First-time mothers and the labor market effects of the earned income tax credit

Peter Shirley

doi:10.2478/izajolp-2020-0007

Table 1

Descriptive statistics for first-time mothers

Variable	All unmarried women	Low-ed unmarried women	All married women	Low-ed married women
Employed	0.48	0.37	0.56	0.37
Monthly earnings (2016 dollars)	1,037	609	2,079	739
Maximum credit (2016 dollars) \| Jul–Dec birth	3,478	3,457	3,467	3,418
Cohort	2004	2004	2003	2002
Fraction low-ed	0.54	1	0.23	1
Fraction black	0.24	0.22	0.06	0.07
Age	24.7	23.4	28.8	25.7
Fraction with Jan–Jun birth	0.49	0.48	0.50	0.48
Fraction with Jul–Dec birth	0.51	0.52	0.50	0.52
Employed in year prior to birth1	0.58	0.46	0.75	0.52
Employed in year prior to birth \| Jan–Jun birth1	0.59	0.47	0.75	0.50
Employed in year prior to birth \| Jul–Dec birth1	0.56	0.45	0.74	0.54
Monthly earnings in year prior to birth1	1,357	793	2,923	1,108
Monthly earnings in year prior to birth \| Jan–Jun birth1	1,396	842	3,008	1,041
Monthly earnings in year prior to birth \| Jul–Dec birth1	1,318	746	2,839	1,164
Number of person-month observations	17,883	9,698	41,834	9,803
Number of unique mothers	2,665	1,516	5,378	1,354

Data Source: Survey of Income and Program Participation, 1996 to 2014 Panels.

1 These measures are created in a similar method to the primary sample but are calculated over the 12 months prior to a woman’s first/only birth.

Table 2

Effects of the ETIC on labor market outcomes, baseline regressions for women’s first child

Functional form	All unmarried women			Low-ed unmarried women			All married women			Low-ed married women
Panel A: Employment

Linear	0.002	0.006	0.047	-0.017	-0.001	0.044	-0.045	-0.007	0.002	0.017	0.006	0.007
	(0.032)	(0.009)	(0.036)	(0.043)	(0.012)	(0.054)	(0.026)	(0.008)	(0.024)	(0.050)	(0.015)	(0.056)
Quadratic	0.045	0.024	0.050	0.012	0.010	0.046	-0.027	0.004	0.003	0.102	0.030	0.023
	(0.057)	(0.016)	(0.038)	(0.079)	(0.021)	(0.055)	(0.034)	(0.013)	(0.025)	(0.096)	(0.027)	(0.058)

Panel B: IHS earnings

Linear	-0.054	0.039	0.446	-0.244	0.036	0.351	-0.428*	-0.065	0.038	0.069	0.032	0.017
	(0.271)	(0.079)	(0.303)	(0.342)	(0.096)	(0.412)	(0.233)	(0.076)	(0.217)	(0.414)	(0.124)	(0.475)
Quadratic	0.297	0.188	0.473	0.072	0.084	0.376	-0.334	0.031	0.038	0.811	0.243	0.155
	(0.453)	(0.125)	(0.320)	(0.622)	(0.168)	(0.427)	(0.310)	(0.128)	(0.231)	(0.804)	(0.229)	(0.495)
N		17,883			9,698			41,834			9,803

Panel C: Log earnings

Linear	0.029	0.023	0.112	0.065	0.037	0.082	-0.062	0.006	-0.078	-0.079	-0.009	-0.226*
	(0.095)	(0.027)	(0.091)	(0.128)	(0.034)	(0.138)	(0.066)	(0.020)	(0.049)	(0.131)	(0.037)	(0.127)
Quadratic	0.050	0.044	0.106	0.083	0.062	0.087	-0.184*	0.005	-0.096*	0.050	0.042	-0.157
	(0.155)	(0.040)	(0.091)	(0.175)	(0.044)	(0.143)	(0.097)	(0.035)	(0.051)	(0.195)	(0.051)	(0.140)
N		8,883			3,805			23,758			3,684
To 1 Disc.	Y	N	N	Y	N	N	Y	N	N	Y	N	N
To Max Cr. Disc.	N	Y	Y	N	Y	Y	N	Y	Y	N	Y	Y
Controls	N	N	Y	N	N	Y	N	N	Y	N	N	Y

[i] Data Source: Survey of Income and Program Participation, 1996 to 2014 Panels.

Each cell shows the parameter of interest from a separate regression. All regressions also include time (month of birth) and time*second-half of the year birth (July-December). Quadratic functional form regressions also include time squared and time squared*second half birth.

Control variables include state and cohort fixed effects as well as age, age squared, and black race indicator. Also included is (the sum of personal tax exemption, CTC, and single to head of household deduction difference*unmarried)*second half of the year birth.

Employment is defined as having monthly earnings of at least $250 in 2016 dollars.

Low-ed is defined as having a high school degree or less.

Standard errors are clustered at the state level.

*p<0.1, **p<0.05, ***p<0.01.

