Red Plenty

"Yes, it is a novel—more or less. That’s absolutely right. It’s set in the 1950s and 1960s—that moment before things started to tank. What’s so brilliant is the way he knits together the economic history and the history of ideas. He talks about some of the economic trends; the Soviet Union was making great strides. Then there were the heated economic debates that went with that. Which side was going to win? Socialism or capitalism? He’s taken that episode in history, those few years, and created this incredible narrative. He’s a brilliant writer. He creates these characters who, to me, feel authentic. He really knows how to bring these characters alive, through their voices. Through these various economic debates, yes, I think he does show why it didn’t happen. There is the problem that planning runs into as the economy gets more complicated. It’s very tricky for a central authority to manage everything—to deal with innovation and all the complicated aspects of an economy as it develops. You can’t. This is a classic debate in comparative economics, which does that better? Is it the price mechanism or the central planners? There’s another episode in there when Khrushchev is driving through New York and he’s amazed by the fact they’re making little hamburgers. That’s a perfect way to feed workers! And these little plasters that you put on when you’ve got a cut! They have glue but not very strong glue so you can rip them off very easily. This capitalism is “a torrent of clever anticipations”—which is just a brilliant way of summing up how the market works. The market somehow is able to do this; how difficult it is for a planner to do it. The way he deftly weaves in these discussions of big questions of economic philosophy is really quite brilliant, I think. That’s why I put the book in there, to show the potential for narrative and the use of characters to create a panoramic view of classic questions. Absolutely, and we’ll get back to that again with the Hodgson book, because that’s really what it’s all about. We forget how heated that debate was among western economists as well. There was the so-called ‘calculation debate.’ It’s a technical, arcane-sounding debate, but he weaves it into the narrative somehow. So you can think about a market economy as a set of equations. The price mechanism solves out all those equations and gives you an optimal set of prices to efficiently allocate resources around the economy. But if that’s the case, couldn’t you solve that set of equations as the planner in the Soviet Union? Couldn’t you solve them and then implement them in your economy? And then tweak them to make things a bit fairer? A huge debate about that was going on through the mid-20th century. Again, it’s not something that students learn much about nowadays, but it was an enormous issue then. Yes, in some ways they’ve got a lot in common. We’ll see that in the Hodgson book, he critiques both of them in the same way. He puts them as actual bedfellows. One thing I really like about the book is the way he brings these issues alive across chapters through different characters. For example, there’s a chapter about the planning algorithm, so figuring out how to tweak prices and the plan to get things to work properly. One of the characters, a woman, turns up to this ‘science city’ to get a job and she goes to a party and meets these slightly tipsy, young mathematical economists who say we need to raise the price of meat. “This capitalism is “a torrent of clever anticipations”—which is just a brilliant way of summing up how the market works.” Then, in the next chapter, you go to a provincial city and you’re in the office of the mayor and there’s a riot breaking out because the price of meat has been raised. So you get this lovely panorama—what seems like this arcane theoretical discussion at a party—and then you see the results on the ground: people not being able to buy meat. There’s a huge riot and lots of people get killed by the police. It was a real historical event."

"It’s a wonderful, gorgeous novel. It has endnotes, which is decidedly odd in a novel. It’s about the ‘socialist calculation debate’, which he says he chose deliberately as the most unpromising topic he could possibly write a novel about. The book lays out the debate over whether it was possible to replicate, using planning mechanisms, the benefits of a market, and describes the Soviet Union’s efforts to realise this in the post-war period, reaching a peak during the Khrushchev years. After that, economic planning underwent a gradual and then rapidly accelerating decline. The book has a couple of characters who pop up here and again, but the narrative structure is really the story of a system. So it begins with a mathematician, Leonid Kantorovich, who has this wonderful insight when he’s sitting in a tram. It’s a beautifully designed scene describing how Kantorovich is stuffed into a tram with all these smelly, sweaty human beings. He thinks about the ways these human beings can somehow magically coordinate themselves so that they all get on and off the tram at the same time; he’s also thinking about the hole that he has in his shoe, which is letting in water, and this extraordinary mathematical idea he has just had. This is the beginning of the notion of linear programming: of how you can take a complex system of variables that looks like it doesn’t have any obvious solution and figure out ways to optimise it. It’s this blending of on the one hand the sweaty reality of human beings, and on the other hand this beautiful, beautiful mathematical idea which seems to have profound consequences. Get the weekly Five Books newsletter The rest of the novel is really working that through, weaving back and forth between the efforts to plan and implement the economy and the systems that this gives rise to, and then the consequences of that for the lives of ordinary people: for a young woman academician, for a woman giving birth, for protesters who are shot because they are demonstrating against the rise in the price of meat which has been planned by these economists. It’s looking at how an abstract system like that works out in reality, and how that reality feeds back into the system. It’s about how it is that the beautiful mathematical insights seem to recede further and further into the distance as the system trundles along and becomes its own thing—its own messy, unpleasant and inefficient human thing. That’s absolutely correct. It all happened, more or less. Spufford is quite clear in the footnotes about what he’s doing. Part of the reason he added them is to say, ‘With this incident here, I telescoped this and that thing together’ or ‘This person is not a real person but has something in common with Raissa Berg, who was a famous geneticist.’ He’s using the tools of a novel to try and probe a social logic, which is an odd, contrary, wonderful thing to do. It’s the kind of thing that shouldn’t work, but does—gloriously. Red Plenty has acquired kind of a cult following among social scientists. It really helps to set the scene for the debates that are happening at the moment. The ways in which information might or might not be used are in many ways recapitulating those that happened 60 or 70 years back, albeit with a different set of technologies and a different set of ways of applying them. On the one hand, we have people in Communist China, like Jack Ma, suggesting that we may not need markets anymore; we may be at the point where planning is actually going to work because we’ve got machine learning. Machine learning is going to provide us with the sophisticated means to achieve what the planners were trying to achieve and where they failed. On the other hand, we’ve got the Silicon Valley model, which is trying to figure out ways to use machine learning techniques to turn raw information into patterned data that can then be turned towards a variety of commercial purposes, with the same kind of enthusiasm that the people like Kantorovich had. This sudden, ‘Oh my God, we have the mathematics to turn all of these complicated miseries of human life into a set of engineering problems that can be optimised, isn’t that wonderful?’ sounds very familiar if you’ve read Spufford’s book. I don’t think we live in a world where it will ever work. One of the off-shoots of Red Plenty is a wonderful piece by a co-author of mine, Cosma Shalizi. He’s a statistical physicist and he goes into the math of Red Plenty and explains why it is, given what we understand about computational complexity, that this stuff simply doesn’t work. Another friend who’s an economist at Columbia, Suresh Naidu, is more optimistic, but hasn’t yet written up the reasons for his optimism. Cosma also talks about how, even if you could somehow get the math to work, the ways in which human beings are likely to respond to these systems invariably mean that they’re going to screw up. There’s this wonderful bit in Red Plenty where there’s a discussion between an economist—who’s really disappointed that they’re not going to apply his beautiful new math—and a somewhat cynical party apparatchik who says, ‘All of this math relies upon the assumption that the producers in the factories are going to give you the information and tell you the truth. We know that’s not going to work. We know that’s not the way people are going to behave. Therefore, we need to have some scope in the system for human beings to respond and figure out ways around it.’ We still have the same thing today. There was a piece by Yuval Harari in the Atlantic about a year-and-a-half ago, saying that authoritarian capitalism is going to beat democracy because authoritarian countries like China are able to use all of these new technologies to run the economy far more efficiently and keep an eye on everyone. “One of the crucial questions we need to understand is how this world . . . of information overload is stressing and straining our political system” What commentators like Harari don’t get is the ways in which these systems are not only incapable of grasping the messiness of actual human social systems, but also able to actually exacerbate the flaws of central planning. For authoritarian countries, China in particular, you have these feedback loops between the categories that people are using to try and understand the world in the central committees, and the actual world they are trying to explain. We know how politics work in these systems. Very often, if you’re not implementing the thought of the beloved chairman, your superiors will decide that there’s something wrong with you and you’re obviously a problematic political element who needs to be eliminated. So the categories you use are likely to reflect the ideas of your superiors, even if you know that they’re wrong. The technologist Maciej Ceglowski describes machine learning as “money laundering for bias.” That can have terrible consequences if machine learning reflects the categories of official thought, and then interprets the policy consequences in terms of these categories too, so that bias compounds bias. This then creates incentives for ever more distorted ways of understanding the world which are implemented through these algorithms and which then create these feedback loops which get worse and worse, and lead, perhaps, to human tragedy, but also to these authoritarian systems not working in the cool, clean, beautiful and efficient way that pundits like Harari expect. There is a wonderful essay by Kieran Healy, a sociologist at Duke University. He notes that when we think about these vast systems of machine learning, we assume that they work as advertised, whether we evaluate them positively because of the wonderful things that they can do, or negatively because they’re creating new forms of authoritarianism and surveillance and control. In practice, we know they sort of work and also sort of don’t. We tend to overestimate the extent to which there’s a single overwhelming logic of efficiency that’s associated with them. Yes. I would also love Silicon Valley entrepreneurs to read Red Plenty en masse and to think to themselves, ‘Which aspects of this apply to how I think about the world, and what aspects of it do not?’ There are important and crucial differences, but there’s also a fundamental similarity between the optimism expressed by these young, excitable Soviet economists and central planners back in the 1950s and the optimism of Silicon Valley people today: that software is going to eat the world, and that this is a really good thing. I think it would be really useful for them to start wondering, ‘Okay, are there aspects of this which simply don’t work in the way we expect?’ And I think that Red Plenty really pokes at these questions in a very, very useful way."