My go-to example of a whole mesh of "accountability sinks" is... cybersecurity. In the real world, this field is really not about the tech and math and crypto - almost all of it is about distributing and dispersing liability through contractual means.
That's why you install endpoint security tools. That's why you're forced to fulfill all kinds of requirements, some of them nonsensical or counterproductive, but necessary to check boxes on a compliance checklist. That's why you have external auditors come to check whether you really check those boxes. It's all that so, when something happens - because something will eventually happen - you can point back to all these measures, and say: "we've implemented all best practices, contracted out the hard parts to world-renowned experts, and had third party audits to verify that - there was nothing more we could do, therefore it's not our fault".
With that in mind, look at the world from the perspective of some corporations, B2B companies selling to those corporations, other suppliers, etc.; notice how e.g. smaller companies are forced to adhere to certain standards of practice to even be considered by the larger ones, etc. It all creates a mesh, through which liability for anything is dispersed, so that ultimately no one is to blame, everyone provably did their best, and the only thing that happens is that some corporate insurance policies get liquidated, and affected customers get a complimentary free credit check or some other nonsense.
I'm not even saying this is bad, per se - there are plenty of situations where discharging all liability through insurance is the best thing to do; see e.g. how maritime shipping handles accidents at sea. It's just that understanding this explains a lot of paradoxes of cybersecurity as a field. It all makes much more sense when you realize it's primarily about liability management, not about hat-wearing hackers fighting other hackers with differently colored hats.
One example that's missing from the list is the TV series 24. A recurring plot point was that, yes, of course torture is bad and it's against the rules and we don't do it, etc etc, but it just so happens that here is such an exceptional, unprecedented, deeply urgent emergency situation where we need to have the information now or horrible things will happen, we need the hero who breaks the rules and goes on torturing anyway. [1]
Fast-forward a few years and you find there were in fact many such "heroes" in reality - in Abu Ghraib and in the Black Sites - and the situation weren't exceptional at all.
So accountability sinks can also be used as calculated ways to undermine your own ostensible ethical guardrails.
[1] https://www.theguardian.com/tv-and-radio/2017/jan/30/24-jack...
I'm now in a stage of my consulting career where I sometimes really get called into big organisation just to find out, that whatever they need to do is already panned out and they all want to do it! Still they call me cause ... it's a big decision and the "higher ups" (which quite often are not even part of the workshop/session then) want an external expert voice. cause the responsibility for this decision lies with them and they can not share it up or sideways, so they share the responsibility partly external.
As the plan quote often (not always) is already very good I mostly end up making sure the goal is measurable in a quantitative and qualitative way, trends towards to and away from the goal are visually available and distributed , and its clear who is responsible to look and report them.
I always remind myself when I have to go to the DMV[0] that I should plan on leaving with nothing more than another action or set of actions to take. I never enter the DMV expecting to complete a process, and the workers behind the counter always have this visible, visceral response when I DONT lose my fucking mind at their response to something. When I continue to be pleasant and understanding it’s like they suddenly come alive. It’s a depressing state of affairs because I understand exactly what they expect and why.
0: for non-Americans and for Americans from other states that may use different terms, the DMV is the department of motor vehicles in many US states and is the central place to get your drivers license, take the drivers test, register your car, get vehicle license plates, etc. Many processes that have many requirements that often are unfulfilled when people show up asking for things.
When dealing with companies, small claims court can be an amazing tool to fix the "nobody is responsible so you hit a wall" issue. The court sends a letter to the company, and either the company figures out who is responsible for dealing with it, or whatever process for collecting unpaid judgements eventually deals with the company (e.g. the famous "sheriff comes to repo the bank's furniture" example).
For companies, this is also fine, because in most cases the built-in processes work well enough, and in others people just give up, that handling the escalations through their legal department is manageable.
Unfortunately, this approach only helps for the subset of cases where the issue is monetary and/or can wait (and only if it happened in a country with a working small claims system).
The squirrel example sounds terrible, but people don't realize the danger that moving pathogen-carrying specimen across ecosystems poses. Introducing a disease into your local environment can have devastating consquences for wildlife or farming or both.
Example: Dairy farms have strict rules about not letting anybody in who was abroad within the last 48 hours because of possible spread of foot-and-mouth disease. There are many such examples and similar examples exist for wild ecosystems.
So, while it may seem cruel to kill a few hundred squirrels, the precaution is justified. The "guilt", if there is any, is with whoever didn't ensure all the paperwork is in order.
