There’s a large literature on bureaucracies, and it has a lot to say that is useful on the topic. Unfortunately, this post manages to ignore most of it. Even more unfortunately, I don’t have time to write a response in the near future.
For those looking for a more complete picture—one that at least acknowledges the fact that most bureaucracies are neither designed by individuals, nor controlled by them—I will strongly recommend James Q. Wilson’s work on the topic, much of which is captured in his book, “Bureaucracy.” I’ll also note that Niskanen’s work is an important alternative view, as is Simon’s earlier (admittedly harder to read, but very useful) work on Administrative Behavior.
Perrow’s work, “Organizational Analysis: A Sociological View” is more dated, and I wouldn’t otherwise recommend it, but it probably does the best job directly refuting the claims made here. In his first chapter, titled “Perspectives on Organizations,” he explains why it is unhelpful to view organizations just as a function of the people who make them up, or as a function of who leads them. When I have more time, I will hope to summarize those points as a response to this post.
I appreciate the response. I poked around a bit looking for an online version of the two works and haven’t found one yet, but will continue looking a bit and look forward to your eventual summary of those points.
(My own motivation for curating this stemmed from having observed a couple bureaucracies forming in realtime over the past year, which did roughly match the “owned by the creator” schema here, as well as encountering various other bureaucracies that seemed to match the “un-owned” schema, which were pathological in the ways you’d expect and that matched the model in the post. The assumption I took as given in this post, although not specified, is that most bureaucracies you encounter will be un-owned)
I think that most bureaucracies are the inevitable result of growth, and even when they were initially owned by the creator, they don’t act that way once they require more than a few people. (See my Ribbonfarm Post, Go Corporate or Go Home)
Comparing the goals of a bureaucracy with the incentives and the organizational style, you should expect to find a large degree of overlap for small bureaucracies, trailing off, at best, around a dozen people, but almost none for larger ones. This isn’t a function of time since formation, but rather a function of size—larger bureaucracies are fundamentally less responsive to owner’s intent or control, and more about structure and organizational priorities. As an obvious case study, look at what happened at US DHS after 2002, which was created de novo with a clear goal, but it is clear in retrospect that the goal was immediately irrelevant to how the bureaucracy worked.
This all makes sense, just doesn’t seem to me to be in conflict with the OP.
I don’t know much about the US DHS, but a few obvious things that pop into mind:
In the schema of the OP, a large bureaucracy is harder to make an effective bureaucracy (for the same reason a large codebase is more likely to have bugs). Especially if that bureaucracy was created quickly. Even if it has a competent owner, it’s just a harder task.
The DHS wasn’t created in a vacuum, it was created a) as part of a weird political situation, b) I suspect it was also to some extent created by existing bureaucracies. I have little reason to believe that the stated goal of the DHS was ever the actual goal. I don’t think it “got immediately compromised”, my guess is it was compromised from conception. (But, I don’t know a whole lot about it and wouldn’t be that surprised if my guesses were off)
Something that the OP doesn’t delve into much (and I do think makes it incomplete) is that bureaucracies might have multiple owners.
Just like a codebase is more likely to run into problems if it’s being created by multiple teams with multiple goals, esp. if those people aren’t aligned with each other, it’d make sense for bureaucracy goals to have degrees of coherence, depending on whether they were created by a single person or as part of political compromise.
I don’t think people should feel obligated to read all that’s been written on a topic before posting their thoughts on that topic to Less Wrong, especially if that writing is not supported by randomized controlled trials (unsure if this is true for the writings you cite).
Fair point about the bar for posting, but this doesn’t read like “posting their [tentative] thoughts,” it reads like conclusions based on extensive review. As a matter of good epistemics, the difference should be made clearer. Similarly, if you dismiss large parts of the literature, it would be good to at least let people know what you think should be ignored, so they don’t waste their time, and even better, why, so they can decide if they agree.
As a side point, I think that considering RCTs as a source of evidence in this domain is a strange bar. There’s lots of case study and other quantitative observational evidence that supports these other approaches, and specifying what evidence counts is, as the phrase goes, logically rude—how would you even design an RCT to test these theories?
If it’s hard to conduct RCTs in a domain, it’s hard to have reliable knowledge about it period. Who’s to say whether your anecdotal observations & conclusions beat mine or someone else’s? One way is to check whether someone’s job is high status enough that their writings on the topic can be considered part of “the literature”. But this is a weak heuristic IMO.
I strongly disagree. There are many domains where we have knowledge with little or no ability to conduct RCTs—geology, evolutionary theory, astronomy, etc. The models work because we have strong Bayesian evidence for them—as I understood it, this was the point of a large section of the sequences, so I’m not going to try to re-litigate that debate here.
Agreed—one is not expected to read all that’s been written on a topic. However, one should acknowledge alternate models when they’re pointed out, and your summaries and conclusions will carry a lot more weight if you can explain why you prefer those over the other ideas.
