Income Volatility and Social Protection in Developing Asia

8/22/2019 Income Volatility and Social Protection in Developing Asia

1/42

November 2006

ERDECONOMICS AND RESEARCH DEPARTMENT

Working PapeSERIESNo.

88

Vandana Sipahimalani-Rao

Income Volatility and SocialProtection in Developing AsiaIncome Volatility and SocialProtection in Developing Asia


2/42

ERD Working Paper No. 88

Income VolatIlItyand SocIal ProtectIon

In deVeloPIng aSIa

Vandana SIPahImalanI-rao

noVember 2006

Vandana Sipahimalani-Rao is an independent consultant/researcher and human development economist. She is a former

economist at the World Bank (South Asia Human Development). The author thanks Rana Hasan and Ajay Tandon fortheir valuable comments and participants of the Asian Development Bank workshop for helpful suggestions.


3/42

Asian Development Bank6 ADB Avenue, Mandaluyong City

1550 Metro Manila, Philippineswww.adb.org/economics

2006 by Asian Development BankNovember 2006

ISSN 1655-5252

The views expressed in this paperare those o the author(s) and do notnecessarily refect the views or policies

o the Asian Development Bank.


4/42

FoREWoRD

The ERD Working Paper Series is a orum or ongoing and recently completedresearch and policy studies undertaken in the Asian Development Bank or onits behal. The Series is a quick-disseminating, inormal publication meant tostimulate discussion and elicit eedback. Papers published under this Series

could subsequently be revised or publication as articles in proessional journalsor chapters in books.


5/42

CoNtENts

Abstract vii

I. INTRODUCTION 1

II. CONSEQUENCES OF VULNERABILITY TO INCOME RISK 2

A. Impact on Productivity and ProtsA. Impact on Productivity and Prots 2 B. Investing in Human Capital 4

C. Unemployment, Income Loss, and Resistance to Reorm 7

III. MITIGATING VULNERABILITY TO INCOME VARIABILITY/SHOCKS:THE ROLE OF SOCIAL PROTECTION 8

A. Inormal Social Protection MechanismsA. Inormal Social Protection Mechanisms 8 B. Formal Social Protection Mechanisms 11

IV. POLICY IMPLICATIONS AND CONCLUSIONS 15

SELECTED REFERENCES 2SELECTED REFERENCES 21


6/42

AbstRACt

Most developing countries in Asia have introduced market-oriented economic

reforms and face increasing exposure to global markets. Households in these

countries thus have greater opportunities to raise their standard of living.

However, they also face riskier environments and typically have to cope with

greater income volatility. This paper undertakes a critical review of the social

protection literature to understand the impact of vulnerability to income risk

in developing Asian economies, and the role of social protection in managing

that risk. The paper highlights the fact that social protection mechanisms areimportant in helping vulnerable households mitigate as well as reduce risk, to

enable them to invest in high-risk but high-return activities. Thus public policy

plays an important role in both protecting and empowering vulnerable individuals.

Key policy implications are suggested based on the critical review.


7/42

I. INtRoDUCtIoN

Most developing countries across the world including those in Asia have introduced market-oriented economic reorms and ace increasing exposure to global markets. While this presentsindividuals and households with opportunities to raise their standard o living, it also implies theyhave to conront riskier environments and, in particular, greater income volatility. In agriculture,

access to higher yielding varieties o inputs and cash crop cultivation is oten accompanied bymore uncertainty in crop yields as higher mean yields are usually associated with greater varianceas well. In industry, increasing domestic and international competition puts pressure on public and

private sector rms to restructure and/or shut down, resulting in increased unemployment and losso income. This paper undertakes a critical review o the social protection literature to understand

the impact o vulnerability to income risk in developing Asian economies, and the role o socialprotection in managing that risk. The central conclusion o this paper is that policymakers in

these countries need to intensiy their eorts to ensure adequate social protection to vulnerablehouseholds with a two-old objective: to empower them to cope with increased risk as well as totake advantage o opportunities to improve their standard o living by investing in high-risk buthigh-return activities.

Social protection aims to assist individuals to transorm the negative cycle o poverty intoa positive cycle o growth and human development (Ortiz 2001, 41). However, almost 80 percento individuals worldwide have little or no social protection coverage (United Nations 2005). This

paper specically highlights the eect o vulnerability to income volatility or the poor ruraland urban households usually employed in the inormal sector as these households are the least

adequately equipped to deal with income-related risk. The inormal sector is dened dierently indierent countries. Most denitions include households that are either sel-employed or work oremployers who hire a small number o workers. However, the inormal sector is a heterogeneousgroup o individuals ranging rom contract workers, to sel-employed, to home-based workers, tothose working or small organizations. According to a recent United Nations report (2005), about

5075% o nonagricultural employment in the majority o developing countries is in the inormalsector. Employment in the inormal sector in developing Asia has increased in recent years, makingprovision o appropriate social protection to them a vital priority or policymakers. (See Felipe andHasan 2006 or an overview o the growing role o the inormal sector in employment).

This paper also ocuses on the social protection needs o the urban ormal sector dealingin particular with unemployment. As developing countries in Asia move toward more open market

policies and strive to be better integrated with the global economy, large-scale retrenchment in theurban ormal sector is becoming a reality that they have to contend with. To be competitive in aglobal economy, many loss-making public enterprises in developing Asian countries are being closed,resulting in large-scale unemployment. Privatization is increasingly becoming a common eature as

well. In addition, more liberal labor laws in some countries are making it easier or the private sectorto hire and re in order to stay competitive. Policymakers need to make it a priority to provide


8/42

adequate unemployment benet compensation and support services to prevent these workers romalling into poverty. Further, protecting workers will have the added benet o reducing resistanceto restructuring programs and a better political consensus or market-oriented reorms.

The paper is organized as ollows. This section, Section I, has introduced the issues the paper

will discuss. Section II discusses the impact o vulnerability to income risk. Section III describesthe inormal and ormal social protection mechanisms available to households in developing Asiancountries or dealing with this income risk. It highlights the relative advantages and disadvantages

o these mechanisms. Finally, Section IV draws policy implications based on the critical literaturereview, suggests priorities or policymakers in these countries, and concludes.

II. CoNsEQUENCEs oF VULNERAbILItY to INCoME RIsK

Households in developing Asian countries are exposed to many kinds o risks. The poor, in

particular, are more vulnerable to both natural and man-made shocks. This paper ocuses on income-related risks. Households that are already poor are particularly susceptible to income shocks andtake decisions to avoid or to cope with them i they have already occurred. Oten these decisions

cause the households, who are already vulnerable, to urther lose income and make them even morevulnerable in the long run. Vulnerability to income shocks and risks can thus make poor householdsspiral into a cycle o poverty and heightened vulnerability (see Ravallion 1988 or an empirical studyo the negative impact o income variability on poverty in India). This section explores some o

the key consequences o vulnerability to income volatility or poor households in Asia. It highlightsthe vicious cycle o vulnerability and poverty.

A. Impac n Prduciviy and Prf

Vulnerability to income variability could lead individuals and households to underinvest in

high-risk activities that maximize productivity and/or prots.

The empirical literature on this issue is airly extensive or the agricultural sector. The maintheme that emerges rom this literature is that agricultural households in developing countries

are usually aced with highly uncertain environments due to risks associated with the weatherand technology. They resort to suboptimal choices to cope with risk due to limited insurance andcredit availability. Some households are orced to make decisions to reduce the income risk they areexposed to ex-ante by taking production or employments decisions to smooth income. This is direct

income smoothing. Other households make nonoptimal ex-post decisions to deal with the eects oincome shocks, i.e., in the ace o reduced income they try to smooth consumption through variousmeans. This is called consumption smoothing. Both these ex-ante and ex-post decisions are usually

not optimal rom the point o view o maximizing current and/or uture productivity and prots.

