Tuesday, 9 July 2013

Economic History and Krugman’s Crib Sheet

One of the positive things about reading blogs is that sometimes you see connections in apparently diverse offerings. So here are two seemingly unconnected posts: Paul Krugman’s discussion of how he came to do his path breaking research in international trade and economic geography, and Kevin O’Rourke’s post on why economics needs economic history.

I remember many years ago being in a large interdisciplinary forum, where Krugman’s research on economic geography came up. The economists in the room were of course very positive, but the geographer there could not hide his disdain. There is nothing in this work that geographers have not actively discussed for the past 50 years, he said. I have no reason to doubt that he was right, but it kind of missed the point. What Krugman and others did was manage to formalise these earlier thoughts in a particularly tractable and useful way.

What is so great about formalism, you might ask. The trouble with just talking and writing about the way the world works is that it is quite easy to become confused or to make mistakes. Macro, because it deals with a highly interconnected system, is full of these pitfalls. The example I use with undergrads when they first come to IS/LM is as follows. Cutting taxes may appear to boost the economy, but if it is financed by more government borrowing, to persuade people to lend more will probably push up interest rates. These higher interest rates reduce output, so as a result tax cuts could end up reducing output. Sounds reasonable, but the reasoning is incorrect. The worst that can happen with a tax cut is that people save it all, in which case output does not change, and neither do interest rates. IS/LM shows us that if interest rates rise it is because output has increased.

So it is good to be able to express ideas about how the world works in terms of simple models. But creating a new type of model for the first time is not an easy thing to do, which is why you get prizes for this kind of thing. Crucially, it may take a long time (decades or more) before someone comes up with that nice simple formalisation that captures those ideas. Yet those ideas are as important before the formalisation as they are afterwards - it is just the reasoning about them that has improved.

How is economics generally taught? In both macro and micro, most of the time we teach the formalisation. This is understandable (it is what has advanced the discipline and made it science like) and to a degree appropriate (understanding models is difficult). However there is a real danger that teaching this stuff crowds out all else. I used not to be concerned about this for macro, because I saw the discipline as inherently progressive, where the data would naturally push advances in the right direction. (My excuse for believing this in part comes from my background in building structural econometric models, where the data really did do that.)

If that is your view, you are likely to be a little dismissive about things like economic history, economic methodology or the history of economic thought. After all, most scientists do not worry too much about these things in their own discipline, and economics tries to be like a science. Even if we take a more realistic view, and think that economists are more like doctors (who fail to understand quite a lot), doctors do not spend too much time thinking about things like the methodology of medicine.

I changed my view in the last few years as a result of both the financial crisis and the subsequent domination of austerity policies. Teaching just what can be currently formalised in what now passes as a rigorous manner excludes too much of what is important. Of course we (hopefully) tell students that there are gaps in what economists can do this way, but perhaps these gaps need to be given a little more space than footnotes. There is a great deal of knowledge and insight in less formal economic reasoning, insight that can too easily be dismissed. Unfortunately it is natural for future academics or policy makers to believe that what is taught in undergraduate or graduate macro is what is important, rather than what has so far been formalised, or what the demands of this particular time and context require formalising, or worse still what political or ideological forces wish to formalise.

The analogy with doctors breaks down because, unlike doctors, an economist does not constantly have the full range of empirical problems thrown in their face. They are also unlikely to have politicians picking and choosing which treatments they like to promote based on the interests of those they serve. In particular, developments in macro over the last few decades have shielded economists from having to explain much of the data. (In my view the dismissal of single equation time series work as a vital component in model building because of identification problems was a crucial mistake.)

As Kevin O’Rourke eloquently argues, teaching economic history provides a useful counterweight to these tendencies. We also need to make room for teaching some elements of methodology and the history of thought, for similar reasons. This is why I have actively supported Diane Coyle’s initiative in the UK (see here, here and here). I think having the occasional option in these subjects misses the point. It is much more about integrating these elements into core courses, although how best to do this remains an open question.





16 comments:

  1. "How to teach economics by Simon Wren Lewis" qouted
    at Lars P Syll blog
    http://larspsyll.wordpress.com/2013/07/09/how-to-teach-economics/

    ReplyDelete
  2. The complaint by at least some geographers carries serious weight. PK loudly claimed that he was the one who brought "formalism" to the analysis of economic geography. This is simply false. See the paper by Masahisa Fujita in 1988 in Regional Science and Urban Econmics, which applied Dixit-Stiglitz to econ geog before PK did so in his widely cited 91 JPE paper, which failed to cite Fujita's paper. Then there was a rather long string of people whom you have probably never heard of publishing in physics, geography, and regional science journals, and even occasionally economics ones such as Econometrica, who did so as well throughout the 80s, but who to this day have never been cited by PK. A partial list includes Y.Y. Papageorgiou and T.R. Smith, Tonu Puu and Martin Beckmann, Peter Allen, Wolfgang Weidlich, Gunter Haag, and Roger White. While many have claimed that PK is the emperor of the new economic geography, he is one without clothes. Sorry.

    ReplyDelete
    Replies
    1. I absolutely second that. This is a pure case of self-inflicted blindness to what has been done outside of economics. To a large extent Krugman really only reinvented the wheel that economic geographer had worked on decades before him. But Krugman of course suits Simon's argument well, since he was "able to express ideas about how the world works in terms of simple models" by "creating a new nice simple formalisation." But there is a heavy price to pay for this gain - much of the realism and relevance of the economic geographers' models got lost in the idealizations that have to be made constructing simple formalised models.

      Delete
    2. You know Fujita and Krugman wrote a book together on economic geography, right?

      Delete
    3. Yes, later, anonymous. Go read Fujita's discussion of PK's work in his piece in RSUE after PK went to Stockholm. It is very carefully worded and praises his work, but it also makes clear that others did crucial work prior to PK.

      I shall quote, anonymous, since you want to make a fight of this, from Fujita and Thisse, 2010, RSUE, in the abstract:

      "Paul Krugman has clarified the microeconomic underpinnings of both spatial economic agglomerations and regional imbalances at national and international levels. He has achieved this with a series of remarkably original papers and books that succeed in combining imperfect competition, increasing returns, and transportation costs in new and powerful ways. Yet, not everything was new in New Economic Geography."

      Barkley Rosser

      Delete
  3. BTW, just for the record, I am somewhat who mostly agrees with PK regarding current macro policy debates, but the record on what went on in economic geography has been ridiculously and wildly mythologized.

    ReplyDelete
  4. Krugman uses the baby-sitting slips 'model' for QE and why it wouldn't produce inflation, and if I remember rightly he called the Wall street and City model makers pre-2008 'nerd savants' for their belief that the maths would predict all.

    Robert Shiller's column 'A Metaphor for Obama' talks about the impact that language has on economics.

    And if you look at page 18 of WEALTH EFFECTS REVISITED 1975-2012 Karl E. Case, John M. Quigley and Robert J. Shiller, with its graph of home prices in California 1975-2012 and its 3 bubbles, you must ask yourself how all but a whole profession could look the other way at that irrational rollercoaster of prices.

    ReplyDelete
  5. Very interesting post, Simon. In my view, economics has gone down the wrong path by trying to ape the natural sciences and treat the subject as a form of applied mathematics. The level of formalism would only be justified if it were based on realistic foundations (as opposed to the patently unrealistic foundations of utility maximization, rational expectations etc.) or consistently yielded accurate predictions. If two models both give broadly accurate predictions, say New and Old Keynesian models in a liquidity trap, then Occam's Razor suggests the simpler (Old Keynesian) model should be preferred. Instead, economists insist on using the more complicated formal model, as it appears more scientific.

    At its worst, this mentality leads to economics becoming a pseudo-science. Too often, evidence is ignored unless it is presented in the form of econometrics or calibration (which can both be highly misleading) and economics exists in a sort of bubble, where hypotheses that are almost certainly wrong are not refuted and instead are praised for their rigor.

    So, I agree that economists should think much more deeply about methodology and the historical evidence, and this should be an integral part of economics education.

    ReplyDelete
  6. I am not an economist - just someone who has been reading a range of econ blogs since the crash to determine how the financial sector (ground zero for capitalism) trashed the economy.

    I would urge all economists to engage in a study of Shakespeare - people are emotional beings and respond emotionally. Look at the stock market on a day when there is unexplained blip in data - panic ensues! Sky is falling! Sell! Sell! Sell!

    And look at how the political beliefs of economists drive their models. (Mankiw's defense of the 1% is one such paper that is filled with political, not rational, explanations for why we are where we are today.)

    The idea of "rational markets" is highly irrational. To build "formal" models on the idea of rational markets, the invisible hand, etc. is mystifying.

    ReplyDelete
    Replies
    1. See, this is exactly one of those cases in which knowledge of the history of the discipline would help understand what you see before you. Math has to be tractable in order to do any good. Knowledge tends to build up in granular fashion. Put those two truisms together and the reason for starting the effort at building models from utility curves and perfectly competitive markets is not mystifying. Come to the economic profession in the early 2000s innocent of what came before and the blinders worn by the Chicago types (which in this context means nearer everybody than it does economists who worked and studied at U Chicago) look pretty mysterious. Simple-minded fresh-waterism is largely the result of mistaking simplifying assumptions for reality.

      Delete
  7. I would urge all economists to engage in a study of Shakespeare - people are emotional beings and respond emotionally. Look at the stock market on a day when there is unexplained blip in data - panic ensues! Sky is falling! Sell! Sell! Sell!

    ReplyDelete
  8. To Anonymous of July 10 05:07

    I'm not sure why you point to the economic profession in the early 2000s as the ones with the blinders on. When speaking before Congress in 2008, Alan Greenspan, who emerged as a leader in economic thinking long before GW Bush's reign, acknowledged the crash led him to discover a "flaw" in his thinking, a flaw that led our economy off a cliff.

    "'Those of us who have looked to the self-interest of lending institutions to protect shareholders’ equity, myself included, are in a state of shocked disbelief,' he told the House Committee on Oversight and Government Reform."(http://nyti.ms/1b0wW2x)

    Was he not aware of Continental Bank bailout in the Reagan era? Wasn't he involved in helping to solve the savings and loan scandal of the GHW Bush era? Didn't he read "Liar's Poker" - written in the late 1980s about a culture where self-interest was focused on maximizing the bonus, not shareholder value?

    History is there to consider and learn from - it is ignored by those who hold the power to actually change things.

    And again, look at Mankiw's paper. It's filled with astonishing statements, like when he claims the children of the chair of Harvard's economic department/Romney advisor have no more opportunity than he had as a young man, the child of people who had no college education. How blind! But such blindness is required to fit ideals into models that absolutely do not reflect the reality of a dynamic economic environment.

    ReplyDelete
  9. Completely agree with this. I'm sure great economists like Adam Smith and Keynes, whose work was thoroughly grounded in the real world and who were so strongly influenced by the wider insights of ethics, history and politics, would be horrified at the mathematical and almost entirely theoretical/model-based nature of modern mainstream economics.

    I've been amazed that it's possible to do undergraduate macroeconomics without even a basic outline of economic history- I'm sure this partially explains why we fail to learn from and repeat mistakes from events like the Great Depression and the Oil Crisis. I've also been amazed that microeconomics has basically no interest in empirical information. It's breathtakingly arrogant to think that, for example, it's possible to get any real insight into competition policy based entirely on models based on shaky assumptions, without looking at how competition policy works in practice at all.

    I think economics is one of the few disciplines where something like RBC models can still be seriously taught to undergraduates because of this arrogant disregard for the real world, in a phoney and futile attempt at 'science'. It's sad that even as significant an event as the financial crisis doesn't seem to have led to a major change in the way academic economics is taught or practiced.

    ReplyDelete
  10. The problem as far as I see it is in the nature of economics. It is not an exact science (cannot be as the reality to be researched is too complicated), but it needs to use the methodology of exact science to present a framework for the field.

    1. Hard to see how to get structure in the topic in other ways than via formalisation. With furthermore as an extra disadvantage of doing it in another way that nothing can be really quantified.
    In that respect both for teaching and use in real life formalisation seems to be the by far best framework for the field.

    2. However the field seems often to forget that the framework itself is also an 'approximation'. It is not an absolute science and likely will never be.

    3. Problems with the formalisation. Mainly simple stats.
    -Samples sizes are small (dealing usually with countries), anyway leaves high sigma.
    -Samples are (might be) different from each other (or the whole thing compensates for that), you have to assume they are similar to be able to do stats.
    -Usually normal distribution is used. Which simply doesnot look always the best method.
    In general things in reality can differ substantially from the formalized norm.

    4. This is usually not properly taken into consideration.
    This is the main weak point.

    A check prior to using the formalised stuff to see if it is applicable in this particular situation is not done. Same with the outcome, if the outcome is possible/relevant in the real world. Simply not done.

    5. As said those pre and post formalisation checks look in the present situation simply necessary but are not made.
    How to make such checks?
    Basically looking at a particular case in the way historians do might be a good method. I (not that that count for much) do not see any better one.
    Students are confronted with and trained in new way of thinking (next to the formalisation) and the field has to place the stuff in the real world (and see where its weaknesses are).

    So both are useful. Formalisation loks to be the by far best way to create a framework and history looks to be (but not properly discussed) to test method and outcome against the real world.

    NB Re the Doctor part.
    Also in medicine it is so much easier when a patient only has one disease. Another problem in current economics is that nearly all are patients and all have more than one disease.
    Continuing from Tinbergen: if you want effective cures you only need to cure one thing.
    NB2 A good lesson may be for politicians to let things not run totally out of control (as they have done) and react much faster on unbalances.
    That Stigy guy wants to extend the field to debt. Imho it should be extended to unbalances in general relavant for the economy. Wider than only debt. Several unbalances are not debt related or dealt with in traditional Macro, but are still highly relevant.
    A more Buddhist style economy: middle road kind of stuff. Or more Tao: the 'there should be a balance between Yin and Yang thingy'.
    NB3 At present a lot of the economy is determined by ad hoc short term political decisions, which make working systematically nearly impossible. Good to bear in mind.

    ReplyDelete
    Replies
    1. "1. Hard to see how to get structure in the topic in other ways than via formalisation."

      Ask a historian. History has FAR more rigor than econ (ever taken a history course? They are HARD.) And no formalizations at all.

      Ask an archaeologist. They have formalizations, but they're subtle and vague, more classification tools than anything else. And their work is very rigorous.

      Ask an anthropologist. Anthrolopology has a lot more rigor than econ, and very few formalizations.

      Ask a psychologist. They actually do have some formalizations, thanks to extensive developments in biology, but they are rarely arrogant enough to believe the formalizations, after the debunking of multiple "explain-everything" schools of psychology from Freud onward over the decades.

      By the way, this is what "interdisciplinary" studies are for. They're for when your subject is flailing around, missing the bloody obvious, and the people in the next building over have figured out how to solve the problem.

      Delete
  11. Rik is quite right about pointing out that economics is not "an exact science." As Thomas Sowell has pointed out, there are the "hard" physical sciences like math, physics, and chemistry; and there are the soft sciences like political science, sociology, and even economics. The distinction can be seen that when you are dealing with inanimate objects, like atoms, numbers, and planetary motion, there are certain "laws" that govern activity. When dealing with human beings, there is little certainty about anything. And the basis of all economics is what the individual citizens are doing!

    If we admit to this dilemma, we can see that postulating universal axioms for economic activity is, if not futile, at least subject to wide ranges of predictability. Compared to astronomy, an economic model would be akin to knowing that the earth revolves around the sun somewhere between one and ten times a year, depending on a variety of unknown forces that vary every few months!

    However, since human nature and behavior has been fairly unchanged for at least a few thousand years, it is possible to examine past experience and see how different economic/political systems worked. Thus, by using the case method in tandem with economic history, one can discover the conditions that accompanied the few truly prosperous societies during the past 4,000 years. And, conversely, those more numerous societies that languished in history can reveal the economic/political conditions that were not conducive to prosperity.

    Such work can formalize at least some general principles that govern the design of the most beneficial institutions, as well as reveal the socio-economic arrangements that retard economic vitality. This may be the most that the "science" of economics can be expected to give us--general principles. Those individuals with the hubris to overstep that limitation may very well continue to do more harm than good.

    We have seen in recent years the coming and going of different abstract economic theories and the failed nature of most "modeling" attempts which should warn us about their too frequent adverse and uncertain outcomes. And the profession's efforts to "fine-tune" the economy have certainly been of dubious value. During the past 40 years, while top economists debated the money supply, inflation, fiscal and monetary policy, a series of monstrous financial debacles recurred over and over again. The savings lost by Americans during the recent mortgage melt-down was a monumental failure on the part of the nation's economists. If you apply a case study approach to each of the past bubbles, panics, banking failures, and recesssions, it becomes clear that our managing economists are without clothes. This failure has been going on since the first land-rush bubbles and banking panics of the 19th century. It has been 200 years of failure to "manage" the economy. Until economic theorists come to grips with this record, new theoretical concepts and models will continue the failed record of the past.

    ReplyDelete