An estimate of teenage abortion demand.
In 1990, forty percent of the one million pregnancies among U.S. teenagers (15-19 years of age) ended in abortion. An estimated 85% of teenage pregnancies are unintended (i.e., occur sooner than desired or are not desired at any time) and 51% of the unintended pregnancies ended in abortion. How adolescent women resolve their pregnancy has profound consequences for both the teenager and society. Studies show that teen mothers have lower educational attainment, lower occupational status, higher rates of poverty, and poorer marital prospects (Kalmuss, Namerow, and Cushman, 1991). Clearly, teenage pregnancy resolution is an issue of national concern.
A number of empirical studies have investigated abortion demand among pregnant teenage women. Leibowitz, Eisen, and Chow (1986) analyzed the pregnancy resolution decisions made by 297 unmarried pregnant California teenagers during the period 19721974. They found that teenage abortion demand was positively related to high school enrollment and labor force participation. Joyce (1988) examined the pregnancy resolution decisions of 31,207 New York City teenagers in 1984. The results showed that abortion attitudes and socioeconomic status increases the likelihood of abortion. Cooksey (1990) analyzed the pregnancy resolution decisions of 1,946 adolescent women over the time period 1979-1986 and found abortion demand was positively related to the education of the parents. King, Myers, and Byrne (1992) used data from 593 unmarried pregnant teenagers in 1984 and found the demand for abortion was positively related to income and high school enrollment and negatively related to religiosity. Lundberg and Plotnick (1995) found that state family planning, abortion, and welfare policies were significant factors in the pregnancy resolution decisions of 1,718 women aged 14-16 in 1979.
The previous studies have several features in common: (1) they all analyzed pre-1990 data due to the unavailability of more current data; (2) they all focused on the impact of personal and family background factors rather than economic cost and socioeconomic variables; and (3) they all used microdata which precluded them from analyzing the effect many important policy variables have on teenage abortion demand.
This study, using current state data, estimates the demand for abortion by teenagers (ages 15-19) for the year 1992. A current estimate of teenage abortion demand is important for several reasons. While most studies investigate the abortion demand of women of childbearing age (15-44), scant empirical research has focused on the socioeconomic determinants of teenage abortion demand. The teenage abortion rate has increased since 1974, and the consequences of any policies designed to reverse this trend will depend on current teenage abortion demand parameters. Furthermore, using aggregate data makes it possible to estimate the effects of these policy variables on teenage abortion demand.
2. Teenage abortion demand model
The model used in this paper stems from the theoretical work of Michael (1973), which argues that the pregnancy resolution decision by a woman is based on a comparison of the perceived costs and benefits expected to be imposed by that child. The perceived costs include the earnings, labor market, and social opportunities foregone from having the child and the benefits include the satisfaction from the child. When the perceived costs exceed the benefits, a woman will engage in fertility control.(1) Abortion is a fertility control method which reduces the probability of a birth to zero. The theoretical model suggests that the primary determinants of abortion demand are the costs of obtaining an abortion, tastes, and the opportunity costs of having a child.
The functional form of the teenage abortion demand equation to be estimated is:
Teenage Abortion [Rate.sub.i] = f (Price [Factors.sub.i], Opportunity Cost [Factors.sub.i], [Tastes.sub.i]) (1)
The dependent variable, the teenage abortion rate, is the number of abortions per 1000 pregnancies of teenage women (15-19 years old) in state i during the year 1992.(2)
The independent variables fall into three categories:
2.1. Price factors
Prior research suggests that abortion obeys the fundamental law of demand - less being purchased at a higher price than at a lower price (Medoff, 1988). The price of an abortion is the average cost of an abortion performed in a non-hospital facility during the first ten weeks in each state in 1992.(3)
Medicaid is a federal program that provides medical care to low-income people. In 1977 Congress passed legislation that prohibited federal Medicaid funding for poor women. States, however, were free to continue funding Medicaid abortions. Previous research suggests that state Medicaid abortion funding increases the demand for abortion, since cost is not a consideration in the use of abortion services (Currie, Nixon, and Cole, 1996). The variable Medicaid is a dummy variable equal to one for those states which continued abortion funding.(4)
2.2. Opportunity cost factors
Leibowitz, Eisen, and Chow (1986) argue that teenagers have a high time-preference for the present and poor information about the future. Teenagers, because of their lack of knowledge and experience, have inadequate estimates of their own opportunity costs, and are heavily influenced by what other women have achieved in the labor market. Teenagers estimate their own opportunity costs by using the opportunity cost set of older cohorts. Accordingly, we use as measures of the opportunity costs of a teenage birth the income, labor market, and marital opportunities of older women.(5)
One measure of a pregnant teenager's opportunity cost is the present value of her net income. Unfortunately, such information is not available. However, previous research established that the average income of women who work year-round full-time is a good proxy for both potential future income foregone and the value of time (Mincer, 1963; Cain, 1966).(6) It is hypothesized that the greater the 1992 full-time income of women in state i, the greater the opportunity cost of a child, and the greater the teenage demand for abortion.
A second opportunity cost factor is the labor force participation rate of women. Female labor force participation measures a teenager's perception of future labor market and economic opportunities. The greater the participation of women in the labor force, the higher the estimates of the opportunity cost of childbirth, and the more likely teenagers will terminate a pregnancy.
Another opportunity cost measure is marital status. The percentage of women in state i who are single measures a teenager's perception of marital market conditions (i.e., the likelihood of remaining single). Single women have a greater opportunity cost of childbearing and childrearing than married women. Also, becoming an unwed single teenage mother will affect both an adolescent's long-run earning potential as well as the opportunity cost of future marital possibilities with its concomitant joint income maximization prospect. It is expected that the greater the percentage of single women in state i the greater the abortion demand.(7)
Religion is a powerful moral and social factor in an individual's "taste for abortion." Most fertility control studies include a variable representing Catholic affiliation because of the Catholic Church's official position that abortion is a mortal sin. The empirical evidence about the impact of Catholicism has been contrary to the predicted negative effect. Deyak and Smith (1976) and Gohmann and Ohsfeldt (1993) found that Catholic affiliation had a positive impact on abortion demand, while Medoff (1988) and Brown and Jewell (1996) found it to be an insignificant determinant of abortion demand. One reason for these contradictory findings may be that many rank-and-file U.S. Catholics feel that the Church's moral teaching does not apply to them (Sander, 1992).
One religious group that is homogenous in their values and beliefs about sexual morality is Fundamentalist Christians. Fundamentalist Christians have a greater level and intensity of religious participation and are more likely to believe in moral absolutes and follow a stricter moral code than other religious denominations (Ammerman, 1987). The Fundamentalism variable is the percentage of each state's population that claims membership in a religious denomination that professes a literal belief in the Bible and its hypothesized effect is negative.(8)
3. Empirical results
The teenage abortion demand model is estimated on 1992 state data using two-stage least-squares (a Hausman test rejected the exogeneity of abortion price in equation (1)). The excluded exogenous variables used as instruments for the endogenous abortion price were: (1) the number of physicians per 100,000 population; (2) the number of abortion clinics; (3) the number of nurses per 100,000 population; and (4) the average weekly wage of employees in the offices of physicians.
The empirical estimates of the teenage abortion demand model appear in Table 1, Column 1.(9) These results provide strong support for the fertility control model. The abortion price is significantly negative - abortion is an economic good that obeys the fundamental law of demand. The degree of responsiveness, the price elasticity of demand, evaluated at the sample mean, was equal to .78. This figure is virtually identical with those reported in previous studies of the abortion demand of women of childbearing age (Medoff, 1988). The full-time income variable is significantly positive; increases in income have a positive effect on teenage abortion demand. The income elasticity of demand of .96 is consistent with prior results (Gohmann and Ohsfeldt, 1993). The Medicaid variable is significantly positive; teenage women have 54 more abortions per 1,000 pregnancies in states that continue to fund Medicaid abortions. As hypothesized, the labor force participation variable is significantly positive, suggesting that the higher value of time from being in the labor force increases teenage abortion demand. Christian Fundamentalism as theorized is found to significantly decrease the demand for teenage abortion.
The only seemingly contradictory finding is that the unmarried status variable was positive, but not statistically significantly different from zero. This result suggests that the likelihood of being a single unwed mother is not a significant determinant in a teenage woman's decision to have an abortion.(10)
4. Policy variables and teenage abortion
During the 1980s many states, in response to high teenage pregnancy rates, enacted parental involvement abortion laws. Parental involvement laws require minors (less than 18 years of age) to notify or get the consent of one or both parents in order to obtain an abortion. [TABULAR DATA FOR TABLE 1 OMITTED] Previous research on this issue tended to analyze the teen abortion rate within a state before and after a parental law became effective (Cartoof and Kleiman, 1986; Rogers et. al., 1991). The impact of parental involvement laws is of considerable policy interest.