Here another two of Sustrik's gems..
Anti-social Punishment: https://250bpm.com/blog:132/
Technocratic Plimsoll Line: https://250bpm.com/blog:176/
seems lesswrong has all of them, older and newer: https://www.lesswrong.com/users/sustrik?from=post_header
I once booked a plane ticket from my home town airport to another country. The purchase notification said something like "PVA" instead of "POV". I looked it up and turned out, the newly built airport that had this exact code was about to open. In a week or so, so I assumed that I'm indeed flying from the new one and forgot about it. The purchase was made through a booking aggregator similar to Expedia.
On the day of travel I took a taxi to the new airport, which is 40 km outside the city. The taxi driver couldn't care less about where I was going. Upon arrival, there was much fewer people than I expected but I shrugged it off. At the entrance though I was asked where I was going and if I was an employee. Apparently the new airport was still closed and my fight was from the old, still functioning one. The one with the code not shown in the ticket purchase receipts.
Panicking since it was only about an hour until departure, I took a taxi back to the old airport, which was a desperate 40-50 minute drive to only realize the plane had already left.
I was flying abroad, with a connection the next morning, about 10 hours later. So I thought that the problem could be solved by just arriving there by any other flight, which I booked almost immediately. However, the airline representative (yes, there was a human to speak to that I could reach easily by phone) told me that a no-show for any segment of the flight invalidates all subsequent ones. There was no way I could convince her that it wasn't my fault. Perhaps there was a rigid process in place that disallowed her from helping, even though I'd make it to the second flight on time.
I ended up buying 2 new tickets, of course more expensive and less convenient ones. This taught me an important and rather expensive lesson on why connected flights with a single airline are sometimes the worst.
Funnily enough, I was bitten by this rule one more time when I didn't show to a flight in to the country due to visa issues (it was covid time) and wasn't allowed on the flight out of it because I didn't show up to the 1st flight, the flights being 1 week apart - but booked in one go.
As to the previous situation, I managed to get compensated by the airline (not even the intermediary!) about a year later after posting a huge rant on Facebook and getting their attention to the situation.
I have a feeling that AI will be used to replace the folks that might get squeamish.
If I understand it correctly, that's what United Healthcare was doing, that got people so mad at the guy that was shot. He brought in "AI Denial Bots," so the company could knowingly cause the death of their customers, without having any "soft" humans in the process.
I read most of this agreeing with everything the author was saying, sometimes in a "I already thought that" but often in a "huh, that's a really cool insight." I quite like the style too.
As a Brit though, I was completely blindsided by the inclusion of Dom Cummings. I'd forgotten he existed. Seeing his and Boris' attitude to PPE provision discussed in a positive light without any mention of the associated scandal[1] made me a bit uncomfortable. Without getting too political, they claimed to have solved a problem, but whether or not it was a justifiable, sensible or legitimate solution is probably going to be debated for decades.
[1] https://en.wikipedia.org/wiki/Controversies_regarding_COVID-...
A few things came to mind as I read this.
1) About 8 years ago I was gifted a copy of Ray Dalio's Principles. Being a process aficionado who thought the way to prevent bureaucracy was to ground process in principles, I was very excited. But halfway through I gave up. All the experience, the observations, the case studies that had led Dalio to each insight, had been lost in the distillation process. The reader was only getting a Plato's Cave version. I used to love writing spec-like process docs with lots of "shoulds" and "mays" for my teams, but now I largely write examples.
2) I live in a Commonwealth country, and as I understand (IANAL), common law, or judge made law, plays a larger role in the justice system here than in the US, where the letter of the law seem to matter more. I used to think the US system superior (less arbitrary), but now I'm not sure. Case law seems to provide a great deal of context that no statute could ever hope to codify in writing. It also carries the weight of history, and therefore is harder to abruptly change (for better or for worse).
3) Are human beings actually accountability sinks? This is only possible if they are causal originators, or in Aristotlean terms, "prime movers", or have pure agency, or are causa sui. But the question is, once we subtract environment (e.g. good parenting / bad parenting) and genetics (e.g. empathy, propensity toward anger), how much agency is actually left? Is it correct for our legal and ethical systems to terminate the chain of causality at the nearest human being?
