Causation Among Socioeconomic Time-Series
Robert T. Michael
NBER Working Paper No. 246
Using annual U. S. time series data from 1950-1974, formal tests of causation are performed among three socioeconomic phenomena: women's labor force participation rates, fertility rates, and divorce rates. Box-Jenkins and other techniques are employed with Granger-Sims type definition of causation based on leads and lags. Women's labor force participation appears to be causally prior to both fertility and divorce; the direction of effect on fertility is negative and on divorce, positive. Additional tests with alternative definitions of variables and a longer (1924-1974) time span also exhibit causal influence from fertility to divorce (with no feedback). When per capita income is also tested for causal influence, it, too, appears causally prior to fertility and divorce.
Document Object Identifier (DOI): 10.3386/w0246
Users who downloaded this paper also downloaded these: