The Unaccountability Machine: a review

In every branch of Timpsons there's a sign from the boss saying "our colleagues in this shop have my total authority to do whatever they can to give you amazing service." Dan Davies' wonderful The Unaccountability Machine is the story of why there are no such signs in other organizations.

It's because, he says, decisions are now made by processes and systems rather than by identifiable individuals, leaving little space for people to use discretion. The result is the emergence of "accountability sinks", complex systems wherein it is difficult to pin blame onto specific people. This is why public inquiries into organizational failures take years and end up making hundreds of recommendations.

Rather than even try to look for accountability, says Dan, we should follow the advice of cybernetician Stafford Beer and not peer into the black box but instead judge organizations by their outputs. Hence the phrase "the purpose of a system is what it does", or POSIWID. Just as market outcomes are independent of individuals' intentions - selfish people can sell great products and nobody intends stock market crashes - so too are organizational outcomes. To take Dan's example, academics might merely want to expand knowledge, but the outcome of academic publishing is citations for academics and profits for John Wiley and Elsevier. POSIWID.

Such accountability sinks are not necessarily bad things. Embedding knowledge in a company rather than merely within individuals permits it to do difficult things and to cope easily with the departure of even the best employees: organizational capital is important. As Alfred North Whitehead said, "civilization advances by extending the number of important operations which we can perform without thinking of them."

But there's a downside. One of the great strengths of The Unaccountability Machine is its history of cybernetics. This teaches us to regard organizations as information-processing devices. And accountability sinks, when combined with striving for profit maximization, can result in disastrous losses of information.

One of Dan's targets is the leveraged buy-out industry. When a private equity company buys a firm and loads it up with debt, it is setting up a system in which the only information that matters is: will we be able to meet the next interest payment? That can force managers to increase efficiency. But it can also cause them to neglect longer-term investment, leaving them with shabby places like WH Smith or KFC that can't compete against rivals.

In a similar vein Roland Bénabou and Jean Tirole have shown (pdf) how "distorted decisions and significant efficiency losses" arise when companies incentivize short-term performance such as sales targets or traders' profits to the neglect of longer-term goals such as investment in training, R&D or goodwill. In cybernetic terms, such companies prioritize information about short-term profits and throw away that about the longer-term.

The classic case of this was of course banks' risky behaviour in the run-up to the financial crisis. Charles Prince, then CEO of Citigroup, said in 2007:

When the music stops, in terms of liquidity, things will be complicated...But as long as the music is playing, you’ve got to get up and dance.

Prince knew Citi was heading for trouble but he was incentivized to discard this information and to focus instead on short-term profits. As Dan says:

Any system which is set up to maximize a single objective has the potential to go bonkers.

The inability of companies to process information wasn't confined to banks, though. Paul Ormerod and Bridget Rosewell have found (pdf) that the pattern of corporate failures is similar to that of the extinction of species, and conclude that "firms have very limited capacities to acquire knowledge about the likely impact of their strategies."

Everybody will learn a lot from Dan's book, and I urge you all to read it.

That said, I have some quibbles. I'd have liked a little more clarity about the precise link between the cybernetic movement and the rise of accountability sinks? Did companies build such sinks deliberately, or did they just emerge (as so many things do) through trial and error? I'd have welcomed some case studies of particular companies to tell us.

A bigger quibble, however, lies in the class aspect. Dan says that the "great unremarked class struggle" of the 70s and 80s was that between capitalism and managerialism, which managers lost. Maybe they did then. But since the 1990s the UK has seen rising CEO pay amidst quite modest stock market gains, suggesting managers have fought back.

And this doesn't mention workers. The rise of accountability sinks was accompanied not only by big rises in CEO pay but also by worsening relative pay and conditions for many workers - be it those who suffer from the power-biased technical change described by Frederick Guy and Peter Skott or frontline workers who must face customers' frustration when impersonal processes lead to them being bumped off of flights, having their trains cancelled or having self-service checkouts that don't work.

If we apply the POSIWID principle to accountability sinks themselves, we'd infer that their purpose is to generate inequality.

There is, though, a massive paradox here. Corporate systems and processes have reduced employees and managers' freedom, agency and autonomy. And yet these values have for centuries been highly prized. "Free will is intrinsically the noblest thing we can have, because it puts us (in a way) on a par with God" wrote (pdf) Descartes in 1647, a view followed by many, including cheerleaders for capitalism such as Hayek and Friedman.

It was of course Marx who saw this paradox. The "laws of capitalist production" he wrote, take the form of "external coercive laws having power over every individual capitalist." And for the worker:

His labour becomes an object, an external existence...that it exists outside him, independently, as something alien to him, and that it becomes a power on its own confronting him. It means that the life which he has conferred on the object confronts him as something hostile and alien.

In this way, social relations between men become "in their eyes, the fantastic form of a relation between things." Man-made systems come to dominate and oppress their creators.

Perhaps Marx's greatest criticism of capitalism was not that it generated inequality and injustice - which existed long before capitalism - but that it created oppression whilst promising freedom.

