29 October 2009

Economics for reality or theory?

Andrew sent this:
I found your lecture today about the theoretical/mathematical vs. the practical application of economics very interesting, and very much along the lines of my own observations. I have noticed that there are two schools of thought in the field, that there seems to be animosity between them, and that we may beyond the peak of usefulness of quantitative analysis (as in, negative utility) - at least until we have some awesome breakthrough in mathematics that allows us to model life more accurately.

Math is cool (even though I hate it most days), but it will always just be math...nothing more. Math to me is like the uniform a basketball (or whatever) team wears...part of the game, but my interest lies with the game and not team uniforms.

I am curious about this because the dynamics between these two schools of thought have implications for me as I move forward.

I enjoy economics so much that I went back to school after a 20 year absence to study it. Over the years I worked for a few large corporations, worked as a consultant, and owned my own retail business. At this point I only have interest in practical application of economic theory. That could be research in a Master's program or finding meaningful work after Berkeley. Whatever I choose it will be real, tangible and practical.

As an aside, I had to laugh at 'mathturbation'...Whenever I hear someone talk in a way that suggests they think math is more than just a tool I hear Sammy Hagar singing "..to me it's all just mental masturbation"....and I think Paul Krugman is awesome.

So my question is this:

Relating to the issue discussed today, do you think there is a meaningful (lasting, permanent) split among economists over the issue of practicality and real life vs theory? If so, is the split analogous to what happened to political science in the 50's and 60's? My understanding is that Polysci used to be the study of actions and consequences...positive statements and scientific research...but it grew to include normative ideas about our political system, which included an expression of values. That was the new, modern direction the field took, and that evolution came about during the political and social turmoil of the 1960's. Some people embraced the new direction and some did not. The ones that did not became part of a minority camp within the field. Do you see a similar crossroads for econ?

All of this has implications for my future so I'd like to hear your thoughts. The more I know about what I'm getting into before I get into it, the better.

Thanks...sorry for the length...I suck at writing short (any)things.
...and here's my answer:
That's a good question. The positive vs normative split is well-known and disparaged in econ. Many thought that we should stick with positive econ, but we all know that normative econ is where the action is at (policy) AND where anyone with an opinion will go. That's why "unstated" biases are so dangerous and prevalent in economics...

That said, I would hope that reality would check thoughtless, inaccurate and misleading math econ, but the self-sustaining nature of mathurbaters (get a phd from a math-econ place; get hired by an m-e place, publish in m-e journals, train more students in m-e) means that it may persist (and perhaps HAS persisted) far beyond its useful (as in real-world relevance) life.

That's just sad, but it's mostly true.

Luckily, we have things like the nobel prize and reality to push back on such silliness...
What are your thoughts?

7 comments:

Eric said...

Interesting. Apparently my models (real world math that you may not like because it is complicated. It is complicated because it has to be.) include both positive and normative aspects.

I am just trying to model what actually happens independent of the reason and without assumptions about values. Values show up as constraints in the model not as assumptions.

africaeconomist said...

maths is great. it is just that the art of thinking before mathematical modelling and the art of translating the results into the real world is often not so great.

pairing real-world people to more abstract maths-people may be the way to go, if we can break our ingrained aversion twds each other..

Eric said...

@africaeconomist

Thinking is coming right now. I am trying to understand the interaction of technology, peak oil, population growth, insolation, levelized cost of energy, and average energy cost. It need thoughts that are clear enough, that I have a decent chance to be right.

Modeling comes later.

Then comes pairing. I would enjoy maths or real world people. I am apparently both and can talk to either group effectively.

Anonymous said...

Thanks for commenting on my note to Dr. Zetland.

I think math is an important part of economics and I respect those who can formulate complicated mathematical models, but I think
math itself is limited by nature. A mathematical model, if it has any hope of being accurate, must account for all variables that may influence a result. It must also accurately represent the influence each variable has on other variables, and must reflect that influence in the final result. This is great for science and physics where we know, for example, how to calculate trajectory of an object based known variables. It also makes sense in some economic situations where all variables are known.

However, economics deals not just with numbers but with people. You can model the decisions people make, but you might be hard pressed to predict those decisions with a model. I think this speaks, at least in part, to the root of why I have issues with math in economics and why I think it has limitations.

There was a hedge fund not too many years ago (I forget the name) that came up with a quantitative model of world markets that worked for a while, and brought in tens if not hundreds of billions of dollars. All the main players on wall street invested heavily, and so did banks, pensions, university endowments, etc. In the end, the model crashed and investors lost everything. The model crashed for the same reason I think mathematical modeling has severe limitations generally: because you can only model something for which you know all the inputs, and therefore models are based on the past. The future, thank goodness, is not static. A static future is good for some...bad for many.

Perhaps the question "What is economics?" is at the root of the issue for me. I tend to focus on the big picture because that is where most interesting questions ultimately lead. I see economics as the study of how people interact not only in financial markets, but in life. The end game - or highest purpose if you like - of economics is to understand the dynamics of human interaction - both in markets and otherwise - and to use that understanding to make decisions that make our collective lives better in the long run. It's the ability to see the difference between a responsible decision and an irresponsible one. It's about doing the right thing given our values, versus the wrong thing. It sheds light on actions and consequences, and serves to clarify and define my decisions as well as a basis from which to defend them. It does, for me, include normative statements of "should". Perhaps economics is a way for me to define, explain and defend my values. Since this is ultimately where policy is made, I think that must be its highest purpose...to shape public policy.

I respect those who can understand complicated math, but it's not my thing. I am not likely to end up crunching numbers in the "demand elasticity" group at a company, and I am not likely to have long conversations with future Professors about the beauty of a particular mathematical equation. Does this mean that I have nothing
valuable to contribute to the conversation within the study of economics? I am confident that the answer is not "yes".

Andrew

Eric said...

@Andrew,
I agree with your normative statements. If you listen well and pay attention to what others are saying, you have a lot to contribute.

People like me can convert your knowledge in equations--not pretty ones but ones that work.

The math that might have a hope of avoiding the Long Term Capital meltdown is brand new.

As to knowing all the variables, that is not really needed. The actual question for a model is does it work in the real world. Approaches to making predictive models without all the data come under the statistics of decision support. As you said, however, a model builder needs to be continually humble and to fix things that don't work. In part, I am an engineer. So fixing things that are broken is part of my mind set.

James Taylor said...

You all miss the point. Have you not noticed all Nobel prize winners in economics are academics? Have you not noticed that no one who actually has to use economics has never been nominated?

The divide in economics is totally academic, and has no relation to reality. It is all about department politics.

Trying to build mathematical models of human behavior, and especially mass behavior is simply silly.

On the other hand, talking economists like, say, Krugman are all caught up in their theoretical models that are also simply silly.

If you are going to get promoted in the department, then you are going to have to publish so you make a choice. Are you more comfortable with numbers or words?

End of story.

Eric said...

@James
You clearly have opinions!
I agree with a number of them.

If you have it, I would like to see your support for the items that you label 'simply silly.'

If you have really good support, both Krugman and I would like to know about it.