Table 3

Effects of the ETIC on labor market outcomes, one child to two children diff-in-disc results

Functional form	All unmarried women	Low-ed unmarried women	All married women	Low-ed married women
Panel A: Employment

Linear	0.023	0.004	0.006	−0.021
	(0.025)	(0.031)	(0.018)	(0.036)
Quadratic	0.024	0.005	0.008	−0.008
	(0.030)	(0.034)	(0.020)	(0.037)

Panel B: IHS earnings

Linear	0.281	0.117	0.089	−0.218
	(0.196)	(0.234)	(0.159)	(0.299)
Quadratic	0.280	0.125	0.098	−0.116
	(0.239)	(0.265)	(0.178)	(0.307)
N	33,043	19,215	86,875	23,130

Panel C: Log earnings

Linear	0.107	0.143	−0.009	−0.140*
	(0.077)	(0.097)	(0.043)	(0.077)
Quadratic	0.105	0.144	−0.033	−0.098
	(0.081)	(0.097)	(0.036)	(0.082)
N	15,647	7,274	46,496	8,448
To 1 Disc.	N	N	N	N
To Max Cr. Disc.	Y	Y	Y	Y
Controls	Y	Y	Y	Y

[i] Data Source: Survey of Income and Program Participation, 1996 to 2014 Panels.

Each cell shows the parameter of interest from a separate regression. All regressions also include time (month of birth) and time*second-half of the year birth (July–December). Quadratic functional form regressions also include time squared and time squared*second half birth. Each of these parameters is included twice, each time interacted with an indicator for one or two children. The coefficient shown is the difference between the discontinuity estimates for one child and two children.

Control variables include state and cohort fixed effects as well as age, age squared, and black race indicator. Also included is (the sum of personal tax exemption, CTC, and single to head of household deduction difference*unmarried)*second half of the year birth.

Employment is defined as having monthly earnings of at least $250 in 2016 dollars.

Low-ed is defined as having a high school degree or less.

Standard errors are clustered at the state level.

*p<0.1, **p<0.05, ***p<0.01.

Table 4

Effects of the ETIC on labor market outcomes, results for women’s second child

Functional form	All unmarried women	Low-ed unmarried women	All married women	Low-ed married women
Panel A: Employment

Linear	−0.020**	−0.022**	−0.005	−0.004
	(0.008)	(0.010)	(0.006)	(0.012)
Quadratic	−0.015	−0.018	0.002	−0.001
	(0.010)	(0.015)	(0.008)	(0.015)

Panel B: IHS earnings

Linear	−0.160**	−0.166**	−0.019	−0.021
	(0.065)	(0.079)	(0.057)	(0.095)
Quadratic	−0.105	−0.102	0.040	0.006
	(0.079)	(0.117)	(0.069)	(0.125)
N	15,160	9,517	45,041	13,327

Panel C: Log earnings

Linear	0.005	0.006	0.003	−0.041
	(0.029)	(0.031)	(0.019)	(0.029)
Quadratic	0.008	0.014	0.017	−0.030
	(0.029)	(0.031)	(0.023)	(0.038)
N	6,764	3,649	22,738	4,764
To 1 Disc.	N	N	N	N
To Max Cr. Disc.	Y	Y	Y	Y
Controls	Y	Y	Y	Y

[i] Data Source: Survey of Income and Program Participation, 1996 to 2014 Panels.

Each cell shows the parameter of interest from a separate regression. All regressions also include time (month of birth) and time*second-half of the year birth (July–December). Quadratic functional form regressions also include time squared and time squared*second half birth. Each of these parameters is included twice, each time interacted with an indicator for one or two children. The coefficient shown is the difference between the discontinuity estimates for one child and two children.

Control variables include state and cohort fixed effects as well as age, age squared, and black race indicator. Also included is (the sum of personal tax exemption, CTC, and single to head of household deduction difference*unmarried)*second half of the year birth.

Employment is defined as having monthly earnings of at least $250 in 2016 dollars.

Low-ed is defined as having a high school degree or less.

Standard errors are clustered at the state level.

*p<0.1, **p<0.05, ***p<0.01.

Table 5

Effects of the ETIC on labor market outcomes, regressions for women with first children by age of child

Functional form	All unmarried women		Low-ed unmarried women		All married women		Low-ed married women
Age of child	0-5 months	6-11 months	0-5 months	6-11 months	0-5 months	6-11 months	0-5 months	6-11 months

Panel A: Employment

Linear	0.026	0.022	0.026	-0.010	0.010	0.002	-0.012	-0.028
	(0.032)	(0.030)	(0.044)	(0.037)	(0.018)	(0.024)	(0.037)	(0.050)
Quadratic	0.024	0.025	0.026	-0.006	0.015	0.001	-0.001	-0.011
	(0.034)	(0.036)	(0.045)	(0.042)	(0.019)	(0.026)	(0.035)	(0.052)

Panel B: IHS earnings

Linear	0.301	0.277	0.331	-0.036	0.123	0.062	-0.093	-0.308
	(0.254)	(0.238)	(0.344)	(0.295)	(0.163)	(0.205)	(0.309)	(0.418)
Quadratic	0.285	0.287	0.330	-0.011	0.154	0.048	-0.027	-0.167
	(0.279)	(0.282)	(0.349)	(0.334)	(0.170)	(0.231)	(0.294)	(0.439)
N	16,129	16,914	9,397	9,818	42,362	44,513	11,005	12,125

Panel C: Log earnings

Linear	0.117	0.092	0.249*	0.022	-0.025	-0.016	-0.164*	-0.151*
	(0.107)	(0.075)	(0.134)	(0.112)	(0.046)	(0.050)	(0.95)	(0.088)
Quadratic	0.119	0.087	0.244*	0.035	-0.047	-0.041	-0.098	-0.122
	(0.108)	(0.085)	(0.126)	(0.122)	(0.043)	(0.042)	(0.098)	(0.098)
N	7,047	8,600	3,222	4,052	22,020	24,476	3,770	4,678
To 1 Disc.	N	N	N	N	N	N	N	N
To Max Cr. Disc.	Y	Y	Y	Y	Y	Y	Y	Y
Controls	Y	Y	Y	Y	Y	Y	Y	Y

[i] Data Source: Survey of Income and Program Participation, 1996 to 2014 Panels.

Each cell shows the parameter of interest from a separate regression. All regressions also include time (mont h of birth) and time*second-haIf of the year birth (July-December). Quadratic functional form regressions also include time squared and time squared*second half birth. Each of these parameters is included twice, each time interacted with an indicator for one or two children. The coefficient shown is the difference between the discontinuity estimates for one child and two children.

Control variables include state and cohort fixed effects as well as age, age squared, and black race indicator. Also included is (the sum of personal tax exemption, CTC, and single to head of house-hold deduction difference*unmarried)*second half of the year birth.

Employment is defined as having monthly earnings of at least $250 in 2016 dollars.

Low-ed is defined as having a high school degree or less.

Standard errors are clustered at the state level.

* p<0.1, ** p<0.05, ***p<0.01.

Table 6

Effects of the ETIC on labor market outcomes, regressions for women’s first child, November to February births only

Functional All form	unmarried women	Low-ed unmarried women	All married women	Low-ed married women
Panel A: Employment
Linear	−0.022	−0.006	0.025	−0.089
	(0.061)	(0.065)	(0.041)	(0.091)
Panel B: IHS earnings
Linear	−0.063	0.094	0.212	−0.851
	(0.511)	(0.546)	(0.357)	(0.748)
N	11,469	6,557	28,612	7,876
Panel C: Log earnings
Linear	0.132	−0.124	0.021	−0.262
	(0.113)	(0.146)	(0.066)	(0.164)
N	5,610	2,596	14,960	2,691
To 1 Disc.	N	N	N	N
To Max Cr. Disc.	Y	Y	Y	Y
Controls	Y	Y	Y	Y

[i] Data Source: Survey of Income and Program Participation, 1996 to 2014 Panels.

Each cell shows the parameter of interest from a separate regression. The coefficient shown is the difference between the discontinuity estimates for one child and two children.

Control variables include state and cohort fixed effects as well as age, age squared, and black race indicator. Also included is (the sum of personal tax exemption, CTC, and single to head of household deduction difference*unmarried)*second half of the year birth.

Employment is defined as having monthly earnings of at least $250 in 2016 dollars.

Low-ed is defined as having a high school degree or less.

Standard errors are clustered at the state level.

*p<0.1, **p<0.05, ***p<0.01.

Table 7

Effects of the ETIC on labor market outcomes, regressions for women’s first child excluding December and January births

Functional All form	unmarried women	Low-ed unmarried women	All married women	Low-ed married women
Panel A: Employment
Linear	0.037	–0.005	0.002	–0.015
	(0.030)	(0.041)	(0.017)	(0.041)
Quadratic	0.023	–0.013	–0.008	0.013
	(0.038)	(0.049)	(0.020)	(0.043)
Panel B: IHS earnings
Linear	0.385*	0.054	0.074	–0.121
	(0.221)	(0.315)	(0.149)	(0.338)
Quadratic	0.249	–0.015	–0.036	0.073
	(0.287)	(0.378)	(0.171)	(0.375)
N	27,153	15,801	72,052	18,787
Panel C: Log earnings
Linear	0.094	0.105	0.003	–0.114
	(0.090)	(0.111)	(0.049)	(0.085)
Quadratic	0.049	0.058	–0.037	0.010
	(0.104)	(0.110)	(0.052)	(0.088)
N	12,714	5,934	38,833	7,012
To 1 Disc.	N	N	N	N
To Max Cr. Disc.	Y	Y	Y	Y
Controls	Y	Y	Y	Y

[i] Data Source: Survey of Income and Program Participation, 1996 to 2014 Panels.

Each cell shows the parameter of interest from a separate regression. All regressions also include time (month of birth) and time*second-half of the year birth (July–December). Quadratic functional form regressions also include time squared and time squared*second half birth. Each of these parameters is included twice, each time interacted with an indicator for one or two children. The coefficient shown is the difference between the discontinuity estimates for one child and two children.

Control variables include state and cohort fixed effects as well as age, age squared, and black race indicator. Also included is (the sum of personal tax exemption, CTC, and single to head of household deduction difference*unmarried)*second half of the year birth.

Employment is defined as having monthly earnings of at least $250 in 2016 dollars.

Low-ed is defined as having a high school degree or less.

Standard errors are clustered at the state level.

*p<0.1, **p<0.05, ***p<0.01.

Table 8

Effects of the ETIC on labor market outcomes, regressions for women’s first child excluding age 0 months

Functional form	All unmarried women	Low-ed unmarried women	All married women	Low-ed married women
Panel A: Employment
Linear	0.022	0.006	0.005	−0.020
	(0.024)	(0.031)	(0.019)	(0.038)
Quadratic	0.024	0.008	0.007	−0.004
	(0.030)	(0.035)	(0.021)	(0.039)
Panel B: IHS earnings
Linear	0.272	0.118	0.088	−0.204
	(0.193)	(0.235)	(0.165)	(0.315)
Quadratic	0.279	0.128	0.091	−0.086
	(0.238)	(0.269)	(0.186)	(0.326)
N	30,556	17,790	80,221	21,442
Panel C: Log earnings
Linear	0.103	0.127	−0.014	−0.147*
	(0.075)	(0.097)	(0.044)	(0.077)
Quadratic	0.101	0.131	−0.039	−0.105
	(0.080)	(0.099)	(0.037)	(0.081)
N	14,648	6,844	42,982	7,914
To 1 Disc.	N	N	N	N
To Max Cr. Disc.	Y	Y	Y	Y
Controls	Y	Y	Y	Y

[i] Data Source: Survey of Income and Program Participation, 1996 to 2014 Panels.

