Friday, May 23, 2014

Econ 101 should have more empirics


At Bloomberg View, I give my prescription for how Econ 101 should change, and discuss further the relationship between high school physics and the Turkish conquest of Constantinople: 
When I was 15, I was probably a little too obsessed with the Turkish conquest of Constantinople. In my teenage mind, that was the moment when the world became modern -- when cannons knocked down the great curtain walls that had kept the world’s greatest fortress impregnable for a thousand years. It was the triumph of science and engineering over medieval brute force. 
So Constantinople was on my mind when our intro physics teacher took us to the lab to test the theory of projectile motion. It involved lobbing metal balls at a piece of tape and recording where the balls hit. When we looked at the tape, it was like magic -- the math had predicted just where the ball would hit. Physics theory really worked. “So that’s how they did it!,” I exclaimed to myself, thinking of the Turkish cannon pounding away at the Byzantine walls. (Little did I know that 15th century gunners found their range by trial and error; mathematical firing solutions would come centuries later.) 
So that’s how I learned Physics 101. But a decade or so later, when I sat down with Greg Mankiw’s ``Principles of Economics'' to teach myself Econ 101, I found myself wondering what kind of equally convincing tests I’d be doing if I were taking the course for real. Fast forward another four years, and I was teaching Econ 101 (actually, Econ 102, but let’s stick with the jargon), right out of Mankiw’s book. And guess what -- there were no demonstrations. There was basically nothing but theory -- all the pretty little theories of comparative advantage and monopoly pricing and loanable funds, and not a whiff of evidence to back them up. Oh, I tried showing my undergrad students a couple of empirical papers, and drew pictures of what a simple regression would look like, but they mostly ignored these, and with good reason -- nothing empirical was on the tests...

24 comments:

  1. Anonymous1:26 PM

    That was a nice piece, but would have held up much better with an example or two to show how econ 101 would be better for it.

    One of my favorites was monopoly pricing of a pharmaceutical at the end of patent protection. That's got a lot of juicy parts: market failure, counter-intuitive result (price goes up), and the data can't be terribly hard to come by.

    - mere mortal

    ReplyDelete
  2. Anonymous1:53 PM

    I was about to agree - I remember some data graphs in a textbook cementing my love of econ in undergrad. But really, at the 101 level, a well-suited example is MUCH more powerful than technical empirics. Examples of things that shift supply & demand. Do price control theory, then look at Venzuela and Zimbabwe. Find examples of firms entering & exiting industries in response to profits/losses. Good economics textbooks and teachers do this all the time, often with just narrative examples. Some data graphs if you must. But save the empirics for Intermediate level.

    ReplyDelete
  3. Anonymous1:55 PM

    This guy has some non-technical examples that undergrads can sink their teeth into: https://www.youtube.com/user/thateconguy

    ReplyDelete
  4. Anonymous3:25 PM

    The hard complicated reality of empirical examples might correct some of the airy "free market priesthood" complaints (which are just inappropriate caricaturing), and show that the profession actually has developed tools to deal with the complex world. It's just that some tools are more powerful than others, and the choice of tool might just be more than freaky-deaky political value judgments.

    Also, empiricism is great for creative problem solving. Companies are hiring in-house economists to make sense of their massive data loads: a statistics plus maths plus programming plus economics skill set.

    ReplyDelete
  5. This is relevant to my interests!

    Weirdly, my school doesn't require students to take econometrics, though that might change in the future. Our graduate program is apparently pretty heavy on empirical analysis, with the aim of churning out solid working economists, but in the undergrad program you can make it as data-driven or as fuzzy and theoretical as you want. I know a lot of people who never want to take econometrics, and are basically just looking for a painless step to business or law school. But, so far my favorite professors are the ones that can bring in a lot of empirical examples. My math econ professor is a trade economist, and made a point to include lost of application examples so we could see how the theory actually holds up. My econometrics prof is a development economist, and has used a lot of examples from his own work, in addition to having us recreate some famous results in class. I think it really opens up a lot of possibilities to see how economists have tried to test their ideas through data. Otherwise it's just a philosophical argument. Plus, I think it's a lot more valuable in the workplace. How else will you be able to tell what business practices get the best results?

    We actually talked about these student complaints in my History of Thought class...the professor is a total heterodox guy, and sort of presented it like, "See? Neoclassical economics is all bunk!" But, I don't really buy that any more than I buy the notion that markets are always perfectly competitive. It seems to me like whatever your economic philosophy is, you've got to have a way to (at least a little bit) prove it or it's useless.

    ReplyDelete
  6. John S6:04 PM

    What about including more emphasis on the results of experimental economics (e.g. Vernon Smith, Herb Gintis)?

    ReplyDelete
  7. bjdubbs10:35 PM

    How about some basic accounting? It seems hard to believe that you can have a good grasp on how the economy operates if you can't read a balance sheet and income statement. The old Samuelson textbooks used to have some stuff about accounting, but that has long since disappeared. The basic problem with econ is that it teaches you "how to think like an economist" when it would be better to teach "how to think like a consumer/CEO/treasury secretary."

    ReplyDelete
    Replies
    1. Anonymous10:51 PM

      Pfft, economics is about overcoming the limitations of arithmetic and accounting.

      Delete
  8. So students should start with economic history?

