All of my IT experience has been in the private sector, except for one project when I worked for IBM and I was sent to the Department for Work & Pensions. That was in 2005. It was a highly educational time, but far from enjoyable. I was shocked by the client's attitude towards risk and accountability. Their managers happily took decisions I knew would wreck the project (if I were to comply) but protect their careers. I wrote about it here. I see that when I wrote it I did not mention IBM or the DWP, but I'm retired now, so what the heck.
On a wider note I've seen similar situations play out elsewhere in large corporations, though never quite as extreme as at the DWP. I occasionally had to give dual advice to executive management when I was an IT security manager; this is what you should do to protect your career, and this is what you should do based on the risk to the corporation we have identified and agreed upon. They were not necessarily the same. Sometimes it was better for a manager to make a loss on the account and annoy the client rather than turn a profit and keep the client happy. I've worked as an IT auditor in an excellent internal audit department for a major insurer. It was part of our job to detect, expose, and challenge such dysfunctional situations.
I've seen more than a few business people build their careers on the backs of bad data architectural decisions made by a consultancy that are pitched as "flexibility." This happens when e.g. marketing agencies design reporting systems for their clients. The cobbled-together result costs no more than the sponsor has in their budget, but keeps someone from the agency employed pushing buttons and manually carrying data. The low-risk and high-reward option is to replace all that with a modern system, but nobody with local authority has an incentive to do that.
From Gemini, very similar to many things you've already covered:
The idea that civil service behavior is primarily driven by the desire to minimize career risk—often termed "blame avoidance" or "risk aversion"—is a central pillar of Public Choice theory and modern public administration.
While several works touch on this, the two most "standard" academic sources that define this perspective are Christopher Hood’s The Blame Game and R. Kent Weaver’s foundational paper on blame avoidance.
1. The Modern Definitive Work: Christopher Hood
Book: The Blame Game: Spin, Bureaucracy, and Self-Preservation in Government (2011)
Christopher Hood is the leading scholar on why modern bureaucracies seem obsessed with "covering their backs." He argues that in an era of high-intensity media and political scrutiny, the primary incentive for a civil servant is not to achieve the "best" outcome, but to ensure that if something goes wrong, the blame cannot be pinned on them.
Key Concept: Negativity Bias. People (voters and politicians) react far more strongly to a single failure than to many successes. Therefore, civil servants prioritize avoiding a "minus" over achieving a "plus."
Three Strategies of Self-Preservation:
Agency Strategies: Creating complex organizational structures or "arms-length" bodies so it’s unclear who is actually responsible for a decision.
Policy/Operational Strategies: Relying on "protocolization"—strictly following rigid rules and checklists so that even if the outcome is bad, the bureaucrat can say they "followed the process."
Presentational Strategies: Using "spin" or "drawing a line" to distance oneself from a crisis.
2. The Foundational Paper: R. Kent Weaver
Article: "The Politics of Blame Avoidance" (1986)
Before Hood, R. Kent Weaver crystallized the theory that politicians and bureaucrats are motivated more by the desire to avoid blame than by the desire to claim credit.
The Argument: Since "voters are more sensitive to out-of-pocket losses than to equivalent gains," the safest career move for a public official is "The Good Soldier" approach: keeping a low profile and avoiding any policy that has a identifiable "loser," even if it has a high collective "gain."
3. The Classic Organizational Model: Anthony Downs
Book: Inside Bureaucracy (1967)
Downs provided the first rigorous economic model of how individuals behave inside a government department. He categorized bureaucrats into types, most notably "The Conservers."
The Conserver: This is the "standard" mid-to-late career civil servant. Their primary goal is convenience and security. Because they cannot "profit" from a successful innovation (unlike a private sector CEO), but they can be fired or demoted for a visible failure, their rational choice is to resist change and minimize any activity that carries career risk.
4. The "Satisfying Superiors" Model: Gordon Tullock
Book: The Politics of Bureaucracy (1965)
Tullock, a co-founder of Public Choice theory, argued that because there is no "market price" to measure a civil servant's success, the only way to get promoted is to please your superior.
The Impact: This creates a "telephone game" where information is filtered as it goes up the chain. Subordinates hide risks and failures to protect their careers, meaning the top of the hierarchy often makes decisions based on sanitized, "safe" information that doesn't reflect reality.
Thanks so much, some of those look good. I guess the problem is that the public choice school never really did much empirical work, so a lot of this stuff language is in the space of "stylised facts that everyone sort of knows" rather than "well investigated phenomena where people know a bit more about the precise conditions in which it does and doesn't appear"
The authors of the paper Marcus endorsed don’t seem to realize it, but they describe generative AI as, in effect, “an anti-viable system” (in the sense of Stafford Beer).
«particularly contexts of government and civil service decision making»
I have noticed in the past an amusing case of where careers have mattered a lot which is why the Royal Navy has arguably many more frigates than it needs:
My summary of that story is that promotions to flag ranks are/were traditionally decided on the basis of performance as captain of a frigate-size ship and to ensure uniformity of evaluation there had to be enough frigate-size ships to be able to evaluate all captains at that point in their careers and frigates were the cheapest option for that ship size. Actually I think a quite reasonable policy in peace time.
Not quite what you're looking for, but J.P. Marquand's 1949 novel "Point of No Return" is primarily (in terms of both plot and metaphor) concerned with career risk for its protagonist (a banker), and is an excellent if sometimes oblique exploration of same.
"Too often we try to quantify what can't actually be measured. [...] David Boyle examines our obsession with numbers. He reminds us of the danger of taking numbers so seriously at the expense of what is non-measurable, non-calculable: intuition, creativity, imagination, and happiness."
I don’t necessarily have any books to cite specifically on this like others do, but one example of this I do find particularly interesting is in military leadership.
My all time management/strategy hero continues to be General Montgomery, whom I can happily wax lyrical about being one of the most effective managers in recent history. However, relevant to this issue, one interesting thing I find about his headquarters was simply his willingness to bring in people from outside the military profession, most notably Bill Williams, his head of Intelligence who was an Oxford don.
This had all sorts of advantages - outside perspectives, different skill sets etc - but crucially it also meant that he wasn’t reliant on a continuing career in the military after the war was over. This basically meant that where the other officers were having to factor in future employment prospects and needing to make sure they didn’t hack off the wrong people, he was largely free to give a completely honest summation of the situation and had carte blanche to basically say whatever he liked to whomever he liked, knowing he was safe in a professorship career afterwards. This amongst other reasons was why he was one of the few people Monty would consistently go to for advice (and had such a good assessment of the situation), and was frequently brought out to say things other people couldn’t.
I know there's been research on real estate agents in this context, and how their behavior when dealing with their own housing is different when they're buying/selling housing for others. (Obviously, they often have good reasons for this!)
The Dictator's Handbook by Bruce Bueno Mesquita has a bit of this (entertaining video below)
TLDR: If you are a dictator you can only be overthrown by a fairly small number of people below you ("Keyholders" e.g. army, police, maybe your own party).... so your actual job to survive is to a) Screw as much money as possible out of the state, b) Hide it, c) Dole enough of it out that the keyholders stick with you rather than taking a risk and overthrowing you.
The critical bit has to be that only you know how to get your hands on the money.
Ideal situation is to have an external sponsor (e.g. French government in the 1980s) who prop you up and ensure that potential coup leaders know that the risk of failure is relatively high, or to have a really good grip on aid donors (e.g. Rwandan government today).
Risk aversion and career safety are often discussed in the context of software engineering and big tech jobs. I don't have a book to recommend, but some authors to start:
There is some research on the concept of “job embeddedness” which might be worth a search around-does being more embedded make you more risk averse?
The angle I was looking at it from was encouraging flighty young risk takers to leave less often so I don’t have anything that directly relates but hopefully this might lead you to some interesting nuggets.
Maybe you could take a look at a specific well studied example like Chernobyl. There should be some interesting stuff, since the conventional wisdom “careerism leads to risk-evasion” obviously needs to be challenged, because careerism seems to have encouraged excessive risk-tasking. In the HBO-TV-series this combination is portrait as the reluctance of the Director (on a good career path, but needs a test to go further) and his second subordinate (who is not on a good path and looking for an opportunity to get ahead) who both want to further their career, while the second one is inclined to take excessive risk and responsibility. The whole series wonderfully plays on this theme of risk, responsibility, guilt, careerism (subverted by fight for remembrance because of inevitable death/cancer) and responsibility sinks (the emergency button).
A quick Perplexity search has revealed some studies on this, but I don’t know how good they are.
Fun and insightful as always. Sorry I can't help with your request but one observation. "How do they do this in other countries" doesn't work so much any more - we have become much more insular. Especially true in Scotland where the correct question is: "Why are we so great comped to England" at least politically anyway!
All of my IT experience has been in the private sector, except for one project when I worked for IBM and I was sent to the Department for Work & Pensions. That was in 2005. It was a highly educational time, but far from enjoyable. I was shocked by the client's attitude towards risk and accountability. Their managers happily took decisions I knew would wreck the project (if I were to comply) but protect their careers. I wrote about it here. I see that when I wrote it I did not mention IBM or the DWP, but I'm retired now, so what the heck.
https://clarotesting.wordpress.com/2020/12/01/the-last-straw-the-project-that-convinced-me-to-resign/
On a wider note I've seen similar situations play out elsewhere in large corporations, though never quite as extreme as at the DWP. I occasionally had to give dual advice to executive management when I was an IT security manager; this is what you should do to protect your career, and this is what you should do based on the risk to the corporation we have identified and agreed upon. They were not necessarily the same. Sometimes it was better for a manager to make a loss on the account and annoy the client rather than turn a profit and keep the client happy. I've worked as an IT auditor in an excellent internal audit department for a major insurer. It was part of our job to detect, expose, and challenge such dysfunctional situations.
I've seen more than a few business people build their careers on the backs of bad data architectural decisions made by a consultancy that are pitched as "flexibility." This happens when e.g. marketing agencies design reporting systems for their clients. The cobbled-together result costs no more than the sponsor has in their budget, but keeps someone from the agency employed pushing buttons and manually carrying data. The low-risk and high-reward option is to replace all that with a modern system, but nobody with local authority has an incentive to do that.
From Gemini, very similar to many things you've already covered:
The idea that civil service behavior is primarily driven by the desire to minimize career risk—often termed "blame avoidance" or "risk aversion"—is a central pillar of Public Choice theory and modern public administration.
While several works touch on this, the two most "standard" academic sources that define this perspective are Christopher Hood’s The Blame Game and R. Kent Weaver’s foundational paper on blame avoidance.
1. The Modern Definitive Work: Christopher Hood
Book: The Blame Game: Spin, Bureaucracy, and Self-Preservation in Government (2011)
Christopher Hood is the leading scholar on why modern bureaucracies seem obsessed with "covering their backs." He argues that in an era of high-intensity media and political scrutiny, the primary incentive for a civil servant is not to achieve the "best" outcome, but to ensure that if something goes wrong, the blame cannot be pinned on them.
Key Concept: Negativity Bias. People (voters and politicians) react far more strongly to a single failure than to many successes. Therefore, civil servants prioritize avoiding a "minus" over achieving a "plus."
Three Strategies of Self-Preservation:
Agency Strategies: Creating complex organizational structures or "arms-length" bodies so it’s unclear who is actually responsible for a decision.
Policy/Operational Strategies: Relying on "protocolization"—strictly following rigid rules and checklists so that even if the outcome is bad, the bureaucrat can say they "followed the process."
Presentational Strategies: Using "spin" or "drawing a line" to distance oneself from a crisis.
2. The Foundational Paper: R. Kent Weaver
Article: "The Politics of Blame Avoidance" (1986)
Before Hood, R. Kent Weaver crystallized the theory that politicians and bureaucrats are motivated more by the desire to avoid blame than by the desire to claim credit.
The Argument: Since "voters are more sensitive to out-of-pocket losses than to equivalent gains," the safest career move for a public official is "The Good Soldier" approach: keeping a low profile and avoiding any policy that has a identifiable "loser," even if it has a high collective "gain."
3. The Classic Organizational Model: Anthony Downs
Book: Inside Bureaucracy (1967)
Downs provided the first rigorous economic model of how individuals behave inside a government department. He categorized bureaucrats into types, most notably "The Conservers."
The Conserver: This is the "standard" mid-to-late career civil servant. Their primary goal is convenience and security. Because they cannot "profit" from a successful innovation (unlike a private sector CEO), but they can be fired or demoted for a visible failure, their rational choice is to resist change and minimize any activity that carries career risk.
4. The "Satisfying Superiors" Model: Gordon Tullock
Book: The Politics of Bureaucracy (1965)
Tullock, a co-founder of Public Choice theory, argued that because there is no "market price" to measure a civil servant's success, the only way to get promoted is to please your superior.
The Impact: This creates a "telephone game" where information is filtered as it goes up the chain. Subordinates hide risks and failures to protect their careers, meaning the top of the hierarchy often makes decisions based on sanitized, "safe" information that doesn't reflect reality.
Thanks so much, some of those look good. I guess the problem is that the public choice school never really did much empirical work, so a lot of this stuff language is in the space of "stylised facts that everyone sort of knows" rather than "well investigated phenomena where people know a bit more about the precise conditions in which it does and doesn't appear"
It would be interesting to see if there are any documented examples of behaviour outside that hypothesised.
A hearty second to a recommendation of any book by C Thi Nguyen, and an apology for popping up with otherwise an unrelated comment.
Here is a Gary Marcus blurb about a paper I think you’ll find fits neatly into your cybernetic management thesis: https://garymarcus.substack.com/p/how-generative-ai-is-destroying-society
The authors of the paper Marcus endorsed don’t seem to realize it, but they describe generative AI as, in effect, “an anti-viable system” (in the sense of Stafford Beer).
«particularly contexts of government and civil service decision making»
I have noticed in the past an amusing case of where careers have mattered a lot which is why the Royal Navy has arguably many more frigates than it needs:
https://www.prospectmagazine.co.uk/magazine/wastedwarships
My summary of that story is that promotions to flag ranks are/were traditionally decided on the basis of performance as captain of a frigate-size ship and to ensure uniformity of evaluation there had to be enough frigate-size ships to be able to evaluate all captains at that point in their careers and frigates were the cheapest option for that ship size. Actually I think a quite reasonable policy in peace time.
Not quite what you're looking for, but J.P. Marquand's 1949 novel "Point of No Return" is primarily (in terms of both plot and metaphor) concerned with career risk for its protagonist (a banker), and is an excellent if sometimes oblique exploration of same.
«on “The Score” by C Thi Nguyen.»
There is a similar argument about metrics in some books by David Boyle (a well-meaning "whig"):
https://david-boyle.co.uk/books
https://www.theguardian.com/books/2025/jul/17/david-boyle-obituary
D. Boyle "The tyranny of numbers" (2000)
https://www.google.co.uk/books/edition/The_Tyranny_of_Numbers/l2-CAAAAIAAJ
"Too often we try to quantify what can't actually be measured. [...] David Boyle examines our obsession with numbers. He reminds us of the danger of taking numbers so seriously at the expense of what is non-measurable, non-calculable: intuition, creativity, imagination, and happiness."
D. Boyle "Tickbox" (2020)
https://www.google.co.uk/books/edition/Tickbox/H6OQDwAAQBAJ
"behind Tickbox lies an insidious philosophy of automation and the misuse of data that weighs heavily on every one of us."
I don’t necessarily have any books to cite specifically on this like others do, but one example of this I do find particularly interesting is in military leadership.
