Barriers to Competition and Productivity: Evidence from India Supplementary Appendix

We assume that our value added production function v = f(l, n, k, ω), is part of a more general production function separable in all intermediate inputs Y = g(f(l, n, k, ω), h(Γ, ω)) where Γ is a vector of intermediate inputs. Let ι be one intermediate input, which LP assume has a demand function of the form:ιit = ιt(ωit, kit). Other possible state variables not explicitly included in the above input demand function include prices of inputs and output(s). We assume input and output prices are fixed across firms within the same industry, but allow for the common prices to change over time by indexing the input demand function by t.1 Assuming monotonicity, i.e. input choice is strictly increasing in productivity for all relevant capital levels,2 the input demand function can be inverted to yield an representation for the unobserved productivity: ωit = ωt(ιit, kit). Then, assuming the monotonicity condition holds, we can estimate the coefficients on the labor inputs by estimating the following regression: