The only way to avoid this is to somehow let the employees share in any efficiency savings. But that comes with problems of its own, and is anyway incredibly ideologically unpopular.

It's a principal-agent problem. Imagine there is $100,000 in front of you. You are absolutely forbidden from taking any of it. However, if you set it on fire, and write a short memo inventing a plausible justification for doing this, your working life will be easier over the next year. A lot of people would do that with only very minor guilt.

From the employer perspective: there's no magic way to find out whether spending is "good" or not. You can add process and reviewers to check. But that also costs money, and slows down all the actual work. That's how you end up spending $1000 of staff time to have a meeting to approve purchasing a pen.

There is no justification except fear of not having enough money. The issue is whether you can meet your peak budget. If you can get by with $x * 0.80 most of the time, but every once in a while you really, really need the full $x then you spend $x/year. Because, if you don't, that extra $x * 0.20 will be cut (possibly even more). This leads to an increasingly tight budget (or it's feared at least) so it's "better" to be wasteful.

It's also useful to request a budget 25% -more- than you need, so some manager can "slash your budget" and feel like he won.

And also ask for stuff you don't need, which you can "give up" during the budget allocation process.

Corporate Budgeting is a game. If you need to play, best you first learn the rules.

The filing out your buget so it doesn't get cut part is understandable. What makes less sense is cutting the max budget based on use wihout leaving any overhead to account for unexpected expenses. Why does that seem so common when it makes the incentives so obviously misaligned?

Because there are always more profitable projects than money. If you can liberate some money by cutting budgets elsewhere, you can run more of them.

Having money left over at the end of the budget year is a failure, as that money could have been used in a profitable projects.

Except you are completely disregarding the obvious outcome: others wasting money to not have their buget cut. And the alternative could be even worse: departments running out of money and imploding because of unexpected costs without any budget headroom.

The point is not that investing the company's money wisely doesn't make sense but that looking hat how much each department used and then basing the next budget on that alone is the opposite of wise because those departments are run by humans that will be looking out for themselves first.

One thing that people often don't get about spending. Agencies are supposed to actually spend the money they were allocated. If they don't that raises very uncomfortable questions. It's maybe worse than spending too much. Where you can come up with all sorts of BS to justify that. Not spending your budget is harder to justify.

The justification is spending money more efficiently, which becomes even more important if you're a government entity and thus "your" money is actually other people's money gathered by taxation.

Put more bluntly, you cannot convince your bean counters to give you money you didn't spend, because if you didn't spend the money you didn't need it regardless your arguments for wanting/needing a seemingly extraneous budget.

Of course, this leads to the incentive of spending your entire budget no matter heaven or hell so you maintain your budget next year which can lead to actually wasting money rather than saving it.