GROWTH, URBANIZATION AND POVERTY IN INDIA Ravallion... · Growth, Urbanization and Poverty...

MARTIN RAVALLION, EDMOND D. VILLANI CHAIR OF ECONOMICS, GEORGETOWN UNIVERSITY

GROWTH, URBANIZATION AND POVERTY IN INDIA

WRI Cities Research Seminar Series — February 11, 2016

Martin RavallionGeorgetown University

Growth,UrbanizationandPovertyReductioninIndia

MartinRavallionDept.Econ.GeorgetownUniversity

PresentationattheWorldResourcesInstitute,Feb.11,2016

Basedon:(i)GauravDatt,MartinRavallion,RinkuMurgai,“Growth,Urbanization

andPovertyReductioninIndia”,2016.(ii)MartinRavallion,TheEconomicsofPoverty:History,Measurement,

andPolicy,OxfordUniversityPress,2016.

Outline• Theurbanizationofpovertyglobally• DebatesinIndia• Dataissues• Overviewofdescriptivestatistics• Decompositionmethodsandresults• Somecomparisonsacrossstates• ConclusionsonIndiaandbroadercommentsontheurbanizationofpoverty

3

Theurbanizationofpovertyglobally

4

Debatesonurbanizationandpoverty

• Developmenteconomicshaslongseenpopulationurbanizationasadrivingforceforpovertyreduction,eventhoughinequalitymightriseasthepopulationurbanizes.

• LewisandKuznetsmodels.Latterintroducedinequalitywithinsectorsbutinastylizedway:populationurbanizationwithoutchangingdistributionwithin eitherurbanorruralareas =>

• Non-neutraldistributionalshifts:theoreticallyambiguousimplicationsforinequalitywithinsectors.

• Longstandingdebatesoverrelativeimportanceofpopulationurbanizationvswithin-sectordevelopment (ruralandurban).

• Inpolicycircles:Risingconcernsabouturbanpoverty.Restrictionsonmigrationintocities;under-servicedruralin-migrants.

5

TheKuznetsHypothesis:Inequality

6

Inequality

0 1

Between-group

Withingroup

Urbanpopulationshare

Totalinequality

TheKuznetsHypothesis:Poverty

7

Poverty

0 1Urbanpopulationshare

Urbanizationhastendedtocomewithlowerpovertyincidence

8

Acrosscountries,wefindthattheoverall(urbanplusrural)povertyratetendstobelowerwhentheshareofthepopulationlivinginurbanareasishigher.

Thisismostlyduetotheassociationbetweenurbanizationandeconomicgrowth.Ambiguousresultsoninequality.

0

20

40

60

80

100

0 20 40 60 80 100

Urban share of the population (%)

Nat

iona

l hea

dcou

nt in

dex

of p

over

ty (%

bel

ow $

2 a

day)

Theurbanizationofpovertymustbeputinanationalcontext

• Fallingnationalpovertyrateswithpopulationurbanizationmaycomemuchwithlessprogressagainsturban poverty.Indeed,wemayseerisingurbanpovertymeasures.

• First-ordergainstoruralmigrantstothecities.• Plusimportantsecond-roundimpacts ofurbanizationon

thelivingstandardsofthosewhoremaininruralareas:– higherremittancesfromurbanareas– thefactthattherearefewerpeoplecompetingforthe

availableemploymentinruralareas.

• Populationurbanizationcouldwelldomoretoreduceruralpovertythanurbanpoverty.

9

Anexampleoftheurbanizationofpoverty

10

Poverty

0 1Urbanpopulationshare

Rural

Urban

National

• Ruralpoormovetourbanareas.• Somemigrantsescapepoverty;therestremainpoor.

BackgroundonIndia

11

Earlyoptimismbutdisappointingprogress

• Post-independenceplannershopedthatIndia'surban-basedindustrializationprocesswouldbringlonger-termgainstopoorpeople,includingthroughrurallaborabsorption.