Each cell shows the parameter of interest from a separate regression. All regressions also include time (month of birth) and time*second-half of the year birth (July–December). Quadratic functional form regressions also include time squared and time squared*second half birth. Each of these parameters is included twice, each time interacted with an indicator for one or two children. The coefficient shown is the difference between the discontinuity estimates for one child and two children.

Control variables include state and cohort fixed effects as well as age, age squared, and black race indicator. Also included is (the sum of personal tax exemption, CTC, and single to head of household deduction difference*unmarried)*second half of the year birth.

Employment is defined as having monthly earnings of at least $250 in 2016 dollars.

Low-ed is defined as having a high school degree or less.

Standard errors are clustered at the state level.

*p<0.1, **p<0.05, ***p<0.01.

Table 9

Effects of the ETIC on labor market outcomes, regressions for women’s first child, post-tax season

Functional form	All unmarried women	Low-ed unmarried women	All married women	Low-ed married women
Panel A: Employment
Linear	0.018	0.017	0.011	–0.002
	(0.030)	(0.036)	(0.024)	(0.052)
Quadratic	0.021	0.021	0.010	0.010
	(0.037)	(0.041)	(0.026)	(0.053)
Panel B: IHS earnings
Linear	0.219	0.177	0.131	–0.064
	(0.243)	(0.281)	(0.204)	(0.427)
Quadratic	0.239	0.209	0.114	0.033
	(0.293)	(0.322)	(0.224)	(0.442)
N	25,583	14,862	67,191	17,940
Panel C: Log earnings
Linear	0.038	0.044	0.008	–0.140
	(0.073)	(0.111)	(0.049)	(0.089)
Quadratic	0.029	0.057	–0.023	–0.129
	(0.079)	(0.116)	(0.043)	(0.098)
N	12,369	5,753	36,150	6,650
To 1 Disc.	N	N	N	N
To Max Cr. Disc.	Y	Y	Y	Y
Controls	Y	Y	Y	Y

[i] Data Source: Survey of Income and Program Participation, 1996 to 2014 Panels.

Each cell shows the parameter of interest from a separate regression. All regressions also include time (month of birth) and time*second-half of the year birth (July–December). Quadratic functional form regressions also include time squared and time squared*second half birth. Each of these parameters is included twice, each time interacted with an indicator for one or two children. The coefficient shown is the difference between the discontinuity estimates for one child and two children.

Control variables include state and cohort fixed effects as well as age, age squared, and black race indicator. Also included is (the sum of personal tax exemption, CTC, and single to head of household deduction difference*unmarried)*second half of the year birth.

Employment is defined as having monthly earnings of at least $250 in 2016 dollars.

Low-ed is defined as having a high school degree or less.

Standard errors are clustered at the state level.

*p<0.1, **p<0.05, ***p<0.01.

Table 10

Effects of the ETIC on labor market outcomes, regressions for women with first children by cohort

Functional form	All unmarried women		Low-ed unmarried women		All married women		Low-ed married women
Cohorts	1995–2005	2006–2016	1995–2005	2006–2016	1995–2005	2006–2016	1995–2005	2006–2016
Panel A: Employment
Linear	−0.019	0.063	−0.044	0.070	0.002	−0.002	−0.032	0.008
	(0.039)	(0.046)	(0.046)	(0.061)	(0.024)	(0.036)	(0.047)	(0.077)
Quadratic	−0.024	0.075	−0.045	0.082	0.007	−0.007	−0.013	−0.009
	(0.042)	(0.049)	(0.049)	(0.060)	(0.029)	(0.034)	(0.048)	(0.079)
Panel B: IHS earnings
Linear	−0.080	0.629	−0.297	0.643	0.072	−0.038	−0.309	0.114
	(0.316)	(0.382)	(0.366)	(0.501)	(0.220)	(0.311)	(0.403)	(0.616)
Quadratic	−0.116	0.699*	−0.301	0.689	0.103	−0.076	−0.159	−0.048
	(0.343)	(0.402)	(0.381)	(0.487)	(0.261)	(0.298)	(0.412)	(0.634)
N	18,365	14,678	11,633	7,582	56,305	30,570	16,757	6,373
Panel C: Log earnings
Linear	0.065	0.070	0.127	0.133	0.042	−0.054	−0.072	−0.237
	(0.083)	(0.161)	(0.121)	(0.157)	(0.053)	(0.044)	(0.089)	(0.215)
Quadratic	0.062	0.094	0.128	0.156	−0.011	−0.042	−0.065	−0.186
	(0.087)	(0.154)	(0.117)	(0.161)	(0.041)	(0.046)	(0.104)	(0.227)
N	8,532	7,115	4,458	2,816	29,037	17,459	6,252	2,196
To 1 Disc.	N	N	N	N	N	N	N	N
To Max Cr. Disc.	Y	Y	Y	Y	Y	Y	Y	Y
Controls	Y	Y	Y	Y	Y	Y	Y	Y

[i] Data Source: Survey of Income and Program Participation, 1996 to 2014 Panels.

