That specific one is an unfortunate side-effect of needing to allocate budgets in advance. Large organizations working with a lot of money need to statically schedule their budgets in advance - and because they don't have crystal balls, they can't actually know in advance how much they'll need, and have to make educated guesses. If a department didn't use all of its allocated budget last year, sure that's a very weak signal that it was given too large of a budget - but it is a signal, and there aren't very many other good ones.
Its the way of the world now to aim for growing into a huge org, hiring as much as you can on the way up. Once those in charge know little more about a part of their company than the budget surplus at the end of the fiscal year the model just doesn't make sense.
And the budget-burning shows that this encourages making over-estimates and then spending the excess to make the actual numbers match the prediction. It's no different than research based on falsified data at that point.