Thongkong et al.

International Journal for Equity in Health (2017) 16:48

DOI 10.1186/s12939-017-0539-5

RESEARCH Open Access

How equitable is the uptake of conditional

cash transfers for maternity care in India?
Evidence from the Janani Suraksha Yojana
scheme in Odisha and Jharkhand
Nattawut Thongkong1,3* , Ellen van de Poel1, Swati Sarbani Roy2, Shibanand Rath2 and Tanja A. J. Houweling3

Abstract
Background: In 2005, the Indian Government introduced the Janani Suraksha Yojana (JSY) scheme - a conditional
cash transfer program that incentivizes women to deliver in a health facility – in order to reduce maternal and
neonatal mortality. Our study aimed to measure and explain socioeconomic inequality in the receipt of JSY benefits.
Methods: We used prospectively collected data on 3,682 births (in 2009–2010) from a demographic surveillance system
in five districts in Jharkhand and Odisha state, India. Linear probability models were used to identify the determinants of
receipt of JSY benefits. Poor-rich inequality in the receipt of JSY benefits was measured by a corrected concentration
index (CI), and the most important drivers of this inequality were identified using decomposition techniques.
Results: While the majority of women had heard of the scheme (94% in Odisha, 85% in Jharkhand), receipt of JSY
benefits was comparatively low (62% in Odisha, 20% in Jharkhand). Receipt of the benefits was highly variable by district,
especially in Jharkhand, where 5% of women in Godda district received the benefits, compared with 40% of women in
Ranchi district. There were substantial pro-rich inequalities in JSY receipt (CI 0.10, standard deviation (SD) 0.03 in Odisha; CI
0.18, SD 0.02 in Jharkhand) and in the institutional delivery rate (CI 0.16, SD 0.03 in Odisha; CI 0.30, SD 0.02 in Jharkhand).
Delivery in a public facility was an important determinant of receipt of JSY benefits and explained a substantial part of the
observed poor-rich inequalities in receipt of the benefits. Yet, even among public facility births in Jharkhand, pro-rich
inequality in JSY receipt was substantial (CI 0.14, SD 0.05). This was largely explained by district-level differences in wealth
and JSY receipt. Conversely, in Odisha, poorer women delivering in a government institution were at least as likely to
receive JSY benefits as richer women (CI −0.05, SD 0.03).
Conclusion: JSY benefits were not equally distributed, favouring wealthier groups. These inequalities in turn reflected
pro-rich inequalities in the institutional delivery. The JSY scheme is currently not sufficient to close the poor-rich gap in
institutional delivery rate. Important barriers to institutional delivery remain to be addressed and more support is needed
for low performing districts and states.
Keywords: Socioeconomic inequality, Conditional cash transfers, Maternity care, Janani Suraksha Yojana, India

* Correspondence: natwutth@gmail.com
1
Institute of Health Policy and Management (iBMG), Erasmus University
Rotterdam, Rotterdam, The Netherlands
3
Department of Public Health, Erasmus MC Rotterdam, P.O. Box 20403000 CA
Rotterdam, The Netherlands
Full list of author information is available at the end of the article

© The Author(s). 2017 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0
International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and
reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to
the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver
(http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.
Thongkong et al. International Journal for Equity in Health (2017) 16:48 Page 2 of 9

Background With JSY being one of the largest conditional cash

Universal Health Coverage, defined as ensuring that transfer programs worldwide [5], a considerable number
everyone has access to affordable and quality health care, of studies have sought to establish its impact on the pro-
features high on the policy agenda of many low and portion of women delivering in a facility. Most notably,
middle income countries. While the focus on maternal Lim et al. (2010) found JSY to have a large positive impact
and child health (MCH) in the Millennium Development on the institutional delivery rate (an increase of 43–49%
Goals (MDG) has brought about much progress in the in the probability of women delivering in a facility), lead-
coverage of MCH care, many countries are lagging be- ing to important reductions in neonatal mortality (a re-
hind on MDG 4 and 5, in particular with respect to re- duction of 2.3 neonatal deaths per 1,000 live births) [10].
ducing neonatal and maternal mortality [1]. This has These large effects have, however, been questioned by
much to do with the difficulties of increasing the rate of Powell-Jackson et al. (2015) and Joshi and Sivaram (2014)
institutional deliveries, especially among the poorest who suggested that, despite an increase in use of mater-
population groups within countries. Possible barriers for nity health services, there was no effect of JSY on
using delivery care can be financial, cultural, or know- maternal and neonatal health due to the low quality of
ledge related. Both demand side programs such as con- care provided in public facilities [11, 12]. Some studies
ditional cash transfers, vouchers and user fee removals, investigating heterogeneity of the effect of JSY on use of
and supply side programs such as performance based fi- maternity care across socioeconomic groups found that
nancing, are increasingly being implemented to increase poorer, less educated women and those belonging to
the coverage of MCH care [2–5]. Scheduled Castes or Tribes (as defined by the Indian Con-
In India, the bulk of health care is still financed by stitution) generally benefit more from JSY [11, 13, 14].
out-of-pocket payments made at the point of health care Yet, these studies do not describe and explain socioeco-
use, leading to problems of inaccessibility and lack of fi- nomic inequalities in JSY receipt in detail.
nancial protection [5]. In India, progress towards MDG4 In this study, we aimed to measure the magnitude of
and 5 has been modest; neonatal mortality has fallen socioeconomic inequality in the receipt of JSY benefits
from 44 per 1000 live births in 2001 to 38 per 1000 live and explain these inequalities using the decomposition
births in 2005 and 28 per 1000 live births in 2015. Ma- method. We use data that was prospectively collected
ternal mortality is still high with 280 per 100,000 and through a demographic surveillance system in the states
174 per 100,000 live births in 2005 and 2015, respect- of Jharkhand and Odisha.
ively [6]. Data from the latest Demographic Health Sur-
vey (2005) indicates that only 40% of women deliver in Methods
an institution [7]. Socioeconomic inequalities in MCH Data
have remained high in the past decade [8]. We used data from a population surveillance system in
With the aim of reducing neonatal and maternal mor- five districts in the states of Jharkhand and Odisha, two
tality and reducing out-of-pocket payments associated of the poorest states in India (Godda, Khunti, and Ran-
with institutional delivery, the Indian Government intro- chi district in Jharkhand, and Mayurbhanj and Rayagada
duced the Janani Suraksha Yojana (JSY) scheme in 2005. district in Odisha). Data were only collected in three out
JSY is a conditional cash transfer program that finan- of 24 districts in Jharkhand and in two out of 30 districts
cially rewards pregnant women for delivery, and espe- in Odisha. The surveillance system was set up as part of
cially for delivery in a facility that is empanelled by JSY. a scale-up of a community-based women’s group inter-
Even though the scheme is centrally sponsored, eligibil- vention to reduce neonatal mortality and improve ma-
ity criteria vary by state. The poorest (low performing) ternal and neonatal health. The intervention had proven
states target all women, while in high performing (less to be successful in a cluster randomised trial in Jhar-
poor) states, only women holding a Below Poverty Line khand and Odisha, and had been scaled up to five dis-
(BPL) card are eligibility. Some states require that deliv- tricts in these states that were not part of the original
ery takes place in a public facility, while other states trial. The ‘scale-up area’ was divided into an intervention
(such as Odisha) also have accredited private providers area in which women’s groups were set up, and a control
[9]. Hospital delivery in a government facility should in area, in which no women’s groups were initially set up
theory be free of charge. Free transport to facility should (at a later stage, this control area for the scale-up be-
also be available for pregnant women and newborns in came the site of a new randomised trial with women’s
both states. It is likely however that these services are groups facilitated by government community health
not always free in practice, and there may still be other workers (ASHAs)) [15]. We analysed the data for the
costs associated with delivery in a facility. In Jharkhand, control arm of the scale-up area, including births in the
during the study period, women also received a small period 1 January 2009 (when the surveillance system
amount of money for home deliveries. was set up) until 31 August 2010 (after which women’s
Thongkong et al. International Journal for Equity in Health (2017) 16:48 Page 3 of 9

groups were set up in the control area as part of the and (iii) transferring JSY uptake from a richer to a
new trial). The control area consisted of 25 clusters in poorer woman reduces the value of the index [18].
five districts, with a total population of around 35,000
(around 7000 per district). Of the study population, Determinants of JSY uptake
20,000 were intensively monitored, collecting informa- To identify the important determinants of JSY uptake,
tion on vital outcomes and secondary outcomes includ- we ran several linear probability models3. First, we ran a
ing receipt of JSY benefits. For the remaining 15,000 model of JSY uptake (yi ) only on the wealth quintile in-
population, only data on vital outcomes and a limited dicators (model 1). Next we added household/mother
set of secondary outcomes, not including JSY uptake, covariates and district indicators (model 2). Thereafter
were collected. We analysed the data for the population we added an indicator for whether the child was born in
that was intensively monitored, which gave us a sample a public facility – one of the most important conditions
of 3682 births. for eligibility for JSY benefits (model 3). Finally we esti-
The main outcome variable of interest reflects whether mated model 2 only on the sample of births that took
or not the woman received JSY benefits for her delivery place in a government facility (model 4).
during the study period (1/0) representing the uptake of
JSY. Other outcomes include the uptake of at least three Decomposition of socioeconomic inequalities in JSY
antenatal care visits (ANC3) and delivery in a public uptake
facility1. Having established the determinants of JSY uptake, we
Our main socioeconomic variable of interest is eco- wanted to identify the most important drivers of socio-
nomic status, using a wealth index derived from princi- economic inequality in JSY uptake. If we assume that
pal component analysis on a range of asset variables as JSY uptake can be written as a linear function of K de-
proxy2. Second, we have an indicator for whether the in- terminants xk ,
fant’s mother belongs to a Scheduled Tribe or Scheduled XK
Caste (ST/SC), who represent the most socioeconomi- yi ¼ α þ β x
k¼1 k ik
þ εi ; ð2Þ
cally disadvantaged groups, recognized by the Constitu-
tion of India [16, 17]. Third, we define an indicator for the corrected concentration index can be written as a
illiteracy reflecting whether or not the infant’s mother weighted average of the concentration indices of the co-
can read and write (1/0), as verified by the interviewer. variates (C xk Þ :
Other covariates include demographics (age, number hXK i
of previous pregnancies) and geographical characteristics CC y ¼ 4 β x C
k¼1 k k xk
þ GC ε ð3Þ
(district indicators).
Where x k is the mean of x, and GC ε is the generalized
Measuring socioeconomic inequality concentration index of the residuals. Equation (3) illus-
Next to summarizing JSY uptake across wealth quintiles, trates that for a covariate to contribute to socioeconomic
we also measured socioeconomic inequality in the up- inequality in JSY uptake it needs to be associated with
take of JSY using a concentration index, which takes into JSY uptake (βk Þ and be unequally distributed across so-
account inequality across the full socioeconomic distri- cioeconomic status (C xk Þ [18, 19].
bution. Erreygers [18] has argued that the standard con- These decompositions were performed for models (2),
centration index has some shortcomings when applied (3) and (4) as introduced in the previous section. All
to bounded variables, most importantly that the bounds analyses were performed in Stata 12. Standard errors are
of the index are dependent of the mean of the indicator. adjusted for clustering on the primary sampling unit.
We therefore applied the corrected concentration index
(CC y ) which, for our binary outcome of JSY receipt (yi Þ Results
, is calculated as: Summary statistics
Table 1 shows the means of our covariates. The charac-
CC y ¼ 8covðyi ⋅Ri Þ ð1Þ teristics of women were largely similar in the two states,
except for a higher percentage of women from Sched-
where Ri reflects a woman i’s fractional rank in the so- uled Castes or Scheduled Tribes in Odisha (86%) than in
cioeconomic distribution. This corrected index shares Jharkhand (62%). In Jharkhand, there were considerable
the same interpretation as the standard concentration district-level differences in the characteristics of the
index. Negative values imply that JSY uptake is more study participants. The percentage of women from
concentrated among poorer women. If all women, irre- Scheduled Casts or Scheduled Tribes varied from 49% in
spective of their socioeconomic status, are equally likely Godda to 91% in Kunthi, and the number of previous
to receive JSY benefits, the index would equal to zero, pregnancies varied from 1.4 in Ranchi to 1.9 in Khunti.
Thongkong et al. International Journal for Equity in Health (2017) 16:48 Page 4 of 9

Table 1 Summary statistics of covariates and distribution across districts and socioeconomic status
Age (years) Number of previous Scheduled tribe / scheduled caste Illiteracy rate (%)
pregnancies (%)
Odisha Jharkhand Odisha Jharkhand Odisha Jharkhand Odisha Jharkhand
Total population 24.9 24.9 1.8 1.6 86 62 69 69
By district
Mayurbhanj 23.9 1.7 91 56
Rayagada 26.1 1.9 81 83
Godda 23.9 1.6 49 78
Khunti 26.4 1.9 91 77
Ranchi 24.9 1.4 56 54
By wealth group
Poorest quintile 25.5 25.1 2.4 1.8 93 68 87 90
Second poorest quintile 25.5 25.4 2.1 1.9 88 70 82 88
Middle quintile 24.5 25.4 1.6 1.8 92 69 78 77
Second richest quintile 24.4 24.6 1.4 1.5 89 57 63 63
Richest quintile 24.2 23.7 1.1 1.1 68 44 30 29
Concentration index (SD) −0.01 (0.00) −0.01 (0.00) −0.16 (0.02) −0.10 (0.01) −0.17 (0.02) −0.21 (0.02) −0.45 (0.03) −0.49 (0.02)
Notes: Table shows the standard concentration index for continuous variables (age, number of previous pregnancies) and the corrected concentration index for
bounded variables (scheduled tribe/scheduled caste, illiteracy rate)

Between-district differences in literacy rate were large in Odisha, there was some pro-rich inequality in Jharkhand
both states. Poor-rich differences in the above character- (CI = 0.18, SD 0.02).
istics were also large in both states. Richer women had The percentage of women that received JSY benefits
on average 1 previous pregnancy, compared with 2 preg- was low in comparison with the percentage of women
nancies among poorer women. Furthermore, the that had heard of the scheme, especially in Jharkhand.
illiteracy rate and percentage belonging to a Scheduled While 62% of women received the benefits in Odisha,
Caste or Scheduled Tribe was substantially lower among only 20% did so in Jharkhand. In the latter state, uptake
the richest quintile in both states. Differences in house- of JSY benefits ranged from 5% in Godda to 40% in
hold wealth between districts were especially large in Ranchi, while such regional differences were small in
Jharkhand, and much smaller in Odisha (Fig. 1a and b). Odisha. In both states, the receipt of JSY benefits was
Table 2 shows average awareness and receipt of JSY disproportionally concentrated among better-off women
benefits across the two states and districts within each (CI: 0.18 in Jharkhand, 0.10 in Odisha).
state, and across wealth quintiles within each state. The Women who took up the benefits, received about
vast majority of women had heard of the scheme (94% 1400 Rs. on average, but reported amounts were some-
in Odisha, 85% in Jharkhand). There was some district- what lower in Jharkhand (1222 Rs.). There was little so-
level variation in Odisha, where only 62% of women in cioeconomic inequality in the amount received, although
Khunti had heard of JSY, compared with 97% in Ranchi. in Jharkhand women in the lowest quintile reported re-
While poor-rich inequalities in awareness were small in ceiving on average 265 Rs. less than those in the highest

Fig. 1 Distribution of household wealth across districts in Odisha (a) and Jharkhand (b). Notes: Figures show the proportion of children in each
wealth quintile in each district, by state
Thongkong et al. International Journal for Equity in Health (2017) 16:48 Page 5 of 9

Table 2 Summary statistics of JSY-related outcomes and their distribution

Heard (%) Received (%) Total amount (Rs.)
Odisha Jharkhand Odisha Jharkhand Odisha Jharkhand
Total population 94 85 62 20 1402 1222
By district
Mayurbhanj 95 59 1405
Rayagada 93 66 1400
Godda 87 5 944
Khunti 62 13 808
Ranchi 97 40 1343
By wealth group
Poorest quintile 91 71 56 11 1398 1015
Second poorest quintile 95 81 62 16 1394 1101
Middle quintile 94 86 59 15 1392 1290
Second richest quintile 96 92 64 25 1418 1256
Richest quintile 96 92 71 33 1408 1280
Concentration index (SD) 0.04 (0.02) 0.18 (0.02) 0.10 (0.03) 0.18 (0.02) 0.00 (0.00) 0.04 (0.01)
Total sample (N) 1345 2337 1345 2337 836 470
Notes: Table shows the standard concentration index for continuous variables (amount) and the corrected concentration index for bounded variables (having
heard about JSY, having received JSY benefits). Total amount received in Indian Rupee (Rs.) was estimated among women who received JSY benefits

wealth quintile. These lower amounts for Jharkhand are care was lower, and inequalities between districts and
arguably explained by the higher proportion of home wealth groups were larger, than in Odisha. In Jharkhand,
births (4% in Jharkhand versus 2% in Odisha) and the only 26% of women made at least 3 antenatal care visits
small benefits associated with home deliveries. and 27% of women delivered in a facility. While district
The above patterns of inequality in the receipt of JSY level differences were small in Odisha, they were large in
benefits reflect inequalities in uptake of maternity care in Jharkhand (8–55% for deliveries in government facil-
(Table 3). In Jharkhand, the overall uptake of maternity ities and 17–44% for 3ANC). In both states, the majority

Table 3 Summary statistics of maternity care-related outcomes and their distribution

ANC3 (%) Institutional delivery (%)
Odisha Jharkhand All Public Private
Odisha Jharkhand Odisha Jharkhand Odisha Jharkhand
Total population 49 26 73 27 64 20 9 7
By district
Mayurbhanj 44 75 74 2
Rayagada 54 71 53 17
Godda 10 17 12 5
Khunti 8 17 12 5
Ranchi 55 44 34 10
By wealth group
Poorest quintile 36 11 65 17 60 13 6 4
Second poorest quintile 43 15 70 17 60 14 9 3
Middle quintile 49 15 68 17 62 14 6 3
Second richest quintile 52 33 76 32 67 27 9 5
Richest quintile 67 57 87 53 73 34 14 20
Concentration index (SD) 0.25 (0.03) 0.37 (0.02) 0.16 (0.03) 0.30 (0.02) 0.11 (0.03) 0.19 (0.02) 0.05 (0.02) 0.11 (0.01)
Total sample (N) 1345 2337 1345 2337 1345 2337 1345 2337
Notes: Table shows the corrected concentration index
Thongkong et al. International Journal for Equity in Health (2017) 16:48 Page 6 of 9

of institutional deliveries took place in a public facility. with JSY receipt in both states, although more so in
The percentage of private sector deliveries was only sub- Odisha (64 PP, versus 37 PP in Jharkhand). In this
stantial for the richest quintile, especially in Jharkhand. model, poor-rich inequalities in JSY receipt disap-
peared, but there still was substantial district level
heterogeneity. In Odisha, women from Rayagada were
Determinants of JSY uptake 21 PP more likely to receive JSY benefits than women
Table 4 presents the regression results of the differ- from Mayurbhanj; in Jharkhand, women from Ranchi
ent models used to identify the determinants of JSY were 26 PP more likely to receive JSY benefits than
uptake. Model (1) confirms the earlier described women from Godda. Model (4) is similar to model
patterns of pro-rich receipt of JSY benefits. These (2) but only included births in government facilities.
poor-rich inequalities were largely explained by Within this sample, the probability of receiving JSY
district-level differences in JSY uptake and by char- benefits only varied by district.
acteristics of the mother (Model 2). In Jharkhand,
the probability of receiving JSY benefits was highest Decomposition of socioeconomic inequality in JSY uptake
in Ranchi (33 percentage points (PP) higher than in Figure 2a and b graphically illustrate the results of
the reference district Godda) and five PP lower for the decomposition of poor-rich inequality in JSY re-
illiterate women than for literate women. In Odisha, ceipt as measured by the concentration index, for
there were no statistically significant district level models (2), (3) and (4) in Table 4 (detailed results
differences in JSY receipt, but women who had had are available in the Appendix). The total height of
more previous pregnancies were less likely to receive the bar represents the magnitude of poor-rich in-
JSY benefits (4 PP). equalities (CI of JSY receipt). Note that a variable
In Model 3, we added an indicator for government contributes to the CI both through its association
facility delivery to the covariates. As delivery in a fa- with JSY receipt (Table 4) and the extent to which it
cility was a condition for receiving JSY benefits, and is unequally distributed by household wealth
there are few private hospitals empanelled in JSY. De- (Table 1). As model (2) and model (3) were esti-
livery in government facility was strongly associated mated on the same sample, the bars are equally

Table 4 Determinants of JSY uptake

Odisha Jharkhand
[1] [2] [3] [4] [1] [2] [3] [4]
Household wealth
Poorest (ref.)
2nd poorest 0.06b 0.04 0.02 0.01 0.04b 0.00 0.01 −0.03
Middle 0.03 −0.00 −0.01 −0.01 0.04 −0.03 −0.02 −0.00
2nd richest 0.08a 0.04 0.02 −0.02 0.14a 0.01 −0.01 −0.01
Richest 0.15c 0.09 0.04 −0.05 0.22c 0.04a 0.01 −0.06
Region
Mayurbhanj (ref.)
Rayagada 0.07 0.21c 0.14b
Godda (ref.)
Khunti 0.08 0.06a 0.27c
Ranchi 0.33c 0.26c 0.48c
Age 0.01 0.00 0.01 −0.00 −0.00 −0.00
Previous pregnancy −0.04 b
−0.01 −0.00 0.00 0.00 0.02
ST/SC −0.02 0.02 −0.02 0.01 0.02a −0.03
Illiterate −0.03 −0.03 −0.02 −0.05 b
−0.03 a
−0.04
Public-institutional delivery 0.64c 0.37c
Observations 1345 1344 1342 860 2335 2329 2315 469
Note: Table shows coefficients from linear probability models with JSY uptake as dependent variable. The fourth [4] model is only estimated on the sample of
women who delivered in a public facility
a
significant at 10% level, bsignificant at 5% level, csignificant at 1% level
Thongkong et al. International Journal for Equity in Health (2017) 16:48 Page 7 of 9

Fig. 2 Decomposition of socioeconomic inequality in JSY uptake in Odisha (a) and Jharkhand (b). Notes: The bars in each figure reflect absolute
contributions to inequality in JSY uptake using model [2], [3] and [4] in Table 4, respectively. In Odisha, CI of JSY uptake equals 0.10 the full sample
(model 2 and 3), and −0.05 for the sample of public deliveries (model 4). In Jharkhand, CI equals 0.18 and 0.14, respectively

high. The decomposition using model (4) only in- Nevertheless, the discrepancy between awareness and
cluded births in a government facility. receipt of benefits was very large in both states, es-
Looking at Fig. 2a for Odisha, we see in the first pecially in Jharkhand. Poor-rich inequalities in JSY
bar that household wealth was the most important receipt reflected substantial pro-rich inequalities in
driver (43%) of poor-rich inequality in JSY uptake; the institutional delivery rate. Yet, even among gov-
the second most important driver was the number of ernment facility deliveries there were considerable
previous pregnancies (34%). The latter contribution pro-rich inequalities in JSY receipt in Jharkhand,
was driven by the higher number of previous preg- which were explained by district level differences in
nancies among poorer women and the negative poverty and receipt of JSY benefits. Conversely, in
association between the number of pregnancies and Odisha, poorer women delivering in a government
JSY uptake. These contributions reduced substan- institution were at least as likely to receive JSY ben-
tially after adding government facility delivery to the efits as richer women.
model (second bar). Within the sample of govern- There are some limitations to this study. First, the data
ment facility deliveries (third bar), receipt of JSY were not collected with the aim of evaluating JSY, so we
benefits was pro-poor (CI = −0.05, SD 0.03); poor could not exactly identify whether women had fulfilled
women were slightly more likely to obtain JSY bene- all the necessary conditions to be eligible for JSY bene-
fits than richer women. fits. We used delivery in a government facility as a
In Jharkhand (Fig. 2b), we get quite a different proxy, as this is the most important requirement. Sec-
story, with district level differences explaining most of ond, the data were only collected in a small number of
the poor-rich inequalities in each of the decompos- districts in both states. The large district level inequal-
ition models (63, 47 and 60% in models (2), (3) and ities in Jharkhand may reflect the fact that data were col-
(4) respectively). Even within the sample of govern- lected in Ranchi district – the capital district- and two
ment facility deliveries, there was considerable pro- rural districts, while in Odisha the capital city district
rich inequality in JSY receipt (CI = 0.14 SD 0.05), was not included. Third, our data relates to births in
which was also driven by district level differences. 2009 and 2010. While the implementation of the JSY
This can be explained by the fact that in Ranchi, the program may have improved in the mean-time, our
richest district, women were far more likely to obtain paper provides one of the most recent estimates avail-
JSY benefits than in the other districts. Once these able for JSY uptake across socioeconomic strata. Fourth,
district level differences were taken into account, JSY given the cross sectional nature of our data, our models
uptake was more concentrated among the poor, which only allow the identification of associations between co-
explains the negative contribution of wealth to the variates and JSY uptake, and should not be interpreted
poor-rich inequalities in JSY uptake (−19%). as causal effects.
Our findings correspond with DLHS evidence that the
Discussion odds of receiving JSY benefits were not always highest
Our analysis shows that inequalities in JSY receipt among the poor. Yet, the poor-rich inequalities in JSY
are substantial in Jharkhand and Odisha, both be- receipt that we report for Odisha and Jharkhand were
tween wealth groups and between districts within much larger than in these national level studies [5, 10].
states. In Jharkhand, these inequalities to some ex- Our finding of large between-state differences in
tent reflected differences in awareness of the scheme. JSY uptake reflects a broader pattern of a highly
Thongkong et al. International Journal for Equity in Health (2017) 16:48 Page 8 of 9

variable implementation process and substantial dif- JSY scheme is currently not sufficient to close the
ferences in JSY uptake across states in the country. poor-rich gap in institutional delivery rate. Important
Interestingly, Odisha was the next-highest performing barriers to institutional delivery remain to be addressed
state in terms of JSY receipt in the country, and Jhar- and more support is needed for low performing district
khand one of the lowest performing states, based on and states.
2007–2009 DLHS data. We found that even within
states, indicators of JSY success were highly variable Endnotes
between districts. Such differences could be due to 1
In Jharkhand, only deliveries in public facilities were
differences in accessible health infrastructure for facil- eligible for JSY benefits, while in Odisha also some pri-
ity delivery, but could also highlight differences in vate facilities were included in the JSY scheme. Unfortu-
state-level government capacity to implement national nately we do not know whether the private facilities
level policies [10]. reported to have been used in our data are accredited by
JSY. Therefore, we cannot identify exactly whether
Implications women have fulfilled all requirements to be eligible for
Our findings imply that the JSY scheme is currently JSY benefits. In Jharkhand, where only births in govern-
not sufficient to close the poor-rich gap in institu- ment facilities are eligible for JSY, we only found 13
tional delivery rate. Low uptake is not so much re- births outside of government facilities that were given
lated to low awareness of the scheme per se, but the cash transfer of 1400 Rs. This does not suggest
rather with remaining barriers to institutional delivery major problems of leakage of the program.
and, in Jharkhand, to district-level differences in per- 2
Assets include: mattress, bed, chair, table, pressure
formance of the JSY scheme. This corresponds with cooker, electricity, fan, radio, television, clock, phone,
findings from another study, which reported that des- animal-drawn cart, bicycle, motor bicycle and agricul-
pite receiving JSY benefits, many families still have to tural land ownership (small size, medium size, large size
borrow money to cover out of pocket expenditures. and mortgage).
Furthermore, non-monetary demand and supply side 3
We have confirmed robustness of results to using
barriers, including quality of care, distance, and a probit models and estimating marginal effects. We
tradition of home delivery remain [5]. The large dif- prefer to present the linear probability model as this
ferences in receipt of JSY and uptake of maternity facilitates the decomposition technique that is used
care between Jharkhand and Odisha and between dis- later. Non-linear extensions of the decomposition
tricts within Jharkhand, suggest that more support is method do exist but require focusing on the latent
needed for low performing districts and states. index rather than the actual uptake variable or im-
Women giving birth in government facilities in Ran- pose approximation errors (Vandoorslaer, Koolman
chi, the richest district, were much more likely to re- and Jones 2004) [21].
ceive benefits as compared to those giving birth in
government facilities in other districts. This might re-
flect a better administration of the scheme in districts Appendix
with more resources. It is reassuring, though, that we
found that poor women were as likely to receive JSY
Table 5 Decomposition of socioeconomic inequality in JSY
benefits as richer women when delivering in a gov- uptake in Odisha and Jharkhand
ernment facility, after taking such district-level differ-
Odisha (%) Jharkhand (%)
ences into account.
[2] [3] [4] [2] [3] [4]
Our findings are in line with results from other evi-
Household wealth 42.78 15.05 −55.01 14.68 2.14 −19.10
dence looking at the equity in the uptake of cash incen-
tive schemes in neighboring countries of India. A review Region −0.30 −0.84 −11.50 63.17 46.92 59.83
of the evidence has revealed that the Safe Delivery In- Age −5.59 −2.96 −5.79 2.39 0.37 0.63
centive Programme in Nepal and the Maternal Health Previous pregnancy 34.34 10.31 2.91 −1.28 −1.47 −3.10
Voucher scheme in Bangladesh failed to target poorer ST/SC 2.01 −1.86 5.36 −0.86 −2.41 2.51
households. The uptake of those schemes was more
Illiterate 8.47 10.64 10.95 11.89 7.56 8.23
concentrated among the better-off [20].
Public-insitutional delivery 0.00 50.27 0.00 0.00 34.95 0.00

Conclusions Error −6.51 −8.07 −8.48 5.72 4.18 6.60

JSY benefits were not equally distributed, favouring In Jharkhand, women received JSY payments in installments. However, women
received 500 Rs. for home delivery event. This was continued during the data
wealthier groups. These inequalities in turn reflected collection period. Number of home delivery might have some influence in
pro-rich inequalities in the institutional delivery. The reducing the average
Thongkong et al. International Journal for Equity in Health (2017) 16:48 Page 9 of 9

