I got some very good comments on this post (thank you all!), which have focused the analysis of the difficulty-of-analysis of bureaucracy on issues relating to imprecise, incomplete, and ambiguous incentives, and selection effects or a lack thereof. I think all of this is substantially correct.

Now I’d like to go a bit further. 

My intellectual interest in the ability of economics to explain bureaucracy pertains to the research I do, which is on a more general question about the structure of economic systems, with the hopes of learning how to build flexible economic mechanisms that are more oriented toward human values. To do this, we need to be confident in our ability to understand the structure of the economic systems that currently exist. Bureaucracies are one such system, and the one most troubling to me from an intellectual perspective, regardless of the good or bad we might associate with them. Bureaucracies play a significant role in our economy, whether operating within a firm or in partnership with or the authority of a government.  And every single part and piece of a bureaucracy would seem to obey economic logic, from the individual utility-maximizing people who work for and with the bureaucracy to the budgets that bureaucracies must manage and the accordant interest in controlling costs and maximizing revenues. And yet, despite all of this economic stuff going on, and despite the effect it has on the economy, economists don’t have a basic theory of bureaucracy, and they don’t even seem to be trying to develop one.

I submit that this situation, although one we are all very used to, is nevertheless slightly weird.

The best way to develop an economic theory of bureaucracy should be to observe bureaucracy until you understand its general underlying logic. But this doesn’t seem to have worked. Instead, the more I’ve thought about this issue, the more my attention has been led instead to the general building blocks of any economic model. And it seems plausible to me that the blocks are not all there when it comes to bureaucracy.

Not that bureaucracy is literally unexplainable in economic terms. But I fear that such an explanation would be analogous to the ability of a physicist to explain anything in terms of the underlying motion of every fundamental particle: true, but not very useful, and awfully hard to implement in practice.

I think that the general structure of economic explanation is to conceptualize a system as preserving the structure of some basic set of relationships through many subsequent interactions. So when we say that a person is utility-maximizing, what we mean is that they have a preference-structure in their brain, and the way they behave is best explained in terms of actions that preserve said preference-structure. For example, if they like drinking water and do not like drinking cyanide, and you give them a labeled glass of each to drink, they will drink the water instead of the cyanide at rates significantly higher than what random chance would predict. The outcome of the interaction, consequently, is consistent with, or preserves, the internal preference-structure motivating the behavior.

Similarly, a firm has a basic revenue-cost relationship, and it acts to keep the relationship at a maximum through interactions with its suppliers, customers, employees, etc. Hence, we call it profit-maximizing. 

And a median-vote-winning democratic politician acts so as to preserve the political demands of their constituents through their legislative behavior and campaign rhetoric.

Of course, these structures are never perfectly preserved. People choose dumb things. Firms have principal-agent problems, and also are made of people, so they do dumb things. Politicians are people, and democracies have principal-agent problems. But by understanding the general basic concept of what is supposed to be preserved, we get a very clear and useful conceptual system for analyzing said failures. 

And, it need not be said, so I had better say it: just because a system preserves a structure doesn’t mean we want that structure to be preserved. Maybe people have bad preferences. Maybe profit-maximizing firms pollute the atmosphere. Maybe political demands are bad and stupid. This is a positive analysis, a way of making scientific sense of complex things. If you don’t like people/firms/governments, it’s useful to know why they do what they do and why that sometimes isn’t what we want them to do.

With bureaucracy then, the implication is that there is no structure that bureaucracies preserve in general. It’s not that they fail to preserve something, the way a person might not make the rational choice or a firm might hire a CEO based on height rather than competence or a politician may win election on a promise that they subsequently break. It’s that there’s nothing in particular to preserve at all. A bureaucracy may blame-minimize on Monday, budget-maximize on the next day, adopt a popular pose on the third day, prestige-maximize on the fourth day, and work-minimize on the fifth day (it's a Friday). No one has to be irrational for this to happen, nor does the bureaucracy have to be badly designed or exhibiting a standard economic "failure". There’s just no general structure with which almost all bureaucracies can be modeled almost all of the time as preserving or failing to preserve...even though bureaucracies are highly recognizable and therefore clearly share a common structure.

This all relates to what the commenters pointed out about the presence of incentives and selection effects, of course, but for my purposes, I like a focus on the basic thing that economic structure does—which, it seems to me, is to preserve something, to guard it against the noise and powerful forces of the universe. Incentives—a tendency for things to go one way rather than another—and selection effects are powerful tools for explaining the basic observation that sometimes, consistently non-random things consistently happen. That, to me, is economics. 

So why don't bureaucracies seem to fit into this view—or do they?—and what are the implications? Well…I have a lot more to say about it. A lot more, in fact. I have written about 150 essays on the subject over the last year or so, and I don’t know quite who to share them with or where to publish them. So if this seems at all interesting, and you’d like to read more about this, do let me know. And if you think that this isn’t a good fit for this site, do sound off as well, and I’ll go elsewhere. 

And for that matter, the first few essays are probably the most difficult, as they are about entering a conceptual world that treats economics a bit differently from the standard rational choice analysis, while still being compatible with it. (To wit, the goal is to take the analysis back a level, to treat concepts like "rational self-interest" as something more like theorems than axioms; economics is something that physical systems do under certain circumstances, hence the username.) I have tried to make them simple and readable, but I should probably have someone act as a beta reader for the first few essays. If this interests you at all, do reach out and I can share a google doc or something...or if there's a way to share a draft privately here, let me know.

But I shall probably want to rewrite the first set of essays (I broke them down into five) a bit anyway if people are in fact interested in them. So in the meantime, tomorrow or sometime soonish, I'll write an essay taking ideas from the comments in the previous essay and playing around with them a bit to emphasize the nature of the kind of general-basic explanation that we already have for people and firms but lack for bureaucracy. Thanks again for all the great thoughts!