Morduch (1995, 104) explains that agricultural households can use production or employment

decisions to smooth income as a sel-preservation technique. He cites various income smoothingmethods to reduce the variability o agricultural output in order to reduce exposure to income risk,drawing on dierent empirical studies. He concludes that mitigating risk through production choices

November 2006

INcome volatIlItyaNd SocIal ProtectIoNIN develoPINg aSIa

vaNdaNa SIPahImalaNI-rao


9/42

can be costly, since typically expected prots must be sacriced or lower risk. A common methodor income smoothing is to use agricultural inputs less intensively than is optimal to reduce theamount o nancial investment in risky inputs. Morduch cites that armers in North India do thiswhen using ertilizer, or example. Farmers could substantially raise expected prots by increasing

applications o ertilizer, but by using less ertilizer, investment losses are reduced in bad times(Morduch 1995, 109). Morduch also cites that armers wait until they have better rainall inormationto make production choices, oten resulting in reduced prots. Waiting or weather inormation can

help smooth income but it can have a negative impact on crop yields as well. Morduch (199) ndsthat households in his sample rom South India had devoted a relatively large share o land tocultivating less risky varieties o crops like rice and less land to higher yielding but riskier seeds.Eswaran and Kotwal (1989) nd that armers who had more access to consumption credit were

quicker to adopt high-yielding variety seeds during the Green Revolution in India in the 1960s.

Selling assets happens ex-post an income shock and is a method used by vulnerable householdsto smooth consumption. Adams and Alderman (1992) nds considerable empirical evidence o

consumption smoothing behavior using panel data rom Pakistan. Rosenzweig and Wolpin (1993)estimate a model o agricultural investment behavior using household data rom India to illustrate aninstance o consumption smoothing. They study the sale and purchase o bullocks that are valuableagricultural assets or these households used or ploughing the elds. They conclude that due to

incomplete credit markets and borrowing constraints, households use the sale o bullocks to smoothconsumption. They use bullock sales or liquidity when income levels drop due to poor harvests orother income shocks. These households thus sacrice agricultural productivity since they do not

buy/sell the optimal number o bullocks rom a prot-maximizing or productivity-maximizing pointo view. This leads to urther loss o uture income since they may not have the optimal numbero bullocks or the next harvest season.

The above studies are part o a growing body o literature that illustrates that consumptionand production decisions may not be separable in agricultural households or vulnerable householdsin developing countries, in the absence o complete markets. Both ex-ante (income smoothing) andex-post (consumption smoothing) are risk coping methods that can be very costly to household rom

a productivity, prot, and medium/longer-term income viewpoint (see Alderman and Paxson 1992 ora review o the income and consumption smoothing literature). While individuals undertaking incomesmoothing behavior to avert risk are likely to sacrice prots and income in the short term, those

resorting to consumption smoothing may orego uture income (due to a less than optimal numbero productive assets, or example). Binswanger and Rosenzweig (1993, 75), or instance, nd thator the bottom wealth quartile, income smoothing would reduce arm prots by 35 percent. Theextent o their vulnerability has been refected in recent years in an increasing number o suicides

by poor armers in India (see Box 1).

erd WorkINg PaPer SerIeS No. 88

SectIoN II

coNSequeNceS of vulNerabIlIty to INcome rISk


10/42

box 1Vulnerable FarmerS commItSuIcIde

Some 8,900 armers committed suicide in the our states o Andhra Pradesh, Karnataka, Kerala,

and Maharashtra in India between 2001 to 2006. This is an alarming statistic. Gujarat and Punjab

have also witnessed high rates o armer suicide in recent years. This is particularly distressing sincethese states (which have seen the largest numbers o armer suicide in recent years) are agriculturally

advanced. The media has largely blamed liberalization, global competition, and new technology orarmers who are increasingly investing in high-value but high-risk crops such as cotton, and are being

caught in a debt trap that they eel they cannot escape when crops ail. As Suri (2006, 1526) suggests,the sequence is more or less the sameheavy investments on inputs, crop ailure, inability to pay back

the loans and mounting debts leading to suicide. Some media analysis has also highlighted the lack oaccess to institutional credit as a key culprit leading to indebtedness, with armers paying exceptionallyhigh interest rates to moneylenders and commission agents (Jodhka 2006). In an interesting survey

conducted in Maharashtra (see Economic and Political Weekly2006, 1544), controlling or other actors,Mishra nds that i outstanding debt increases by Rs.1,000 then the odds that the household is one

with a suicide victim increases by 6 percent and i the household owns bullocks then the odds that it isa household with a suicide victim decreases by 65 percent.

It is clear rom these observations that indebtness is the major cause o suicide or poor andvulnerable armers. While new technologies and global competition present opportunities or agriculture,most poor armers vulnerable to income volatility do not have the resources and acilities needed to take

advantage o these opportunities without alling into a debt spiral. It is imperative that policymakersocus on increasing access to credit at reasonable terms and better protection rom risk in the orm o

subsidized agricultural insurance where possible to enable them to ree themselves rom being dependenton moneylenders and commission agents who charge exorbitant rates and oten lock input prices at

unreasonably high levels (commission agents oten double as input suppliers to poor armers).

Source: Based on a series o articles in the Economic and Political Weekly(2006).

b. Inveing in Human Capial

Risk aversion and vulnerability to income shocks/variability can also dissuade other kinds oinvestments with potentially high returns. Vulnerable households tend to discount the uture and anyinvestment decisions o an intertemporal nature could be negatively aected by this discounting.

Households sometimes hesitate to invest in the education o their children or pull them out oschool due to economic shocks. This can have a detrimental impact on the economy in the longrun where human capital investments are suboptimal. In addition, these kinds o decisions can

also have the same entrapment eect discussed earlier where vulnerability and poverty perpetuatesurther vulnerability and poverty due to lack o education.

The literature on this issue is not very extensive. The child labor literature emphasizes the

schoolinglabor choice that parents make or their children in poor households. The literatureocuses, or the most part, on the link between poverty and an increased supply o child labor.While vulnerability is, o course, linked to poverty, most o the studies dealing with child labor andschooling do not address vulnerability, exposure to risk, and their relationship with human capital

investment. Nevertheless, some studies do explore this connection.

The traditional investment model o household decision making regarding education or childrenassumes perect capital markets in order to enable parents to borrow against uture earnings to

November 2006




11/42

nance current schooling. Baland and Robinson (2000) propose a theoretical economic model othe household decision where they choose how to divide their childrens time between work andschool. One o their central conclusions is that child labor is more prevalent where capital marketsare imperect since households are credit-constrained and cannot borrow easily in lean income

periods (see Parsons and Goldin 1989, and Ray 2001 or similar theoretical models).Grootaert and Patrinos (1999)include as one o the key economic variables determining the

supply o child labor household ownership o income generating assets. As they explain, the more

assets owned by a household and the more diverse they are, the better its ability to manage riskand the less likely its need to use child labor to insure against income fuctuations (Grootaert andPatrinos 1999, 20). The empirical studies described in this book nd it dicult, or the most part,

to distinguish between the impact o poverty and that o vulnerability to income variability on thechild labor versus schooling decision. All the studies, including one rom Asia on the Philippines,nd that belonging to the lowest income/expenditure quintile signicantly increases the probabilityo a child working rather than going to school. The authors venture that this explanatory variable

can be an indicator o poverty but also o susceptibility to risk, since these would be the householdsmost constrained in terms o access to credit and insurance. However it is dicult to empiricallyseparate the two eects. Nevertheless since other household characteristics related to poverty arecontrolled or, it is likely that this nding does imply that risk aversion and income variability

(and lack o protection rom it) can increase the probability o a child being sent to work insteado school.

Jacoby and Skouas (1997) is one o the ew rigorous studies that explores this relationship.They use panel data rom rural India to examine the impact o seasonal changes in the incomes oagricultural households on childrens school attendance. They estimate a human capital investmentmodel with uncertainty under various scenarios. The expected theoretical outcomes are summarized

in Table 1 below.

table 1

a taxonomyoF Income eFFectSon School attendance

marKet Structure

IdIoSyncratIc ShocKSaggregate ShocKS(VIllage leVel)

antIcIPated unantIcIPated antIcIPated unantIcIPated

Complete intravillage markets

Complete intervillage markets None None None None

Intervillage credit only None None None Yes

Village autarky and:

Complete intravillage markets None None Yes Yes

Intravillage credit only None Yes Yes YesHousehold saving only1 Asymmetric Yes Yes Yes

Household autarky Yes Yes Yes Yes1 Also i borrowing is possible but at an increasing marginal cost.Source: Jacoby and Skouas (1997, 318).


SectIoN II



12/42

Table 1 illustrates the expected impact o idiosyncratic and aggregate village-level incomeshocks on school attendance under dierent nancial market scenarios. For instance, i householdsace complete nancial (credit and insurance) markets, they are able to reallocate resources acrossboth time and space at xed prices it implies a separation between a households human capital

investment decisions and its consumption decisions (Jacoby and Skouas 1997, 316). Incomevolatility would have no impact on school attendance. At the other extreme, i households arecompletely autarkic (sel-dependent), all kinds o income shocks would be expected to aect school

attendance since they would have no means o borrowing to tide over an income downturn.

They conclude rom the empirical estimations that parents oten do pull their children outo school during times o nancial crisis in the ace o incomplete nancial markets. There are

dierences in the impact o idiosyncratic and aggregate income shocks across villages but both havea signicant negative impact on schooling attendance or the ull sample. This is usually done sothat children can earn some wages to supplement the household income. This is particularly trueor unanticipated income shocks or small arm households that are more vulnerable to seasonal

income variability. Thus child labor is used as a consumption smoothing mechanism in the aceo lack o other ormal or nonormal means by which a household can get credit to tide over thenancial downturn. This aects their schooling and, in the long run, could have a substantiallynegative impact on their education attainment (though direct empirical evidence o the link between

attendance and attainment is not available).

Another rigorous empirical investigation o this issue is not or an Asian country but or Brazil.

Duryea (1998) use panel data rom Brazil and arrive at similar conclusions as Jacoby and Skouas door India. They nd that unanticipated income shocks in the orm o unemployment or the atherand the resulting loss o transitory income orce many households to withdraw rom school or notgo on to the next grade (an even stronger conclusion than the impact on school attendance since

this paper deals directly with progression through school).

Dehejia and Gatti (2002) empirically study the correlation between vulnerability to income-

related risk and schooling/child labor at a macro level using cross-country data. They use panel childlabor data or 172 countries and estimate a Tobit model. They conclude that the availability o credithas a strong signicant negative impact on child labor even ater controlling or income povertyand several other key independent variables. They also examine the eect o income volatility onchild labor and nd a strong positive association particularly in countries where access to credit is

poor, suggesting that poor, vulnerable households use child labor as a means to smooth income inthe ace o incomplete nancial markets. This has a negative impact on child schooling (and leisuretime). They nd that the coecient on secondary schooling is, in particular, strongly signicant,

emphasizing the negative relationship between child labor and schooling (or primary schoolingit is marginally signicant as perhaps children can combine work and school more easily at thislevel). This would, in the long run, have a negative impact on the income o the next generation(assuming positive returns to education).

Strauss and Thomas (1994) summarize a set o studies that examine the eect o incomevulnerability in the orm o income shocks on other human capital investments such as health.They cite Foster (1995) who nds that a major food in rural Bangladesh, which raised rice prices,

caused lower child weight growth among the vulnerable, landless households. They emphasize thatthere are very ew studies that have attempted to measure eects o resource shocks on human

November 2006




13/42

capital outcomes and how existing assets may condition those eects (Foster 1995, 1982). Clearlymore empirical research is warranted in this important area.

C. Unemplymen, Incme L, and Reiance Rerm

The above sections have considered vulnerability to income volatility in terms o how this aects

household decisions. As discussed earlier, many o these decisions urther enhance the vulnerabilityo these households to income variability and sometimes deepen the poverty trap.

Vulnerability to income risk can also arise or worsen due to job loss and resulting unemployment.

The income loss rom becoming unemployed can itsel be a crucial blow or vulnerable households(oten setting in motion some o the decisions like pulling children out o school, as examinedearlier). Labor market insecurity is increasing in transition and developing Asian economies. Economicreorm programs are orcing governments to shut down, restructure, or privatize many public sector

enterprises. In addition, layos are increasingly becoming more common in the private sector inthese countries as well, as increasing global competition and market orientation put pressure onrms to cut costs.

Another consequence o being vulnerable to income shocks and not having adequate mechanismsto deal with income loss due to unemployment has been widespread resistance to restructuringand privatization reorms. In 2004, 74 million workers were unemployed in Asian developing

countries, and another 426 million were underemployed (Felipe and Hasan 2006). The unemployedare a particularly vulnerable section o the population since most o these countries do not haveadequate social saety nets or those who lose their jobs. In the past, job security was guaranteedor workers in the public sector and or many in the private sector in many Asian countries like

India. As Felipe and Hasan (2006, 268) comment, provisions or job security in Indias labor codehave been a de acto substitute or a system o unemployment insurance insoar as organized sectorworkers are concerned. However with economic restructuring programs being put in place in manyAsian countries as globalization and competitive pressures mount, job security is no longer the

norm and downsizing is increasingly becoming common.

The reorm and economic restructuring process is quite oten signicantly slowed down and

sometimes even halted due to opposition by workers and trade unions whose jobs are threatened.In most developing Asian countries, the social saety nets provided or those losing their jobs areinadequate (see a more detailed discussion o this issue in Section III) and this heightens theirvulnerability, causing them to oppose the restructuring process. Workers sometimes even resort to

violence in their rustration, as when retrenched workers in Malawi burned down 10,000 hectareso plantations (Adam Smith Foundation 2006). The 2006protests in France against a proposed lawthat would make ring o youth easier illustrate that these are burning issues even in developedcountries. The widespread labor disputes in Peoples Republic o China (PRC) in the late 1990s

despite government eorts to provide increased saety nets refects the vulnerability o workers

who lose their jobs, even in the ormal sector, and the reality o resistance to reorm that manygovernments have to contend with (Nielsen, Nyland, Smyth, and Zhu 2005).

As the above discussion elucidates, vulnerability to income risk and shocks can have a debilitatingeect on households in developing Asian economies. Shocks can oten push households on thebrink o poverty into becoming poor i they do not have adequate ormal or inormal mechanisms


SectIoN II



14/42

to deal with them. Exposure to risk can also cause vulnerable households to make decisions thatreduce risk at the expense o productivity and enhanced uture incomein the orm o agriculturaldecisions or those regarding the health and education o their children. All o these decisions otenpush these households in a downward spiral o poverty and vulnerability, sometimes trapping them

in a vicious cycle that is dicult to break out o.

III. MItIGAtING VULNERAbILItY to INCoME VARIAbILItY/sHoCKs:tHE RoLE oF soCIAL PRotECtIoN

The discussion above provides a compelling argument on why governments should provide

adequate social protection mechanisms to mitigate and reduce vulnerability. In addition, this sectionwill make the argument that it is critical or Asian governments to have policy rameworks thatsupport useul inormal social protection mechanisms that are already prevalent as well as thoseprovided by the private or nongovernmental sectors.

Traditionally, the advocates o social protection schemes have ocused on their copingaspect, i.e., the act that they enable people to deal with the adverse eects o shocks. More

recently, international literature has emphasized providing various orms o social protection toenable individuals, particularly those who are oten exposed to shocks, to undertake risky activities.The World Bank Social Protection Sector Strategy paper (World Bank 2001) sets out a social riskmanagement ramework where social protection policies act not only as saety nets but also as

springboards to enable vulnerable households to break out o the povertyvulnerability trap byinvesting in building human capital and making prot-maximizing decisions. Needless to say thesaety net aspect continues to be important as well.

A. Inrmal scial Precin Mechanim

Poor and vulnerable households in developing countries, including those in Asia, are capable odealing with some income risk without ormal social protection provided by the government. Skouasand Quisumbing (2004)use ve household data sets rom developing countries (including Bangladeshand Russia) to understand the extent to which households are capable o insuring consumption

rom economic shocks to real income using inormal means. They conclude that households use aportolio o strategies, but that dierent types o households may have dierential ability to usethese strategies. Besley (1995) summarizes a variety o inormal risk coping and sharing mechanisms

and uses the term non-market institutions as an overarching term or the variety o mechanismsused in practice.

The inormal methods discussed by Skouas and Quisimbing (2004)and Besley (1995) includethe kinds o income and consumption smoothing methods such as the sale o productive assets, useo child labor, and other decisions to avert risk discussed in the earlier section on vulnerability. Inaddition, transers and remittances among amily, neighbors, and riends; and loans or the use osavings and household stores, are other inormal means used by households to deal with income

alls due to shocks. Paxson (1992) nds using data rom Thailand that household savings respond totransitory income shocks. Platteau (1991) explains that transers and loans among amily membersor neighbors are used when there are contingencies uncorrelated across the group (Ahmad etal. 1991, 141). Unni and Rani (2001)nd that social networks in Gujarat, India, help in times onancial crisis. In a survey they designed or a sample o households in the state o Gujarat, 90%

8 November 2006




15/42

of households reported the existence of such a social network to bail them out in times of nancialcrisis (Unni and Rani, 2001, 39). Agarwal (1991) illustrates how vulnerable households diversifyincome sources by, for example, migrating seasonally, using mixed cropping, and drawing on thelabor of women and children in lean seasons. Rosenzweig and Stark (1989) conclude using data

from rural India that households that have more variable farm prots have a greater probability ofhaving a member employed in nonfarm activity as a means of income diversication.

Besides these individual and ad-hoc informal arrangements, there are also more institutionalized

informal arrangements that vulnerable households use to mitigate income risk. Most of theseinvolve some kind of group risk sharing. Platteau (1991) discusses a more institutionalized informalrisk pooling method he labels co-operative labor-exchange pools where pool members commit to

helping each other out during any misfortune without expecting payment. What they expect, inreturn, is similar help in their time of need. Bardhan (1984) highlights the continued presenceof some tied labor contracts even in the modern age and argues that they exist because they aremutually benecial to both employers and employeesthey enable the laborers to mitigate their

vulnerability to low income during lean seasons when other employment may be hard to nd, andenable landlords to pay lower wages than they would otherwise have to in peak seasons.

Besley (1995) describes mechanisms like credit cooperatives, rotating savings and credit

associations (ROSCAS), and group lending. Credit cooperatives, he explains, take advantage oflocal enforcement and informationthey borrow from a bank and then on-lend to their members.ROSCAS are a variant of this informal lending mechanism found in many countries including Asia

(e.g., India; Republic of Korea; Taipei,China), where groups get together and pool their funds in apot of funds that goes to one member by lot or bidding every period. ROSCAS can assist memberswith risk sharing in the event of income or health shocks that occur during the rotation cycles(Calomiris and Rajaraman 1993). Group lending (which has become more widespread as inspired

by the success of the Grameen Bank in Bangladesh) involves loans being given to groups that arethen jointly liable to pay them back. As Besley suggests, group lending acts like a kind of forcedrisk-sharing arrangement since each members payoff depends on all the other members productiveuse of the loan.

Most studies on the informal sector and/or informal social protection mechanisms are descriptivein nature. It is imperative to empirically observe whether these methods are effective solutions forthe informal sector in terms of mitigating their vulnerability to income risk. However, there are

only a few studies that attempt to do this in a rigorous manner. Nevertheless, several studies dodocument the advantages that informal mechanisms have, in some cases, over more formal systems.They also discuss the weaknesses and limitations of these mechanisms.

Agarwal (1991) quotes a study that found empirical evidence that living in a joint family setupas prevalent in many developing economies can actually help improve agricultural productivity.Informal coping mechanisms help these households cope better with agricultural risk enabling

them to make more risky choices. All the risk-sharing institutions discussed above (cooperatives,ROSCAS, group lending etc.) share the feature that it is easier for individuals/households to monitoreach other when there is repeated contact between them and they know each other well, as isthe case in the villages/neighborhoods where these methods are seen used. Belsey (1995) calls

this the peer monitoring view. Agricultural insurance, for example, sometimes fails, whereas thesemore informal institutions are successful in reducing vulnerability to risk for households in villageswhere information problems abound and legal contracts are often difcult to enforce. Formal bank


SectIoN III

mItIgatINg vulNerabIlIty to INcome varIabIlIty/ShockS:

the role of SocIal ProtectIoN


16/42

loans also ail or the same reason. Enorcement is also oten easier when inormal mechanisms areused or similar reasons o better local knowledge. Hence the adverse selection and moral hazardproblems aced by ormal institutions in providing social protection are minimized when inormalinstitutions are used or mitigating income risk.

Many studies conclude that amily patronage systems and inormal systems o social protectionwork well when the shocks experienced by households are not o a catastrophic nature and/orare experienced or an extended period (see World Bank 2001, Agarwal 1991). These (inormal)

mechanisms seemingly work or most but not all o the population i the shock is idiosyncraticand not catastrophic, but they ail under a protracted and severe crisis (World Bank 2001, 7).Households that do not have social protection other than relying on their inormal networks are

usually still very vulnerable to shocks that aect an entire community or village such as aminesor foods since inormal risk sharing breaks down under these circumstances.

Townsend (1995) summarizes the results rom ew empirically rigorous analyses o risk sharing

in the inormal sector. He reports studies using data rom Cote DIvoire, India, and Thailand. Hisstudy nds that, in general, incomes o a village or region are much less covariate than would be

expected. The data reveal that there is, indeed, risk sharing among households in a village or regionusing inormal means. However, this is not complete risk sharing. Studies rom various countries

nd, using household data, that while there is some risk sharing, it is oten imperect and poorhouseholds are usually still vulnerable to risk. These include Townsend (1995) or India (discussedabove), Deaton (1992) or Cote DIvoire, Townsend (1995) or Thailand, and Jalan and Ravallion

(1999) or the PRC. Skouas and Quisimbing (2004) also use data rom ve countries (includingBangladesh and Russia in Asia) and conclude that households are only partially able to insure.

In addition, across the board these studies nd that richer households with assets/landholdings

were better able to deal with idiosyncratic risks whereas poorer households and those withoutany land were oten vulnerable even to idiosyncratic shocks. It is oten these households that areorced to deplete their meager assets when aced with an emergency, and this pushes them into

poverty. They are oten the households that use the income and consumption smoothing strategiesdiscussed in Section II. However, as argued earlier, these can be costly ways o protecting oneselagainst risk since households oten have to sacrice current or uture income when they make thesedecisions. In addition, some o these choices, such as using household stores o ood grains, canbe detrimental to women and girls in the household as gender biases in ood allocation in times

o shortage are well documented in the literature (Agarwal 1991). As Morduch (1995) observes,even where inormal insurance is well developed, public actions that displace inormal mechanismscan yield net benets.

The inormal risk sharing mechanisms detailed above such as ROSCAS and group lendingschemes do not have the disadvantages o individual coping methods that households use to smoothincome or consumption. In addition, they have the advantage o reducing the adverse selection and

moral hazard issues aced by ormal institutions. Nevertheless, in practice, they have seen mixedsuccess. Cooperatives, or example, become nancially unviable in many countries especially whentheir design does not take into account mass deaults due to covariant risk actors (Besley 1995).Also, as economies develop, the presence o these traditional inormal institutions declines over

time especially as better monitoring technologies emerge, indicating that they are not the preerredoption in modern economies.

As the above discussion reveals, research indicates that not being covered by ormal social

protection programs does not necessarily imply that an individual is substantially more vulnerable

10 November 2006




17/42

to risk. Households have various risk coping mechanisms that work well to buer them againstshocks and make them less vulnerable. However, many o these mechanisms do not work wellwhen a catastrophe strikes or negative shocks such as amine occur or a long period. More ormalmechanisms and state intervention are required to ensure that transient poverty does not rise and these

households do not suer. Also as some o the articles mentioned above argue, inormal risk sharingand coping arrangements can have costs that are not desirable such as decisions that compromiseproductivity and investment in human capital. As some o the studies discussed illustrate, inormal

risk sharing is not perect and may still expose households, particularly those that are very poor,to substantial risk. The poorest households are still oten vulnerable even to idiosyncratic shocks.Sometimes inormal mechanisms do not work well because o problems o enorcing repaymentamong households, or because a household is unable to reciprocate a transer. Morduch (1995) states

these (inormal) mechanisms can be highly eective in the right circumstances, but most recentstudies show that inormal insurance arrangements are oten weak. Thus public social protectionprograms are required and need to be well targeted using indicators o the ability o households tosel-insure against shocks where inormal mechanisms may not buer certain households.

b. Frmal scial Precin Mechanim

What ormal social protection mechanisms exist in developing Asian countries? A wide range oprograms are in place in dierent countries. These are usually grouped under ve categories: socialinsurance programs (such as unemployment and health insurance, pensions etc.); social assistance

and welare programs; labor market programs; micro and area-based schemes usually aimed at theinormal sector; and child protection programs. There are several books and articles that discussall these mechanisms in detail (see Ortiz 2001 or a good overview or Asia). This paper ocuses on

those mechanisms (similar to the discussion o inormal mechanisms) that are specically designedto mitigate income risk and, in particular, empower vulnerable individuals and households to takeintertemporal decisions without compromising productivity and uture income to enable them toescape the povertyvulnerability trap.

(i) Social assistance mechanisms

Social assistance mechanisms play an important role in developing Asian economies. Theseare critical to the most vulnerable individuals and households and include cash transers, short-term cash, and in-kind transers to cope with negative shocks and longer-term poverty reduction

measures such as in-kind subsidies (see Ortiz 2001 or more details). Ortiz cites how several socialassistance mechanisms have been successul in Asian and Pacic countries and can support growthwith equity and bring about success in poverty reduction or vulnerable groups (Ortiz 2001, 307).This contradicts widespread claims that they create households that become permanently dependent

on the state and are too expensive to implement. However, many in-kind subsidy schemes (e.g.,ertilizer subsidies) have not been eective and research indicates that they have beneted thebetter-o, rather than the very poor (World Bank 2001). Most countries are attempting to reducethese subsidies.

Nevertheless, there are certain subsidies such as scholarships or school ees and ree healthcare or the poor that are essential to continue. Another useul subsidy being increasingly used in

developing Asia is ree midday meals or children in public schools. These subsidies help vulnerablehouseholds avoid having to resort to child labor as a means o income support and withdrawingtheir children rom school when they experience an income shock. The schooling and child labor

erd WorkINg PaPer SerIeS No. 88 11

SectIoN III




18/42

literature is extensive and there are a host o policy measures that governments take to curb childlabor and encourage vulnerable households to send their children to school (or good reviewssee OECD 2003, Grootaert and Patrinos 1999). The measures ocused on here are those related tomitigating income-related risk or vulnerable households to enable them to send their children to

school rather than to work.

(ii) Economic incentives

Economic incentives are provided in various countries to mitigate income poverty and enablepoor households to send their children to school instead o work. These include scholarships, ree

school meals, ree uniorms, books, transportation etc. While these are not necessarily targeted athelping households deal with vulnerability to risk, they nevertheless help in this regard as well,since households do not have the cash to spend on their childrens schooling particularly duringlean periods or when they have suered an income shock (OECD 2003, Sipahimalani 1999, Dreze

and Kingdon 1999). Midday meals in India have met with great success in some states such asTamil Nadu in terms o their impact on increasing enrolment in school (Sipahimalani 1999). The

school nutrition program in Pakistan and the Food or Schooling program in Bangladesh have alsobeen successul in increasing school enrolment and reducing dropout (OECD 2003). Some programsin Latin America have linked cash transers to households to school enrolment or attendance, andhave been very successul in increasing human capital investment by increasing both enrolmentand attendance by mitigating household vulnerability to income risk. These include the Bolsa Scola

program in Brazil and the Progresa program in Mexico (OECD 2003). Sometimes school ees can bewaived during an economic crisis to prevent people rom pulling out their children rom school andalso rom alling into a transient poverty trap. This was done with success in Indonesia during

the nancial crisis (Sayed and Filmer 1998). Expanding the reach o subsidized health insurancecontinues to be an important priority in developing Asian countries, and vulnerable households canbe orced not to seek medical help i they are not insured and have suered an income loss.

(iii) Public works

Public works are another ormal social protection mechanism useul in helping households cope

with temporary income shortalls. These are programs where the unemployed and underemployed aregiven temporary jobs usually in building and repairing inrastructure. They have been used airlyextensively in India and Korea, or example, and have the advantage o being sel-targeting. They

could, once again, be useul in helping households tide over lean income periods instead o havingto resort to selling assets or pulling their children out o school to send to work.

(iv) Social funds

Social unds can be seen as a recent variant o the public works approach to providing

employment or the chronic poor or communities hit by economic or other disasters (Ortiz 2001).They are agencies in government that are set up to assist local communities by nancing small-scaleprojects, usually in inrastructure, based on local demand. They are relatively new in Asia, startingas recently as 1999. While it is too early to evaluate their impact, they are useul in post economic

crisis situations such as the social unds set up by the World Bank in Indonesia and Thailand inorder to help the communities cope with large-scale economic crises where inormal mechanisms

1 November 2006




19/42

usually ail. Their key advantage is the autonomy they enjoy compared to usual government programs,and the act that they are designed based on the demand o local communities, and can be welltargeted. Evaluations o social unds in other countries have revealed that the poverty impact o

improving schools, clinics, water supplies, access roads etc. are well established (Ortiz 2001, 444).

However, their sustainability is uncertain and the act that they divert resources rom strengtheningother mainstream government interventions remains controversial.

(v) Agriculture insurance

Agricultural insurance (including micro insurance) is an important tool used by policymakersto provide social protection to mitigate income risk, particularly or small armers. It enablesvulnerable armers to choose the optimal mix o inputs and make productivity-maximizing decisions.In practice, however, it is plagued by many implementation obstacles such as problems o adverse

selection and moral hazard mentioned earlier (pitalls that are, in act, avoided by inormal risksharing methods such as ROSCAS). In addition, risks that beall the entire group that is insured(covariant risks) oten make these schemes nancially unviable. As Abada (2001) notes, schemes

implemented so ar in the developing countries have not been particularly successul and somehave had to be discontinued or substantially modied. Abada (2001) quotes a study that oundhigh loss ratios in seven developing countries (including India, Japan, and Philippines). Thusagricultural insurance almost always has to be publicly subsidized and private agricultural insurance

is not usually viable.

Vulnerability to income shocks resulting rom unemployment in the ormal sector is anotherissue that policymakers in Asia increasingly have to deal with. It has resulted in privatization and

economic restructuring programs usually being met with widespread resistance by labor unions andoten rom the public at large, as discussed in Section II. Most o those who lose their jobs havebeen accustomed to job security in the past. In addition, most Asian developing countries do nothave well-developed social protection systems or the unemployed. As losing ones job becomes an

inevitable reality o the 21st century in these economies, governments need to ocus on developingsocial protection mechanisms that protect the unemployed and make job loss more palatable, whilestill keeping these mechanisms aordable. These mechanisms are reerred to as labor adjustment

programsdened by the Adam Smith Institute (2006, 1) as government led programs to helpreduce the negative consequences o enterprise reorms.

In developed economies, unemployment insurance and assistance are controversial policies.

Most studies have concluded that unemployment benets increase unemployment since they reducethe incentive to search or a new job. Calmors and Holmlund (2002) summarize the evidence romvarious developed countries to conclude that there is considerable support or the hypothesis thatlower benet levels and shorter entitlement periods associated with unemployment insurance reduce

unemployment.

However, developing countries oer a dierent scenario since they are usually not near ullemployment level and as discussed earlier, oten have a majority o workers working in the inormal

sector. I viable unemployment benets are oered to workers in such a situation, it is expectedthat this would be benecial as it would enable workers, particularly in the inormal sector, totake risks and look or better paying jobs. This would help create a more fexible labor market. In

addition rms would be more willing to hire workers during boom periods i labor laws allow them


SectIoN III




20/42

to lay o workers more easily. The opposition to restructuring programs and reorming rigid laborlaws would also decrease i better social saety nets are in place.

Hu (1994) nds in his study on the PRC that reorming the social saety net system has been

crucial to increasing labor market fexibility there. He claims that Chinas experience suggests

that, without removing rigidity in the labor market, the social and economic costs o necessaryenterprise restructuring in terms o employment would be too high to be politically sustainable(Hu 1994, 4). He goes on to explain that the traditional system o social protection in the PRC or

the large public sector was rm-based, tying the employee to a particular rm and discouraginglabor mobility. Loss-making state-owned enterprises had substantially higher labor costs than theirprivate counterparts due to overstang and unsustainable social benets; moreover, competitive

pressures were orcing them to restructure and lay o redundant sta. In 1986 a big step towardlabor market fexibility was introduced where a policy decision was taken to hire all new workersas contractual workers. In addition, SOE managers could now legally lay o workers. However,laying o is not a common practice largely because this has aced widespread opposition due to

the tradition o social saety nets being associated with the rm an employee works or. Thus,increasing labor market fexibility, a necessary condition to transorm the state-owned sector, hascalled or overhauling Chinas social security system (Hu 1994, 19). Unemployment insurance atthe local government level has been introduced as the cornerstone o this new system. Further

evaluations will be needed to study the impact o this scheme on resistance to reorm.

Fretwell (2004, page???) reviews various social support programs or workers losing theirjobs due to privatization in Brazil, transition economies in Eastern Europe, and Western Europe.He concludes that while a broad spectrum o mechanisms is used in dierent countries, the key tosuccessul programs is direct dialogue between key stakeholders (including workers) both beoreand during the restructuring program. A working document by The Adam Smith Institute (2006)

reiterates the importance o dialogue, particularly with workers and unions early on in the processo restructuring or privatization in order to build support or reorm.

What unemployment beneits can be oered in lieu o job security in developing Asianeconomies as job losses become a reality or many workers? A range o options are available topolicymakers in Asian developing countries to provide better social saety nets or unemployedworkers. The choices they make will have to balance the benets to workers and their amilies aswell as the benets o reduced resistance to privatization and more fexible labor laws with the

costs o providing social saety nets.

In most developed countries unemployment benets are provided on a regular basis in the

orm o unemployment insurance or unemployment assistance. However most developing countriesincluding those in Asia do not provide unemployment insurance/assistance on a regular basis, whichare expensive options or most o these economies during periods o restructuring/privatization.These are long-term institutional developments that need to be considered on a country-by-country

basis. The unemployment benets usually considered by developing countries are specic to thecontext o enterprise restructuring and privatization that leads to large numbers o workers beingretrenched.

These include income support programs and nonincome-related support programs or labor services.Most labor programs consist o some income component in the orm o a severance payment and/orpension payments as well as other labor services including retraining and redeployment/counseling.

1 November 2006




21/42

Some programs also involve employees owning shares in the new rm, i any. Severance can includewage arrears, regular severance, and/or special severance to encourage workers to quit voluntarily. Asthe Adam Smith Institute (2006, 18) states, in developing a severance scheme the main challengelies in devising severance payments that are both attractive or workers and nancially aordable

and sustainable. The details o alternative types o severance schemes are not relevant here. Thekey issues involved are appropriate targeting o packages to prevent adverse selection and puttingin place nancing arrangements to ensure speedy implementation. Felipe and Hasan (2006) nd,

in an extensive review o labor market policies in Asia, that severance pay and the cost o ringworkers is generally higher in Asian countries compared to Arica and industrialized countries(and about the same as in Latin America). Rama (1999) explains that the voluntary approach toretrenchment can be quite expensive and can also lead to the most skilled workers with best outside

employment options leaving (adverse selection). Targeting by oering severance packages only toworkers identied as redundant through studies and analysis might be a better option.

Redeployment programs or labor services are a key element o support to the unemployed that

Asian developing country governments use to varying extents. They are meant to assist workers wholose their jobs to be reemployed gainully in a short period o time. Fretwell (1999) and Betcherman,Olivas, and Dar (2004) nd that these programs have a signicant positive impact i careullytargeted and well run. The caveat that the programs should be well run is an important one. The

programs provided need to be demand-driven to prove useul to the workers. Services provided asredeployment programs range rom training/retraining, counseling, job search assistance, placementservices, and assistance or sel-employment or small entrepreneurs.

IV. PoLICY IMPLICAtIoNs AND CoNCLUsIoNs

In the past, policymakers viewed social protection mechanisms as tools to help vulnerablehouseholds cope with and mitigate risk. This paper illustrates that the conventional thinking on therole o social protection mechanisms has evolved so that they are now also considered important or

reducing risk to enable vulnerable individuals to invest in high-return activities. This is particularlyimportant in a increasingly competitive and market oriented global environment where new butoten riskier technologies and opportunities are available. Enabling vulnerable households to takeadvantages o these would not only improve their welare but also stimulate economic growth

through more productive use o assets and inputs and higher human capital accumulation.

In the medium to long term, policymakers need to reduce chronic poverty which is, notsurprisingly, one o the strongest correlates o vulnerability to income risk. Nevertheless, policies

aimed at empowering vulnerable individuals and households to mitigate risk and deal with incomevariability in the short term also helps lower transitory poverty, which is oten a major componento overall poverty. The previous section detailed some pros and cons o various inormal and ormal

social protection methods. This section draws lessons or policy based on that discussion.


SectIoN Iv

PolIcy ImPlIcatIoNSaNd coNcluSIoNS


22/42

(i) thinking uide he x: creaive hinking and cr ecralapprache deigning cial precin prgram

Traditionally, social protection mechanisms have entailed either social insurance or assistance

o some kind. However, as discussed earlier, it is oten the case in developing countries that some othese methods are either nancially unviable and/or ineective in protecting vulnerable individualsrom income risk due to various reasons. For example, unsubsidized weather insurance does nothelp armers undertake riskier activities with higher returns since there are cheaper alternatives to

smoothing consumption. In act even subsidized agricultural insurance may be beset with problemssuch as adverse selection, moral hazard, and covariant risk. In such situations a conventionalsocial protection policy o agricultural insurance may not be the answer to mitigating and reducingagricultural risk.

On the other hand, creative solutions such as helping provide consumption credit (so thatarmers do not get caught in debt traps with high-interest charging moneylenders as their only

option or borrowing) may be eective in enabling armers to invest in prot-maximizing behavior.Consumption credit would help smooth consumption, allowing the armer to take riskier decisions

to maximize prots without worrying about alls in consumption or the household. Eswaran andKotwal (1989) demonstrate that in agrarian economies, consumption credit can, in act, sometimes

assume the role o insurance. They explain that consumption credit markets allow risk pooling, thusenabling individuals to smooth consumption without resorting to nonoptimal choices. They nd thatarmers who had access to consumption credit were more likely to adopt high-yielding variety seedsduring the Green Revolution in South Asia, or example. This is a good option to insurance schemes

since it enables risk pooling across time, and usually the random actors generating uncertaintythat may be covariant across armers are not correlated across time. In addition, credit contractseconomize on inormation and are usually easier to implement than agricultural insurance contracts

that have proven to be very inormation-intensive and an administrative burden. Thus, Eswaran andKotwal (1989, 50) conclude that although credit is used exclusively or consumption stabilization,

it has a positive impact on production behaviour. Biswanger and Sillers (1983) nd similar resultsin their study where they engaged armers in several developing countries in small gambles as

an experiment. They ound that the dierences among armers in their readiness to adopt newtechnologies was less due to dierences in risk preerences and more due to the availability o, inthat case, production credit.

Similarly, in order to help income smoothing, or example, increasing opportunities or armhousehold members to get jobs with assured salaries may increase the eciency and stability oagricultural output (Rosenzweig and Wolpin 1993). Having a household member who has a stableincome could enable a armer to make prot-maximizing agricultural decisions without worrying as

much about income variability due to the high risk nature o any o these choices.

In their subsequent study on the relationship between incomplete nancial markets and

childrens schooling, Eswaran and Kotwal conclude that complete credit markets can have a benecialimpact on human capital investment as well. This nding is supported in other empirical studiesdiscussed in Section III. Thus, while providing access to consumption credit is not usually denedas a social protection measure, it has the benecial impact o allowing vulnerable households to

take decisions without compromising productivity and current/uture income, and decisions onhuman capital investment. This nding is echoed by Grootaert and Patrinos (1999 , 161) when theystate programs are needed to provide credit to poor households without collateral so that theyno longer need to rely on their children working as insurance against alling income. Dehejia and

1 November 2006




23/42

Gatti (2002) is the only empirical study to investigate the relationship between nancial marketsand child labor decisions in a rigorous manner at a macro cross-country level. As detailed earlier,they nd that income volatility has a signicant impact on child labor. In addition, they split theirsample o countries into a high credit and low credit group (by access to credit) and conclude

that countries with better access to credit income shocks do not signicantly aect child labordecisions. More importantly, extending the ability o households to insure against income shocks hasa distinct advantage over legal restrictions and direct bans as it can decrease child labor without

lowering amily welare (Dehejia and Gatti 2002, 12). Decreasing child labor would, o course, helpin increasing human capital investment, as more children are sent to school and have increasedleisure time as well.

In the example discussed here, the link between nancial markets and human development hasbeen emphasized. Policymakers may provide scholarships or other economic incentives to encouragehouseholds to send their children to school instead o work (in addition to the regular programs obuilding more schools and hiring teachers etcetera). However, these policies are unlikely to increase

human capital investment or households that are very vulnerable to income risk. Cross sectorallinkages in designing programs could involve, or example, providing these economic incentives butalso improving household access to micro credit or consumption credit through ormal means likecooperative banks, or encouraging inormal means like ROSCAS.

