faculty, full-time, at the the University of Victoria in Canada. However, I'm expecting that my new link will enable me to play a more active role in the profession in New Zealand (where I studied and lived for many years) than has been possible in recent times.

So, a big "thank you" to those who made this wonderful opportunity possible!

© 2014, David E. Giles

Reading for the New Year

Back to work, and back to reading:

• Basturk, N., C. Cakmakli, S. P. Ceyhan, and H. K. van Dijk, 2013. Historical developments in Bayesian econometrics after Cowles Foundation monographs 10,14. Discussion Paper 13-191/III, Tinbergen Institute.
• Bedrick, E. J., 2013. Two useful reformulations of the hazard ratio. American Statistician, in press.
• Nawata, K. and M. McAleer, 2013. The maximum number of parameters for the Hausman test when the estimators are from different sets of equations.  Discussion Paper 13-197/III, Tinbergen Institute.
• Shahbaz, M, S. Nasreen, C. H. Ling, and R. Sbia, 2013. Causality between trade openness and energy consumption: What causes what  high, middle and low income countries. MPRA Paper No. 50832. 
• Tibshirani, R., 2011. Regression shrinkage and selection via the lasso: A retrospective. Journal of the Royal Statistical Society, B, 73, 273-282.
• Zamani, H. and N. Ismail, 2014. Functional form for the zero-inflated generalized Poisson regression model. Communications in Statistics - Theory and Methods, in press.

© 2014, David E. Giles