Love and life between the censuses: a model of family decision making in Puerto Rico 1950-1960.

Abstract

The model of behavior reported consists of several relationships designed to explain or describe several interdependent household decisions that are reflected in differences among aggregate data for 75 small geographical regions in Puerto Rico (termed muncipios) for the census years 1950 and 1960 and to some extent for the intervening years. The key variables of Puerto Rican demographic and economic structure in the 1950s are: family size decisions as reflected in crude birthrates; female labor force participation rates; marriage patterns; income differences; and migration decisi ns. The chief variables that drive the system i.e. those treated as exogenous are death rates both crude and for infants; educational attainments and current opportunities for education; certain characteristics of the industrial structure of each region; and the tightness or looseness of the labor market as reflected in the unemployment rate during the Census week. The model contains 5 equations describing births female labor force participation income and form and prevalence of marriage. A rather highly disaggregated set of 20 equations describing migration in various age-sex categories is included. The equations wre estimated using cross-section data for 1950 and 1960 and vital statistics for intervening years. On the whole the statistical results are quite variable and sensitive to what estimation procedures are used as well as to the inclusion or exclusion of certain variables but mostly regressions using the same sets of variables and the same estimation procedures for the 2 census years are quite similar. Birthrates were strongly influenced by past death rates form and prevalence of marriage and the opportunity costs of children as measured somewhat indirectly by female labor force participation rates. Age composition income and educational attainments had disappointingly variable and frequently unimportant effects. The female labor force participation rate was negatively related to income in analyses using methods considering the simultaneous nature of the determination of both variables but not in ordinary least squares regressions. Although this connection was weak at best the opposite findings do illustrate the importance of accounting for simultaneity in models of this sort. The effect of young children was negative. The most powerful explanatory variables were the index of the demand for female labor services and the unemployment rate. Educational and age composition variables had erratic but sometimes significant effects. The explanation of regiolan cariations in income showed the importance of differences in educational attainments rural-urban differences and the effects of marriage and sex differences. Legal marriages were positively and consensual unions negatively related to income. The migration equations revealed mixed results for different age groups but income and cost differences tended to show up very strongly. Firm policy conclusions draw from this analysis would be premature.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 35 citations exist upstream.