Is Economics Science? Are Economists Scientists? The war continues
The long anticipated Pro/Con letters in the NYT were published and no one’s mind was changed.
Harvard Nobel laureate in Economics, Dr. Eric Maskin, had written a challenge letter last week affirming that the discipline of economics is no different from seismology or meteorology. Neither of those make predictions of any accuracy, so why pick on econ theory?
In response, professional economists said they are doing true Science while the others said not so.
This post is the follow up to our Economics can never be right and the point is, Economics is not just a non-science, the results from the discipline can never be exactly right the way a screw can be manufactured to exactly fit a threaded hole.
This box summarizes 4 successively more difficult ways to control a process. Each requires monitor stations for measurements to gauge the procedure. Measurement inaccuracy affects predictions of all 4 types of methods.
Type(1) Normal Statistical Process Control (SPC) techniques can assure extremely high quality in the end product, because both the raw materials used for input and the final products are measured; full communication occurs between all monitor stations.
Type (2) It is possible to generate new products that survive operational uncertainties using well analyzed destructive testing of many prototypes. Safety systems were developed for steam boilers and rocket engines compensate for turbine misbehavior. Measurement-based simulations are increasingly more important in Type (2) prototype development.
Types (3) These require in-operation monitors; post-Now feedback that is imperfectly simulated by past experience.
One class are those that might have been Type (2) but destructive testing is not feasible.
Another class are those systems with many mutually interacting parts … systems whose complexity can boarder on the chaotic (meteorology and seismology are examples). Predictions are necessarily of a statisticalnature, and success develops slowly, because events are infrequent and measurement error inevitable.
Type (4) These activities have all the issues of Type (3) but with many more interconnected subsystems or very strongly interacting elements. You rarely never have exactly repeated crises and no post-Now simulation is reliable feedback. Examples are Battlefield management and economics. These activities have no true control of events.
Responders to the NYT used meteorology and seismology as their standard comparisons. Both are classed as sciences; is econ suffering from discrimination?
This field started up as a science during WW-II to aid battlefield management. Over the decades the monitor grid has tightened to fractions of a kilometer across, all over the world. Meanwhile, codes have evolved to handle this tidal wave of data. The point is the more measuring points that could be used for input monitors, the more realistic the weather models could become. Currently, prediction accuracy is good for at least three days and reasonable for five.
Meteorology is a Type (3) science because it is a quest for quantitative understanding with predictions that can be disproved. Currently, there is no Theory of Weather but if/when it is developed it could predict – rain will fall on you 3 days hence, but never how many kilograms of water will strike your car.
Cal Tech’s Charles Richter along with Beno Gutenberg started the science in 1935 with an indicator of earthquakes using their specific equipment in their small region. It has since grown into an international effort to quantify ground shocks all around the globe, to understand their causes and, ultimately, to produce accurate predictions. Currently, the workers cannot tell you when an earthquake will happen, nor how big it will be, but are successful in identifying places with high quake susceptibility.
Seismology is a Type (3) science that is more difficult than meteorology due to the (relatively) few monitor stations and the points that should be monitored are deep underground and not directly accessible. Any Theory of Earthquakes with accurate statistical predictions is for the future. With fewer available data sets, results are more tentative, with schools of thought for competing models; progress is slow.
Economics is a Type (4) process, similar to Battlefield Management
If economics did little more than tell us what happened in the past, Econ professors would rank as History professors; very few contracts flow from the Cato Institute or the Heritage Foundation to History guys.
Economists apply what they are pleased to call ‘theories’ (better called models) to predict what will happen if things continue on as they are. But none of those models have a chance of working all the time.
Economics is a Type (4) activity studying strongly interacting networks of human behavior. Specific realizations of a situation are not capable of being repeated due to complexity involved. Collected data has very large measurement errors. Data can not be fully trusted due to
- incomplete information,
- intentional misinformation.
These effects spread a miasma of mistrust over received information. See part 1 of this series (Economics can never be right) for more on these points. This is similar to the dilemma facing battlefield officers – the fog of information in battle, no true duplication of events to test the model, and no ability to try again to see if things might change. Economic models cannot be verified the way weather models can.
Type (4) activities will always generate sharply competing models with schools of thought that hold diametrically opposite viewpoints. Some economists call any stray thought a scientific theory but they are not.
Some models will always be founded on emotional belief, since you can rarely disprove a model. Some try to generalize from past events to correct current decisions. But bad advice is pervasive. As some Generals will always advocate frontal charges against emplaced machine guns, some Economists will always be for heavy taxation on our poorest citizens with bonuses to the rich. The structure of the disciplines generates emotional clouds over thinking that will never dissipate. Economics is not a science.
Charles J. Armentrout, Ann Arbor
2013 Sep 2
Listed under General …thread General > Musings
Have a comment? Click on the title of this post, go to bottom, let us know.
Related posts: Click the INDEX button under the Banner picture