The approach you suggest is very common in the public sector, almost to a rule.
Public services are just massive, costly, complicated, and cater to all sorts of vulnerable populations, and the public have pay for it. The time horizon on some complicated projects is so long that knowledge can change in the meantime and make the project/strategies look incoherent. E.g., some countries decommissioned public transport options over decades when cars became more widespread, and are undoing it now because cars became too widespread, and it'll take decades to rectify. Then there are world events that upend everything, such as Ukraine or COVID-19. It's just very complicated, far more complicated than anything in software development in my experience.
Public services are just massive, costly, complicated, and cater to all sorts of vulnerable populations, and the public have pay for it. The time horizon on some complicated projects is so long that knowledge can change in the meantime and make the project/strategies look incoherent. E.g., some countries decommissioned public transport options over decades when cars became more widespread, and are undoing it now because cars became too widespread, and it'll take decades to rectify. Then there are world events that upend everything, such as Ukraine or COVID-19. It's just very complicated, far more complicated than anything in software development in my experience.