Development Economics: Micro-based vs. Macro-based

[Econ2011]

I see that most people perceive development to be a micro-based area of specialization. I talked to a few graduate students and they all said that development-focused students mainly work with professors specializing in microeconomics/labor etc. Having come from a developing country, I can see where people are going with that. However, I feel like we need to look at it from a macro-perspective as well. What do you guys think? Why do you think there is less of an interest on the macroeconomics side of it?

 

Inputs are appreciated!

[Andronicus]

I think the macro side of it is typically called growth. There is plenty of interest, just under a different name.

[Econ2011]

I am trying to convince the adcoms through my SOP that I am for development economics and areas complementing my interest would include growth, trade, international finance etc. Perhaps, I am being a bit too broad?

[dreck]

I see that most people perceive development to be a micro-based area of specialization. I talked to a few graduate students and they all said that development-focused students mainly work with professors specializing in microeconomics/labor etc. Having come from a developing country, I can see where people are going with that. However, I feel like we need to look at it from a macro-perspective as well. What do you guys think? Why do you think there is less of an interest on the macroeconomics side of it?

 

Inputs are appreciated!

 

Macro development is a massive field; I'm not sure whether there are more people in micro or macro development, but it's not like the latter is something everyone ignores. I'm a little shady on my history of economic thought, but I think development is a traditionally macro field; the micro approach came along later.

In addition, part of the reason there's so much micro development going on right now is due to innovative techniques in experimental methods and randomized trials, which provide new ways of answering the questions of micro development. If the grad students you talked to are at a school where the experimental approach is big, that's probably why more of them are into micro development.

[thewhiterabbit]

I'm a little shady on my history of economic thought, but I think development is a traditionally macro field; the micro approach came along later.

 

Yep.

 

Development was historically macro, and I think the standard narrative on the issue was that macro dev had become somewhat of a moribund field until the micro people came along and injected a new perspective and some vitality. In my view the main point of contention between the two sectors is what constitutes an appropriate standard of evidence. The people doing micro dev tend to work on experiments and trials that can more conclusively determine causality than you could ever be able in macro dev. Macro dev people counter that their work is of broader significance and importance than micro dev.

 

I'm biased (micro dev person here) but a good place to look for commentary on the issue is Dani Rodrik's writings and blog; I'm quite sure he has a post addressing this out there, in fact, but don't have time to look for it now. He's more of a macro person but is in Cambridge where micro-dev tends to dominate, so he has thoughtful things to say.

[dreck]

 

In my view the main point of contention between the two sectors is what constitutes an appropriate standard of evidence. The people doing micro dev tend to work on experiments and trials that can more conclusively determine causality than you could ever be able in macro dev. Macro dev people counter that their work is of broader significance and importance than micro dev.

 

This is being picky, but I don't know if I'd go so far as to say they "determine causality." They attempt causal inference that would not be possible doing emprical macro development, but there are some big obstacles, namely the generalization problem and the third-cause problem--which usually takes the form of participants gaming the system in some way, or acting a certain way because of cultural norms rather than the reasons the researcher wants to ascribe. Of course, in the best work, the researchers make pretty compelling arguments for their results' validity. But I would say that a big part of the contention between micro and macro development is the fact that "mediocre/bad" field work papers are arguably much more likely to get published right now than "mediocre/bad" empirical macro development, despite concerns about generalization, etc.

[Andronicus]

Finding a treatment effect in a well-controlled experiment is pretty much as close to determining causality as anyone can ever get.

[EnvDep2011]

Just depends on what want to argue gets caused. If you're arguing that "intervention X caused Y in place Z under conditions A, B, and C," then yes, experimental evidence puts you on extremely solid ground. But if you want to say "intervention X causes Y generally," then it's a whole different story.

[michaelmas]

Development Macroeconomics is mostly Endogenous Growth Theory (I said mostly), without prescribing any specific policy recommendations, other than the stylized ones.

 

The reason, I think, that Development Economics is mainly affiliated with Micro instead of Macro is that the 80's growth engineering attitude has largely disappeared after so many failed attempts.

[Econ2011]

I guess causality is easier to interpret in a micro-setting but it would be difficult to generalize given the differences that persist between different geographical regions, income levels etc. I still think we would have to analyze from a macro-setting if we are to understand the whole picture and general relationship.

[dreck]

The reason, I think, that Development Economics is mainly affiliated with Micro instead of Macro is that the 80's growth engineering attitude has largely disappeared after so many failed attempts.

 

While the theory of growth has changed little since the 80s, and the "growth engineering" attitude has faded, the empirics of growth/macro development is still an active field for research, which has important things to say about policy. See, for example Acemoglu's Settler Mortality Paper, and the wealth of imitations it has inspired. I attended a seminar a while back on culture and development using a similar approach. So while micro development is definitely hot right now--and who knows how long that will last--there's still plenty of work to be done on the macro side of things.

[greatexpect]

I think Dani Rodrik does a good job of explaning the "bifurcation" in this talk

 

[OneArmedEcon]

Personal opinion: nearly everything in economics is applied micro, including the various fields contained in modern Macro.

[thewhiterabbit]

This is being picky, but I don't know if I'd go so far as to say they "determine causality." They attempt causal inference that would not be possible doing emprical macro development, but there are some big obstacles, namely the generalization problem and the third-cause problem--which usually takes the form of participants gaming the system in some way, or acting a certain way because of cultural norms rather than the reasons the researcher wants to ascribe. Of course, in the best work, the researchers make pretty compelling arguments for their results' validity. But I would say that a big part of the contention between micro and macro development is the fact that "mediocre/bad" field work papers are arguably much more likely to get published right now than "mediocre/bad" empirical macro development, despite concerns about generalization, etc.

 

I don't really disagree but would like to clarify (as others have done) that I don't see the generalization problem as a causality problem. I just see it as a generalization problem :-). I.e. we can say "this treatment had this effect in this setting" and then the degree of generalization from there is somewhat conjecture.

 

I agree generalization is a problem, though. One of the great things about the existence of the Poverty Action Lab, imho, is there is a nonacademic structure for repeating field experiments in other settings, helping a lot with generalizability, whereas a full professor would not want to do something that's been done before.

[Zeno]

Personal opinion: nearly everything in economics is applied micro, including the various fields contained in modern Macro.

 

Including micro theory? :P

[philecon]

Dani Rodrik's weblog: Re-uniting development economics

 

Here's the link to Rodrik's blog post.

[petecheese]

Applied micro FTW

[appl09]

Yep.

 

Development was historically macro, and I think the standard narrative on the issue was that macro dev had become somewhat of a moribund field until the micro people came along and injected a new perspective and some vitality. In my view the main point of contention between the two sectors is what constitutes an appropriate standard of evidence. The people doing micro dev tend to work on experiments and trials that can more conclusively determine causality than you could ever be able in macro dev. Macro dev people counter that their work is of broader significance and importance than micro dev.

 

I'm biased (micro dev person here) but a good place to look for commentary on the issue is Dani Rodrik's writings and blog; I'm quite sure he has a post addressing this out there, in fact, but don't have time to look for it now. He's more of a macro person but is in Cambridge where micro-dev tends to dominate, so he has thoughtful things to say.

 

I'd generally agree with what you said - I think in the Harvard/MIT development clan, there is a recent shift to even higher standards of identification...which has lead to research questions tend to be micro-based. Obviously, the identification approach taken can significantly limit the types of questions one can address.

 

It baffles me why causality should be of primary importance in addressing research questions. True, RCTs and other micro-based identification approaches can answer questions with more internal validity but should that come at the price of less external validity (usually the downside of RCTs along with the GE effects) or the type of questions one should address?

 

You make a point of JPAL replicating experiments to address the external validity concerns but replications can not address three major problems: one that has to do with the aggregation bias from micro theory; the other that has to do with a concern that a relationship that holds at the individual level may not hold on a more macro level (say, country- level). For example the health-income relationship is a perfect example of a how the relationship between health and income can be addressed with an RCT and even multiple RCTs in multiple contexts but an RCT can never address this relationship on a macro level (does better health lead to higher economic growth, which may be the more relevant and important level to address the policy question at) which is why the micro studies generate significantly different predictions than the macro ones on this relationship. And the third concern is the ever-so-present concern of the GE effects. Again, RCTs alone can not address this GE concern. A scaled-up intervention can generate a completely different relationship between two variables than the one predicted by a great LATE estimate from an RCT.

 

In many aspects of development content validity (the actual measure of what we intend to measure) can also be a very serious concern.

[triple_sec]

Personal opinion: nearly everything in economics is applied micro, including the various fields contained in modern Macro.

 

Ditto. I think that thinking in terms of the traditional micro/macro splitting hardly makes any sense any longer. By now, what is labeled collectively as economics has become a great unified theory of sorts organized around the basic principle of equilibrium behavior of rational agents who optimize. In this sense, what you call micro today could be equally legitimately labeled as "foundations of economics" and macro as "applied general equilibrium theory." Just my two cents.

[kipfilet]

Ditto. I think that thinking in terms of the traditional micro/macro splitting hardly makes any sense any longer. By now, what is labeled collectively as economics has become a great unified theory of sorts organized around the basic principle of equilibrium behavior of rational agents who optimize. In this sense, what you call micro today could be equally legitimately labeled as "foundations of economics" and macro as "applied general equilibrium theory." Just my two cents.