Each cell shows the parameter of interest from a separate regression. All regressions also include time (month of birth) and time*second-half of the year birth (July–December). Quadratic functional form regressions also include time squared and time squared*second half birth. Each of these parameters is included twice, each time interacted with an indicator for one or two children. The coefficient shown is the difference between the discontinuity estimates for one child and two children.

Control variables include state and cohort fixed effects as well as age, age squared, and black race indicator. Also included is (the sum of personal tax exemption, CTC, and single to head of household deduction difference*unmarried)*second half of the year birth.

Employment is defined as having monthly earnings of at least $250 in 2016 dollars.

Low-ed is defined as having a high school degree or less.

Standard errors are clustered at the state level.

*p<0.1, **p<0.05, ***p<0.01.

Table 11

Effects of the ETIC on labor market outcomes, baseline regressions for all numbers of children

Functional form	All unmarried women			Low-ed unmarried women			All married women			Low-ed married women
Panel A: Employment
Linear	-0.043*	-0.014***	-0.015***	-0.054*	-0.011**	-0.011**	0.0004	-0.009***	-0.012***	0.037	-0.0002	-0.003
	(0.022)	(0.004)	(0.004)	(0.027)	(0.005)	(0.005)	(0.015)	(0.002)	(0.002)	(0.026)	(0.004)	(0.004)
Quadratic	-0.021	-0.015***	-0.010*	-0.049	-0.010*	-0.008	0.020	-0.012***	-0.012***	0.059	-0.003	-0.001
	(0.035)	(0.006)	(0.005)	(0.042)	(0.006)	(0.006)	(0.027)	(0.004)	(0.004)	(0.044)	(0.004)	(0.006)
Panel B: IHS earnings
Linear	-0.399**	-0.119***	-0.127***	-0.506**	-0.100**	-0.099**	0.004	-0.071***	-0.112***	0.301	0.005	-0.018
	(0.184)	(0.035)	(0.032)	(0.222)	(0.040)	(0.037)	(0.134)	(0.022)	(0.022)	(0.215)	(0.030)	(0.034)
Quadratic	-0.196	-0.125***	-0.085*	-0.376	-0.088*	-0.064	0.143	-0.106***	-0.112***	0.478	-0.016	-0.004
	(0.290)	(0.045)	(0.044)	(0.339)	(0.047)	(0.049)	(0.240)	(0.037)	(0.038)	(0.349)	(0.034)	(0.044)
N		42,200			25,285			111,580			32,497
Panel C: Log earnings
Linear	0.010	0.002	-0.020**	0.008	0.016	-0.008	-0.018	-0.005	-0.020***	-0.027	0.007	-0.001
	(0.052)	(0.008)	(0.009)	(0.070)	(0.013)	(0.012)	(0.041)	(0.008)	(0.005)	(0.082)	(0.013)	(0.013)
Quadratic	0.011	0.002	-0.020	0.002	0.024	-0.004	-0.035	-0.008	-0.022**	0.012	0.019	0.024
	(0.099)	(0.010)	(0.013)	(0.108)	(0.017)	(0.018)	(0.060)	(0.012)	(0.008)	(0.119)	(0.017)	(0.017)
N		19,318			9,304			56,118			11,093
To 1 Disc.	Y	N	N	Y	N	N	Y	N	N	Y	N	N
To Max Cr. Disc.	N	Y	Y	N	Y	Y	N	Y	Y	N	Y	Y
Controls	N	N	Y	N	N	Y	N	N	Y	N	N	Y

[i] Data Source: Survey of Income and Program Participation, 1996 to 2014 Panels.

Each cell shows the parameter of interest from a separate regression. All regressions also include time (month of birth) and time*second-haIf of the year birth (July-December). Quadratic functional form regressions also include time squared and time squared*second half birth.

Control variables include state and cohort fixed effects as well as age, age squared, and black race indicator. Also included is (the sum of personal tax exemption, CTC, and single to head of household deduction difference*unmarried)*second half of the year birth.

Employment is defined as having monthly earnings of at least $250 in 2016 dollars.

Low-ed is defined as having a high school degree or less.

Standard errors are clustered at the state level.

*p<0.1, **p<0.05, ***p<0.01.

Table A1

Effects of the ETIC on labor market outcomes, two children to three children diff-in-disc results

Functional All form	unmarried women	Low-ed unmarried women	All married women	Low-ed married women
Panel A: Employment

Linear	0.004	0.004	0.017***	−0.001
	(0.006)	(0.007)	(0.006)	(0.008)
Quadratic	0.009	0.017	0.016*	0.006
	(0.011)	(0.012)	(0.009)	(0.013)

Panel B: IHS earnings

Linear	0.027	0.022	0.167***	0.002
	(0.055)	(0.059)	(0.050)	(0.066)
Quadratic	0.074	0.0149	0.166**	0.052
	(0.091)	(0.098)	(0.079)	(0.105)
N	24,317	15,587	69,746	22,694

Panel C: Log earnings

Linear	0.017	0.013	0.027**	0.002
	(0.016)	(0.027)	(0.013)	(0.022)
Quadratic	0.016	−0.001	0.024	−0.001
	(0.018)	(0.031)	(0.020)	(0.031)
N	10,435	5,499	32,360	7,409
To 1 Disc.	N	N	N	N
To Max Cr. Disc.	Y	Y	Y	Y
Controls	Y	Y	Y	Y

[i] Data Source: Survey of Income and Program Participation, 1996 to 2014 Panels.

Each cell shows the parameter of interest from a separate regression. All regressions also include time (month of birth) and time*second-half of the year birth (July–December). Quadratic functional form regressions also include time squared and time squared*second half birth. Each of these parameters is included twice, each time interacted with an indicator for two or three children. The coefficient shown is the difference between the discontinuity estimates for two children and three children.

Control variables include state and cohort fixed effects as well as age, age squared, and black race indicator. Also included is (the sum of personal tax exemption, CTC, and single to head of household deduction difference*unmarried)*second half of the year birth.

Employment is defined as having monthly earnings of at least $250 in 2016 dollars.

Low-ed is defined as having a high school degree or less.

Standard errors are clustered at the state level.

*p<0.1, **p<0.05, ***p<0.01.

Table A2

Effects of the ETIC on labor market outcomes, one to two children and two children to three children difference-diff-in-disc results

Functional form	All unmarried women	Low-ed unmarried women	All married women	Low-ed married women
Panel A: Employment

Linear	0.029	0.021	−0.012	−0.007
	(0.026)	(0.035)	(0.018)	(0.038)
Quadratic	0.021	0.005	−0.012	−0.012
	(0.033)	(0.039)	(0.018)	(0.036)

Panel B: IHS earnings

Linear	0.341	0.243	−0.083	−0.100
	(0.214)	(0.263)	(0.150)	(0.321)
Quadratic	0.260	0.094	−0.103	−0.152
	(0.260)	(0.296)	(0.158)	(0.309)
N	42,200	25,285	111,580	32,497

Panel C: Log earnings

Linear	0.103	0.156*	−0.059	−0.040
	(0.073)	(0.085)	(0.048)	(0.089)
Quadratic	0.106	0.179**	−0.069	−0.038
	(0.082)	(0.083)	(0.046)	(0.103)
N	19,318	9,304	56,118	11,093
To 1 Disc.	N	N	N	N
To Max Cr. Disc.	Y	Y	Y	Y
Controls	Y	Y	Y	Y

[i] Data Source: Survey of Income and Program Participation, 1996 to 2014 Panels.

Each cell shows the parameter of interest from a separate regression. All regressions also include time (month of birth) and time*second-half of the year birth (July–December). Quadratic functional form regressions also include time squared and time squared*second half birth. Each of these parameters is included three times, each time interacted with an indicator for one, two, or three children. The coefficient shown is the difference between the discontinuity estimates for one child and two children minus the difference between the discontinuity estimates for two and three children.

Control variables include state and cohort fixed effects as well as age, age squared, and black race indicator. Also included is (the sum of personal tax exemption, CTC, and single to head of household deduction difference*unmarried)*second half of the year birth.

Employment is defined as having monthly earnings of at least $250 in 2016 dollars.

Low-ed is defined as having a high school degree or less.

Standard errors are clustered at the state level.

*p<0.1, **p<0.05, ***p<0.01.

Table A3

Effects of the ETIC on labor market outcomes, results for women’s first child ages 1 to 5 months

Functional form	All unmarried women	Low-ed unmarried women	All married women	Low-ed married women
Panel A: Employment

Linear	0.023	0.032	0.011	−0.009
	(0.031)	(0.044)	(0.018)	(0.037)
Quadratic	0.023	0.033	0.015	0.006
	(0.033)	(0.045)	(0.020)	(0.036)

Panel B: IHS earnings

Linear	0.269	0.351	0.131	−0.053
	(0.245)	(0.346)	(0.165)	(0.311)
Quadratic	0.271	0.353	0.153	0.042
	(0.272)	(0.354)	(0.174)	(0.304)
N	13,642	7,972	35,708	9,317

Panel C: Log earnings

Linear	0.109	0.215	−0.035	−0.181*
	(0.107)	(0.134)	(0.049)	(0.098)
Quadratic	0.111	0.211	−0.058	−0.116
	(0.108)	(0.128)	(0.045)	(0.098)
N	6,048	2,792	18,506	3,236
To 1 Disc.	N	N	N	N
To Max Cr. Disc.	Y	Y	Y	Y
Controls	Y	Y	Y	Y

[i] Data Source: Survey of Income and Program Participation, 1996 to 2014 Panels.

Each cell shows the parameter of interest from a separate regression. All regressions also include time (month of birth) and time*second-half of the year birth (July–December). Quadratic functional form regressions also include time squared and time squared*second half birth. Each of these parameters is included twice, each time interacted with an indicator for one or two children. The coefficient shown is the difference between the discontinuity estimates for one child and two children.

Control variables include state and cohort fixed effects as well as age, age squared, and black race indicator. Also included is (the sum of personal tax exemption, CTC, and single to head of household deduction difference*unmarried)*second half of the year birth.

Employment is defined as having monthly earnings of at least $250 in 2016 dollars.

Low-ed is defined as having a high school degree or less.

Standard errors are clustered at the state level.

*p<0.1, **p<0.05, ***p<0.01.

