Reconstrução de histórias de nascimentos a partir de dados censitários: aspectos teóricos e evidências empíricas

Abstract

Brazil has experienced a steady fertility decline during the last fifty years, despite the lack of socioeconomic and structural changes or any explicit family planning policies. The low fertility levels leads to new challenges, one of them related to fertility measurement. Although Brazil has very good quality census data, the debate on future fertility trends and their political/economic consequences has been limited by the lack of good quality birth histories and vital registration. The only empirical evidence used by the experts is the measurement of period TFRs (incidence rates). As fertility reaches lower levels, new dimensions assume an important role in fertility analysis for instance, the current debate about tempo and quantum. Taking those dimensions into consideration is crucial for understanding the fertility transition in Brazil. The main objective of this dissertation is to determine the more satisfactory methodology of birth history reconstruction to the Brazilian case in 2000, based on two methodologies one being developed at Cedeplar, and one elaborated by Luther and Cho (1986). Data come from the Brazilian Demographic Census. The specific objectives of this dissertation are: (i) to apply the more satisfactory birth history reconstruction methodology to the 1980 and 1991 Brazilian Demographic Censuses; (ii) to apply the Own Children Method to the last three editions of the Brazilian demographic censuses; (iii) to compare the TFR series produced by the birth history reconstruction methodology to the series produced by the Own Children Method ; (iv) to apply the Köhler and Ortega method to disentangle tempo and parity composition effects from the observed TFR; and (v) to determine alternative period fertility measures, based on birth probabilities. Results suggest that the methodology of birth history reconstruction developed at Cedeplar HNM is more satisfactory to the Brazilian case, as it produced better results when applied to the 2000 Brazilian Demographic Census. Data from the 1980 and 1991 censuses also yielded satisfactory results when compared to the Own Children Method results. The application of the Köhler and Ortega method and the determination of the alternative period fertility measures suggest that the fertility decline in Brazil was accompanied by a negative tempo effect and a significant parity composition effect. At the end of the 1990s, the results indicate that Brazil will probably enter into the fertility postponement process with a dominant tempo effect. If this scenario is confirmed, Brazil will experience very low levels of observed TFR in the near future.