The more general point is that policymakers should break the sector silos when designingprograms or social protection, as such coordination across sectors and ministries could have a

very benecial impact on risk reduction and mitigation. Thinking creatively out o the box couldhelp policymakers ocus on designing mechanisms that are not conventionally considered socialprotection and, sometimes, eliminate schemes such as some agricultural insurance schemes thatare ineective and nancially unsustainable.

(ii) scial aiance: he imprance argeingand an inegraed apprach

While it is imperative that policymakers take into account cross sectoral linkages andunconventional social protection mechanisms, some established social assistance schemes to protect

the most vulnerable households rom income volatility continue to be critically important. It is crucialin most developing Asian economies that child protection schemes be prioritized to enable sustainedhuman capital investment by the poor, particularly during lean income periods. Careul analysis andplanning is needed to ensure that social assistance and child protection programs are well targeted

and prioritized in a cost-eective manner. Administration costs also need to be curtailed to makethe programs nancially sustainable. Policymakers could seek out publicprivatenongovernmentorganization (NGO) partnerships to make delivery and implementation more cost-eective and better

targeted; since area-based NGOs oten have better access to local inormation to enable targeting.

As Van Ginneken (2003) explains, accurate income gures are dicult to obtain in developingcountries. Thereore targeting using other indicators such as assets and sel-selection has beenmore successul. This paper also emphasizes that targeting methodologies have to be simple so that

they are transparent and easy or local government ocials to implement. Geographical targetingis useul during regional crises (Ortiz 2001).

Social protection programs in general and social assistance schemes in particular are usually

not very well integrated within the development ramework o Asian developing economies. This


SectIoN Iv



24/42

makes targeting and evaluation o the eectiveness o schemes very dicult. In addition, havinga prolieration o disjointed schemes makes it complicated or policymakers to make implementingagencies accountable or outcomes. As Justino (2003, 7) comments, the ocus o social securitypolicies in developing countries should thus be on the reduction and mitigation o structural

orms o vulnerability and on the implementation o ways o coping with all types o risk and beintegrated within the overall development strategy o the country rather than implemented asindividual programmes.

(iii) Frmal and inrmal mechanim: defning relaive rleand uilding linkage

Section III elaborated upon various traditional or inormal social protection mechanisms usedby households in developing Asian countries. Some methods o mitigating risk used by householdssuch as depleting assets to smooth consumption or making nonoptimal production decisions to

smooth income are clearly detrimental to households and the economy at large. Nevertheless as seenin Section III, there are other inormal risk-sharing methods such as transers and gits, rotating

savings, and credit schemes or group insurance schemes that are very eective at protectinghouseholds without sacricing productivity and prots particularly in the case o idiosyncraticshocks. In some cases they work better than ormal schemes because o the advantages o localinormation and peer monitoring. Nevertheless, as Ahmad et al. (1991) highlight, and the WorldBank (2001) Social Protection Sector Strategy reiterates, inormal social protection mechanisms

usually break down in the ace o catastrophes or shocks o a protracted nature. Van Ginneken(2003) observes many unemployment and poverty rates soared in many East Asian countries duringthe 1997 nancial crisis partly because households relied largely on inormal social protection

mechanisms. Asian governments and multilateral institutions have since then reconsidered andbegun to ocus on the role o ormal social protection arrangements particularly to mitigate incomerisks when catastrophes strike. Dhanani and Islam (2002) discuss the case o Indonesia during andater the 1997 nancial crisis and conclude that government social protection mechanisms o rice

subsidies and education scholarships were key to sotening the impact o the crisis even thoughthey were scaled to a national level only a year ater the crisis hit. They conclude that ormalsocial protection mechanisms need to be in place beore a crisis hits (even those measures that areintended as short-term ones to enable risk coping during a crisis need to have enabling policies

and legislations in place beorehand) to enhance their impact. Public work programs continue tobe an important but temporary measure governments can use to help tide over dicult economicperiods or households in the inormal sector.

In some situations, however, substituting inormal mechanisms might be socially optimal ithese mechanisms are harmul to the households. Also, certain regions in a country may not havewell-developed inormal mechanisms. The role o public saety nets is more critical here. In addition

to designing and implementing ormal social protection schemes, policymakers should also create

a regulatory and institutional ramework to encourage appropriate private coping mechanisms anddiscourage others that are harmul to productivity or have other negative consequences. Morduch(1999) emphasizes this point, explaining that policymakers can help improve regulation and strengthen

institutions to help households in the inormal sector get access to better credit, insurance, andsavings options. Providing access to good savings instruments is essential as is creating useulregulations or an enabling environment or banks and NGOs to help households save. Direct

18 November 2006




25/42

insurance programs by governments have not been successul as cited earlier. Governments need tourther exploit opportunities to acilitate private and NGO-provided insurance to reduce the moralhazard problem, as NGOs oten have better local inormation. Governments can provide appropriatesubsidies/guarantees, enabling legislation, and reinsurance support to create a supportive ramework.

Another interesting area or public policy to support inormal mechanisms is by improving legalrights o certain groups like women to own assets such as land to improve their access to credit(World Bank 2001). The World Bank summarizes this public policy role succinctly when it states

that the development community needs to nd ways to support, or crowd in pro-developmentinormal mechanisms to make them more eective and inclusive and discourage, or crowd outthe detrimental ones (World Bank 2001, 35).

Policymakers can also attempt to link some ormal schemes to inormal ones to enhancethe joint impact o both kinds o programs on vulnerability. Social unds, which are ormal socialprotection mechanisms set in place by governments, or example, have been successul when theyare community-driven since they are created to nance small-scale projects on the demand o

communities. In act, as Abada (2001) cites, empirical evaluations have revealed that social unds arenot well targeted in practice and are not very eective or the very poor and vulnerable. Involvinglocal NGOs in partnership with the local governments is a good way to ensure more communityinput into their design and implementation to improve their perormance. Newer social unds such

as one in Zimbabwe (the Community Action Project) are attempting to combine government andinormal insurance mechanisms. The local government head o the village keeps a plot o land asidewhose produce is meant or vulnerable households to cope with shocks. The community provides

the agricultural labor required or production (World Bank 2001, 37). Policymakers should undertakemore experimentation and pilot projects and evaluate these projects to nd ways or ormal andinormal social protection schemes to work together to maximize their infuence.

(iv) scial precin r he unemplyed: maximizing impac in ac-eecive manner

The discussion in Section IIIB reveals that unemployment is a growing concern or ormaland inormal sector workers as most developing Asian economies are now pursuing market-orientedreorms. The labor adjustment programs that governments usually put in place to deal with these mass

layos have been described in that section. It is vital that policymakers structure these programs sothat they maximize impact in terms o benets or the workers (and the related impact o reducedresistance to market-oriented reorms) while being cost-eective. Very high severance packages, asmentioned earlier, are common in Asian economies and these have the disadvantage o sometimes

being nancially uneasible and oten causing the best workers to leave (adverse selection). Toovercome these disadvantages the ormula to calculate severance pay should be careully thought ousing one o several approaches: setting minimum and maximum levels; reducing payments to workers

near retirement; or a loss-based ormula that takes into account actors such as seniority, gender,

and education to predict the welare loss o each worker and compensate the worker accordingly(Adam Smith Institute 2006). Asian governments need to study these alternatives and choose theappropriate method or each case.

Section IIIB also mentions redeployment programs that usually orm part o the labor adjustmentprograms. These include labor services such as training, counseling, and job search assistance, amongothers. These can be cost-eective while at the same time very benecial to unemployed workers.


SectIoN Iv



26/42

However, the kinds o services oered and method o implementation are crucial to their success.International experience has shown that demand-driven and well-targeted services are more successulin improving the re-employment opportunities or laid-o workers (Adam Smith Institute 2006).

Traditionally the government has tried to undertake all the training through its own institutions in

most developing countries in Asia, sometimes diluting the quality and useulness o the training.Implementation should be done by well-qualied agencies, whether private, nongovernmental, or

public. While they are usually not very expensive, unding needs to be in place to ensure that thereare no delays in implementation. One option would be to use unds rom privatization (Fretwell 2004).The PRC experience (Betcherman et al. 2004) illustrated that in longer-duration training programs,workers unding part o the training costs and programs other than those oered by the ocial

labor bureaus were associated with higher gains in employment. Fretwell (2004) also emphasizesthe importance o ongoing evaluations o these programs and cautions against generalizing acrosscountries. There are very ew evaluations in developing countries, which is an essential requirementin the uture to enable policymakers to make more inormed choices. Betcherman, Olivas, and

Dar (2004) summarize the available evaluations o retraining workers ater mass layos in seven

developed countries, one transition economy, and one Asian country, the PRC. Results are mixed,with the PRC study showing uncertain impact on employment. In the PRC, the evaluation revealed

that participation in retraining led to increases in employment (in one o the cities), while, in theother city, the employment eect was insignicant (Betcherman, Olivas, and Dar 2004). Clearlymany more rigorous evaluations o these experiences are required or developing countries.

