There are many ways to understand the political chaos of our time. Many of those recall the famous folktale of the blind men and the elephant, while a handful offer more comprehensive takes. With its focus on the problem of information overload, Dan Davies' book “The Unaccountability Machine” provides an unexpected example of the latter, with a distinctive twist: Davies is a former banking regulator with a clear sense of what needs to be done to restore a sense of order on more equitable foundations. 

I referenced “The Unaccountability Machine” in a recent article asking what a progressive alternative to Elon Musk’s DOGE might look like. That hardly did justice to the value and complexity of Davies’ analysis of how our world has come unstuck since the 2008 financial crisis, and how the triumph of neoliberal economics led to what he calls a “brutal class war” and global chaos.

Davies writes that his central subject is “the biggest problem of modern industrial life — the problem of being overloaded with information, of ‘trying to get a drink from a firehose.’” But as he unfolds his explanation, it has much broader application in understanding our current worldwide political situation. He tells the story of three successive intellectual and organizational revolutions: the all-but-forgotten managerial revolution, the promising, but partially derailed cybernetic revolution — which could have made sense of the managerial revolution and built on it constructively — and then the neoliberal revolution that overtook us instead.

There’s much more to the book than that, including a reflection on the protean musical genius Brian Eno, whose work has involved cybernetic ideas for more than half a century. Alas, to avoid information overload, I set Eno to one side for this conversation with Davies, which has been edited for clarity and length.

Your book is about information overload, and specifically the problems of unaccountability that come out of that. You start by developing the idea of an “accountability sink.” What does that mean, and how does it work? And how is it sometimes beneficial?

The accountability sink was the original concept that I started thinking about when doing the research for this book. It's just a social, organizational, legal and managerial mechanism whereby a decision is created that has no identifiable human being as its owner. You have a decision that nobody seems to have made, and consequently nobody can be held accountable for it. I wanted to write a book about how terrible this all was and what bad people managers were. Nobody took accountability for anything anymore, and isn't that terrible. 

I spent about a year writing, and then I got an attack of intellectual honesty which caused me to be delayed by about a year, because I started to realize that actually you can't run a modern industrial society on the basis of individual human beings taking all the decisions and being personally accountable for them. You have to build systems, because otherwise you can't run anything remotely as complicated as a modern industrial economy. Also you need consistency in some things. I talk about people avoiding individual accountability and delegating their decision-making responsibilities to a written policy — but in a court of law, for instance, that's exactly what you want the judge to do. There are plenty of times you don't want to be exposed to the individual prejudices and caprices of human beings. 

So where did that lead you?

That made me start thinking, well, is there any way we can distinguish the good or bad accountability sink, or arrangements where someone is actually doing something necessary versus someone just trying to pass the buck. That's what got me into a deep and long involvement with management cybernetics, Stafford Beer and all this wonderful 1970s information economics that seemed to get abandoned when economics took its right turn with Milton Friedman and the neoliberal revolution. I think it's all related. 

OK, so what is management cybernetics, and how does it help clarify unaccountability problems? 

Cybernetics is basically information theory or the mathematics of information applied in contexts of control. In electrical engineering there were clear applications for this kind of control theory and cybernetics. But then people started saying, “Could we apply this to human systems?” 

When you start scaling up from an electrical engineering problem to a human system, you have the problem that you no longer have a clear measure of information. You can no longer think about things in terms of kilobytes or megabytes per second, because the amount of information is context-dependent. Also, if you start thinking about the complexity and the information in any human system, it's such an astronomically big number that it starts not making any sense to talk about it that way. 

So the question is, can you still say anything that is mathematically useful and mathematically valid? Everyone realized the answer is yes, you can, because all these things have the exact same analogies with economics, and there's a lot you can do with indifference curves and supply-and-demand curves, even when you know that you're not actually going to be able to measure human preferences. 

