Close

Complete Captcha

• 10 years - 155 posts - Latest - RSS
• Thread: 597 Goods vs 226 No Goods

1. Economist
   2fa7
   

   Spoke with this Stanford phd from my cohort a month ago. He has moved away from structural and gone into reduced form labor stuff. He did structural labor but couldn't get his stuff published. Even at economic inquiry, he has trouble (where peter ardiciano is an editor). He has had a much better time publishing reduced form stuff including an ReStat.

   Thought I might mention it. The debates over structural and reduced form are interesting, but the above anecdote makes me grateful I didn't train on structural econometrics. It seems like it's nice to know that stuff, but ultimately you are taking risks.

   10 years ago # QUOTE 40 Volod 21 Vlad !

2. Economist
   c54e
   

   Arcidiacono no longer edits EI.

   Write better papers.

   10 years ago # QUOTE 20 Volod 4 Vlad !

3. Economist
   b644
   

   N = 1.

   OP sounds like a star.

   10 years ago # QUOTE 32 Volod 8 Vlad !

4. Economist
   3364
   

   Wahh, nested fixed point pseudo ML too hard!!

   10 years ago # QUOTE 4 Volod 3 Vlad !

5. Economist
   2fa7
   

   Not saying I'm a star. Just recounting an anecdote. Had never heard of someone reverse on their training so soon after graduating, and then talk about the challenges like that. Structural has a smaller target. Personally, the appeal of it diminishes everyday.

   10 years ago # QUOTE 28 Volod 1 Vlad !

6. Economist
   b017
   

   Spoke with this Stanford phd from my cohort a month ago. He has moved away from structural and gone into reduced form labor stuff. He did structural labor but couldn't get his stuff published. Even at economic inquiry, he has trouble (where peter ardiciano is an editor). He has had a much better time publishing reduced form stuff including an ReStat.
   Thought I might mention it. The debates over structural and reduced form are interesting, but the above anecdote makes me grateful I didn't train on structural econometrics. It seems like it's nice to know that stuff, but ultimately you are taking risks.

   If an economist can program structural econometics, why would doing reduced form analysis be a problem?

   10 years ago # QUOTE 5 Volod 17 Vlad !

7. Economist
   2fa7
   

   ^ I don't follow. The person was moving away from structural not because there was a problem, but because methodologically he had problems with publishing in it. He had a diversified portfolio due to coauthors, and his ReStat plus an AEJ applied were both econ of education studies that were, for lack of a better, traditional compared to structural estimation. He wasn't saying anything about which one was eaiser for him; he was noting that structural papers were harder to publish. And when he spoke about it, he spoke as though it were in the past for him. He was clearly investing in a different human capital from the conversation we had. My point is only that the number of editors with preferences for that is shrinking. I didn't realize ardiciano at duke had left EI, but that's even one less person. Whereas a study that identifes causal effects using reduced form methods has a range of top field (JHR, JOLE, Jpube, JHealthEcon, JLawEcon) and ReStat, AEJ applied, and top general. Outside of IO, structural seems less and less appealing.

   10 years ago # QUOTE 7 Volod 1 Vlad !

8. Economist
   0455
   

   tenure clock has killed serious work, macro or structural work

   10 years ago # QUOTE 32 Volod 12 Vlad !

9. Economist
   1d86
   

   I suspect your friend just isn't bright enough to do good structural work. There's no reason to go on and claim that the approach iss dead.

   10 years ago # QUOTE 11 Volod 36 Vlad !

10. Economist
    2f4d
    

    maybe I'm crazy and maybe its the company I keep, but econ is getting less insular, and more similar to the non-stupid branches of other social sciences. we apply stats to social science data, and our results don't matter unless they are relevant to the world at large.

    structural is worthless because nobody really believes in economic theory, nor should they.

    if you are smart enough to learn structural, learn real stats, not just econometrics.

    10 years ago # QUOTE 19 Volod 53 Vlad !

