Specialization, Household Production, and the Measurement of Economic Growth

Our concern in this paper is with the measurement of economic growth, or more precisely, the mismeasurement of growth that may occur from ignoring household production. As an economy develops, production shifts from households to markets.' Traditional measures of economic growth which ignore most household production may therefore be biased upward. The existence of household production also complicates international comparisons of real GNP as there are significant differences in the relative size of the household sector across developed and developing economies. There is an extensive empirical literature on the size of the household sector. Most estimates place the value of household production in the United States for recent years at 30-40 percent of GNP.2 Previous studies have found that the relative size of the household sector has remained fairly constant over time, implying that measured growth rates provide a reasonable proxy for overall growth rates. Existing estimates of household production, however, have little theoretical grounding. For the most part, they also ignore the use of capital in the household sector. In addition, most estimates relate to a single point in time.3 Finally, existing studies relate to time periods after which the large movements out of the household and into the market may have taken place. In this paper, we provide new estimates of the value added in household production for selected years from 1930 to 1985. We find that the relative size of the household sector in recent times is smaller than is commonly estimated. We also find, contrary to the results of others, that the relative size of the household sector has fallen continuously since 1930. We conclude that real per capita output has grown from 1930 to 1985 at a rate that is significantly below the rate implied by conventional measures of output. We develop in the next section a formal model of household production which we hope will help us to clear up such issues as the appropriate value to impute to household labor, how to treat time spent transacting, and how to deflate household production in order to obtain real output. The model is not a general-equilibrium model. Its role is not to derive implications for growth, but rather to guide us in our "accounting" exercise.