The Spacing and Limiting Components of the Fertility Transition in Latin America
In this chapter we present the results of an analysis of trends in marital fertility within categories of key socio-economic factors using data from six Latin American countries: Colombia, Dominican Republic, Ecuador, Mexico, Peru, and Trinidad and Tobago. These countries were selected for analysis because they have completed high quality surveys for two points in time, as part of the WFS and the DHS programmes. Our analysis of trends and differentials builds upon earlier work on all WFS countries by Rodríguez and Cleland ( 1987).
The focus of our analysis is on marital fertility--or more precisely, fertility following first union--rather than overall fertility. This choice reflects our view that an essential feature of the fertility transition is a change in childbearing within marriage, which in turn results from changes in spacing and limiting fertility behaviours. This is not to deny the fact that part of the observed decline in fertility in the region is due to rising ages at marriage, nor the fact that women in higher social strata tend to marry at later ages than women in lower strata. It turns out, however, that the effect of changes in age at marriage on fertility is relatively modest, compared to the magnitude of changes in marital fertility.
The socio-economic factors selected for analysis are three: type of place of residence, wife's education, and husband's occupation. These factors proved to be the three most important socio-economic determinants of marital fertility in our earlier analyses of WFS data, which considered in addition husband's education and husband's and wife's work status. The two work-status variables could not be included in the present study because they cannot be constructed from the information collected in the DHS. As regards husband's education, previous work has found that it adds rather little explanatory power to wife's education, so its introduction would be rather repetitious at this point.
Our main methodological tool is a model of the period marital fertility of individual women. The model has been extensively used with individual and aggregate WFS data ( Rodríguez and Cleland 1987, 1988), and quite recently with aggregate DHS data ( Moreno 1990). This chapter represents its first application to individual level DHS data. The model expresses fertility in terms of a spacing component, which is assumed to operate equally at all union durations, and a