Population, households, and CO2 emissions

The model linking environmental impact to population, affluence, and technology, or I = PAT, is reformulated in terms of households (i.e., I = HAT) as opposed to persons. Such an approach may be preferable in the case of environmental impacts that arise from activities, such as residential heating and automobile transport, for which there exist significant household-level economies of scale. Because of changes in average household size, the I = HAT model gives rise to a very different decomposition of the sources of historical growth of environmental impacts than does I = PAT. Taking growth of global energy consumption as an example, the authors find that I = PAT attributes 18 percent of the annual increase (in absolute terms) over the period 1970-90 to demographic increase in more developed regions, whereas I = HAT attributes 41 percent because the number of households grew faster than the number of persons. The I = PAT and I = HAT models also give rise to substantially different projections of CO, emissions in the year 2100. The authors conclude that decomposition and projection exercises are sensitive to the unit of demographic account chosen. Until more is known about the nature of the many activities that give rise to environmental impacts, it would be unwise to draw far-reaching conclusions from one choice of model without a substantive justification of that choice.