1 Who are the Beneficiaries of the Poverty Alleviation Programs in Rural India? Takahiro Sato (Osaka City University) A Preliminary Analysis on the Consequences.

1

Who are the Beneficiaries of the Poverty Alleviation Programs in

Rural India?

Takahiro Sato (Osaka City University)

A Preliminary Analysis on the Consequences of the Decentralization

in the 1990s

2

Outline• Section 1: Introduction• Section 2: The institutional background of the 73rd

Constitutional Amendment and poverty alleviation programs.

• Section 3: Implications of the decentralization by referring to the earlier studies. (But, today I skip this section)

• Section 4 : Empirical investigation of the allocation of the poverty alleviation programs before-and-after the decentralization.

• Section 5: Concluding remarks.

3

1 Introduction• Does greater decentralization improve the

living condition of the poor, the women, or the minority groups in rural India?

• The 73rd Constitutional Amendment in 1993 →The elected Panchayats determine the beneficiary selection for the poverty alleviation programs

4

Advantage and disadvantage of decentralization

• Advantage:• →Enhanced accountability• →Information advantage

• Disadvantage:• →Capture by local elite

5

From the fields (1)

• “Decentralisation in Karnataka yielded paradoxical results. The number of people involved in corrupt acts increased significantly. But the overall amount of money stolen almost certainly decreased - at least modestly” (Richard and Manor 1998: 61)

6

From the fields (2)• “A further step in democratization seems to have

taken place with the reformation of gram panchayat in 1993 with the 73rd Amendment to the Constitution ･･･ The reformation of panchayat system in 1993, together with the decline of factional politics after liberalization policy, made the gram panchayat meeting a more democratic place in terms of representation and functioning.” (Tanabe 2004)

7

This paper’s approach

• To use (1) household survey data from the National Sample Survey Organization (NSS), and

• (2) regionally aggregated election data from the Election Commission of India (ECI).

8

This paper’s approach

• To focus on those who were the beneficiaries of the poverty alleviation programs in rural India before-and-after decentralization.

• Difference-in-Difference Approach.

9

Section 2 Institutional Context

• 2.1: Features of the 73rd Constitutional Amendment in 1993 aiming at the decentralization

• 2.2: Nature of poverty alleviation programs in rural area.

10

2.1 The 73rd Constitutional Amendment

• Three-tier “Panchyati Raj” (local self-governance) system.

• →“Zilla Parishad” at district level, “Panchayat Samiti” at block level, and “Gram Panchayat” at village level.

• Persons selected by direct election fill all of the seats in a Panchayat.

11

2.1 The 73rd Constitutional Amendment (Continued)

• Elections of Panchayats are held regularly every five years.

• Political reservation for ST/SC and women.

• The Eleventh Schedule in the Constitution.

• →Poverty alleviation programme as one of responsibilities of Panchayats.

12

2.2 Poverty Alleviation Programs • Integrated Rural Development Programme

(IRDP): • →aiming at generating sufficient income to

enable the rural poor to cross the poverty line.

• →providing the government subsidy and bank credit to the poor for encouraging the self-employment activity.

13

2.2 Poverty Alleviation Programs(Continued)

• Public works programs:• →National Rural Employment Programme

(NREP), Rural Landless Employment Guarantee Programme (RLEGP), Minimum Needs Programme (MNP), and other schemes

• →aiming at providing wage employment to the poor.

14

Section 4 Statistical Analysis• Data:• →Main unit of observation: household• →National Sample Survey (NSS)'s 50th round

(1993-94), and 55th round (1999-2000)• →including detailed socioeconomic information on

approximately seventy thousand rural households• In addition, 　→ election data from the Election Commission of