My all time management/strategy hero continues to be General Montgomery, whom I can happily wax lyrical about being one of the most effective managers in recent history. However, relevant to this issue, one interesting thing I find about his headquarters was simply his willingness to bring in people from outside the military profession, most notably Bill Williams, his head of Intelligence who was an Oxford don.
This had all sorts of advantages - outside perspectives, different skill sets etc - but crucially it also meant that he wasn’t reliant on a continuing career in the military after the war was over. This basically meant that where the other officers were having to factor in future employment prospects and needing to make sure they didn’t hack off the wrong people, he was largely free to give a completely honest summation of the situation and had carte blanche to basically say whatever he liked to whomever he liked, knowing he was safe in a professorship career afterwards. This amongst other reasons was why he was one of the few people Monty would consistently go to for advice (and had such a good assessment of the situation), and was frequently brought out to say things other people couldn’t.
I know there's been research on real estate agents in this context, and how their behavior when dealing with their own housing is different when they're buying/selling housing for others. (Obviously, they often have good reasons for this!)
The Dictator's Handbook by Bruce Bueno Mesquita has a bit of this (entertaining video below)
TLDR: If you are a dictator you can only be overthrown by a fairly small number of people below you ("Keyholders" e.g. army, police, maybe your own party).... so your actual job to survive is to a) Screw as much money as possible out of the state, b) Hide it, c) Dole enough of it out that the keyholders stick with you rather than taking a risk and overthrowing you.
The critical bit has to be that only you know how to get your hands on the money.
Ideal situation is to have an external sponsor (e.g. French government in the 1980s) who prop you up and ensure that potential coup leaders know that the risk of failure is relatively high, or to have a really good grip on aid donors (e.g. Rwandan government today).
https://www.youtube.com/watch?v=rStL7niR7gs
Risk aversion and career safety are often discussed in the context of software engineering and big tech jobs. I don't have a book to recommend, but some authors to start:
https://blog.pragmaticengineer.com/what-silicon-valley-gets-right-on-software-engineers/
https://lethain.com/durably-excellent-teams/
There is some research on the concept of “job embeddedness” which might be worth a search around-does being more embedded make you more risk averse?
The angle I was looking at it from was encouraging flighty young risk takers to leave less often so I don’t have anything that directly relates but hopefully this might lead you to some interesting nuggets.
Must be something in the air! FT had this great piece on obsessive measurement over xmas > https://www.ft.com/content/5e98d100-5d0e-4cc5-8ce7-1b365f239f91
He came on to my podcast for a great chat back in 2021 or so. https://www.thendobetter.com/arts/2021/6/19/c-thi-nguyen-on-games-philosophy-agency-real-world-gamification-podcast
Maybe you could take a look at a specific well studied example like Chernobyl. There should be some interesting stuff, since the conventional wisdom “careerism leads to risk-evasion” obviously needs to be challenged, because careerism seems to have encouraged excessive risk-tasking. In the HBO-TV-series this combination is portrait as the reluctance of the Director (on a good career path, but needs a test to go further) and his second subordinate (who is not on a good path and looking for an opportunity to get ahead) who both want to further their career, while the second one is inclined to take excessive risk and responsibility. The whole series wonderfully plays on this theme of risk, responsibility, guilt, careerism (subverted by fight for remembrance because of inevitable death/cancer) and responsibility sinks (the emergency button).
A quick Perplexity search has revealed some studies on this, but I don’t know how good they are.
Fun and insightful as always. Sorry I can't help with your request but one observation. "How do they do this in other countries" doesn't work so much any more - we have become much more insular. Especially true in Scotland where the correct question is: "Why are we so great comped to England" at least politically anyway!
Career driven risk aversion is definitely a big thing in sport.
Here's a book about how the England cricket encouraged the batters to be less risk averse in T20 matches
https://amzn.eu/d/ex33G7v
And here's Clifford Asness talking about risk aversion in ice hockey
https://www.aqr.com/Insights/Perspectives/Perhaps-the-Most-Important-Essay-I-Will-Ever-Co-Author