Thanks to everyone who commented on my views. I very much appreciate the discussion. I am about to throw myself heavily into moving ahead on a book I am working on, but I thought that before I did so, I would respond to the latest set of comments. I will likely not to respond to the next set if there are any, because it is too much fun, and it draws me away from the drudgery of writing my book. 



Re Sumitra:
 Sumitra, writes that Mill saw policy as the ultimate goal of science. I don't deal with ultimate goals, but I don't think it matters what the ultimate goal is--I still favor freeing scientists from policy, although I fully agree that thinking about policy and the way things work and do not work can often stimulate good science. In a complex world, we don't know what knowledge will be useful for policy or not, and the way breakthroughs are made in thinking for policy is to let people think about issues in a totally abstract way without regard for policy--simply for understanding. Some of those thoughts may have implications for policy, but to make them have it too soon leads one not to ask the creative questions one should. Example--macro--as macro developed to make the macro models have policy implications, the economics profession didn't pick up on those economists such as Goodwin who were working on interpreting Keynes' macroeconomics as leading to a study of systems with complex dynamics where turbulence could lead economies to an unattractive basins of attraction. Instead, macroeconomists studied models with linear dynamics to keep the models more policy relevant. This meant that macro theory never really developed the complex models that would be able to incorporate a financial crisis. So I don't think Mill and I are necessarily saying anything different. I just argue that the best way to get creative thinking about policy is to encourage theorists to think about systems without having to worry about policy.  

Most economists are not theorists--they are applied economists or teachers of economics--I prefer not to call these "scientists" because it leads people to think of scientific methodology, and to define what they are doing in reference to science. I think it is far better for applied economists to think of themselves as engineers--and think of them as developing heuristics--general rules for guiding policy. These heuristics are based in science but are  not strictly speaking scientific rules. Developing these heuristics involves integrating insights gained from economic theory and empirical work with those gained from history, past writing, and a knowledge of institutions, and thus requires a broader education than does preparing young people to become scientists. As long a people think they are training scientists, they can justify the existing training (although even then I don't find their justification convincing) and current research practices as doing science. 

I encourage people interested in this issue to take a look at  Billy Koen's work on the methodology of engineering and heuristics, and to ask whether his description of engineering methodology better fits what most economists do than does the methodology of science. Engineers don't call themselves scientists, and the don't call their knowledge scientific knowledge. They solve problems using science in whatever manner works, and the leave a large fudge factor to account for their lack of scientific knowledge. 

Re James:
	James writes: "Dave's definition of science as the statement of how things work should lead to his evaluating his alternative theories of inflation by empirical evidence.  Indeed, careful deductive reasoning affirms the quantity theory of money as the reliable principle from which to account for inflation.  I think historians of economics invite their being ignored as practitioners of economic science by their refusal to recognize that."
	I think James and I are using different conceptions of the term science. When I am talking about science, I mean something that can be rigorously defined, measured and scientifically empirically tested in a way that meets the current scientific heuristic. All that must be done before it becomes scientific knowledge.  As a scientific fact, I'm not sure what money is, so I'm not precisely sure what the quantity theory of money is telling us. We just don't have the empirical data to test quantity theories in a scientific way. We do have  heuristics about monetary control and inflation and those heuristics include controlling some variable that has moneyness, but these heuristics are contextual and depend on the situation--changing institutional structures can change the heuristics. 

Re Roger.
He argues "that the effect of introducing the science/art distinction is to provide an additional argument with which to defend a narrow view of economics."  My view is that is does exactly the opposite, by pointing out that what 98% of the researchers do is much better described as art, that is best described as having a methodology similar to engineering rather than science, one is not providing an argument for a narrow view of economics--it just provides a narrow view of economic science. Using the term art or engineering to describe what most economists do stops people from protecting their work and views under the mantle of science. By stating that clearly, one can align oneself with the true economic scientists out there, who are often as concerned with economic findings being called science as I am.   

Re Pat:
Pat writes "I believe that David's history of economics leaves out a large chunk of economics. His recommendations regarding how to use the phrase "economic science" reflect a gap in his treatment of the history of economics. It is not ideological to take account of the classical observation that the division of labor tends to expand productivity or of the early neoclassical transformation of that observation into the view that the action of the entrepreneur role tends to be in the interest of the consumer role. It is an acknowledge of the reality discovered by our predecessors. The choice is not between ideology and science but between ignorance and knowledge." My view is that Classical economists had a much broader theory than did neoclassical economists--and dealt with issues that went far beyond what became neoclassical models. But it was all informal--the Classical science of economics never developed, because it got waylaid by a "neoclassical" revolution that, in part because it wanted to relate economics to policy, and in doing so stopped doing creative work in theory that would start to incorporate the complexities of increasing returns, feedback effects and non-linear dynamics into our general understanding. Economists such as Hayek, Veblen and Boulding recognized this, but they were not followed. Instead those who wanted to make economics policy relevant were followed. 

Re Ric
Ric writes: "If heterodox economists push the distinction that Dave makes and really claim that most of what economists do is not science, but something else they are dismissed as not being "serious" economists as Paul Krugman would say. So in many ways heterodox economists have had to play the game of doing economic "science," if they are to be taken seriously or even to advance in the profession.  When we point out that what many traditional economists do is not "science" than we are looked at as being quacks."  I agree that the science rhetoric of often played, and as I have written elsewhere, I believe that a good strategy for heterodox economists doing theoretical work is to work jointly with mathematicians,  physicists and statisticians. There are gains from trade to be had there, as many of them share the concerns that heterodox economists have. Work in complexity theory is moving our theory forward---it is bringing feedbacks and network effects into the core understanding. Epistemic game theory is bringing back a consideration of norms, and the limits of traditional econometrics is being recognized by economists such as David Hendry.   


Re John:
In response to my use of Viner's "economics is what economists do" John writes: It seems to me that such a definition is circular at best, and dangerous at worst. For example, in the documentary "Inside Job," economists took large fees to validate their clients pre-determined views. Thus, economics is a fraud in which hucksters give a pseudo-scientific cover for thieves. I don't think we want to say that. For me, a definition is not about what the professionals do so much as about what they ought to be doing."

Our views here differ significantly. I specifically do not see economics as a fraud in which hucksters give cover for thieves. I see the large majority of them as thoughtful people trying to do the best job they can. I have been active on the ethical code issue, and I do not believe that Inside Job showed that economists took large fees to validate their client's predetermined views. 

In my view the ethical issues involved are far more complicated than any such characterization. In my view, the views that Mishkin and Hubbard expressed had little to do with the money they received for their writings. The problem is deeper than money. The problem is that they fully believed what they were writing, and they believed it with the conviction of a scientific understanding, when at best I could see them believing it as an engineering heuristic. What's the difference--engineers leave a large fudge factor to account for what they know they don't know. Thus if theory says-build a wall 4 inches thick, their heuristic will be to  build it 6, 8, or even 10 inches thick to account for known unknowns.  The problem with economists thinking they have scientific knowledge is that they don't build in the fudge factors into their policy advice 



David Colander
CAJ Distinguished Professor of Economics
Department of Economics
Middlebury College
Middlebury, Vermont, 05753
(802-443-5302)