Even cap&trade is unnecessary complexity. Just tax it. Keep increasing the tax until the pollution levels drop. Use the tax revenue to offset the taxes on productive behavior.
Taxing would involve playing referrer, collecting taxes for this specific issue, and then allocating that tax to either public or private parties to work on productive behavior.
Taxing sounds more problematic given how dysfunctional most governments are, especially when it comes to non-regulatory procedures. More failure points in the market generation.
A variable tax to disincentivize pollution combined with cap and trade would make the most sense to me.
Not needing to figure out the relationship between input and output is the beauty of feedback systems, and why they are so prevalent in engineering.
For example, one of my jobs at Boeing was working on the hydraulic actuators that moved the elevators up and down. The goal was to move the elevators in proportion to the control column movements. Nobody ever bothered to figure out how far the actuator would move for a particular input. How it worked was a mechanical linkage provided feedback so when the elevator was too far, the linkage would open the valve to move it back. If it wasn't far enough, the linkage would open the valve the other way, which would move it forward.
It behaves beautifully. It also makes the 3 redundant actuators work together, despite variations in tolerances.
They are for many purposes.
> if the central planners get things wrong people can literally die.
That's a problem with any system.
Eg. "The tax per gram of PM2.5 is $1. Each year, if average PM2.5 levels across America are above 5 ug/m^3, the tax will increase 10%. Otherwise, it will decrease 10%."
Then lawmakers once need to put in place the tax and set the target, and they never need to intervene again.
In fact they'd have no reason to improve anything provided that their projection for the cost of improvement looked marginally more expensive then just paying projected future cost of the tax.