External Shocks, Financial Crises, and Poverty in Developing ......external shocks explain 30 percent of cross-country variation in GDP volatility in Latin America. When terms of trade,

47

DEVELOPING COUNTRIES HAVE BECOMEincreasingly integrated into globalgoods and financial markets over thelast decade. Their export volume increased by9 percent per year during the 1990s, up from2 percent during the 1980s. Net long-termcapital flows, even after declining in 1998,remained almost three times the 1990 level.As discussed in previous issues of Global Eco-nomic Prospects (World Bank 1993, 1997,1999a), globalization provides developingcountries with significant benefits and spurseconomic progress. GDP growth in develop-ing countries (excluding the transition econo-mies) averaged 5 percent during the 1990s,compared with 3 percent during the 1980s.Poverty—the number of people living on lessthan $1 a day—fell from 29 percent in 1990to an estimated 24 percent in 1996. But thefinancial crisis of 1997–99 has also shownhow globalization, and in particular greateropenness to external capital flows, can exposedeveloping countries to increased volatilityfrom international financial and goods mar-kets. The poor are especially vulnerable to thisvolatility.

This chapter reviews the evidence aboutthe impact on poverty of the external shocksand volatility to which developing countriesare exposed. It then presents and assesses evi-dence of the impact of the 1997–98 financialcrisis on poverty in the most affected EastAsian countries. Finally, it discusses lessonsand policy conclusions.

The chapter reaches the following con-clusions:

• The financial crisis has underlined howglobalization, especially financial integra-tion, exposes developing countries to ex-ternal shocks. These shocks often reducethe gains in poverty reduction from open-ness and increase poverty significantly inthe short to medium term. This fact un-derscores the importance of addressing theissue of volatility in order to maximizethe positive effects of growth on povertyreduction.

• The countries most affected by the EastAsian crisis illustrate the asymmetric im-pact of changes in per capita income onpoverty and the negative effects of vola-tility on growth. Though less dramaticthan early predictions suggested and veryheterogeneous, the negative social impactof the East Asian crisis and consequentcrises in Russia and Brazil has been enor-mous. The increase in consumption pov-erty has been significant. In addition, thecrisis has resulted in large and costly re-allocations of people and sharp declinesin middle-class standards of living. Un-like the situation in Latin America whereincome inequality increased significantlyduring crises, in East Asia the effects onincome distribution have been small andhighly differentiated. The extent of theseeffects depends on the country’s income

2External Shocks,Financial Crises, and Povertyin Developing Countries

48

G L O B A L E C O N O M I C P R O S P E C T S

level and the impact of the crisis on dif-ferent economic sectors.

• Urban poverty increased in all countries,particularly the Republic of Korea, wheretotal employment declined and open un-employment grew more than in othercountries in the region. Falling real wagesin the urban formal sector affected mostlyhigh-income groups. In Thailand the im-pact was felt mostly in rural areas becauseof the large inflows of workers from ur-ban areas and the relatively small increasesin agricultural prices.

• The crisis demonstrated the flexibility oflabor markets in developing countries.These markets help absorb the effects ofshocks through reduced wages and labormobility within and between urban andrural areas. Thus the decline in total em-ployment in Thailand and Malaysia waslimited, and employment actually rose inIndonesia. Labor was reallocated from theformal (urban) sector to other activities,particularly the informal sector and agri-culture, where exchange rate depreciationsimproved incentives.

• Even where public spending on safetynets increased significantly, the impact onpoverty was limited for several reasons.These included the absence of safety netsbefore the crisis, response lags, institu-tional problems, and low levels of spend-ing relative to the scale of poverty. Insome cases evidence suggests that well-functioning programs were underfundedrelative to the potential impact of shockson poverty.

• The severity of the crisis in Indonesia isreflected in the strong responses of house-holds to increase consumption as a shareof income, adjust their asset holdings, andincrease the share of staple foods in theirconsumption baskets to cope with theshock. In the Republic of Korea and Ma-laysia the response of households was toincrease the savings rate. The composi-tion of consumption expenditures changedsignificantly. Households spent more, pri-

marily on essential items such as food,fuel, housing, health, and education.

• Real public expenditures on education andhealth fell in most countries. The extentto which households were able to adjusttheir spending to offset this decline var-ied across countries as well as incomegroups. In Thailand families and govern-ment programs acted to cushion the im-pact of the crisis in order to avoid declinesin school enrollment rates or in access tohealth services. In Indonesia, however, theseverity of the crisis led to significant de-clines in poor households’ access to botheducation and health services, particularlyin urban areas. Such setbacks can haveirreversible effects on human development.

• Any development strategy for stable andsustainable growth must include bothadequate safety nets and appropriatepolicies and institutions designed toprevent financial crises and respond whencrises do occur. Prospects for povertyreduction depend not only on futuregrowth but also on countries’ capacityto manage volatility and reduce growthfluctuations.

External shocks andpoverty in developing countries

Discussions of the link between growth andpoverty reduction in developing countriesimplicitly take the view that long-term growth(and therefore poverty reduction) is a stableprocess. But as the financial crisis of 1997–99shows, the process of growth is neither smoothnor linear and is often subject to sharp changes(especially major slowdowns and recessions)from a variety of external or internal shocks(World Bank 1999a). The asymmetric effectsof income growth on poverty during expan-sions and downturns, however, imply thatthese changes often have profound, long-last-ing effects on the poor. A decline in per capitaincome tends to have a negative effect on pov-erty that is much greater than the improve-ment generated by an equivalent increase.

49

While economic crises hurt both poor andrich, the poor have less leeway to respond tothe crises. If domestic capital markets wereperfect and the economic downturn tempo-rary, all economic agents could borrow tosmooth consumption and maintain welfare.But capital markets are imperfect and seg-mented. Credit or insurance is typically notavailable to the poor. With few savings andlow or subsistence incomes, the poor becomeeven more vulnerable to shocks. Crises andrecessions can result in irreversible negativeeffects on the poor through their impacts onhealth, schooling, and nutrition. Volatility ingrowth also tends to create more uncertaintyand risk for investors. That fact alone tendsto reduce the rate of economic growth, fur-ther dimming prospects for poverty reduction.Thus the volatility of the growth process indeveloping countries matters a great deal forboth immediate and long-term poverty reduc-tion and income distribution.

In general, the growth process is muchmore volatile in developing countries thanin industrial countries. Sudden reversals andother changes in international financial flows

are only one source (albeit an important one)of external shocks that can lead to crises andrecessions in developing countries. Fluctua-tions in the terms of trade are another im-portant and long-standing source, reflectingdeveloping countries’ reliance on primarycommodity exports and price variability ininternational markets. Volatility in the termsof trade was almost three times greater indeveloping countries than in industrial coun-tries during 1961–97 (Pritchett 1998; East-erly, Islam, and Stiglitz 1999) (figure 2.1).Volatility is particularly significant for theMiddle East and North Africa, LatinAmerica, and Sub-Saharan Africa. Usingsimulation models that replicate the range ofobserved economic fluctuations, Mendoza(1995) finds that disturbances in the termsof trade account for about one-half of theobserved variability in GDP and real ex-change rates, and that the share is greaterfor developing countries than for industrialcountries. Policies to mitigate and cope withvolatility in growth and the consequent ef-fects on the poor are therefore essential inall developing countries.

Figure 2.1 Terms of trade and GDP growth volatility, 1961–97

14

Standard deviations (percent)

3.9

11.3

7.4 8.1

9.7

11.413.1

2.5

5.34.2

2.5

8.6

4.75.6

12

10

8

6

4

2

0Terms of trade Real GDP

High-income

Low- and middle-income

East Asia and Pacific

South Asia

Middle East and North Africa

Latin America and the Caribbean

Sub-Saharan Africa

Note: Unweighted average of countries' standard deviations of relative distance from Hodrick-Prescott filter trend.Source: World Bank staff calculations.

E X T E R N A L S H O C K S , F I N A N C I A L C R I S E S , A N D P O V E R T Y

50


External shocks, long-term growth, andpovertyExternal shocks, such as variations in theterms of trade, volume of trade, and externalfinance, are highly correlated with variationsin GDP growth. They account for a signifi-cant share of the volatility in developing coun-tries (Easterly, Islam, and Stiglitz 1999).According to Hausmann and Gavin (1995)external shocks explain 30 percent of cross-country variation in GDP volatility in LatinAmerica. When terms of trade, export vol-umes, external finance, and interest rateshocks are taken into account, developingcountries experience more and larger exter-nal shocks than industrial economies. Theincidence of small and medium-size shocks isabout the same for both (World Bank 1993).During the 1970s and 1980s it was not un-usual for developing countries to suffer unfa-vorable shocks equivalent to 4 percent of GDPor more.

Volatility of growth and other macro-economic variables is also much larger indeveloping countries than in industrial coun-tries (Pritchett 1998; Easterly, Islam, andStiglitz 1999). Figure 2.1 shows that volatil-ity in GDP growth is more than twice as highin developing countries as it is in high-in-come countries of the Organisation forEconomic Co-operation and Development(OECD). The volatility of GDP growth ishigher for all developing regions, except forSouth Asia, and it is more than three timeshigher for the Middle East and North Af-rica. GDP growth in developing countries ishighly unstable, with large shifts over timeand low correlation of per capita growth ratesacross decades (Easterly and others 1993;Pritchett 1998).

Volatility has a negative impact on pov-erty in part because it reduces long-termgrowth (box 2.1). For instance, a large degreeof volatility makes “stop and go” policies morelikely, slowing growth and leading to low-quality policies such as those in Sub-SaharanAfrica, especially during the 1970s and 1980s.(Guillaumont, Jeanneney, and Brun 1999). Ex-

ternal negative shocks can also interact withsocial conflicts and weak domestic institutionsfor conflict management to produce growthcollapses (Rodrik 1998). After controlling forother factors, Hausmann and Gavin (1995)find that a higher standard deviation of realGDP is associated with higher rates of pov-erty. They estimate that if Latin Americancountries had the same GDP volatility as in-dustrial countries, poverty would decrease by7 percentage points.

External volatility and fluctuations inpovertyVolatility does more than simply increase pov-erty. Short-term fluctuations in income growthalso cause sharp variances in the incidence ofpoverty, even in the short to medium term.For example, in Venezuela poverty decreasedby 10 percentage points between 1989 and1991, rose by 20 percentage points between1991 and 1994, then fell again in 1995 androse in 1996 (Lustig and Deutsch 1998).Mexico is another striking example of this ef-fect, as box 2.2 describes.

Fluctuations in commodity prices may in-duce short- to medium-term changes in bothgrowth and poverty. During the boom yearsgrowth is faster and poverty declines, but dur-ing busts, which are usually more sudden, pov-erty increases. Fluctuations in commodityprices have a significant, direct impact on per-sonal incomes and an indirect impact on gov-ernment social expenditures and GDP. Earlierstudies argued that commodity price boomsdo not significantly affect real GDP(Cuddington 1988; Gelb and Associates1988). But more recent empirical work hasshown that changes in terms of trade havesignificant effects on real output growth.1 De-clining trends in real commodity prices havea negative effect on real income growth in thelong term in developing countries (see chap-ter 4).2 In addition, a slowdown in growth inthe bust years may become more severe as in-vestments made during the boom years areoften less productive (Collier and Gunning1996).

51

Empirical findings also support the no-tion that economic cycles have an asymmetriceffect on poverty. They show that a contrac-tion will have a greater impact on the povertyrate than an expansion of the same size(Morley 1994; Londoño and Székely 1997a;De Janvry and Sadoulet 1998).10 It has beenestimated that a 1 percent decline in per capitaincome during recessionary episodes in LatinAmerica in the 1980s reduced earlier gains by3.4 percent of per capita income growth inurban areas and 2.2 percent in rural areas (DeJanvry and Sadoulet 1998). One explanationfor this phenomenon is that during recessionsthe unskilled are the first to lose their jobs,because firms tend to hoard their skilled em-

ployees. As a result, income distribution be-comes more inequitable, amplifying the effectof declining incomes on poverty (Agénor1998).

Income poverty and inequalityduring the East Asian crisis

The recent crisis in East Asia has under-lined the risks for developing countriesof reversals in private capital flows and thedramatic social impact of the resulting finan-cial crises. The East Asian crisis had a sub-stantial impact on output and poverty in1998, although these effects began to lessenin 1999.


irreversibilities or asymmetric adjustment costs ininvestments increase uncertainty and lower invest-ment (Pindyck 1991; Aizenman and Marion1999).7 Second, costs increase because productivefactors move among sectors in response to morefrequent shifts in price signals. And third, the riskof inappropriate monetary, fiscal, trade, and finan-cial policies increases.

Terms-of-trade volatility has been found tohave a negative effect on long-term growth in de-veloping countries. Commodity price uncertainty,as measured by the standard deviation of forecasterrors from some statistical models, reduces growthrates (Dehn and Gilbert 1999).8 Most empiricalstudies have used direct volatility of terms of tradeas a proxy for uncertainty and have found negativeeffects on long-term growth (Mendoza 1994;Hausmann and Gavin 1995; Guillaumont,Jeanneney, and Brun 1999; Easterly and Kraay1999).9

The overall evidence also indicates that overthe long run the dependence of many developingcountries on commodities with volatile prices has anegative impact on long-term growth and thereforeon poverty. Dehn and Gilbert (1999) find a signifi-cantly negative effect of commodity price uncer-tainty on poverty, as measured by infant mortality.

Box 2.1 Volatility, growth, and poverty

W hen growth proceeds smoothly over timeand income inequality improves (or at leastdoes not worsen dramatically), poverty declines asper capita income and real wages rise. The elastic-ity of poverty, as measured by the headcount in-dex—for example, with respect to the growth ofper capita income3 —is estimated to be between–1.5 and –3.5.4 The size of the effect is greater incountries where income is more evenly distributed(Ravallion 1997).5

To the extent that volatility creates uncer-tainty, it has negative effects on growth and there-fore on poverty. Recent empirical evidence supportsthis view and contradicts the early literature. Usingbalanced panel data for a sample of 92 countriesfor 1960–85, Ramey and Ramey (1995) find that aunit increase in the standard deviation of innova-tion in GDP (innovation to GDP growth is used asa measure of uncertainty) implies a lower GDP percapita growth of 0.2.6 Similarly, from a growthregression of 130 countries for 1960–95, Easterlyand Kraay (1999) find that the standard deviationof growth has a strong negative effect (-0.18) onaverage per capita growth (after controlling forother variables).

There are three likely explanations for thenegative link between volatility and growth. First,

52


The impact of the crisis on povertyIncome poverty almost invariably increasesduring a crisis. Household surveys conductedin Latin America during recessionary periodsin the 1980s and 1990s provide evidence ofthis effect. They show that the incidence ofpoverty increased during the first year of therecession in 9 out of 11 cases, and remainedhigher for one or more years after the reces-sion in 19 out of 21 episodes (Lustig 1999).

Poverty also increased during the first yearof the crisis in the most affected East Asian

countries (table 2.1). Evidence from Korea il-lustrates the asymmetric impact of crises onpoverty. During stable growth in 1990–97, theestimated elasticity of the percentage of poorwith respect to per capita GDP was –3.5(Kakwani and Prescott 1999). But during thecrisis in 1998 the incidence of poverty in-creased by 123 percent. Real per capita GDPdeclined by 6.7 percent, and consumption percapita declined by 10.4 percent. In Indonesiaas well, the rate of increase in poverty wasabout 10 times the rate of the decline in con-

ing a Gini index of 0.54 in 1989 that remainedunchanged until 1996 (Székely 1998). Total pov-erty rose from its lowest point of 28 percent in1984 to 36 percent in 1989, or from 20.7 to 29.6million poor. Infant and preschool mortality causedby nutritional deficiencies increased from 1982onward, and educational indicators for the poordeteriorated (Lustig 1998).