The first guy who started doing this practically was the British management consultant Stafford Beer. After World War II he got a job in a British steel company, and started doing what they then called operations research, which is just applied mathematics in the context of attempting to improve the efficiency of processes. He started looking at patterns of communication and organization, and spotting bottlenecks in the same way that you’d spot bottlenecks in the production process. 

The fundamental law of motion of cybernetics is that if you have a control system, it has to have at least as much complexity as the thing it's trying to control. If you are trying to model or regulate something at a higher level of complexity than you're capable of representing, it's not going to work. The system is going to blow up and become unregulated. 

In a management context, that's easy to recognize. Everyone who's worked in real life on a management consulting assignment will say the No. 1 problem of all managers is that they have a flood of information coming in, and they need to trim that down to literally make it manageable. so you attenuate and reduce the flow of information to something you can deal with and control. 

One building block of this approach is the famous concept of a black box, where you know the inputs and outputs, but not what’s going on inside. Could you explain the importance of recognizing black boxes, and how systems are built around them?

This is something that everyone does all the time, because you can't function in any other way. You actually gain a lot by paying conscious attention to what you're doing. 

The example I used, talking to some investment managers a couple of weeks ago, is that you have a government treasury secretary who needs to finance fiscal deficits by selling treasury bonds. They don't know the mechanics of how treasury bonds are bought and sold — they just know that there's a treasury bond market and they need to get some dollars from that. One layer down, you've got the bond traders who buy and sell the bonds. They do not usually care about the mechanics of how those trades are settled, and how the money flows between different bank accounts. There's another group of people who have to deal with that. They don't care too much about how the telecom network works, and even the telecoms people don't necessarily care how the electricity is generated. 

At one level, financing the U.S. deficit or managing His Majesty's treasury is a set of instructions to move electrons over wires, but at every level, the people taking the top-level decisions don't care about the electrons. Everyone is working at a level of representation that they can handle, and everyone is treating the level above them as a black box that gives them instructions and the level below them as a black box that does the things that they say will happen. 

That's what everyone does, because you can't do things any other way. If you start trying to contemplate the inside complexity of anything, beyond a very simple textbook example, you will go raving mad, or more likely you'll spend all your time doing analysis and never get any work done. 

Another important concept that emerges is that of a “resource bargain,” which mediates how the different parts relate to each other. Explain what that is and how it functions. 

The resource bargain is really the principle of the black box tuned into an active method of management. The black box, being defined as part of a subsystem, is allowed to organize itself however it likes as long as it fulfills its function in the system. The resource bargain tells parts of the system, "You have autonomy to do what you want, as long as you only use the agreed level of resources in doing so, and as long as you perform your function in the system of taking information from these sources and outputting decisions via these sources." 

It's very similar to the normal principle of management by exceptions: As long as something appears to be working normally, and appears to be performing its function without demanding too much finance, without having too many people and without demanding too much time and effort of other managers in the system, then you let it get on with its business. 

So this is all weird 1970s electrical engineering jargon for some straightforward commonsense management theory: Set clear objectives, set budgets, don't micromanage. All these things can be traced down to mathematically rigorous principles about information processing. 

Beer's management model has five organizational levels, which include operations, regulation, optimization and intelligence. At the top level is the one he calls “identity” or “philosophy,” which seems like the most important to understand. 

Yeah, in his model you have the here-and-now management system responsible for optimizing things as they are, and you have the intelligence system that is responsible for change and reorganization. You can't fail to respond to the outside environment, but if you try to restructure too rapidly or you change things too much, you break the organization. So you need an ultimate court of appeal to balance the here-and-now with the future and the outside. 

He calls that identity or philosophy, because it's in managing the balance of present and future where the identity of the organization is created. In setting the principles of how it balances information that's coming in now with the information it expects in the future, that's how an organizational system defines its purpose. 

