INTERNATIONALBANK FOR WORLD BANK R E T C N O E N STRUCTION PM AND DEVELO April 2003 No. 22 A regular series of notes highlighting recent lessons emerging from the operational and analytical program of the World Bank`s Latin America and Caribbean Region THE IMPACT OF VULNERABILITY AND SHOCKS ON THE POOR THE GUAPA PROGRAM IN GUATEMALA Emil Tesliuc and Kathy Lindert "As for the earthquake (of 1976), it affected the families because the majority of houses were destroyed. Most had to build shelters between the fields and in the road.... Many still haven't repaired their houses." - Kaqchiqel villagers in Guatemala The Guatemala Poverty Assessment report was part of a proactive approach (given vulnerability, poverty and risks, broader, multi-year program of analytical work and technical what can be done to get help poor people escape poverty and assistance that signals the commitment of the World Bank to reduce the likelihood that others will fall into poverty?). poverty reduction and seeks to: Using a combination of quali- (a)contribute to filling the crucial in- tative and quantitative data, the formation gaps on poverty and living RVA analyzed both vulnerabil- conditions; (b) deliver timely outputs ity to shocks and vulnerable on a regular and on-going basis in groups. response to the Government's re- quests and data availability; and (c)provide longer-term partnering Combining Qualitative and collaboration on poverty analysis and Quantitative Data and strategy. This program is collectively called A unique feature of the RVA the Guatemala Poverty Assessment work in Guatemala was the Program (GUAPA program, see Box combination of quantitative 1), which is intricately linked to the and qualitative data on vulner- World Bank's support for the ability. First, since quantitative MECOVI program in Guatemala. The main counterpart panel data were not available, a module on risks and shocks agencies for the GUAPA program have been the Instituto was introduced into the first Living Standards Measurement Nacional de Estadística ­ Guatemala (INE) and the general Survey (ENCOVI 2000). This module included retrospective planning secretariat under the presidency (SEGEPLAN) in questions to try to capture dynamic issues via cross-sectional collaboration with the university of Rafael Landívar (URL). data. These data include information on: (a) 26 types of shocks (economic, social, life-cycle, natural); (b)whether the One of the major components of the study examined how the shocks triggered a loss in consumption, income or wealth; (c) limited assets of the poor make them particularly vulnerable the main coping strategies used to compensate for these to adverse shocks. In the wake of recent shocks in Guate- losses; (d) whether the households were able to compensate mala (Hurricane Mitch in 1998, the recent coffee crisis, for the welfare loss; and (e) the estimated time until successful droughts and deaths from extreme acute malnutrition), the resolution of the situation. Second, a complementary qualita- issue of vulnerability has taken center stage in policy discus- tive study was conducted in 10 rural villages from the sions. As such, the World Bank conducted a Risk and ENCOVI sample, covering 5 different ethnicities. This Vulnerability Assessment (RVA) as part of the broader Qualitative Poverty and Exclusion Study (QPES 2000) in- GUAPA. The RVA brings a "vulnerability lens" to poverty cluded a module of open-ended questions on perceptions of analysis. Instead of a passive, reactive approach (given shocks (occurrence, main coping strategies/response, and poverty, what can be done to reduce it?), it takes a dynamic, types of assistance received if any). 1 Vulnerability to Shocks The concept of vulnerability has two elements: (a) a person or household's resilience to shocks, which is largely based on their assets (higher resilience means lower vulnerability); and (b) the severity of the shock's impact (the more severe the impact, the higher the vulnerability). The ENCOVI 2000 reveals that poor households are more exposed to natural shocks, reflecting their dependence on agriculture as a source of living and their location (e.g., more likely to live in mar- ginal areas). The non-poor are more often the victims of eco- nomic shocks. (idiosyncratic) shocks. Most households experienced mul- tiple shocks (shocks tend to hit in bunches). Strong empirical The ENCOVI also shows that poor have lower resilience than evidence shows that the impact of a shock on household the rich. Some 88 percent of the extremely poor and 86 welfare is worse if there are other shocks as well (see Figure percent of the poor suffered losses in response to shocks, 1). compared with 83 percent of the non-poor. This is particularly notable for natural shocks, which caused welfare Shocks are difficult to predict, but key future sources of losses for half of the top quintile and two thirds of the bottom vulnerability include worsening terms-of-trade, reduced quintile of affected people. The probability of restoring remittances, and natural disasters ­ all shocks that could hurt incomes to pre-shocks levels rises with income. the poor. Coffee shocks and lost remittances would have severe and lasting impact. Natural disasters would have high and lasting impact, particularly for the poor. Shocks: characteristics and effects The effects of shocks are multi-dimensional. In addition to While Guatemala was spared major "macro" shocks in the economic effects on wealth and income, reported impacts in year 2000, households reported a high incidence of localized the QPES include: (a) psychological, such as the demoraliz- ing impact of job loss, the traumatic impact of vio- Box 1 - The four main interconnected "prongs" of the GUAPA program include: lence (fear, post-traumatic stress syndrome or susto); 1. "GUAPA collaborative," which seeks to foster institutional development and capacity (b) social, destroying trust building in counterpart agencies for greater ownership and sustainability of the analysis and and social capital within results. As such, the program adopts a collaborative approach, providing technical villages; (c) damage or de- assistance and hands-on training to counterpart agencies for poverty measurement, data struction of community as- collection, analysis, and policy and strategy formulation; sets (loss of road access, school destroyed, water 2. "GUAPA analytical," which seeks to conduct a thorough, multi-dimensional analysis of tank damaged); and (e) im- poverty building on the framework of the 's world development report for 2000/2001 using pacts on health (death, ill- both quantitative and qualitative data (the main product is this present report); ness) and education (chil- dren cannot attend school). 3. "GUAPA policy," which seeks to contribute to the design of the bank's upcoming Idiosyncratic shocks are al- country assistance strategy (CAS), as well as the upcoming country economic memoran- most always associated dum (cem), the government's poverty reduction policies and strategies, and the poverty- with income or wealth effectiveness of interventions and policies in Guatemala losses. Social shocks (e.g., violence, unrest) are less 4. "GUAPA operational," which seeks to forge linkages to lending operations currently likely to cause income or under preparation to improve their poverty focus, such as those in social protection, wealth losses but clearly education, and transport have psychological and so- cial impact. Some shocks, 2 like earthquakes, fire, or hurricanes, mainly affect household income loss of 28 percent for job loss, 19 percent for wealth and community assets. Other natural shocks primarily accidents of the breadwinner, and 17 percent for lower affect income. earnings and bankruptcy. Natural agricultural shocks had an important but less severe impact on household income: 11 The duration of impact varies by type of shock. Many QPES percent for harvest loss, 10 percent for pest infestation, and 9 villages report that families still live in homes that were badly percent for drought and worsened terms-of trade. damaged by the Earthquake of 1976 (over 25 years later). The shock worsened household living conditions that were already poor; there has been no recovery. Hurricane Mitch Vulnerable Groups in Guatemala had catastrophic consequences on some villages, completely wiping out their main productive base. The social and psychological impact of the conflict of the 1980s is also Certain characteristics of households can indicate which clearly long-lasting. groups (or characteristics) might be more vulnerable to shocks (due to structural features or lower resilience). A "vulnerability profile" predicts the probability that households will be poor in the future. Main Coping Strategies Chronic vs. Transient Poverty and Vulnerability Guatemalan households rely on their own assets and collec- tive (community) action as their main coping strategy; few The profile of vulnerability is similar to that of poverty, but receive public assistance. For most shocks, the main coping there are differences particularly for urban areas. strategies include reducing consumption, and self-help (sup- Vulnerability to consumption-poverty was estimated using a plying more labor, selling or mortgaging assets, drawing stochastic model of consumption and its variance, taking into down savings). Few households reported receiving Govern- account household characteristics and the likelihood of ment or NGO/donor assistance. Lower consumption was the experiencing shocks. The analysis reveals that 56 percent of main coping strategy in the face of falling household incomes the population was poor in 2000, but 64 percent had a greater (earnings) and public protests. Informal coping mechanisms than 50 percent probability of being poor in the future (or ("social capital"), such as borrowing or receiving help from `vulnerable to consumption poverty'). Most notably, higher friends, relatives or neighbors, were the primary coping vulnerability-poverty ratios were observed in the strategy for family disputes, accidents or death of household Metropolitan region (2.2 times higher) and urban areas (33 members. Formal insurance and credit were most common percent higher). In other words, poverty is lower in urban for insurable risks, such as fire, earthquakes, hurricanes, Table 1 - Classification of Poverty and Vulnerability: Transient vs. Chronic? and land slides. Numbers = % of total population The poor are less equipped Observed Poverty Status Based on Consumption than the rich to fight shocks. V Poor56.2% Non-Poor43.8% u For some shocks, such as in- Vulnerable l >50% chance of being C come or job losses, poorer n poor infuture Chronic Poor Vulnerable to Chronic Expected consumption E o e 64.2% (LMvulnerable) 44.4% Poverty(LM vulnerable)6.9% < poverty line51.3% x n households are more likely to r (LM = 51.4%) p s a e u reduce consumption than to (HV = 12.8%) b c m i use other strategies. Expected consumption t p l > poverty line e t i Frequently Poor Vulnerable to frequent = 12.8 (frequently poor) + d i t (HVvulnerable) 4.1% poverty(HV vulnerable)8.7% 35.9 (non-vulnerable) o Not all income or wealth y =48.7% n losses result in a reduction in Non-Vulnerable< 50% Non-vulnerable non- Infrequently Poor7.7% chance35.8% poor28.2% consumption. Most house- Poor = Chronic poor + transient poor; transient poor = frequently poor + infrequently poor. Vulnerable group = LM vulnerable + HV holds are able to smooth their vulnerable. LM vulnerability group = chronic poor + vulnerable to chronic poverty; HV vulnerability group = frequently poor + vulnerable to frequent poverty. Classification scheme adapted from Bidani and Richter (2001). Estimates from World Bank calculations consumption even when faced using data from the ENCOVI 2000, Instituto Nacional de Estadística - Guatemala. Numbers may not add exactly to 100% due to rounding. with shocks. In fact, just over a quarter of all shocks that resulted in income or asset losses forced households to cut areas and the capital but a significant share of the urban their consumption as a way of coping. In most cases, population is vulnerable to poverty. households were able to mitigate the effects of shocks or use coping strategies other than reducing consumption. In terms Poverty can be: (a) chronic, with a mass of people statically of severity, economic shocks had the highest negative impact living in poverty and transmitting it across generations; (b) on household income, consumption and wealth. transient, with many moving in and out of poverty; or (c) affect a specific set of sub-groups that are chronically poor The cost of shocks can be significant. The most severe and vulnerable due to specific features or circumstances. impact is associated with economic shocks, with an average Each has different implications for policy and targeting. 3 For policy formulation purposes, the pool of vulnerable were and promote early childhood development -- is crucial divided into two mutually exclusive groups: (a) those made to avoid inter-generational transmission of poverty and vulnerable by low expected mean consumption (79 percent) vulnerability; (labeled `LM vulnerable') and (b) those vulnerable to high volatility of consumption (13 percent) (labeled the `HV · Building the assets of the poor. The chronic nature of vulnerable'). The HV vulnerable are divided further between poverty and vulnerability implies that interventions the `frequently poor' among the current poor and the should aim to build the assets of the poor, as `vulnerable to frequent poverty' among the current non-poor. emphasized in the Peace Accords and the Government's poverty reduction strategy. Social · Chronic poverty dominates: 56 percent of Guatemala's protection programs can play an important role. population was poor in 2000, the majority of these (79 Specifically, when designed properly, conditional cash percent) were chronically poor (44 percent of the total transfer programs can be quite effective in helping ease population), a fifth were transient poor (12 percent of demand-side constraints, which have been shown to the total population). constitute important limitations for improved coverage of key assets, such as education and health. · Chronic vulnerability dominates: 64 percent of Guatemala's population was estimated to be vulnerable Disaster management and relief should be expanded and to poverty in the future; the main reason for their improved, given the disproportionate exposure of the poor vulnerability is low expected mean consumption, and vulnerable to natural disasters and agriculture-related which accounted for 79 percent of total vulnerability shocks. The introduction of catastrophic insurance may also (or 51 percent of the total population), whereas only a merit consideration. Interventions should be well-targeted to fifth are vulnerable due to high volatility of the poor and delivered in a timely manner. Since exposure to consumption (13 percent of the total population). some natural disasters seems to be largely determined by location and geography, maps of vulnerability to natural The chronic nature of poverty and vulnerability has important disasters could be useful for risk management planning. policy implications: interventions should build the assets of Their use, in conjunction with poverty maps, would greatly the poor, as emphasized in the Peace Accords and in the enhance ability to target limited funds for disaster relief. Government's poverty reduction strategy. Since natural disasters often damage or destroy community infrastructure (in addition to income and wealth losses at the Chronically Vulnerable Groups and the Life-Cycle Approach household level), social funds could be an institutional channel for relief and infrastructure rehabilitation. Certain groups are highly vulnerable due to special circumstances. There are other sources of vulnerability in ********** Guatemala besides consumption poverty. The RVA classifies risks over the life-cycle (by age group), assessing More Information vulnerability in nutrition, education, health, access to basic services and exposure to natural disasters. It also looks at the Learn more about GUAPA and our work in Guatemala here: number of poor people at risk and whether the risks or circumstances are likely to have lasting, even inter- http://www.worldbank.org/gt generational effects on poverty (e.g., education and Visit the Bank's Social Protection Website at malnutrition-related risks). Among the risks of particular concern are: malnutrition; low school enrollment, late school http://www1.worldbank.org/sp/ entry and grade repetition; child labor; low earnings; low health coverage of the elderly; lack of access to basic services among the poor; and higher exposure to natural disasters. About the Authors Seasonal migrants and their families also appear to have higher poverty and vulnerability rates than those who migrate Emil Daniel Tesliuc is a Social Protection Specialist in the permanently or the general (non-migrating) population. Human Development Network at the World Bank. Kathy Lindert also works at the Bank and is a Senior Economist in the Social Protection Group of the Human Development Policy Implications Department within the Latin America and Caribbean (LCR) Region. The analysis suggests that strategies to reduce vulnerability should emphasize: About "en breve" · Children. A strategic emphasis on children ­ particu- To subscribe to "en breve" send and email to larly child-focused interventions to reduce malnutrition "en_breve@worldbank.org" 4