Cash Transfers and Fertility

Notes: Estimates of discontinuity at the cutoff in an RD–DiD framework. Weekly data on the 50 Spanish provinces in the year of announcement (treatment year) and in the preceding year (control year). Dependent variables are logarithm of number of abortions per day in each week among women aged 15–44 years (Panel A) and the corresponding abortion rate, expressed per 100,000 women aged 15–44 years (Panel B). Forcing variable is the week of abortion; weeks are created such that they start on July 4 (Column 1), May 13 (Column 2), and August 1 (Column 3). Data are restricted to eight weeks on each side of the cutoff. Data in Column 3 cover eight weeks prior to May 13 on the left side of the cutoff (that is, March 14–May 8) in order to exclude the problematic period of May 13–July 31. Abortions related to fetal deformations are excluded from the sample. Standard errors are clustered at the province level.

Notes: OLS regressions. Monthly data on the 50 Spanish provinces between 01/2000 and 12/2017. Dependent variable is logarithm of number of births per day in each calendar month among women aged 15–44 years in the specific subgroup. Fully interacted model: all variables are interacted with dummy variable “younger,” “not married,” and “no parent is high-skilled” in Columns 1–3, respectively, and the dummy variable itself is also included. Division into younger (15–32 years) and older (33–44 years) age groups in Column 1 is based on median maternal age in 2000–2017. Category “not-high-skilled” includes low-skilled individuals and those out of the labor force. Births in 12/2010 and 01/2011 are set to missing. (Un)employment rates are included with a lag of three quarters. Standard errors are clustered at the province level.

Notes: Estimates of discontinuity at the cutoff in an RD-DiD framework. Estimations equivalent to those in Table 2 but on data from different subsamples. Fully interacted model: all variables are interacted with dummy variable “younger,” “not married,” “less than high school,” “poor province,” and “stronger crisis” in Columns 1–5, respectively, and the dummy variable itself is also included. Division into younger (15–32 years) and older (33–44 years) age groups in Column 1 is based on median maternal age in birth data in 2000–2017. Data are restricted to eight weeks on each side of the cutoff. Abortions related to fetal deformations are excluded. Sample size is 3,200. Standard errors are clustered at the province level.

Notes: OLS regressions. Monthly data on the 50 Spanish provinces between 01/2000 and 12/2017. Dependent variables are logarithm of number of births per day in each calendar month among women aged 15–44 years (Columns 1 and 3) and the corresponding birth rate, expressed per 100,000 women aged 15–44 years (Columns 2 and 4). Columns 1–2 show interaction terms with a dummy variable that takes a value of one for the 25 provinces with lower GDP per capita in 2007 and 2010, and zero otherwise. The dummy variable is also included in the model but not shown. Columns 3–4 show interaction terms with a dummy variable that takes a value of one for the 25 provinces more affected by the economic crisis, and zero otherwise. Births in 12/2010 and 01/2011 are set to missing. (Un)employment rates are included with a lag of three quarters. Standard errors are clustered at the province level.

Notes: OLS regressions. Monthly data on the 50 Spanish provinces between 01/2000 and 12/2017. Dependent variable is the logarithm of number of births per day in each calendar month among women aged 15–44 years. Births in 12/2010 and 01/2011 are set to missing. (Un)employment rates are included with a lag of three quarters. Standard errors are clustered at the province level.