Click for next page ( 265


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 264
CHAPTER 10 ESTIMATES OF PUBLIC COSTS FOR TEENAGE CHILDBEARING: A RE`rIEW OF RECENT STUDIES AND ESTIMATES OF 198 S PUBLIC COSTS Martha R . 8u rt with Frank levy . Teenage childbearing is a shot topics, not least because legisla- tors, programofficials, and the generalpublic are beginning to real- ize the tremendous costs it imposes. Dur ing the last 5 years several attempts have been made to obtain good estimates of these costs. Most interest has focused on public costs--AFDC (welfare), food stamps, medical assistance and social services. However, at least one study looks also at private costs. Two studies explore these costs at a national level; four others focus on a single state, county, or other local geog raphical area. These studies use quite varied methodologies and arrive at quite different cost estimates. Some studies calculate the public cost over a 20 year per iod of the family begun by a single teen birth, and the total cost for the sane 20 year per lad of the cohort of families begun by a teen birth in a single year. They ask, for example, what will the family of a woman who has her first baby as a teenager in 1985 cost the public by the time her f irst baby reaches adulthood? Estimates range from S13, 852 per family (Mecklenburg County, NC, in 1979 dollars) to S18. 710 (nationally, in 1979 dollars) . The model study of this type is the one conducted by SAI International (1979} for the Population Resource Center. Th is study estimated that the family begun by each first birth to a teenager in 1979 would eventually cost taxpayers S18, 710 (value in 1979 Collars); further, it estimated that all fami- lies begun by first births to teenagers in that year alone would eventually cost taxpayers at least S8.3 billion ~ in 1979 dollars) . Throughout the rest of this paper, these f igures will be referred to as Single birth cost. and Single cohort cost,. respectively. All but the ~ inal section of this chapter was or ig inally prepared by Burt for the Center for Population Options and published by COO under the title, reestimates of Public Costs for Teenage Childbearing: A Review of Recent Studies and Estimates of 1985 Public Costs,. and is reprinted with the permission of the Center for Population Options. The f inal section on the uses of these costs in benef it-cost studies was jointly wr itten by Burt and Levy. 264

OCR for page 264
265 This methodology uses as its population base &~1 teenagers experi- encing a first birth. Many of these teens will nearer receive welfare. The cost estimates are averages, spreading the public cost incurred by some teenagers over all teenagers. Therefore, they will be consider- ably lower than estimates which include only welfare recipients as the i r base . Other studies estimate the total public outlay in a given year that is attributable to teenage childbearing, including expenses for fami- lies in which the first birth occurred when the mother was ~ teenager, even though she may now be considerably older. Since approximately half of the AFDC caseload at any time consists of families begun by teen birth, but only 4 percent of these families are headed by women who are currently teenagers, this f igure~-total public cost for a single year--is very large. The model study of this type, Moore, Wertheimer and Holden (1981), estimated that the public spent S8.6 billion in the single year of 191S on AFDC, Medicaid and food sta - e for families begun by a teen birth. This f igure will be referred to as ~ single year cost. throughout this paper. Table 1 summarizes the costs estimated by the studies reviewed. The remainder of th is paper descr ibex the studies reviewed, and dis- cusses the methodological issues they raise. PART I THE STUDIES For each study reviewed, the title of the appropriate document, its authorts), the jurisdiction studied, the data sources and years used, and the dollar estimates developed are given first, followed by a be ief descr iption of the report' s contents. National Stud ies Single Birth and Single Cohort Cost Estimates Title: Can Analysis of Government Expenditures Consequent on Teenage Childbirth. Menlo Park, CA: SRI International, 1979. (37 pages) Authorts): None listed Jurisdiction: Entire United States Data Sources and Year(s): 1976 U.S. nasality statistics; published studies of welfare and other service costs, and studies of probabil~ ties that families begun by a teen birth will receive these services All expressed in 1979 dollars.

OCR for page 264
266 TABLE 1 Summary of Cost Estimates for Teenage Childbearing from Local and Nat ions ~ Stud ies Single Single Single In ~ irth Cohort Cost Year Cost Constant Study Cost (in millions} {in millions) 1985 S7 SRI Internationall 18, 710 8, 300 Not 27, 272 est. imated Mecklenburg Co., NC2 13, 852 12 estimated 20 ,191 St. Louis3 Not 17, 675 82 estimated 21, 463 Moore, Werthe imer and Not Not Holden4 estimated estimated 8, 550 16.33 bil Wertheimer and MooreS Not Not estimated estimated 5, 830 11.14 bil Monroe Co., NY6 Not Not 26 (1977) 0.44 bil estimated estimated 24 (1978) 0.38 bil lFor U. S . as a whole; expressed in present value 1979 dollars. 2 For Mecklenburg Cc ..~ty (Charlotte), NC ; expressed in present value 1979 dollar s. 3For St. Louis city and 8 surrounding counties; expressed in present value 19 81 dollar s. 4For the U .S . as a whole, expenditures estimated for 1975, in '975 dollars. SFor U. S . as a whole, expenditures estimated for 1990, in constant 1982 dollars, incorporating the more restrictive eligibility criteria for welfare programs legislated by the Omnibus Budget Reconciliation Act of 1981 (P.L. 97-35). 6For Monroe County (Rochester), NY, actual expenditures for 1977 and 1978. 7Us ing OMB def lator for payments to ind ividuals. Single Birth Cost: S18,710 S ingle Cohort Cost: at least S8 .3 billion Th is report serves as the model for many of the local studies den scribed below. SRI used a variety of published statistics to develop its estimates of cost and welfare and service utilization. Most juris- d lotions can develop parallel local statistics and calculate similar cost estimates for local decision making. The major contribution of thi s study 1 ies in the alar ity of its assumptions and in its methodol- ogy. Because these have been adopted in other studies, they are den so ribed here in some detail.

OCR for page 264
267 SRI used all births to teenagers, whether married, premarital con- ceptions or out-of-wedlock. on the assumption that births to teenagers under any circumstances increase the risk of public dependency. Their single birth cost is calculated by dividing the total single cohort cost by the number of f irst births to teenagers. All higher order births to teenagers are treated as extended costs of a f irst birth to a teenager . They calculated costs separately for three age g ro'~ps--ld and younger, 15-11, and 18-19--because teenagers giving birth at each age encounter signif icantly different risks and probabilities for using dif ferent services. SRI calculated medical costs for the year of pregnancy and birth and for the next 9 years (but not for years 11-20), because the overwhelming majority of medical costs occurred in those years. Wel- fare costs are calculated for 20 years, including the year of the birth, and adjusted for the probability that the household would still be on AFOC for each year after the first birth. Medical costs included charges for prenatal care, normal, C-section and~complications. den liveries, normal pediatric care, and pediatric care required because of the special medical problems encountered by some proportion of children born to teens. Welfare costs included AEDC, food stamps and other food programs, social services, and public housing expenditures. Further, SRI discounted furler costs. Discounting takes into con- sideration that a dollar in hand today is worth more than a dollar 10 years f ram now, because it can be used to earn money in the meantime, thereby of f setting some future costs. SRI assumed that ~ dollar in- vested today could earn an average of 10 percent a yeas, but that in- f ration would absorb 6 percent of that (on the average), leaving a net discounting factor (or real interest rate) of 4 percent, compounded annually. All future year costs of teenage childbirth were discounted by this rate. SRI ' s f igures for Single Birth Cost (S18, 710) and Single Cohort Cost (88.3 billion) are therefore expressed in represent value. (i.e., discounted) dollars--the amount one would need to Connie now to meet the present and future costs associated with a single f irst birth or a single cohort of f irst births. Finally; SRI calculated f_ cost., not martinet costs. See the discussion of Walentik (1983), done for the Danforth Foundation with respect to St. Louis and surrounding counties, to understand the pl icat ions of th is me~chodolog ical dec is ion. S ingle Year Cost Estimates Title: Teenage Childbearing: Public Sector Costs. ~ Final Report to the Center for Population Research, NICHD, on Contract No. NO1-HD-92822 . Washington, D.C .: The Urban Institute, 1982. (205 pages, plus appendices). Author(s): Richard F. Wertheimer and Kristin A. Moore Jur isd iction: Entire United States Data Sources and Year ~ s): Method 1: Current Population Survey (CPS) for March, 1976; Survey of Income and Education, 1976; and AFDC Surrey,

OCR for page 264
268 1975 were the data bases for cost estimates. The National Long itudinal Surrey of Young women and the Panel Study of Income Dynamics were used to establish probabilities of receiving AFDC, given specific household characteristics. TRIM, the Transfer Income Model, is a computer s~mu- lat ion model des igned to ass ign benef its on the teas is of household characteristics and income. It was used to produce the cost estimates based on the CPS and ache SIE. Method 2: Following this work, a dynamic simulation was run to ~ age. the population to the year 1990 and assess the welfare status and other character istics of individuals at that time. Single Year Cost: by Method 1--S8.55 billion for 197S; by Method 2 --85. 83 billion for 1990. This research study's primary goal was to project public sector costs associated with teenage childbearing to the year 1990, using assumptions about welfare eligibility based on the Omnibus Budget Reconciliation Act of 1981's more restrictive provisions, and to simu- late the ef facts of alternative scenar ios on these costs. Among the alternative scenar ios tested were: What if births to teenagers under 20 (or under 18) were cut in half? What if only married teens had babies? What if second births to teen mothers were delayed? What if all teen mothers completed high school? For our purposes, the important and widely cited part of this study Is its estimate of APDC, Medicaid and food stamp expenditures in the single year of 1975 that went to families begun be a teen birth ~ rem cardless of the mother's current age) . Throughout the rest of this paper th is f igure will be referred to as ~ single year cost . ~ In 1975 that cost was estimated to be S8. S5 billion. This estimate was developed using computer simulations (TRIM, the Tr ansfer Income Model) and several net tonal data sets ~ the March 1976 Current Population Survey of nearly SO,OOO households: the Survey of Income and Education, 1976; and the 1975 ALEC Survey) to der ive costs for AFOC and food stamps. Var ious ad justments were made for the SIE, wh ich underestimates the number of babies present in households-- notably, the STE was ad justed to conform to the number and distr ibu- t ion of birth- recorded in vital statistics records for 1976, and for the known total budget for AFDC for that year. Medicaid costs were estimated by applying average Medcaid expenditure for recipients of different ages to the AF9C Survey date. The methodology used in making these estimates is described briefly in Section I and Appendix A of the study' s Final Report. This methodology, and the methodology of the larger dynamic simula- tion leading to the figures for 1990, is certainly more sophisticated, and more expensive, than feasible for most local jurisdictions or states to use. Nor is it necessary for local purposes, since this study provides a conservative guideline for calculating local AFDC

OCR for page 264
269 expenditures. Regardless of which survey was used as the cats base, or whether estimates were adjusted or unadjusted, all methods of cal- culating costs resulted in estimates that 50-56 percent of AFDC costs, or AFDC and food stamp costs, were spent on households begun by a teen birth. Therefore, local efforts to quantify single year costs attri- butable to teenage childbear ing could simply assume that 53 percent (or 50 percent, or S6 percent) of the AEDC expenditures for ~ given year were attributable to teenage childbearing. The same assumption could also reasonably be applied to that proportion of food stop expenditures that went to AFOC households during ~ given year. Alternatively, many states are now calculating the proportion of their AFDC caseload attributable to teenage childbearing (that is, the proportion of current cases that involve a family begun by a f irat birth to a teenager) . If a jurisdiction has local f igures available on this proportion, the local f igures should be used. Local Stud lea Single Birth and Single Cohort Cost Estimates T isle: IF inane ial Report: Adolescent Pregnancy. ~ Charlotte, NC: Mecklenburg Council on Adolescent Pregnancy, node (probably 1980). (7 pages) Author t s): None listed. Jurisdiction: Mecklenburg County, NC (Charlotte) Data Sources and Yearts): 1979 births and welfare/service coat data from county and state agencies and private studies. Single Birth Cost: S13,8S2 Single C:,burt Cant: S-11.746.283 This report is very brief and to the point, presenting a Balance sheet. on the public cost of f irst births to teens in 1979, projected for 20 years. The report follows the methodology developed by SRI International for projecting these costs, including d~wountinq, using only f irst births in the base year, treating teenagers 14 and younger, lS-17 and 18-19 as separate categories, and using first births to all teenagers (not just out-of-wedlock births). Public costs included were maternity, pediatric and other Medicaid benefits, WIC, AFCC, food stamps, school lunch and day care. (The report did not specify how day care costs were calculated). Total d iscounted (present velue) public costs for 848 f irat births to teenagers in 1979 were S11.746.283. or S13.852 per first birth. The report also presents another figure-the amount of money that should be used aside. with each adolescent birth (not just first birthe) to cover present and future benefit payments. This figure was calculated at S10,2SO in 1979. This latter figure is somewhat misleading. It is derived by dividing the Single Cohort Cost (611,746,283) by the total number of births in 1979 (1,146), whereas the Single Birth Cost

OCR for page 264
270 (S13,852) is derived by dividing the Single Cohort Cost by the number of f irst births ( 848 ) and assuming that second and higher order birth costs are part of the consequence of a first birth that occurred in an earlier year. I f, as the report states, the authors followed the SRI me~chodology, then the Single Cohort Cost should represent the cost of first births, with all their consequences over the next 20 years in- cluding subsequent births while the mother is still a teenager. S13, 852 is the amount that should be reset aside. for all first births, and it should cover the cost implications of subsequent births. Title: Teenage Pregnancy: Economic Costs to the St. Louis - Cononunity.. St. Louis, Mo: l~anforth Foundation, 1983. (37 pages) Author ( s): David S. Walentik Jurisdiction: St. Louis SMSA, consisting of the city of St. Louis, four surrounding Missouri counties and four surrounding Illinois count ~es. Data Sources and Yearts): 1981 local and Missouri birth statistics, local, Missouri and national costs figures and probabilities of using d ifferent services. S ingle ~ itch Cost: Sl4, 041 ~ for each birth, including f irst, second and higher order births) (S17.~S if you use only first births but the report does not calculate this ~ igure) S ingle Cohort Cost: S81. 8 million This study applied the SRI methodology, with three except ions, to 1981 data from the St. Louis area. The first exception is that the cost calculations were based on an 18 year rather than a 20 year prom Section. The second exception is that the calculation of Single Birth Cost was based on the total number of births to teenagers in that year, rather than on first births to teenagers. However, a Single Fourth Co=. figure comparable to SPI's and to Mecklenburo County's :. be calculated from data in the report. The third exception is more critical, and raises an important issue for all studies deriving cost estimates. The St. Louis methodology calculated ~ marginal cost to teenage pregnancy, whereas the SRI study calculated full cost. The difference is this. Although teenage mothers have an increased probability of welfare dependency, women who have their first birth at age 20 or later still have some probability of receiving welfare, and having other costs of the it pregnancy arid childrearing borne by the public. Further, most interventions will at best delay births until teenagers reach the age of 20 or more; they will not entirely prevent births. Many personal characteristics of these women predict to some level of welfare depender~y even if they succeed in delaying childbearing into their 20's. The St. Lou is study sought the Snare inal increase in public cost attributable to a teen birth, over and above the probable cost to the public of a birth to a non-teenager. The SRI study calculated the full cost borne by the public of a birth to a teenager, without taking into consideration that the public would have some probability of

OCR for page 264
271 incurring welfare costs regardless of the age of the mother at f irst birth. Both figures are important. The full cost figure will obviously be larger than the marginal cost figure; but it would mislead policy makers to infer that the public would save the full cost if teenagers delayed childbearing. Rather, they would save the marginal cost. Wertheimer and Moore ~1982) estimate that reducing births to women under 20 by 50 percent would save only 20 percent of welfare expendi- tures. Presumably the savings would be clone to 40 percent if all such births were eliminated~an unlikely scenario. The discrepancy in rem duced welfare costs occurs because women who experience their first bi rth when they are 20 or older still have some probability of rem ceding welfare. Therefore, both f iguzes might be calculated, but used for different purposes. The St. Louis report provides the reader with much more information about teenage sexual activity, pregnancy, pregnancy resolution, child- bearing, marriage and separation, and decreased educational attainment than is contained in the cost calculations. It gives an overview of issues involved in teenage childbearing using national accounts, illu- strated with local data where these are available. By doing this it can serve as a ~primer. in the issue. and consequences of teenage childbearing for local decision makers. S ingle Year Estimates Title: Research on the Societal Consequences of Adolescent Ch ildbear ing: Welfare Costs at the Local Level. ~ Final report on NICHD Contract No. NO1-HD~02838. Washington, D.C.: Bokonon Systems, Inc ., 19 81. (92 pages, plus appendices) Author ~ s): A. Harvey Block and Susan Dubin Jurisdiction: Monroe County, New York (Rochester) Data Sources and Year(s): 1977 and 1978 county data on all public l assistance payments to cases open at any time during those 2 years. S inale Year Costs 1971: fir~t-birth-as-teen families: S26.3 million; S5, 083 per case no~teen-birth families: S17. 5 million; S4, 900 per case 1978: first-birth-as-teen families: S24.0 million; SS. 533 per case no~teen-birth families: S15. 5 million; S5, 127 per case This study took advantage of Monroe County's highly automated public assistance data system to make estimates of single year costs attr ibutable to teenage childbear ing for the 2 years 1977 and 1978. During those 2 years 11,500 families, with an average 2.5 children

