Machines of Loving Grace (and their Masters).
A shorter version of this article was originally published by ORGzine.
If there was a fault with the first episode of Adam Curtis’ All Watched Over By Machines of Loving Grace, which aired Monday night, it was his analysis of the eponymous machines. In an hour long programme which was, by turns, brilliant, visionary, terrifying, and, just a little, frustrating, Curtis zipped from Ayn Rand’s Objectivism to the New Economy bubble, the Credit Crunch and a barn full of people playing a collective game of Pong.
Curtis’ thesis, that computers were meant to liberate us from traditional forms of hierarchy, government, and, if we believe Rand, the need even for altruism, but have instead provided false comfort and in fact helped to concentrate and consolidate power has some validity, and, as always, Curtis makes the case with considerable aplomb. But it’s not the computers themselves that made us slaves, it’s the financial and governmental elites who utilised them. That might seem like a slight point, but Curtis’ repeated claims that is was computers that created the financial models that failed, risks absolving those who wrote the programmes, interpreted (or misinterpreted) the results and failed to adequately manage or regulate their implementation. The financial crisis wasn’t caused by machines, it was caused by the greed and hubris of men (and in some cases women*).
Of course, all that isn’t to say that machines, and perhaps more importantly, our perception of what they could achieve, didn’t play a significant role, and it’s worth examining that in a little more detail.
Anarchism, left and right
The Objectivists believed that government was evil. At a time when both left and right on the mainstream believed that giving people too much freedom would lead to anarchy and destruction Rand and her followers advocated a radically individualist philosophy, society existed, but it shouldn’t have, the weak didn’t deserve love and altruism was something to apologise for.
They thought that, through technology, all government could be abolished and replaced with a system of free markets and free people, self-organised by profit and their own self-interest. Computers would allow people to communicate directly and in the absence of interfering, irrational politicians people could govern themselves.
But the problem here isn’t the machines, it’s the (borderline sociopathic) philosophy. Computers can help us to self-organise, and it’s not only the hard right that think we can do without central government. Those on the libertarian left have long held that we can emancipate ourselves from arbitrary government authority without descending into barbarism, but to avoid the vast inequality and privilege of the Randians anarcho-capitalism, we have to also liberate ourselves from corporate power.
As Mikhail Bakunin put it:
Freedom without socialism is privilege and injustice. Socialism without freedom is slavery and brutality.
Computers can help us to organise non-hierarchically and without central authority, but they’re only a tool. How we choose to use them is up to us. Across the Arab world (and, indeed, all around the world) people have protested this Spring for genuine democracy and economic justice. They’ve done it using Twitter and Facebook, communicating and working in distributed networks, and they’ve not done it because they believe in Objectivism. But neither have they done it because of the tools they had available. As Michael Hardt and Antonio Negri observe of the protests:
The prevalence in the revolts of social network tools, such as Facebook, YouTube, and Twitter, are symptoms, not causes, of this [non-hierarchical, network] organisational structure. These are the modes of expression of an intelligent population capable of using the instruments at hand to organise autonomously.
Machines are tools. How we use them is almost always down to the beliefs we already hold.
But perhaps, that’s obvious when we’re talking about political organisation, how about that other area in which Curtis theorises computers have come to dominate our lives, finance.
Efficient Markets and Risk
Finance changed dramatically over the last thirty years, during that period where Alan Greenspan, one of Rand’s most vociferous acolytes, and the neo-liberal hard right has been dominant. As markets were deregulated, and computing power increased, vast swathes of new assets into which money could flow were created. Securitisation became ever more complex and products more exotic. And as this happened new models were needed to keep track of all this information. Mathematicians, computer scientists and physicists became more important and the salesmen with their inside connections ever less important to financial companies. Managers, who still came from business, rather than science, backgrounds, began to believe that these new quantitative analysts could create models that would guarantee them profits they could only have dreamed of in the past. And behind all of this was the belief of the Objectivists (and others) that markets were rational and efficient.
The efficient market hypothesis (EBM) states, at its strongest, that all information about an asset is already in the price. What that means is that events that have already occurred cannot affect the future price of an asset, be that a share, a security or a piece of property. So if shares in a company have been increasing for the last month, that should have no bearing on what the price does tomorrow, because the events that led to those rises have already been taken account of and future changes will be in response to future events.
So how does the EBM help in financial modelling? doesn’t it say that prices in the future are unknowable? doesn’t that make prediction harder? Well, while the EBM says we can’t simply extrapolate future prices from past prices what it does do, formally, is to tell us that asset prices should be what are known mathematically as Markov processes. That is they should be stochastic (random) processes where to predict (probabilistically) the future state of the system we need only know the current state, not the whole history of the process. This makes the maths much easier, and opens up the possibility of utilising decades of work in mathematics and physics to calculate the statistical properties of the market. It makes it possible for complex financial models to put numbers on the risk of prices increasing or decreasing, and, hence, for traders to hedge against their transactions and, hopefully, eliminate those risks. If the EBM is right and the market behaves itself (there are no external shocks or crises) traders should be able to virtually guarantee not to lose money.
Of course the EBM isn’t true. Markets are not efficient. A cursory examination of the historical data reveals that there is short term positive and long term negative correlation in prices. If a share has been increasing in price for a week, it probably will go up tomorrow, though only on average and to a small extent, while over a longer period it will revert back to the mean. Knowing when it is likely to switch is, unfortunately, almost impossible.
