On experiments in economic governance: to plan or not to plan?

Chandre Dharmawardena, in the Island, 1 February 2012

I wish to comment on an interesting article by Leelanada De Silva (Sunday Island of Jan. 28), from a scientist’s perspective. Articles by Usvatta-aratchi and Carlo Fonseka (The Island Oct. 19), and by Gunadasa Amarasekera et al treated related topics. When my article entitled `Some lessons from China’ (The Island of Jan. 12 Jan.) appeared some readers wrote that `planning is better than blind process’. Leelananda suggests that effective planning is possible in both socialist and capitalist systems.

Social and physical sciences: Planners need to predict and choose among various possibilities. The task of scientists is also prediction. The subject known as `man-body theory’ deals with the statistical mechanics of many agents influencing one another. Hence, mathematical physicists find mathematical economics to be familiar reading. Indeed, the success of quantitative models in science prompted quantitative approaches to economics. Samuelson, von Neumann, Morgenstern et al are names familiar to mathematical physicists and economists alike. Thus, began the rise of `econometrics’ and model building, using elementary dynamical and statistical concepts like time-series analysis, bell curves, Z-scores, regression, input-output matrices, Markov and Non-Markov processes, game theory, Hilbert spaces, etc.

However, in physics, theories have to make successful predictions of controlled experiments before they are used. Controlled experiments are impossible in the social sciences even for dictators. Nevertheless, ‘economic planning’ as an extension of budgeting and fiscal policy soon became the norm. The centralized ‘socialist’ economies seemed best for planning optimal growth, but risky for individual liberties. Economists and bankers began to dictate what governments ‘should do’, while collecting fat bonuses and failing to predict financial crises staring in their faces.

When Carlo Fonseka says that ‘Cuban communists produce 500 doctors while a comparable Capitalist society produces only 50’, he applauds centralized ‘socialist’ planning. However, the 500 doctors find shortages of medicine in the ‘planned economy’ and the patients die. In contrast, drugs are plentiful in the capitalist economy and doctors prescribe lavishly to enrich the companies and clinics. The over-medicated patients become permanent cash cows.

Scepticism about predictability: Eclipses could be foretold with uncanny accuracy and celestial mechanics is the paradigm of predictability. However, the great 19th-century mathematician, Henri Poincare, proved that some simple clock-work like systems as well as all complex systems were beyond prediction. Practical indeterminism has risen out of formal determinism. Historical evolution is not the dialectical doing of big forces, but a devil dancing on details.

Mathematicians ignored Poncare’s results as a rare pathology. It was only in the 1960s, with the availability of computers, that Poincare’s results came to the fore as ‘chaos theory’. Hard-headed economists had already realized that econometrics was mostly a misleading exercise, responsible for the financial mistakes on the part of the big powers in the 1970s (a period discussed by Leelananda). Here, the warnings of the economist Frederich Hayek are most appropriate. In his Nobel address some 37 years ago, Hayek referred to economic planning, be it by Marxists or capitalists, as nothing but “Pretense of Knowledge”. Readers may Google Heyak and read the Nobel oration of a master.

Physicists and chemists also realized the importance of Poincare. Science is successful because it uses a ‘reductionist’ approach limiting a system to a few variables. Although we can precisely predict isolated quantum systems with even a thousand or so particles, we cannot predict the exact conformation of a protein at room temperature. Complex systems are governed by differential equations having the indeterminacies discovered by Poincare a dozen decades ago.

Statistically, virtually all swans should be white. Social theorists began to realize that social planning has to deal with ‘black-swans’ – Nicholas Taleb’s terminology for the occurrence of the utterly unanticipated. Chaotic systems do not obey normal distribution, Z-scores, extrapolations from regression analysis, Markov chains etc. They are full of black swans, intensely sensitive to initial conditions and defy prediction.

Complex systems: Physicists and mathematicians were humbled when Yakov Sinai proved in the 1960s that a class of billiard problems could be unpredictable and chaotic! So the scientist trying to predict the behaviour of a bacterium with millions of interacting proteins needs new tools. Such tools can be tested by experiments. Fund managers and economists also have to make predictions. But they cannot openly do experiments.

Nature solves the problem of finding the best seeds by scattering a million seeds and allowing Darwinian ‘co-opetition’ (such co-opetition often involves profitable cooperation and competition). Scientists model this using `Monte-Carlo’ type computer simulations. By analogy, capitalist societies allowing Darwinian co-opetition are more likely to succeed. This was what Deng Xiaoping realized in 1977, in fixing a failed Maoist China.

Such simulations have to be constrained by limiting the ‘allowed moves’ to energy controlled ‘Metropolis-Teller’ moves. By analogy, we need controlled economies with capitalism balanced by social checks, liberty balanced by authority, human-rights moderated by social rights, and so on. Many competing plans and planners (many board rooms!) are needed for the task.

Social experiments – revolution or devolution – are full of dangerous black swans. Social planners should continue to study societies, but avoid any centralized plan implementation.