Identifying the vulnerable: new evidence from Peru

A joint publication by the Pol icy Research and Poverty and Social Policy Departments of the World Bank No. 6 March 1997 Identifying the Vulnerable: Education. Better educated individuals may adapt more easily to changing eco- nomic circumstances; they may use New Evidence from Peru assets more effectively, obtain better credit arrangements and exploit new Protecting vulnerable groups dur- experienced by different households. income opportunities more quickly. ing episodes of macroeconomic The study controls for a range of socio- Low education may hinder adaptationi, contractions is vital to poverty economic factors -- education, employ- thus raising vulnerability. reduction in developing countries. ment, wealth, and household Well-targete-d emergency programs can composition -- and identifies those fac- Empobeuiymgent. rSabliyvty ofeamings and provide teniporary income support to tors that predisposed households to job secuoradtypgeneal var alording tof the most needy, while longer-rnm social suffer consumption losses during the thge saetor n andytgiveof e hpousehofd investment funds offer the potential for period of macroeconomic shock. wAge eacroners in an civechumsehold. removing thie cniosthtpredispose change, unstable employment arrange- certain hous;ebolds to vulnerability. Yet, and a rieahoshl' K\~~~~~~~~~~~~~~et ma rais e a househLd's because it is not always clear who is Povertyanvuerblttocsmpindln. vulnerable and why, it has been difficult Vulnerabilityvunrblttocsmpindlne to design and target these poverty alle- Assets, Savings and Access to Credit. viation programs so that they reach The terms "poverty" and "vulnerability" Households hit by unexpected income those who nieed help the most. are commonly misinterpreted. Poverty shocks may adapt by selling assets, Drawing on the experience of Peru refers to an individual's current socio- drawing down savings, or taking out between 19,85 and 1990, a recent World economic status; vulnerability, on the credit, and through such means may Bank study explores the determinants of other hand, refers to potential changes in maintain consumption levels. House- vulnerabilily. The period of study co- this status. The poor are not necessarily holds without such buffers may be more incides weith an unusual set of vulnerable; for example, subsistence vulnerable to consumption loss. 'heterodox' policies implemented in farmners in remote areas are almnost al- Peru that prompted tremendous macro- ways poor, but are not particularly Labor Force Participation. House- economic shocks to households. Per vulnerable to macroeconomic shocks. holds may absorb economic shocks by capita output fell by 15%, wages de- It is important to distinguish poverty increasing labor force participation. rined by 30-60%; public sector from vulnerability because when coun- Households with fewer employable expenditures were slashed by 50%, with tries suffer economic shock, certain members, or with members that face concomitanyt cuts in the social sector, households that were not poor may risk discrimination in the labor market ranging from 45% (education) to 75% becoming so. Policy analysts focusing (women, minorities), may have less ca- (housing). attention on the current poor may over- pacity to adapt in this manner, thereby Data from the Peru Living Standards look important segments of the suffering greater consumption loss. Studies (PALSS) in 1985-86 and 1990 population who, while not currently show that in Lima, average per capita impoverished, are nonetheless suscep- Participation in Inter-HouseholdAssis- consumption fell by 50%, and the pov- tible to poverty. tance Networks. Access to income erty rate rose from 13% to 54%. How are vulnerable groups best support from other households may Exploiting. a rich subset of panel data identified? Common sense suggests the provide a fom of consumption insur- from these surveys that tracked 699 following household characteristics are ance, particularly if the network households over the period, the study likely to be associated with vulnerabil- includes households outside the area hit examines the change in consumption ity: by economic shocks.  Gender of Head of Household. House- additional member offset by 2.5% the findings suggest that a temporary in- holds headed by women may be average household consumption de- come subsidy targeted to female-headed disadvantaged due to fewer labor mar- cline. However, households with a households would be less effective in ket opportunities and additional larger number of children were at a dis- protecting vulnerable households than a caretaking responsibilities that prohibit tinct disadvantage, suffering a 7% per-child income subsidy. employment. Such factors may raise greater decline in consumption for each In terms of the underlying detenni- the chances of consumption loss. child. nants of vulnerability that can be Finally, there was no evidence that addressed by public policy, the study The Evidence from Peru networks of exchange between house- concludes that education is perhaps the holds within Peru protected against most important factor in determining a A recent World Bank study has shed consumption decline. However, ex- household's capacity to adapt to eco- light on .tereltion t thoeaove change networks with households in nomic shocks. Raising levels of scharaictvulnerbistycsa. ahousehold's foreign countries did reduce consuimp- educational attainment may be the sin- economic vulnerability. The report of tion loss significantly, by 13%. gle most effective long-term strategy for fers a mixed assessment of thesereuighshodvleblty relationships, indicating the importance of drawing on hard empirical evidence Conclusions Finally, the study suggests that inter- when formulating policy. Poverty alleviation programs -- both household networks established within a The data from Peru offer strong sup- short- and long-term -- require data that country cannot be relied on to protect port for the correlation between level of can help to accurately identify vulner- tion loss during periods of economic education and reduced vulnerability. able groups. During periods of shock. The only households to be so Each year of education attained by the economic shock, such programs may protected in Peru were those that had head of household produced a 2% re- then be able to better target households conectins wit eoshos in fign duction of consumption loss. that will most require temporary income counetris witrepresengles tn 3of - . O . ~~~~~~~~~~~~~countries -- representing less than 3%/ of Female-headed households were more support in order to maintain their cur- households in the study. While the likely to be poor at the outset of the pe- rent economic status. Policy analysts overseas transfer network survived, the riod, but, surprisingly, were no more may also use this data to understand domestic network apparently ceased to vulnerable to consumption loss than better the conditions under which vil- function as the economic crisis took male-headed households over the pe- nerability arises in the first place. hold. In light of these findings, policy riod. Further, there was no evidence A case study using panel data for 6599 makers should not readily assume that that the sector of employment of the Peruvian households during a particu- households can escape poverty inde- household, nor the household's access larly volatile economic period (1985- pendently of govemmental aid. to credit and savings, significantly influ- 1990) provides some valuable insights enced consumption loss. for the formulation of poverty allevia- There is evidence that larger house- tion programs. Sources: Paul Glewwe and Gillette Hall. 1995. holds were at an advantage over the First, and somewhat surprisingly, fe- Who is Most Vulnerable to Macroeconomic period, being able to diversify and in- male-headed households exhibited no Shocks? Hypotheses Tests Using Panel Data from Peru. LSMS Working Paper #117. The World crease participation in the labor market, disadvantage, but the number of child Bank. Washington, DC. pp.34. while also taking advantage of econo- dependents in a household was strongly mies of scale in consumption. Each correlated with vulnerability. These 5 * Poverty Lines is prepared by Gillette Hall with assistance from Andrea Ramirez. This issue was edited by Brian Ollihane. For information or a subscription, contact. Visit our World Wide Web sites: LSMS Administrator (Attn: Poverty Lines) PSP Poverty Home Page: http://hc:o.worldbank.org/htnl/hcovp/poverty/contents.html World Bank, N8-058 LSMS Home Page: http://www.worldbank.orglhtmVprdph/lsms/lsmshome.html 1818 H Street, N.W. Washington, D.C. 20433 Fax: (202) 522-1153 Internet: Isms@worldbank.org