I'm stating that it may not be worth doing. You'd have to actually do the maths to confirm either way.
Disclaimer: as a public transport user I have no dog on this fight.
I can try a back of the envelope I guess...
If I assume a time in a car is around 0.9x as good as average time (which feels about right to me and indeed seems to match actual usage with risk level tolerance from 40 years ago), assume moderately large speed limit reduction enough to increase travel time by 20% and use a figure of 10kg of CO2 per hour (seems about right looking at estimates of modern car gas consumption) and the NICE QALY value of £30k we get around £100 worth of QALY loss per ton of CO2. This is a bit over the largest estimates for a CO2 tax that would cover CO2 emission externalities ($100 per ton) (using the latest estimates for the costs of climate change over the next 100 years).
Given the large error bars in the rough work above, it looks quite plausible that the real calculation could go either way, with a speed limit possibly being net good or net bad.