From 1989 to 1994 growth resumed, largelydue to economic and financial liberalization, realper capita GDP growth averaged 2 percent.Although the total poverty headcount index haddeclined slightly to 34 percent by 1994, thenumber of poor had increased to 30.7 million(Lustig and Székely 1998; Székely 1999a).According to Székely (1999b) 86 percent of therise in poverty trends in Mexico from 1984 to1994 resulted from the increase in inequality,while the rest was the result of the drop in GDPper capita.

From 1989–94 poverty rose among ruralworkers in the primary sector and in the southernand southeastern regions. This increase was theresult of the appreciation of the peso and thedecline in institutional support for agriculture,including the loss of subsidies, the collapse ofguaranteed prices for major crops, and high inter-est rates (Lustig and Székely 1998). The financialcrisis that hit Mexico at the end of 1994 hadconsiderable repercussions for growth and total

Box 2.2 External shocks and fluctuationsin poverty in Mexico

Mexico’s experience since the 1970s shows howpoverty declines during periods of economicgrowth and increases during periods of crisis andadjustment. External shocks contribute to thesevariations.

During the 1970s Mexico experienced rela-tively high and sustained growth. The increase inreal GDP per capita averaged 3.8 percent per year,despite the short-lived financial crisis of 1976 (seebox figure). Income inequality declined: the Giniindex dropped from 0.58 in 1970 to 0.51 in 1977(Londoño and Székely 1997b). Total poverty fellsignificantly, dropping from 49 percent in 1968 to34 percent in 1977, or from 23.3 to 21.3 million.Further gains were realized during the second halfof the decade (reflected in the numbers for 1977–84), spurred partly by favorable terms-of-tradeshocks and rising oil production.

In the early 1980s the international environ-ment became unfavorable for Mexico. Thecountry’s terms of trade declined, and real interna-tional interest rates increased. The resulting debtcrisis, the adjustment of the 1980s, and the col-lapse of oil prices in 1986 resulted in a sharp de-cline in incomes. Between 1982 and 1988 real GDPper capita growth averaged a negative 1.9 percentper year, and real wages fell by 36–46 percent from1983–88. This decline in incomes contributed to adramatic increase in poverty (Lustig 1998). In-equality increased sharply in the late 1980s, reach-

53


Table 2.1 Growth, poverty rates, and Gini coefficients in East Asia, 1996–98

Indonesia Malaysia Rep. of Korea Thailand

Real per capita GDP growth (percent)

1997 2.9 5.4 4.5 –1.41998 –15.1 –9.2 –6.7 –10.3

Real per capita consumption growth (percent)1998 –5.5 –12.4 –10.4 –11.6

Headcount poverty indexa National National Urban National1996 11.3 — 9.6 11.41997b 11.0 8.2 8.6 9.81998 16.7 — 19.2 12.9

Gini index1996 0.380 — — 0.4771997 — 0.496 0.290 —1998 0.370 — 0.294 0.481

— Not available.a. Figures for Indonesia are based on consumption expenditures, with a national poverty line equivalent to about $1 a day in 1985international purchasing parity (IPP) dollars; data are from February 1996 and December 1998. Figures for the Republic of Koreaare based on consumption expenditures, with a national poverty line equivalent to about $4 a day in 1985 IPP dollars. Figures forThailand reflect national income poverty, measured at around $2 a day. Figures for Malaysia reflect income poverty.b. The 1997 figures for Indonesia and Thailand are estimates based on precrisis trends in declines in poverty.Source: Kakwani 1999; Kakwani and Prescott 1999; World Bank staff calculations.

Box 2.2 (continued)poverty. Real GDP growth declined to a negative6.2 percent in 1995 and averaged 2.6 percentfrom 1995 to 1998. Total poverty increased dra-matically, rising to 45 percent in 1996, or 41.7

million (Székely 1999a). Mexico has adjusted tothe crises primarily through downward flexibilityin real wages rather than through increases inunemployment.

