Last winter, several intersections around my city had problems with snow accumulating and blocking the stop lights, because the new LED bulbs don't generate enough heat to melt the snow that accumulates. I guess hindsight is always 20/20, but it's surprising the people paid to make those decisions didn't think of it beforehand.
Does the new acquisition suit the environment it will be operating in.
These requirements ought be stated in the contract of sale.
If the supplier makes claims that it does, and it turns out it doesn’t then they are in breach of the contract and can be tasked with making good.
Of course, sometimes we get these things wrong.
I imagine the case for street lighting is even stronger, especially when you start to factor labour costs for bulb replacement as well. And traffic lights would be stronger again, because they're cycling instead of being constantly on, which shortens the filament life significantly.
Now if by "years" you meant "two years", then fine, I can believe it- but two years is still a fantastic payoff period.
Don't forget HPS burns out pretty regularly too. Tactic my muni is using is simply to replace burnt-out HPS with LED instead of repairing it. Lo, after a few years we're probably half LED.
Are you somehow surprised that an infrastructure project would take years to pay off?