Should pollution reductions count as productivity gains for agriculture?

Productivity changes have been used to gauge economic performance for at least fifty years. The rate of change (over time) in a total factor productivity index measures whether the pace of increase in one or more outputs exceeds a weighted index of the rates of change in the inputs used to construct that output. Positive values of this measure are often associated with technical change or exogenous sources of improvements in efficiency. Negative values create puzzles, such as the U.S. productivity slowdown of the 1970s and 1980s. For the most part, the focus of these measures has been on marketed outputs and inputs. As a result, changes in conditions affecting a firm's (or farm's) activities that cannot be measured as an output, input, or one of their prices are often suggested as explanations for productivity puzzles. Because productivity measures have been so closely linked to changes in living standards, it is natural to ask whether net increases (i.e., after taking account of any increase in inputs) in marketed outputs are the only things that should 'count' as gains to our standard of living. The articles by Fare and Grosskopf (FG) and Gollop and Swinand (GS) consider several different technical aspects of addressing this question. The purpose of this article is to comment on their proposals. Both articles implicitly accept the notion that changes in commodities that are not available in markets should be considered in evaluating performance. Fare and Grosskopf focus on how they should be 'valued' in the productivity indexes, while Gollop and Swinand define conditions when pollution reductions can be allowed to 'count.' The best overall summary of this comment on both papers repeats an overworked phrase-'the devil is in the details.' To see why, consider first the Gollop-Swinand analysis. A central conclusion of their proposed welfare-based approach to productivity measurement is that reductions in pollution should 'count' as increasing productivity rates only if the absolute magnitude of the marginal disutility of pollution (measured in monetary units) exceeds the marginal abatement cost. If the reverse is true (i.e., the magnitude of the marginal disutility is less than the marginal abatement cost), then reductions in pollution should reduce the rate of total factor productivity increase. Should the absolute marginal disutility exactly equal the marginal control cost, we can ignore the non-market sector.