• ThathopewaslargelyshatteredbytheslowpaceofpovertyreductionintheperiodfromIndependenceuntilthe1980s.

• Why? Someobserverspointedtotheslowpaceoflaborabsorptionfromagricultureassociatedwiththemoreinward-lookingandcapital-intensivedevelopmentpathofthisperiod.

12

RelativelyslowpaceofurbanizationinIndia

• TheurbanpopulationsharehasbeenrisingsteadilyovertimeinIndia,from17%in1950to31%today.

• India’spaceofpopulationurbanization(proportionateincreaseintheurbanpopulationshare)hasbeenlessthaneitherSouthAsiaasawhole,orlowermiddle-incomecountriesasawhole.

• AndmarkedlyslowerthanforChina.TheurbanpopulationsharesofChinaandIndiawereaboutthesamearound1990,butthesharenowexceeds50%inChina.

13

Pre-1991• EconomicgrowthinIndiatendedtocomewithlowerpoverty

measures.• Theelasticityoftheincidenceofpovertywithrespectto

meanhouseholdconsumptionwas-1.3over1958-1991(DattandRavallion).

• Giventhemodestrateofgrowthoverthisperiod,successatavoidingrisinginequalitypriortothe1990swaskeytothisfinding.

• Higherabsoluteelasticitiesformeasuresofthedepthandseverityofpoverty,indicatingthatthosewellbelowthepovertyline havebenefitedfromeconomicgrowth,aswellasthosenearthepovertyline.

14

MajorpolicyregimechangeinIndia:Dismantlingofthe“licenseraj”

15

• Whilethereformprocesscanbedatedbacktotheearly1980s,“large-scale”reformstartedintheaftermathofthemacroeconomiccrisisof1991– Tradeliberalization

• Reductionintariffandnon-tariffbarriersonimports• Flexibleexchangerateandconvertibilityoftherupeeonthecurrentaccount

– Easingofrestrictionsondomesticandforeignprivatesector– Dilutionofstatecontrolofbankingandinsurance– Dismantlingofpublicsectormonopolies

• Evidenceofrisinginequalityinthewakeofthesereforms.

Newgrowthpathsinceearly1990s

• ThetrendrateofgrowthinIndia’sNetDomesticProduct(NDP)percapitaintheperiod1958-1991wasunder2%perannum,

• Butitwasmorethandoublethisrateintheperiodsince1992.

• TherewasmuchhopeinIndiathatthehighergrowthratesattainedinthewakeoftheeconomicreformsthatstartedinearnestintheearly1990swouldbringafasterpaceofpovertyreduction.

• However,thesignsofrisinginequalityinthepost-reformperiod,raisingdoubtsabouthowmuchthepoorhavesharedinthegainsfromhighergrowthrates.

16

TighteningofIndia’scasuallabormarkets

• Tighteningofruralcasuallabormarkets.– Risingrealwagerates.– Narrowingoftheurban-ruralwagegap(Hnatkovska andLahiri,

2013).• Why?

1. Schoolinghasexpanded,reducingthesupplyofunskilledlabor,especiallyinruralareas.

2. Therehasalsobeenadeclineinfemalelabor-forceparticipationrates.

3. Constructionboom=>

17

Constructionboom

• ConstructionboomacrossIndia,especiallyin(ruralandurban)infrastructure,whichhadbeenneglectedforalongperiod.

• Risinglabordemandfromconstruction=>higherwagesofunskilledlaborrelativetoskilledlabor within ruralareas,aswellasrisingruralrelativetourbanwages(formaleworkers).

• Unclearhowpermanentthischangewillprovetobe.– Itmaybeconjecturedthat(likeChina)IndiahasreacheditsLewis

TurningPoint.– However,otherfactorsleadingtohigherwagesevenwhilethereisstill

ruralunderemployment.– Andreversalsmightbeexpected.

18

Thispaper

19

NewdataseriesonpovertyinIndia

• Forthepurposeofthispaperwehavecompiledanewdataseriesonpovertyandrelateddataspanning60years,extendingtheperiodofanalysisinpastresearch.

• Withthebenefitofnearlytwodecadesofpost-1991data,webelievethereisnowsufficientdataforthepost-1991periodtorevisittheearlierfindingsonthepovertyimplicationsoftherateandpatternofgrowthinpost-reformIndia.

• Attribution toreformsperse isproblematic,butafurtherscrutinyoftheemergentpropertiesofthechanginggrowthprocesswithrespecttopovertyreductionisclearlyimportant.

20

Decompositionsofthechangesinpovertymeasures

• Decompositionsidentifyingcontributionofurbanversusruraleconomicgrowth,aswellaspopulationurbanization.

• DecompositionofpovertyreductionbysectorofNDP.

• Newdecompositionmethodthatallowsustoidentifythedifferencebetweenpopulationurbanizationeffectswithconstantwithin-sectordistribution(asintheKuznetsprocess)versuschangingwithin-sectordistributions.

21

Dataissues

22

23

Cross-countrycomparisonscanbedeceptive

• Pastresearchhasreliedoncross-countrycomparisons– Singlecross-sections(suchasinthemanytestsoftheKuznets

hypothesis)– Sometimesusingpaneldata,thoughthetypicallyshorttime-

serieshasmeantthatthecross-countryvariabilityisdominant.

• However,developmentovertimewithincountriesneednotaccordwiththecross-countrypatterns.

• Desirabletohaveareasonablylongtimeseriesofsurveys;ashortseriescanbedeceptiveforinferringatrend.

24

DataissuesforIndia• Amongstdevelopingcountries,Indiahasthelongestseries

ofnationalsurveyssuitablefortrackinglivingconditions.– Highlycomparablesurveysupto1999/2000– Changesinsurveydesignin1999/2000createdaserious

comparabilityproblem– Newsurveysreasonablycomparableto1993/94andpriorsurveys– Uniformvs.mixed(post-early)recallperiods– Surveycomplianceproblems,esp.,urbanareas– WorryingdiscrepanciesbetweenNSSandNAS.

• Twopovertylines:– LowerlineanchoredtohistoricallinesofthePlanningCommission– Upperlineanchoredtotheinternationallineof$1.25adayat

2005PPP(about$1.90adayat2011PPP)• Threepovertymeasures:

– Headcountindex,povertygapindexandsquaredpovertygapindex(Foster,Greer,Thorbecke,1984)

Overviewoftimeseries

25

GrowthandpovertyinIndiaover60years

0

20

40

60

80

100

7.0

7.5

8.0

8.5

9.0

9.5

10.0

10.5

11.0

1950 1960 1970 1980 1990 2000 2010 2020

Net domestic product per capita(log, right axis)

Headcount index(%, left axis)

Urban population share(%, left axis)

Total

Primary

Second-ary

Tertiary

Lowerline

Upperline

26

Anti-Kuznets1:fallingthenrisinginequalitywithin sectors

25.0

27.5

30.0

32.5

35.0

37.5

40.0

42.5

45.0

1950 1960 1970 1980 1990 2000 2010 2020

Gin

i in

de

x (%

) Urban

Rural

27

Anti-Kuznets2:U,notinvertedU!Risingbetween-sectorinequality

since1970

1.05

1.10

1.15

1.20

1.25

1.30

1.35

1.40

1.45

1950 1960 1970 1980 1990 2000 2010

Rat

io o

f urb

an m

ean

to ru

ral m

ean

(bot

h in

con

stan

t rur

al p

rices

)

Signsoflevellingoff

28

Convergenceofruralandurbanpovertymeasures

0

10

20

30

40

50

60

70

1950 1960 1970 1980 1990 2000 2010 2020

RuralUrbanNational

Hea

dcou

nt in

dex

(%, l

ower

line

)

