Book recommendations on empirical methods in economic research and econometrics?

Question

This question is more so geared toward anyone with research experience within economics, though of course, anyone is welcome to respond.

I'm an undergraduate math major with a minor in economics. I'd like to know which textbooks are decent for introducing one to research methods in economics? Graduate level is fine, perhaps even preferred. If a student was to take his research to "the next level" (publishing worthy), which textbooks would provide the rigor and insight to develop said research?

Also, any introductory recommendations for econometrics?

Thank you in advance.

ANNENDUM: From the limit experience I have in economics, I've found the following areas to fascinate me on multiple occasions (in no particular order):

1. Game theory
2. Macroeconomics (monetary policy)
3. Microeconomics (i.e., commodity pricing)
4. Pricing theory

Answer 1

Firstly, for basic introductory econometrics, the following is quite good:

"Introduction to Econometrics" by Stock and Watson.

It is light on technical details, and heavy on intuition. This might not appeal to you if you are a math major, but it's the most important thing to get straight if you are interested in applying the empirical methods rather than developing new econometric methods.

Stock and Watson is very basic, so you will probably want to supplement it. Good choices are likely to be

"Mostly Harmless Econometrics" by Angrist and Pischke

(this is another book that puts intuition first, but now with more formal detail), and for an advanced graduate treatment

"Econometric Analysis of Cross Section and Panel Data" by Wooldridge.

I don't know Wooldridge very well, but I know that his book is popular and well-loved among econometrics grad students.

These books have a bias in favour of microeconometrics and deal less, if at all, with time series methods. More comprehensive coverage can be found in

"Econometrics" by Hayashi

or

"Econometric Analysis" by Greene.

The book by Hayashi is probably better for self-study; Greene is more of a comprehensive reference book.

Answer 2

In case you want to go into Dynamic Stochastic General Equilibrium models, the book

"Dynamic General Equilibrium Modeling" 2009 (2nd ed.) by B. Heer and Al. Mausser is a very useful companion.

For econometrics I would also suggest

"Probability Theory and Statistical Inference: Econometric Modeling with Observational Data" 1999, by A. Spanos. It provides the statistical foundations of econometrics in a way no other book does.

To comment on some of the suggestions related to econometrics in the other answers:

"Econometric Analysis of Cross Section and Panel Data" by Wooldridge. has a very fresh approach, (chapter 2 "Conditional Expectations and Related Concepts in Econometrics" was long missing from any Econometrics book).

"Econometrics" by Hayasi, except of presenting a new synthesis focused around Extremum Estimators (i.e. Maximum Likelihood and Generalized Method of Moments), and of giving space to Time Series, Unit root econometrics and Co-integration, it has theoretical depth alongside very practical applications, a combination which is not usually found.

"Econometric Analysis" by Greene is indeed a compendium of results -very useful for lookup purposes, but not really suitable for self-study.

Answer 3

In addition to the notes from @Ubiquitous, I'd recommend these excellent and free/open resources from a number of academics. These are all posted freely and openly by the authors, and are all high-quality (and all at the "graduate level").

• John Cochrane's Time Series for Macro and Finance monograph pdf
• John Stachurski's intro grad textbook, online
• Michael Creel's graduate econometrics notes, with code
• Kenneth Train's course and textbook on simulation-based econometrics
• Bruce Hansen's all-around graduate econometrics textbook

Cochrane's monograph should be your first stop for time series, period. Stachurski covers some of the very cool theory behind 'metrics. Creel provides a very nice, slide-style set of notes -- gives you a chance to "quickly peruse" graduate econometric theory (as well as very cool parallel computation estimators). Train has nice chapters on simulation-based estimators. Hansen is an all-around textbook with reasonable depth.

Somewhat further afield, but perhaps useful for the intro level, are the "Think X" series of books from Green Tea Press:

• Think Stats for basic statistics second ed in progress
• Think Bayes for Bayesian estimation approaches

These are truly intro-level books -- and very excellent as such. I use the "Train Problem/German Tank Problem" from Think Bayes whenever I need to illustrate the usefulness/intuition of Bayesian approaches to a beginner.

I will note that for self-study, Hayashi (noted in another answer) has a lot to offer.

Edit: If you are looking for broad examples of methodologies, I would strongly recommend "Mostly Harmless Econometrics." The book has a particular focus and mild methodology bias (see Andrew Gelman's nice review, which points out some of the topics missing), but it is still excellent at diving directly into application of their ideas, via examples from the literature and replication. I strongly believe that replication is key for learning about research methods, so I like their approach quite a lot. One minor warning: to get to the applications, you have to get through their very long intro chapter on the basic of regression. But hang in there -- that chapter is great as well, and you'll then get to enjoy the fruits of application/replication after it.

Also, if you want to know a little more about what economists do, check out these two non-technical books: "Passion and Craft" and "Lives of the Laureates." ("The making of an economist, redux," is also good, but perhaps less ...exciting of a read. Take that for what its worth, since the first two books aren't necessarily action page turners.)

Edit 2: as @cc7768 notes in the comments, John Stachurski and Tom Sargent's website, QuantEcon, is absolutely fantastic. They essentially work through the major basic concepts you need for computational economics (and so many things are computational these days), and give you clear, working code examples for a vast number or major problem-types. You should very definitely check it out if you are interested in economic methodologies -- it's probably the first place I'd recommend you stop, actually. Very easy way to dip your feet into serious methodology examples. I learned an enormous amount from early versions of these notes.

Answer 4

Mostly Harmless Econometrics is probably what you're looking for. However, that's because you're a math major--it's probably not a good recommendation for your average undergrad Econ major (the book itself is quirky and offbeat, which belies the sophistication of the mathematical methods presented).

In addition to the Econometrics textbooks mentioned by others, you might also want to look at A Guide to Econometrics by Peter Kennedy--it works well as a supplemental text and, like "Mostly Harmless", it's pretty cheap compared to those textbooks (ex: Wooldridge @ $250 on Amazon).

One last thing you need to check out (I've been recommending this as often as I can): Hal Varian--the chief economist at Google--recently wrote this article for the Journal of Economic Perspectives ("Big Data: New Tricks for Econometrics"), detailing some of the techniques of Machine Learning most relevant to Econometrics (there's also a data set and R code which accompanies the article). And, if that piques your interest, you might want to look at The Elements of Statistical Learning by Hastie, et al.--or, for those looking for something less math-heavy--An Introduction to Statistical Learning, both of which expand on some of the methods detailed by Varian (and both are available in Hardback or as a free pdf from the authors' sites).

Answer 5

Here is an introductory set of self-published notes by Tony Cookson from an honors-level undergraduate course. They're clear in distinguishing causation from mere statistical relationships.