Monday, November 21, 2005
Feeling More and More Hetero
A while ago, a stack of books in a Strasbourg bookstore caught my eye: a collection of essays dissenting from mainstream economics. I read it with interest, although I found the quality of the arguments very uneven. I dislike the banner "Post-Autistic Economics" which the movement bears, after its origins as "Autisme-Economie"; I much prefer "Heterodox Economics".
As an operations researcher, in a field which has specialized in techniques originating in mathematical economics, I am well aware of the many limitations and inadequacies of "optimization technologies". These include tractability of formal models, but many, many others: one solves a model (if one can) but a model is an incomplete and selective representation of a reality; there may be multiple criteria of success which are incommensurate (or worse, as Arrow proved, no consensus on preference ranking trade-offs); they may be subject to incomplete and noisy data.
Different stakeholders may, and usually do, have different perceptions of what the situation is, what the options are, what the future should be. This is not new, nor solved, in economics. What is the "right" view/decision/outcome is relative (wasn't it W.V. Quine who pointed out that "anything can be held true if we are willing to make sufficient adjustments elsewhere in the system"?). Pareto wrote about "ophelimity", a word he coined to extend the notion of "utility" to other desires than the purely useful, to which Fisher (Is "Utility" the Most Suitable Term for the Concept It is Used to Denote?, 1918) preferred "wantability". Price indexing and measures for cross-country comparisons of quality of life remain problematic, to say the least.
Operations research was divided by these issues, in the 1970s if not earlier. Cooke demonstrated that some problems might never be tractable ("NP hard"), and Karp and others soon added many problems of equivalent hardness to the list. "Optimization" is now reserved for cases where money (cost, profit) is the only criterion of performance. Constraint satisfaction, data enveloppement, and multi-objective (Pareto) methods facility discovery of the options in cases which cannot be reduced to a single quantitative criterion. Alternative approaches for considering organizational performance include Balanced ScoreCard, Team Syntegrity, Reengineering, Soft Systems Methodology, although not all of these would be considered descendants of O.R. by practitioners. The schools that grant degrees in O.R. now specialise in combinatoric optimization, with probabilistic extensions for the more sophisticated of them (Financial Engineering at Princeton, e.g.). Much as a mathocratic approach dominates in economics, operations research has specialized in computing of models, and abandoned the epistemological challenges I (and others?) believe essential to its effectiveness.
Now, there has been an uprising in economics, particularly of economics grad students, who insist that economics reduced to standard mathematical models discards a lot of knowledge that economists should have. I wish the same were true in O.R., but the field is so much smaller, newer, and ill-defined, that I fear it has already passed the tipping point. Inasmuch as O.R. as it was taught to me was mainly economic and in the service of production, anyway, I hereby proclaim myself an economist, and a heterodox one at that!
Could all this (above) be recast in a medical analogy? Diagnosis, prognosis, treatment are all part of medical praxis; heck, I read somewhere that in some cultures one pays the doctor while one is healthy, not ill--a healthcare system rather than the sickness care system predominent today. Be that as it may, if there were only, say, surgeons or pharmacists, we might have the "best of" treatment for many ailments. But we would surely be missing the benefit of the diagnoses and prognoses of general practitioners and specialists. I fear that this is what has in fact happened in economics (and its offshoot, O.R.)-- the pharmacists have taken over. The obsession is with the choice of treatment and consideration of possible side effects; the pursuit may be reactive (take your medecine) or "proactive" (take your vitamins), but rarely or never projective and wholistic [should provide phrases for these last two, too, but they are not coming to mind].
Feeling more and more hetero(dox economist), until I'm ready to found hetero(dox) scheming research, would I fit neatly into one of the specialties? Do I even understand what they comprise? According to the heterodox economics portal...
Researching my affinities with each of these should keep me busy for a while!
As an operations researcher, in a field which has specialized in techniques originating in mathematical economics, I am well aware of the many limitations and inadequacies of "optimization technologies". These include tractability of formal models, but many, many others: one solves a model (if one can) but a model is an incomplete and selective representation of a reality; there may be multiple criteria of success which are incommensurate (or worse, as Arrow proved, no consensus on preference ranking trade-offs); they may be subject to incomplete and noisy data.
Different stakeholders may, and usually do, have different perceptions of what the situation is, what the options are, what the future should be. This is not new, nor solved, in economics. What is the "right" view/decision/outcome is relative (wasn't it W.V. Quine who pointed out that "anything can be held true if we are willing to make sufficient adjustments elsewhere in the system"?). Pareto wrote about "ophelimity", a word he coined to extend the notion of "utility" to other desires than the purely useful, to which Fisher (Is "Utility" the Most Suitable Term for the Concept It is Used to Denote?, 1918) preferred "wantability". Price indexing and measures for cross-country comparisons of quality of life remain problematic, to say the least.
Operations research was divided by these issues, in the 1970s if not earlier. Cooke demonstrated that some problems might never be tractable ("NP hard"), and Karp and others soon added many problems of equivalent hardness to the list. "Optimization" is now reserved for cases where money (cost, profit) is the only criterion of performance. Constraint satisfaction, data enveloppement, and multi-objective (Pareto) methods facility discovery of the options in cases which cannot be reduced to a single quantitative criterion. Alternative approaches for considering organizational performance include Balanced ScoreCard, Team Syntegrity, Reengineering, Soft Systems Methodology, although not all of these would be considered descendants of O.R. by practitioners. The schools that grant degrees in O.R. now specialise in combinatoric optimization, with probabilistic extensions for the more sophisticated of them (Financial Engineering at Princeton, e.g.). Much as a mathocratic approach dominates in economics, operations research has specialized in computing of models, and abandoned the epistemological challenges I (and others?) believe essential to its effectiveness.
Now, there has been an uprising in economics, particularly of economics grad students, who insist that economics reduced to standard mathematical models discards a lot of knowledge that economists should have. I wish the same were true in O.R., but the field is so much smaller, newer, and ill-defined, that I fear it has already passed the tipping point. Inasmuch as O.R. as it was taught to me was mainly economic and in the service of production, anyway, I hereby proclaim myself an economist, and a heterodox one at that!
Could all this (above) be recast in a medical analogy? Diagnosis, prognosis, treatment are all part of medical praxis; heck, I read somewhere that in some cultures one pays the doctor while one is healthy, not ill--a healthcare system rather than the sickness care system predominent today. Be that as it may, if there were only, say, surgeons or pharmacists, we might have the "best of" treatment for many ailments. But we would surely be missing the benefit of the diagnoses and prognoses of general practitioners and specialists. I fear that this is what has in fact happened in economics (and its offshoot, O.R.)-- the pharmacists have taken over. The obsession is with the choice of treatment and consideration of possible side effects; the pursuit may be reactive (take your medecine) or "proactive" (take your vitamins), but rarely or never projective and wholistic [should provide phrases for these last two, too, but they are not coming to mind].
Feeling more and more hetero(dox economist), until I'm ready to found hetero(dox) scheming research, would I fit neatly into one of the specialties? Do I even understand what they comprise? According to the heterodox economics portal...
The heterodoxy includes (but is not limited to)
- Austrian economics,
- Behavioural economics,
- Black political economy,
- Ecological economics,
- Evolutionary economics,
- Feminist economics,
- Georgist economics,
- Historical economics,
- Institutionalism,
- Marxism,
- Post Keynesian economics,
- Postmodern economics,
- Postcolonial economics,
- Rhetorical economics,
- Social economics and
- Sraffian economics.
Researching my affinities with each of these should keep me busy for a while!
# posted by Maurice Lanselle @ 4:00 PM 