Changes in poverty, inequality, and per capita GDP in Mexico

15

1968–77

1977–84

1984–89

1989–92

1992–94

1994–96

10

5

0

–5

–10

–15

Percent

Change in total poverty

Average per capita GDP growth

Change in Gini coefficient

54


sumption per capita—much higher than theusual elasticity during expansions. While theselosses have been reversed somewhat since1999, the extent and sustainability of the re-covery remains to be seen (see chapter 3). Evenreturning to the precrisis level of poverty, how-ever, is likely to require more time and incomegrowth.

The severity of the impact of the EastAsian crisis varied across countries. Differencesin national poverty levels and the distributionof the income of the poor around these levelsmay explain some of the variances. For in-stance, in Korea the poverty line is around $4per day, while in Indonesia it is around $1 perday. If the individuals whose incomes droppedsignificantly are clustered above the povertyline in Korea and below the poverty line inIndonesia, the impact of the crisis on povertymay well appear lower in Indonesia. But otherfactors are also responsible for these differ-ences.

Korea’s experience was strikingly differ-ent from the others. Korea had the largest in-crease in open unemployment, a decline in theeconomically active population, and a largedrop in real wages that was second only toIndonesia’s. Labor mobility from the informalsector was also more limited than in othercountries. Korea is also the most urbanizedEast Asian country, and the negative impactof recessions has been found to be most dev-astating for poor urban dwellers (Morley1994; Lustig and Deutsch 1998; De Janvry andSadoulet 1998). The increase in urban pov-erty in 1998 was huge in Korea: the headcountindex, based on consumption expenditures,reached 19.2 percent, an increase of more than10 percentage points.11 The increase was evengreater (15 percentage points) between the firstquarter of 1997 and the third quarter of1998—the lowest (7.5 percent) and highestpoints (23 percent), respectively (Kakwani andPrescott 1999). The incidence of poverty de-clined to 15.8 percent in the last quarter of1998.

In other countries the increases weresmaller than had been anticipated, given the

magnitude of the crisis (table 2.1). In Indone-sia the impact of the crisis on poverty was stillsignificant. Estimates for Malaysia are notavailable, but welfare declines were wide-spread, presumably leading to increases inpoverty in both urban areas and traditionallypoor rural states.

Urban and rural poverty. Urban povertyincreases during crises owing to a combina-tion of lower real wages, higher unemploy-ment, and increases in the relative price offoods. The impact of a crisis on poverty willbe smaller if workers can move easily fromthe formal sector to other activities, particu-larly agriculture, and if exchange rate depre-ciations lead to improved incentives foragriculture. Even under those conditions, how-ever, it is still likely that urban poverty willincrease.

The relative impact on urban and ruralareas was different in Indonesia and Thailand.In Indonesia the crisis had a strong urban bias,even though the percentage changes in pov-erty rates were similar in urban and rural ar-eas. Poverty in urban areas rose from 9.7percent in 1996 to 15.4 percent in 1998, andin rural areas climbed from 12.3 percent to17.6 percent. Average per capita spending inurban areas fell 34 percent in real terms,whereas rural expenditures fell only 13 per-cent. A survey of expert respondent views sug-gests that urban areas were, on average, muchharder hit than rural areas (Poppele, Sumarto,and Pritchett 1999). Of the 20 hardest-hit ar-eas, 14 were urban, while of the 20 that suf-fered the least impact, 13 were rural. In nearlyevery province, region, and island, the nega-tive impact of the crisis was consistently higherfor urban than for rural areas. In Thailand,however, the impact on poverty was more se-vere in rural areas than in urban. Poverty ratesrose from 11.8 to 17.2 percent in rural areas,but only from 1.2 to 1.5 percent in urban set-tings.12 One possible explanation for the dif-ference in the impact of the crisis on the twocountries is the higher price incentives for ag-ricultural production in Indonesia, whichstimulated production.

55

Regional effects. The impact of the crisisvaried considerably across subnational regions.In the northern region of Thailand, for ex-ample, the poverty ratio actually dropped from10.2 percent in 1997 to 9.2 percent in 1998.In the northeastern and southern regions it rosedramatically, climbing from around 15 per-cent to 23.2 percent and from 8.6 to 14.8percent, respectively. In Indonesia per capitareal expenditures declined by 42 percent inWest Java and by 30 percent in Jakarta, re-gions that were better off before the crisis. Butreal expenditures declined between 10 and 20percent in other regions. These differences aremost likely linked to the behavior of producerprices. In Indonesia, areas that produced ex-port crops benefited from the sharp exchangerate depreciation. This fact combined withseveral reforms (such as clove marketing) toput more benefits in the hands of farmers.13

Similarly, in Thailand the poor performanceof the southern region may be linked to thefall in rubber prices during the crisis period.An important aspect of the crisis in Indonesiais that it does not appear to have affected poorareas disproportionately. Rather, the impactvaried in both well-off and poor areas.

The impact on income distributionGiven the significant drop in GDP normallyassociated with economic crises, poverty rateswill increase unless there is a massive reduc-tion in inequality. But income distributiontends to worsen during crises. Inequality inhousehold incomes or consumption increasedin most of the countries in Latin America dur-ing crises and recessions in the 1980s (WorldBank 1999a). For 10 recession episodes forwhich data are available in Latin America,inequality rose in 6 cases during the recessionyear (Lustig 1999). In Argentina the Gini co-efficient for the greater Buenos Aires area in-creased from 0.44 to 0.53 during the recessionof 1989. Inequality was higher after the re-cession than it had been before in 15 out of22 episodes. In Chile the Gini coefficient ontotal household incomes is estimated to haveincreased from 0.52 in 1979 to 0.55 in 1984

because of the impact of high open unemploy-ment and (by some measures) deep real wagecuts. However, the Gini coefficient in Chilehad declined to 0.53 by 1988 (Riveros 1994).

In Brazil income inequality increased, de-spite a successful defense of real wages in theformal sector and little increase in open un-employment, in part because of inflation anddeclining incomes in the informal and agri-cultural sectors (Fox, Amadeo, and Camargo1994). In Argentina average real wages oscil-lated wildly with episodes of inflation duringthe 1980s, though unemployment was nothigh. However, the gap in earnings betweenthe top and bottom deciles of income earnersin Buenos Aires widened steadily from 1980through 1988 as younger adults increasinglyentered the informal sector (Riveros andSanchez 1994).

Analyses of the effects of crises on incomedistribution suggested that the impact differedin middle- and low-income countries(Bourguignon, de Melo, and Suwa 1991). Dur-ing most economic crises and subsequent struc-tural adjustments in middle-income countries,income distribution worsens because wage cutsand layoffs in the formal sector tend to be bi-ased toward unskilled workers. The impactof crises on inequality in low-income coun-tries is more difficult to predict. Wage and em-ployment losses in the urban formal sectoraffect workers with relatively high incomes,and the rise in food prices hurts the urban poor.But the rural areas where most of the poorlive tend to gain because of currency depre-ciation and higher prices for agriculturalgoods. Bourguignon, de Melo, and Suwa(1991, 359) find, from simulations, that “inthe standard adjustment package, inequalityincreased significantly for the Latin Americanarchetype but decreased significantly for theAfrican archetype.” A major reason for thisdifference is that there are few formal sectorwage earners in the bottom half of the incomedistribution ladder in very poor countries—for instance, those in much of Sub-SaharanAfrica. Because crises hit the formal sectorhardest, the poor are less affected. In Latin


56


America, however, where formal sector work-ers come from all income brackets, poor peopleare hit more directly in a crisis.

Compared with Latin America in the1980s, the distributional impact of the EastAsian crisis was limited for high-income coun-tries (Korea), upper-middle-income countries(Malaysia), and lower-middle-income coun-tries (Indonesia and Thailand). Changes inoverall inequality, as measured by the Gini co-efficient, were minor between 1996 and 1998(table 2.1).14 In Thailand there may have beenweak redistribution from middle- to high-in-come groups (Kakwani 1999). But early stud-ies for Korea and Thailand suggest that thoseat the bottom of the income distribution lad-der—the “ultrapoor”—were hit harder thanothers with incomes below the poverty line(Kakwani 1999; Kakwani and Prescott 1999).The evidence is more mixed for Indonesia(Poppele, Sumarto, and Pritchett 1999).