29

Convergenceofruralandurbanpovertymeasures

-4

0

4

8

12

16

20

1950 1960 1970 1980 1990 2000 2010 2020

Upper line Lower line

Rur

al m

inus

urb

an h

eadc

ount

inde

x (%

)

Upper line

Lower line

30

UrbanizationofpovertyinIndia

0

4

8

12

16

20

24

28

32

36

1950 1960 1970 1980 1990 2000 2010 2020

Lower line

Upper line

Urb

an

sh

are

of t

he

po

or

(%)

31

Growthelasticitiesofpovertyreduction

32

Post-reformgrowth;acceleratedpovertyreduction,butrisinginequality

• Significantspurtineconomicgrowth,drivenbygrowthinthetertiaryand(toalesserextent)secondarysectors.

• Thepaceofpovertyreductionalsoacceleratedpost-1991,witha3-4foldincreaseintheproportionaterateofdeclineinthepost-91period.

• Theaccelerationinruralpovertydeclinewasevenhigherthanthatforurbanpoverty.

• Thishappenedalongsideasignificantincreaseininequalitybothwithinandbetweenurbanandruralareas.

33

Highergrowth+highergrowthelasticitiespost-1991

• Despitetheincreaseininequality,wefindgreaterpost-91responsivenessofpovertytogrowthintheaggregate.

• Thisholdsregardlessofwhethergrowthismeasuredbasedonnationalaccountsorsurvey-basedconsumption.

• AlsorobusttoanallowanceforcorrelatedmeasurementerrorsusingIVsdrawnfromotherdata.

34

Elasticities(Regression)

Meanconsumptionpercapita(NSS)

OLS IV Elasticity t-stat Elasticity t-statHeadcount:higherline Wholeperiod -1.45 -10.8 -1.32 -19.3Pre-1991 -1.13 -18.2 -1.11 -31.4Post-1991 -1.99 -34.2 -1.98 -37.8H0:pre-91=post-91 prob>F(1,35)orF(1,34) 0.00 0.00Povertygap:higherline Wholeperiod -2.34 -17.8 -2.26 -26.0Pre-1991 -1.99 -15.1 -1.96 -23.0Post-1991 -2.79 -30.3 -2.71 -40.0H0:pre-91=post-91 prob>F(1,35)orF(1,34) 0.00 0.00Squaredpovertygap:higherline Wholeperiod -3.00 -24.0 -2.98 -31.5Pre-1991 -2.65 -13.2 -2.57 -18.0Post-1991 -3.30 -26.1 -3.18 -35.7H0:pre-91=post-91 prob>F(1,35)orF(1,34) 0.01 0.00

35

Regression-baseddecompositions

36

37

Urban-ruralregressiondecomposition

• Meanincome:• Growthrate:

• Testequation:

• Nullhypothesis:

ut

ut

rt

rtt nn µµµ +=

rt

ut

rt

ut

rt

ut

ut

rt

rtt nnnssss ln)]/([lnlnln Δ−+Δ+Δ=Δ µµµ

tit

it

it ns µµ /=

trtu

t

rtu

trt

nut

ut

urt

rt

rt n

nnssssP εηµηµηη +Δ−+Δ+Δ+=Δ ln).(lnlnln 0

H0: ηη =i for i=r,u,n Populationurbanization:Kuznets+within-sectordistributionalshifts

EncouragingsignsthatIndia’surbaneconomicgrowthisbenefitingtheruralpoor

• Regime1(Pre-1991): Urbaneconomicgrowthhelpedreduceurbanpovertybutbroughtlittleornooverallbenefittotheruralpoor.Themaindrivingforceforoverallpovertyreductionwasruraleconomicgrowth.

• Regime2(Post-1991): Asbefore,urbangrowthreducedurbanpoverty,andruralgrowthreducedruralpoverty.