According to Dutch law, you lose your Dutch citizenship if you accept another nationality. The Dutch embassies (who are responsible for renewing Dutch passports abroad) are well aware of this law and have processes in place to refuse a passport renewal if you can’t provide proof of temporary residence in the country you reside in. The local institutions however, don’t have these processes in place and are generally not aware of this law because it only happens to a tiny little percentage of the population. And nobody updates the national registry with your new nationality because that’s the responsibility of local municipalities, not the Department of Foreign Affairs. So if you decide to simply renew your passport in the Netherlands instead of abroad, they’ll just give you a new passport because you’re still registered as a Dutch citizen at the local level and they don’t have a process in place to check your foreign nationality.
Don’t ask me how I know :) It is one of the few accountability sinks that doesn’t affect me negatively.
I think this is one of the most fascinating aspects of solo non-stop around the world sailing: You have no one to blame other than yourself. It puts you into a mindset that is unique in this day and age. The sailors, when interviewed after their ordeal, also mention it a lot.
Another fun one is asking for a higher salary - for obvious reasons moderately sized companies have formal systems that make it logically impossible to do on an employees initiative (the boss doesn't control salaries, payroll doesn't control salaries and all the formal systems point to the boss and payroll). The real approach is that a worker has to somehow convince one of the people with serious power to overrule the default systems.
But the important thing to recognise is there are always people who can overrule a given formal process and they are being held accountable to something. The issue becomes what their incentives are. In the success stories in this article (like the one where the doctor saves a bunch of people) the incentives lead to a good outcome when the formal system is discarded. In the leading ground squirrel example someone without doubt had the power to prevent the madness and didn't because their incentives led them to sit quietly in the background hidden from history's eye. Ditto the Nazi example - obviously there was someone (probably quite a few someones) who could have stopped the killing. They didn't override the system because they through it was performing to spec, and it is probably difficult to prove they were in hindsight because informal systems don't get recorded.
Interesting article, but picking Johnson and Cummings's handling of Covid as a positive example is a very odd choice, given their falling out and the numerous corruption allegations and parliamentary inquiries into their actions since then.
This reminded me of my favorite David Greaber book: The utopia of rules (https://en.m.wikipedia.org/wiki/The_Utopia_of_Rules).
Greaber, if I remember right, argues that modern bureaucracy started with efficient means of communication. He squares the Deutsche Post as the milestone, as they made the whole population available to be controlled. Now the state could send them letters, count them, enlist them in the military etc.. It's a brilliant observation: communication technology is the main tool of the bureaucracy. The tangent he takes fron there is even more brilliant: we have been heavily focusing and improving the communication tech (telephone, fax, tv, radio, internet, social media) but not necessarily the tech to reduce thr burden of work for the masses (robots!). If you would ask someone 100 years ago how the future would look like, people would almost invariably say they would need to work less in the future, abd at some point they invariably expected to have robots do all the work. Yet, all we got is smartphones that watch every movement of us, makes us available to the employer anywhere and anytime, hence more means to control us by state or, exceedingly, private bureaucracies. There's a reason why AI boom is happening, as this is the next tech on the bureaucracy tree.
This being said, none of these tech are bad by themselves. It is the shape they took and the way they are used in contemporary society. To tie with the OP: we have communication tools available to us that is billions of times more efficient and effective yet the customer service, or any interaction with any big corporation (as a customer or employee) or state got so much worse and impersonal. Impersonal as in, individual cases do not exist anymore, only policies. One could have expected to escalate a claim back in late 19th century by just writing letters and eventually get to someone, or even just show up at the offices of a company and get their problem resolved (this is still the case in developing countries). Can we expect this now?
> The card design only allowed for 24 characters, but some applicants had names longer than that. They raised the issue with the business team.
> The answer they've got was that since only a tiny percentage of people have names that long, rather than redesigning the card, those applications would simply be rejected.
Long names are a pain. This happened to me when I tried to open a bank account in Vietnam. Similarly bank tellers in China were always puzzled and needed to call supervisors when having to enter the information. Also airport auto gates frequently fail for me, and systems that want me to enter your full name in a form will reject my input more often than not. When I'm asked to sign my full name with my signature, it hardly fits and I need to write in tiny letters.
If I ever have children I'll name then with something short, with no special characters. Something like Tim, Kim, Leo... Otherwise they will always end up the edge case.
Interestingly if you’re denied a credit card in Europe a subject access request can be very helpful for understanding why
> Bad people react to this by getting angry at the gate attendant; good people walk away stewing with thwarted rage.