OCR for page 264
272 each, had open cases with We Monroe County Department of Social Ser- vices, the vast bulk of wh~.~h (81 percent) were AFDC cases. 7,021 families appear in both years, suggesting long-term or recurrent wel- f are dependency . Th is also means that 4, 4 79 f Emil ies appear on the public assistance rolls during only one of the 2 years, suggesting only short-term dependency. Families begun by a f irst birth to a teenager comprised S6 percent of the public assistance caseload (59 percent of the AFOC caseload), and absorbed 60 percent of the cost for public welfare programs. They had slightly more children, on the average, than no~teen-birth families (2.6 vs. 2.3 children}. Fire percent of welfare families were multi- unit families--&chat is, the youngest family member was the grandchild, rather than the child, of the head of house hold, and all three genera- tions comer ised a single welfare case . Despite encountering some data missing from the computerized system (case and individual identification numbers were not entered on all transactions) the researchers were able to calculate public assistance costs separately for families begun by a teen birth, and for families in which the f irst birth occurred when the mother was 20 or older. These costs include AFDC payments, Medicaid, single issue, emergency and vendor payments, and estimated food stamp costs. 1977 costs for families begun by a teen birth totaled 826.3 Nonillion, while no~teen- birth families absorbed S17.S million in public outlays. Per case costs in 1977 were SS, 083 and 84, 900 respectively, with the slightly larger families of teenage parents accounting for the difference. 1978 costs were 624.0 million (65,533/case) for families begun by a teen birth, and S15.4 million (SS, 127} for no~teen-birth families. The foregoing camper ison may imply to some readers that f amities begun by a teen birth and families begun by ~ later birth absorb almost the same amount of public funds. After all, a 60/40 ratio does not seem so extreme. This impression is misleading. To provide an accu- rate camper ison, per case public expenditu res would have to be weighted by the probability that any household of that type would became an AFDC case. According to Moore (1978, Table 2), households begun by a teen birth have a 1 in 4 probability of receiving AFDC, whereas households begun by a later birth have a 1 in 10 probability of receiving AFDC. Applying these ratios to the Monroe County; per case costs, the average public expenditure in 1977 for a f irst-birth-as-teen household would be S1.271 (multiply SS.083 by .25}, but the average public expenditure in 1977 for a no~teen-birth household would be only 8490 {multiply 84,900 by .10) . Public expenditures are thus 2.6 tines a. high for households begun by a teen birth as for other households. An important contribution of this study is its detailing of costs to non-AFDC families. These turned out to be so small, in comparison to payments to AFOC families, that future studies can justifiably ignore them. Non-AFDC families received only 2 percent-of the total public assistance outlays for 1977 and 1978.

OCR for page 264
273 Another component of public costs revealed by this study is outlays for emergency, single issue and vendor patents (e.g., for day care, special counseling, WIN training) for AEDC families, over and above the basic AFT: grant, Medicaid and food stamps. In Monroe County this f igure amounted to approximately 12 percent of the total outlay for AE DC families, and 13 percent of the combined cost of AEDC (basic grant), Medicaid and food stamps. Jur isdictions vary widely in their use of emergency, single issue and vendor payments. Some jurisdictions never use these ~extras.--an AFOC family gets only its basic grant, rem gardless of circumstances. Other jurisdictions use special payments to a greater or lesser extent. Future efforts to estimate public costs might want to use these f figure to ad just a basic estimate, if the jurisdiction for which costs are being calculated makes significant use of special payments to AFDC families. This report contains many other elaborate analyses comparing the costs incurred for families begun by a teen birth and families for whom childbearing was delayed beyond the teen years. They are potentially interesting to ache specialist, but probably not to the lay reader who wants just the basic cost information. Title: Teenage Pregnancy and Teenage Parenthood in Illinois: 1979-1983 CostsEvanston, IL: Northwestern University, Center for Health Services and Policy Research, 1984. (131 pages} Author ~ s): Janet Reis Jur isdiction: Illinois Data Sources and Year ~ s): estimates for probabilities for 68 cost items were taken f rom 14 state or national sources; estimates for cost, and distribution of cost among state, federal, local go~rern- ments, businesses and private individuals were taken from 9 local and state sources, both public and private. All costs are translated into 1983 dollars. Costs for teen births in the years 1979-1983 were estimated. S ingle Year Cost (total cost, both public and Private): S853 million In 1983 dollars for births to teens that occurred only during the 5 year per iod 19?9-1983 . Costs include all costs incurred through 1983 of the five birth cohorts born in the years 1979-1983 (i.e., 5 years of costs for 1979 babies, 4 years of costs for 1980 babies, etc. ~ . Of this amount, 8425 million i. public expenditure, financed through taxi tion to private individuals and businesses. Businesses are estimated to pay an additional S72 million; priorate individuals pay the differ- ence ~ 83 56 million} . This is an ambitious study, departing in many ways from the concept tualizations, pur~ses and methodologies used in any other study rem viewed in this paper. Ma jor points of departure are: Inclus ion of all costs of raising children, whether borne by private individuals, businesses or governments. Th is makes the cost estimates much closer to the Cost of kinds. research (see Espenshade, 1984) than to other studies of the costs of teenage childbear ing .

