effect of abortion on illegitimacy. Women in abortion states in the West and Northeast clearly used legal abortion more frequently than women in abortion states in the South and North Central regions. In the West and Northeast, six of the eight abortion states had rates of 16 or more legal abortions per 1000 women aged 15 to 44 years; in four states-New York, California, Hawaii, and Washington-the rates were 20 or more. In contrast, women living in abortion states in the South and North Central regions were obtaining legal abortions at much lower rates, ranging from 2.8 in Arkansas to 13.7 in Delaware. Further, it appeared that in those southern states where abortion was legal, many women must have found it difficult or undesirable to obtain an abortion locally. The data show that substantial numbers of legal abortions performed on women living in these states had in fact been obtained elsewhere. For example, in Virginia the overall rate of legal abortions for residents was 6.8 per 1000 women aged 15 to 44, but nearly half (3.1 per 1000) had been obtained in another state. The role of out-of-state abortion is even more striking in South Carolina, where almost two-thirds of all legal

Table 3. Estimated legal abortion rates for 1971; estimated illegitimate birth rates for 1965, 1970, and 1971; and average annual percent change in the illegitimate birth rate 1965 to 1970 and 1970 to 1971 in abortion states grouped by geographical region. Legal abortions and illegitimate births are related to women aged 15 to 44 years; ISR, in state of residence; OSR, outside state of residence. [Data from table 4 in (1); see also (4)]

Fig. 2. Percent change in illegitimate birth rates in abortion and nonabortion states in the United States, 1970 and 1971 Age groups, in years, are shown at be right.

were larger in the regions with states of high abortion use than in those with states of low abortion use. Indeed, rain of illegitimate fertility in most of the abortion states in the West and North east fell anywhere from 15 to 19 percent, whereas in the North Central and South, the individual abortion state, with the exception of Delaware, showed relatively modest declines or small rises

Among the obstacles that could have depressed abortion use in the South and North Central abortion states were a possible lesser interest on the part of the medical profession and the public in making abortion services widely avail able, particularly in regard to providing public financial assistance. Although data are not available concerning the extent of public payment for abortion services in most of the abortion states, we do know that 42 percent of all res dent abortions performed in 1972 in New York City and 37 percent a California were publicly financed (19. In California the proportion publicy financed was 60 percent for black women and 32 percent for white women. It is unlikely that the extent of public assistance for abortion was anywhere nearly this high in abortion states of the South and North Central regions.

Since the highest legal abortion rates and the greatest illegitimacy declines occurred in the West and Northeast regions, it could be argued that regional differences rather than differences in the availability of legal abortion are responsible for the decline in illegitimacy This is not the case, however, for our data show that in each region of the country, including the South and North Central, women living in nonabortion states showed less decline or more rise in illegitimacy than women living in abortion states.

If our conclusion is correct, legalized abortion has had a substantial effect co illegitimate fertility in the United States. Can the magnitude of this effect be measured?

• See (26). † See (27). * The rates shown differ slightly from those previously published (2) because in the study described herein we used somewhat different methods for estimating the numbers of unmarried women. The rate for 1965 was assumed to be equal to the rate in 1966 since California birth data by legitimacy are not available prior to 1966. § Abortion rates shown for New Mexico are based on revised figures for resident and nonresident women obtained from the Center for Disease Control.

Illegitimate Births Prevented

In 1971 there were an estimated 416,000 illegitimate births in the United States and an estimated 272,000 legal

abortions performed for unmarried women (17). Thus about 40 percent of the recorded pregnancies (births plus legal abortions) were terminated through legal abortion. However, it cannot be assumed that every pregnancy terminated by abortion would have resulted in an illegitimate birth if abortion had not been legalized. Some would undoubtedly have ended as illegal abortions, others would have been legitimized by a "forced" marriage, and still others would have ended as fetal deaths. An important question, therefore, is, What proportion of the pregnancies currently being terminated by legal abortion would have resulted in an illegitimate live birth or one of the other three possibilities? Although there is no certainty as to what would have happened without legal abortion, we have estimated the number of illegitimate births that would have occurred based on what we believe are reasonable and conservative assumptions. To make these estimates, we have compared the actual number of out-of-wedlock births with the number that would have occurred if the previous upward trend had continued. Projecting the trend during 1965 to 1970 in age-race-specific illegitimate birth rates to the population of unmarried women in 1971 (18), we arrive at an expected 455,000 illegitimate births in 1971. However, the actual number fell 39,000 short of this estimate. Presumably, then, the difference between the expected and actual number of illegitimate births represents approximately the number of those births averted by legal abortion. The abortions that prevented these births comprised about 14 percent (19) of the total legal abortions performed on unmarried women in the period affecting 1971 births. Of the remaining 86 percent of the abortions, we estimated that most would have ended as illegal abortions or as legitimate births (20, 21).

27,000 fewer than actually occurred. Most of this additional saving would have come from nonwhite women who apparently used out-of-state abortion less than did white women. Of the 27,000 additional illegitimate births that could have been prevented in 1971, 18,000 would have occurred to nonwhite women and 9,000 to white

women.

Legitimate Fertility

Since most births in the country are legitimate, the general trend in overall fertility is influenced mostly by the reproductive trends among married women. Legitimate birth rates reached their highest postwar peak in 1957, and then began to fall. In the period considered here, 1965 to 1971, legitimate birth rates continued to decline through 1968, rose slightly in 1969 and 1970, and then, like illegitimate birth rates, dropped sharply in 1971 (Table 4), reaching their lowest level since the downturn started in 1957. As Table 4 shows, there were decreases for every age group, with the rate for women aged 35 to 44 years decreasing the most. Teen-agers who were the only women with rising legitimate fertility between 1965 and 1970, had the least decline between 1970 and 1971 (2 percent).

Both white and nonwhite women participated virtually evenly in the 1971 drop. While white and nonwhite women of every age group contributed to

the decline, the greatest reduction for both groups occurred in the oldest age group, 35 to 44 years.

Although in 1971 legitimate and illegitimate fertility declined equally (6 percent), it was not readily apparent that legal abortion had played a major role in the decline in legitimate births. The ratio of abortions to births for married women is so low that it alone could not explain the 6 percent drop in the legitimate birth rate. Consequently, a separate examination of the trend in marital fertility for abortion and nonabortion states is a more feasible method for measuring the impact of legalized abortion on legitimate birth

rates.

Figure 1 shows that up to 1970 the two groups of states showed similar trends in legitimate birth rates (average decline of 1.4 percent per year in abortion and nonabortion states) but that they diverged in 1971. Although both groups showed a substantially greater drop in birth rates than for the previous 5 years, the drop was somewhat larger in abortion than nonabortion states (8 percent as opposed to 5 percent; see Table 5). This differential is much smaller than that for illegitimate birth rates in 1971 (a decline of 12 percent in abortion compared to 2 percent in nonabortion states). Thus, although legal abortion evidently had some impact on the decline in marital fertility between 1970 and 1971, it was considerably less than on nonmarital fertil. ity.

7.6

Age (years)

Average annual percent change

15-441

15-19

136.6 131.0 424.1 433.5

20-24

293.3 283.3

25-34

147.3 137.2

35-44

- 5.8

0.5

- 1.8

-2.1

-2.2

22.3

19.9

-7.6

-6.6 --10.5

15-441

15-19

131.3 126.3 407.1 416.7

121.9

392.2

116.9

-1.1

- 5.9

The relatively lesser influence of legal abortion upon marital fertility for both white and nonwhite women, as seen in Table 5 and Fig. 3, characterized virtually all age groups. The only exception is married white teenagers. Their fertility necessarily reflects the influence of pregnancy before marriage (22); they were the only group of married women to experience rising birth rates in the previous 5 years. For this group, marital fertility in 1971 continued to rise in nonabortion but declined in abortion states. This indicates that legal abortion had a depressive influence not only on white teen-age marital fertility but on the incidence of pregnancy-related marriages as well.

Differences between abortion and nonabortion states in the decline of marital fertility permit us to estimate the number of legitimate births prevented by legal abortion (23). While legal abortions equaled in number about 40 percent of the recorded pregnancies among unmarried women, this was not true for married women. In 1971 there were 3,137,000 legitimate births in the United States, but only 134,000 legal abortions for married women. Thus, legal abortions equaled about 4 percent of the recorded pregnancies among married women. Not all these aborted pregnancies would necessarily have ended as live births if abortion had not been legal: some would have ended in legitimate births, others in illegal abortions, and still others in fetal deaths. An estimate of the number of legitimate births

of the 134,000 legal abortions to mat ried women would have ended as live births, most of the remaining four-fifths would have ended as illegal abortions (24). Since legitimate births declined by 164,000 between 1970 and 1971, and since we estimate that 28,000 of the decline can be attributed to the availa bility of legal abortion, it is clear the most of the decline in legitimate fert ity resulted from influences other than legal abortion.

Summary and Conclusion

In sum, it appears that legal abortion depressed overall fertility, but partice larly illegitimate fertility, by giving women an opportunity to terminat their pregnancies when other means of birth control either had not been used or had failed. If legalized abortion had not been available, an estimated addtional 39,000 illegitimate babies and 28,000 legitimate babies would have been born in 1971 in the United States.

While this makes up a small part of total births (3,500,000), the illegitimate births prevented represent almost onetenth of all out-of-wedlock children born in the country in that year. In addition > preventing these births the legalization of abortion appears to have reduced the incidence of pregnancyrelated marriages and thereby may have helped to limit subsequent marital disruption. Finally, legal abortion prevented large numbers of illegal abor

tions from occurring. Our data indicate that well over half-most likely between two-thirds and three-fourths-of all legal abortions in the United States in 1971 were replacements for illegal abortions.

Further declines in illegitimate birth rates for the country as a whole will depend, in considerable part, on the extent to which legal abortion becomes more readily available and more widely used. Theoretically, greater use of efficient contraception could also cause illegitimate fertility to decline. But there are many reasons why women do not use efficient contraception even when they know about it and have access to the materials (25). Even though the use of abortion throughout the nation is now legalized by the Supreme Court decision, this does not necessarily mean that services will in fact be everywhere more readily available. Our interstate analysis suggests that should the liberalization of abortion laws be reversed, not only would there be an upturn in illegal abortions and pregnancy-related marriages, but also a marked rise in illegitimacy, particularly among women who do not have the means to obtain an illegal abortion.

References and Notes

1. Abortion Surveillance Report-Legal Abortions, United States, Annual Summaries (Center for Disease Control, Department of Health, Education, and Welfare, Atlanta, Ga., 1970 and 1971). The 15 states which we have characterized as having legalized abortion are those in categories V, VI, and VII of the 1970 report (table 21, p. 39). The 15 states exclude Georgia. Although Georgia had a liberalized law, this law had been invalidated in 1970 and thus Georgia was placed by the Center for Disease Control in a different category. In 1971, the number and rate of legal abortions performed in Georgia was low (1579 legal abortions or 1.5 per 1000 women aged 15 to 44 years).

2. B. Berkov and J. Sklar, The Impact of Legalized Abortion on Fertility in California (International Population and Urban Research, Univ. of California, Berkeley, 1972); J. Sklar and B. Berkov, Stud. Fam. Plann. 4, 281 (1973).

1. The following studies have shown that illegitimate children experience more social and bealth handicaps than do legitimate children: K. Davis, "The birth rate and public welfare in California," testimony before the State Social Welfare Board, 1972 (unpublished); E. Crellin, M. L. K. Pringle, P. West, Born Illegitimate: Social and Educational Implications (National Foundation for Educational Research in England and Wales, London, 1971); H. Forssman and I. Thuwe, Acta Psychiatr. Scand. 42, 71 (1966); B. Berkov, in California's Twenty Million: Research Contributions to Population Policy, K. Davis and F. G. Styles, Eds. (Institute of International Studies, Univ. of California, Berkeley, 1971).

4. Methodological details of rate calculations and a list of specific sources of birth and population data are available on request. 5. A. Clague and S. Ventura, Trends in Illegitimacy: United States, 1940-1965, National Center for Health Statistics, Series 21, No. 15 (Government Printing Office, Washington, D.C., 1968).

6. The following is a comparison of our estimates of illegitimate births and total births. with figures published by the National Center

7. This yields lower illegitimate and higher legitimate birth rates than shown by the NCHS data, particularly for nonwhite women over age 25. The reasons for considering separated women as exposed to the risk of illegitimate childbearing are discussed in B. Berkov and P. Shipley, Illegitimate Births in California 1966-1967 (State of California Department of Public Health, Berkeley, 1971).

8. National Center for Health Statistics, Vital Statistics of the United States, 1968 (Government Printing Office, Washington, DC., 1970), vol. 1, table 1-6.

9. P. Cutright, in Demographic and Social Aspects of Population Growth, C. F. Westoff and R. Parke, Eds. (Commission on Population Growth and the American Future, Washington, D.C., 1972).

10. K. Davis, in ibid., pp. 235-265. 11.

in Contemporary Social Problems, R. Merton and R. Nisbet, Eds. (Harcourt Brace Jovanovich, New York, 1971). pp. 313-360; M. Zelnik and J. Kantner, Soc. Sct. Res. 1, 335 (1972).

12. M. Diamond, P. Steinhoff, J. Palmore, R. Smith, J. Biosoc. Sci. 5, 347 (1973); K. Luker, thesis, Yale University (1974).

13. There were some states in which abortion was legalized prior to 1970. The lack of apparent effect on illegitimacy by passage of abortion legislation during this period was due in part to the relatively short time during which the statutes had existed and in larger part to the fact that despite "legalization" many restrictions remained which prevented the widespread use of abortion at that time. It was not until 1970 that New York, Washingon, Alaska, and Hawaii adopted the first laws with virtually no restrictive requirements. For an analysis of the effects of California's abortion law during 1967 to 1970, see E. Jackson, in California's Twenty Million: Research Contributions to Population Policy, K. Davis and F. G. Styles, Eds. (Institute of International Studies, Univ. of California, Berkeley, 1971).

14. The lesser decline in illegitimacy in 1971 for nonwhite women in abortion states does not reflect a lesser use of legal abortion; in California and New York City legal abortion rates in the period 1970 to 1971 were higher for nonwhite than for white women. The lesser decline in illegitimacy for nonwhite women reflects their higher pregnancy rate [see our data in (2) and the paper by Tietze (15)). 15. C. Tietze, Fam. Plann. Perspect. 5, 36 (1973). 16. New York City Department of Health, Bureau of Maternity Services and Family Planning, unpublished data prepared in 1973; State of California Department of Public Health, special tabulations of reports of 1972 therapeutic abortions.

17. We estimate that in the fiscal year period July 1970 to June 1971 (the period when a pregnancy would have had to be terminated to prevent a birth in 1971) there were 406,000 legal abortions performed in the United States. This estimate for the fiscal year period was derived from the total in calendar year 1971 [see Center for Disease Control (D)] by applying the ratio of fiscal to calendar totals available for New York State and California. It was assumed that the percentage of abortions performed on married and unmarried women was the same in the fiscal as in the calendar year period.

18. This method yields a more conservative estimate of the number of nonwhite illegitimate births prevented, and a more realistic estimate of white illegitimate births prevented, than the method used for legitimate births prevented (if one applies the relative decline in nonabortion states to the abortion states). Since white unmarried women living in nonabortion states appeared to have made extensive use of out-of-state abortion, if one applies their fertility rates to women in abortion states one obtains an unrealistically Jow estimate of white illegitimate births prevented.

19. The 14 percent assumes that one

abortion

is required to prevent one live birth. The figure rises to 17 percent if we apply to the United States Tietze's estimate for New York City that 1.2 abortions are required to replace one live birth. This takes account of the fact that a woman having an abortion is more quickly reexposed to the risk of preg nancy than a woman having a live birth [see Tietze (15)).

20. Using data from the Health Insurance Plan of Greater New York (HIP) on fetal death probabilities adjusted for undetected fetal deaths [see Shapiro et al. (27)), and assuming that on the average legally aborted pregnancies had already progressed through the eighth week of gestation, we estimated that approximately 13 percent of the legally aborted preg nancies would, without interruption, have ended as fetal deaths. Although the HIP fetal death experience may be more favorable than that for the general population of unmarried have women obtaining legal abortions, we attempted to compensate for this by assuming a shorter average gestation period than is actually the case for legally aborted pregnancies. For selected states in the United States in 1971 (excluding California and New York), the median gestation period was 10.5 weeks. For California in the first quarter of 1971, the median gestation period was 11.2 weeks. Data for New York do not permit calculation of a median gestation period (see (7). Even if we assume that the 13 percent fetal loss is an underestimate, it remains that if unmarried women currently using legal abortions were to try to carry their pregnancies to term, a relatively small part of these pregnancies would have ended as fetal deaths.

true

21. S. Shapiro, H. S. Levine, M. Abramowicz, Adv. Planned Parenthood 6, 62 (1971). 22. National Center for Health Statistics, Monthly Vital Statistics Report 18 (supplement) (Government Printing Office, Washington, DC., March 1970).

were

23. This estimate includes legitimate births that premaritally conceived. Unfortunately. data are not available that would permit us to estimate this group separately from legitimate births conceived after marriage. 24. See (20) for discussion of fetal deaths. 25. E. Sandberg and R. Jacobs, Am. J. Obstet. Gynecol. 110, 227 (1971); see also (12).

26. The average annual changes were calculated by computer from unrounded estimated rates and will therefore differ slightly from average annual percent changes calculated from the rounded figures shown for the annual birth rates.

27. To calculate the average annual percentage change, we assumed that the rate of change between 1965 and 1970 was constant; that is, we determined the value for in the formula (rate in 1970/rate in 1965). 28. Rates were calculated by relating illegitimate (or legitimate, Tables 4 and 5) births regardless of age of mother to the estimated number of unmarried (or married, Tables 4 and 5) women aged 15 to 44 years. 29. We acknowledge support from the Center for Population Research, National Institute of Child Health and Human Development (NIH-NICHD-73-2728). We thank Hazel Anderholm for graphic art work, Karen Corson for typing the manuscript, Arlene Guerriero for administrative and research assistance in preparation of the data, Sarah Lee Tsai for technical advice and the development of complex computer programs, Jan Seibert for editorial suggestions, and Kingsley Davis, who directed this study, for his invaluable guidance and encouragement. We also thank the National Center for Health Statistics and the many State Health Departments that provided data for this study.