Some Official Data Come With Standard Errors!

Without intending to, I seem to have been on a bit of a rant about data quality and reliability recently! For example, see here, here, and here.

This post is about a related topic that's bugged me for a long time. It's to do with the measures of uncertainty that some statistical agencies (e.g., Statistics Canada) correctly report with some of their survey-based statistics.

A good example of what I have in mind is the Labour Force Survey (LFS) from Statistics Canada.

What (Some of) My Colleagues Are Up To

There's plenty of empirical research going on in the Department of Economics at the University of Victoria, where I work. Readers of the blog get to read plenty about what I've been doing, but what about other empirical work by some of my colleagues?

The following is a small cross-section of some of the quantitative papers that have been produced recently in this department. I've limited myself to very papers that are readily available for downloading, so not all of my empirically oriented colleagues are represented here - sorry!.

Confidence Intervals for Impulse Response Functions

An impulse response function gives the time-path for a variable explained in a VAR model, when one of the variables in the model is "shocked". We get a "picture" of how the variable in question responds to the shock over several periods of time.

An impulse response function (IRF) is essentially a type of conditional forecast. It's a messy function of the estimated coefficients in the VAR model, and the data. So, it's really just a point estimate, period by period. There's some uncertainty associated with the IRF, of course - this comes from the uncertainty associated with the estimated coefficients in the model. So, we really need to report a confidence band, period by period, to go with the IRF.

More on the Quality of Economic Data

Yesterday I posted two pieces relating to the quality of economic data, in general terms, and with reference to China.

I'm firmly of the view that we need to be paying more attention to data quality than we currently do as economists. We also need to keep in mind that data are frequently revised, and this has implications for policy conclusions based on preliminary figures.

To help you with your reading on this topic, here's a small selection of papers that touch on different aspects of this topic:

Now That the Semester is Over....

Another teaching term is done, and the exams are all graded!

HT to my colleague, Emma Hutchinson, for this timely item:

Bias Reduction Paper Published

Another of our papers on bias reduction for Maximum Likelihood estimators has now been published. This one is titled, "On the Bias of the Maximum Likelihood Estimator for the Two-Parameter Lomax Distribution", and is co-authored with Ryan Godwin and Helen Feng. It's in Vol. 42 (11) of Communications in Statistics - Theory and Methods, and is available here.

This paper stems from an ongoing research program with Helen, Ryan, and others. Other posts relating to this program can be found here and here. There's more of this on the way!

© 2013, David E. Giles

The Reliability of China's Economic Data

There have long been concerns about the reliability of published macroeconomic data for China. About 3 months ago, the U.S. - China Economic and Security Review Commission published a timely report, titled, "The Reliability of China's Economic Data - An Analysis of National Output". The report certainly makes interesting reading.

In a post earlier today I warned about the importance of data quality. When the data relate to an economy that's size and importance as that of China, then it's time to sit up and take notice!

Data Quality is Paramount

Yesterday, in a post on the Worthwhile Canadian Initiative, Frances Woolley rightly drew attention to some rather disturbing issues associated with the upcoming release of the 2011 National Household Survey (NHS), by Statistics Canada. In a nutshell, she asks the question, "How can we be sure that the NHS information about the religious beliefs of Canadians is accuarate?"

Recently, I made the comment: Data - the econometrician's lifeblood! Can't function without it." I wish I'd been more specific, and said "reliable data."

The T. D. Dwivedi Memorial Lecture

Yesterday, I was greatly honoured to present the 2nd. invited T. D. Dwivedi Memorial Lecture, in the Department of Mathematics & Statistics at Concordia University, in Montreal. The late Try Dwivedi was instrumental in the establishment of statistics at Concordia, and he also played a leading role in the development of the statistical profession in Canada generally

I'm linked to Dr. Dwivedi through the work that each of us did with V. K. (Viren) Srivastava. So, I have a "Dwivedi number" of 2.  You'll find more about this in an earlier post about Viren here.

The talk that I gave was titled, "Bas Adjustment for Nonlinear Maximum Likelihood Estimators", and you can download my slides from here if you're interested. The material for the lecture was based on a research program that I've been involved with in recent years, jointly with Helen Feng, Ryan Godwin, Jacob Schwartz, and others. A previous post on this blog discussed some of this research.

I'd like to thank the Dwivedi family, Yogen Chaubey (Department Chair), and the faculty of the Department of Mathematics and Statistics at Concordia University, for the kind invitation, their outstanding hospitality, and for a memorable visit to Montreal.

© 2013, David E. Giles

A First Encounter With Monte Carlo Simulation

In my second-year undergraduate course on Statistical Inference for economists, I use Monte Carlo simulation with EViews to illustrate the notion of the "sampling distribution" of a statistic, such as an estimator. This is hardly unusual. However, before we get started I have to persuade the students that this whole Monte Carlo thing might actually work!

So, we go through an exercise where we use simulation to approximate the value of π.

You Can Quote Me on That

The other day I came across the Empirical Quotes page on Mark Byran's blog. Some of his quotes related specifically to econometrics, and I thought I'd share a few others. That certainly doesn't mean that I agree with them all!

In Praise of Quandl!

Data - the econometrician's life-blood! Can't function without it.

So, when a new source of data becomes available - especially one that's sophisticated, reliable, and FREE - it's time to sit up and take notice. Quandl is a recent Canadian start-up that delivers economic and financial time-series data, and then some.

It's an interesting business model. When you go to Quandl, you link to the original sources of the data, you know when the data were last updated, and you get some basic graphical analysis before you download the numbers. 

If you're an R, MATLAB, ........... user, it's a breeze to import the data and get busy with the analysis. I've tried out the associated R package, and it's just great. As you can probably tell - I'm hooked!

I can see myself using Quandl a lot for teaching purposes, as well as for my research, and I suspect that my students will be equally enthusiastic.

© 2013, David E. Giles

Star Wars

Today, Ryan MacDonald, a UVic Economics grad. who works with Statistics Canada, sent me an interesting paper by Abel Brodeur et al.: "Star Wars: The Empirics Strike Back". Who can resist a title like that!

The "stars" that are being referred to in the title are those single, double (triple!) asterisks that authors just love to put against the parameter estimates in their tables of results, to signal statistical significance at the 10%, 5% (1%!) levels. A table without stars is like champagne without bubbles!

Mark Thoma on Empirical Macro

Mark Thoma has a really nice post today on his blog, Economist's View. It's titled, "Empirical Methods and Progress in Macroeconomics".

Students of econometrics, and anyone doing empirical work in (macro)economics, would benefit from reading what Mark has to say about the use of historical data vs. experimental data.

I won't spoil the story by repeating it here, but his bottom line is:

"I used to think that the accumulation of data along with ever improving empirical techniques would eventually allow us to answer important theoretical and policy questions. I haven’t completely lost faith, but it’s hard to be satisfied with our progress to date. It’s even more disappointing to see researchers overlooking these well-known, obvious problems – for example the lack of precision and sensitivity to data errors that come with the reliance on just a few observations – to oversell their results".

© 2013, David E. Giles

Being Unbiased Isn't Everything!

When we first learn about estimation, we encounter various properties that estimators might possess. Unless your first course in statistics or econometrics takes a fully Bayesian stance, then these properties will be ones based on the sampling distribution of the statistic that is being used as the estimator.

There are plenty of unsettling things that can be raised against the notion of the sampling distribution, but let's put those to one side here. In elementary courses, attention usually focuses on just the mean and variance of an estimator's sampling distribution. I'm not endorsing this - it's just a fact of life.

UVic Economics Honours Class

With all of the great work that our Ph.D. and M.A. students are doing, it's easy to overlook an equally important group of students in our department. Each year we have a small group of undergraduate students taking our "Honours" program, and they deserve special mention at this time of year.

This week, with classes over, and final exams underway, the students in the Honours class are making presentations of the research that they've been undertaking over the past few months. Their research projects are always interesting and well executed. Past Honours students have gone on to some of the best doctoral programs in Canada, and have acquitted themselves extremely well.

Here are the presentations given yesterday and today:

And in the Red Corner.........

Here's a paper that I think all students of Econometrics will benefit from reading: "The Widest Cleft in Statistics - How and Why Fisher Opposed Neyman and Pearson", by Francisco Louçã (2008).

National Centre for Econometric Research

Australia has one, but Canada doesn't. What is it?