With that we move from the managerial revolution and the cybernetic revolution — which tried to make sense of the managerial revolution — to the neoliberal revolution of the 1970s onward. Your essential insight about neoliberalism, as I read it, is that it emerged in a time of crisis with a strong identity tied to systematically throwing away, or “attenuating,” an enormous variety of data, basically everything except present-tense costs and prices. That was a way to simplify the world for top decision-makers who faced what Alvin Toffler described as "Future Shock." Is that a fair reading? 

It's a very fair reading. I'm not sure I could have put it better myself. The one thing I would add is that this was class war. This was real, brutal class war between the owners of capital and the managers of enterprises. 

Around the time of the 1970s when the leveraged buyouts, Milton Friedman, all those people got going, there were still plenty of people like J.K. Galbraith and Herbert Simon who were writing about the world of industry as if we were in a new period where “capitalism versus socialism” was no longer particularly relevant, because the modern industrial world was controlled by a techno-structure, and the people who were now interesting to look at were not the capital owners, the bourgeoisie and the investors, they were the managers, soldiers and civil servants who were the people who actually controlled the means of production in the world. That was something in the air at that time. 

Now we come forward to 2020 and it seems almost weird or quaint to believe that anyone could be controlling industry other than shareholders and investors. Everyone has to agree, it's really been drummed into us — and I went to business school, I've been through this indoctrination process — that everyone works for the shareholder. 

---

Want a daily wrap-up of all the news and commentary Salon has to offer? Subscribe to our morning newsletter, Crash Course.

---

There's even this doctrine that everyone has a fiduciary duty to maximize value to the shareholder, which is an interesting thing to say, because it's pure ideology. It's simply not true. That's not what's the fiduciary duty of a manager means. There is no necessary duty to maximize anything. There are certain duties that you owe when you are managing a company that someone else owns, but it's nothing like the ideology taught to us post-Milton Friedman. 

So the neoliberal revolution was great for some managers because it meant you had to think about a lot fewer things. You only have to think, if the number goes up, if the share price goes up, then I'm doing well, which solved a huge problem of trying to run a complex economy in an incredibly complicated world. 

But while it was good for some of the managers, it was absolutely fantastic for non-manager investors because they were being put back in charge. And lots of the neoliberal process and particularly the leveraged buyout revolution, in my view, was explicitly disciplinary. The vast growth in debt and debt-financed takeovers had a pretty clear purpose, which was to discipline the managerial class and to remind them that they had to work for the investor class. 

It also split the managerial class: You have CEOs and other top management getting rewarded with stock options, so they become completely aligned with the owners of capital.

Absolutely. At the same time, the middle managers, the men in the gray flannel suits, on occasion they got doses of rhetoric that the Cultural Revolution would've been proud of. They were deadwood. They were the inefficiencies in the system, and they had to be cut out ruthlessly. You really could, at some point, do a quiz: Who said this, was it Mao Zedong, or some guy working in a leveraged buyout firm? 

I think that was kind of a crisis, because the middle managers were the brains of the firm, and what we actually did over that period of 50 years was to just sit around cutting out bits of the cerebellum of the global economic system, and try to work out how much of it we could do without. 

Which got even worse when it came to government.

Yeah, absolutely. I still don't quite understand why Friedmanism caught on so much in government, except that it was such an incredibly strong and simple ideology. The market is a fantastic accountability sink. If you’re worried about making decisions and worried about being held accountable for the decisions you make, it's great to have the bond market there. You can now blame everything on the bond market. You can effectively pretend that all the things you are doing are no longer political choices, they're just the inevitable physical consequences of the market system, which covers absolutely all possible things.

That system seemed to work — at least to those on top — right up until around 2008, with the financial crisis and a bunch of other things that started to go visibly wrong. You had a wave of anti-democratic  movements, things fell apart. Regarding that, you say: "The medium itself is the message; what liberal society ought to be responding to is the fact of mass distress, not its content." Can you say a bit more about that, and how you think society should respond?

