7
$\begingroup$

My institution is updating some of its bachelor and master study programs in economics and business administration, and the question of choice of software for econometrics and related classes has come up. I wonder what the leading institutions are using and where the field is headed. Hence my questions:

  1. What software is currently most used for teaching econometrics and related subjects in leading study programs of economics (bachelor and master, or undergraduate and graduate)?
  2. What software is currently most used for teaching econometrics and related subjects in leading study programs of business administration (bachelor and master, or undergraduate and graduate)?
  3. Are there any pronounced tendencies of an ongoing switch to another software or a switch to take place in the near future (say, during 2020-2025)?

The question is intended to be objective. It does not ask for opinions on your favorite software (though I would not mind hearing some core points on that as long as they contribute to choosing an appropriate software for our study programs).

Edit (after the bounty was added): Please try to make the answer specific by including

  • the year to which your knowledge refers;
  • study program: economics, business, ...;
  • type of institution/deparment: university vs. something else, economics department vs. business school vs. ...;
  • tier: R1, R2 etc. or similar
  • country;
  • level: undergraduate/graduate or bachelor/master/Ph.D.;
  • course: econometrics or some other (in case you may wish to share your experience in a related but different course).
$\endgroup$
  • $\begingroup$ Could I at least get some anecdotal evidence from professors and students of some top schools? There must be some visiting Economics Stack Exchange. (Due to my own lack of activity here, I am not entirely sure about professors, but I expect at least some students show up here every now and then.) Contributions of anyone else with relevant knowledge will be equally appreciated. $\endgroup$ – Richard Hardy Jun 8 at 9:16
  • $\begingroup$ You might be interested in economics.stackexchange.com/questions/5428/…. $\endgroup$ – FZS Jun 24 at 12:54
  • $\begingroup$ Julia has been talked about in my department but this is at the PhD level $\endgroup$ – EconJohn Jun 30 at 16:10
5
$\begingroup$

In 2008 the American Economic Association (AEA) created a reproducible research requirement about posting the code and the data for their papers. According to analysis done by the AEA Data Editor Vilhuber, Turrito, and Welch (2020), Stata and Matlab were by far the most dominant software used in these submissions (see figure below, the raw data are here). Given the elite status of this family of journals in the profession, I think this is a reasonable proxy for the software used by elite economists and therefore in elite economics PhD training programs. For what it is worth, Stata and Matlab (and in that order) were also the software I was trained to use in my PhD econometrics classes and my work for my economics PhD, about ten years ago, at an R1 school in the USA. The only exception was my Bayesian econometrics class which I believe used R. As an undergraduate (also at a R1 school in the USA), I used Eviews in my econometrics classes and R in my probability and statistics classes.

Figure 5. Software by Supplement across Years, Files Migrated in October 2019

| improve this answer | |
$\endgroup$
  • $\begingroup$ Thanks for your answer! It looks like Stata is dominating by a large margin among the academic researchers. Given the age distribution in the profession, this cannot tell much about current teaching trends, though. The lag is between a few a few dozen years. On the other hand, there is substantial natural inertia (professors not being so quick to change software), so if you learn Stata today, your skills are unlikely to become obsolete fast. $\endgroup$ – Richard Hardy Jun 24 at 20:08
  • $\begingroup$ I also wonder about the popularity of Stata in the private sector. Do companies buy Stata licences for their employees, or do the freshly minted economists switch to some freeware, say, R? I know this is another question, but perhaps you have some knowledge of that as well? $\endgroup$ – Richard Hardy Jun 24 at 20:10
  • 2
    $\begingroup$ I think this is another question. That said, among firms employing PhD economists to do economic analysis, like for the litigation consultants, I believe it is common to pay for Stata or Matlab licenses. $\endgroup$ – BKay Jun 25 at 17:59
4
+50
$\begingroup$

At McMaster University in Canada (#4 in Canada for research and #98 worldwide) both undergraduate and graduate (MA and above) economics programs utilize STATA for econometrics for its ease of use (real coding is not needed to get started) and for its accessibility i.e. low cost. R and R studio is utilized as well but to a far lesser degree and only in Econometrics II (Graduate level).