The effect of parental involvement laws depends on a variety of factors. (1) The laws vary widely in their restrictiveness. Some states require notification by physicians before they can perform teenage abortions, some states allow grandparents, clergy or other responsible adults to give their permission. (2) The intensity in state enforcement of the laws varies considerably. For example, many states permit minors to bypass parental involvement by appearing before a judge. Some judges routinely approve the procedure and others rarely approve it. (3) The laws may be merely symbolic if most teenagers routinely involve a parent. Henshaw and Kost (1992) found that 61% of unmarried minors said one or both of their parents knew about the abortion.
To test the efficacy of state parental involvement laws, a dummy variable equal to one if the state enforced the law in 1992 is added to equation (1). The empirical results appear in Table 1, column 2. The coefficient of the parental involvement law was negative, but not statistically significantly different from zero. Parental involvement laws are found to have very little impact on teenage abortion demand.
Whether the availability of welfare provides a financial incentive for teenagers to become pregnant is a much disputed issue. A related empirical question is whether Aid to Families with Dependent Children (AFDC) benefits affects the abortion decision of pregnant teenagers. AFDC benefit levels are set by each state and vary from state to state. To determine the role of the welfare system on teenage abortion demand, the average monthly state AFDC payment for a recipient and one child is included in equation (1). The empirical results appear in Table 1, column 3. The AFDC variable is negative, but not statistically significantly different from zero, suggesting that decreases in AFDC benefits to unwed mothers will, in the short-run, not increase the demand for teenage abortion.
The impact of unemployment on teenage abortion is an issue that has been overlooked in the abortion literature. The unemployment rate is another opportunity cost factor being a measure of the perceived likelihood of being able to find a job. Economic theory suggests that if a woman's labor supply is procyclical, the higher the unemployment rate, the lower the value of a woman's time, and the lower the demand for abortion.
The impact of unemployment on teenage abortion is empirically tested by replacing the labor force participation rate variable in equation (1) by the unemployment rate of women, sixteen years or older. The empirical results in Table 1, column 4 show that teenage abortion demand is coincident with the business cycle: increasing during expansions when opportunity costs are higher and decreasing during recessions when opportunity costs are lower.
Michael (1973) argued that education may operate through several independent channels of influence. Educational attainment may increase the demand for abortion through three channels. (1) School enrollment disrupted by a birth has long-run earnings implications. (2) The well-established positive relationship between education and lifetime earnings suggest greater education is associated with a higher value of time. (3) More education may be associated with greater preferences for physical goods relative to children. In order to ascertain the impact of educational attainment on teenage abortion the percentage of women, twenty-five years or older, who have completed at least twelve years of schooling, a measure of the perceived probability of finishing high school, is added to equation (1). The empirical results appear in Table 1, column 5. The educational attainment variable is negative, but not statistically significantly different from zero. Educational attainment is found not to have an independent influence on the teenage abortion decision.
Very little research has been done on the effect of abortion attitudes on abortion demand. One reason is that it is usually very difficult to directly measure abortion preferences. On November 3, 1992, the General Election Exit Poll randomly surveyed 15,490 voters. Voters in each state were questioned about their abortion attitudes. An abortion attitude index was formed by arbitrarily assigning the value 4 to the attitude "abortion should be legal in all cases"; the value 3 to the attitude "abortion should be legal in most cases;" the value 2 to the attitude "abortion should be illegal in most cases;" and the value 1 to the attitude "abortion should be illegal in all cases."
A State's Abortion Attitude Index was computed by weighting each abortion attitude value by the proportion of the state's voters with that value and adding up all the weighted values. A State's Abortion Attitude Index score can range from the value 1 (a state in which the populace believes abortion should be illegal in all cases) to the value 4 (a state populace which believes abortion should be legal in all cases). The State Abortion Attitude Index variable was added to equation (1) and the empirical results appear in Table 1, column 6. The Abortion Attitude Index variable is positive, but not statistically significantly different from zero, suggesting that a state's abortion attitudes have no impact on teenage abortion demand.
5. Pooled time-series estimates of teenage abortion
This section estimates the teenage abortion demand model using state level data pooled over the two years 1992 and 1980. There are several methodological advantages in estimating a pooled time-series model. First, pooling data increases the sample size yielding more consistent parameter estimates. Secondly, pooled data yields more efficient parameter estimates than a singe year cross-section model. Finally, many variables exhibit more variation over time than during a particular survey year.
Equation (1) was reestimated using two-stage least-squares for the 50 states over the years 1992 and 1980. Added to equation (1) is a time trend variable set equal to one in 1992 and zero in 1980 in order to control for year-specific fixed effects. Table 2, column 1 presents the pooled time-series estimates of the teenage abortion demand model. The estimated coefficients are both quantitatively and qualitatively consistent with the cross-section estimates previously reported. Abortion price is significantly negative and the price-elasticity of demand is .75. Full-time income is significantly positive and the income elasticity of demand equals .82. Teenage women have 46 more abortions per 1,000 pregnancies in Medicaid states. Labor force participation is significantly positive while Christian Fundamentalism is significantly negative. Consistent with the cross-section results, marital status is not a significant determinant in a teenager's abortion decision. The time trend variable indicates that, ceteris paribus, there has been a secular decline in teenage abortion demand over the period 1980-1992.
Columns 2-6 of Table 2 report the empirical results when the previously analyzed policy variables are included in the pooled time-series teenage abortion demand model. The parental involvement, unemployment, and education variables are consistent with the previous cross-section results. The parental involvement law (Table 2, column 2) was negative, but not statistically significantly different from zero. Parental involvement laws have no long-run impact on teenage abortion demand. The unemployment rate (column 4) is significantly negative - teenage abortion decreases during recessions and increases during expansions. The education variable (column 5) is positive, but not statistically significantly different from zero. Education is not a significant determinant of teenage abortion demand.
The AFDC benefits variable (column 3) in the pooled time-series model is positive and statistically significantly different from zero. In the long-run decreases in AFDC benefits decrease teenage abortion demand. This result is consistent with the work of Akerloff, Yellen, and Katz (1996) who argue that decreases in the financial incentives for having a child may result in a long-run modification in teenage sexual behavior.
The state Abortion Attitude Index (column 6) is found to be significantly positive. This suggests that while short-run abortion attitudes have no impact on teenage abortion demand, in the long-run, teenage abortion demand is positively related to a state's preference for abortion.
[TABULAR DATA FOR TABLE 2 OMITTED]
This study estimated the demand for teenage abortion in the United States using state data for the year 1992. The results showed that teenage demand for abortion is price inelastic and increases with income. Teenage abortion demand is also found to be positively related to state Medicaid funding, women's labor force participation and negatively related to Christian Fundamentalism.
The teenage abortion model was also used to examine the impact of various policy variables. Parental involvement laws, AFDC benefit levels, and educational attainment, and state abortion attitudes are found to have, in the short-run, no impact on teenage abortion demand. The empirical results also showed that teenage abortion demand is coincident with the business cycle.
This study also estimated the teenage abortion demand model using state data pooled over the years 1992 and 1980. The pooled time-series estimates of the teenage abortion demand model were both quantitatively and qualitatively consistent with the cross-section estimates. One of the advantages of using pooled time-series data is that it allows examination of the impact of various policy in the long-run. Consistent with the short-run results, in the long-run, parental involvement laws and educational attainment have no impact on teenage abortion demand. Teenage abortion demand was also found to be coincident with the business cycle. However, the long-run results also showed that teenage abortion demand is positively related to a state's abortion attitudes and to the level of AFDC benefits.
1. The study of teenage abortion demand does have one potential problem. The theory of fertility control implicitly assumes that decision-makers are rational. This only means that teenage women are well-informed about the consequences of their decisions and are more than likely to be correct about the consequences of their decisions (Duncan and Hoffman 1990). This holds whether the pregnant teenager makes her own decision, has the decision forced on her, or made the decision in consultation with others.
2. Miscarriages and stillbirths are excluded since they do not involve a pregnancy resolution decision.
3. Over 90 percent of all abortions are done in the first trimester and this percentage has been stable since 1980. In 1993, 93 percent of all abortions were performed in a non-hospital facility.
4. Teenage abortion rates, abortion prices, and Medicaid states were obtained from the Alan Guttmacher Institute, which is a national research foundation affiliated with Planned Parenthood. The Guttmacher Institute is widely recognized as having the most comprehensive available information about abortions in the U.S. and by state.
5. There is no bias in using the characteristics of women rather than pregnant teenagers since the opportunity cost for a teenager is the perceived foregone opportunities of women in the labor market.
6. The opportunity cost of an additional child includes the income foregone from leaving the labor market and is directly related to the time involved in childrearing. An increase in a woman's time value raises the price of children, who are relatively time-intensive, inducing a substitution towards fertility control methods such as abortion.
7. All the economic variables in the abortion model are exogenous at the time of the abortion since abortion is an ex post event.
8. The data on all economic variables were obtained from the U.S. Bureau of the Census, State Reports (1993). The data on religious affiliation was from the National Council of Churches survey, Churches and Church Membership in the United States (Quinn 1993). The mean and standard deviation (in parentheses) for the variables in equation (1) are Price: 305.14 (48.76); Full-Time Income: 18416 (2703.18); Medicaid: .26 (.43); Labor Force Participation: 74.68 (3.96); Single: 34.81 (4.19); Fundamentalism: 14.55 (13.8).
9. Equation (1) was also estimated with the number of abortions per thousand live births and the number of abortions per thousand women 15-19 years. The empirical results were virtually identical to these reported in the text.
10. It has been suggested that black teenage women have different opportunity costs or marital market opportunities than white teenagers. In order to determine whether race had an impact on teenage abortion demand, equation (1) was reestimated with the percentage of teenagers who are black included. The empirical results showed that the race variable was not statistically significantly different from zero and the coefficients of the other variables in the model were virtually identical to those previously reported.
Akerlof, G., Yellen, J., & Katz, M. (1996). An analysis of out-of-wedlock childbearing in the United States. Quarterly Journal of Economics, 111 (2):277-318.
Ammerman, Nancy T. (1987). Bible Believers. New Jersey: Rutgers University Press.
Brown, Robert W., & Jewell, R. Todd. (1996). The impact of provider availability on abortion demand. Contemporary Economic Policy, 14(2):95-106.
Cain, Glen C. (1996). Married Women in the Labor Force: An Economic Analysis. Chicago: University of Chicago Press.
Cartoof, V., & Klerman, L. (1986). Parental consent for abortion: Impact of the Massachusetts law. American Journal of Public Health, 76(4):397-400.
Cooksey, Elizabeth C. (1990). Factors in the resolution of adolescent premarital pregnancies. Demography, 27(2):207-218.
Currie, J., Nixon, L., & Cole, N. (1996). Restrictions on medicaid funding of abortion. Journal of Human Resources, 31(1):159-188
Deyak, Timothy A., & Smith, V. Kerry. (1976). The economic value of statute reform: The case of liberalized abortion. Journal of Political Economy, 84(1):83-100.
Duncan, G. J., & Hoffman, S. D. (1990). Welfare benefits, economic opportunities, and out-of-wedlock births among black teenage girls. Demography, 27(4):519-535.
Gohmann, S. F., & Ohsfedt, R. L. (1993). Effects of price and availability on abortion demand. Contemporary Policy Issues, 11(4):42-54.
Henshaw, S. K., & Kost, K. (1992). Parental involvement in minor's abortion decisions. Family Planning Perspectives, 24(5):196-213.
Joyce, Theodore. (1998). The social and economic correlates of pregnancy resolution among adolescents in New York City, by race and ethnicity: A multivariate analysis. American Journal of Public Health, 78(6):626-631.
Klamuss, D., Namerow, P. B., & Cushman, L. F. (1991). Adoption versus parenting among young pregnant women. Family Planning Perspectives, 23(1): 17-23.
King, R. H., Myers, S. C., & Byrne, D. M. (1992). The demand for abortion by unmarried teenagers. American Journal of Economics and Sociology, 51 (2):223-235.
Lundberg, S., & Plotnick, R. D. (1995). Adolescent premarital childbearing: Do economic incentives matter? Journal of Labor Economics, 13(2): 177-200.
Medoff, Marshall H. (1988). An economic analysis of the demand for abortions. Economic Inquiry, 26(2):353-359.
Michael, Robert T. (1973). Education and the derived demand for children. Journal of Political Economy, 81 (2):5128-5164.
Mincer, Jacob. (1963). Market prices, opportunity costs, and income effects. In Carl F. Christ, ed., Measurement in Economics: Studies in Mathematical Economics and Econometrics (pp. 67-82). Stanford: Stanford University Press.
Quinn, B., Bradley, M., Goettinghall, P., & Shriver, P. (1993). Churches and church Membership in the United States. Atlanta: Glenmary Research Center.
Rogers, J. L., Boruch, R. F., Stoms, G. B., & Demoya, D. (1991). Impact of Minnesota parental notification law on abortions and births. American Journal of Public Health, 81(3):294-298.
Sander, William. (1992). Catholicism and the economics of fertility. Population Studies, 46(3):477-489.
Singh, Susheela. (1986). Adolescent pregnancy in the United States: An interstate analysis. Family Planning Perspectives, 18(5):210-220.
U.S. Bureau of the Census. (1993). State Reports, Social and Economic Characteristics, 1990. Washington D.C.: U.S. Government Printing Office.
|Printer friendly Cite/link Email Feedback|
|Author:||Medoff, Marshall H.|
|Publication:||The Journal of Socio-Economics|
|Date:||Mar 1, 1999|
|Previous Article:||Two kinds of order: thoughts on the theory of the firm.|
|Next Article:||A visibility analysis of the scientific production of Ukrainian economists: 1969-94.|