I disagree, slightly. We have to expect some degree of ethical behaviour from everyone, even those who nominally have no room to manoeuvre. If everyone in such positions were to disobey unjust orders the orders would eventually have to change.
Walking away stewing in rage does nothing except fill you with damaging hormones.
The KLM squirrel problem is arguably the opposite problem as all the other examples. It would have been a simple matter to call up KLM's corporate counsel and have them figure out how to both comply with the government's order and the country's animal welfare laws.
So really a case of not enough bureaucracy rather than too much.
As sad as things turned out for the squirrels it’s bizarre to worry too much about 440 squirrels dying in a country with lots of meat farming…
> If a culture of finger pointing and shaming individuals or teams for doing the "wrong" thing prevails, people will not bring issues to light for fear of punishment.
Not only this. I had a particularly toxic manager many years ago and during some major disaster he was fuming around the office demanding to know who was responsible. We had to spend time determining that instead of how to fix the issue. Someone might have been responsible for initiating the disaster, but he was responsible for extending it (substantially).
It's only related to what he wrote but it reminded me of something that low-key annoys me whenever I hear Americans talk about the Holocaust.
I know he only touches on it very slightly and indirectly raises a related point to what annoys me about most coverage about it.
It's pretty simply that the people that were systematically slaughtered during that time period were classified to be Jews, Gypsies and other "undesirables", but they were first and foremost German and identified as such. Nazi Germany didn't kill "other" people, it systematically alienated groups of the population to then eradicate them, by first walling them off to make communication impossible, then spreading enough propaganda to make the average Joe no longer consider them his neighbor.
Seeing the social climate all over the world change, chief among them Americas, does make me think this lesson hasn't been taken in whatsoever.
The first step to atrocities is always to cut of communication between the groups, and people nowadays are actively doing that themselves now - not artificially enforced like it was back then.
In my experience, the credit card example is _usually_ solved in a practical way which is still somewhat bad, but allows the person to at least get a card: They abbreviate one or both names in some way for the card.
As crappy as the system with its max length for people's names, it's common to allow first initial + surname. It also works very badly for non ASCII names - to my understanding, I _think_ people in East Asia just have to use romanisations if they want to have a Mastercard. This all sucks, but it's a bit more than "the card design" - it's quite fundamentally baked in to how the whole system works. There aren't a lot of systems out there which are based on more aged and legacy technology than card networks.
David Graeber has written a really good book about this exact topic and one that I highly recommend. He explores why and how bureaucracy crept up on us.
“ The Utopia of Rules: On Technology, Stupidity, and the Secret Joys of Bureaucracy”
The conclusion of Davies' second extract — about e.g. being bumped off a flight — is recognisable but the conclusions are actually wrong. The situation in these cases is actually more subtle. The person you're speaking to does normally have some capacity to escalate in exceptional cases. But they can't do it as a matter of course, and have to maintain publicly that it's actually impossible.
The people who get what they want in these situations are the ones who are prepared to behave sufficiently unreasonably. This is a second order consequence of 'unaccountability' that Davies misses. For the customer, or object of the system, it incentivises people to behave *as unpleasantly as possible* — because it's often the only way to trigger the exception / escalation / special case, and get what you want.
Generalizing here, but it’s a sign of a bad process when it’s thrown out in an emergency.
Crisis is when well-thought out, tested procedures should be used, at least as a starting point.
The TV series Yes Minster and Yes Prime Minister have a lot to say about accountability in the governments.
> Eventually, employees noticed a problem: The card design only allowed for 24 characters, but some applicants had names longer than that. They raised the issue with the business team.
I'm looking at you, ANA Mileage Club card! 24 characters should be enough for anyone according to their database. They even have a whole page dedicated to how you should work around it (I tried, this procedure & indeed it lets you truncate your name, but then you won't be able to associate any tickets you purchase in your real name with the card). https://www.ana.co.jp/en/jp/amc/reference/merit/procedure/in...
The moment I saw 250bpm, my mind took me to ZeroMQ and indeed it was Martin. Then ofcourse Pieter Hintjens came to my thoughts next. I just loved ZeroMQ but I don't use it anymore. Good to find Martin's blog and a great writeup.
Also... there was this discussion several months ago: https://news.ycombinator.com/item?id=43877301
Both that item and this item add the unique perspectives of the authors, but both are about issues raised by Dan Davies' Unaccountability Machine. So if you like this thread, you might like that thread.
