My Talk at the JSM

Tomorrow morning (Wednesday 6 August) I'll be presenting at the Joint Statistical Meetings in Boston.

The title for my talk is "Modelling Asymmetries in the Market for Gasoline in Western Canada", and it's based on some research that I have underway on this topic. 

The question that's addressed in this work is: "Are the upward and downward movements in retail gasoline prices, that follow increases and decreases in crude oil prices, symmetric?" There's a common perception that the "flow-through" from oil prices to gasoline prices is faster when prices are rising than when they are falling. I use ARDL models with an explicit allowance for possible asymmetry to test this hypothesis.

I'll have more to say on this when the work is completed, but in the meantime you can see some partial results by downloading the slides for my talk here.

© 2014, David E. Giles

The 7 Pillars of Statistical Wisdom

Yesterday, Stephen Stigler presented the (ASA) President's Invited Address to an overflow, and appreciative, audience at the 2014 Joint Statistical Meetings in Boston. The title of his talk was, "The Seven Pillars of Statistical Wisdom".

I'd been looking forward to this presentation by our foremost authority on the history of statistics, and it surpassed my (high) expectations.

The address will be published in JASA at some future date, and I urge you to read it when it appears. In the meantime, here are the "seven pillars" - the supporting pillars of statistical science - with some brief comments:

Estimation & Accuracy After Model Selection

This was the title of Brad Efron's invited paper at the 2014 Joint Statistical Meetings in Boston this morning. It was a great presentation, with excellent discussants - Lan Wang, Lawrence Brown, and Soumendra Lahiri.

The paper and discussion are scheduled to appear in the September 2014 issue of JASA.

A lot of what Brad and his discussants had to say related to one of the main points in one of my recent posts. Namely, if you search for a model specification, then this affects all of your subsequent inferences - and usually in a rather complicated way. Typically, even after searching for a preferred model, we tend to "pretend" that we haven't done this, and that the model's form was known from the outset. Naughty! Naughty!

What Brad has done is to address the "pre-test" issue in a rather nice way. You won't be surprised to learn that bootstrapping features heavily in the methodology that he's developed. Using two examples - one non-parametric, and one parametric - he showed how to take account of model selection via Mallows' C_p statistic, and the lasso (respectively), when constructing regression confidence intervals.

One important feature of his analysis involves "smoothing" the results to take account of the discontinuities that inherent in model selection. Although it wasn't mentioned in Brad's talk, these discontinuities are the source of some of the most important problems associated with pre-testing in general. For example, traditional pre-test estimators of regression coefficients (based, say, on a prior test of linear restrictions on those coefficients) are inadmissible under a range of standard loss functions. This inadmissibility is entirely due to the fact that these pre-test estimators are discontinuous functions of the random sample data.

All in all it was a great session, with some nice take-away quotes:

• "The discussants actually discussed my paper."
• "Simulations are hard to do."
• "Model averaging is perfectly easy to do, but model selection is not."

I took some comfort from the last two of these comments!

© 2014, David E. Giles

Correlation and Causation

A hat-tip to Judea Pearl, whose e-newsletter alerted me to this interesting post on the EvaluationHelp blog. It shows the original sixteen diagrams in Sewall Wright's classic 1921 paper on correlation and causality.

Philip and Sewall Wright were responsible for seminal contributions to the basic notions of instrumental variables estimation and parametric identification, though there is still some debate over their relative contributions to these important concepts.

Reference

Wright, S. (1921). Correlation and causation. Part I Method of path coefficients. Journal of Agricultural Research, 20, 557-585.

© 2014, David E. Giles

Summer Reading List

While I'm at the lake, fishing, this is your big chance to get on with some reading.

There won't be a quiz, but I know that you'll thank me for this later on:

• Aberdie, A., S. Athey, G. W. Imbens, and J. Wooldridge, 2014. Finite population standard errors. Mimeo.
• Boero, G., J. Smith, and K. F. Wallis, 2014. The measurement and characteristics of professional forecasters' uncertainty. Journal of Applied Econometrics, in press.
• Liu, C-A., 2014. Distribution theory of the least squares averaging estimator. Journal of Econometrics, in press.
• Manzan, S., 2014. Forecasting the distribution of economic variables in a data-rich environment. Journal of Business and Economic Statistics, in press.
• Sanderson, E. and F. Windmeijer, 2014. A weak instrument F-test in linear IV models with multiple endogenous variables. Discussion Paper 14/644, Department of Economics, University of Bristol.
• Yang, Z., 2014. A general method for third-order bias and variance corrections on a nonlinear estimator. Journal of Econometrics, in press.

© 2014, David E. Giles

More on Step-(Un)Wise Regression and Pre-Testing

I've been meaning to do a decent post on Pre-test Estimation for some time. It just hasn't happened!

The general issue of pre-testing came up in my recent post on Step-Wise Regression, (I prefer the term, "Step-Unwise Regression"). I want to add a few things to what I said there.

First, a reminder of this post from April 2013:

Promoting Econometrics Through Econometrica

Regular readers of this blog will know that I have an interest (but negligible talent or authority) in the history of econometrics. Actually, the same applies to the history of the discipline of statistics. I find it difficult to appreciate where we are without knowing something about where we came from, and I try to convey this to my students.

Olav Bjerkholt (University of Oslo) has provided me with a lot of very valuable material and insights in recent months, and I've been delighted to have drawn on his contributions in previous posts (e.g., here, here, and here).

Olav wrote to me yesterday, as follows:

Step-wise Regression

Some time ago, Haynes Goddard emailed me suggesting that I post something about step-wise regression. He also put me on an interesting paper by Peter Flom and David Cassell.

Many statistical and econometrics packages include stepwise regression. I wish they didn't! Here's why.

Rejected Economists

If you're feeling "down" as a result of a recent rejection of your work by a nasty journal editor, you may find some comfort in this paper from The Journal of Economic Perspectives in 1994: "How Are the Mighty Fallen: Rejected Classic Articles by Leading Economists", by Joshua Gans and George Shepherd.

In it, you can read about the many famous economists - Nobel laureates included - who have struggled to get their work published. There are some great stories here.

In the process, you'll learn why the Tobit model is so-named - and it's not just for the obvious reason that comes to mind!

I promise that you'll feel much better after reading what Gans and Shepherd have to say.

.

© 2014, David E. Giles

Price Indices Based on Scanner Data

Among the many interesting presentations that I attended at the recent Annual Conference of the N.Z. Association of Economists was one by Frances Krsinich, from Statistics New Zealand.

The paper that Frances gave was titled "Price Indexes From Online Data Using the Fixed Effects Window Splice (FEWS) Method". Here's the abstract: