The Relationship between Early Pregnancy and Wages

Abstract

The purpose of this research is to investigate the existence of a possible relationship between early pregnancy and wages. Findings within my research may provide policymakers with critical information required to make decisions that may avert premature pregnancy. Furthermore, I hope the findings of my investigation can help motivate policymakers to focus their efforts on groups that are harmed more due to early pregnancy. The regression analyzes cross-sectional data from 2017 which includes all fifty states. Within the study, I control for factors that include insurance, race, region, marital status and education. I also control for the bottom 1% and top 1% to restrict my findings from being biased by the extremes of the population. I found that there is no statistically significant increase in the log of wages from an increase in the mothers age at birth.

Introduction

Early pregnancy is both an obstinate social issue and a difficult matter to address because it is constantly perpetuated within the same households. Furthermore, early pregnancy represents a huge drain on society. According to the study conducted by Moore and Wertheimer (1984), to measure the impact of hypothetical intervention into prevention of teenage pregnancy, if the fertility of all teenagers is reduced by 50%, the public sector costs for Aid to Families with Dependent Children, Medicaid, and Food Stamps for families of women 20-29 would be reduced by $1.4 billion.

Another area of concern for early pregnancy is the loss in educational attainment of early parents. The research conclusions by Card and Wise (1978) proved that young mothers were less likely to be in school and more likely to have greater educational setbacks than their peers that delayed childbirth. This claim supports the summary statistics that approximately 66% of people within the dataset do not have a bachelor’s degree or more. This detrimental loss to education has greater repercussions in women’s ability to attain higher paying and receive more prestigious jobs.

The next area of concern is the loss to wages that results from early pregnancy. The loss to wages is highly correlated to the reduction in educational attainment present in adolescent mothers. According to the conclusions made by Card and Wise, “Adolescent mothers have less prestigious jobs, have lower incomes, and are less satisfied with their jobs…” This claim supports the findings of Moore and Wertheimer (1979) that found for year a woman delays the birth of their first child, their income increases by $200. The findings in my data are contrary to these claims. I found for each additional year a mother is older, the log of her wages and salary is decreased approximately 0.94%, holding all else constant. However, this coefficient was deemed insignificant within the regression.

A final area of interest is the racial discrepancy present between a mother’s wages and the age she has her first child. According to the two studies of Moore and Wertheimer (1979, 1984), the age of the mother at first birth does not seem to be as statistically significant for black women as it does for white women. Conversely, black women have more work experience and greater earnings by the time they are 27 than white women. This claim conflicts with my regression findings because the p-value of my race variable for blacks is 0.003, a good indicator that the variable is significant. According to my analysis, the claim cannot be substantiated that the age of the mother does not affect blacks the same way it does whites.

This research aims to analyze the effects of early motherhood on the log of wage and salary income by examining cross-sectional data from 2017. In conjunction with the literature, my investigation aims to support the claim that early pregnancy is harmful to wages, controlling for variables like race, insurance, region, marital status, and educational attainment. A point of interest is the difference in wages and salary of mother’s, varying by region. In my literature review I did not find much analysis broken down by region. While I understand variations occur region to region, and it is impossible to control for all exogenous differences, it may be an initial step in further research comparing the regional effects on mother’s wages and salary. My goal of any regional discrepancy findings would be possible reform in poorly performing regions to better address the issue of early pregnancy, modeled on the approaches of better performing regions. The paper is structured as follows: model specifications, results, and conclusion. The model specification will describe the data and methodology, a summary of the data and variables, and the estimation technique used. The results section will address the bias found within the data and the findings of the research. Finally, the conclusion will reiterate the findings of the investigation, discuss the limitations of the data and suggest the policy implications of the study.

Data and Methodology

The cross-sectional data collected for this analysis was provided by the Integrated Public Use Microdata Series (IPUMS) covering 2017. The data is aggregated to the national level excluding only those considered in the bottom 1% or top 1% of the wage and salary earners to avoid bias from the extreme tails of the population. The dependent variable of interest for this regression is the log of total pre-tax wages and salary, denoted by log_wage. As illustrated in Table 1, the mean, minimum, maximum and standard deviation of the variable is 10.54761, 1.386294, 13.50899, and 1.085094, respectively. The independent variable of interest is the age of the mother and the age of the mother squared.