 

Seconded!

[michaelmas]

Applied micro FTW

FFS, GTFO u SOB.

 

But seriously, I think there is a rule on this forum against this sort of "textease", or whatever it's called.

 

Just had to point that out.

[dreck]

I don't have a problem with an ftw occasionally.

 

I don't really think macro is just applied micro though. Having micro-foundations is not the same as doing microeconomics. But this is definitely just an opinion.

 

Wrt the casuality/generaliztion. If you define causality narrowly, to mean that the treatment effect was caused by the treatment, then I don't disagree. But most researchers aim higher; the whole point of randomized trials is to try to design research in such a way that if a treatment effect is observed, we have evidence to say something like, "better education causes better health outcomes," which uses a more general concept of causality. So I do think generalization and causation are intertwined, but this more semantics than a substantive disagreement. I'm not hating on randomized trials; there's a chance micro development will be my field.

[thewhiterabbit]

@dreck I think we're in general agreement.

 

@appl09 I agree with much of what you wrote except for this:

 

It baffles me why causality should be of primary importance in addressing research questions. True, RCTs and other micro-based identification approaches can answer questions with more internal validity but should that come at the price of less external validity (usually the downside of RCTs along with the GE effects) or the type of questions one should address?

 

Without attempting to determine causality the interpretation of research is often too strong given the evidence available. To take your example, health and income at a nationwide level, isn't causality of utmost importance in regards to policy interpretation? If health is necessary for income and growth then we need to focus on health; if it's the opposite we need to focus on spurring the economy. In truth it's probably both, which is why this is such a tough question to answer, but if one effect dominates, that's extremely relevant to actual policy choices. And yet it's nearly impossible to answer this using standard macro approaches (though if you know of such a paper that you think does so I would be happy to be corrected).

 

I absolutely agree with you that there is a tradeoff here between a higher standard of evidence and a higher level of (attempted) relevance. I think I can hazard to guess that we might find substantive disagreement in our visions for the field. Most micro dev people are completely aware of both the promises and limitations of the randomized approach (etc - there are other ways to do micro development in this 'clan' without randomization) but would say that it's a better foundation upon which to build the field than the 'old way.' The longer-term goal from the micro perspective I believe would be to achieve results that are externally valid and address GE effects when relevant (not always relevant depending on the question), but this is in the long-term, and for now we're starting from the bottom up, with rigor. Whereas others would rather not sacrifice a lot of pretty good research because it doesn't meet this new, higher bar, especially since it's addressing larger questions.

 

To go back to the health and growth example, a micro dev person would find a result demonstrating that better general health has a causal effect on an individual's income to be very interesting because it demonstrates that there are physical limitations to a person's ability to do productive work that stem from general health. I agree with you that in no way can you then say 'improve health and we'll improve growth' because of GE and other problems. But it provides a micro foundation, as it were, to understanding the macro dynamics. It's a piece of a large puzzle. It doesn't address the question posed above - does growth require health or health require growth? - any better than the macro peeps can, but it does say something conclusive if interpreted in a properly humble manner.

 

In truth I don't see as much of a conflict as there is sometimes made out to be. I believe there is room for all of us in development and I think you probably do as well. I do think that the added vitality and attention and fresh perspective is generally good for all.

 

Though I would also hazard that academic trends are annoying :). Even as an 'insider' I feel this way.

[dreck]

It baffles me why causality should be of primary importance in addressing research questions. True, RCTs and other micro-based identification approaches can answer questions with more internal validity but should that come at the price of less external validity (usually the downside of RCTs along with the GE effects) or the type of questions one should address?

 

Ok, so whiterabbits post is longer and I mostly agree, but I think the answer to this boils down to two things.

1) Causality is important because the goal of much of development economics is policy recommendations. Without knowing which direction causality goes, you can't make good policy.

2) This doesn't mean that we have to ignore general equilibrium effects or macroeconomic concerns; the research is meant to inform policy not dictate it based on the basis of an experimental finding. I agree that replication doesn't fix all the problems with the approach, but it's not as if randomized folks advocate that we simply ignore GE effects or the aggregation problem.

 

I know it already works this way much of the time, but this discussion seems to me like good evidence why micro and macro dev people should be talking to each other instead of quibbling about methods. It's not as if both sides don't know the problems with their own approach, but you've got to try to answer the important questions in as many ways as you can.

[OneArmedEcon]

I'll just add that people who do development really well (e.g., Acemoglu) tend to present both "macro" and "micro" arguments (not necessarily in the same paper, but in fact have a series of papers on a central topic using various methodologies that complement each other).