Salvador Allende was elected President of Chile in 1970. Most people know how that story ends: the coup in 1973, which led to Allende's death and brought the dictator General Pinochet to power. But far fewer people know that Allende contributed to an amazing chapter in the history of information.
'Project Cybersyn' used a 'supercomputer' called the Burroughs 3500 in an attempt to coordinate decision-making in an increasingly nationalised economy. Allende, a Marxist, hoped that with modern technology he could succeed where the Soviet Union had failed. The computer would centralise and coordinate the tremendous complexity of a national economy. Allende was helped by a larger-than-life cybernetic theorist named Stafford Beer, who supported the socialist project but also demanded a generous fee and plenty of chocolate, cigars and Chilean wine.
It would be unfair to blame Chile's economic collapse on Cybersyn — strikes, opposition from the US and sweeping nationalisations should take the blame — but it's safe to say the project was not a success. Factory managers did not send in accurate information for Cybersyn to crunch. Ironically the project worked best when strikes were paralysing Chile and those managers started to panic and ask for help.
My father worked on Burroughs 3500 machines and he tells me they were not quite supercomputers even by 1970 standards. But Cybersyn's essential problem wasn't that the computer wasn't powerful enough. It was that computers actually seem to be a force for decentralisation, not centralisation.
An influential study published at the height of the dotcom boom by Lorin Hitt of Wharton and Erik Brynjolfsson of MIT showed that companies who bought pricey new IT systems only became more productive if they reorganised, pushing more responsibility and more decision-making power to people further down the organisation. Neither the fancy computers nor the "power to the people" reshuffle helped alone: they had to go hand in hand to be effective.
This may explain why researchers such as Raghuram Rajan, formerly chief economist of the IMF, have concluded that large US businesses are tending to flatten their hierarchies, cutting out interim bosses and giving more responsibility, again, to junior staff. (Another simultaneous explanation is globalisation: in fast-moving, competitive and diverse markets, a decentralised company has an advantage. Decentralisation has costs and benefits, but the balance is tipping in its favour.)
Even the US Army, chastened by its experience in Iraq, is re-emphasising the importance of junior officers responding to local conditions, rather than gosh-wow computerised control rooms. It's a lesson many businesses have already learned.
Tim Harford's new book, Dear Undercover Economist
, is published this month in paperback
blog comments powered by
