Productivity growth has long been an important concern of economists, industrialists, and government officials because it is recognized as the key to business profitability and economic welfare. The apparent lag in productivity growth in the U.S. economy, relative to other industrial countries and to our own past record, has generated many diagnoses and diverse policy prescriptions.

These concerns emerge from conventional measurements of productivity change that encompass only marketed outputs and inputs. Labor productivity, for example, measures output per worker. A broader measure, multifactor productivity (MP) (sometimes also called total factor productivity), measures output per unit of an index of labor, capital, and intermediate materials inputs. In this analytical framework, productivity change is defined as the difference between the growth rate of output and that of the index of inputs.

Almost no attempt has been made to measure environmentally related outputs (such as emissions) or inputs (such as natural resource services) that are not marketed or to assess their significance for economic productivity. What follows is an exploratory step in that direction, using the private electric power industry in the United States as an example. (For an update of this study and two additional cases including the pulp and paper and agriculture sectors, see Repetto et al., 1996.)

A typical 500-MW coal-fired power plant produces more than just 3.5 billion kWh of electricity per year. The 1.5 million tons of coal and 0.15 million ton of limestone it uses as inputs reappear in some form as outputs. Emissions to the atmosphere include 1 million tons of carbon in the form of carbon dioxide, 5,000 tons of sulfur as sulfur dioxide, 10,000 tons of nitrogen oxides formed largely from air drawn into the combustion process, and a variety of other compounds. Solid outputs include 140,000 tons of ash and 193,000 tons of scrubber sludge, which contain 5,000 and 40,000 tons of sulfur, respectively.

A more general measure of economic productivity, recognizing the conservation of matter and energy, would assess the extent to which the industrial transformation of materials has yielded outputs with greater economic benefits—or lower economic costs—than the costs of the inputs. Many of the power plant's unmarketed outputs have economic significance. Airborne sulfur emissions, for example, affect human health, plant growth, and the durability of materials. As recent experiments with marketable emissions rights indicate, the extent to which such outputs are marketed is largely an institutional arrangement. Productivity measures restricted to a subset of economically significant inputs and outputs can misrepresent technological progress in the industry.

As a step toward a broader measure, we developed an index of atmospheric emissions by weighting each of three pollutants (sulfur and nitrogen oxides and particulate matter) according to their estimated economic significance, defined as