OCR for page 264
274 o Estimates of costs to businesses, other than the business share a_ the tax burden for public costs. a Inflation of costs estimates for services receiving full or partial federal government support by 21 percent, to compensate for what the author calls the federal income tax penalty--the fact that Illinois citizens pay federal taxes, but don't see all those taxes coming back to the state. O Inclusion of elements of both ~ single year cost. methodologies and Single birth cost. and Shingle cohort cost. methodologies in the Awe study (which makes comparing cost figures from this study with those of any other study very difficult). O Recommendations for programs to reduce teen births, and esti- mates of the cost savings these programs would generate. Private as Well as Public Cost of Children Tt takes a lot of money to raise children. For most children, their parents pay the expensed of raising them, both in direct cash outlays and in opportunity costs (Espen~hade, 19841. Rich people spend more to raise their children than poor people, therefore in this con- text rich people's children ~cost. more than poor people's children. Bus inesses pay the it share, too, through insurance for employees and their families, through taxes for schools and other services for chil- dren, through the cost of crimes committed by children, and so on. The higher the employee's salary, the greater the probable contribu- tion of the employer to the welfare of the e~nployee's children. Chil- d ren require expenditures whether or not they are born to teenage bothers. While there is nothing intrinsically wrong with trying to estimate the toted expenditures for children, public or private, esti- mates of the total cost of teenage childbearing are better done in the f lacework of mare ine1 costs. 1 Combining public and private expenditures for children appears very confusing from ~ policy perspective, because the technique prom duces an estimate of total cost that will increase as the population improves its economic situation (to which reduced teenage childbearing is predicted to contribute), because economically better~off families will spend more on their children. Yet the public share of this in- creasing total will shrink, due to decreased reliance on public wel- fare among more economically viable families. The public focus rem mains the principle concern of policy makers, and studies of cost would best serve both public and other purpose. by keeping the dis- tinct~ between public and private costs clear. Adding to the inflated cost estimate derived from full vs. marginal cost is the researcher' s assumption that if a birth does not happen to a teenager, it will never happen at all, and therefore all the cost

OCR for page 264
284 TABLE 3 Public Cost of First Births to Teenagers in 1985, and Potential Savings Associated with Delaying these Birthsl Age at First Birth 15 15-17 18-19 All Teens F first Births2 S ingle ~ irth 9,638 144,308 217,18S 311,131 Full Cost 17, 724 l7, 689 11,214 13, 902 Potential Savings3 7,089 7,076 20-vear S ingle Cohort Full Cost ~ in billions) Potential Savings ~ in billions} 4,485 5,560 .~7 2.55 2.44 5.16 .07 1.02 0. 97 2.06 . 1 Twenty year projection, covering the years 1985-2004. Cost is expressed in 198S represent value. dollars, which means that this is the amount that would have to be set aside in 1985 to cover the 20 year cost of families begun by first birth to a teenager after taking into consideration inflation and the future earning power of a dollar invested in 198 5. 2 1984 nasality statistics were the latest available. The same numbers have been used in the 1985 estimates shown in this table. 3 Calculated at 40 percent of full cost. first birth: in Mecklenburg County, NC in 19790 All other things being equal, one would blare expected the 1985 estimate to be higher than any of these earlier estimates, simply due to inflation. The major reason for reduced cost is a reduced probability of receiving welfare (see below for an explanation of why this occurred}. ~us, changes in the availability of welfare support make a substantis1 difference in the long-term fiscal consequences of teenage childbearing.

OCR for page 264
285 A Comparison Figure: Continuous Welfare Dependency Many people may be surprised that the single birth cast of S13, 902 is so low. As explained above, the methodology used to reach this figure was based on all teenagers having a first birth. Many of these teenagers, either as teens or a" older waken, never go on welfare, or else they depend on welfare for relatively short periods of time (e.g., less than 2 years). This reality contradicts the popular perception that having a baby as a teenager dooms the resulting family to in med late and continuous welf are dependency. For many teenager s, such d ependency does not happen. Nevertheless, what many people want to know is, allow much does ~ family begun by a teen birth cost the public when it is continuously dependent on welfare?. One can develop this comparison projection using a simple modification of the methodology used to arrive at the S13,902 figure. As always, this comparison projection rests on certain assumptions which, if changed, would change the fine1 result. Let u s assume: A teenager 14 or younger, who has a f irat birth; immediately receives welfare for her baby and for subsequent babies; also gets welfare for herself once she become. an independent head of household; and stays on welfare continoualy for 10 years; 2. A teenager 15-17, who has a f irst birth; immediately receives welfare for her baby and for subsequent babies; also gets welfare for herself once she becomes andindependent head of household; and stays on welfare continuously for 7.5 year.; 3 . A teenager 18-19, who has a f irst birth, immediately receives welfare for her baby and for subsequent babies; also gets welfare for herself once she becomes an independent head of household; and stays on welfare continuously for 5 year s. D iscounting public costs in the same way we did for the basic 20- year pro Section, we ar rive at the following f igures: the 14 year old would cost the public S46,456 over the assumed 10 year period of wel- fare dependency; the 15-17 year old would cost the public 84d,2-01 over the assumed 7.5 year period of welfare dependency; and the 18 to 19 year-old would cost the public 830,955 over the assumed S year period of welfare dependency. Averaging these results--after weighting the f inures for the numbers of 14 year olds, 15-17 year olds arid 18-19 year olds exper ienc ing a f irst birth one err ives at an average, dis- counted f igure of 836, 502 for ~ teenager giving birth for the f irst time and immed irately beg inning an extended per iod of continuous wel- fare dependency (as described in our assumptions). This figure of S36, 502 may strike many readers as more ~reasonable. than the earlier figure of S13.902 that is, as more in line with their

OCR for page 264
286 expectat ions . As noted above, howeve r , it is t rue only for those teen- agers who actually do use welfare immediately and continuously. Such teenagers do not represent the majority of teenagers having children. SOU=ES AND ASSUMPTIONS E OR SINGLE: BE RTH AND SINGLE COMORT PUBLIC COST E STIMATE: S The approach used was, first, to estimate the cost for a single family begun by a first birth to a teenager in 198S, projected over the 20 year per iod 1985-2004, and then to multiply that estimate by the number of first births to teens in 1985 to get the cohort cost esti- mate. Following the SRI methodology, cost was estimated separately for teenagers experiencing their first birth at age 14 or younger, at age 15-17, and age 18-19, order to reflect the very different health and dependency risks entailed by births at these different ages. Basic Assumptions Completed Family Size Since a 20 year projection traces a teenager's childbearing career from the tine of her first birth, counting subsequent children {and their associated costs a. part of the cost of an initial teen birth, some assumption about the number and timing of subsequent births is necessary. Wertheimer and Moore's (1982, Table 20) estimates were used for this purpose. These estimates are that teens who had their f irst baby at 17 or younger would have 2.1 children by age 22, 2.8 children by age 27, and 3 .2 children by age 32 . Teens whose f irst child was born when they were 18 or 19 would have 1.6 children by age 22, 2.4 by age 27, and 2.8 by age 32. These figures were used to calculate annual average increases in family size, which in turn were used when calcu- lating the number of household members eligible for AFDC, food stamp and Medicaid support, and when to add extra medical expenses Lesson c iated wth pregnancy and childbearing. Basic Assumptions -- Probability of Receiving Welfare For teenagers ~ age 19 or younger), the probability used is .34 {Moore, 1978~. The rest of the probabilities come frown Wertheimer and Moore (1982, Table 41-A, Baseline scenario), which does not estimate the probability for teenagers. For women whose first child was born when they were 17 or younger, these probabilities are: .20 when they are 20-24; .16 when they are 25-29; .02 when they reach age 30 and older. For women 18 or 19 at f irst birth, the probabilities are: .15 when they are 20-24; .09 when they are 2S-29; and .03 when they are 30 and older. Applying these probabilities is quite straightforward if one is estimating costs only for 18 year olds, or only for 15 year olds. How- ever, the methodology used in this paper combines teenagers of differ- ent ages together and estimates costs for 15-17 year olds and for 18-19

OCR for page 264
287 year olds. This approach requires some adjustments in the Transition years. ~ For example, among 1S-17 year olds, in the third year of the projection, when the 17 year olds have turned 20 (and thus would be subject to a probability of welfare receipt of .20), the teens who were initially IS and 16 years old are only 18 and 19, respectively (and thus the .34 probability of welfare receipt is stil applicable to them). Therefore, the probability of welfare receipt used for the third year of the projection will be the average of the probabilities for each age, weighted by the proportion of the larger group (15-17 year olds) who are each age ~ i.e., 15, or 16, or 17) . The same sort of adjustment must be made for each year that part of the group is old enough for the next applicable probability, but part of the group is not. In the years of the projection when women are 20 and older, these probabilities are lower (by 4-1S percent, depending on the woman's age and age at first birth3 than those used by SR: and the other studies using the SRI method, because the estimates incorporate the more re- strictive welfare eligibility criteria required by the Omnibus Budget Reconciliation Act of 1981. The biggest difference made by these new criteria involves the amount a head of household could earn and still retain her eligibility for AFDC and Medicaid. When the stricter cri- teria took effect in 1982, this amount was reduced significantly, and many households receiving some income through earnings were removed from the AFDC rolls. In the present calculations, the changes in eli- gibility requirements for AEDC affect the estimate of total cost and the estimate of mare inal savings. Basic Assun~tions--Number of First Births to Teenagers 9,638 first births to girls 14 and younger, 144,308 first births to 15- 17 year olds, and 217,185 f irst births to 18-19 year olds. me se f ig- ures are taken f ram 1984 nasality data (NCIlS, 1986, Table 2) . 1984 data are the latest available, and 1985 f igure. have been assumed to be similar. Basic Assumptions--Inflation and Discount Rates Inf ration is assumed to average 5 pe Scent per year over the 20 year e involved in these projections. Money invested in 1985 is assumed to grow at a rest interest rate (after accounting for inflation) of 4 percent per year, compounded annually. Thus 4 percent is used as the discount rate. This practice follows the discounting example of SRI I nternat tonal ~ 19?9 ~ .

OCR for page 264
288 BeS ic Assumptions-- Potent ial Saving 8 The potential savings from reducing or eliminating teenage childbear- ing is assumed to be 40 percent of the full public cost of teenage childbearing. This figure (40 percent) is derived from Wertheimer and Moore (1982), who used dyr~aJnic stimulation techniques to estimate the probable effect on welfare expenditure. of several scenarios related to teenage childbearing. me greatest savings (20 percent) were achieved under a scenar to in which births to women under 20 were rem duced by half. Double savings ~ i.e., 40 percent rather than 20 per- cent) are assumed to result from doubling the reduction in teenage childbearing {from a 50 percent reduction to a 100 percent reduction). Coat Data AFDC cost: S113 per recipient per month, estimated for 1985 (FACE, Section 8, Table 17} . The same source indicates an anticipated 1.8 percent annual increase in the benef it per recipient. Since inflation is assumed to be 5 percent. per year, this means that the value of the A*DC benefit loses ground to inflation at a rate of 3.2 percent per year. Its value has been reduced appropriately for the years 1986-2004. Food stamp cost: S42.80 per recipient per month in 1984 (WHOP, Appendix G. Table 1~. Assumed to rise 2 percent per year (actual increase indexed to the food component of the Consumer Price Index), therefore the 1985 f igure used was 843. 66. For years following 1985, the per recipient cost was reduced by 3 percent per year {5 percent inflation minus 2 percent increase in food stamp allocation). All adjusted for 12 percent administrative costs, and for a 7S percent participation rate of AFDC households in the food sta" pros ram (WHOP, Appendix G. Table 1 and Section 8, Tab_- 18~. Annual Medicaid cost: 1984 was the latest year for which figures were available from the Medicaid Hotline (Social Security Administra- tion), which summarizes the states' annual 20-A-2 reports and gives information over the telephone. Each AFDC child cost Medicaid 6414 in 1984, while each AFOC adult cost Medicaid S847. These figures have been increased to 843S and S889 for 1985, based on an assump- tion of a 5 percent increase in Medicaid outlays. As explained in relation to Table 2, this is ~ conservative assumption, since real medical costs, and Medicaid costs, have been increasing an average of a-12 percent per year. The 5 percent inflation rsee and the 5 percent increase in medical cost outlays cancel each other out, so 8435 and S889 will be used as Medicaid cost outlays for all 20 years of the present projections. Medicaid costs associated with births : These calculations used 83,107 as the cost borne by Medicaid ~ including prenatal care costs) for

OCR for page 264
289 birth to a teenager 14 or younger; 82,867 a. the cost for 15-17 year olds; S2,397 as the cost for 18-19 year olds, and S2,446 as the cost for women 20 and older. These figures are an average of costs estimated by three different studies: Walentik (1983~; Wertheimer and Moore (1982, Appendix B); and Burt et al. (19841. These costs were multiplied by the probability of a birth to produce the birth cost associated with each year of the 20 year pro ject ion. Social services cost: was assumed to be 5 percent of AFoC cost, fol- lowing Welentik's (1983) logic and example. Publ ic hous ing cost: Was assumed to be 2 percent of AFOC cost, fol- lowing Walentik' s (1983) and Say International' s (1979) examples. AFDC/food sta - /Medicaid administrative cost: AFDC and food stop administrative costs were calculated using 12 percent of direct benef its, based on histor teal precedent. Medicaid administrative costs were calculated at 5 percent of benefits, also based on historical data. USING THE COSTS IN BE:NEFIT-COST ANALYSI S2 The public costs of teenage childbearing are rarely calculated as ends in themselves. Typically they are gathered to bolster the argu- ment for preventive programs--sex education, counseling, providing contraception, and so on. But which programs should be chosen? From an economic perspective, certain rules apply for making these choices even when budgets are plentiful. Suppose we def fine the Net Benef its of a prog r am in the following way: A program's = net benef its the dollar value of the program' s benef it - the dollar value of the program's costs No matter how large an agency's budget, no program should be under- taken unless its Net Benefits are positive~-i.e., unless the value of its benef its exceeds its costs. When budgets are tight, the problem becomes harder. Now the administrator should select those particular programs that produce the greatest Net Benefits. While all thin sounds fine in theory, in practice it is not so easy. In fact there are no perfect formulas, but it is possible to give some general ideas and numbers that will help administrators to examine their own options. In most situations, the JI105t defensible perspective to adopt for cost-benefit analysis is the taxpayers' perspective: how many public dollars will this program save? This means that if a program benefit is to be counted, it must be translated into reduced public dollars. Lower ing the rate of pregnancy must be translated into reduced welfare

OCR for page 264
290 and medical costs. Staying in school longer must be translated into postponements of pregnancy and higher likelihood of employment which must again be translated into reduced welfare and medical costs. And so on. While this Strategy may be confining, it provides a consistent way to think about program evaluation. M!:ASVRING PRoG RAM BENEFITS A typical program does not keep records in terms of public costs saved.. Rather it is likely to generate information of the following kind: A) When compared to normal expectations (or to a control 9 Coup) 16 year-old girls were 20t less likely to have a first pregnancy during the year they were in the program. Or, B) When compared to normal expectations (or to a control group), 17 and 18 year-old mothers were 158 less likely to have a second pregnancy dur ing the 18 months following prog ram pert ic ipation. How do either of these statements translate into dollar benefits? R.ain with statement (A). Recent estimates by Burt suggest that the Dresent discounted value of future public costs associa- in 1985, ~ ~ ~ ted with a f it st teenage bi rth are: TABLE 4 Present Discounted Value of Future Public Costs Associated with a First Teenage Birth Age at First Birth Public Expend itures 15 16 17 18 19 S18, 130 S17,851 S17, 464 S12,214 S10. 671 These coats inc3.ude assumptions about subsequent fourths, the likes lihood the woman receives AFoC, etc., as explained e~-.:;~ier in this chapter. They reduce with each year a first birth i^. postponed be- cause of reduced probabilities of receiving welfare, caller completed f amily size, and fewer medical complications for later childbearers.

OCR for page 264
291 Statement (A) does not say what happens after the year the girls were in the program. We will begin with ~ conservative interpreta- tion. Suppose that the program only succeeds in postponing the girl's first pregnancy for a year while her total number of children, her chances of completing high school, etc., are the same as those of teenagers not participating in the program. What public costs have been saved? The answer is straightfoward, but also involves discounting. If the girl had her first child this year, we would be confronted with public costs whose present value totals S17.851. Thin means that by putting S17,851 in the bank this year, we would generate enough revenue to cover all future costs. I f the g irl has her f irst child next year, at age 17, the 68~e logic holds except that now the process is postponed for ~ year and we need to deposit S17,464 in the bank for next year to cover all future publ ic costs. With cur rent interest rates at about ~ percent, it follows that to build up the S11, 464 for next year, we need to deposit an amount this year, C, which solves the equation: or C x ( 1. 01) - S17, 464 C = 17,464 ~ S16.321 1.07 (1) (2) Thus, if the girl has her first birth this year, (at age 16), we have to commit S17,851 this year to cover future costs. If the girl has her first birth next year, we have to conceit S16.321 this year which will g row into S17, 464 by next year and will be Buff iciest at that Time to corer the costs. Thus postponing the first birth by one year has saved us S1. 530 (S17,851 - S16,321}, assuming we really do commit the money this year and actually get a 7 percent rate of interest. To com.- plete the example, if girls in the program are 20 percent leas likely to have a first pregnancy during the year they were in the program, then the expected savings per girl is .2 x S1,530 ~ 8306 and if the program is worth any consideration, program costs per participant must be less than this amount. ~ f we make the same calculations for ~ postponement f ram age 16 to 18, or from 17 to 18, we obtain public savings of S7, 182 and 86,049, respectively. ~ t is apparent f tom these cats that the biggest petri - s will be associated with postponing a first birth until 18 or 19 (or even later). A delay from age 16 to age 17 will save S1.530. But a delay from 16 to 18, implying also high school completion, will save S7, 182: S12,214 / ~ I. 07} 2 ~ 10, 668; S17, 851 - S10. 668 ~ S7, 182. Even the one~year delay f rom 17 to 18 is Mu ite .valusble, ~ saving S6,049: SI2,214 / 1.07 ~ S11.415; S17, 464 - S11.~15 ~ S6,049. Since many more 17 year aids than IS year olds give birth, this implies that program efforts will have more payoff in terms of cost effectiveness if they focus on helping 17 year aids delay childbearing for one year.

OCR for page 264
292 Turning to statement (B)--the postponement of second pregnancies-- we can see that the methodology is the same but the costs are dif- ferent. The future public costs of noosing from one to two children are less than the costs of moving from no children to one. These costs have not been estimated. but the same methodology could be adapted to do so. St MPLE B REAR-EVEN ANALYSI S The previous examples began with a distinct advantage. We knew what impact the program had on postponing pregnancies. In many cases, this information is precisely what we do not know. In this case, a limited analysis is still possible. We cannot do a complete estima- tion of Net Benefi*$--tbat is, the dollar value of benefits less costs. But we can examine the possibility that benefits exceed costs and therefore that the program is worth undertaking at all. Suppose that we estimate that it costs S16 to have a young woman participate in a series of family life education classes. And suppose (following the discussion above) that we determine the public cost savings of postponing a first pregnancy for a year for a 16 year old is S1.530. It follows that if the program benefit. are to exceed costs, the program must cause at least 1 out of every N participants to postpone a pregnancy where N solves the following equation: or S1, 530 ~ N x S16 = 96 (3) Put differently, if 96 girls participate in the program, the total program costs will be 96 x S16 or S1, 536. As long as one of these par- tic-~ants postpones a pregnancy, the savings in public cost will be S1, ~ :~. In this case, the program would just break even in terms of net: benefits. If more than one participant postpones pregnancy, the Net -enefits are that much larger. In many situations, this kind of ~break-even. analysis may be helps ful. For while an administrator may not know a program's precise effectiveness, she or he may have a rough sense of its effectiveness and so may know whether the break-even N. calculated in (3), is at all plausible. For exile, if a particular program leads to ~ break-e~ren N of one postponed pregnancy for every three participants, the adminis- trator may know that such effectiveness is impossible and so the prom gram should not be undertaken. Conversely, if ~ different program has a break-exren N of one postponed pregnancy for every 600 participants, the program administrator may feel that number is quite plausible and this will make the program worth pursuing.

OCR for page 264
293 GLOSSARY D iscounting -- Reducing a nominal cost for ~ future year by the amount that money invested now could earn by that year, adjusting for inf let ion. Full Cost -- The actual, or present value, public dollar outlays for families begun by a teen birth, not adjusted for the probability that families begun by births to women 20 or older also have some probability of receiving welfare and other publicly supported services. Marg inal Cost -- The cost to the public of teen births, over and above the costs the public would incur for families begun by a later birth. This marginal cost is thus the full cost of teen births, minus the full cost for later births. S ingle Birth Cost -- The cost to the public of the family begun by a single teenage birth, from the pregnancy through the time the baby becomes an adult (20 years, in most calculations). S ingle Cohort Cost -- The cost to the public of all such families begun by teenage births in a single year, for the 20 years that the teenage mother and her baby may be dependent on public support. This figure equals the single birth cost multiplied by the number of first births to teenagers in a given year (cohort). Single Year Cost -- The cost to the public during a single yes: to support all families begun by a teen birth (although the mother in most cases will no longer be a teenger). Notes 1 For example, Dillard and pal (1982) assume tht teenage childbearing results in mars inally larger families, less education and poorer earn- ir~g capacity of teenage mothers. Their estimates of the cost of teen- age childbearing, over and above the cost of any childbearing, are based on the extra child the teen mother is likely to have, the oppor- tunity cost associated with not being able to enter the labor force as soon as a mother with fewer children, and the lower wages she will earn when she is in the labor force. This marginal cost turns out to be S37,050 for a poor teenage mother, in addition to the S61,613 cost for a poor mother who did not have ~ baby as ~ teenager. This marginal cost, while still substantial, is considerably less than the almost SIOO,OOO that would be implied by using the methodology of the Illinois study. 2 Frank Levy contributed the basic outline and concepts of this reck tion. Burt contributed the estimates for public costs of teenage childbear ing .

OCR for page 264