The Early Days of a Better Nation

Posted 1:54 pm by Ken

Red Plenty

I recently got a review copy of Francis Spufford's new book Red Plenty, and, like Brad DeLong, immediately dropped everything to read it. It's a fictionalised account, or a non-fiction novel, about the project in the early 1960s to use computers to plan the Soviet economy. A key figure is the genius Kantorovich, who invented the mathematical technique of linear programming in 1938. (We follow his mind as the idea dawns on him, on a tram.) He and other real characters such as Kosygin and Khrushchev mingle with fictious characters - some based on real people, some not, but all convincing.

It's a bit like reading a novel by Kim Stanley Robinson, Neal Stephenson, or Ursula Le Guin - or maybe a mashup of all them; full of arguments between passionate and intelligent people, diverting (in both senses) infodumps, and all about something that actually happened - and, more significantly, about something that didn't happen, and why it didn't.

Computer scientist Paul Cockshott, a prominent advocate of cybernetic socialist planning, has written a comprehensive and enthusiastic review:

This is a marvelous and unusual book. It sits in a remarkable way in between science popularisation, social history and fiction. The author describes it variously as a novel whose hero is an idea and a fairytale. The hero idea is that of optimal planning. The idea of running a planned economy in just such a way as to ensure that resources are optimally used in order to deliver the ’red plenty’ of the title.

[...]

The author shows real skill as a science populariser, explaining such diverse topics as how the Pentode valve logic of the early BESM computers worked, to the molecular mechanics of the carcinogenesis mechanism that eventually killed its designer. He vividly portrays the enthusiasm and self confidence of the USSR in the late 50s when Khrushchev’s boasts that they would overtake the USA by 1980 and achieve communism seemed plausible. He gives a good didactic account both of the basic mechanisms of the Soviet Economy, and, through the lives of incidental characters paints a picture of its real operation that is more detailed and convincing than any academic history.

He traces the idea of cybernetic economic management from the hope of the 50s and early 60s to its sidelining under Kosygin, and the eventual relegation of Kantorovich to the less ambitious task of optimisating steel tube output for the oil and natural gas industry. Ironically, says Spufford, as growth rates slipped in the 70s, it was only the exploitation of petroleum for export that allowed Soviet living standards to rise.

[...]

All in all, let me say again, this is a book that should be read by anyone with a serious interest in economic alternatives.

Labels: Marxism, reviews, skiffy

27 Comments:

I'm not sure how computers are relevant to the calculation problem. They're relevant to the "millions of equations" problem, but that problem assumes that one has the data in usable form and just needs to perform calculations on it, which makes it downstream from the two main problems -- the Hayek problem (namely the impossibility of collecting the date in the first place, since the information exists in largely inarticulate form) and the Mises problem (even if one could collect the relevant data about preferences, the preferences would be ordinal rather than cardinal and so absent prices there'd be no way of combining different individuals' value scales).

By Roderick T. Long, at Sunday, May 23, 2010 5:19:00 pm

Roderick - these perfectly valid points you raise are extensively discussed in various papers by Cockshott et al. One aspect of the Hayek problem appears in Spufford's book, in a chapter titled 'Working From the Photograph'.

(I should add that Spufford's book presents a by no means flattering picture of Soviet society, politics, and economics.)

Anon - a great deal of progress within capitalism itself is the result not of entrepreneurs but of ideas developed by employees, whether in R&D or otherwise. Even in Soviet-type economies there was technical innovation and progress. Whether there was enough is a different question, to which the answer is no.

By Ken, at Sunday, May 23, 2010 5:50:00 pm

This is the Spufford who wrote that book on the British space programme, right?

Ken, have you read Hilary Wainwright's _Arguments for a New Left_? She tries to take on Hayek's argument about the nature of economic knowledge and its inconsistency with the concept of a planned economy by arguing that diffuse economic knowledge is not limited to entrepreneurs but can be accessed by workers at the point of production. Therefore a new socialism could engage with that knowledge by basing itself on workers' autonomy. . .

By D.J.P. O'Kane, at Monday, May 24, 2010 11:15:00 am

Actually, you're right about ideas from workers - companies which have a suggestions scheme tend to do very well - even IBM had that "banana badge" system! The problem with suggestions is that which was encountered by the chief test pilot of Rolls Royce, Bill Beaumont. He asked for a rise, his boss asked why he thought he deserved one, he pointed out that he had thought of the Spey-engined Phantom, and his boss said "That's your job!" Gah...
The problem with trying to model an economy is that this would depend on collecting a lot a information that is none of the state's darned business!
As to entrepreneurs, I recall the suggestion in one of your books that the early Soviet space programme was really run by mavericks who used the State's ambitions to do what they wanted to do - sounds cool to me! But I think that the way ahead for a British independent spacefaring capability is the private enterprise route - we've got to get back to making things, a service-industry (especially financial services!) economy will just fall to pieces!
Grif

By Anonymous, at Monday, May 24, 2010 12:26:00 pm

Someone needs to deal with this whole calculation problem for a centrally-planned socialist economy by finding ways to get around the "centrally-planned" part. Perhaps the sketch provided by D.J.P. O'Kane could be filled in by relying on market mechanisms to distribute the calculations, as it were. We could call it "socialistic marketism."

Again, though, in my largely-ignorant perusals of such things, I guess I still don't understand why "efficiency" is such a bugaboo. It's not as if plenty of other systems don't muddle along without perfect efficiency. For that matter, it's not as if existing capitalism perfectly obeys the Efficient Markets Hypothesis. So couldn't distributed planned economies potentially do "well enough"? I'm sure this is amusing naïveté on my part.

By mds, at Monday, May 24, 2010 2:48:00 pm

Thanks for the review and recommendation - it's on the wish list.
Especially interesting since one of the main applications for massive amounts of computing power is complex financial modelling and high-speed trading, which are also economically dubious if not actually destructive of value. Not quite as bad as using them to model nuclear explosions, but considerably less productive and conducive to the welfare of humanity than using them to model big blue aliens for James Cameron, or global weather pattterns for the Met Office.

In answer to Mr O'Kane: yes, same Spufford. He also wrote the very good "I May Be Some Time: Ice and the English Imagination".

By ajay, at Tuesday, May 25, 2010 11:09:00 am

Chris Williams' reference to Cybersyn intrigued me, since I knew nothing about it but had read one thing by Stafford Beer, an introduction to a book on a biological world-view that many consider to be cranky but I do not. Dr. Beer's introduction seemed pretentious, so I dismissed him for two decades. But Chris's reference led me to look at Wikipedia's entry on Beer. It is fascinating, it changed my opinion of Dr Beer, and it is relevant to this blog topic. http://en.wikipedia.org/wiki/Anthony_Stafford_Beer .

By George Berger, at Thursday, May 27, 2010 7:20:00 am

To Ken -- thanks, will read.

To George -- thanks for the plug!

To Anonymous and mds -- Whatever Hayek and Mises themselves may have thought, nothing in their argument requires entrepreneurs to be employers rather than workers. Indeed, Kevin Carson argues here that the arguments against central planning by the state also apply to centralised hierarchical corporate firms; and in his other writings (here and here) he goes on to draw the conclusion that the real upshot of Mises' and Hayek's arguments should be to favour worker-controlled firms over both state planning and traditional capitalist corporations.

By Roderick T. Long, at Thursday, May 27, 2010 5:49:00 pm

I disagree profoundly with the idea that entrepeneurs are the sole agents of progress. Without defending the Soviet Union or China, the former developed a space programme while the latter was able to raise average life expectancy by thirty years between 1949 and 1975. Cuba's gains in literacy, healthcare etc. are, as far as I'm concerned, examples of progress, though of a social rather than scientific nature. Even then state organisations like the NHS were able to discover basic yet profound facts like the link between health and exercise.

By Budapestkick, at Thursday, May 27, 2010 8:55:00 pm

Roderick made the point "They're relevant to the "millions of equations" problem, but that problem assumes that one has the data in usable form and just needs to perform calculations on it, which makes it downstream from the two main problems -- the Hayek problem (namely the impossibility of collecting the date in the first place, since the information exists in largely inarticulate form) and the Mises problem (even if one could collect the relevant data about preferences, the preferences would be ordinal rather than cardinal and so absent prices there'd be no way of combining different individuals' value scales)."

A key figure in Spuffords book is Kantorovich, and as I argue here Kantorovich essentially produced an answer to Mises' impossiblity of calculation in kind argument. The fate of Kantorovich's solution was a major theme of the book under review.

By Paul Cockshott, at Thursday, May 27, 2010 10:23:00 pm

Something along those lines, known as "Project Cybersyn" , was attempted in Chile between 1970-3, under the direction of Anthony Stafford Beer
By modern standards, it was very crude - a network of telex machines that linked factories with a single computer centre in Santiago.
After the military coup on September 11, 1973, the control centre was destroyed.

Given the subsequent development of the internet and the vastly more powerful processors available today, the feasibility of such a system would be far better.
It would only be an extension of what's currently used by private companies for just-in-time production, financial transactions and share trading.
It's really just a question of balancing production and consumption across a whole economy, without bottlenecks.
The obstacles are more social and political than technical.

Prianikoff

By Anonymous, at Sunday, May 30, 2010 8:59:00 am