Economists measure the cost of protection in terms of static efficiency, growth rates and firm-or industry-level productivity effects. This survey is devoted exclusively to the literature on the static efficiency. A key preliminary point to note is that estimates in this literature are not derived econometrically from pre-and post-liberalization data. Instead, they are based on simulations of partial-or general-equilibrium models that are parameterized using estimates from the literature and calibrated around a pre-liberalization base year.
I. "Small" Triangular LossesThe natural starting point for our purpose is Harry Johnson (1960) whose central message was that static costs of protection, measured as a proportion of GNP, are inevitably low. Tariffs give rise to the triangular efficiency losses whose magnitude can be summarized in the formula:(1)Cost of protection as a proportion of GNP = (½) α.ητ 2 where α denotes the imports-to-GNP ratio at the tariff-ridden equilibrium, η the absolute value of the arc elasticity of demand for imports as we move from protected to free-trade equilibrium, and τ the ad valorem tariff rate at the domestic price. Because α is usually well below 1, estimates of η revolve around 1 or 2, and τ 2 is the square of a number less than 1, calculations based on this formula are expected to yield small numbers.Johnson's message was echoed by a number of empirical studies published during and around 1960s and summarized in Panagariya (2002). These studies nearly uniformly found the cost of protection for different countries in different years to be no more than 1% of GNP. Many subsequent studies, which employed computable general-equilibrium models generated similar numbers, leading Richard Harris (1984) to remark, "It is well known that conventional calculations of the costs of protection give numbers which are quite small; often in the order of 0.5 to 2.0 percent of GNP. This result holds for almost all known studies based on the competitive neoclassical model, either partial or general equilibrium."
II. High Protection, High CostsIt is easy to show, however, that low costs of protection in the conventional model are not inevitable. The impression that these costs are low stems from calculations that have themselves been done for situations involving low initial tariffs. That small deviations from the optimum are accompanied by small losses is hardly surprising.From formula (1), the cost of protection rises at an increasing rate with the tariff. Therefore, it is not inconceivable that high tariffs might lead to high costs. This is suggested to some degree by an early calculation by Arnold Harberger (1959) for Chile: based on an average tariff rate of 50%, he calculated the cost of protection in Chile during 1950s to be 2.5 percent of GNP. ThoughHarberger and his contemporaries viewed this estimate as small, it was well above the other estimates obtained in the 1960s, based on tariff reductions of less than 15 percentage points.Challenging Harberger-Johnson calculations, Bhagwati (1968) ha...