• Butmuchstrongerevidenceofapositivefeedbackeffectfromurbangrowthtoruralpovertythaninthepre-1991data.

38

39

Urban-ruraldecompositionbeforeandafter1991

(Lowerline;headcountindex) Pre-1991 Post-1991

Growthrateofmeanruralincome(share-weighted)

-1.28 -1.96(-11.1) (-3.59)

Growthrateofmeanurbanincome(share-weighted)

0.20 -6.40(0.30) (-4.26)

Populationshifteffect(logpoints) -0.23 0.26(-2.76) (1.30)

R2 0.87

Povertyreductionandtheurban-ruralcompositionofgrowth

trtu

t

rtu

trt

nut

ut

urt

rt

rt n

nn

ssssP εηµηµηη +Δ−+Δ+Δ+=Δ ln).(lnlnln 0

Effectsacrosssectors:Headcountindex

Change in log poverty measure National Urban Rural Coeff. t -stat Coeff. t -stat Coeff. t -stat HEADCOUNT: Higher line

Pre-91 Urban growth -0.023 -0.04 -0.624 -7.54 0.694 1.39 Pre-91 Rural growth -0.940 -12.2 -0.014 -0.65 -0.921 -15.4 Pre-91 Pop. urbanization -0.146 -3.47 -0.015 -1.84 -0.121 -3.32 Post-91 Urban growth -3.590 -4.37 -1.244 -9.01 -2.423 -3.24 Post-91 Rural growth -2.076 -7.41 -0.165 -1.68 -1.918 -7.37 Post-91 Pop. urbanization 0.195 2.47 -0.039 1.04 0.173 2.66

R-squared 0.905 0.850 0.900 Pre91=Post91 prob.> F(2,33) 0.000 0.006 0.000 Pre91=Post91 prob.> F(3,33) 0.000 0.005 0.000

40

Effectsacrosssectors:SPGindex

Change in log poverty measure National Urban Rural Coeff. t -stat Coeff. t -stat Coeff. t -stat SQUARED POVERTY GAP: Higher line

Pre-91 Urban growth -1.212 -1.10 -0.394 -1.30 -0.387 -0.41Pre-91 Rural growth -1.990 -7.68 -0.130 -2.71 -1.876 -7.64Pre-91 Pop. Urbanization -0.259 -1.59 -0.051 -2.16 -0.221 -1.48 Post-91 Urban growth -8.482 -3.44 -3.609 -9.05 -5.311 -2.79Post-91 Rural growth -1.772 -2.03 -0.356 -1.84 -1.388 -2.00Post-91 Pop. Urbanization 0.269 1.04 0.252 3.46 0.080 0.40

R-squared 0.840 0.819 0.810 Pre91=Post91 prob.> F(2,33) 0.020 0.000 0.063 Pre91=Post91 prob.> F(3,33) 0.012 0.000 0.071

41

Insummary• Structuralbreakaround1991intherelationshipbetween

povertyandthecompositionofgrowth.

• Bothurban-ruralandsectoral(output)decompositionsaresuggestiveofstrongerinter-sectorallinkages,wherebygrowthinonesectortransmitsitsgainselsewhere.

• Post-91,urbangrowthhasemergedastheprimarydriverofpovertyreduction– Directly,asurbanpovertyhasbecomesignificantlymoreresponsiveto

urbangrowth,

– Indirectly:urbangrowthhasbecomesignificantlymoreruralpovertyreducingsince1991.

42

Unifieddecomposition

tttut

rtt IKNGGPE ++++=Δ )ln(

• rG and uG :ruralandurbangrowthinconsumption.

• N:effectofthepopulationshiftcontrollingforgrowthinmeanconsumptionwithineachoftheurbanandruralsectors.Thusthistermalsoreflectsanywithin-sectordistributionaleffects.