It's not just this, however, that brings the legitimacy of capitalism into question. Dan's book raises a question posed 50 years ago by Stephen Marglin: what do bosses do? We used to be able to say that they invented things and built companies, but in accountability sinks they are mere functionaries presiding over processes they barely understand. People might admire and defer to entrepreneurs, but why do so to overpaid bureaucrats?

This was the question raised by Joseph Schumpeter back in 1943 when he wrote (pdf) that large corporations replaced entrepreneurs with "rationalist and unheroic" bureaucrats who are "ill equipped to face the problems, both domestic and international, that have normally to be faced by a country of any importance," thereby anticipating the political backlash against managerialism:

The capitalist process, by substituting a mere parcel of shares for the walls of and the machines in a factory, takes the life out of the idea of property...Dematerialized, defunctionalized and absentee ownership does not impress and call forth moral allegiance as the vital form of property did. Eventually there will be nobody left who really cares to stand for it—nobody within and nobody without the precincts of the big concerns.

People could tolerate unfreedom, inequality and rule by faceless bureaucrats when capitalism was at least delivering higher living standards. But after almost two decades of stagnation, the question becomes more forceful: what's the point of capitalism?

Comments

[Kevin A. Carson]

As James Scott pointed out, Taylorism was invented to minimize worker discretion and reliance on the worker's distributed knowledge -- even at the cost of reduced efficiency.
The reason is that discretion cannot be safely entrusted in someone who knows their interests are directly opposed to yours.

[Blissex]

«Did companies build such sinks deliberately, or did they just emerge (as so many things do) through trial and error?»

That often is a false dichotomy: a lot of things emerge semi-spontaneously, and then those that are welcome by whoever is in charge get endorsed and expanded, and those that are unwelcome get opposed and repressed. So my impression is that many things are both spontaneous at the beginning and become deliberate later.

Consider "wokeism": it used to be the spontaneous obsession of a fringe, then when corporate executives found it was useful it became endorsed and expanded by corporate media and corporate HR departments and corporate "sponsored" politicians.

[Blissex]

«Corporate systems and processes have reduced employees and managers' freedom, agency and autonomy.»

For whom? Whether it is for everybody or whether they enhance the power of those who can define "systems and processes" to suit themselves is a big deal.

«And yet these values have for centuries been highly prized.»

For whom? They have been prized for "people of quality" and not so prized for the "little people".

«that brings the legitimacy of capitalism into question»

I guess that our blogger reckons that capitalism superceded feudalism thanks to its greater legitimacy and not because of a hard class struggle by capitalists against feudalists where the capitalists won because they were stronger rather than more legitimate.

What brings ("private") capitalism into question is whether it is still the "natural" social form for the industrial mode of production now that the industrial mode of production is now longer expanding thanks to increasing and more efficient use of mineral fuels.

[Blissex]

«"rationalist and unheroic" bureaucrats who are "ill equipped to face the problems, both domestic and international, that have normally to be faced by a country of any importance,"»

This and other parts of this blog post seem to me to refer implicitly to the idea of "hypernormalisation", where real world complexity has become so great that the ruling classes of developed countries have given up trying to handle it and have tried to manage it by repression and deception.

https://en.wikipedia.org/wiki/HyperNormalisation

[rsm]

Why isn't it as obvious to anyone else that the purpose of capitalism to manufacture money (capital)?

In 2008 when the market had a collective panic attack, why shouldn't the Fed have created the liquidity that the private sector had, irrationally and emotionally, withdrawn due to contagion and spreading panic effects (not due to changing values of the underlying assets themselves which were minor compared to the contagion effects on assets that were good before the crises and are good again now)?

Why should traditional old economic growth be our goal? Why not encourage finance as offering distractions that employ humans in less destructive pursuits of fictitious goods, rather than trying to make money by creating real goods that inevitably come with environmental externalities that are far more consequential than mere stock trading?

Can financial capitalism save the planet by distracting enough people who would otherwise mess up the real world, with the trading of purely virtual products?

[Boy]

I’ve been reading the book of late and it’s helpful to see your thoughts here. The most thought provoking and alarming insight for me is that these systems are under nobody’s control - and in a sense are artificial intelligence, which in complex systems, we don’t really understand either. Some points to add:

Firstly, in the context of climate change, this is interesting: leaders posture, creating a pretense of control, but in reality even Biden, or Starmer, or Xi have as little power as a pet goldfish over emissions levels. They are as much bystanders, despite the nice speeches and suits, as you or I.

Secondly, there is a sort of meta interaction of out of control systems. Which feels like chaos. Can we apply POSIWID to this?

Lastly, POSIWID has a feel of self-fulfilling prophecy. With a shrug of our shoulders, we resign ourselves to the little agency we have and we thereby participate in the prophecy, leading to the question: is a system of thought which encourages people to believe it is outside of any man’s control, the very purpose of that system? POSIWID!