> And it turns out that the German soldiers faced surprisingly mild consequences for disobeying unlawful orders.
Huh. Franz Jägerstätter was executed for refusing to fight in the war.
Reminds me of a Rickover quote:
"If responsibility is rightfully yours, no evasion, or ignorance or passing the blame can shift the burden to someone else. Unless you can point your finger at the man who is responsible when something goes wrong, then you have never had anyone really responsible."
I don’t see anyone mentioning it, so: I was disappointed that an otherwise interesting post was turned political.
> This is why even the well-off feel anxious and restless. We may have democracy by name, but if the systems we interact with, be it the state or private companies, surrender accountability to the desiccated, inhuman processes and give us no recourse, then the democracy is just a hollow concept with no inner meaning.
> You can't steer your own life anymore. The pursuit of happiness is dead. Even your past achievements can be taken away from you by some faceless process. And when that happens, there’s no recourse. The future, in this light, begins to feel less hopeful and more ominous.
> It’s eerie how much of today’s political unrest begins to make sense through this lens.
No, your past achievements aren’t taken away from you. When you’re wronged, you almost always have recourse, up to and including making a big stink on social media. Private companies aren’t meant to be political democracies. They’re in fact almost explicitly designed to be authoritarian, because it works well. You don’t suddenly live in a not-democracy just because the companies have a CEO and middle managers that set up processes.
I wish the rest of the post wasn’t called into question by this hyperbole, but it is. It makes some interesting points, but ultimately it feeds into a natural desire to be pessimistic. Which means it’s entertainment rather than an analysis.
I didn’t realize Martin is blogging again! Hurray!
> Bad people react to this by getting angry at the gate attendant; good people walk away stewing with thwarted rage.
Notwithstanding the rest of the column, this particular example brings the following thought to mind:
It could actually be argued that getting angry at the gate attendant is not a "bad people" response. Suppose that under those circumstances, the typical individual passenger would demand the gate attendant to either let them onto the flight, or compensate them reasonably on the spot, and if denied - even with a "it's not within my authority" - inform their fellow passengers, which would support the demand physically to the extent of blocking boarding, and essentially encircling the gate attendant until they yield (probably by letting the original passenger onto the plane), and if security gets involved - there would be a brawl, and people on all sides would get beaten. Now, the individual(s) would would do such a thing may well suffer for it, but in terms of the overall public - gate attendants will know that if they try to do something unacceptable, it will fail, and they will personally face great discomfort and perhaps even violence. And airports would know that such bumps result in mini-riots. So, to the gate attendant, such an order would be the equivalent of being told by the company to punch a passenger in the face; they would just not do it. And the airport would warn airlines to not do something like that, otherwise they would face higher airport fees or some other penalty. And once the company realizes, that it can't get gate attendants to bump passengers this way, it will simply not do it, or authorize decent compensation on the spot etc.
Bottom line - willingness to resist, minor ability to organize, and some willingness to sacrifice for the public benefit - can dismantle some of these accountability sinks.
a good "collective response" would be to deny the non-agency of the gate attendant. That is,
Another terrific write up on this subject is Jen Pahlka (cited in the article)’s essay on the “cascade of rigidity”.
> Bad people react to this by getting angry at the gate attendant; good people walk away stewing with thwarted rage.
> You ask to speak to someone who can do something about it, but you're told that's not company policy.
People somewhere in between realise that the point of the gate attendant (or Level 1 tech support person) is to shield management from customers, so you have to outflank the shield.
Being yelled at by a customer is bad for the Level 1 support person, although there's usually a policy in place for phone support that you can hang up if the customer is getting aggressive. What's much worse is saying to management "hey here's something you might want to look at" and being super yelled at by their boss for not doing their duty of keeping the customer away from the higher-ups. That kind of thing can get you fired.
But you can hack the system in many ways. The point is to find someone higher up without going through the person who's not allowed to help you, and without blaming them for doing their job.
Some possibilities: find the higher-ups on linkedin, speak to a company rep or executive personally at an event if your professional circles overlap, send a printed physical letter to someone in control, and so on.
Something I've seen work many times: if you're a student, find out about the university's management structure and ask for a personal meeting with the Dean of X of whoever sits above the department admin person who's assignment is "we've taken this decision, now make the students happy with it". A dozen students asking to personally speak with the Dean or President lets them know something's up and the shield was ineffective. Since there's usally some kind of statement of values about how the "student experience" is central to everything they do (read: "students are paying customers"), they can't just turn you away.
Does Kafka, the writer[1], come to mind for anyone?
[1] I dislike that I have to specify. I wish there were still only one common reference for this name.
Related discussion (517 pts):
https://news.ycombinator.com/item?id=41891694
https://aworkinglibrary.com/reading/unaccountability-machine
(A very short overview of Dan Davies' book, quoted in TFA, that came up with the term)
EDIT: complementing book mentioned in that thread
Cathy O'Neil's "Weapons of Math Destruction" (2016, Penguin Random House) is a good companion to this concept, covering the "accountability sink" from the other side of those constructing or overseeing systems.
Cathy argues that the use of algorithm in some contexts permits a new scale of harmful and unaccountable systems that ought to be reigned in.
I find some startup leaders really struggle with this if they come from a Big Corp environment. If you're a CTO or VP or Director of Engineering for like 20 people, you're actually going to have to decide things. Yes at your previous roles you could follow a flow chart or whatever but here you have to actually take accountability. Watch them tap dance trying to avoid committing to anything.
All the Republican Senators and Congressmen/Women .. I am looking at you.
You can't get a credit card because your name is too long?
You can't pass immigrations because you don't have a last name?
The future is not made for you, because progammers and designers didn't get requirements that match the diversity of this beautiful world.
It remindes me how someone I know often makes complicated food orders in restaurants (modfying or replacing items on the menu), and then they get disappointed or complain because their wishes are forgotten or screwed up. I never make changes to a menu item, because I assume they are unable to accommodate me (either due to stress, lack of intelligence/memory, bad process e.g. not writing down customers' orders etc.). As a result, I get disappointed less often on average - make your oder "compatible" with the realities of this world to avoid disappointment and stress.
There is actually an official procedure for U.S. Immigrations dealing with people who have names that cannot be split meaningfully into first/last names, e.g. some people from India. Assume your name is "Maussam", then you are permitted and expected to fill in that string in BOTH fields, first name and last/family name, when booking a flight or applying for visa. (A similar hack could be devised for names that are "too long".)
Overall, these examples are reminiscent of the movie Brazil (1985), which is about a dystopian future in which a plumber that helps people fix their toilets gets hunted as a terrorist because he didn't fill in the right form.
My theory is the world has been gradually converging towards the absurd state parodied in that movie.
Airlines are not the only ones that get less and less accountable. We should stop spending our money with companies that communicate with us using email spam services the address of which begins with [email protected].
>The SREs were accountable to the higher ups for the service being up. But other than that they are not expected to follow any prescribed process while dealing with the outages.
That's because hard work and being serious about your tasks do not get you promoted.
The pyramid on the dollar bill is built of human bricks which believe that they are free from repercussions of their actions under orders from above. But real karma is much more of a bitch than even HN moderation
Another major accountability sink is employment. Employee is shielded from financial responsibility for the damage he incurs while working. While he may be punished for disobeying orders or acting criminally, he's not financially responsible for the fallout (especially if he was only doing the things he was ordered to do and/or reasonable things). Doing a job is inherently risky behavior. If you are doing it in a context of financial amplifier (a company) in a regulated society that can quickly hunt you down and destroy your life if you misstep then in the absence of accountability sink protections barely anyone would be brave enough to get employed. That's also why LLC exist. To enable risk taking by promising to not hunt you to the bottom if you fail.
The reasons we suffer these accountability challenges are often rooted in that anyone holding someone else accountable, may experience negative consequences to self...and those are often estimated as too high to do "the right thing".
If the governing part at the time of the Nazi trials actually held each and every person involved accountable, would they win the next election?
If a company holds their employees to the actual standards laid out by their policies or guidelines, what would attrition look like? Would they immediately be short staffed critial roles? Would they loose a key employee at a very inconventient time?
These are the real reasons preventing us from holding people accountable.
The conclusion of Davies' second extract — about e.g. being bumped off a flight — is recognisable but the conclusions are actually wrong. The situation in these cases is actually more subtle. The person you're speaking to does normally have some capacity to escalate in exceptional cases. But they can't do it as a matter of course, and have to maintain publicly that it's actually impossible.
The people who get what they want in these situations are the ones who are prepared to behave sufficiently unreasonably. This is a second order consequence of 'unaccountability' that Davies misses. For the customer, or object of the system, it incentivises people to behave as unpleasantly as possible — because it's often the only way to trigger the exception / escalation / special case, and get what you want.