11. Economist
    cbfe
    

    Please describe real stats in great detail, I only know MHE and Hayashi

    10 years ago # QUOTE 27 Volod 0 Vlad !

12. Economist
    b017
    

    maybe I'm crazy and maybe its the company I keep, but econ is getting less insular, and more similar to the non-stupid branches of other social sciences. we apply stats to social science data, and our results don't matter unless they are relevant to the world at large.
    structural is worthless because nobody really believes in economic theory, nor should they.
    if you are smart enough to learn structural, learn real stats, not just econometrics.

    Yes, but the problem is that "Real" stats wasn't developped to provide protocols social scientists can use for working with nonexperimental data the way econometrics has. This is why when you start talking to other social scientists doing reduced form work outside of the economics profession you quickly realise they have no idea what they're doing and when you confront them on not being critical about model assumptions they invoke the hand-wavey rhetoric of only caring about "results". E.g., most SPSS users I've met think that endogeneity issues can be solved with simultaneous equation models even when the model is under-identified.

    10 years ago # QUOTE 40 Volod 2 Vlad !

13. Economist
    94e7
    

    I concur with OP. I spent 3 years trying to publish my structural dissertation and extensions. 0 pubs. ZERO!

    Fortunately, a reduced form side project with one of my colleagues changed my perspective on my publication strategy and the next four years produced 10 pubs from top fields to solid B+ outlets (second/third tier general interest). Upgraded my placement. Nailed tenure.

    Sure, you CAN publish in structural. But if you are smart enough to actually do structural well you can retool and crush the reduced form crowd...and at the end of the day it's pubs that matter!

    10 years ago # QUOTE 39 Volod 7 Vlad !

14. Economist
    2f4d
    

    I've seen economists make lots of stupid assumptions. causality ain't causal when your model makes heinous functional form assumptions, and people forget (or never knew) that a LATE interpretation is different from a ATE interpretation.

    10 years ago # QUOTE 14 Volod 3 Vlad !

15. Economist
    b017
    

    I gooded you 5f5e, but what you're saying still made me depressed.

    10 years ago # QUOTE 11 Volod 1 Vlad !

16. Economist
    baa2
    

    I personally find reduced form labor very very boring. Therefore I do structural. But I'm not in the NA tenure system, so tenure clock doesn't matter as much.

    10 years ago # QUOTE 10 Volod 6 Vlad !

17. Economist
    caf5
    

    I don't think from a top school will help much for a top field journal publication in structural work ...

    Spoke with this Stanford phd from my cohort a month ago. He has moved away from structural and gone into reduced form labor stuff. He did structural labor but couldn't get his stuff published. Even at economic inquiry, he has trouble (where peter ardiciano is an editor). He has had a much better time publishing reduced form stuff including an ReStat.
    Thought I might mention it. The debates over structural and reduced form are interesting, but the above anecdote makes me grateful I didn't train on structural econometrics. It seems like it's nice to know that stuff, but ultimately you are taking risks.

    10 years ago # QUOTE 0 Volod 1 Vlad !

18. Economist
    0455
    

    For reduced form work, the danger is at naive functional form assumption (linearity can be far from truth, but most empirical work won't spend time on it.), and identify from something from variations far from representative, even though it may be exogenous. (Frankly, identifying return to schooling for everyone from high school dropouts is quite rediculous.)

    For structural work, the problem is that you have a model in mind, and you want to push it to the furthest, neglecting other possible sources of heterogeneity.

    10 years ago # QUOTE 6 Volod 4 Vlad !

19. Economist
    e034
    

    People overrate the difficulty of doing structural work. The real obstacle is time. But because computation is getting faster and easier, this will become less prohibitive in the long run. In the meantime, you might as well do some reduced form work so you can get tenure.

    10 years ago # QUOTE 8 Volod 8 Vlad !

20. Economist
    9a05
    

    What's the rough proportion of structural vs reduced form work recently?

    10 years ago # QUOTE 3 Volod 2 Vlad !

Reply »