The University of Guelph (#16), also in Canada, introduces students to R studio in undergrad econometrics. This is largely due to its open-source status. STATA is used in the upper-level courses (fourth year and above) in econometrics.

My personal preference, however, would be to integrate the use of SQL and Python for further analysis. This would offer students the ability to manage larger datasets as well as harness the necessary skills to data wrangle and become more adept with economic's research best practices. In doing so an institution would be able to set students up (in my opinion) for further success not only in research but in the public/private sectors as well.

I hope this helps.

| improve this answer | |
$\endgroup$
  • $\begingroup$ This is the kind of answer I was expecting from the beginning: concrete and with the right amount of detail. Thank you! $\endgroup$ – Richard Hardy Jun 30 at 7:48
  • $\begingroup$ You are most welcome :) $\endgroup$ – KeithRoberts Jun 30 at 19:58
3
$\begingroup$

As an Economic Sciences student, I can only testify for this sector.

We principally use EViews, for applied econometrics, time-series analysis. Some professors like to use the R programming language and Stata.

All three are extremely powerful tools.

| improve this answer | |
$\endgroup$
  • $\begingroup$ Thank you for your answer. Could you be more specific (please see my edit)? $\endgroup$ – Richard Hardy Jun 24 at 13:01
  • $\begingroup$ @RichardHardy such details are on my profile $\endgroup$ – the_rainbox Jun 24 at 15:40
3
$\begingroup$

Im currently in grad school of a R1 university in the US (Econ Dept). Here's my take on the issue:

  1. STATA is a very helpful tool in cleaning data and performing some preliminary analysis (enough at the undergraduate level).
  2. Beyond that people move to either R (ha very good graphics packages and very easy integration with Latex) or to Python/Julia/Matlab.
  3. During my coursework in grad school, we did our coding mostly in matlab or R. (I used python instead of matlab). Stata is taught at the undergraduate econometrics courses, but not in our grad courses.
| improve this answer | |
$\endgroup$
  • $\begingroup$ Thank you for your answer! Point 3. answers my question directly; it was Stata for the undergrad and Matlab or R in grad school. Regarding point 2, could you be more specific? Who are the people there? $\endgroup$ – Richard Hardy Jun 27 at 7:21
  • $\begingroup$ Ah yes, it includes a whole bunch of people - professors in our department (working on IO and related fields), professors with whom I have worked as a research assistant (in grad school, as well as undergrad days), and a number of my peers who are working on IO or trade. $\endgroup$ – Tomcat Jun 27 at 19:46
3
$\begingroup$

In my experience in Ecomomics I would say Stata and Matlab. Stata is widely used at the undergrad level and for microeconometrics, while Matlab is introduced more at the postgrad level for general econometrics classes.

In my personal opinion there is a clear tendency in schools to push students more into programming based software (it is a useful tool for several reasons including job market skills). Stata is a great software but once you have to deal with analyses outside of the traditional microeconometrics it has some very evident shortcomings.

Edit: this refers to my experience in BSc/MRes/Phd in a research University in the Netherlands and a Russell Group University in the UK through 2014-2020.

In terms of Econometric courses (in Economics departments) my experience in my undergraduate degree in the Netherlands was that STATA was the most taught software by far. I took classes in data analytics and big data in computer science departments and the choice of software there was R/Python and Unix (with the use of R through the terminal).

During my postgraduate degree in the UK Econometrics courses were taught using Matlab (computational macroeconomics was taught using Julia but it was always our own choice to pass this on and do the coursework in Matlab).

For my PhD I am still in an Economics department, however, I am also in a more finance focused topic and I find that my supervisors and people I collaborate with are more keen on Python than any other language.

| improve this answer | |
$\endgroup$
  • $\begingroup$ Thank you for your answer. Could you be more specific (please see my edit)? $\endgroup$ – Richard Hardy Jun 24 at 13:01
  • $\begingroup$ Your edit is great, thank you! $\endgroup$ – Richard Hardy Jun 30 at 7:50

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.