Economic Logic: a survey and variations on the theme

Introduction

Economic logic is used in colloquial language as the applictions of certain forms of economic thinking in policy making and management. At the end of September 2005 the phrase 'economic-logic' gives rise to 175.00 results in www.google.nl. The related phrase 'economical-logic' features only 479 results. (There are 43 results for 'economic-algebra' and just 4 results for 'economical-algebra'.)

Leaving out the '-' yields more results: for 'economic logic' one finds (end of September 2005) 19.600.000 results with the first pages very similar to 'economic-logic' and in both cases Mark Skousen leading the field of results.

In April 2006 'economic-logic' produces 283.000 hits, with Mark Skousen at the top and this page nr. 2. 'Economical-logic' now has 480 hits. 'economic logic' has risen to 36.900.000 hits with the same two pages on top.

In this document I will review and comment a number of occurrences of the phrase 'economic logic' in existing literature.

I Economic logic: some uses of the phrase

1 Economic logic, Austrian Economics and the calculation debate

Many documents that can be downloaded from the homepage of the Ludwig van Mises Institute in Auburn, Alabama suggest that von Mises viewed economics as a science that can be successfully executed as a branch of applied logic. This perspective is due to the Austrian School an constitutes a part of Austrian Economics. However, this form of logic seems to be limited to a philosophical theory of action mainly taking into account an 'Austrian selection' of authors. Formalized logic is apparently not advocated, probably because the logic is primarily aimed at supporting 'Austrian views'. Indeed a rather purely mathematical development of a formal economic logic may or may not turn out to support the 'Austrian dogmas' in various areas. In the Austrian School calculation plays a more important role than logic.

A cornerstone of these developments is that socialist state planning requires calculations which are impossible because market prices cannot be calculated. This the logic of von Mises leads to the insight of the inadequacy of apriory calculation on which centralized socialism is supposedly built. The proof of impossibility seems not to take the logical form of a liar paradox or a more sophisticated version of it. Rather it seems to stronlgy doubt the plausibility of these computations being feasible. But modern computational science can reconstruct the experimental outcome of statistical mechanics in some case and the case might still be open.

1.1 General options for concluding the calculation debate

Why is it impossible for a socialist planner to run a capitalist simulation of its future before installing a plan, thereby optimizing the plan beyond the simulated future. The answser must lie in the fact that the capitalist future could use computers of equal strength to those doing the simulation for the socialist planner and somewhere here lies the opportunity of a formalized diagonalization. Still the argument is non-trivial. If the planner has a 5-year planning horizon, (s)he must in one go plan in such a way as to outperform the decisions that otherwise millions of agents would make during these 5 years. The argument can go in different ways: is it a complexity problem because these millions of agents each using massive computers simple apply far more processing power than the planner may hope to perform, or is it the very problem that the five year planning cannot take into account the full combinatorial explosion of developments that subsequently occur.

The complexity problem arising in the first case can take a trivial form: following Turing the termination of a computation cannot be predicted. One needs to rely on observation of the computatation and cannot do any better than that. This in itself already constitutes a formidable obstacle to the planning calculation but perhaps the termination behavior of computations cannot be embedded in economic forecasting in a useful way. In the second case the problem is merely one of planning involving prediction to an impossible extent, i.e. an application of chaos theory to the issue of economic planning. In the latter case there is no argument needed concerning the role and possibility of price calculations. Whatsoever sophisticated pricing strategy cannot defeat the combinatorial explosion of possible futures, while a dynamic planning with simple pricing strategies may be able to easily deal with an unpredictible future. In the case of the first argument (against the possibility of performing the 'socialist' planning) it might be necessary to use a sophistcated real time pricing strategy as a means of communication between millions of independent agents in order to develop a mechanism that provably cannot be beaten by planning.

The caclulation debate has another dimension with respect to both the feasibility and the relevance of predicting equilibrium market prices. Although the position of von Mises was aimed against socialist economists in the 1920's it had to be defended against the arguments of market equilibrium oriented economist in a later stage. Such defensive arguments were convincingly put forward by Hayek in the 1940's. By and large it seems not to be the case that any of these arguments concerning the feasibility of 'calculation' can be understood as a mere application of formal logic. Again that does not imply or deny the possibility of logical arguments concluding the issue on the long run.

1.2 Logic of computation and the calculation debate

Concerning the logical problems of prediction in joint work with Alban Ponse I have worked along two lines: in 'Execution Architectures for Program Algebra' (Utrecht Logic Group Preprint Series no. 220, June 2004, to appear in the Journal of Applied Logic) we have investigated versions of the halting problem (predicting events in a computable and deterministic system) in the case of finite state automata, coming to the conclusion that this is paradoxical and hence impossible. As a consequence the impossibility result of Turing is a mere diagonalization argument that is independent of the computational universality of the famous Turing tape as a datastructure. Relations with Newcomb's paradox and the prisoners dillemma are formulated. These results suggest that in finite economies forecasting might be provably impossible. However, in our 'A bypass of the Cohen impossibility result' we investigate the observation made by F. Cohen in 1984 that it is impossible for a computer to detect wheter or not a program it is running is a computer virus. After formalizing this matter in program algebra it turns out to be a forecasting problem. But unlike the halting problem mentioned above this forecasting problem is not paradoxical and the impossibility phenomenon only arises if full Turing completeness of the computing system is assumed. That is of course an unrealistic assumption in practice and in the finitary case virus detection is feasible in principle. If economic forcasting after formalization appears to be rather like virus detection (perhaps too much to hope for), then planning is not paradoxical, though it might still be computationally unfeasible.

2 The ecomnomic logic of "open science"

In The Economic Logic of "Open Science" and the Balance between Private Property Rights and the Public Domain in Scientific Data and Information: A Primer Paul David (Stanford University and University of Oxford, 2004) describes the economical aspects of open science as its "economic logic". No explanation of the phrase logic, or justification of its use, is given but the balance mentioned in the title is dealt with in meticulous detail. None of the 19 cited references has 'logic' in its title. This suggests that the phrase logic conveys the sense of completeness felt by the authors after bringing a multitude of viewpoints together in a homogeneous framework.

3 The economic logic of executing computer hackers

This refers to a column posted by Steven E. Landsberg (May 2004). He performs a long calculation demonstrating that it is in fact preferable to execute malicious hackers than murderers. As he agrees the conclusion rests on numernous assumptions and Landsberg does not at all commit himself to the idea of capital punishment. In the document 'economic logic' is used where 'economic arithmetic' might have been preferable and for both terms the justification is clearly defective. I conclude that talking about 'the economic argument for executing computer hackers' sounds far too drastic and serious and that 'economic logic' is used instead with the purpose of providing a sarcastic undertone that is helpful to avoid being taken seriously and literally. (I have a copy of the page available which is of couse not made public for reasons of IPR.)

II Microeconomic logic and macroeconomic logic

At the end ofseptember 2005 Google finds 222 results for 'microeconomic-logic' and 136 for 'macroeconomic-logic'. I will first highlight some of the more informative hits concerning microeconomic logic.

Concerning 'macroeconomic logic': this phrase seems to be used as an alternative for 'macroeconomic' theory on a limited but increasing scale. There is no sign of a justification of the use of the phrase.