This Week's Reading

This past week I've been somewhat pre-occupied with the final exams for my undergraduate Economic Statistics course, and graduate Econometrics, courses. However, I've still managed to get some reading done, including the following miscellaneous papers:

Half a Million

Getting close to half a million page-views ........ Thanks!

© 2012; David E. Giles

Seminar on Pre-test Estimation & Testing

Last Friday I gave a seminar in the Department of Mathematics and Statistics, here at UVic. The Statistics seminar series is always very enjoyable, and I really enjoy interacting with this friendly and capable group.

My talk was titled, "The Effects of Prior Hypothesis Testing on the Sampling Properties of Estimators and Tests: An Overview". Preliminary test ( or pre-test) estimation (& testing) was a research topic that I was heavily involved in for about a decade, from the mid 1980's to the mid 1990's. A lot of that work was done with Judith Clarke. I've been looking at some related problems again recently.

If you're interested in this topic, you'll find the slides from my talk here.

© 2013, David E. Giles

Piecing the Puzzle Together

One of the things that we all hope for is that our students will make connections between the material they encounter in one course, and the things they learn about in another course. You shouldn't "forget" what you know about macroeconomics when you go to your econometrics class, etc. In other words, we hope that students will look for the "big picture" as they learn.

Admittedly, that's often easier said than done. When you're sitting in class listening to someone talking about a particular statistical test, it's hard to see how this might tie in with something that your micro. prof. was talking about last week.

Connecting the sots across different subjects in the one discipline is difficult enough, but it's even more difficult to do this across different disciplines while you're still learning the material. 

Today, I gave my last undergraduate "statistical inference" class for the the term. It couldn't have ended on a better note. Here's why.

After the class, one of the students stopped to ask a question - not about today's material, but about a paper he'd been reading in the journal, Economic Modelling. The student had been writing an essay for an English course and had chosen a topic relating to the Greek financial crisis. The paper he'd been reading was an applied econometrics piece, and he realized that this all related to what we'd been doing in our introductory treatment of the linear regression model. His specific question was "what is a VAR model, and how does it relate to a simple regression model?"

A good question, of course, but what was even more rewarding for me was to see him putting the pieces of the puzzle together, by thinking across the different subjects that he's studying. Good job!

© 2013, David E. Giles

Bruce Hansen's Econometrics Textbook

Well-known econometrician, Bruce Hansen (U. Wisconsin) has a free first-year Ph.D. econometrics textbook, titled simply Econometrics, that has been going through various drafts since around 2000.

Here are the chapter titles for the latest draft (January 2013):

1. Introduction
2. Conditional Expectation and Projection
3. The Algebra of Least Squares
4. Least Squares Regression
5. A Introduction to Large Sample Asymptotics
6. Asymptotic Theory for Least Squares
7. Restricted Estimation
8. Hypothesis Testing
9. Regression Extensions
10. The Bootstrap
11. Nonparametric Regression
12. Series Estimation
13. Quantile Regression
14. Generalized Method of Moments
15. Empirical Likelihood
16. Endogeneity
17. Univariate Time Series
18. Multivariante Time Series
19. Limited Dependent Variables
20. Panel Data
21. Nonparametric Density Estimation 

Why not download Bruce's textbook and take a good look at it? I think you'll like what you see.

© 2013, David E. Giles

George E. P. Box

March 28 saw the passing of a great statistician, George E. P. Box. Econometricians know his work well, such as through Box-Jenkins analysis of time-series data, the Box-Cox transformation, and many other major contributions. A brief discussion of George Box can be found here. Bradley Jones' piece, George Box: A Remembrance, says it all.

I've written about George Box previously on this blog. For example, see A Bayesian and Non-Bayesian Marriage, Highly Cited Statistical Papers for Econometricians, and Busking for Business.

© 2013, David E. Giles

How NOT to Plot Your Data

I've long been a staunch advocate of graphing our data in creative and informative ways - preferably before  we subject those data to some fancy statistical interrogation. Indeed, I've made this point previously in this blog - e.g., here.

So, I was pleased to come across Karl Broman's collection of Top Ten Worst Graphs. 

I'm sure that you have your own favourites that you could add to this list. If not, you probably haven't been keeping your eyes open!

© 2013, David E. Giles