Published February 2004
Experience curves suggest that unit capital costs ought to fall considerably over the coming 1-2 decades. In several cases it will be cheaper to build a plant in 2020 than it is now, and the 2020 plant will have twice as much capacity!
Contrary to conventional wisdom, capital costs do not increase relentlessly — in fact they fall. And there is more to such reductions than just scale factors, a lot more.
This report presents capital-cost experience curves for five products: polypropylene, styrene, methanol, ethylene and polyethylene. Their doubling-slopes are rather steep, significantly higher than the rule-of-thumb range for industry in general of 20-30%. Excepting methanol, experience-based cost cuts are substantially greater than what would come from scale increases alone. Data correlations (of experience to cost) were fairly-good to very-good.
Somewhat surprisingly, the best measure of experience is cumulative production, not cumulative capacity. On second thought this is not so surprising, because capacity data are notoriously inaccurate, at least relative to production statistics. Production is easier to measure, and it is often documented by governments, which have unparalleled access to raw data.