    ReplyDelete
  9. ArmChairEconomist1:34 PM

    Let us analyze Noahpinion's recent acts in the econoblogosphere: Noah trolled Tyler Cowen in a petty attempt to be included in the Pikketty debate, and succeeded in being recognized. Noah pleaded that Steve Levitt pay attention to a not-very-well thought out critique of an anecdote Levitt casually mentions in a pop-econ book and Noah got Levitt's attention. Now Noah is begging for Mankiw's attention and seems to have succeded there as well. This is all in the past ten days. As the king of EconoTrolls, I think it is past time for Noah to update his very own (and probably best) post on EconoTrolls & include some sort of new category.

    ReplyDelete
    Replies
    1. Actually I'm thinking of turning EconoTrolls into a book! :-)

      Delete
    2. ArmChairEconomist8:47 PM

      You should do it. Based on this post, we know the implied importance you place on empirics (even if your actions tell us that empirics are important only when it fits your agenda). And, you undoubtedly have first hand experience with the empirics of EconoTrolling.

      Back to the teaching, when I've taught intro (macro)economics, I use Mankiw's text only as a guide. And, it's a great guide. I then I apply empirics to the text. When I teach, approx one-quarter of the semester is devoted to macro history since the 60's-ish & primarily revolving around $policy (Stagflation, the Phillips Curve mistake, Volcker/s disinflation, Great Recession, etc...). It's disappointing to me (based on your comments) that other instructors apparently aren't doing the same. But, do we know others aren't doing the same? Maybe YOU rely too heavily on a textbook and don't introduce empirics (I haven't a clue how you teach, but am assuming this post is based on what you know best - the way that YOU teach, because obviously anyone wouldn't just assume that others blindly follow a text word-for-word and then write a Bloomberg piece about it).

      Anyways, if nothing else, hopefully this book will finally provide another line to your CV. Any idea of how are you planning to expand the EconoTrolls to include a new domain for someone with anything but Noahpinion?

      Delete
    3. Econotroll category: Empiricist

      "Econ 101 should have moar empirics"

      How they see themselves:
      http://archives.thestar.com.my/archives/2012/8/24/metrobiz/m-hair.jpg

      How the world sees them:
      http://3.bp.blogspot.com/-Drcwk2z-aMc/UGEmHchfLQI/AAAAAAAAC1s/ctjwnQZ9jAM/s1600/Super+Talent.JPG

      Favorite blog: Anyone who will link back

      Favorite dead economist: David Hume

      Will appear in response to posts regarding: theory vs. evidence

      Craziest idea: The bad old days are over, economics is a science now.

      Special attack: Net.Troll

      Secret weaknesses: Unfalsifiable models with multiple free parameters are impervious to empirics

      Delete
    4. Hehe. Excellent.

      Delete
  10. The biggest surprise is that you seem to think Mankiw's useless textbook is in any way worth elaborating on. That would be like a Physics professor illustrating the principle of a flat earth.

    ReplyDelete
  11. Anonymous7:12 PM

    You state "Kay is right when he says that “heterodox economics” isn't a real alternative to what they teach in Econ 101. To understand why, you will have to spend a month or so immersed in the writings of the “heterodox,” "

    Wow! You must be a genius (and a quick reader).

    ReplyDelete
  12. The first course should be on Finance and Investments without any Econ 101 or 102. Econ should be only in grad schools as it is mostly arm waving and it is fun to do that in grad schools.

    Just about everybody has first had knowledge of how they are paying for their education and it is real - like the real walls and canon balls.

    ReplyDelete
  13. But this is precisely why Economics is lagging so far behind Physics as a science. Wanna know where a ball is going to land? Try throwing it a few times and see what happens. Wanna know if the Higgs Boson exists, just try smashing a few trillion particles together and see what happens. Wanna know how a change in oil prices will affect the interest rate 5 years from now... Assume stationarity...

    ReplyDelete
    Replies
    1. You're right, why didn't us economists think of doing randomized controlled trials? Exogenous variation solves all our problems.

      Pass me 100 identical global oil markets, and I'll get to work.

      Delete
    2. ArmChairEconomist12:31 PM

      Psssssh, silly economist thinking a sample size of 100 is sufficient............

      Delete
  14. Anonymous4:30 AM

    Noah, is your Stonybrook email the best way to contact you privately?

    ReplyDelete
  15. ArmChairEconomist4:12 PM

    Looks like this trolling attempt also got the attention of Tyler Cowen? hahahaha http://marginalrevolution.com/marginalrevolution/2014/05/assorted-links-1140.html

    ReplyDelete
  16. Anonymous4:55 PM

    Mankiw's textbook is filled with claims that are far from reality and evidence from, at least, experiments would have toned down the governing ideology behind the book's major claims. Of course it is no surprise to read his reaction to this post.

    ReplyDelete
  17. Rather than wade into empirical testing, I focus on the challenge of putting theory to work to solve problems. It is one thing to say "people are motivated by incentives." It is quite another to say "I am trying to get someone to do X, so here is how I will pay them." Students very quickly see the many challenges of applying theory--far harder than trying to hit a wall with a rock. The first challenge is to recognize the difference between the theoretical constructs (like incentive and motivation) and how we create and measure them (like a specific bonus plan and hours worked or output produced). Empirical challenges come later.

    If anyone is interested, my book, What Counts and What Gets Counted, is available for free at http://ssrn.com/abstract=2427106, with links to over 100 videos. It covers Managerial Reporting, not Econ 101, but draws heavily from microeconomics.

    ReplyDelete