Well, the first is a lot easier to do than the second. My diagnosis is that 2008 was the year when the neoliberal revolution reached the point where too much information processing power had been cut out of both corporations and governments. The use of debt is a greatly underappreciated information-processing technology, because if you're going to load something up with debt, you immediately reduce the information set down to the question of "Can we service the debt?" 

So we built up a system which had to handle increasing complexity. We had cut out a lot of its ability to manage that complexity by forcing it to only care about prices, quantities and returns on investments. And 2008 was kind of the crisis moment at which those lines crossed and the system became unable to regulate itself. We've been living with the consequences of that ever since. 

Meaning what?

My diagnosis of the last 10 years in politics is that everyone is living in Alvin Toffler's world of future shock. Everyone is living in a world where they feel out of control because they correctly perceive that they are part of a system which is no longer able to handle the information or handle the complexity that it has created. 

Into that gap step lots and lots of people who have a family resemblance between them across wildly different geographies and cultures. All of them are saying, "This can be made simple. I have a simple solution to this. I am going to bring things back" — you know, in the U.K. we had Brexit, which literally had the slogan "Take back control." All of these populist authoritarian figures are all trying to say that the world can be made simple and comprehensible again. Which is kind of tricky because you can't turn back the clock. We can't go back to the 1970s and do things over again. 

So back to that trickier second question: What to do?

We have to learn to deal with the complexity of the system that we've got in different ways. It has to mean less use of debt, because the use of debt in the economy just smoothes things over and makes systems fragile. Similarly, we need to have less emphasis on shareholders and investors as the overall goal of economic activity, because we can't afford anymore to have organizations that are concentrating on single values of price and quantity.

People have to be given space to make longer-term plans, to pursue a greater diversity of objectives, and shareholders and investors are going to have to deal with that. Which is, politically, a difficult pill to swallow when you've been the ultimate goal of the entire system for 50 years, but this is how economics used to be. Herb Simon used to write that profits aren't the goal of the company, they're one of the constraints on the company. The company has things to do, and the only reason it cares about profits is that it has to give its investors a reasonable return on their investments, not to divert all of its efforts towards maximizing it. 

But at the most fundamental level, what needs to be built back is not accountability per se. It’s the broken feedback link, where there is no longer any way to communicate from the bottom level of the system to the higher levels of decision-making. It’s that feedback link breaking that becomes intolerable. Firstly, because without that feedback link the higher levels of the management system are working in an information environment which will inevitably depart from reality. The systems will go crazy. But also, from the point of view of the people at the bottom of the system, being ignored is psychologically intolerable. 

There have got to be more and better ways for people to communicate with those who make decisions about them. Building back those communication links is difficult. I suspect it's going to be different in every different system. But what we have at present, across the world, is a management system that just doesn't listen to its customers, or doesn't listen to its voters, in anything like the way it used to. That communication link is broken, and in some way that needs to be restored. 

One thing you talk about is using — you call them “robots,” but we could say algorithmic systems — to help gather that information. I’m partial to something that's less opaque to the ordinary citizen, things like citizens' assemblies or “micropublics,” which are more popular in Europe but not well known in America. They have shown themselves able to facilitate depolarization.

Yes! Not just micropublics, but consultations in general. I'm working at the moment on problems of infrastructure planning and procurement, and I think the key to the whole problem of sclerosis is not "too much regulation." It's too little communication, which means that people just dig their heels in and use the regulatory and legal system because they have no other way of communicating their priorities.

Finally, what’s the most important question I didn’t ask? And what’s the answer?

I'd say the important question is "How is this any different from the rest of the management science literature, which is pretty depressing stuff?" My answer would be that looking at things in information terms really clarifies your analysis. It gives you a language to talk about things which are otherwise difficult to describe. It helps you understand why things don't work. Usually, the answer is that they have been asked to deal with a problem they're not capable of representing.