Slashdot Mirror
Maximizing Economic Output With Linear Programming...and Communism (medium.com)
Slashdot reader mkwan writes: Economies are just a collection of processes that convert raw materials and labour into useful goods and services. By representing these processes as a series of equations and solving a humongous linear programming problem, it should be possible to maximize an economy's GDP. The catch? The economy needs to go communist.
"[P]oorest members would receive a basic income that gradually increases as the economy becomes more efficient, plateauing at a level where they can afford everything they want to consume," argues the article, while "The middle classes wouldn't see much change. They would continue to work in a regular job for a regular -- but steadily increasing -- wage... Without the ability to own real-estate, companies, or intellectual property, it would be almost impossible to become rich, especially since the only legal source of income would be from a government job."
"[P]oorest members would receive a basic income that gradually increases as the economy becomes more efficient, plateauing at a level where they can afford everything they want to consume," argues the article, while "The middle classes wouldn't see much change. They would continue to work in a regular job for a regular -- but steadily increasing -- wage... Without the ability to own real-estate, companies, or intellectual property, it would be almost impossible to become rich, especially since the only legal source of income would be from a government job."
This isn't a new idea. Kantorovich (one of the inventors of linear programming) considered this venue of economic optimization himself, but the technology of the day wasn't up to the task and the bureaucracy didn't want to be displaced either. Some of his suggestions inspired the reforms that later got implemented by Kosygin, but the Soviet economy was rather distorted by subsidies at that point, so a lot of those reforms got rolled back.
There was also the fear that linear programming, with its shadow prices, would covertly smuggle capitalism into communism. See also Red Plenty for a half-fictionalized account of Kantorovich's attempts (or the Crooked Timber post, In Soviet Union, Optimization Problem Solves You).
Beyond that, there's Towards a New Socialism which is an idea/plan of how to run a socialist centrally planned society with modern technology. It uses sparse linear programming for the plan construction part and is based on sortition for government to diminish the inevitable corruption that comes with concentrating economic power like any CPE does. Would it work? Who knows? It may be interesting in the utopian sense anyway.
Tangentially related (speaking of scientific communism/socialism), there's also Project Cybersyn, the project to use cybernetics to run socialist Chile. That wasn't based on linear programming, though. If linear programming is the neat route, Cybersyn would be the scruffy route. Again, who knows whether it would have worked; if Medina's Cybernetic Revolutionaries is anything to go by, a considerable part of the problem was that of bureaucracy and what the people were used to. Managers didn't use the system because it felt cumbersome to do so, etc.
I am an economist. Economists have already extensively studied this kind of approach. It's called an Input/Output Model. Communist countries used it in their approach to central planning during the 1970's. It failed miserably for two reasons:
1) It assumes zero substitutability between inputs. E.g., to make a car you need exactly 1.35 tons of steel, 52.7 kg of rubber, 217 kg of glass, 1.73 KW of electricity, 29.4 hours of labor, etc. No other formula is possible, you can't use more energy and less labor, for instance. For reference, the production function is known as a Leontief production function. To be fair, adding any kind of substitutability between inputs results in a completely intractable problem. However, without substitutability this is a lousy way to actually model an economy.
2) It assumes perfect information on the part of the central planner. While this is an oft-used simplification in economic models, it's a lousy reflection of reality. It's simply impossible for a central planner to gather and correlate sufficient information to make it work.
Yet another piece-of-crap opinion article written by someone who couldn't be bothered to do an hour's research on Wikipedia.