Acknowledgements 10. Lim SS, Dandona L, Hoisington JA, James SL, Hogan MC, Gakidou E. India’s
We thank Rica Garde for her contributions to the development of the initial Janani Suraksha Yojana, a conditional cash transfer programme to increase
plan for this study. births in health facilities: an impact evaluation. Lancet. 2010;375(9730):2009–23.
11. Powell-Jackson T, Mazumdar S, Mills A. Financial incentives in health: New
Funding evidence from India’s Janani Suraksha Yojana. J Health Econ. 2015;43:154–69.
TAJH was primarily supported by the Economic and Social Research Council 12. Joshi S, Sivaram A. Does it pay to deliver? An evaluation of India’s safe
and the Department for International Development (grant number ES/ motherhood program. World Dev. 2014;64:434–47.
I033572/1). TAJH and EvdP were funded by an Erasmus University Rotterdam 13. Amudhan S, Mani K, Rai SK, Pandav CS, Krishnan A. Effectiveness of demand
Research Excellence Initiative grant. EvdP also acknowledges support from and supply side interventions in promoting institutional deliveries–a quasi-
the Netherlands Organization for Scientific Research, Innovational Research experimental trial from rural north India. Int J Epidemiol. 2013;42(3):769–80.
Incentives Scheme, Veni project 451-11-031. 14. Gupta SK, Pal DK, Tiwari R, Garg R, Shrivastava AK, Sarawagi R, Patil R,
Agarwal L, Gupta P, Lahariya C. Impact of Janani Suraksha Yojana on
institutional delivery rate and maternal morbidity and mortality: an
Availability of data and materials
observational study in India. J Health Popul Nutr. 2012;30(4):464.
Data will not be shared as this data belongs to Ekjut.
15. Tripathy P, Nair N, Mahapatra R, Rath S, Gope RK, Bajpai A, Singh V, Nath V, Ali
S, Kundu AK, et al. Community mobilisation with women’s groups facilitated
Authors’ contributions by Accredited Social Health Activists (ASHAs) to improve maternal and
NT performed the statistical analysis and wrote the manuscript. TAJH and newborn health in underserved areas of Jharkhand and Orissa: study protocol
EVDP participated in coordination and helped to draft the manuscript. SS for a cluster-randomised controlled trial. Trials. 2011;12:182.
and SR gave advice and insight into the JSY scheme. All authors read and 16. Ministry of Tribal Affairs. Government of India. http://www.tribal.nic.in/
approved the final manuscript. Content/DefinitionpRrofiles.aspx. Accessed 13 June 2014.
17. United Nations in India. http://in.one.un.org/task-teams/scheduled-castes-
Competing interests and-scheduled-tribes/. Accessed 13 June 2014.
The authors declare that they have no competing interests. 18. Erreygers G. Correcting the concentration index. J Health Econ. 2009;28(2):
504–15.
Ethics approval and consent to participate 19. Wagstaff A, Van Doorslaer E, Watanabe N. On decomposing the causes of
Ethical approval for the study was obtained through an independent ethical health sector inequalities with an application to malnutrition inequalities in
research committee chaired by Alok K Debdas in Jamshedpur, India, and Vietnam. J Econom. 2003;112(1):207–23.
through University College London’s Research Ethics Committee (UK). 20. Witter S, Somanathan A. Demand-Side Financing for Sexual and Reproductive
Permission was obtained from local community representatives (village Health Services in Low and Middle-Income countries: A Review of the
headmen in Jharkhand and Panchayati Raj institution leaders in Odisha) to Evidence (Policy Research Working Paper 6213). The World Bank, East Asia and
collect data in their areas. Individual informed consent was sought from all the Pacific Region, Human Development Department; 2012.
participants and recorded though either a signature or thumbprint. 21. Van Doorslaer E, Koolman X, Jones AM. Explaining income-related
inequalities in doctor utilisation in Europe. Health Econ. 2004;13(7):629–47.
Author details
1
Institute of Health Policy and Management (iBMG), Erasmus University
Rotterdam, Rotterdam, The Netherlands. 2Ekjut, Plot 556B, Potka,
Chakradharpur, West Singhbhum, Jharkhand PIN: 833102, India. 3Department
of Public Health, Erasmus MC Rotterdam, P.O. Box 20403000 CA Rotterdam,
The Netherlands.

Received: 18 May 2016 Accepted: 21 February 2017

References
1. Lozano R, Wang H, Foreman KJ, Rajaratnam JK, Naghavi M, Marcus JR,
Dwyer-Lindgren L, Lofgren KT, Phillips D, Atkinson C. Progress towards
Millennium Development Goals 4 and 5 on maternal and child mortality: an
updated systematic analysis. Lancet. 2011;378(9797):1139–65.
2. Ranganathan M, Lagarde M. Promoting healthy behaviours and improving
health outcomes in low and middle income countries: a review of the impact
of conditional cash transfer programmes. Prev Med. 2012;55:S95–105.
3. Grainger C, Gorter A, Okal J, Bellows B. Lessons from sexual and reproductive
health voucher program design and function: a comprehensive review. Int J
Equity Health. 2014;13:33.
4. Witter S. Performance-based financing for strengthening delivery of sexual
and reproductive health services in low-and middle-income countries: an
evidence synthesis paper. Washington DC: World Bank; 2012. Submit your next manuscript to BioMed Central
5. Modugu HR, Kumar M, Kumar A, Millett C. State and socio-demographic and we will help you at every step:
group variation in out-of-pocket expenditure, borrowings and Janani
Suraksha Yojana (JSY) programme use for birth deliveries in India. BMC • We accept pre-submission inquiries
Public Health. 2012;12(1):1048. • Our selector tool helps you to find the most relevant journal
6. The World Bank. http://data.worldbank.org. Accessed 16 Jan 2016.
• We provide round the clock customer support
7. STATcomplier The DHS Program. http://www.statcompiler.com. Accessed 16
Jan 2016. • Convenient online submission
8. Pathak PK, Singh A, Subramanian S. Economic inequalities in maternal • Thorough peer review
health care: prenatal care and skilled birth attendance in India, 1992–2006.
• Inclusion in PubMed and all major indexing services
PLoS One. 2010;5(10):e13593.
9. Government of India: Janani Suraksha Yojana: features and frequently asked • Maximum visibility for your research
questions and answers. In: Ministry of Health and Family Welfare MHD,
editor. Nirman Bhavan, New Delhi: Government of India; 2006. Submit your manuscript at
www.biomedcentral.com/submit