[Boy]

@Blissex

> where real world complexity has become so great that the ruling classes of developed countries have given up trying to handle it and have tried to manage it by repression and deception.

I’m not sure that for most participating bureaucrats it’s deliberate repression and deception, but more a case of their believing the stories they’re told due to conformist personality-type, possibly combined with a cognitive dissonance. As humans we cling to beliefs in our agency and free will. But these might well be fictions.

[Blissex]

"I’m not sure that for most participating bureaucrats it’s deliberate repression and deception, but more a case of their believing the stories they’re told"

There are huge differences between ruling, governing and administering and modt bureaucrats merely administer so which stories they believe do not matter.

The trusties who govern and make short term decisions and the masters/oligarchs who make long term decisions are quite deliberate and well briefed by strategists and think-tanks.

[Jan Wiklund]

Another book of the same theme is Jerry Muller's The tyranny of metrics: When the bosses don't have any clue at all, (because "boss" is thought to be a profession in itself that doesn't need any knowledge of what the organization is doing) they rule by numbers, like the drunk looking for his key at the lamppost because it is lightest there.

Old technocrats were at least raised in the businesses they led, and knew how they worked, by silent knowledge. They were replaced by economists when the rentiers took over in the 70s-80s, and they had to rely on automatic, and therefore unaccountable, processes which they didn't understand.

[Blissex]

«they had to rely on automatic, and therefore unaccountable, processes which they didn't understand»

For people of power and wealth "understand" is a rather secondary concern, "makes me money" is a rather stronger concern, and they made a lot of money over the past few decades. Lots and lots of money, POSIWID!.

What is still surprising for me is how many people including our blogger talk about abstract generalities like "austerity", "decline", etc. despite the overwhelming evidence that a large minority of the population, those in power for the past 4-5 decades, has been doing very well. POSIWID!

The side effects on others like employees and customers are rather less important than even "understand" for those who have been in power during the past 4-5 decades.

«When the bosses don't have any clue at all, (because "boss" is thought to be a profession in itself that doesn't need any knowledge of what the organization is doing)»

Which "bosses"? Perhaps operations managers "don't have clue at all", but the private fund managers, their accountants and lawyers, and the CFOs they place into their targets are as a rule very sharp and competent and understand very well all the fine details of asset stripping, tax loopholes, offshore havens, shell companies, political donations, PFI deals, accounting dodges; these people have been making a lot of money.

[Boy]

@Blissex: I simply can’t accept there is a group of oligarchs somewhere cynically and intentionally working to repress and deceive. As Day’s
book suggests, the truth is worse: nobody is in control. In concession to you, maybe the great deceit are people with power pretending they do in fact have a modicum of control when in fact they are as clueless as the rest of us.

[Blissex]

«I simply can’t accept there is a group of oligarchs somewhere cynically and intentionally working to repress and deceive.»

But that does not reflect my argument either which is that there are *several* cliques of oligarchs trying to control things in different but similar ways, and not always but often they succeed in controlling many important things, in particular those most of them agree on; it is not entirely by chance that thatcherism has lasted for decades, and that there were "everybody who is anybody against Corbyn" attacks in 2015-2019, or that after 2016 many of the newspaper editors who campaigned for brexit were almost stealthily replaced by others who supported remaining.

«As Day’s book suggests, the truth is worse: nobody is in control.»

Control is not something that is either 100% or 0%, "control" does not mean that every single whim gets immediately done, it can be good enough for some clique of oligarchs to get 30% over the decades of what they want, by being skilled and persistent operators who support trends that help their interests and oppose trends that damage their interests.

I am a bit astonished by people who think that either some clique has absolute control or nobody has any control, because they seem to have never worked in a corporate office, an university, a government department.

[Blissex]

«nobody is in control»

While I reckon that some cliques have partial control (even if sometimes they work at cross purposes), I am more sympathetic to a different absolutist claim that "nobody knows anything", because when confronted with complex systems it is very difficult to ascertain their state, never mind figure out how to achieve some outcomes.

However I reckon that some cliques keep winning beyond the odds, so there must be a way to achieve partial control without knowing much (or even anything), and also by observation of some skilled operators I reckon that can be done with "support trends that help their interests and oppose trends that damage their interests".

The absence of a magic wand of perfect control or a tome of complete knowledge can be partially overcome with persistence pushing.

[Blissex]

«astonished by people who think that either some clique has absolute control or nobody has any control, because they seem to have never worked in a corporate office,»

Hank Greenberg of AIG/Starr apparently once said "All I wish for is a little unfair advantage", for example.

Most experienced operators probably realize that having just 5% better "luck" than competitors is enough to become fantastically rich over some time, and it is also rather safer than absolute control or power.

«being skilled and persistent operators»

The campaigns against Sanders, Trump, Corbyn, Johnson for example did not seems "entirely spontaneous" to me...

"gutta cavat lapidem", and big stones can be cut with high-pressure water jets, and there are many options in between drops and high pressure water jets. :-)