NUTRITION-RELEVANT ACTIONS

I - DIRECT NUTRITION INTERVENTIONS
II - FOOD INTERVENTIONS
III - POVERTY ALLEVIATION PROGRAMMES

I - DIRECT NUTRITION INTERVENTIONS

A - The Integrated Child Development Services Scheme (ICDS)
B - The Tamil Nadu Integrated Nutrition Programme (TINP)
TINP-II
C - The Midday Meals Programme
D - Vitamin A Prophylaxis Programme
E - Anaemia Prophylaxis Programme

Direct nutrition interventions currently operational in the states of Andhra Pradesh and Tamil Nadu are summarized in Tables 2 and 3 (objectives, components, beneficiaries, coverage and costs). The following sections discuss the design and implementation of each of these programmes in some detail.

A - The Integrated Child Development Services Scheme (ICDS)

The ICDS is probably the largest programme of its kind in the world. It aims to achieve four objectives:

i) to improve the health and nutrition status of children 0-6 years by providing supplementary food to beneficiaries 300 days per year and by coordinating with state health departments to ensure delivery of required health inputs;

ii) to provide conditions necessary for child psychological and social development through early stimulation and education;

iii) to enhance the mother’s ability to provide proper child care through health and nutrition education;

iv) to achieve effective coordination of policy and implementation among the various departments to promote child development.

Initiated in 1975 on an experimental basis in 33 blocks, ICDS now covers around 2200 out of a total of 5500 rural blocks in India. (Blocks have an average population of 110,000. While a majority (92% of "fully operational" projects in September 1991) of the ICDS projects in the country are sponsored/funded by the central Government of India (GOI), through the Department of Women and Child Development (WCD) in the Human Resources Development Ministry, some of the states have initiated state-sponsored ICDS projects. The GOI and the states share ICDS costs. The GOI provides training and operating costs including salaries, equipment, supplies, play materials, petrol and oil expenses, and medical kits, estimated at Rs. 0.9 million per block-year. State governments meet the costs of supplementary food, currently estimated at around Rs. 1.7 million per block-year. As of September 1991, Andhra Pradesh had a total of 141 "fully operational" ICDS projects - 132 centrally sponsored and 9 state-sponsored projects - covering a population base of about 14 million. Corresponding figures for Tamil Nadu are 81 (all sponsored by the central government) and a population of about 8 million.

Table 4 summarizes the implementation of ICDS in Andhra and Tamil Nadu with reference to national averages. Using the percent of projects operational to those sanctioned as an indicator of the state government’s commitment to ICDS, Andhra Pradesh has a distinct advantage over Tamil Nadu, as well as over the national average. However, both Tamil Nadu and Andhra Pradesh fall much below the commitment level exhibited by other states such as Gujarat (96%), Karnataka (95%), Haryana (98%), and even Madhya Pradesh (95%). Quality of implementation, as assessed by percent functionaries (anganwadi workers or AWWs) appointed and trained, is superior in Tamil Nadu as compared with both Andhra Pradesh and the national average.

The GOI has followed a gradual ICDS expansion policy, based on two kinds of targeting: first to the most disadvantaged areas and, second, within them to vulnerable pre-school children and pregnant and nursing women. The initial geographic focus was on tribal, drought-prone areas and blocks with a significant proportion of scheduled caste population. The programme is also targeted towards malnourished children, but in practice most beneficiaries of supplementary feeding are not selected through nutritional screening. Selection criteria generally depend upon a combination of individual worker’s perceptions of which are the poorest households and overall quotas which the Central and individual state governments set for supplementation. Thus there tends to be more area than individual targeting in ICDS.

According to estimates², Andhra Pradesh and Tamil Nadu have about 2.64 and 2.40 million poor preschool children respectively. ICDS covers about 1.03 million children in Andhra and 0.44 million children in Tamil Nadu³. These figures indicate that even with perfect targeting to the poor households (using poverty as a criterion), ICDS could cover only about 39% of the needy children in Andhra and 18% in Tamil Nadu. Corresponding figures for the country as a whole are close to 50.6%.

². Total population in the two states is estimated at 63.3 and 55.6 million in the 1991 census. According to the Planning Commission, poverty incidence (1987-88) is 31.61% in AP and 32.80% in TN yielding an estimated 20.0 and 18.2 million poor people in the two states respectively. Assuming that preschoolers constitute 13.2% of the total population, this works out to 2.64 million poor children in AP and 2.40 million in TN.

³. The Department of Women and Child Welfare reports that in September 1991, 0.90 million children in Andhra Pradesh and 0.41 million preschoolers in Tamil Nadu were covered under the ICDS. After adjusting for the number of centres reporting (87.2% in AP and 92.6% in TN), the numbers of children covered is estimated at 1.03 and 0.44 million respectively for the two states.

TABLE - 2
MAJOR NUTRITION RELEVANT ACTIONS IN ANDHRA PRADESH AND TAMIL NADU

		Beneficiary groups covered
Program	Implementing agency (s)	Preschool children	Women	School-children	Others	STRATEGY/OBJECTIVES
DIRECT NUTRITION INTERVENTIONS
1. Integrated Child (ICDS) Development Services	Women and Child Welfare	0-6 Years	Pregnant and lactating women			Reduce child-malnutr. through integrated nutrition/health/preschool education inputs
2. Mid-Day Meals Program (MDM)	Education			2-14 Yrs		Reduce malnutr & increase school attendance through supplementary feeding in schools
3. Tamil Nadu Integrated Nutrition Program (TINP)	Dept. Social Welfare	6-36 Months	Pregnant & lactating women			Reduce malnutr. through integrated nutr/health & strong growth-monitoring & IEC inputs
4. Vitamin A Prophylaxis Program	Health and Family Welfare	1-5 years				Reduction of Vit A def through oral mega-dosing
5. Nutritional Anemia Prophylaxis Prog	Health and Family Welfare	1-6 Years	Preg. & lact. women		Sterilized women	Reduction of anemia through iron/folic acid supplements
6. Goitre Control program	Civil Supplies				Gen. population	Iodized salt distribution to reduce IDD
7. Mobile Food & Nutr. Extension Education	Food and Nutrition Board				Women	Nutrition/health education to improve dietary practices
HEALTH INTERVENTIONS
1. Expanded program of Immunization (EPI)	Health and Family Welfare	0-6 Years	Preg. Women			Morbidity reduction through immunization
2. Diarrhoeal Disease Control Program (CDD)	Health	0-6 years				Diarrhea-control through ORT & education
3. School Health Services	Health and Family Welfare			6 - 14 yrs.		Provision of improved health facilities
4. Maternal & Child Hlth Program (MCH)	Health & Family Welfare	-	Pregnant women			Provision of improved MCH facilities to reduce IMR & MMR

TABLE 3
COVERAGE AND COSTS OF MAJOR NUTRITION-RELEVANT ACTIONS IN ANDHRA PRADESH & TAMIL NADU

INTERVENTION	ANDHRA PRADESH		TAMIL NADU
	COVERAGE	ANNUAL OUTLAY	COVERAGE	ANNUAL OUTLAY
1. ICDS	0.402 Mill 0-3 Yrs* 0.501 Mill 3-6 Yrs* 0.194 Mill Preg/Lact*		0.1 85 Million 0-3 Yrs* 0.223 Million 3-6 Yrs* 0.077 Mill. Preg/Lact.*
2. MDM	NA (Discontinued in 1985)		1.8 Million Preschoolers 6.4 Million School children	1950 Million Rs (1989-90)
3. TINP	NA	NA	0.64 Mill 6-36 Mon children (1988) 0.26 Mill Preg/Lact women (1988)	130.6 Million Rs (1988-89)# (984.5 Million Rs for 1980-89 #)
4. VIT. A PROPH	1.1356 Million$		I Dose: 0.32 Mill (1987-88) II Dose: 0.373 Mill (1987-88)
5. ANEMIA PROPH	1.635 Million children$ 1.479 Million women$		0.2427 Million Children 0.1281 Million Preg Women
6. GOITRE CONTROL			None
7. F & N Extn Edu.			NA	NA
HEALTH:
1. EPI/UIP	$1.64 Mill DPT, 1.45 Mill Measles, 1.70 Mill BCG, 1.64 Mill OPV, 1.08 Mill DT, 1.06 Mill Tetanus children 1.78 Mill Tetanus pregnant women		1.03 Million Tetanus, 1.14 Mill DPT, 1.14 Mill Polio, 0.826 Mill BCG & 0.86 Mill Measles	6.493 Million Rs (1988-89)
2. CDD			NA	2.286 Million Rs (1988-89)
3. School Hlth.			0.69 Million Children	7.13 Million Rs
4. MCH			0.714 Million Mothers regd 0.410 Million deliveries	8.787 Million Rs (1986)

SOURCES:

i) Data for Tamil Nadu extracted from: Children and women in Tamil Nadu, A situational analysis, 1990, UNICEF.

ii) * Extracted from Dept. Of Women & Child Dev. Status Report on ICDS, Sep 1991.

iii) * From Shekar, 1992.

iv) $ Communication from Directorate of Health & Family Welfare, Govt of Andhra Pradesh, 1992 (Data for 1990-91).

Several other unpublished studies that have examined the issue of community involvement reinforce this (Agarwal & Lata; Gandhi; Paranjpe & Bhagwat; Kumar, Prakash & Lal; Dev & Lal; Sharma & Chand: all quoted in Punhani & Mahajan, 1989). The only form of community contribution noted is in terms of buildings for village centres, a trend relatively more typical in rural than in urban or tribal areas. Community involvement could be strengthened through more regular home visits by the AW. The task of eliciting community participation is admittedly difficult in villages stratified by social barriers and economic differentials; therefore, health and nutrition activities are rarely conducted in women’s working groups, which were envisaged as a major vehicle for community participation.

There is considerable variability not only in ICDS impact but also in the quality of services provided. Both training and supervision can play an important role in quality assurance. In particular, four training areas are highlighted in evaluations as requiring strengthening: i) there is considerable unevenness in the training imparted by over 200 centres spread all over the country; ii) despite good training manuals, the training syllabus and materials need to be adapted to suit differing rural, urban and tribal conditions; iii) supervisory training is very weak. Consequently, although many supervisors are aware of ICDS deficiencies in their sector, vigorous actions to remedy the situation through more frequent and more extended visits to AWs are not undertaken. The need to supervise a large number of AWs and inadequate mobility also limit their attention to quality issues; iv) existing funding places inadequate emphasis on in-service training of all categories of functionaries.

ICDS sets feeding quotas per block which usually are fully subscribed. Most AW nutritional screening is to identify severely underweight children, who are entitled to food supplementation but because of listlessness and lack of appetite rarely can consume a double ration. Once enrolled, most child beneficiaries continue to receive food regardless of nutritional status until they reach school age. ICDS maintains with some justification that in locales where poverty levels approach 75%, such as many tribal blocks, virtually all pre-school children are at nutritional risk and warrant continued supplementation. However, the cost-effectiveness argument for area targeting rather than individual nutrition screening is weaker in ICDS blocks where poverty levels are substantially lower.

The present supplementation system has other drawbacks. Once a feeding quota is filled, the AWW has little incentive to seek out additional malnourished children for other AW services. Area targeting thus tends to reduce the emphasis on monitoring individual child growth and can detract from worker focus on case management of malnourished children through health check-ups and referral. Evaluations have indicated that long-term supplementation may substitute for food which the child otherwise would receive at home and thus run counter to the development of maternal understanding of the special needs of malnourished children and improved family food behaviour. On the other hand, although the direct impact of the present feeding programme on child nutrition may be less than desirable, it may induce mothers and children to come to AWs (NFI 1988).

Given the documented inadequacies in programme implementation, monitoring and supervision, the likelihood of a favourable impact is small. Further, the studies that do suggest an improvement in nutritional status (such as Tandon, 1983 & 1984 and the Punhani & Mahajan, 1989 review of several studies) are based on pre-post designs which fail to account for improvements in nutritional status that may be attributable to factors other than ICDS programme inputs such as secular/temporal trends, changes in socio-economic conditions, and impact of programmes other than ICDS. Where non-ICDS control groups are used, little attention is paid to the comparability of the two groups, and corrections for non-comparability (such as differences in socio-economic status) are not accounted-for in research design or appropriate statistical analyses. Furthermore, none of the evaluation studies reported were double-blind studies, thus leaving the results open to observer bias. The NFI study (1988) suggests a possible reduction in severe underweight prevalence rates in almost all states, an increase in the ‘percent normal’ in Tamil Nadu and Haryana, and a decrease in the ‘percent normal’ in Andhra Pradesh, Jammu & Kashmir and Karnataka among ICDS beneficiaries between 1981 and 1986. Nevertheless, the study cautions that the apparent improvement may be an artefact of the change in the populations captured by the ICDS, rather than an improvement in the nutritional status of beneficiaries - an argument, which if tenable, negates the ‘observed impact’ of all the before-after studies mentioned above.

The most recent (July 1992) national evaluation of the ICDS reports that "minor differences were noticed in the nutritional status of children from ICDS and non-ICDS areas" among under-three year old children suggesting minimal impact on nutritional status. The impact among older children (3-6 years) is reported to be "more prominent", but no statistical tests are reported to support this hypothesis. Further, impact is reported to be highest in urban areas followed by rural and tribal areas. However, the study states that "...it may be noted that ICDS areas continued to have a large percentage of malnourished children". The numbers "malnourished" are: 64.5% 0-3 year olds and 64.1% 3-6 year olds in ICDS areas as compared with 68.9% 0-3 year olds and 70.7% 3-6 year olds in non-ICDS areas (all classifications based on the Indian Academy of Paediatrics (LAP) classification which is a modification of the Harvard classification).

Subbarao (1989) has reviewed evidence from several sources to arrive at two cogent conclusions - first that available evidence cannot measure ICDS impact accurately, and second, that there is enormous inter-project (and therefore inter-state) variation in impact which is not apparent in aggregate analyses.

District-wise ICDS coverage data from Andhra Pradesh indicate totally inadequate targeting in relation to poverty criteria (Figure 9). The Pearsons correlation coefficient for percent of households below the poverty line in 1987/88 and the percent (of total eligible) beneficiaries covered under ICDS is -0.68 p < 0.001; the Spearman rank correlation being similar at -0.67. Thus, ICDS coverage (area targeting) seems to be in inverse proportion with ‘need’ as identified by poverty criteria. Unfortunately, anthropometric data are available only for the eight NNMB-monitored of the twenty two districts. For these eight districts, the correlation between underweight prevalences and ICDS coverage is relatively high (Pearsons r = +0.59; Spearman’s rank order correlation = +0.43). It seems probable that the "sentinel districts" selected by the NNMB for nutrition monitoring may have been selectively picked up for ICDS implementation. This review concludes that ICDS may have contributed to some improvement in the nutritional profile of young children, but in the absence of convincing data on such impact, conclusions must remain tentative.

Figure 9. ICDS AND POVERTY - AREA TARGETING IN A.P.

By and large though, the basic design of the ICDS is sound in as much as it is designed to target the most vulnerable groups. Nevertheless, a greater emphasis (in both design and implementation) on the younger 0-3 year age group will help tighten access of ICDS services to those most needy. Where ICDS needs major reorientation is in beneficiary and area targeting (to better match poverty ratios and nutritional criteria where these are available on a reliable and regular basis), in area-specific planning and flexibility, and implementation quality issues such as training, supervision and support. If ICDS efforts could be better targeted to those most ‘nutritionally needy’, and if the delivery of care inputs (growth monitoring and information-education-communication) were to be strengthened, perhaps a much greater impact could be expected.

These efforts would need to be complemented with a greater flexibility in programme implementation to cater for local needs - a direction that would also facilitate the much-needed sense of community involvement and empowerment. Such flexibility at the grass-roots level could be introduced through several innovations.

For example, ICDS workers should spend more time in the community before starting a new ICDS project/centre. This time should be spent in identifying location/village-specific problems that could be addressed by the ICDS functionaries, in collaboration with and with the active participation of local leaders. Innovative techniques such as Participatory Rural Appraisal (PRA) could be used during this mutual-familiarization process. A small discretionary fund could be made available for actions to be initiated as a response to the "familiarization exercise". For example, if the PRA exercise in a village suggests that the most pressing need is water, ICDS workers should have the flexibility to respond to this need, perhaps even before any mention of nutritional concerns. Once some positive steps have been made in this direction (perhaps ICDS can coordinate with the relevant water department to solve the problem), then and only then should nutritional issues be introduced. Such an approach will help to make ICDS more relevant to local needs, help improve the credibility of the ICDS worker and the programme, and increase community participation. Operational research in this context is essential to test the feasibility of such approaches.

Tamil Nadu has recently introduced a new innovation - the ICDS programme in one rural district (Puddukottai) has been merged with the Noon Meals Programme. This ICDS-NMP merger has resulted in a two-worker ICDS model whereby one worker is responsible for nutrition and health activities, and the second one for the pre-school education activities. Specific evaluations of this new model are not yet available.

To sum up, ICDS has succeeded in reaching around 40% of rural India, mainly the poorest areas, an achievement in itself. ICDS is well-conceived: provision of an integrated package of health and nutrition services through village-based AWs has considerable potential for improving child nutrition and health in India, and promote early childhood development. However, available studies suggest that ICDS impact varies a great deal from block to block. Its training and supervision, health linkages, referral services, mother counselling and nutrition education, coverage of under-threes and degree of community participation all need to be strengthened for ICDS to achieve its full potential. The key issues to be addressed to improve the impact of ICDS on child development are how to ensure that (a) the whole range of ICDS services is provided to the targeted beneficiaries, particularly, under-threes and pregnant women, (b) nutrition supplementation does not become largely a substitute for a part of the home meal, (c) health and nutritional competence of the families is increased accompanied by better weaning and child feeding practices, and (d) coverage of necessary health services is increased, referrals for malnourished children and at-risk pregnant and nursing women are completed and severely underweight children are nutritionally rehabilitated.

B - The Tamil Nadu Integrated Nutrition Programme (TINP)

The Tamil Nadu Integrated Nutrition Project (TINP), a World Bank/IDA assisted health and nutrition intervention offers a package of health and nutrition services to young children and pregnant and lactating women in rural Tamil Nadu. The first TINP project has been followed up with a second generation TINP-II (unless otherwise specified subsequently TINP refers to TINP-I).

TABLE 4
IMPLEMENTATION OF ICDS IN ANDHRA PRADESH AND TAMIL NADU

NO. PROJECTS			% STAFF APPOINTED			% STAFF TRAINED			COVERAGE
SANCTIONED	OPERATIONAL*	REPORTING*	TO SANCTIONED			TO APPOINTED			TOTAL	CHILDREN	WOMEN
			CDPO	SUP	AWW	CDPO	SUP.	AWW	POP$	(0-6 yrs)
									(Mill)	(Millions)	(Millions)
ANDHRA PRADESH 169	141 (83.4% of sanctioned)	123 (87.2% of operational)	47	45	64	71	74	82	14 (21.1%)	0.90	0.19
TAMIL NADU 111	81 (73.0% of sanctioned)	75 (92.6 of operational)	56	63	61	94	80	74	8 (14.4%)	0.41	0.08
INDIA 2424	1962 (91.8% of sanctioned)	1986 (83.0 of operational)	55	59	71	80	86	71	242 (28.7%)	12.9	2.7

Notes:

1. # Figures in parentheses are percentages to total as per 1991 census.

2. * Refers to "Fully operational" projects and "Projects reporting" respectively as reported by Dept. of Women & Child Welfare.

3. $ Estimated approximately @100,000 population for each project sanctioned.

4. Performance figures pertain to the projects reporting, and not to total operational projects and will therefore need to be scaled upwards by the % reporting.

5. State and Centre sponsored projects are clubbed together in this table.

6. CDPO is Child Development Project Officer, SUP. is Supervisor and AWW is anganwadi worker...functionaries of the ICDS scheme.

Source: Computed from Dept. Of Women & Child Welfare, Status Reports on ICDS, Sep 30 1991.

TABLE 5

IMPACT OF THE TAMIL NADU INTEGRATED NUTRITION PROJECT

MEAN WEIGHTS OF BENEFICIARY CHILDREN: NOV 1980 VERSUS NOV 1986

AGE	NOV’80			NOV’86			DIFFERENCE	P value*
(Months)	MEAN WT.	STD.DEV.	N	MEAN WT.	STD. DEV.	N	NOV’ 86 - NOV 80
	(Gm)			(Gm)			(Gm)
6	5157.5	1065.8	20	5915.1	616.6	83	757.6	.006**
7	5884.6	1124.0	76	6178.2	601.5	71	293.6	.049*
8	6179.4	1431.8	27	6516.5	806.1	85	337.0	.252
9	6239.5	1146.4	76	6772.0	764.5	75	532.2	.001**
10	6271.4	1106.8	29	6722.2	858.5	54	450.8	.063
11	6596.9	975.6	48	7208.2	936.4	73	611.3	.001**
12	6905.8	1021.3	39	7297.9	752.3	73	392.2	.039*
13	6982.0	1065.5	82	7434.1	1019.5	64	452.0	.010**
14	7159.5	1113.6	21	7569.4	851.0	80	409.8	.128
15	7082.4	1104.2	202	7900.5	970.3	94	818.1	.000**
16	7198.4	1327.0	16	7717.6	891.0	91	519.1	.150
17	7302.9	573.5	17	7971.0	1268.1	69	668.1	.002**
18	7452.2	1250.0	46	8101.3	1105.6	76	649.1	.005**
19	7796.4	1459.6	56	8228.6	994.5	84	432.1	.056
20	7440.9	661.4	11	8486.3	901.3	51	1045.4	.000**
21	7934.1	1119.5	210	8550.8	886.9	63	616.6	.000**
22	8126.9	1239.7	13	8548.6	923.7	69	421.6	.262
23	7543.7	1084.1	8	8608.2	996.0	73	1064.5	.028*
24	8188.9	881.0	9	8852.6	1008.6	76	663.7	.060
25	8548.6	1312.4	93	9031.0	1220.3	71	482.3	.016*
26	8947.1	1388.7	17	8893.8	1142.8	81	-53.2	.884
27	9003.3	1181.1	289	9167.8	875.0	107	164.5	.134
28	10061.9	1312.6	21	9460.0	1087.6	115	-607.9	.059
29	9390.9	1250.1	22	9689.9	1094.9	99	299.0	.308
30	10083.3	1712.8	24	10022.4	1074.2	105	-60.9	.869
31	9496.9	1373.5	65	9927.4	953.8	93	430.5	.031*
32	9980.0	1066.3	5	10124.7	1062.4	81	144.7	.781
33	10148.5	1355.0	185	10114.1	1110.7	64	-34.5	.840
34	10967.9	1336.2	53	10568.5	1106.2	73	-399.4	.078
35	9930.9	1958.3	36	10651.9	1080.3	53	721.0	.056
36	10768.1	1513.8	51	10389.5	1091.3	36	-378.6	.164

* Differences significant at p <.05
** Differences significant at p <.01

SOURCE: Shekar, 1992.

- to halve malnutrition among children under four years of age;
- to reduce infant mortality by 25%;
- to reduce vitamin A deficiency in the under fives from about 27% to about 5%; and
- to reduce anaemia in pregnant and nursing women from about 55% to about 20%.

Official evaluations of the TINP (Govt. of Tamil Nadu, 1989) have suggested a favourable impact of the programme on nutritional status of beneficiaries. The order of impact attributed to TINP is between one-third to a half reduction in prevalence of severe malnutrition among 6-24 month olds, and a reduction of about 50% in severe malnutrition among 6-60 month olds. However, data and analyses to support this claim are weak. A recent assessment by Shekar (1991) provides more convincing evidence for programme impact. The study shows a statistically significant improvement in nutritional status of programme beneficiaries over time of implementation of the programme. Further, as evident in Table 5 the nutritional impact is more pronounced among children less than 24 months of age - precisely the age that needs to be targeted on the basis of "nutritional need". Plausibility analyses (Shekar 1991) add to the conviction that the observed improvement in weights of young children can be attributed to TINP inputs rather than to competing explanations such as secular trends and overlap of beneficiaries with the Noon Meal Programme. Recent unpublished data from the NNMB (NIN 1991b) which show a greater improvement in nutritional status of young children in TINP versus non-TINP areas lend further credibility to the conclusions of a favourable impact (see Table 6).

Table 6: Changes in Percentage of Under-Five Children Below 75% NCHS weight-for-age, between 1979 and 1990 in TINP and non-TINP districts of Tamil Nadu

District	1979	1990	Difference
TINP
Chengulpet	66.2	48.5	17.7
Ramanathapuram	70.5	58.7	11.8
Tirunelveli	67.6	48.3	19.3
N. Arcot	64.1	55.0	9.1
Madurai	70.2	48.6	21.6
Non-TINP
The Nilgiris	67.4	58.7	8.7
Thanjavur	62.4	59.5	2.9
Kanyakumari	56.5	44.2	12.3

Source: NIN 1991b

The weakest component of the TINP has been its efforts at eliciting community participation (Shrimpton, 1989) - most likely a consequence of the top-down approach in programme implementation, still a characteristic of most development programmes in India.

Also, the project failed to reach its health goals. Although infant mortality declined by 12 to 26 percent in different areas of the project, similar declines also occurred in non-project areas. Just under a quarter of all eligible children received full doses of vitamin A, while 28% of children had not received even one. Half of all eligible pregnant women had not received anaemia prophylaxis.

TINP demonstrates that it is possible to reach a high proportion of younger children who are nutritionally the most vulnerable, and significantly to reduce the prevalence of severely underweight children through well targeted health and nutrition services aided by communication and community mobilization activities. While it did not achieve all of its goals, the project has an unusual number of lessons for the design and implementation of outreach programmes, particularly in the area of training, supervision and monitoring. Key features include carefully defined recruitment criteria for local workers; limiting field worker tasks to those which are manageable and high priority; specification of daily and monthly work routines; decentralized training systems; supervisory practices which facilitate on the job training; the use of local women’s groups to support project activities; the display of performance information to clients and workers at the village nutrition centre; and a management information system which could rapidly detect performers falling below established norms. On the health side, the main lesson is that large scale investment in health infrastructure and supplies is not sufficient to improve performance. Complementary software measures are needed to optimize health workers’ performance. With a few design changes in the supplementation criteria, and more focus on maternal nutrition and improved health-nutrition coordination, it may be possible to reduce the prevalence of the moderately (grade II) underweight, resulting in a greater proportion of children in normal and grade I category.

Comparison between ICDS and TINP is difficult; ICDS covers children under six years of age and includes pre-school education whereas TINP focuses on children under three with nutrition and health services only. But several of the observed deficiencies of ICDS - the relative failure to reach under threes, the poor health and nutrition education, and the neglect of home visits and community participation - can be remedied using systems similar to those in TINP.

Costs per beneficiary (omitting pre-school education, and hence comparable) in TINP have been estimated to be close to half those incurred in the ICDS (Dapice 1986, quoted in Berg 1987, p124). The total financial outlay in 1988-89 for TINP was over Rs 130 million with a beneficiary coverage of 0.64 million children in the 6-36 month age group and 0.24 million pregnant and lactating women averaging about Rs 145 per beneficiary per annum. Berg (1987) mentions a cost per beneficiary of US$ 9.41 for 1984-85, with about 30% of the expenditure on food and the remaining 70% on other recurrent costs.

Given the features of TINP design and implementation discussed above, and the documented evidence for impact, it is concluded that the TINP was successful in reducing nutritional deprivation among young children in Tamil Nadu. A major part of this impact is likely to have occurred via improved ‘care’ factors, complemented by some improvement in access to health facilities. This beneficial effect could, nevertheless, be amplified by strengthening community-involvement in the programme.

TINP-II

TINP-II is designed to cover, in a phased manner, 316 of the total 385 rural blocks in Tamil Nadu, with an estimated total population of 32.8 million. The target group has been extended to encompass young children from birth until six years of age (as against 6-36 month old children in TINP-I). Furthermore, in recognition of the duplication (geographic and age-group) of the services of TINP-I and the Noon-Meal Programme, TINP-II has planned for a merging of TINP and NMP centres, to promote complementarity. TINP-II centres are being opened in select non-TINP-I areas, while existing TINP-I centres are simultaneously being converted to TINP-II centres. TINP blocks will be converted to ICDS depending on governmental allocations for ICDS. A total of 9194 Community Nutrition Centres currently function under TINP-II, covering 98 new blocks, over and above the 177 TINP-I blocks. An additional 5257 centres are to be opened in the coming phases. An IDA loan of US $ 95.8 million spread over an eight year project implementation period (1991-98) has been approved for the second Tamil Nadu Nutrition Project. The specific objectives of TINP-II include:

- reduction of severe malnutrition among 0-36 month old children by 50% in new project areas, and 25% in TINP-I areas;

- increasing the proportion of children classified as "normal" by 50% in new and 35% in TINP-I areas;

- contribute to a reduction in infant mortality to 55 per thousand live births;

- contribute to a 50% reduction in incidence of low birth weights.

- the target age group is 0-60 months (as against 6-36 months);

- the addition of pre-school education under TINP-II;

- the merger of TINP with NMP, so that two field-level workers will be present at each TINP-II centre. The first will be a TINP designated worker, essentially catering to the health and nutritional needs of beneficiaries, while the second worker will be responsible for pre-school education activities;

- changes in selection criteria for supplementary feeding, thus diluting the relatively tight beneficiary targeting for feeding that was successfully implemented in TINP-I;

- greater emphasis on the communications component, with special attention to developing and operationalizing a specific communications strategy;

- greater emphasis on health service delivery, including the referral system, and structural changes to facilitate this;

- strong emphasis on coordination between health and nutrition service delivery;

- increased outreach and coverage of beneficiaries;

- changes in the training plan to cater to the new programme design.

Worker training seems to be lagging behind TINP-II phasing-in, so that in many of the newer project areas, the essence of TINP-I is missing. Without the planned training, there are several anomalies relating to the duties of the NMP worker and the TINP worker who have simply been instructed to function from one centre. The two workers are not clear about the demarcation of duties, and/or the changes in the objectives or design of the TINP. The coming years will unfold the actual fate of TINP-II.

C - The Midday Meals Programme

The Midday Meal Programme (MDM) or school lunch programme operating in some of the states in India has a dual objective of improving nutritional status and school attendance concurrently. A small nutrition education component is also supposed to be included, but is rarely implemented. Beneficiaries are provided with one meal, which supplies about one-third of the daily requirement of energy and protein, either at school or at preschool centres.

Tamil Nadu has a massive school lunch programme commonly referred to as the "Chief Minister’s Noon Meal Programme" (NMP). The NMP emerged as a populist scheme with little attention to prioritization of resources. Children between the ages of 2-14 years are fed daily through 63,000 NMP centres, under the scheme at an expense of Rs 0.44 - 0.90 per beneficiary, and a total cost of Rs 2200 million (in 1984 i.e. about 10% of the state’s budget and 20% of its annual expenditure in the sixth five year plan). Expenditures in the seventh plan (1985-90) were estimated at about 1750 million Rs per annum and the coverage is estimated at 6.4 million school children in addition to 1.8 million preschoolers - an average expenditure of Rs 213 per beneficiary per annum. UNICEF (1990) reports an outlay of Rs 1950 million for 1989-90.

While some efforts have been made to evaluate and assert the positive nutritional impact of the scheme (Devdas 1987), the results do not stand up to scientific scrutiny. Several lacunae in the design of the programme point towards the unlikelihood of any significant impact on nutrition. The programme caters only to school children thereby excluding the poorest who cannot attend school, and the target age group does not include the most vulnerable 0-2 year olds. Two other counter arguments are relevant here. First, much of the grain resources for the NMP come from those diverted from the Public Distribution System, and often the rural works programmes (discussed in later sections), thus potentially reducing the poor’s access to food through the PDS. This may also be associated with the fact that the meal provided at school often becomes a substitute rather than a supplement. Second, a major part of the revenue for the programme was generated by the state government from liquor revenues - a not so desirable promotion for the liquor industry in Tamil Nadu. Harriss (1991) concludes from a case study that "the scheme reaches the poor, but not all poor, probably not the poorest, and not only the poorest children."

Andhra Pradesh’s flirtation with a midday meals programme (MDM), as with other states, was short-lived. This programme was initially introduced in a few deprived areas in 1980, was extended by the then new state government to cover all school children in Andhra in grades one to five in 1982-83. Expenditure on the programme rose from Rs 3.1 million in 1981-82 to 292.4 million in 1982-83 accounting for nearly 97% of the total direct expenditure on nutrition support in that year. In 1983-84, total expenditure on the midday meals programme and PDS together in Andhra Pradesh added up to over 12% of the annual plan outlay (Subbarao, 1989) - a level of expenditure that could not be sustained. The MDM, initiated as a populist political gesture, without adequate prioritization and policy formulation, was consequently withdrawn in 1985.

The nutritional benefits of the MDM in Andhra have not been evaluated, but given that the scheme was untargeted, and, as in the case of Tamil Nadu, that it would have diverted food grain from PDS and rural works programmes, and severely constrained the state budget, the sum toto benefits are likely to have been minimal, if any. In conclusion, the chances of the NMP having any nutritional impact are negated by the following:

- there is no targeting so that "all children between the ages of 2-14" attending balwadis/schools are covered. Clearly the opportunity costs of reaching the most needy (also the ones most likely to show a nutritional impact), are extremely high;

- only children attending balwadis/schools are covered by the programme. But the most needy group is least likely to be found in balwadis/schools, so that this strategy in fact purposively excludes the most needy;

- the food given in the school/balwadi is given as a substitute rather than as a supplement to the home meal;

- the costs per beneficiary of NMP are nearly double those of comparable programmes.

Other feeding programmes operational in India at various times include the Balwadi Nutrition programme, the Creche Nutrition Programme, and the Special Nutrition Programme. These together make a negligible contribution to state-level figures owing to coverage of small population groups.

D - Vitamin A Prophylaxis Programme

Prevalence of mild vitamin A deficiency in the world ranges between 20-40 million cases at any one time, nearly a half of which is in India. Other sources (Ministry of Health and Family Welfare, undated) report a 5-7% prevalence of "eye-signs" of vitamin A deficiency among children in India, while NNMB (NIN 1991) (which covered eight states in the country) reports a 0.7% incidence of Bitot’s spots among children in 1988-90, the figures being 1.0% for Andhra Pradesh and 0.6% for Tamil Nadu. WHO’S cut-off for identifying a public health problem is 0.5% thus identifying both states as vitamin A deficient.

The Government of India has initiated a two-pronged approach to combat vitamin A deficiency in India:

i) Fortification of vegetable oils. It is mandatory by law for all vegetable oils marketed in India for human consumption, to be fortified to the level of 25 IU retinol per gram of oil. 60% of the vitamin A utilized in the country is used for fortification of vegetable oils or animal feeds. However, in view of the low level of consumption of vegetable oils by poorer/vulnerable sections, much of this fortification benefits the less vulnerable sections of the population.

ii) The National Prophylaxis Programme for Prevention of Blindness due to Vitamin A Deficiency. This was initiated by the government in 1970, to target children 1-5 years of age. A recent review of the situation in 1989 has led to the inclusion of 6-12 month old children with a single dose of 100,000 IU of retinol, linked with the Universal Immunization Programme (UIP). The prophylaxis programme comprises a long-term and a short-term strategy. While the short-term strategy focuses on administration of prophylactic mega-doses of vitamin A periodically, the long-term strategy aims to improve dietary intakes as the ultimate solution to the problem. The four major thrusts of the programme are:

- Promotion of regular consumption of dark-green leafy vegetables or yellow fruits and vegetables;

- Promotion of breast-feeding and colostrum to protect against vitamin A deficiency;

- Oral prophylactic doses of vitamin A as follows: one dose of 100,000 IU to infants 6-11 months, and six-monthly doses of 200,000 IU to children 6-60 months;

- Treatment of vitamin A-deficient cases by administrating: a single oral dose of 200,000 IU of vitamin A immediately at diagnosis, and a follow-up dose of 200,000 IU 1-4 weeks later.

The prophylaxis programme is implemented through Primary Health Centres and sub-centres. Prophylactic doses of vitamin A, (supplemented with nutrition advocacy) are administered by para-medical staff manning the PHC. In areas where the ICDS is in operation, vitamin A administration is conducted under the auspices of the ICDS. Records of administration of doses are kept in registers/weight cards/health cards maintained by ICDS functionaries. In Tamil Nadu, vitamin A prophylaxis (and nutrition education) is implemented through the ICDS and the TINP in areas where these programmes are operational. There has been a successful health education component of TINP directed to vitamin A prophylaxis.

Assessments by the Ministry of Health and Family Welfare (1988) claim that 85% of the target for vitamin A prophylaxis was met in 1987-88. However, these estimates of programme performance seem unrealistically optimistic especially in view of the limitations in supply and logistics of delivery of the prophylactic dose. Further, none of the assessments pertain to information/education/communication efforts. One programme review found a "low" level of awareness of the prophylaxis programme and it’s benefits among health workers and the general public.

In Tamil Nadu, coverage in 1987-88 is reported at 0.32 million children for the first dose and 0.37 million children for the second bi-annual prophylactic dose (UNICEF, 1990). In Andhra Pradesh, coverage was reported at 0.37 million children in 1985 (Rao et al, 1988). More recent data (Govt of Andhra Pradesh 1992), estimate coverage in Andhra Pradesh at 1.14 million children i.e. about 14% of the total preschool population in the state.

E - Anaemia Prophylaxis Programme

Incidence of anaemia in India is high despite a near-adequate iron intake as discussed in Part I. Inadequate intake and absorption of dietary iron, and the high prevalence of parasitic infections are the primary causes of anaemia in India. The anaemia prophylaxis programme is targeted to young children and pregnant and lactating mothers through the health delivery system or ICDS/TINP. Coverage rates are low, compared to targets - about 1.64 million children and 1.48 million women in Andhra in 1990-91, and 0.24 million children and 0.13 million women in Tamil Nadu in 1987-88 (See Table 3). Other problems with the programme, recognised in the Ministry of Health and Family Welfare/UNICEF 1989 workshop report, included irregular and short supply of iron tablets to distribution centres, inadequate training of health workers and awareness among beneficiaries (MOHFW/UNICEF 1989).

II - FOOD INTERVENTIONS

The Public Distribution System (PDS)

The Public Distribution System (PDS)

The PDS in Andhra Pradesh
The PDS in Tamil Nadu

The PDS is the major food-subsidy/income transfer programme in the country, administered by the Ministry of Food and Agriculture. The per capita energy consumption of the lowest three deciles has been shown to be much below the recommended norms in all states, the situation being exaggerated in the chronically food deficit states, and having deteriorated over time between 1961 and 1983 as judged by results reported by the National Sample Survey Organization (Gupta 1987). (NSSO results are based on consumer expenditure patterns and not on food intake). This situation persists despite national food security. Inequitable access to food due to structural poverty, exacerbated by seasonal variations, underscore the need for an effective household food security system, based on targeted food distribution.

The PDS provides essential food grains (and sometimes other additional items), to poor households at subsidized rates. It is a three tiered network with the Food Corporation of India (FCI) as a national agency, wholesale at the State/District level, and Fair Price Shops (FPS) at the retail level. Regarding the middle level, each State has its own trading corporation, either working on its own account (e.g. Tamil Nadu) or acting as an agent of FCI (e.g. Andhra Pradesh). Allocations of subsidized food are made to the PDS by the central government, these allocations being boosted by the states’ own efforts in procuring grain for the PDS. At all-India level, the Fair Price Shops increased from 48,000 in 1960 to about 350,000 in 1990, with increasing de-centralization, covering almost 85% of population. As the number of FPS increased substantially, the amount of food grain distributed through PDS also increased from 13 million tonnes in 1980 to 18 million tonnes in recent years.

The main objective of the PDS system is to make grain available according to need. A recent study however on the utilisation of the PDS by NSSO showed that 44% of the population depend on PDS for kerosene as opposed to only 14% for rice, for example (Sarvekshana 1990). A higher percent of rural than urban population purchase their requirements from sources other than PDS, implying the PDS is tilted more towards urban consumers. Non-availability of the items is one of the major reasons for not purchasing or part-purchasing. As reflected in Figure 10, ironically, the poorest states account for the smallest shares of PDS food supplies from the central government, while a disproportionately large share is allocated to relatively prosperous urban areas. In regions covered by the PDS overall consumption undoubtedly increases, but this may be at the expense of equally poor unincluded households in regions not supplied by the PDS, where consumers pay higher prices directly as a result. Further, past efforts at either household or individual targeting have been minimal, if non-existent. This is particularly relevant in the context of the remarkably high subsidy cost of PDS at over Rs 7500 million per annum. However, both Tamil Nadu and Andhra Pradesh have experimented with targeted variations of the PDS each of which is discussed below.

Figure 10. PDS AND POVERTY - By State

Data: 1984 PDS, 1983-84 hh poverty

Experience with the PDS in some states has shown that while it may be possible successfully to tighten targeting to a great extent, "perfect targeting" may be unrealistic in practice. But, recent governmental efforts in structural adjustment have necessitated a reduction in subsidies, and a consequent shift in policy (though action on this issue has yet to materialize) towards better area and beneficiary targeting through the PDS. The possibility of commodity targeting (using coarse grains) is also now seriously being considered by the central government (Bhatt 1991a).

The PDS in Andhra Pradesh

Andhra Pradesh, with 7.6% of India’s poor, receives 12% of the total subsidized food grains distributed under PDS by the central government (Subbarao 1989). The state has a loosely targeted PDS wherein eligible households are designated as "Green or Yellow card holders" based on income criteria. According to recent estimates (Ministry of Civil Supplies 1991), of all households allocated ration cards, some 75% had the preferential green cards while the balance 25% had yellow cards in 1990. Corresponding figures for 1983 were 80% and 20% respectively, implying some slight tightening in the targeting criteria over the years. Before January 1992, rice, the principal food component of the existing PDS, was supplied to green card holders at Rs 2 per Kg and to yellow card holders at Rs 3.50 per Kg.

The major criticism of the PDS in Andhra is the loose targeting despite shortages in PDS supply so that present coverage of the poorest groups is relatively small. Another problem is the fact that the scheme is socially regressive since rice is disbursed twice a month while the poor subsist from day to day. It is clear from data in Figure 11 that poverty incidence and PDS resource allocation are unrelated. Thus, a situation already demonstrated at state level is also evident at the district level. Further, area-targeting did not improve between 1983 and 1989. The Pearson’s correlation coefficient between percent of households below the poverty line (1987/88) and percent households issued PDS ‘green cards’ (1989) is +0.34 (p < 0.13) with the Spearman’s rank correlation for the same data at +0.30. When data for the eight districts surveyed by the NNMB are analyzed, PDS targeting however appears extremely efficient (Pearsons’s r = + 0.76, Significant at p < 0.03; though Spearman’s r = + 0.38). These districts seem to have received specialized attention for PDS targeting, possibly as a consequence of their selection by the NNMB as indicator districts. Ironically, the correlations are negative (Pearsons r = -0.09, Spearmans r = -0.25) though statistically non-significant, when ‘nutritional need’ (as identified by child underweight prevalences available for the eight selected districts) is used instead of poverty indicators. Correlations between ICDS targeting and PDS targeting are also poor (Pearsons r = +0.06 p <0.81; Spearmans r = -0.01), suggesting a probable lack of convergence of area-targeted services offered by the two potentially complementary programmes.

Radhakrishna and Indrakant (1987) have shown the beneficial effects of the PDS on welfare levels of the poor, while also illustrating that this beneficial effect could have been much greater if targeting had been limited to the lowest 40%. This level of targeting would also ensure greater sustainability and cost-effectiveness. In 1983-84. the cost of rice subsidy was estimated at 10% of the total tax revenue of Rs 9650 million. Tax revenue is estimated at Rs 32,580 million in 1992-93, and the cost of the rice subsidy is estimated to be 25% of this - a level that is likely to be unsustainable in the current economic situation. This lack of sustainability is reiterated by the January 1992 announcement by the state government to reduce the rice subsidy and increase the cost of PDS rice from Rs 2 per kilogram to Rs 3.50 per kilogram, and to restrict access to a maximum of 16 Kg (as compared with 20 Kg earlier) per green card family per month.

Figure 11. PDS AND POVERTY - AREA TARGETING IN A.P.

In 1989, about 2.1 million tonnes of rice were distributed through the PDS in Andhra Pradesh. Subbarae (1989) has estimated that if PDS was to meet the entire ration quota of only the poorer green card holders, about 2.5 million tonnes of rice would have been required in 1988 - a less than 25% increase in supplies. On the same lines, current resources would be adequate if access is restricted to the poorest 60 percent - a feasible level of tightening - given that poverty level are estimated at 31.6 percent in the state.

In summary, the existing PDS system in Andhra Pradesh fulfills some of the need for food security among the poor. To that extent, the PDS is likely to have contributed towards protecting nutrition within beneficiary households. Nevertheless, there is tremendous scope for enhancing impact by tightening targeting criteria (area and beneficiary targeting), for including nutritional criteria in PDS targeting (thereby reducing leakages to the non-needy), and for tying in complementary PDS and ICDS inputs in areas where poverty and malnutrition co-exist.

The PDS in Tamil Nadu

Tamil Nadu receives less than it’s fair share of grain for the PDS (see Figure 10) from central allocations, so that much of the state’s grain for PDS is made available through state-level efforts. With 7.4% of the all-India poor resident in Tamil Nadu, the state receives only 4% of the total grains mobilized under PDS by the central government. This is in direct contrast with Andhra Pradesh, which accounts for 7.6% of the poor and yet is apportioned 12% of the total PDS grain. This inequity in PDS allocations is further underscored by Kerala which has only 2.6% of the poor but over 16% of the PDS grain.

Total food grain channeled through the PDS (state and central efforts together) was 2.14 million tonnes in 1985 (Bulletin Food Statistics, 1986). The targeting criteria used in the Tamil Nadu PDS system is outlined in Table 7. Of the 11.8 million PDS entitlement cards in operation in the state, approximately 8.6 million (73%) are issued to households with incomes less than Rs 6000 per annum (Annadurai, undated), who receive food and other entitlements at prices that are between 40-60% of the prevailing market prices. An estimated 85% of the retail PDS outlets are now located in rural areas (UNICEF, 1990). It should be noted that there is a major need for PDS to operate effectively in urban, as well as rural, areas. If the urban population were to depend totally on the market for food, prices would rise, possibly out of reach of many of the urban poor, and food would be sucked in from rural areas. However, as elsewhere, appropriate targeting to the poor in urban areas is crucial.

While the Tamil Nadu PDS system has been rated as among the best targeted in the country, there are still some limitations. While the area-targeting and general outreach of PDS to poor and remote villages is good, the beneficiary-targeting is clearly less so. The fact that 73% households qualify as ‘low-income’ may have as much to do with under-reporting of incomes as poverty (which demonstrates the difficulty of accurate means assessment through income estimation). Other limitations based on micro-level studies, discussed by Harriss (1991), include: the constrained availability of food grains through the system - the major factor governing access to supplies - and the insensitivity of the system to the needs of large households with the largest dependency ratios. Further, in times of shortages, the poorest households were also the ones most likely to lose access to PDS supplies through the mortgage of their entitlement cards.

Currently, area-targeting of PDS is based on poverty criteria only. With the central government’s new focus on revamping the PDS to make it more effective, the following suggestions are tendered:

i) 1700 blocks have been selected by the government (on poverty criteria) for launching the "revamped PDS". It is suggested that nutritional data be superimposed on poverty data to identify further blocks for inclusion. Blocks with the highest poverty levels as well as the highest preschool underweight prevalences should be priority areas for the revamped PDS. Nutritional data for this exercise could be made available from the records of the Ministry of Welfare, ICDS monitoring cell (or the TINP in Tamil Nadu). The ICDS monitoring cell collects monthly nutritional status data from ICDS blocks all over the country. Despite some inadequacies in these data, they could be used for this initial exercise. In fact, this demonstrated use of nutritional data for policy and planning may help to reinforce the need for better nutritional data for the future.

ii) Efforts should be made to link the revamped PDS with other developmental programmes such as the ICDS/TINP, to promote complementarity of inputs (i.e. PDS attacks household food insecurity while ICDS/TINP deal with the ‘care’ factors). The suggestion in (i) above may be the first move in this direction. The possibility of other innovations should also be explored. One example is the allotment of an extra PDS quota to families certified by the ICDS/TINP worker as having a "malnourished child" in the family.

iii) Greater use of coarse cereals in the PDS to promote self-selection of the poor, since the rich are less likely to purchase coarse grains. However, this must be accompanied by great caution, so that the food commodities selected are acceptable to the poorest in that region/area.

III - POVERTY ALLEVIATION PROGRAMMES

A - The Integrated Rural Development Programme (IRDP)
B - The National Rural Employment Programme (NREP)
C - The Rural Labour Employment Guarantee Programme (RLEGP)

The declining association between agricultural growth and poverty in the post-Green Revolution era presents a strong case for poverty alleviation programmes. Two kinds of poverty. alleviation strategies are evident in India. The first directed through asset-investment, resource development and self employment, and the second through wage-employment schemes and/or rural works programmes. The important schemes in each of these two areas are described below, before available evidence for impact is reviewed and the balance weighed between the two strategies in the current socio-economic environment.

A - The Integrated Rural Development Programme (IRDP)

The IRDP, initiated in selected districts in India in 1978-79, was rapidly extended nationally during the sixth Five Year Plan in 1980-81. The programme, targeted at rural families below the poverty line, is designed to provide a capital subsidy and complementary credit at low interest rates to finance productive investments in income generating assets.

Subsidies for asset acquisition range from 25% for small farmers, 33.3% for marginal farmers, agricultural labourers and rural artisans, and as much as 50% for scheduled castes and tribes. Block-level staff select potential beneficiaries in consultation with the village council (Gramsabha), help them select viable investments (such as animal husbandry, agriculture, horticulture, weaving, handicrafts etc), and provide back-up support when needed.

In the sixth Five Year Plan, IRDP coverage (including coverage of scheduled castes/tribes) exceeded targets. However, resource allocations between states were not based on incidence/intensity of poverty, and even the investment per beneficiary varied tremendously among states - partly as a consequence of inadequate credit mobilization in many cases. While both Andhra Pradesh and Tamil Nadu performed slightly above the national average in credit mobilization, states with large tribal populations performed poorly.

In the seventh Five Year Plan (1985-90), the IRDP was expected to cover 20 million beneficiaries, of which 30% were supposed to be women. Based on experience in the sixth plan, several modifications in the IRDP were proposed, including a greater involvement of voluntary agencies and peoples representatives, a higher investment per Beneficiary to facilitate adequate income generation, decentralized planning to ensure better coordination between other sectoral projects and IRDP, establishment of forward and backward linkages for the investments promoted, and improving the role of financial institutions.

The data provided in the mid-term appraisal of the Seventh Plan (GOI 1988) outline the progress of IRDP between 1985-88. While implying a fairly adequate coverage of targets, these data however do not highlight some of the important implementation issues ~ such as mode of selection, degree of targeting, incremental incomes accruing, etc - which vary widely by state. The following can be gleaned from the GOI 1987 evaluation data, based on a survey between October 1985 and September 1986:

i) Targeting was generally good with a majority of the beneficiaries (95% in A.P., 86% in T.N.) concentrated in the less than 3,500 Rs/annum income category;

ii) Using the level of decentralization of beneficiary selection as a criteria for assessing community involvement, Andhra Pradesh did much better than the national average - 87% beneficiaries being selected by the local Gram Sabha, as compared to 55% nationally. By contrast, community involvement was almost non-existent in Tamil Nadu, with only 1% selection at local level.

iii) The average assistance given to beneficiaries ranged between Rs 2,930 to 4,035 on a sliding scale - ironically, with the assistance per beneficiary increasing with increasing income i.e. in direct contrast with ‘need’. This, despite the fact that the poorest groups appeared to have relatively high potential for generating income from the IRDP assets as reflected in the Incremental-Capital Output Ratios (ICOR). Again, inter-state variations are large, but both Andhra Pradesh and Tamil Nadu seemed to be less inequitable in the provision of assistance than either the national average or, states such as Madhya Pradesh, Maharashtra and Jammu and Kashmir.

Table 8: Impact of IRDP on beneficiaries

Study	% hhs that crossed poverty line	% hhs that received incremental incomes
Institute for Financial Management and Research	NR	84
Reserve Bank of India	17*	51*
National Bank for Agriculture and Rural Development	47	82
Programme Evaluation Organisation (GOI)	49	88

* Incomes of beneficiaries at current prices were discounted by 27% to arrive at real incomes at 1981 prices.

Source: Bandopadhyay (1989) as quoted in Gaiha (1990)

District-wise IRDP coverage data for Andhra Pradesh are presented in Figure 12. The significantly negative correlation between district-wise incidence of household poverty (1987/88) and percentage of beneficiaries targeted under IRDP (1991) for the 22 districts in Andhra Pradesh is a clear indication of the perverse targeting that has evolved wherby IRDP services seem to be concentrated in better-off areas (Pearson’s r = -0.76, p <0.00; Spearman’s rank order correlation is similar at -0.85). The correlations are slightly lower when data are analyzed separately for the eight districts selected by the NNMB for nutrition surveys (Pearson’s r = -0.67, p <0.07; Spearman’s r = -0.57). Inter-district allocations for IRDP are clearly divorced from poverty criteria, allocations being in inverse proportion to need. Correlations between district-level ICDS and IRDP targeting for the 22 districts are high in Andhra Pradesh, suggesting a convergence of services offered by the two complementary programmes in non-poor areas (Pearson’s r = +0.64, p <0.002; Spearman’s r = +0.74). With respect to IRDP, such non-poor areas are more likely to be well supplied with credit institutions with the capacity to absorb credit to the extent envisaged - one factor possibly contributing to the mismatch.

Figure 12. IRDP AND POVERTY - AREA TARGETING IN A.P.

National data reviewed here indicates that the IRDP may have had some impact on the incomes of the poor as reflected in Incremental-Capital Output Ratios and the percentage increase in family income - 38% families increased their incomes by over 50% in the year from October 1985 (GOI 1987) and only 16% reported no increase in incomes. This indication of positive impact is further reinforced by the high level of "investments remaining intact" even in the relatively poorer states. Nonetheless, evidence for supporting the claim for "large scale success" is thin. Further, in Andhra Pradesh, the irrationality of IRDP area targeting does not auger well for plausible impact.

Better area and beneficiary targeting to reach the poorest income groups, improved implementation and back-up support and aftercare, an increase in investments per beneficiary, reduction of leakages due to corruption, and greater involvement of the community, possibly through decentralized administration, are some of the changes that are immediately warranted to facilitate greater impact in the future. Further, as Kakwani & Subbarao (1990) point out, "for sustained income generation, it would be necessary to ensure continued access to institutional credit for the very poor". For this, it is suggested, access should be continued for a period beyond a one-time grant, especially for the very poor, and it should be accompanied by access to guidance and support. Stronger management information systems for programmes such as IRDP would also help to reduce the corruption associated with the programme⁴. If such credit schemes are genuinely to reach the poorest, they may well benefit from the experience in Bangladesh with the highly successful Grameen Bank - a public sector credit institution, providing loans to the poor on reasonable terms and with an excellent repayment record (see description in Basu 1991 p361-3)

⁴ Anecdotal evidence does exist to support this. For instance, loans may be sanctioned using another person’s name to avoid repayment; or the beneficiary may only receive 40-50% of the sanctioned loan, the rest being usurped by middlemen.

Two other schemes that have evolved out of the IRDP need mention here - the scheme for Training of Youth for Self Employment (TRYSEM) and the DWCRA (Development of Women & Children in Rural Areas) which started in 1979 and 1982-83 respectively. A review of the TRYSEM suggests that the scheme had a low impact in poorer states with low potential demand for services, while in better-off regions the percent of youths trained to those self employed was high (Subbarao 1989) suggesting that the poorer states may need alternative interventions. The seventh plan mid-term appraisal of the DWCRA notes inadequate implementation of the programme.

B - The National Rural Employment Programme (NREP)

The NREP initiated in 1980 to replace the food-for-works programmes, aimed to "generate additional gainful employment for the unemployed and underemployed persons in rural areas, to create productive community assets for direct and continuing benefits to poverty groups and to strengthen the rural, economic and social infrastructure to bring about a general improvement in the overall quality of life in rural areas. It also aims to improve the nutritional standards of rural poor through the supply of food grains as part of wages" (GOI Seventh Plan Mid-Term Appraisal, 1988).

The percent distribution of employment under the NREP closely parallels the state-wise percent distribution of the ‘ultra-poor’ (see Figure 13). Kakwani & Subbarao (1990) calculate a rank correlation of r = +0.74 (significant at p < 0.01) between the distribution of the ultra poor and the man-days of employment generated. Nonetheless, the states of Tamil Nadu and Andhra Pradesh (as well as Uttar Pradesh, Rajasthan, Kerala & Karnataka) seem to have received a greater share of NREP employment than would be justified by the percent of the ‘ultra poor’. The eastern states of West Bengal, Assam and Bihar (the more needy) have received less than their fair share.

Figure 13 - EMPLOYMENT DISTRIBUTION UNDER NREP (1987-88)

Though the mid-term appraisal of the seventh plan suggests that achievements have surpassed targets, some deficiencies are evident. Andhra Pradesh and Tamil Nadu are the only two states that have employed over 40% women under the NREP, the national figures being about 20%. Women’s participation is the lowest in the poorest central and eastern states. Furthermore, women’s wages are discriminately lower than men’s wages (which are lower than the minimum wage). Participation of Scheduled castes and tribes and that of the rural landless is also estimated to be low (between 17-33%) (Gillespie, 1988). The aim to provide one kilogram of food grain per person per day as part of the wages was not achieved, the average figure being 0.45 kg/day, presumably because of inefficiencies in supply of food grains channelled through the PDS. There is also a mis-match between the grains demanded (rice & coarse grains) and the grains supplied (predominantly wheat). The nature of rural works undertaken under the NREP revealed a strong bias towards construction of rural roads, schools and other infrastructure likely to benefit the ‘less poor’.

An annual report of the Department of Rural Development has noted that "employment provided under NREP is of short duration, often sporadic, and hence fails to generate enough income to create a visible impact on the living conditions of the poor". The lack of a clear understanding of the programme as a transitionary one, and inadequate planning to coordinate infrastructural requirements with NREP works are some of the other major shortcomings of the programme. The more recently observed move towards decentralized planning and implementation may help in reducing some of these weaknesses.

Kakwani & Subbarao (1990) while conceding the deficiencies of the NREP, summarize the main desirable features of employment programmes i.e. the self-targeting (through relative unattractiveness of the remuneration rates for the less poor); and the capacity to substitute for a social security system, at least for those able to work. They profess that if deficiencies are tackled, and if women are attracted to work sites (as has been done in both Andhra Pradesh and Tamil Nadu), these programmes "can reach out to the poorest fifth more readily than most alternatives". This thesis is supported to varying degrees by many others (Gillespie, 1988; Gaiha, 1991; Bhatt, 1991b; Seventh Plan, 1985-90). Further, if NREP area targeting within Andhra Pradesh is as poor as the IRDP targeting (a plausible assumption, given that the two programmes are implemented by the same agency), chances of having reached the poorest are low, the only saving grace being the built-in beneficiary targeting whereby, even in the more prosperous districts, only the poorest would be attracted to the NREP.

It is this self-selection mechanism, providing for a built-in targeting to the poorest which gives wage employment programmes a greater nutritional potential than asset-endowment schemes like IRDP.

Bhatt (1991b) argues for a guarantee of at least 100 days of employment to the most needy, rather than the provision of 50 days work to 40 million households. This would require at an all-India level, an annual financial outlay of Rs 60,000 million (2000 million person-days at unit cost of Rs 30 per person-day). While governmental concern in this area is high, recent allocations of Rs 6,300 million (1989-90 plan outlay), Rs 21,000 million for 1991-92 and Rs 18,250 million for 1992-93 fall far short of this.

C - The Rural Labour Employment Guarantee Programme (RLEGP)

The RLEGP was launched in 1983-84 to provide an employment guarantee to at least one member of every landless labour household up to 100 days in a year and create durable assets for strengthening rural infrastructure. Programme design and implementation is almost identical to the NREP, and most of the issues discussed above are relevant here as well. Employment targets have been overshot as in NREP, but the ‘guarantee’ has not been implemented, because of non-feasibility in small trials. The RLEGP was merged with the NREP in the 1989-90 annual plan. According to recent press reports, the 1992-93 annual plan has dropped all employment guarantee schemes from the Central Planning Budget. However, Tamil Nadu has announced an employment guarantee scheme similar to the widely acclaimed Maharashtra Employment Guarantee Scheme. A sum of Rs 100 million has been allocated for this scheme which is to be initiated in four districts i.e. Dharamapuri, Puddukottai, Pasumpon and Ramanathapuram (The Times of India, Feb 8 1992). The scheme also envisages payment of a small dole (Rs 2/day) to each registered person for whom employment can not be provided.