Why the different results from previous empirical analysis? One reason is sluggish capital adjustment: previous approaches using regressions of first-order conditions (prices) probably underestimate the elasticity, while the supply-side approach I use likely overestimates it. 7/8