The Mastery of Passions Socioeconomic Status and the Fertility Transition in Stockholm, 1878-1926 Joseph Molitoris & Martin Dribe Centre for Economic Demography Lund University Funded by:
Motivation Urbanization and modernization play implicit or explicit roles in theories of fertility decline Few studies of urban populations during fertility decline Social groups have role in most theories as vehicles of change Diffusion: behaviors, innovations, preferences pass within and between groups Adjustment: SES group determines response to new structural conditions
Data: Roteman s Archive Period: 1878 1926 Extraction: All females age 15 to 49 ever living in Stockholm and all individuals connected to them 3.7 M observations of 970,940 individuals (59.6% women) HISCLASS used to create SES groups from HISCO classification (Van Leeuwen, Maas & Miles, 2002)
Data: Variable Management Socioeconomic Status Husband s/head of Household s applied to wives/cohab Single women retained own SES Missing dates were filled in providing maximum exposure (i.e. Jan 1 or Dec 31 if month and day were missing) Overestimates of exposure time Created 8 time invariant districts by consolidating districts
Stockholm s Demographic Context City expanded at mean annual rate of 2.2% from 1870 to 1930 High adult male mortality 1881-1890: Stockholm e 25 =32.6; All of Sweden e 25 = 40.6 IMR was as high as 200 deaths per 1000 births in 1880 Swedish IMR 1880 =120 per 1000 Unique marriage arrangements
Stockholm s Population Growth Population 500 000 450 000 400 000 350 000 300 000 250 000 200 000 150 000 100 000 50 000 Total Population and Population Growth in Stockholm by Year 10 8 6 4 2 0 % Change 0 1878 1883 1888 1893 1898 1903 1908 1913 1918 1923 Population rate of change -2 Source: Folkmängden i Stockholm 1252-2012. Stockholms Stad.
Methods Descriptive: TMFR20 & TFR Multivariate Analysis Piecewise constant hazard rate models» Baseline hazard functions were assumed to have been constant for 6 month spells» Parity 2+ model used shared frailty of mothers» DVs: Age at first birth/time since last birth
Methods Multivariate Analysis (cont.) Logistic Regression» DV: Pr(interval>5 years) Sensitivity: Pr(interval>8 years); do not count failstopping as true stopping» Std. Errors: clustered by mother
Descriptive Results Births per woman 5 4.5 4 3.5 3 2.5 2 1.5 1 0.5 0 Total Fertility Rate by SES and Period 1878-1889 1890-1901 1902-1913 1914 1926 Higher occupations Lower managers Skilled Workers Lower Skilled Unskilled TOTAL
Descriptive Results 9 TMFR20 by SES and Period Births per married woman age 20 to 49 8 7 6 5 4 3 2 1 0 1878-1889 1890-1901 1902-1913 1914-1926 Higher occupations Lower managers Skilled Workers Lower Skilled Unskilled TOTAL
Piecewise Constant Hazard Models & Logistic Regression Model First Births Parity 2+ P(Interval> 5 years) HR HR OR SES: Higher Occupations 1.00 1.00 1.00 Lower managers. professionals 1.00 1.00 1.03 Skilled Workers 1.06 1.08*** 0.87*** Lower Skilled Workers 1.16*** 1.16*** 0.79*** Unskilled Workers 1.47*** 1.31*** 0.68*** Missing 1.41*** 2.14*** 0.43*** Period: 1878-1889 1.00 1.00 1.00 1890-1901 0.97 0.75*** 1.37*** 1902-1913 0.98 0.58*** 2.08*** 1914-1926 0.62*** 0.31*** 3.53*** District: Gamla Stan 1.00 1.00 1.00 Norrmalm 0.76*** 0.93*** 1.01 Kungsholmen 1.61*** 1.21*** 0.83*** Östermalm 0.99 1.13*** 0.86*** Södermalm East 1.76*** 1.29*** 0.79*** Södermalm West 1.62*** 1.24*** 0.82*** Brännkyrka 3.78*** 1.62*** 0.74*** Mother's Birthplace: Stockholm City 1.00 1.00 1.00 Stockholm county 1.28*** 1.03** 0.92*** Other Sweden 1.20*** 1.07*** 0.92*** Outside Sweden 0.94 1.15*** 0.98 Previous Child Dead 1.17*** 0.77*** chi2 377 093 360 200.8 21 314.7 Women at Risk 130 325 124 668 105 445 Births/Intervals 34 021 168 065 250 650
Multivariate Results Hazard Ratio 2 1.8 1.6 1.4 1.2 1 0.8 0.6 0.4 0.2 0 Net Effects Between SES Groups on Age at First Birth 1878-1889 1890-1901 1902-1913 1914-1926 Higher Occupations Lower managers, professionals Skilled Workers Lower Skilled Workers Unskilled Workers
Multivariate Results 1.8 Net Effects Within SES Group on Age at First Birth 1.6 1.4 1.2 Hazard Ratio 1 0.8 0.6 0.4 0.2 0 1878-1889 1890-1901 1902-1913 1914-1926 Higher Occupations Lower managers, professionals Skilled Workers Lower Skilled Workers Unskilled Workers
Multivariate Results 1.6 Net Effects Between SES Groups on Time Since Last Birth Hazard Ratio 1.4 1.2 1 0.8 0.6 0.4 0.2 0 1878-1889 1890-1901 1902-1913 1914-1926 Higher Occupations Skilled Workers Unskilled Workers Lower managers, professionals Lower Skilled Workers
Multivariate Results 1.4 Net Effects Within SES Groups on Time Since Last Birth 1.2 1 Hazard Ratio 0.8 0.6 0.4 0.2 0 1878-1889 1890-1901 1902-1913 1914-1926 Higher Occupations Lower managers, professionals Skilled Workers Lower Skilled Workers Unskilled Workers
Multivariate Results 1.6 Net Effects Between SES Groups on Pr( Interval>5) Odds Ratio 1.4 1.2 1 0.8 0.6 0.4 0.2 0 1878-1889 1890-1901 1902-1913 1914-1926 Higher Occupation Lower Managers Skilled Workers Lower Skilled Unskilled
Multivariate Results 8 Net Effects Within SES Groups on Pr(Interval>5) Odds Ratio 7 6 5 4 3 2 1 0 1878-1889 1890-1901 1902-1913 1914-1926 Higher Occupation Lower Managers Skilled Workers Lower Skilled Unskilled
Discussion Some evidence that elites were regulating fertility earlier Start of period: Highest age at first birth, longest time between births, highest odds of a long interval (stopping) Behavior may have spread throughout population Differences between groups became smaller over time Most dramatic change among middle class (skilled workers) End of period: highest age at first birth, longest birth intervals, and highest risk of long interval