India’s Statistical Report on General Elections

15

Table 1 Descriptive Statistics of the Main Variables

1993 1999variable NOB Average SD Min Max NOB Average SD Min Max

irdpdmy: IRDP=1 68923 0.05 0.21 0.00 1.00 71252 0.06 0.23 0.00 1.00pwdummy: public works=1 69301 0.04 0.19 0.00 1.00 71099 0.03 0.18 0.00 1.00pwirdp: poverty alleviation programs(IRDP orpublic works)=1 69301 0.09 0.28 0.00 1.00 70959 0.08 0.28 0.00 1.00illiteracy: illiterate=1 69219 0.48 0.50 0.00 1.00 71413 0.46 0.50 0.00 1.00hohsex2: female headed household=1 69225 0.09 0.29 0.00 1.00 71466 0.10 0.30 0.00 1.00anyownld1: landed=1 69230 0.95 0.23 0.00 1.00 71146 0.93 0.25 0.00 1.00religion2: Muslim=1 69230 0.09 0.28 0.00 1.00 71392 0.11 0.31 0.00 1.00sgrp1: ST=1 69230 0.13 0.34 0.00 1.00 71349 0.14 0.35 0.00 1.00sgrp2: SC=1 69230 0.18 0.38 0.00 1.00 71349 0.18 0.38 0.00 1.00hhtype2: agricultural labour household=1 69230 0.24 0.43 0.00 1.00 71327 0.26 0.44 0.00 1.00hhtype4: agricultural self employmenthousehold=1 69230 0.43 0.50 0.00 1.00 71327 0.38 0.49 0.00 1.00agehoh: age of the head of household 69230 44.59 13.72 0.00 99.00 71461 45.27 13.91 0.00 99.00numbadu: number of adult person per household 69230 3.28 1.76 0.00 45.00 71466 3.37 1.83 1.00 39.00pcgini: Gini coefficient of own land 68773 0.69 0.08 0.41 0.95 70968 0.71 0.09 0.41 0.93voter turnout: voter turnout ratio 67952 0.57 0.12 0.22 0.85 70968 0.61 0.09 0.34 0.82effnphd2: two party competitiveness index 67952 1.52 2.16 0.00 9.05 70968 0.87 1.52 0.00 15.98

16

4.2 Estimation strategy (1)

• To find the characteristics of the beneficiaries and to examine the influences of the socio political environments.

• →estimating the Probit model for 1993 as before decentralization and 1999 as after decentralization.

• →investigating the estimated coefficients.• →comparing the results of 1993 and 1999.

17

Estimation Equation

18

Table 2 Basic Probit Model

dependent variable IRDPyear=1993 year=1999NOB=67642 NOB=70252Wald chi2(30)=1389.14 Wald chi2(31)=802.74Prob>chi2=0 Prob>chi2=0Log pseudolikelihood=-12551 Log pseudolikelihood=-14794Pseudo R2=0.0555 Pseudo R2=0.0262

variable coefficient t-value p-value coefficient t-value p-valueeffnphd2 -0.12 -15.07 0.0% 0.01 1.62 10.6%voterturn -0.80 -4.60 0.0% -0.32 -1.78 7.6%pcgini -0.46 -3.10 0.2% -0.15 -0.94 34.8%illitercy 0.03 1.74 8.1% 0.03 1.71 8.8%hohsex2 -0.25 -7.02 0.0% -0.05 -1.74 8.1%anyownld1 0.40 7.83 0.0% 0.10 2.74 0.6%religion2 -0.06 -1.74 8.2% 0.00 0.06 94.9%sgrp1 0.24 8.78 0.0% 0.24 9.91 0.0%sgrp2 0.31 13.76 0.0% 0.14 6.34 0.0%hhtype2 0.11 4.28 0.0% 0.01 0.42 67.6%hhtype4 0.03 1.52 12.8% 0.03 1.45 14.7%agehoh 0.00 -1.50 13.5% 0.00 -1.13 25.8%numbadu 0.03 6.88 0.0% 0.01 1.08 28.0%

19

Table 2 Basic Probit Model

dependent variable public worksyear=1993 year=1999NOB=67938 NOB=70105Wald chi2(30)=1739.56 Wald chi2(31)=911.07Prob>chi2=0 Prob>chi2=0Log pseudolikelihood=-9833 Log pseudolikelihood=-9645Pseudo R2=0.0915 Pseudo R2=0.0502

variable coefficient t-value p-value coefficient t-value p-valueeffnphd2 -0.05 -6.60 0.0% -0.03 -3.28 0.1%voterturn -1.67 -8.38 0.0% -0.09 -0.40 68.8%pcgini 1.07 6.46 0.0% -0.82 -4.26 0.0%illitercy 0.17 7.89 0.0% 0.00 0.02 98.6%hohsex2 -0.13 -3.65 0.0% -0.16 -4.32 0.0%anyownld1 0.13 2.86 0.4% 0.05 1.15 24.9%religion2 0.05 1.40 16.3% 0.01 0.31 75.8%sgrp1 0.10 3.55 0.0% 0.28 9.88 0.0%sgrp2 0.10 3.96 0.0% 0.04 1.61 10.7%hhtype2 0.08 3.03 0.2% 0.09 3.45 0.1%hhtype4 -0.20 -8.33 0.0% -0.22 -8.99 0.0%agehoh 0.00 -4.60 0.0% 0.00 -2.62 0.9%numbadu 0.03 4.57 0.0% 0.02 3.55 0.0%

