by From Gerald Holtham (originally a comment) The biggest problem, it seems to me, is not the tools that economists use but the failure to start from real, concrete questions and issues in the real world. Someone builds a model of a real situation. It may be useful or not but if it is ingenious it gets published. People then play tunes on the model; they seek to “generalise” it in various ways. We embark on a process of theoretical development for its own sake, increasingly divorced from the original problem and all too often from any real application at all. The usual destination is then sterility. I have seen mathematical queuing theory fruitfully applied in the study of development aid and mathematical game theory successfully applied in auction design. The secret of success was that
Topics:
Editor considers the following as important: Uncategorized
This could be interesting, too:
John Quiggin writes RBA policy is putting all our futures at risk
Merijn T. Knibbe writes ´Extra Unordinarily Persistent Large Otput Gaps´ (EU-PLOGs)
Peter Radford writes The Geology of Economics?
John Quiggin writes Suggestions for a small experiment
from Gerald Holtham (originally a comment)
The biggest problem, it seems to me, is not the tools that economists use but the failure to start from real, concrete questions and issues in the real world. Someone builds a model of a real situation. It may be useful or not but if it is ingenious it gets published. People then play tunes on the model; they seek to “generalise” it in various ways. We embark on a process of theoretical development for its own sake, increasingly divorced from the original problem and all too often from any real application at all. The usual destination is then sterility.
I have seen mathematical queuing theory fruitfully applied in the study of development aid and mathematical game theory successfully applied in auction design. The secret of success was that the axioms of the model were ad hoc – designed for a real, specific situation that was successfully analysed.
An example of the opposite process is Dornbusch’s application of rational expectations theory to exchange rates. It was a neat, ingenious model and helped to create a fashion that made it obligatory to specify rational . i.e. model-consistent, expectations in every application. Few seemed to notice that the Dornbusch model implied exchange rates would move in a saw-tooth pattern of sudden abrupt moves and gradual reversions while in fact exchange rates tend to follow a more symmetrical cyclical pattern of ups and downs. The reason is Dornbusch assumed similar agents with similar information sets. In practice we have diverse agents with diverse information sets generating endogenous as well as exogenous uncertainty. By ignoring reality and concentrating on development of “the theory” macroeconomcs went mad and led to the sterility that pfeffertag notes.
Deductive logic is a necessary part of the tool kit and so is mathematical modelling but you have to tailor your starting assumptions to a specific situation and keep your eye on the real world. You also have to accept empirical refutation when it smacks you in the face.