Technology and learning in a global MAC curve: evidence from the phase-out of ozone depleting substances
I estimate a marginal abatement cost (MAC) curve using detailed, project-level data from the global phase-out of ozone depleting substances (ODSs) across 145 low-income countries, including China, Brazil and India.
The data cover an estimated 85 per cent of the ODSs by weight phased out by these countries. Exploiting a quasi-natural experiment in the implementation of the Montreal Protocol, I find that average project unit abatement cost fell in the initial period of the phase-out before eventually rising, forming a convex curve overall.
One explanation for this, which test, is that pre-existing technology possibilities that could be adapted the phasing out ODSs, flooded into the abatement arena in the early years, pushing down the cost of abatement through a kind of meta-learning. Holding the appearance of these new technology “possibilities” constant, I also find that each repeat “application” of each possibility associates with lower abatement cost, as through a learning-by-doing effect.
These findings elaborate the theoretical basis for predicting how the cost of abating greenhouse gasses (GHG) may change over time, including why the cost of abatement now is so low.