20

Table 2 Basic Probit Model

dependent variable poverty alleviation programs(IRDP or public works)year=1993 year=1999NOB=67938 NOB=69972Wald chi2(30)=2028.98 Wald chi2(31)=1165.84Prob>chi2=0 Prob>chi2=0Log pseudolikelihood=-18709 Log pseudolikelihood=-19267Pseudo R2=0.0534 Pseudo R2=0.0293

variable coefficient t-value p-value coefficient t-value p-valueeffnphd2 -0.09 -15.59 0.0% 0.00 -0.37 71.4%voterturn -1.22 -8.57 0.0% -0.30 -1.84 6.6%pcgini -0.19 -1.59 11.2% -0.60 -4.19 0.0%illitercy 0.11 6.90 0.0% 0.01 0.53 59.8%hohsex2 -0.20 -7.00 0.0% -0.09 -3.43 0.1%anyownld1 0.25 6.88 0.0% 0.09 2.84 0.4%religion2 -0.01 -0.37 70.9% 0.00 -0.13 89.6%sgrp1 0.17 7.23 0.0% 0.30 14.00 0.0%sgrp2 0.24 12.21 0.0% 0.12 5.96 0.0%hhtype2 0.09 4.35 0.0% 0.04 1.84 6.6%hhtype4 -0.09 -5.24 0.0% -0.08 -4.33 0.0%agehoh 0.00 -4.49 0.0% 0.00 -2.10 3.6%numbadu 0.04 9.03 0.0% 0.01 2.48 1.3%

21

Summary of the Main Results

• (1) Regional socio political environment is likely to affect the allocation of the poverty alleviation programs: Greater inequality and less political competition, less the provision of the poverty alleviation programs.

• (2) The disadvantaged groups were not necessarily likely to be the primary beneficiaries of the poverty alleviation programs over others. The female headed household and the landless group remained in a disadvantaged position in participating such programs throughout the entire period.

• (3) But, SC, ST, and agricultural labour household have been in advantaged position to get the programs.

22

4.3 Estimation strategy (2)

• Difference-in-Difference Approach

• Madhya Pradesh as ’treatment 　 group.’

• Karnataka, Kerala, and West Bengal as ’control group.’

23

Why is MP treatment group?• ’Article 243 G of the Constitution empowered

the state legislatures to give panchayats so much power as to make them the institutions of self-government with powers to prepare plans for economic development and social justice including the subjects listed in the 11th Schedule of the Constitution. But, with some exceptions in Kerala, Madhya Pradesh,Tripura and West Bengal nothing worthwhile has been devolved to the panchayats’ (Pal 2001) .

24

Why is MP treatment group?• Before the Amendment, no serious decentralization in M

P. → The 73rd Constitutional Amendment treats MP.• Chief Minister of MP, Digvijay Singh’s political will.• The first state to conduct the Panchayats election in 199

4 under the provision of the 73rd Constitutional Amendment.

• Vacancy rate of the representatives of the Panchayats is quite low.

• Only state to introduce of right to recall the member of Gram Panchayats.

• The advanced state in preparing the District Planning Committee and the right to information.

• See the Table 3.

25

Table 3 Progress of the Decentralization at the State Level

DistrictPlanning

Committee

Statefinancialresource functions staff

AP 17% 45% 7% NoArunachal Pradesh 0% 0% 0% No

Assam 0% 0% 0% NoBihar 0% 0% 0% No

J harkhand 0% 0% 0% NAGoa 0% 0% 0% No

Gujarat 0% 0% 0% NoHaryana 0% 55% 0% Yes

HP 7% 79% 24% NoKarnataka 100% 100% 100% Yes

Kerala 52% 100% 52% YesMP 34% 79% 31% Yes

Chhattisgarh 34% 79% 31% NAMaharashtra 62% 62% 62% No

Manipur 0% 76% 14% YesOrissa 17% 86% 10% YesPunjab 0% 24% 0% No

Rajasthan 0% 100% 0% YesSikkim 100% 100% 100% Yes

Tami Nadu 0% 100% 0% YesTripura 0% 41% 0% Yes

UP 41% 45% 31% YesUttarakhand 41% 45% 31% NAWest Bengal 41% 100% 41% YesA & N Island 0% 0% 0% YesChandigarh 0% 0% 0% No

D & N Haveli 0% 10% 10% YesDaman & Diu 0% 100% 0% No

Delhi 0% 0% 0% NoLakshwdeep 0% 21% 0% YesPondicherry 0% 0% 0% No

J K NA NA NA NoMeghalaya NA NA NA NoMizoram NA NA NA NoNagaland NA NA NA No

progress of devolution to thePanchayats under the EleventhSchedule of the Constitution

26

Why is Karnataka, Kerala and West Bengal control group?