Table A4

Effects of the ETIC on labor market outcomes, weighted regressions for women’s first child

Functional form	All unmarried women	Low-ed unmarried women	All married women	Low-ed married women
Panel A: Employment

Linear	0.038	−0.001	0.006	−0.043
	(0.024)	(0.034)	(0.018)	(0.032)
Quadratic	0.037	−0.006	0.009	−0.033
	(0.029)	(0.035)	(0.020)	(0.035)

Panel B: IHS earnings

Linear	0.386*	0.050	0.092	−0.367
	(0.200)	(0.261)	(0.154)	(0.272)
Quadratic	0.375	0.009	0.103	−0.291
	(0.264)	(0.274)	(0.176)	(0.297)
N	33,043	19,215	86,875	23,130

Panel C: Log earnings

Linear	0.106	0.109	−0.051	−0.188***
	(0.090)	(0.106)	(0.060)	(0.069)
Quadratic	0.113	0.120	−0.077	−0.134*
	(0.093)	(0.110)	(0.056)	(0.077)
N	15,647	7,274	46,496	8,448
To 1 Disc.	N	N	N	N
To Max Cr. Disc.	Y	Y	Y	Y
Controls	Y	Y	Y	Y

[i] Data Source: Survey of Income and Program Participation, 1996 to 2014 Panels.

Each cell shows the parameter of interest from a separate regression. All regressions also include time (month of birth) and time*second-half of the year birth (July–December). Quadratic functional form regressions also include time squared and time squared*second half birth. Each of these parameters is included twice, each time interacted with an indicator for one or two children. The coefficient shown is the difference between the discontinuity estimates for one child and two children.

Control variables include state and cohort fixed effects as well as age, age squared, and black race indicator. Also included is (the sum of personal tax exemption, CTC, and single to head of household deduction difference*unmarried)*second half of the year birth.

Employment is defined as having monthly earnings of at least $250 in 2016 dollars.

Low-ed is defined as having a high school degree or less.

Standard errors are clustered at the state level.

Weighting is performed using provided sampling weights.

*p<0.1, **p<0.05, ***p<0.01.

Table A5

Effects of the ETIC on labor market outcomes, regressions for women’s first child with alternate employment definitions and trimmed earnings

Functional form	All unmarried women	Low-ed unmarried women	All married women	Low-ed married women
Panel A: Employment, $500 monthly earnings

Linear	0.028	0.012	−0.0001	−0.035
	(0.022)	(0.027)	(0.018)	(0.034)
Quadratic	0.030	0.015	0.001	−0.026
	(0.027)	(0.031)	(0.020)	(0.035)

Panel B: Monthly earnings

Linear	176.76**	53.20	78.36	−118.33
	(80.38)	(57.53)	(104.52)	(107.47)
Quadratic	183.67*	58.68	45.68	−75.81
	(92.26)	(60.56)	(109.20)	(115.01)
N	33,043	19,215	86,875	23,130

Panel C: Trimmed log earnings

Linear	0.109	0.139	−0.004	−0.124*
	(0.078)	(0.101)	(0.038)	(0.065)
Quadratic	0.106	0.137	−0.024	−0.083
	(0.080)	(0.102)	(0.034)	(0.069)
N	14,732	7,169	35,657	8,080
To 1 Disc.	N	N	N	N
To Max Cr. Disc.	Y	Y	Y	Y
Controls	Y	Y	Y	Y

[i] Data Source: Survey of Income and Program Participation, 1996 to 2014 Panels.

Each cell shows the parameter of interest from a separate regression. All regressions also include time (month of birth) and time*second-half of the year birth (July–December). Quadratic functional form regressions also include time squared and time squared*second half birth. Each of these parameters is included twice, each time interacted with an indicator for one or two children. The coefficient shown is the difference between the discontinuity estimates for one child and two children.

Control variables include state and cohort fixed effects as well as age, age squared, and black race indicator. Also included is (the sum of personal tax exemption, CTC, and single to head of household deduction difference*unmarried)*second half of the year birth.

Employment is defined as having monthly earnings of at least $500 in 2016 dollars in Panel A. In Panel C, the top 10% of earners from each sample are excluded.

Low-ed is defined as having a high school degree or less.

Standard errors are clustered at the state level.

*p<0.1, **p<0.05, ***p<0.01.

Table A6

Effects of the ETIC on labor market outcomes, regressions using phase-in rate for women’s first child

Functional form	All unmarried Women	Low-ed unmarried women	All married women	Low-ed married women
Panel A: Employment

Linear	0.026	0.016	0.008	−0.021
	(0.022)	(0.027)	(0.015)	(0.032)
Quadratic	0.023	0.014	0.009	−0.011
	(0.028)	(0.031)	(0.018)	(0.032)

Panel B: IHS earnings

Linear	0.301	0.207	0.096	−0.220
	(0.180)	(0.205)	(0.133)	(0.265)
Quadratic	0.269	0.179	0.089	−0.153
	(0.231)	(0.238)	(0.159)	(0.269)
N	33,043	19,215	86,875	23,130

Panel C: Log earnings

Linear	0.108*	0.145*	−0.011	−0.131*
	(0.059)	(0.076)	(0.036)	(0.071)
Quadratic	0.104	0.143*	−0.036	−0.100
	(0.064)	(0.077)	(0.032)	(0.078)
N	15,647	7,274	46,496	8,448
To 1 Disc.	N	N	N	N
to Phase-in Disc.	Y	Y	Y	Y
Controls	Y	Y	Y	Y

[i] Data Source: Survey of Income and Program Participation, 1996 to 2014 Panels.

Each cell shows the parameter of interest from a separate regression. All regressions also include time (month of birth) and time*second-half of the year birth (July–December). Quadratic functional form regressions also include time squared and time squared*second half birth. Each of these parameters is included twice, each time interacted with an indicator for one or two children. The coefficient shown is the difference between the discontinuity estimates for one child and two children.