I do also think that Samo has engaged with a large part of the literature, and it just doesn’t come through in this post. (My model is that he has dismissed a large part of the existing literature for pretty decent reasons, but am not confident)
There’s a large literature on bureaucracies, and it has a lot to say that is useful on the topic. Unfortunately, this post manages to ignore most of it. Even more unfortunately, I don’t have time to write a response in the near future.
For those looking for a more complete picture—one that at least acknowledges the fact that most bureaucracies are neither designed by individuals, nor controlled by them—I will strongly recommend James Q. Wilson’s work on the topic, much of which is captured in his book, “Bureaucracy.” I’ll also note that Niskanen’s work is an important alternative view, as is Simon’s earlier (admittedly harder to read, but very useful) work on Administrative Behavior.
Perrow’s work, “Organizational Analysis: A Sociological View” is more dated, and I wouldn’t otherwise recommend it, but it probably does the best job directly refuting the claims made here. In his first chapter, titled “Perspectives on Organizations,” he explains why it is unhelpful to view organizations just as a function of the people who make them up, or as a function of who leads them. When I have more time, I will hope to summarize those points as a response to this post.
I appreciate the response. I poked around a bit looking for an online version of the two works and haven’t found one yet, but will continue looking a bit and look forward to your eventual summary of those points.
(My own motivation for curating this stemmed from having observed a couple bureaucracies forming in realtime over the past year, which did roughly match the “owned by the creator” schema here, as well as encountering various other bureaucracies that seemed to match the “un-owned” schema, which were pathological in the ways you’d expect and that matched the model in the post. The assumption I took as given in this post, although not specified, is that most bureaucracies you encounter will be un-owned)
I think that most bureaucracies are the inevitable result of growth, and even when they were initially owned by the creator, they don’t act that way once they require more than a few people. (See my Ribbonfarm Post, Go Corporate or Go Home)
Comparing the goals of a bureaucracy with the incentives and the organizational style, you should expect to find a large degree of overlap for small bureaucracies, trailing off, at best, around a dozen people, but almost none for larger ones. This isn’t a function of time since formation, but rather a function of size—larger bureaucracies are fundamentally less responsive to owner’s intent or control, and more about structure and organizational priorities. As an obvious case study, look at what happened at US DHS after 2002, which was created de novo with a clear goal, but it is clear in retrospect that the goal was immediately irrelevant to how the bureaucracy worked.
This all makes sense, just doesn’t seem to me to be in conflict with the OP.
I don’t know much about the US DHS, but a few obvious things that pop into mind:
In the schema of the OP, a large bureaucracy is harder to make an effective bureaucracy (for the same reason a large codebase is more likely to have bugs). Especially if that bureaucracy was created quickly. Even if it has a competent owner, it’s just a harder task.
The DHS wasn’t created in a vacuum, it was created a) as part of a weird political situation, b) I suspect it was also to some extent created by existing bureaucracies. I have little reason to believe that the stated goal of the DHS was ever the actual goal. I don’t think it “got immediately compromised”, my guess is it was compromised from conception. (But, I don’t know a whole lot about it and wouldn’t be that surprised if my guesses were off)
Something that the OP doesn’t delve into much (and I do think makes it incomplete) is that bureaucracies might have multiple owners.
Just like a codebase is more likely to run into problems if it’s being created by multiple teams with multiple goals, esp. if those people aren’t aligned with each other, it’d make sense for bureaucracy goals to have degrees of coherence, depending on whether they were created by a single person or as part of political compromise.
I don’t think people should feel obligated to read all that’s been written on a topic before posting their thoughts on that topic to Less Wrong, especially if that writing is not supported by randomized controlled trials (unsure if this is true for the writings you cite).
Fair point about the bar for posting, but this doesn’t read like “posting their [tentative] thoughts,” it reads like conclusions based on extensive review. As a matter of good epistemics, the difference should be made clearer. Similarly, if you dismiss large parts of the literature, it would be good to at least let people know what you think should be ignored, so they don’t waste their time, and even better, why, so they can decide if they agree.
As a side point, I think that considering RCTs as a source of evidence in this domain is a strange bar. There’s lots of case study and other quantitative observational evidence that supports these other approaches, and specifying what evidence counts is, as the phrase goes, logically rude—how would you even design an RCT to test these theories?
If it’s hard to conduct RCTs in a domain, it’s hard to have reliable knowledge about it period. Who’s to say whether your anecdotal observations & conclusions beat mine or someone else’s? One way is to check whether someone’s job is high status enough that their writings on the topic can be considered part of “the literature”. But this is a weak heuristic IMO.
I strongly disagree. There are many domains where we have knowledge with little or no ability to conduct RCTs—geology, evolutionary theory, astronomy, etc. The models work because we have strong Bayesian evidence for them—as I understood it, this was the point of a large section of the sequences, so I’m not going to try to re-litigate that debate here.
Agreed—one is not expected to read all that’s been written on a topic. However, one should acknowledge alternate models when they’re pointed out, and your summaries and conclusions will carry a lot more weight if you can explain why you prefer those over the other ideas.
I do also think that Samo has engaged with a large part of the literature, and it just doesn’t come through in this post. (My model is that he has dismissed a large part of the existing literature for pretty decent reasons, but am not confident)