Most developing Asian economies do not have ongoing active labor market programs (ALMPs) othe kind seen in industrialized countries. As these countries increasingly grapple with unemploymentissues, it is important to ask the question, should ALMPs be introduced in these countries? Theevaluations o ALMPs in developed economies show mixed results. Betcherman et al. (2004) review

over 150 evaluations o ALMPs and conclude that a wide range o results can be ound with someprograms demonstrating positive labor market eects and others showing either no impact or even

negative eects. Nevertheless, the small number o studies and ALMPs available in developing countriesindicate that the impacts o ALMPs here may be dierent rom that in developed economies. Whileweaker implementation capacity sometimes lowers the quality and impact o services provided, thepotential to train large numbers o workers who need the training remains. More research needsto be done, however, beore governments embark on a large scale to undertake ongoing ALMPs in

developing Asian economies since these can oten involve substantial nances.

An issue that is o critical importance in developing Asian economies is the absence o socialprotection in the orm o any kind o unemployment benet or workers in the large and growing

inormal sectors. Labor market reorms will have to take this sector into account to provide at leasta basic level o protection to these workers. Some developing countries have begun to take stepsin this direction. In Chile, or example, unemployment benets or the inormal sector nanced

mainly rom value-added tax revenues have been introduced with moderate success (World Bank2004). In India, the parliament is likely to discuss a bill to introduce social security or theunorganized sector. The bill proposes to nance this scheme using a combination o nominalworker contributions, employer contributions (where possible), and central and state government

contributions. Policymakers in developing Asian economies will need to ocus on providing someunemployment benets to the inormal sector workers who are growing in number.

0 November 2006




27/42

This paper addresses the issue o providing adequate social protection against income volatilityor those particularly vulnerable to income shocks and risk. It draws on available literature to ocuson the negative impact o vulnerability. We conclude that vulnerable households that do not haveaccess to adequate protection rom income risk are likely to get caught in a vicious cycle o poverty

and vulnerability. In addition, they oten hesitate to invest in high-return activities that have higherrisk such as adoption o new technology in agriculture. Sometimes, as in the case o many armersin India who have committed suicide, they suer terrible nancial consequences i they do adopt

the higher-risk technologies but have a bad harvest. In the urban ormal sector, the reorm processmay be slowed down as workers without adequate social protection oppose restructuring. This papersummarizes the various inormal and ormal social protection methods and programs in developingAsian countries. An analysis o the pros and cons o these schemes and their operation leads us

to recommend that policymakers need to intensiy eorts to provide adequate social protectionto households in the inormal sector and to those who might get unemployed in the urban ormalsector as well. It is vital that policymakers in developing Asia think creatively and across sectorbarriers about how to achieve a higher level o good quality social protection, and empower people

to take advantage o the opportunities oered by new technologies and global competition. A

more integrated approach to planning and implementation along with better targeting is criticalto the success o schemes. Furthermore, policymakers should provide an enabling environment or

inormal mechanisms that are benecial to fourish. At the same time policy should discouragethose methods that are harmul or that sacrice prots and productivity, while providing appropriateormal alternatives. More empirical research and evaluation is certainly warranted in this area tounderstand linkages across sectors and evaluate the impact o ormal and inormal programs.

sELECtED REFERENCEs

Abada, J. 2001. Micro and Area-Based Schemes. In I. Ortiz, ed., Social Protection in Asia and the Pacifc.

Asian Development Bank, Manila.

Adam Smith International. 2006. Labor Issues in Enterprise Restructuring: An Operational Brie or ADB.

Adams, R. H., and H. Alderman. 1992. Sources o Income Inequality in Rural Pakistan: A Decomposition

Analysis. Policy Research Working Paper No. 836, The World Bank, Washington, DC.

Agarwal, B. 1991. Social Security and the Family: Coping with Seasonality and Calamity in Rural India. In

E. Ahmad, J. Dreze, J. Hills, and A. Sen, eds., Social Security in Developing Countries. Oxord: Clarendon

Press.

Ahmad, E., J. Dreze, J. Hills, and A. Sen, eds. 1991. Social Security in Developing Countries. Oxord: Clarendon

Press.

Alderman, H., and C. Paxson. 1992. Do the Poor Insure? A Synthesis o the Literature on Risk and Consumption

in Developing Countries. Agricultural Policies Working Paper Series, No. 1008, The World Bank, Washington,

DC.

Asian Development Bank. 2001. Social Protection Strategy. Manila.

Atkinson, A. B., and J. Micklewright. 1991. Unemployment Compensation and Labor Market Transitions: ACritical Review.Journal o Economic Literature 29(4):1679727.

Baland, J-M., and J. A. Robinson. 2000. Is Child Labor Inecient?Journal o Political Economy108:663

79.

Bardhan, P. K. 1984. The Political Economy o Development in India. Delhi: Oxord University Press.


Selected refereNceS


28/42

Besley, T. 1995. Nonmarket Institutions or Credit and Risk Sharing in Low-Income Countries. Journal o

Economic Perspectives 9(3):11527.

Betcherman, G., K. Olivas, and A. Dar. 2004. Impacts o Active Labor Market Programs: New Evidence rom

Evaluations with Particular Attention to Developing and Transition Countries. World Bank Social Protection

Paper No. 0402, Washington, DC.

Binswanger, H., and M. Rosenzweig. 1993. Wealth, Weather Risk and the Composition and Protability o

Agricultural Investments. Economic Journal 103:5678.

Binswanger, H. P., and D. A. Sillers. 1983. Risk Aversion and Credit Constraints in Farmers Decision Making:

A Reinterpretation.Journal o Development Studies 20:521.

Calmors, L., and B. Holmlund. 2000. Unemployment and Economic Growth: A Partial Survey. Swedish

Economic Policy Review7:10753.

Calomiris, C., and I. Rajaraman. 1998. The Role o Roscas: Lumpy Durables or Event Insurance? Journal

o Development Economics 20716.

Deaton, A. 1992. Savings and Income Smoothing in Cote dIvoire. Journal o Arican Economies 1(1):1

24.

Dehejia, R. H., and R. Gatti. 2002. Child Labor: The Role o Income Variability and Access to Credit in a Cross

Section o Countries. Policy Research Working Paper No. 2767, The World Bank, Washington, DC.Dhanani, S., and I. Islam. 2002. Poverty, Vulnerability and Social Protection in a Period o Crisis: The Case

o Indonesia. World Development30(7):121131.

Dreze, J., and G. G. Kingdon. 1999. School Participation in Rural India. The Development Economics Discussion

Paper Series No. 18, London School o Economics STICERD.

Duryea, S. 1998. Childrens Advancement through School in Brazil: The Role o Transitory Shocks to Household

Income. Inter-American Development Bank Working Document No. 376, Brazil.

Eswaran, M., and A. Kotwal. 1989. Credit as Insurance in Agrarian Economies. Journal o Development

Economics 31: 3753.

Felipe, J., and R. Hasan. 2006a. The Challenge o Job Creation in Asia. ERD Policy Brie No. 44, Economics

and Research Department, Asian Development Bank, Manila.

. 2006b. Labor Markets in Asia: Issues and Perspectives. London: Palgrave Macmillan or the Asian

Development Bank.

Foster, A. 1994. Prices, Credit Constraints and Child Growth in Low Income Rural Areas. In J. R. Behrman

and T. N. Srinivasan, eds., Handbook o Development Economics, Vol. 3. Amsterdam: North Holland.

Fretwell, D. H. 2004. Mitigating the Social Impact o Privatization and Enterprise Restructuring. Social

Protection Discussion Paper Series No.0405, The World Bank, Washington, DC.

Van Ginneken, W. 1998. Social Security or All Indians. Delhi: Oxord University Press.

. 2003. Extending Social Security: Policies or Developing Countries. International Labor Review

143(3):119.

Grootaert, C., and H. A. Patrinos, eds

Documents

Income Volatility and Social Protection in Developing Asia