But because of all that scrutiny, the government often has to behave suboptimally: a) Processes can be a matter of law, requiring an extraordinarily heavy bure

But because of all that scrutiny, the government often has to behave suboptimally: a) Processes can be a matter of law, requiring an extraordinarily heavy bureaucratic lift to adapt to new information or fix bugs. b) Processes are calibrated to minimize trust. Contracting must be as obtuse as it is, because the alternative is to let civil servants make judgement calls, which the public cannot abide. This probably mitigates some serious downside risk, but it also precludes a lot of potential upside. Other institutions succeed when competent and trustworthy people just do what they think is right. (I daresay this is the only way an organization ever accomplishes anything difficult). Government agencies are only allowed to do that in emergencies, or in tiny edge cases like special forces. c) Where leaders do have latitude, the fear of embarrassment weighs unusually heavy on their decision-making. It's harder to take good risks. I'm sure there was waste at IBM that would not survive in government. I'm equally sure there was innovation at IBM that could not have happened in government. In some ways "government can't do anything right" is a self-fulling prophecy. The controls we implement in response to that perception make it come true. More fundamentally, I think any engineer can identify with how having to balance so many stakeholders could be incompatible with doing their best work. Making government less accountable probably isn't the answer either.