• Commitment to the decentralization before the 73rd Constitution Amendment.

• Good practice case of the decentralization in India since in these states the Panchayats have worked very well.

• The decentralization of Karnataka is regarded as a model case in preparing the 73rd Constitutional Amendments.

• Implementing decentralization both before and after the 73rd Constitution Amendment.

27

Estimation strategyTreatment Group Control Group

1993 MP Karnataka, Kerala, West Bengal

1999 MP Karnataka, Kerala, West Bengal

• Pooling the sample restricted to Karnataka, Kerala, West Bengal, and MP in both reference years.

• Estimating Probit model by employing difference-in-difference approach.

28

Estimation Equation

29

Table 4 Probit Model of the difference- in-difference approach

dependent variable IRDP publc works poverty alleviation programsNOB=29847 NOB=29846 NOB=29929Wald chi2(40)=230.35 Wald chi2(39)=184.94 Wald chi2(40)=248.36Prob>chi2=0 Prob>chi2=0 Prob>chi2=0Log pseudolikelihood=-5867 Log pseudolikelihood=-2822 Log pseudolikelihood=-7661Pseudo R2=0.0198 Pseudo R2=0.032 Pseudo R2=0.0165

variable coefficient t-value p-value coefficient t-value p-value coefficient t-value p-valueeffnphd2 0.00 0.01 99.5% 0.01 0.16 87.2% 0.01 0.23 81.7%

voterturnout -0.16 -0.67 50.3% 0.94 2.54 1.1% 0.27 1.21 22.8%pcgini -1.07 -4.21 0.0% -0.67 -1.83 6.7% -0.95 -4.10 0.0%

illitercy × DMPT 0.18 1.58 11.4% -0.01 -0.04 96.6% 0.09 0.91 36.2%hohsex2 × DMPT 0.00 -0.01 99.0% 0.09 0.31 76.0% 0.01 0.06 94.9%

anyownld1 × DMPT 0.26 1.07 28.4% 0.34 1.25 21.2% 0.31 1.55 12.2%religion2 × DMPT 0.28 0.91 36.2% 0.24 0.83 40.6%

sgrp1 × DMPT 0.26 1.62 10.5% -0.15 -0.65 51.4% 0.23 1.55 12.1%sgrp2 × DMPT 0.25 1.78 7.6% -0.25 -1.25 21.2% 0.17 1.38 16.8%

hhtype2 × DMPT -0.32 -2.13 3.3% -0.01 -0.04 96.7% -0.19 -1.41 15.8%hhtype4 × DMPT -0.21 -1.46 14.5% 0.05 0.23 81.6% -0.14 -1.10 27.1%

DMPT -0.42 -1.66 9.7% -0.37 -1.30 19.3% -0.44 -2.11 3.5%

Note: See Table 4 in my paper for checking the results of the other variables.

30

Summary of the Main Results(1) The provision of the poverty alleviation program is

depressed by the decentralization. (2) The decentralization allocates less IRDP to the agricultural

labour households. The absolute poor in India are mainly from the agricultural labour households. This suggests that there is likely to be the local capture. The decentralization does not necessarily contribute the improvement of the welfare of the social disadvantaged group.

(3) However, the decentralization allocate more IRDP to SC. This may indicate the effect on the political reservation of the Panchayats for SC.

(4) There is no influence of the decentralization on the provision of the public works. In general, the public works contain the self-targeting mechanism. There may less discretionary manipulation of the public works by the local elite.

31

Section 5 Concluding Remarks • (1)The disadvantaged groups were not necessarily likely to

be the primary beneficiaries over others. →The female headed household and the landless group remained in a disadvantaged position throughout the entire period. (Section 4.2)

• (2) The possibility of local capture by the decentralization. But, public works are likely to be free from local capture. (Section 4.3)

• (3) The political reservation for the disadvantaged group may have the positive effect on the allocation of the program in favor of that group. (Section 4.3)

• (3) Regional socio political environment affects the allocation of the poverty alleviation programs.→ Greater inequality, lesser the provision of the poverty alleviation programs. (Section 4.2 and 4.3)

32

Thank you.