Results

The main results of the regression can be found at the end of the report labeled Table_2. In column 1, I present the naïve regression of the log of wages as a function of the mother’s age and mother’s age squared. In column 2, I control for the possibility of healthcare coverage by including the variable insured. Theoretically, having health insurance should positively bias the regression because having insurance should be positively correlated with age and the log of wages. However, healthcare decreases the constant in the regression equation. In column 3, I additionally control for marital status, by including the marriage variable. I theorize this should negatively bias the equation. Expert analysis by Jeremy Staff and Jeylan T. Mortimer (2012), found marriage has a negative correlation of a 5% decrease in wages as compared to non-married women. While the change in the constant is small, it is not negative, conflicting with the evidence presented. In column 6, I control for the educational dummy variables, some_college and bachelors_or_more. While the magnitude of the bias may be different, both variables should play a role in having a positive bias on the regression. This bias is seen in a significant increase in the constant term.

My initial hypothesis was that mother’s age would have a positive relationship to the log of wages. Furthermore, I thought that the mother’s age squared would have a negative relationship to wages. The naïve regression fit my hypothesis of relationship between mother’s age and the log of wage. The t-score was t=4.14, both statistically significant and having the same sign as the null hypothesis. Also, mother’s age squared was negative, as hypothesized, and statistically significant with a t=-4.04. However, as I began to control for additional variables the statistical significance of the age of the mother began to fall. After controlling for insurance and marital status, the coefficient of mother’s age became statistically insignificant, t = -1.21. Furthermore, the variable for mother’s age squared became significant only at the 5% level, with a t = -2.36. When I controlled for the seasonal regional dummies, race, and the education dummies, the signs of the coefficient were opposite to what I had hypothesized and statistically insignificant. These results show that mother’s age may be a proxy for other variables that are more correlated with wage like race, region, and educational attainment.

Conclusion

Before I began the investigation, I hypothesized that the increase in mother’s age would have a positive correlation on the wages of the individual. Furthermore, I theorized that there would be a point where a mother’s age would begin to have negative effects on wage. The naïve regression confirms the assumption of a positive relationship between mother’s age and the log of wages and salary. The naïve regression also reveals that approximately 29 years above the mean of mother’s age, the positive effects of mother’s age become negative. However, once I controlled for other variables including: insurance, marital status, race, region, and education, the variable for mother’s age was proven to be statistically insignificant and had the opposite sign to what I hypothesized. According to the controlled equation, a one-year increase in the age of the mother, decreases the log of wages and salary by 0.94%, holding all else constant. Furthermore, after approximately 21 years above the mean, the negative effect on wages become positive. With the insignificant t-test and conflicting sign direction, I failed to reject the null hypothesis that mother’s age would negatively affect wages.
When running F-Tests of joint significance on race and region, the conclusion of the tests confirmed the variables were jointly significant at the 5% level. The results of these tests confirm that the variables jointly explain the variation in the log of wages. Another important finding was the significance of the schooling variables. The t-statistics of both these variables are astronomically high and show a strong correlation in the variation of the log of wages.

The limitations of the data were based on an inability to effectively distinguish between the age of the mother and the age of the child. This limited my results because I had trouble generating a variable that accurately described the age of the mother when her first child was born.

The implications of my analysis show there is not a statistically significant relationship between early pregnancy and wages, other factors like region, race, and education are more strongly correlated with wage concerns for young mothers. When only controlling for mother’s age the data suggests that early motherhood has a significant impact on the wages of mothers. However, other variables explaining this variation than mother’s age alone. I urge policymakers to address the root of the problem of low wages for mothers which include low educational attainment and racial discrepancies present when compared to the wages of whites. I would also urge policy makers to look at the New England region for possible solutions to the problem. When comparing regional variables to New England, every one of them was negative, suggesting that the New England region may have found a solution, or at least a better way to deal with the problem at hand.

Bibliography

1. Hofferth, S. L., Hayes, C. D., & National Research Council. (1987). Social and economic consequences of teenage childbearing. In Risking the Future: Adolescent Sexuality, Pregnancy, and Childbearing, Volume II: Working Papers and Statistical Appendices. National Academies Press (US).
2. Card, J., & Wise, L. (1978). Teenage Mothers and Teenage Fathers: The Impact of Early Childbearing On the Parents' Personal and Professional Lives. Family Planning Perspectives,10(4), 199-205. doi:10.2307/2134267
3. Staff, J., & Mortimer, J. T. (2012). Explaining the motherhood wage penalty during the early occupational career. Demography, 49(1), 1–21. doi:10.1007/s13524-011-0068-6

Did you like this example?

724

38