• K: theKuznetseffectofpopulationshiftholdingwithin-sectorpovertylevelsconstant.

• I :theinteractioneffectsbetweensectoralpovertychangeandpopulationshift.

43

Unifieddecompositionresults Componentsofpredictedchangeinpoverty: GR GU N K I

Totalpredictedchangeinpoverty

Ruralgrowth

Urbangrowth

Populationshiftwithintra-sectordistributional

change

Populationshiftholdingintra-

sectoraldistribution

constant(Kuznetseffect)

Interactionbetween

sectoralpovertychangeand

populationshift

Headcount:higherline Pre-91 Annual%agechange -1.16 -0.53 0.01 -0.60 -0.04 0.00Shareofpred.annualchange 100.0 45.7 -1.0 52.0 3.2 0.1Post-91 Annual%agechange -4.67 -3.26 -2.93 1.55 -0.04 0.01Shareofpred.annualchange 100.0 69.8 62.7 -33.1 0.9 -0.2

Povertygap:higherline Pre-91 Annual%agechange -1.86 -0.86 -0.12 -0.89 0.00 0.00Shareofpred.annualchange 100.0 46.3 6.3 47.7 -0.2 -0.2Post-91 Annual%agechange -6.49 -2.71 -5.32 1.52 0.01 0.01Shareofpred.annualchange 100.0 41.7 82.0 -23.4 -0.2 -0.2 44

45

SectoralNDPdecompositions

(Lowerline;headcount) Pre-1991 Post-1991Primary(share-weighted) -1.14 -1.97

(-4.59) (-0.22) Secondary(share-weighted) 4.50 -1.64

(2.71) (-0.32) Tertiary(share-weighted) -3.58 -1.51

(-3.20) (-1.07 Primary+Secondary+tertiary

-1.55 (-2.79)

R2 0.66

t

n

iititit YsP εππ +Δ+=Δ ∑

=10 lnln

Povertyreductionandthesectoralcompositionofgrowth

Change in log Headcount

(Lower Line) Change in log Headcount (Higher Line)

Unrestricted

model Restricted model Unrestricted

model Restricted

model Variable or statistic Coeff. t-ratio Coeff. t-ratio Coeff. t-ratio Coeff. t-ratio Pre-91 variables: Primary sector growth -1.138 -4.59 -1.143 -4.58 -0.823 -3.41 -0.817 -3.43 Secondary sector growth 4.496 2.71 4.510 2.75 2.941 2.64 2.921 2.66 Tertiary sector growth -3.582 -3.20 -3.581 -3.28 -2.395 -3.30 -2.396 -3.42 Post-91 variables: Primary sector growth -1.965 -0.22 -2.053 -0.33 Secondary sector growth -1.637 -0.32 -0.548 -0.15 Tertiary sector growth -1.510 -1.07 -1.360 -1.33 NDP growth -1.551 -2.79 -1.199 -2.95 Variables common to both periods: Change in log ratio of CPI to NDP deflator 1.224 4.44 1.235 4.22 0.847 4.08 0.831 3.76 Change in binary var. for an MRP estimate -0.320 -4.73 -0.321 -5.45 -0.219 -4.91 -0.224 -5.56 Number of observations 40 40 40 40 R-squared 0.656 0.655 0.660 0.658 Wald test of restrictions: prob > F(2, 32) Pre-91:π1=π2=π3 0.002 0.008 Post-91:π1=π2=π3 0.999 0.977

Note: The Table gives least squares estimates with heteroscedasticity and autocorrelation-consistent standard errors.

46

Change in log Poverty Gap

(Higher Line) Change in log Squared Poverty Gap (Higher Line)

Unrestricted

model Restricted

model Unrestricted

model Restricted

model Variable or statistic Coeff. t-ratio Coeff. t-ratio Coeff. t-ratio Coeff. t-ratio Pre-91 variables: Primary sector growth -1.378 -5.47 -1.367 -5.56 -1.763 -5.13 -1.751 -5.30 Secondary sector growth 4.931 2.66 4.899 2.72 6.354 2.38 6.319 2.45 Tertiary sector growth -4.128 -3.50 -4.129 -3.64 -5.376 -3.30 -5.377 -3.43 Post-91 variables: Primary sector growth -0.217 -0.02 1.747 0.15 Secondary sector growth -1.349 -0.25 -2.104 -0.30 Tertiary sector growth -1.856 -1.28 -2.138 -1.18 NDP growth -1.624 -2.86 -1.880 -2.71 Variables common to both periods: Change in log ratio of CPI to NDP deflator 1.427 4.46 1.403 4.73 1.864 4.38 1.837 4.94 Change in binary var. for an MRP estimate -0.354 -4.68 -0.351 -5.59 -0.458 -4.32 -0.448 -5.19 Number of observations 40 40 40 40 R-squared 0.668 0.666 0.652 0.648 Wald test of restrictions: prob > F(2, 32) Pre-91:π1=π2=π3 0.005 0.019 Post-91:π1=π2=π3 0.981 0.945

Note: The Table gives least squares estimates 47

Decompositionofchangeinnationalmeasures

Componentsofpredictedchangeinpoverty:

Totalchange

inpoverty

Totalpredictedchangeinpoverty

Totalpredictedchangeinpovertywithout

populationgrowth

PrimarysectorNDP

growth

SecondarysectorNDP

growth

TertiarysectorNDP

growth

NADeflator-CPIdrift

Headcount:higherline Pre-91 Annual%agechange -0.90 -0.78 -1.80 -0.78 3.41 -4.47 0.05Shareofpredictedannualchange 100.0 43.1 -189.6 249.0 -2.6Post-91 Annual%agechange -4.79 -5.87 -7.89 -0.73 -1.87 -4.99 -0.25Shareofpredictedannualchange 100.0 9.3 23.7 63.3 3.1SquaredPovertyGap:higherline Pre-91 Annual%agechange -2.68 -1.89 -4.23 -1.66 7.37 -10.04 0.10Shareofpredictedannualchange 100.0 39.3 -174.1 237.2 -2.4Post-91 Annual%agechange -8.16 -9.35 -12.52 -1.14 -2.94 -7.82 -0.54Shareofpredictedannualchange 100.0 9.1 23.4 62.5 4.3

48

Insummary• Post-1991datasuggestsector-neutrality inthepoverty

reducingeffectofgrowthinnetdomesticproduct.– Unlikethepre-91period,whenonlyprimaryandtertiarysectorgrowthcontributedtopovertyreduction,after91allthreesectorshavehadasignificantimpact.

• Thetertiarysectorhasthehighest(absolute)growthelasticityofpovertyreduction,abouttwiceashighasthosefortheprimaryandsecondarysector.

• ThisreflectsboththechangingnatureofthegrowthprocessaswellasthelargestructuraltransformationoftheIndianeconomyoverthelasttwodecadeswiththesecondaryandtertiarysectorsnowaccountingformuchlargersharesofnationaloutputandemployment.

49

50

Comparisons across states: The economic geography of

poverty reduction

51

Trend rates of poverty reduction by state (1970-2000)

0.4

0.5

0.6

0.6

0.8

0.8

0.9

0.9

1.1

1.1

1.2

1.2

1.3

1.8

0.0 0.5 1.0 1.5 2.0

Assam

Bihar

Punjab & Haryana

Uttar Pradesh

Karnataka

Madhya Pradesh

Rajasthan

Maharashtra

Orissa

Andhra Pradesh

Gujarat

Tamil Nadu

West Bengal

Kerala

% points per year

India:Somelessonsfromsub-nationaldata

52

Whyhaspovertyfallensomuchfasterinsomestatesthanothers?

• Higher average farm yields, higher public spending on development, higher non-farm output and lower inflationwere all poverty reducing in India

• Agricultural growth, development spending and inflation had similar effects across states

• However, the response of poverty to non-farm outputgrowth in India varied significantly between states.

• The states with initially higher levels of human development saw greater poverty impacts from non-farm growth. Better infrastructure also helped.

εηπγβββ itiiitiitGOViit

YLDiit

NFPiit + + t INF +GOV YLDNFP = P +++ lnlnlnln

53

India: Elasticities of poverty to non-farm economic growth

0.0

0.5

1.0

1.5

2.0

2.5

2 4 6 8 10 12 14

H PG SPG

Elasticitiesofpovertytonon-farmoutput

Kerala WB

Bihar

AP

54

Initial conditions matter to the impact of growth on poverty

• Low farm productivity, low rural living standards relative to urban areas and poor basic education all inhibited the prospects of the poor participating in growth of India’s non-farm sector.

• Rural and human resource development appear to be strongly synergistic with poverty reduction through an expanding non-farm economy.

55

Conclusions

FallingpovertywithchangingsectoralpatternofgrowthinIndia

• Economicgrowthhasnotonlycomewithalowerincidenceofabsolutepovertybuttherehasbeenanacceleration inthepaceofprogressagainstpovertypost-1991.

• ThenewpatternofgrowthhasbroughtgreaterbenefitstoIndia’spoor.

• Whiletherehasbeenrisinginequalitywithintheruraland(especially)urbansectors,growthwithinsectorshasdeliveredsufficientgainstoIndia’spoortomitigatehigherinequality.

• TheKuznetsprocesshasplayedlittlerole.OnconsideringthesectoralpatternofNDPgrowthwefindanindicationthatthesectoralpatternofgrowthmatterslesstoprogressagainstpovertythanwasthecaseinthepre-1991period.

56

Newsectoralpatternofpovertyreduction

• Thecontributionofprimarysectorgrowthhasrapidlydwindledfromaccountingforabouttwo-fifthsofthetotalpovertydeclinepre-91tolessthan10percentofthetotal(andlarger)povertydeclinepost-91.

• Thetertiarysectoralonehascontributedover60%ofthepost-91povertyreduction.

• Thesecondarysectorgrowthhascontributedaboutaquarter.India’sconstructionboomsince2000hasclearlyhelpedassureamorepro-poorgrowthprocessfromthesecondarysector,althoughthesustainabilityofthischangeisunclear.

57

Urbanizationofpoverty

• Urbaneconomiescreatenewopportunitiesthatpoorpeopleinruralareashaveoftensoughtouttoimprovetheirlives.

• Distortedurbanlabormarketscanreadilycreateexcessiveurbanization,

• …ascanthelackofeffectivepubliceffortstopromoteagricultureandruraldevelopment;indeed,manydevelopingcountrieshavegoneevenfurtherin(explicitlyorotherwise)taxingtheruraleconomytosupporttheurbaneconomy.

• However,theurbanizationofpoverty—wherebypovertyratesfallmoreslowlyinurbanareasthaninruralareas—istobeexpectedinadevelopingcountrythatissuccessfulinreducingpovertyoverall.

58

Betweenarockandahardplace

• Poorpeopleareoftentrappedasthevictimsofpoliciesthatsimultaneouslyrepressagriculturewhilemakinglifedifficultforruralmigrantstothecities.– Removinglong-standingpolicybiasesinbothtaxationand

publicspendingremainsahighpriorityforpro-poorgrowth.– Nolessmisguidedarerestrictionsonmigrationandurban

policiesthatunder-supplyservicestopoorurbanresidents,includingruralmigrants.

• Morepro-poordevelopmentpolicieswillprobablyentailanurbanizationofpoverty,butthatshouldnotbeacauseforalarmaslongaspovertyisfallingoverall.

59

Formoreonthistopic:economicsandpoverty.com

Thankyouforyourattention!

60