Niskanen's bureaucrats

Some more about Niskanen's budget-maximising bureaucrat this time:

The simple model describes a bureaucrat as a maximising character: A bureaucrat strives to maximise budget and subordinates among other things (such as power and recognition). This costs money - taxpayers' money usually, though that's different in some countries (Kuwait, for example).

In the beginning, there was no bureaucrat. The first bureaucrat got hired by the society, paid and having a great choice of possible actions, some occupation was chosen which paid back much more to the society than the bureaucrat's service did cost. The difference between his utility and his pay was the "rent" for society, its gain.
The story was quite similar with the next bureaucrats until at some point there were so many bureaucrats and so few choices of sensible bureaucracy occupations that the one additional bureaucrat did cost exactly as much as he was worth.
This is something economists like because it makes the model accessible to math (increment = 0). In fact, this specific bureaucrat with exactly identical cost and utility may or may not exist. Maybe he was skipped.

Budget-maximising model graph
What's important is that the bureaucracy wants to grow (its leaders want it to grow). Further bureaucrats are added, even though the poor choice of bureaucracy occupations means that these additional bureaucrats are not worth their pay.
The top bureaucrats get away with this, though: They point at all the utility generated by the bureaucracy as a whole and how this is (still) greater than its costs. They deceive away from the fact that some part of the whole is inefficient excess. Again, this is economic theory only, for in reality it's practically impossible to determine the utility generation properly.The model does nevertheless adequately describe that and why bureaucracies can grow too big.

That's not where the story ends, though. The bureaucracy can point at its net utility (positive society rent) up until the point at which an additional bureaucrat means that all the utility gains of the bureacracy have been compensated by its wasteful parts. The theory claims that the bureaucracy must not go beyond this point, or else it will be cut. Again; it's impossible to determine the utility gain and thus impossible to determine whether a specific bureaucracy is past this point or not.

This means that the 'feel' of it may actually explain the size of a bureaucracy. We tolerate bureaucracy growth until we feel it's become too large. This point may be anywhere - some people don't even recognise that the bureaucracy is capable of delivering greater utility than it costs ever.

This political economic model (Budget-maximising theory) is of great use when looking at national security bureaucracies; armies, navies, air forces, coast guards, domestic spying agencies, foreign intelligence agencies, national-level police agencies.

The current series of NSA scandals is an almost humorous example for this: The NSA and other intelligence agencies sprawled and - under cover of secrecy - escaped the judgement of their cost/benefit ratio. According to the model, they may very well have grown past the point of negative net benefit. This doesn't only apply to its size and budget, but also to its thirst for power(s).
Its defensive behaviour was predictable; the thirst for power(s) and budget requires it to fight for as much as it can get. A voluntary withdrawal is almost unheard of.

(Another economic model, the principal agent problem, can also help to explain the mess.)

We can expect such overreach outcomes according to Niskanen's model whenever the public has no clue about the policy. This doesn't even require formal secrecy as evidenced by the cancer-like spread of subsidies and other benefits for farmers in Germany*.

All in all, it's from an economic theory point of view unreasonable to trust a spy agency in its defence of its powers or a military in its defence of its budget. It's almost assured that both operate well past the optimum and instead grab as much money, powers, personnel or ship hulls as possible.


P.S.: I understand that some people will seize on this as it supports their "small government" ideology. Let me be clear: I'm for "good government", not "small government". "Small government" is lazy, simplistic ideology for people who are easily exploited by the group nowadays known as "the 1%". "Small government" is basically a rallying cry for crony capitalism.

*: Read the German mineral oil taxation law if you don't get this reference.


  1. Are you aware of any studies to determine the optimum size of a civil service? For example, has anyone ever determined a theoretically ideal civil service to population ratio?

    Similarly, are you aware of any governments applying sunset clauses to their ministries, whereby they periodically review mandates and resources (personnel and money) with the aim of either adjusting allocations based on new circumstances, or collapsing the ministry and folding its resources into other, more relevant sectors?

    I imagine the Sir Humphries of the world would be aghast at such suggestions, though from a taxpayer's perspective they might be worth investigating.

    1. No, it seems they're rather evading the problem. Bureaucracy growth is being outsourced to contractors, and thus even less transparent to the citizens. Same goes for certain government-owned banks and some parafisci; techniques to get things done more with less hassle and less transparency than with the official government budget.

      An optimal civil service is pretty much impossible to determine; as mentioned in the blog text, we cannot properly measure their utility. We could at most find correlations in international comparisons, but there's too much noise in those.

  2. Actually, there's a school of thought that contracting out most civil service responsibilities is the way to go. There are no pension liabilities, it's easier to shift resources as contracts expire, and easier to remove low-performing individuals in comparison with the heavily unionised civil service of most developed nations.

    It will never happen, of course, and as you note, contracted services will continue to represent stealthy bureaucratic growth.

    1. To tie personnel closely into a civil service career (forfeiting pensions if they leave etc.) can serve well against revolving door corruption.

      A mere evasion of established rules as with outsourcing is rarely a better option than to address the real problem inside the established rules.

      And in the end, the government and its activities need to be transparent at least to the most interested part of the public. No transparency is a recipe for society getting ripped off by special interests.

  3. You can find the optimum by reduction. Reduce until performance declines and then let it grow again.

    1. That's true for the model, but the noise and the suboptimal order of choices make this inaccurate at best. Add in a changing environment and a once-found approximate optimum will be obsolete soon thereafter.

      Besides, trial & error is rather unpopular in regard to life or death situations. You don't want to learn the hard way that your maintenance effort for an airliner has been too low. Nor do you want to learn the hard way that you hadn't enough deterrence against a nuclear power.