Social Capital, Gender Competition, and the Resurgence of Childlessness

Abstract

As fertility in much of the developed world remains far below replacement level, it is appropriate to look closely at trends and patterns in childlessness. The rationale for doing so is reinforced by a multidimensional scaling analysis of 80 countries, which finds that contemporary fertility patterns are largely determined by two factors: the overall level and the proportion childless. Parity status life tables for 24 low fertility nations for periods since 2000 show that 15 of them have period parity progression rates implying that over 20% of women will never have any children. Commonly used figures on proportions at parity zero in cohorts completing their reproductive years have understated the level of childlessness inherent in recent data by ignoring the behavior of younger cohorts. Still, even those cohort data reveal an upward trend in childlessness. The likelihood of a resurgence in childlessness is bolstered by steady increases in the mean ages at first birth observed in all 24 study populations. Looking ahead, high proportions childless can be consistent with stable, egalitarian unions, as children now bring few resources to parents while making great demands upon them.

Appendices

Appendices

1.1 Appendix A: Multidimensional Scaling

We began by collecting recent data on completed parity distributions for countries included either in the UNdata website (72 countries) or in the Human Fertility Database (8 countries: Canada, Japan, the Netherlands, Norway, Portugal, Sweden, Taiwan and the United States). A list of all countries and their abbreviations in Fig. 2.1 is given below. Although the countries represented in the UNdata website give data for varying numbers of parity categories, we were able to gather data for all 80 countries in our analysis for parities zero through parity 5 and above. From these data we constructed five parity progression ratios for each of the 80 countries.

We analyzed the 80 by 5 matrix of parity progression ratios using Stata’s mds command (see https://www.stata.com/manuals13/mvmds.pdf), and began by calculating standardized Euclidean distances between the set of parity progression ratios for each pair of countries. The mds command then carried out a principal coordinates analysis (“classical metric scaling”) of the 3160 Euclidean distances calculated in the first step. Plotting the eigenvalues associated with the five dimensions produced by this analysis showed that there was little decline after the third dimension, suggesting that a two dimensional representation of the distances is appropriate (Davison 1983, p. 69). Figure 2.1 presents the 80 countries’ positions on the first two dimensions yielded by our analysis.

Abbreviations for Countries in Fig. 2.1

Albania	Alb	Luxembourg	Lux
Australia	Aus	Macedonia	Mac
Austria	Aust	Malawi	Mala
Azerbaijan	Azer	Malta	Malt
Bahamas	Bah	Mauritius	Maur
Bahrain	Bahr	Mexico	Mex
Barbados	Barb	Moldova	Mol
Belarus	Bela	Montenegro	Mont
Bhutan	Bhu	Morocco	Mor
Bolivia	Bol	Mozambique	Moz
Botswana	Bots	Nepal	Nep
Brazil	Braz	Netherlands	Neth
Bulgaria	Bul	New Zealand	NZ
Burkina Faso	B-F	Norway	Nor
Burundi	Bur	Palestine	Pale
Canada	Can	Paraguay	Para
Chili	Chl	Peru	Peru
Columbia	Col	Philippines	Phil
Costa Rica	C-R	Poland	Pol
Croatia	Cro	Portugal	Por
Czechia	Cz	Romania	Rom
Dominican Rep.	D-R	Russian Federation	RFed
Ecuador	Ecu	Rwanda	Rwa
Estonia	Est	South Korea	SKor
Finland	Fin	Serbia	Serb
France	Fr	Slovakia	Slvk
Georgia	Geo	Slovenia	Slvn
Ghana	Gha	Spain	Sp
Greece	Gr	Sri Lanka	SL
Hungary	Hun	Suriname	Suri
India	In	Sweden	Swe
Indonesia	Indo	Taiwan	Tai
Ireland	Ire	Tajikistan	Taj
Israel	Is	Thailand	Thai
Jamaica	Jam	Trinidad & Tobago	T&T
Japan	Jpn	Turkey	Tur
Kyrgyzstan	Kyr	Ukraine	Ukr
Latvia	Lat	Uruguay	Urug
Lesotho	Les	United States	US
Lithuania	Lit	Venezuela	Ven

1.2 Appendix B: Fertility Data

TFR values in Table 2.1 are based on yearly TFRs reported in the Human Fertility Database (HFD). We report the mean of the period TFR values for each group of years shown. Similarly, PSLT P0 values in Table 2.1 are means of proportions of women age 49 at parity zero for each group of years reported in the HFD period fertility life tables. Values for mean age at first birth are also means of annual values for each group of years reported for each country in the “table mean ages at birth” section of the HFD period fertility tables.

We estimated the cohort childlessness proportions in Table 2.1 using HFD data on exposure by year, age and parity. For each country and group of years, we summed the exposure values for ages 45 through 49 and divided that sum by the sum of the exposure values for all parities for those ages. In a few cases exposure values were not available for all five ages. In these cases we averaged the ages reported by the HFD.