some econ literatures are still crammed with mutually contradictory models for which the scope conditions are neither known nor specified. And the stock of existing theories is still enormous. In some areas, especially in macro, economists really do have theories that make almost any prediction, with no real way to choose between them except priors and politics. And many economists still have very little problem using modeling assumptions that have already been taken to the data with discouraging results.
Pointer from Mark Thoma. In his post, Smith tries to “score” various criticisms of economists. His post made me want to recycle a quote from Herbert Stein:
1. Economists do not know very much.
2. Other people, including politicians who make economic policy, know even less about economics than economists do.
[typo corrected]
Non-economists are responsible for many of the critiques of economists to which Smith gives a low score.
I have come to believe that economics is epistemologically difficult. That is, it is difficult to answer the question, “How do you know that?” Non-economists do not have much insight into this issue. Unfortunately, many economists lack insight as well.
The appeal of the mathematical approach is that it provides rigorous connections between assumptions and conclusions. The weakness of the mathematical approach is that it places tremendous pressure on one’s choice of assumptions. And, as Smith has pointed out, these choices are more arbitrary than they are in the hard sciences.
Economists can almost never directly test their assumptions. Milton Friedman famously suggested not worrying about direct testing. Instead, he proposed the indirect approach of testing predictions. In practice, however, this does not work, or at least it does not work cleanly.
One problem is that you can have two interpretive frameworks that both “predict” one observed phenomenon yet have different predictions about other phenomena about which we do not have precise observations. Consider the vast array of candidate explanations for the financial crisis, with widely varying implications about how one might try to prevent a recurrence.
Another problem is that when an anomalous observation appears to confound an interpretive framework, this fails to result in a decisive rejection of that framework. Instead, the framework is tweaked in order to accommodate the observation. So, when the huge fiscal contraction in the United States at the end of World War II did not lead to another Great Depression, the explanation might be “pent-up consumer demand.” When the inflation rate failed to obey the Phillips Curve in the 1970s, the explanation might be “supply shocks” and/or “higher expectations of inflation.”
If assumptions cannot be tested directly, and Friedman’s proposal to test predictions does not work, how will assumptions be chosen? The answer, all too often, is a combination of mathematical tractability and faddism. Economists will jump all over a model because it is fun to play with, regardless of how silly or irrelevant the set of assumptions may be. The overlapping-generations model of money would be a prime example.
My main concerns with mainstream economics include:
1. A bias toward “engineers” rather than “ecologists.” That distinction comes from Greg Ip’s new book, Foolproof. The engineer is like Adam Smith’s man of system, who ignores evolution, both as a factor that may permit markets to over come their own failures and as a factor that may cause government “solutions” to become obsolete.
2. A bias toward simplifying the phenomenon of specialization. Macroeconomists live in a world with one producer and one consumer (the “representative agent.”) Microeconomists live in a 2x2x2 world, with two factors of production, two goods, and two producers. They miss important differences between those worlds and the real world of millions of tasks being performed to lead to a final product.
Stop giving engineers a bad name! What engineers do is write down their non-ideal assumptions and don’t pretend them away!
Economists are usually on firm ground when they observe a factor and then explain the effect of that factor on the market. Where they get into trouble is when they try to explain results. That requires an understanding of all the relevant factors and their relative weights, and that is usually impossible.
Attributing engineer like mentality is not an accurate description of certain Economists.
Engineers understand the schematics of the apparatus, they apply specific scientific laws to a structure they fully comprehend. They also study the underlying mechanisms before attempting a fix, of course they cannot afford to not have concrete results.
Engineers also do not look at correlations within metrics to do root cause analysis. Also, unlike economists, they tend to be more modest in their attempts to “engineer”.
My aphorism may apply:
Information is not Knowledge
Knowledge is not Wisdom
Wisdom is not understanding
It can be taken that Knowledge begins with the perception of information, followed by perceptions of connections of those bits of information producing meaning to the perceiver.
Thus some particular field of enquiry (economics, e.g.) may shape perceptions of information and the perceptions of the connections or relations of those bits of information to form what serves as Knowledge, but may be no more than information (at best) dependent upon how and why the connections have been perceived.
It is the problem of finding some “universality” of meaning in what is perceived.
Sorry –
Thus much of what passes for “normative” economics consists of determinations of what meaning certain information conveys (as knowledge).
Markets often share these biases or we wouldn’t have so much production occurring within companies.
Yes, the common implicit argument is:
“Economics isn’t a science, therefore my personal beliefs about economics are true”
Rather than drawing the correct conclusion of global skepticism.
Let’s say someone were asking someone else with an Arnold Kling-like outlook on Economics for advice on how to compose a paper that made the case for various interpreting recent macroeconomic phenomena through a different framework of explanatory interpretation.
“Here is a candidate for the root, primary driving phenomenon, and a narrative for how it is causing many of the other various, but derivative, phenomena we observe.”
A popular way of going about that, after introducing the idea in plain language, is to motivate the discussion with some simple toy Micro and/or Macro mathematical model that demonstrates the plausibility of some key relationship through manipulating equations, (“See, given the toy model assumptions, when the root thing goes up, all those other things always go down.”) and then trying to marshal up a lot of corroborating empirical evidence consistent with the model that bolsters the strength of the interpretive framework.
But would something like that get past the “Arnold Kling reviewer”? What would that reviewer think would be an optimal presentation of something like that?
I imagine most readers have realized this, but I’ll point out that Stein’s #2 is misquoted in the post, probably a typo, but it does change the meaning. It should be
2. Other people, including politicians who make economic policy, know even less about economics [not “economists”] than economists do.