Control variables include state and cohort fixed effects as well as age, age squared, and black race indicator. Also included is (the sum of personal tax exemption, CTC, and single to head of household deduction difference*unmarried)*second half of the year birth.

Employment is defined as having monthly earnings of at least $250 in 2016 dollars.

Low-ed is defined as having a high school degree or less.

Standard errors are clustered at the state level.

*p<0.1, **p<0.05, ***p<0.01.

Table A7

Effects of the ETIC on labor market outcomes, using only federal variation for women’s first child

Functional Form	All unmarried women	Low-ed unmarried women	All married women	Low-ed married women
Panel A: Employment

Linear	−0.026	−0.073	0.024	−0.013
	(0.043)	(0.049)	(0.021)	(0.045)
Quadratic	−0.020	−0.070	0.031	0.008
	(0.050)	(0.056)	(0.023)	(0.047)

Panel B: IHS earnings

Linear	−0.126	−0.570	0.271	−0.152
	(0.351)	(0.390)	(0.184)	(0.370)
Quadratic	−0.088	−0.535	0.317	0.011
	(0.403)	(0.438)	(0.203)	(0.388)
N	33,043	19,215	86,875	23,130

Panel C: Log earnings

Linear	0.208**	0.034	0.018	−0.184**
	(0.092)	(0.143)	(0.057)	(0.085)
Quadratic	0.209**	0.034	−0.019	−0.130
	(0.098)	(0.146)	(0.051)	(0.092)
N	15,647	7,274	46,496	8,448
To 1 Disc.	N	N	N	N
To Max Cr. Disc.	Y	Y	Y	Y
Controls	Y	Y	Y	Y

[i] Data Source: Survey of Income and Program Participation, 1996 to 2014 Panels.

Each cell shows the parameter of interest from a separate regression. All regressions also include time (month of birth) and time*second-half of the year birth (July–December). Quadratic functional form regressions also include time squared and time squared*second half birth. Each of these parameters is included twice, each time interacted with an indicator for one or two children. The coefficient shown is the difference between the discontinuity estimates for one child and two children.

Control variables include state and cohort fixed effects as well as age, age squared, and black race indicator. Also included is (the sum of personal tax exemption, CTC, and single to head of household deduction difference*unmarried)*second half of the year birth.

Employment is defined as having monthly earnings of at least $250 in 2016 dollars.

Low-ed is defined as having a high school degree or less.

Standard errors are clustered at the state level.

*p<0.1, **p<0.05, ***p<0.01.

Table A8

Effects of the ETIC on labor market outcomes, regressions for women with first children by race

Functional form	All unmarried women		Low-ed unmarried women		All married women		Low-ed married women
Race	Black	Non-black	Black	Non-black	Black	Non-black	Black	Non-black

Panel A: Employment

Linear	0.091	−0.007	0.010	−0.002	−0.055	0.012	0.011	0.0003
	(0.062)	(0.031)	(0.046)	(0.040)	(0.080)	(0.018)	(0.112)	(0.036)
Quadratic	0.081	−0.004	0.005	0.002	−0.041	0.014	−0.001	0.014
	(0.071)	(0.036)	(0.048)	(0.045)	(0.086)	(0.020)	(0.109)	(0.036)

Panel B: IHS earnings

Linear	0.800	0.067	0.149	0.075	−0.528	0.155	−0.032	−0.031
	(0.534)	(0.250)	(0.366)	(0.319)	(0.681)	(0.157)	(0.947)	(0.295)
Quadratic	0.734	0.076	0.119	0.101	−0.402	0.156	−0.153	0.073
	(0.610)	(0.289)	(0.392)	(0.355)	(0.740)	(0.175)	(0.912)	(0.301)
N	9,088	23,955	5,061	14,154	5,147	81,728	1,541	21,589

Panel C: Log earnings

Linear	0.089	0.108	0.178	0.080	−0.057	−0.006	0.214	−0.168**
	(0.112)	(0.100)	(0.210)	(0.104)	(0.116)	(0.044)	(0.379)	(0.078)
Quadratic	0.094	0.104	0.136	0.093	−0.046	−0.036	0.062	−0.128
	(0.118)	(0.103)	(0.200)	(0.104)	(0.118)	(0.038)	(0.384)	(0.082)
N	4,306	11,341	1,797	5,477	3,005	43,491	571	7,877
To 1 Disc.	N	N	N	N	N	N	N	N
To Max Cr. Disc.	Y	Y	Y	Y	Y	Y	Y	Y
Controls	Y	Y	Y	Y	Y	Y	Y	Y

[i] Data Source: Survey of Income and Program Participation, 1996 to 2014 Panels.

Each cell shows the parameter of interest from a separate regression. All regressions also include time (month of birth) and time*second-half of the year birth (July–December). Quadratic functional form regressions also include time squared and time squared*second half birth. Each of these parameters is included twice, each time interacted with an indicator for one or two children. The coefficient shown is the difference between the discontinuity estimates for one child and two children.

Control variables include state and cohort fixed effects as well as age, age squared, and black race indicator. Also included is (the sum of personal tax exemption, CTC, and single to head of household deduction difference*unmarried)*second half of the year birth.

Employment is defined as having monthly earnings of at least $250 in 2016 dollars.

Standard errors are clustered at the state level.

*p<0.1, **p<0.05, ***p<0.01.