PHN-871 3 THE AMANPEMTAI INCENTIVE FAMILY WELFARE AND CHILD SPACING PROGRAM by Carl M. Stevens June 1987 Population, Health and Nutrition Department World Bank The World Bank does not accept responsibility for the views expressed herein which are those of the author(s) and should not be attributed to the World Bank or to its affiliated organizations. The findings, interpretations, and conclusions are the results of research supported by the Bank; they do not necessarily represent official policy of the Bank. The designations employed, the presentation of material, and any maps used in this document are solely for the convenience of the reader and do not imply the expression of any opinion whatsoever on the part of the World Bank or its affiliates concerning the legal status of any country, territory, city area, or of its authorities, or concerning the deliminations of its boundaries, or national affiliation. PHN Technical Note 87-13 THE AMMANPETTAI INCENTIVE FAMILY WELFARE AND CHILD SPACING PROGRAM ABSTRACT The Ammanpettai incentive family planning program was launched in February 1985 in the state of Tamil Nadu, India. This paper reviews the experience of the program in its first 10 months of operation. The main hypothesis being tested is that an incentive program of this type will, per unit of resources committed to it, result in a far greater proportion of women (in the defined population) effectively practicing contraception than would do so under alternative family planning schemes which entailed the same resource commitment, but which did not feature incentives. The major intented output of the program is reduction of the total fertility rate in the target population. Married women 18-35 years old who have not been sterilized and whose husbands have not been sterilized are eligible to participate. The women are advised on the various types of contraceptives available and which would best .suit their needs. Contraceptive supplies are provided free. Participants visit the clinic once a month and, if not pregnant, receive an incentive payment of Rs. 30. This type of incentive program gives participating mothers a direct, material and immediate reason to learn how to prevent pregnancy and focuses consumer education on how a small family can be achieved. Program experience at the time of this study was too brief to generate sufficient data on how successful the program was at averting births. However, the data available indicate that the program was able to motivate large numbers of women to use modern temporary methods of contraception who had no previous history of contraceptive use. Of the 343 women enrolled in the program in February 1985, 300 were still in the program in December 1985. Of the 43 women who left the program, 14 had become pregnant, 20 had moved from the program area. Prepared by: Carl M. Stevens, Consultant to the World Bank Reed College, Oregon June 1987 wr 2 Introduction The Ammanpettai Program (located near Thanjavur in the State of Tamil Nadu, India), based upon a novel incentive strategy, was launched as a pilot program February 1985 and has been in operation for about one year at the time of this writing. It is' anticipated that findings from the operation of this pilot program will contribute significantly to strengthening family planning programs in India and elsewhere. Following a brief description of this Program, this discussion will explain its relationship to the Government of India's regular, ongoing family planning effort. As will be seen, the Ammanpettai Program complements that effort in a number of important ways. This discussion will then address some economic, financial and fiscal implications of the Ammanpettai Program, reaching the conclusion, among others, that were an Ammanpettai-type Program extended to cover the population of Tamil Nadu as a whole, the fiscal burden would be within acceptable bounds. Finally, there will be a report on the performance of this program to date. Although it is early to reach firm conclusions on this score, it can be said that the evidence to date is most encouraging. The Ammanpettal Program This pilot program operates in two ethnically and economically similar, geographically proximate, rural areas. The experimental (incentive) group is based in the Stella Maris Charity Clinic in Ammanpettai and includes the entire Ammanpettai catchment area of approximately 6,000 people. 3 The control group is based in the government primary health center at Nadukkaveri 30 km distant and consists of all women using this clinic for their health care and family planning services. The distance between the two areas allows separation of the populations, but still permits comparison of groups living in similar conditions and adequate surveillance by program personnel. Married women between 18-35 years of age who have not been sterilized and whose husbands have not been sterilized are eligible to participate in the Ammanpettai Program incentive group. Eligible women from the catchment area who want to join visit the Stella Maris Clinic, give information on the number and ages of their children and on their past and current contraceptive practices (if any), and provide a urine specimen for pregnancy testing. If they test not pregnant, they will be enrolled in the program as available places permit. Clinic staff will discuss with them the various modes of temporary contraception and will make a recommendation based on the participants' needs and preferences. Contraceptive supplies are provided free of charge. The participating woman will continue to visit the clinic once each month and on each visit she is not pregnant she receives an incentive payment of Rs. 30/-. The clinic director met with women's groups in the villages of the Stella Maris catchment area and with community leaders to explain the program before it was started. However, the Ammanpettai Program employs none of the special education/motivation workers characteristic of most family planning programs. Indeed, most of the 300 women now participating in the program learned about it from health workers and oral and printed announcements posted in their villages which explained the program and instructed them to 4 report to the Stella Maris Clinic if they were interested in participating. The response to such notices clearly showed that women will come in large numbers to participate in such an incentive program and this in spite of the fact that fewer than 2.0% of the women now participating in the program had, according to their own testimony, ever used a modern contraceptive before. Only resource constraints prevent increasing enrollment in the Ammanpettai Program to two or three times the number of women now participating, i.e., to more than 50% of the eligible women in the catchment area. The overall hypothesis to be tested by operating the Ammanpettai Program trial is that an incentive program of this type will, per unit of resources committed to it, result in a far larger percentage of women (in the defined population) effectively practicing contraception than would effectively practice contraception under alternative family planning regimens which entailed the same resource commitment, but which did not feature incentives of this type. The following subsidiary hypotheses inform the expectation of this result, viz: (1) The Ammanpettai Program features incentives of a kind which will motivate relatively large numbers of women in the eligible group to take the initiative to join the program and will motivate them to want to continue participation in the program once they have enrolled. Thus, the program is an exceptionally effective way to introduce women to modern methods of contraception. (2) The motivation factors adduced in (1) give the participating women an immediate, tangible reason to want to learn how to prevent pregnancy, and this immediate "need to know" will be very effective im promoting acquisition of 5 the requisite knowledge--indeed, far more effective on this score than are the information/publicity components conventionally associated with family planning programs which frequently fail to promote acquisition of the requisite knowledge. (3) Having been introduced to modern methods of contraception, the women participating in the program will discover that child spacing is in fact very rewarding for various reasons in addition to merely the incentive payment featured by the program. These additional reasons and rewards provide strong, ongoing motivation to continue child spacing. (4) The women who participated in the program will effectively be introduced to fertility control in consequence of their participation. In addition, the fertility control behavior of other women in the community (defined population) from which the participants have been drawn may also exhibit some favorable changes. Broadly speaking, this "rub-off effect" will result from the mechanisms of social learning by imitation facilitated by the information resource represented by the women who have been participants in the program. Program Parameters: Overview of Implications for Cost Effectiveness All family planning programs which can be characterized as of the Ammanpettai type will share the central, distinctive feature of incentive family welfare payments to participants who present periodically not pregnant. Beyond this, however, different versions of programs of this type may have 6 different operating parameters and these differences can be expected to be important for program performance and program cost. A major intended output of the program (a major measure of program performance) is a reduction in the total fertility rate for the defined population from which the direct program participants are drawn.1 It is not the intention of the program to maintain participating women on incentive family welfare payments throughout their reproductive lives. Rather, a central principle which informs programs of this type is that women who participate in the program for relatively short periods of time will, owing to this experience, have a significantly higher probability of continuing with contraception after they leave the program than .would otherwise (i.e., without the program experience) have been the probability that they would have adopted contraception at similar ages and parities. Performance of the program in terms of reduction in the fertility rate for the defined population (FRpop) from which the direct program participants are 1Additional important intended outputs of the program include improvements in mother and child health and enhanced education attainment for the children of participants. For most family planning programs, in practice, to measure the performance of -the program over fairly short periods of time, resort is had to some surrogate "indicator" of the fertility rate, typically some measure of contraceptive prevalence, e.g., couple-years protection (CYPs). The same approach may be adopted with the instant program, although it is an advantage of the Ammanpettai- type program that the actual pregnancy status of the participants is known simply in the course of administering the program. 7 drawn will depend upon the strength of the direct participants' commitment to effective contraceptive practice and the "rub-off effect" (alluded to foregoing and anticipated to operate through the mechanism of social learning by imitation facilitated by the information resource represented by the women who have been direct participants in the program). Theory suggests that program performance will be a function of several operating parameters, viz: The rate of family welfare payments to direct participants (PAY): The periodic-reporting interval (in months) for, direct program participants (INTERVAL) The length of time direct participants are in the program (TIME). The larger PAY and TIME and the smaller INTERVAL, the greater the expected impact in increasing the strength of direct participants' commitment to effective contraceptive practice. The ratio of the number of direct program participants (PARTIC) to the size of the defined population from which they have been drawn (POP)--i.e., the ratio PARTIC/POP. FRpop will be a function of PARTIC/POP in a straightforward arithmetical sense (assuming that the program has a favorable impact in reducing the fertility rate for direct participants). This parameter may also tend to increase the rub-off effect. 8 The number of program participants handled per year (PARTIC) depends upon the number of participants in the program at any one time (SET) and the length of time each SET is in the program (TIME). Turning to the cost side, we need a proximate, cost-consequential measure of program output--namely, the number of direct participant contacts per year (PC). Generally, total cost of the program (TC) as a function of (PC) may be represented as: (1) TC(PC) = A + bPC Here A is fixed cost (e.g., clinic facilities, core staff). This sets the "scale" of the program. The coefficient b is average variable cost which, more particularly speaking, may be decomposed as: b1 - supplies per participant contact (e.g., pregnancy tests). b2 - staff time per participant contact (e.g., to administer tests). b3 - the incentive payment per visit (PAY). Since PC=SET x 12/INTERVAL, any given PC can be achieved with different combinations of SET and INTERVAL. Summarizing the foregoing discussion, it is hypothesized that program performance measured in terms of reductions in FRpop (in practice, perhaps represented by some surrogate variable) will be a function of various program 9 parameters as follows (holding the scale of the program and the b1 and b2 components of variable cost constant): (2) FRpop (SET, TIME, PAY, INTERVAL, PARTIC/POP) Generally, a definite function to represent equation (2) would be fairly complex. Thus, some of the parameters influence FRpop in more than one way, e.g., TIME, which contributes directly to the strength of direct participants' commitment to effective contraceptive practice and also is one of the determinants of the ratio PARTIC/POP. As the program is operated over time, successive sets of participants will be processed through such that the ratio PARTIC/POP will increase as a function of time (in a way that depends upon SET and TIME)- An appropriate definite function to represent (2) would probably contain both additive and multiplicative terms. As will appear from the discussion to follow, we do not, at this juncture, recommend attempting actually to estimate a definite function representation of (2). As equation (1) suggests, various versions of the program may be designed entailing the same total cost for each, but featuring different values for the parameters and hence, it may be anticipated, different program performance in terms of impact on FRpop. The cost-effective formats for the program will be those which, for given total costs, have the most favorable impact on FRpop. The program parameters can be regarded as "resource using" (in the sense that increasing PAY, TIME and SET and decreasing INTERVAL would all require increased resources). The general condition for cost effectiveness is that the yield on the margin, in terms of decreased FRpop, be the same for the resources allocated to each of the program parameters. That is, the trade-offs 10 within any given program budget (e.g., increase SET and compensate by appropriate decrease in TIME) are to be evaluated in terms of the marginal impacts on TRpop. It is easy to state the formal conditions for selection of the optimal (cost effective) program format. Actually to evaluate program formats in these terms would be quite another matter however. Indeed, a formidable observation regimen would be required for comprehensive evaluation. In any event, at this juncture, at the very outset of field experience with Ammanpettai-type programs, a more modest objective will be in order. We may realistically hope, if the field trial now underway (and a few variations on it) can be continued, to test the impact upon FRpop of a few plausible values for a couple of the principal program parameters (e.g., TIME and PAY). If on the basis of such initial field testing, the program proves as promising as we now anticipate, more extensive field testing will be in order. ai Relationship of the Ammanpettal Program to Ongoing Fertility and Fertility Control Events in India Fertility and Contraceotive Prevalence2 For some years, birth rates have been falling in India, more in some states than in others. Table 1 (following page) shows the average number of children born per woman by age and state. According to these data, the state of Tamil Nadu has the lowest fertility rate among Indian states. It is the plan of the Government of India to reduce the crude birth rates for India as a whole from its present approximately 33.0 to the low 20s by the year 2000. Currently, about 29% of married couples in India are protected against pregnancy in some fashion, about 80% of these by sterilization (i.e., a sterilization rate of about 23% of eligible couples) according to official reckoning. Estimates of contraceptive use based on household-survey findings which take account of not only methods supplied through the Government's program, but of all methods, including traditional ones, are considerably higher than the official estimates. Thus, a 1981 survey reported that 35% of eligible couples used some form of contraception (see USAID DELHI 1985, p. 19). Big differences were found by state. 2See: "Population in India's Development: USAID's Population Assistnace Activities." USAID, New Delhi, May 1985 (henceforth USAID DELHI 1985). This report provides an excellent discussion. 11a Table 1 Average number of children born per woman by age, 1981 Total Age group S1. Rural No. India,State Urban 15-19 20-24 25-29 30-34 35-39 40-44 45-49 50+ 1 2 3 4 5 6 7 8 9 10 11 INDIA* Total 0.17 1.13 2.41 3.46 4.26 4.71 4.99 4.74 Rural 0.19 1.19 247 3.53 4.35 4.81 5.07 4.80 Urban 0.13 0.89 2.24 3.22 3.98 4.38 4.68 4.50 1 Andhra Pradesh Total 0.28 1.27 2.39 3.15 3.79 4.08 4.34 4.06 Rural 0.31 1.30 2.41 3.17 3.80 4.10 4.35 4.08 Urban 0.22 1.18 2.33 3.10 3.78 3.98 4.29 3.93 2 Bihar Total 0.23 1.18 2.32 3.33 4.06 4.40 4.59 4.34 Rural 0.23 1.18 2.30 3.31 4.05 4.40 4.59 4.34 Urban 0.19 1.18 2.44 3.47 4.11 4.41 4.62 4.34 3 Gujarat Total 0.07 0.95 2.34 3.46 4.28 4.81 5.15 4.93 Rural 0.07 0.97 2.43 3.59 4.44 5.00 5.36 5.11 Urban 0.07 0.90 2.17 3.21 3.96 4.37 4.65 4.46 4 Haryana Total 0.15 1.20 2.68 3.94 4.94 5.74 6.21 6.08 Rural 0.16 1.30 2.79 4.10 5.17 5.97 6.46 6.36 Urban 0.10 0.93 2.36 3.43 4.22 4.93 5.26 5.06 5 Karnataka Total 0.20 1.16 2.44 3.47 4.31 4.71 5.07 4.67 Rural 0.22 1.23 2.53 3.55 4.38 4.78 5.11 4.66 Urban 0.15 1.01 2.25 3.25 4.13 4.53 4.97 - 4.71 6 Kerala Total 0.06 0.74 1.94 2.91 3.89 4.49 4.99 5.03 Rural 0.06 0.75 1.99 2.96 3.96 4.57 5.07 5.11 Urban 0.06 0.68 1.74 2.71 3.60 4.16 4.63 4.70 7 Madhya Pradesh Total 0.24 1.36 2.70 3.84 4.74 5.24 5.51 5.09 Rural 0.25 1.42 2.75 3.90 4.81 5.32 5.57 5.13 Urban 0.18 1.17 2.54 3.63 4.45 4.90 5.20 4.93 8 Maharashtra Total 0.16 1.11 2.35 3.31 4.10 4.50 4.96 4.96 Rural 0.18 1.20 2.47 3.35 4.26 4.80 5.15 5.12 Urban 0.12 0.97 2.17 3.05 3.78 4.12 4.49 4.53 9 Orissa Total. 0.16 1.23 2.52 3.51 4.25 4.59 4.75 4.32 Rural 0.16 1.24 2.52 3.51 4.26 4.60 4.75 4.33 Urban 0.15 1.16 2.56 3.52 4.17 4.50 4.71 4.15 10 Punjab Total 0.06 0.81 2.29 3.45 4.28 4.79 5.28 5.37 Rural 0.05 0.83 2 37 3.56 4.40 4.95 5.46 5.50 Urban 0.06 0.78 2.13 3.16 3.97 4.34 4.80 4.91 11 Rajasthan Total 0.18 1.25 2.71 3.95 4.93 5.55 5.90 5.63 Rural 0.18 1.27 2.73 3.99 5.00 5.64 5.99 5.73 Urban 0.17 1.22 2.65 3.78 4.66 5.16 5.51 5.23 12 Tamil Nadu Total 0.11 0.95 2.17 3.09 3.79 4.03 4.18 3.90 Rural 0.11 0.98 2.21 3.11 3.79 3.98 4.10 3.83 Urban 0.09 0.89 2.10 3.04 3.80 4.14 4.36 4.06 13 Uttar Pradesh Total 0.19 1.21 2.55 . 3.68 4.57 5.03 5.28 4.99 Rural 0.20 1.24 2.58 3.70 4 61 5.07 5.32 5.05 Urban 0.14 1.06 2.44 3.57 4.38 4.83 5.02 4.69 14 West Bengal Total 0.22 1.19 2.43 3.47 4.33 4.85 5.11 4.58 Rural 0.24 1.32 2.62 3.76 4.68 5.20 5.40 4.73 Urban 0.14 0.85 1.97 2.75 3.45 3.91 4.27 4.14 *Excludes Assamn. Source: Census of India 1981, Series -1 India / Paper-2 of 1983, "Key Population Statistics Based on 5 Percent Sample Data" 11 12 The officially reported couple-protection rate for Tamil Nadu 1983-1984 was 34.2%. Targets and achievements for that year were:3 Family Welfare Method Target Fixed Achievement Sterilization 399,000 496,780 IUD 168,000 55,847 CC Users 244,000 65,815 Oral Pill Users 90,000 11,877 Thus, as in India as a whole, in Tamil Nadu about 80% of acceptors were accounted for by sterilization. Just as fertility events and fertility-control events differ among the various states in India, so may they differ among the various districts which comprise each state. The Ammanpettai Program operates in the District of Thanjavur, where, for 1984-1985 these data are reported.4 Number of ECs - 684,379 Number of F. W. acceptors by method: Sterilization 167,792 IUD 8,960 CC 11,963 Oral Pill Users 2,930 TOTAL 191,645 Thus, as with India as a whole, about 25% of ECs have been sterilized. However, in Thanjavur District, the total acceptor rate is lower than for Tamil JSee Seven Years Progress of Health and Family Welfare Services in Tamilnadu Since 1977 Under the Leadership of Honorable Dr. M. G. Ramachandran. Chief Minister of Tamilnadu 1977-1984, Dr. H. V. Hande, Health Minister, Government of Tamil Nadu, Tenth Joint Conference of Central Council of Health and Central Family Welfare Council, at New Delhi 9-11 July, 1984, pp. 17 et seq. 4See E. C. Annual Statement, Proforma 11, Medical Services and Family Welfare Department, Thanjavur District. 13 Nadu as a whole, namely about 28% as compared with 34%, such that sterilization accounted for a higher percentage of all acceptors in this district (about 88%) than for Tamil Nadu as a whole or India as a whole (in each case about 80%). Thus, th.e Ammanpettai Program, designed to encourage use of temporary methods of contraception, in operating in Thanjavur District is operating in a district which, compared to Tamil Nadu as whole, has exhibited relatively poor performance in resort to temporary methods. Fertility Control: Plans and Prospects According to USAID DELHI 1985 (p. 19): "The task the Government has set for itself is thus a strenuous one. To reduce the birth rate to the low twenties by the end of the century will, by the Government's reckoning, require a doubling in the prevalence of contraception to 60% of eligible couples." The .Government has thus set an ambitious goal for India's family planning program. Some features of the fertility-control effort in India are such that the Ammanpettai Program will complement that effort in ways that will enhance the prospect for success pursuant to this goal. For example, one such feature is the conviction, in virtually every quarter, that in future, fertility control in India will have to rely far more upon recruiting acceptors of temporary methods and far less upon sterilization than has been true of the effort to date. Thus, according to USAID DELHI 1985, p. 18: "Most observers of the Indian program and the Government itself have long since recognized that the contraceptive mix will inevitably have to shift away from sterilization toward greater reliance on other methods since sterilization is approaching an upper bound among couples who by age and parity can be considered the market." 14 The Ammanpettai Program is, of course, designed precisely to encourage just such a shift in the contraceptive mix. Generally speaking, a number of features of the current social and economic order in India augur well for fertility regulation in the coming years. Thus, the demographic transition has already hit the urban middle class and shows evidence of generalizing not only in urban, but also rural areas where the representative rural family finds itself increasingly stressed by increasing pressure on land. More generally, the status of women in the Indian social order has in recent years been changing in ways that change attitudes with respect to appropriate family size, in favor of smaller families. If full advantage is to be taken of these and other trends, however, it will be necessary that India's family-planning program function efficiently. Unfortunately, in light of the track record of this program, one cannot be sanguine on this score. The program has exhibited various malfunctions which will need to be remedied if the government's ambitious fertility-reduction goals are to be achieved (see USAID DELHI 1985, pp. 24 et seq.). Among these malfunctions, and of peculiar interest for evaluation of the role of the Ammanpettai Program, is the finding (see USAID DELHI 1985, pp. 24,25): (3)Knowledde about contraception - several recent studies reveal surprisingly low levels of awareness of contraceptive, methods other than sterilization, and even lower levels of knowledge on how to use them. For example, less than half of currently married women living in rural areas know much about the common spacing methods.... While these findings are surprising after so many years during which the Government has claimed to have created awareness, if not use, of contraception, the situation nonetheless has its positive side. It suggests that bridging this knowledge gap through well- 15 designed energetically executed IE&C program may have a material impact on contraceptive use...." A promotional program of the Ammanpettai-type probably represents a far more powerful strategy for "bridging this knowledge gap" than even the most energetically executed IE&C program is apt to prove. "...There is an important need for the (Government) program to (1) increase its level of IE&C activities and (2) change the content of these activities by shifting the emphasis from why birth control is essential (most couples are now aware of the officially propagated advantages of small families) to how a small family can be achieved by methods less drastic than vasectomy, tubectomy and laparoscopy." An incentive program of the Ammanpettai type which gives participating mothers a direct, material and immediate reason to learn how to prevent pregnancy, represents what is in many ways the ideal response to the need to focus consumer education in this domain on the problem of how a small family can be achieved. It is also an ideal response to the problem of how to cope with and counteract pervasive misinformation in India among consumers concerning the various contraceptive methods. Another feature of the present family planning program in India of peculiar interest for evaluation of the role of the proposed incentive program is the extent to which family planning in India already incorporates incentives of one kind or another. Thus, sterilization and IUD acceptors are compensated monetarily with such payments accounting for about one-fifth of total program expenditure. More generally, according to USAID DELHI 1985 pp. 30-31: "The Government's program seems to have thought of everything...it does not pay family allowances; it has made donations for family planning purposes tax deductible; it offers incentives for the adoption of certain methods 16 of contraception and offers all methods at subsidized prices; the Central Government offers salary increases and low interest housing loans to its employees who are sterilized after two or three children; some states have reduced maternity benefits for its women employees after the birth of the second child; cash awards are made to the states that have the best record of family planning performance; it issues "green cards" to those who accept sterilization after two children which entitles the holder to various kinds of preferential treatment and benefits, including free lottery tickets. Individual states have sometimes supplemented these measures with some of their own devising, including facilitated school admission for the children of families in 'which one member is sterilized after two or three children'." Clearly, the family-planning. environment in India is incentive oriented and in this sense provides a natural home for a trial of an Ammanpettai-type program. Although the Government's program may seem "...to have thought of everything..." it at least up to now has not, to our knowledge, implemented an incentive program of the Ammanpettai-type. Should this program prove successful, the pro-incentive environment in which family-planning already takes place in India should augur well for the adoption of programs of the Ammanpettai-type more widely throughout the system. 17 Some Economic, Financial and Fiscal Implications of Ammanpettai- type Programs Introduction As was explained in the Introduction to this report, a family-planning program of the Ammanpettai-type might feature any of various formats depending upon the mix of operating parameters adopted. Some of these formats will be more efficient (cost effective) than others. Field trials will be required to determine the most cost effective program formats--and, in the longer run, it will be analysis of the economic, financial and fiscal implications of what have been determined to be cost effective program formats that will be of greatest interest. The small program as now implemented and operating in Ammanpetta has, of course, adopted certain parameters--values which seemed reasonable for such an initial trial (although, of course, it cannot be known without additional field trials whether these are the cost effective parameters). In any event, this analysis of economic, financial and fiscal implications will in part address the program now operating. Beyond this, however, this evaluation will address some aspects of economic, financial and fiscal implications for what might be regarded as a prototypical program of this type operated on a large scale. Since no such program is now operating, we will assume some values for certain parameters that such a program might feature. s a 18 The Fiscal Burden imposed by an Ammanpettai-Type Program The distinctive feature of an Ammanpettai-type program is the promotional family welfare payments for non-pregnancy. We may focus the analysis initially on this feature. In evaluating an Ammanpettai-type program, it is important to recognize that these payments are not an economic cost of the program, they do not represent use of real resources by the program, i.e., they do not represent a claim on real resources which might have been allocated to some alternative economic activity. Rather, these payments are transfer payments from whoever would pick up the tab for such a program (say, the tax payers) to the participating women who receive these incentive payments. This does not mean that these payments impose no burden, however. These transfer payments do represent an additional claim on fiscal capacity, and since fiscal capacity is always scarce relative to the many competing claims upon it, we should consider the fiscal burden imposed by these payments. The magnitude of this burden will, of course, depend upon the particular configuration of the program. Let us assume a program with incentive family welfare payments at the rate of Rs. 30/- per month (this is the rate for the program now operating in Ammanpettai). Another parameter critical for the fiscal burden imposed by the program is that of the scale of the program, the number of participant-years to carry in the program each year. The fewer this number, the less the burden on the fisc. But, the fewer this number, the longer it will take to give all of the eligible women an opportunity to participate in the program and have their fertility-control behavior modified thereby. (Program formats featuring the same participant-year loads may nevertheless differ with respect to other operating parameters which are important for program 19 performance. See the Introduction to this report for discussion of these matters.) Let us assume a program maintaining an enrollment throughout the year equivalent to 10% of the eligible women in the defined population--for purposes of this exercise, that of, the state of Tamil Nadu. Thcse eligible for the program, it will be recalled, are married women between the ages of 18-34 (with certain other qualifications). In the state of Tamil Nadu for 1981 there were about 6.7 million women in this age group.5 Enrolling 10% of this number, the program would have 670,000 participants. With incentive payments at the rate of Rs. 30/- per month, or Rs. 360 per year, the total incentive bill would come to about Rs. 240 million, or about Rs. 5.0 per capita for the population of Tamil Nadu. In evaluating this fiscal burden we may ask how it compares with fiscal capacity and related fiscal efforts in Tamil Nadu. The following table exhibits some relevant data.6 Consolidated Fund of Tamil Nadu Revenue Account Disbursements Account Name 1983-1984 Accounts (Rs. millions) 280 Medical 1,059.7 281 Family Welfare 207.4 Total Social and Community Services 8,386.5 Total Disbursements-Revenue Account 19,108.0 Excess of Revenue Over Disbursements 517.0 TOTAL RECEIPTS 19,625.1 bOr about 13.8% of the 1981 Tamil Nadu population of 48.4 million. See Census of India, 1981, Population by Quin-quennial Age Groups, based on 5% sample. 6Source: Annual Financial Statement 1985-1986, Budget Publication 22, Government of Tamil Nadu. 20 The Rs. 240 million fiscal burden which would be imposed by the instant version of an Ammanpettal-type program comes to about 1.2% of Tamil Nadu's total receipts (and expenditures) on revenue account and, as such, this burden can fairly be considered to be modest relative to Tamil Nadu's total fiscal capacity and total fiscal effort. The burden is still quite modest relative to total expenditures for Social and Community Services, it comes to about 2.9% of this amount. Expenditures for family planning are represented mainly by account 281 Family Welfare although a significant proportion of account 280 Medical, say 30%, should also be reckoned as expenditure on family planning.7 On this basis, we would have a total expenditure on family planning of about Rs. 525 million. As compared to this fiscal effort, the burden implied by the instant incentive program is relatively very large, namely about 46%. What does all of this mean for evaluation of the fiscal feasibility of an incentive program of the type here under consideration? Appropriate inquiries in the field in Tamil Nadu might turn up information to answer this question. Pending such inquiry, we make a few observations. Budget-making processes are notoriously rather inflexible over considerable periods of time such that large changes in the proportion of total operating expenditure going to any given activities or programs are hard to realize, particularly in the short run. From this point of view, the fact that the fiscal burden implied by the promotional program is large relative to state budget expenditures on family planning would ordinarily be regarded as an ominous sign not boding well for fiscal feasibility. This interpretation may be less warranted in this case than in the usual case, however. This is so because, although the family planning program in "According to the USAID DELHI 1985, p. 30, note 77 time allocation studies indicate that about 40% of the time of health workers is spent on family planning, which might amount to, say 30% of the health budget. 21 India is implemented at the local government level, the program is planned and largely funded from the center. Hence, the additional fiscal burden implied by the incentive program should be considered relative to fiscal capacity and fiscal effort on family-welfare account at the centre and, presumably, if the centre responded affirmatively to this fiscal challenge, it would be with funds earmarked for family planning such that rigidity in the state-level budget-making process would not present the usual problem. On the bright side here is the fact that Family Welfare Budgets in India have been increasing very rapidly in recent years, e.g., the planned average annual outlay for the period 1985-1890 at Rs. 7200.0 million is 2.6 times the average annual outlay for the period 1980-1985 (see USAID DELHI p. 29). In thinking about the additional burden these incentive payments would impose upon fiscal capacity, it is very important to recognize that if the program performs as anticipated and makes a significant contribution to reducing the fertility rate it will thereby also tend to reduce claims on fiscal capacity. Government expenditure for basic needs (such as education, medical, housing, sanitation, water supply, urban development and the like) are a function of the rate of population growth such that the slower population growth, the less rapidly expenditure for basic needs must increase to maintain given service levels. The basic-needs programs comprise most of those accounts aggregated under social and Community Services in Tamil Nadu's revenue account. As noted foregoing, Rs. 240 million of incentive payments comes to about 2.9% of the amount budgeted for these programs. Consequently, even a very small reduction in claims on fiscal capacity to fund the basic needs programs would recover a good bit of the cost of the incentive payments, i.e., the net burden of these payments is very probably much less than the gross burden. 22 All in all, one is tempted to the view that the Government of India might well be induced to make the relatively small additional fiscal effort called for by an incentive program of the type here proposed if a credible case were made that the program would have a significant impact on fertility. The only way to make a credible case that this program would have a significant impact on fertility behavior is to run field trials. This is, of course, just what we are attempting to do with the recently launched Ammanpettai Incentive Family Welfare and Child Spacing Program. Adding the Economic Costs So far the discussion of program costs has focussed just upon the incentive-transfer payments which, although representing the largest part of the program budget, are not to be reckoned as economic costs of the program. (See discussion on this point foregoing.) An Ammanpettai-type program will also, of course, incur economic costs. For the version of the program now operating at Ammanpettal, direct resource-use costs come to about Rs. 15/- per participant per month, i.e., about one-half of the transfer-payment outlay. (See budget on following page.) In calculating the fiscal burden that would result from generalization of an Ammanpettai-type program to the state of Tamil Nadu as a whole, the reader may increase the percentages already presented for the incentive-transfer-payment burden by half again to reflect the economic costs-- e.g., from 1.2% to 1.8% of Tamil Nadu's total receipts (and expenditures) on revenue account, and so on for the other percentage burdens. Some Performance Features of Family Planning Programs in India: Is An Ammanpettal-Type Program Apt to be Cost Effective? Whether an Ammanpettai-type program is apt to be cost effective depends, of course, not only upon program costs, but also program outputs. Pending the outcome of extensive field trials, one cannot speak with confidence 23 about the probable outputs of the Ammanpettai program. Nevertheless, an examination of some aspects of the performance of currently operating family- planning activities in India can help to cast some light on the question whether there exists a prima LacLe case, at least, than an Ammanpettai-type program is apt to be cost effective. In India, the Government's family planning program comprises a complex organization format which involves funds and actors at the federal, state and local government levels and from private groups. We do not attempt to afford a comprehensive description of this program here. Rather, we direct attention to certain features of this program which are directly relevant for evaluation of the proposed incentive program. What can -be said about unit costs for the family planning program as it now operates in India? Because of the multitude of actors and budgets involved in one way or another in this program, it is hard to assemble data on program costs in which one has confidence. Nevertheless, a few exemplary figures in this domain may be of some help in thinking about the relative cost effectiveness of an Ammanpettai-type program. As reported herein foregoing, for Tamil Nadu 1983-1984, the Family Welfare Budget was Rs. 207.0 million and for that year 630,319 acceptors were recruited. Thus these acceptors were recruited at about Rs. 330 (or U. S. $28) per head. If we include, say, one-third of the Medical Budget (on the grounds that these resources were allocated to family planning) in the outlay on family planning, the price per head to recruit acceptors goes to about Rs. 830 (or U.S. $69). These findings for Tamil Nadu are in line with findings for India as a whole reported elsewhere. For example, according to USAID DELHI 1985 (p. 30), in 1982-1983, taking account for just the Family Welfare Budget, acceptors 24 Ammanpettai Incentive Family Welfare and Child Spacing Program Budqet (Assume 300 Participants) Per Month Itemrn Amount (Rs) Personnel: Management and Admin. Project Director 1,200 Assist. Proj. Dir. (Project Supervisor) 500 Accountant, Part-time 50 Office Boy TOTAL Rs. 2,000/month ($167.00) Supplies Pregnancy tests @ Rs. 8/-x (say) 300 2,400 Office 200 POL 120 TOTAL Rs. 2,700 ($225.00) Incentive Payments to Participants @ Rs.30/mo.x 300 Rs. 9,000 Summary ($750.00) Direct resource costs Rs. 4,700 or approximately Rs. 15.7/month per participant. Incentive (transfer) payments Rs. 9,000 or Rs. 30.0/month/participant. NQIE: Additional resource costs not included in the foregoing accounting include: * Staff time for clinical activities, mainly administering the pregnancy tests and some counseling of mothers with respect to both f.p. and general health. * Facility and equipment costs. * In reckoning these costs, it must be kept in mind that both in the SMC and other ,similar facilities which might serve as a base for this kind of program, excess capacity of staff/facilities (owing to low utilization rates) may be such that the opportunity cost of these additional resources is very low. 25 were recruited at the equivalent of over U. S. $20 per head. And (see note 77), if an appropriate portion of the health budget is included in family planning costs, we find a cost per acceptor of the equivalent of U. S. $65. USAID DELHI 1985 remarks that the program: "...is even more high cost when looked at in t9rms of the number of births averted. Because very few births are averted among the older women where the bulk of the sterilization cases are concentrated, and because of the poor continuation rates recorded for other methods, the ratio of contraceptors to births averted is on the order of 6 to 1." According to these findings, under the present family planning program in India, it may cost at least Rs. 4,800 (i.e., 6 acceptors at Rs. 800 each) to recruit the acceptors necessary to avert one birth. It is perhaps a fair evaluation of the position to remark that an Ammanpettai-type incentive program would not have to perform very spectacularly to better this kind of track record. Counting both the incentive payments and the economic costs, the Rs. 4,800 would finance about 106 participant months--say five participants for about 20 months each, or 10 participants for about 10 months each, or some other configuration of numbers of participants and durations of participation adding up to 106 months. Since the Ammanpettai-type program not only recruits but also tends to commit the acceptors, it would seem plausible to anticipate that significantly more than one birth would be averted per 106 months of participant exposure to the incentives. In any case, the prim facie supposition on this score would appear 26 compelling enough to justify extensive field trials to test the cost effectiveness of the Ammanpettai-type program.8 bOther data on the cost of family planning in India yield similar conclusions. Thus, it is reported (see USAID DELHI 1985, p. 30, noted 76) that the cost per sterilization in 1982-1983 was about Rs. 570, say about Rs. 760 in current Rs. This would maintain a participant in the Ammanpetai program on incentives for more than two years. In attempting to form some initial judgments with respect to relative cost effectivenss of these two programs, one could start by comparing the average age and parities of participants in each. 27 Cost Effectiveness Where Program Budgets Entail Both Transfer Payments and Economic Costs An Ammanpettai-type program budget features a mix of economic costs and non-economic (transfer payment, budget) costs. In evaluating such a program to determine cost effective formats, there are two scarcities to take into account, i.e., scarce fiscal capacity which must be economized and scarce real resources which must be economized. There is an interaction between the scarcities. To command real resources to deploy, the program must use scarce fiscal capacity. Transfer payments likewise use scarce fiscal capacity, but they may also result in some saving in real resources. For example, with the incentive payments attracting participants in the Ammanpettai-type program and strongly motivating them to themselves learn how to avoid pregnancy, this kind of program might well save the resources which, in the existing program, are aflocated to IE&C activities, scores of "motivators" and the like. The transfer payment (incentives) and the real resources are both inputs to the production function which produces the desired output--namely, births averted. At the practical level, cost effectiveness evaluation of family planning programs is an uncertain business at best owing to uncertainties with respect to program production functions, i.e., uncertainties on the score of what inputs how configured will produce what outputs in the form of births averted. As the foregoing discussion suggests, cost effectiveness evaluation of family planning programs which use both transfer payments (for incentives) and real resources as inputs entails additional conceptual problems. The government is choosing between programs which entail different mixes of resource costs and transfer- payment costs. In deciding which mix will produce any given output (number of births averted) at least cost, there are two genre of costs to take into account, 29 Ammanp2ttai Program Progress Report: De-ember 1985 Enrollment and attrition data are as follows: Number in program since 2/85* 343 Moved out of catchment area 20 Pregnant 14 Elected permanent sterilization 4 (two M; 2 F) Dropped because over age 3 Religious objection 1 Died (natural causes) 1 Total attrition 43 Remaining in program 300 *55 women were deliberately dropped from the program after participation for six months (February-August 1985). They are not included in the figure above, but are being followed-up at six month intervals in the absence of monthly incentives. See attached. Evaluating the pregnancy record The table on the following page exhibits relevant characteristics for the 14 pregnancy cases. These are the attrition group which must be examined from the point of view of program success or failure. Of these cases, eight (all but one enrolled 8/85) made only one initial visit to the program and then left the program because they self-identified themselves as pregnant (i.e., rather than because they presented and failed the pregnancy test). The circumstances strongly suggest a high probability that these pregnancies occurred shortly before the first clinic visit and hence before participation in the program began (our pregnancy test does not become positive until 45-60 days after conception). In interpreting the pregnancy record, it will be far less misleading 29a Ammanpettal Incentive Family Welfare and Child Spacing Program Progress Report: December 1985 Beneficiary Pregnancy Record Beneficiary Date Enrolled .# visits Age Children Age youngest I.D. Number child 61 8/85 1 21 2G/13 2.0 69 " 1 22 10/13 2.5 108 3/85 7 20 10 2.5 124 8/85 3 21 1G 0.5 144 * 1 25 2G 1.5 152 9/85 4 20 1G 0.5 159 3/85 7 27 23/10 1.5 208 * 6 22 1G 0.8 248 2/85 7 23 2G 1.3 259 6/85 1 24 iG 1.5 271 8/85 1 30 /lB 3.0 273 * 1 24 10/13 .0 275 * 1 26 G 3.5 282 1 27 23/10 3.0 30 - to exclude these eight cases than it would be to include them in the record. This would leave six pregnancies which resulted from about 150 beneficiary-years of beneficiary participation. Of these six pregnancies, three occurred in women with only one female child and one in a beneficiary with two female children. Experience with the program has been too brief to generate findings which can give a reliable indication of program success in averting births. Nevertheless, and while recognizing this, one can at least say that the track record to date augurs well for the future of program performance. Particularly notable has been the manifest capacity of the program to introduce large numbers of women to the use of modern temporary methods of contraception who had had no prior history of such contraceptive practice and who, moreover, were drawn from a population in which only on the order of 2.0% of eligible couples were acceptors of temporary methods. The Participants Who Were Dropped from the Program In August 1985 29 women recruited from the village of Poondi and 26 women from Thanjavur, all of whom joined the program in February 1985 were dropped from the program because they were from outside the catchment area. They were urged to continue with their contraceptive use and both pills and condoms continued to be provided free of charge by the clinic or multipurpose worker at local level. On January 17, 1986 all of these women were asked to come back to the clinic for interview and pregnancy test. It was agreed to pay Rs. 30 to each if not pregnant and Rs. 15 to pregnant women (to defray cost of journey and loss of day's work). Forty-three women reported to the clinic, one of whom was found to have a positive pregnancy test (she had removed her IUD two months previous (cramping) but not taken new precautions). Two women who were absent were reported by worker to be still receiving pills monthly. Three women have moved out of area and we have no follow-up. Seven have 31 been dropped from the program because of pregnancy and did not come in. Of these seven, one had made only a single initial visit to the program. Forty-two of the original 55 or 76% of women introduced to contraceptives in February 1985 and who received monthly Rs. 30 through July 1985 have continued on contraceptives for the six months since incentive payments were discontinued. This was ascertained by direct interview of all 42 (by J. Stevens) and through discussion with the local health worker. Women Who Apply to Participate But Cannot be Accommodated Since the inception of this program in February 1985, many more eligible women have applied to join the program than can be accommodated with existing resources. Up to our visit in January 1986, these women were told they could not unfortunately be included. Following a series of discussions with clinic and project staff during January, we decided to offer all women who apply for admission to the incentive program, free access to the same contraceptives and counseling as our 300 currently enrolled beneficiaries. Names and biodata (age, ages and number of children, previous contraceptive use, income, education) will be recorded exactly as for beneficiaries and these new women will be followed to determine how many are ready to start contracepting without the incentive payment. This list of women may be used for replacement of dropouts from the incentive roster on a serial or lottery basis. An additional advantage of the program deserves attention. The monthly visits by these women to the clinic accompanied in most cases by their youngest child has given us an unparalleled opportunity to provide regular health surveillance, education and disease prevention and treatment. Thus malnutrition, infections, severe parasitosis and in some cases correctable severe neurological and medical conditions for which no help has been sought 32 have been observed during these regular visits to the family welfare clinic. Appropriate treatment and preventive strategies have been implemented.