And that’s not the only problem with our financial models. Because, as I alluded to earlier, all this only works, even approximately, when markets behave. And it turns out markets misbehave far more often than most financial models expect. If most of the time your strategy makes you £1000 but occasionally it loses you £1m, it matters if that risk is 0.01%, in which case you should expect to make money in the long run, or 1%, in which case you won’t. But, once again, the mathematics becomes far simpler when approximations which are untrue, and underestimate these extreme events, are used instead of more realistic distributions.
But, for all that, these failures only mattered so much because we believed them, because the regulators and managers, who often didn’t understand the models or the risks, failed. If we blame the models, we risk thinking that if only we can get the model right, everything will be fine, in fact the problem lies much deeper, in an economic and philosophical system that believes society can be perfected through rationality, that politics isn’t important and if only markets were freer everything would work itself out. If nothing else, the events of the last three years should have laid that belief to rest. Depressingly, they haven’t.
*For the record I don’t buy the idea that if finance or politics were more gender balanced, but all else remained the same, we’d have seen anything different, but it’s worth noting the disparity nonetheless.
What I took from the documentary is how at first it seems ridiculous that powerful people like Greenspan would govern with a world view that comes from a novel that is basically an expression of the autistic mind. But when you look at the man’s career it does seem to suggest that it was exactly this form of fantasy that was driving him, and all his right-wing cheerleaders.
Curtis documentaries (all on youtube) focus on how right wingers use power to treat the powerless with disdain. Intellectually what they’ve been doing since the cold war is to misuse mathematical models to constrain the agency that people should exercise in their communities – something right wingers can only envisage as ‘markets’. Academics failed to critique this form of misanthropy presumably because they lacked the skills or had similar inclinations. John Nash admitted his work for the US military during the cold war had been misused…after being treated for schizophrenia for 10+ years (Curtis – The Trap).
While obviously we can’t blame computers any more than we can blame guns, we do have to appreciate what they can be used for. Our general lack of understanding of the computer modeling that the powerful are doing will mean they will continue to get away with land and financial heists we are just waking up to. They have no qualms about treating the rest of us like mugs.
Its very hard to sit in a room with someone who has an aggressive black and white world view and talk about subtlety and why the ideas that could bring them millions are dangerous and inhuman. Especially when they can bring up the free-market argument: well if it doesn’t happen here, it will somewhere else, so we might as well take the lead.
Again, I bring up the word autism, not because of a lack of respect for people with this kind of mind, but because I don’t think people who find it difficult or impossible to understand how other people think should be making decisions about society and how we want to relate to each other. Of course the right-wingees try to dismantle communities and atomise relationships. For them this is something that really doesn’t exist. They can’t imagine it, they have no use for it, it is just unwelcome noise that muddies the Platonic ideals that fill the void where their imaginations might otherwise reign.
I found the programme interesting and am also wondering what the next episodes will say. I agree with Alasdair that the users of the financial modelling tools are much more to blame than the tools themselves, and also that the managers of the forecasters were even more culpable for accepting the predictions uncritically. The point about Alan Greenspan backing down from calling a warning seemed particularly worrying. Does anyone know if this is accurate? Also, I think the senior managers in the banks didn’t understand the packaged debt products, as well as not understanding the models.
There is an interesting interview with Adam Curtis on the excellent Little Atoms podcast this week:
http://feedproxy.google.com/~r/littleatomspodcast/~5/xZcnFwF6y8w/technoutopia.mp3
There are some interesting arguments about how our conception of social action is wrong. Rather than focusing on the actors, we should focus on the networks that produce social change. In this understanding objects constitute change as much as being tools of change.
That is to say that it is not just an individual using an object that changes things, but the result of a negotiation between individual(s) and object.
A good example is barbed wire. Without barbed wire it would have been impossible to enclose the American West (there are neither trees to build fences, nor readily available stones to build walls). Livestock would have roamed freely preventing the settlers from owning land. This meant that the invention of barbed wire allowed a network relationship between the settlers and the object that transformed the soical relations of the East and West coasts of America and with the First Nation people.
You can read more in the work of Bruno Latour. It will be interesting if this is where Adam Curtis is going. I was expecting a bit more Latourian analysis when I saw how he was starting out…
http://en.wikipedia.org/wiki/Aramis,_or_the_Love_of_Technology
I was quite impressed by the documentary, but agree that it was a bit simplistic to imply that the extreme faith in market efficiency came from computers, because the ideas were around for a long time before we had machines to do the calculations. Computers were just a convenient prop for people with a particular agenda, who used shiny technology to make their theories look more credible. Well, that and the wonders of a positivist approach to science, which meant that the people who were putting the ideas forward never gave proper consideration to any evidence that countered their views, and so they probably believed in it as well.
Incidentally, the whole idea of computers being able to magically produce more knowledge than has been programmed into them is one of the things satirised in The Hitchhikers’ Guide to the Galaxy. Read into that what you will, but I find it vaguely amusing to cast the Randian economists in the role of the white mice.
I didn’t see the programme, but you’re unquestionably right about machines only being tools. It seems quite obvious to me so no wonder you were frustrated if this was the argument being put forward.
Computers haven’t created or caused anything – their users have. Yes, technology has created new opportunities, but it’s still people who decide what these opportunites are, and which ones to take.
If someone gets shot, do you blame the gun? No, but you could blame the people who took the opportunity to develop the weapon, and you can certainly blame the person who pulled the trigger.