Labor incomesTo a large extent the impact of the crisis onconsumption poverty reflects changes in thereal incomes of households. The channelsthrough which the impact of a crisis reacheshouseholds can be traced to the sources ofhousehold income—that is, wages, returns onassets, profits from self-employment, andtransfers (Ferreira, Prennushi, and Ravallion1999). These sources tend to vary with house-hold income level—for example, poor house-holds tend to depend on self-employmentincomes and transfers, whereas the rich receivemuch of their income from assets. For thisreason, changes in the overall composition ofnational income can move households up ordown the distribution ladder.

Labor markets have the most profoundeffects on poverty, however. Labor demandshocks hurt households by lowering realwages, increasing unemployment, and reduc-ing self-employment earnings. While reducedlabor demand almost always raises the inci-dence of poverty, different kinds of labor de-mand–shocks have different effects on incomeinequality. In a recession, real wages fall.

Households at the low end of the distributionladder in developing countries are affected theleast, because they receive little or no wageincome. But labor demand shocks have astrong impact on those formal sector workerswith the lowest skills, who are more likely tolose their jobs than their more skilled coun-terparts.15 They then either become unem-ployed or move to the informal sector, wheretheir earnings are likely to be lower. As a re-sult households at the middle to lower-middlerange of the income distribution ladder arepushed further down, swelling the numbersof households with low incomes.

The crises in East Asia followed a patternsimilar to those seen earlier in other countriesfaced with sharp reversals of external capitalflows. A comparative analysis of the impactof similar crises on labor markets offers thefollowing conclusions (Fallon and Lucas1999):

• Wages fall sharply during the crisis or inensuing years, usually by more than theGDP. In 22 recessionary episodes in LatinAmerica during the 1980s and 1990s, realwages fell in 16 cases during the year ofrecession, and in 18 cases remained lowerthan precrisis levels after two years (Lustig1999). This wage drop was also a strik-ing feature of the East Asian crisis.

• Total employment growth drops in thecrisis year, but usually by less than thedecline in GDP growth.

• Employment in manufacturing is alwaysadversely affected, though less spectacu-larly than are wages.

• The effects on agricultural employmentare more muted. In some cases (for ex-ample, Indonesia in 1998 and Turkey in1994), employment increased despite anabsolute decline in GDP.

• Rising unemployment is an important fea-ture of many crises. In Latin America un-employment increased during the year ofrecession in 24 of 31 episodes and re-mained higher in 24 cases two years intothe recession (Lustig 1999). The most sig-

57

nificant increases were in Argentina in1995 (6 percentage points) and Chile in1982 (11 percentage points).

Experience thus far in East Asia broadlysupports these conclusions. Real wage growthdropped sharply in 1998 and became nega-tive in all affected countries (table 2.2). Indo-nesia saw particularly spectacular wagedeclines that were broadly similar across sec-tors. Although wage cuts can moderate theimpact of a recession on employment, duringthe East Asian crisis nonagricultural employ-ment fell in all countries. Only in Indonesia,where agricultural employment increased con-siderably, did overall employment rise. Theconstruction sector was the most affected, witha dramatic drop in employment of 15 to 35percent, but manufacturing employment alsofell significantly. With the exception of Ko-rea, however, falls in employment in 1998 werenot large despite substantial decreases in GDP.(Korea saw a large decline in employment aswell as in real wages.) The inactive popula-tion increased by 9 percent between the sec-ond quarter of 1997 and the fourth quarter

of 1998, with women representing three-fourths of the increase. In Thailand 18.5 per-cent of the overall decline in per capita incomewas the result of wage cuts, whereas only 2.7percent was attributable to higher unemploy-ment (Kakwani 1998).

Labor force mobility. To some degree, theimpact of the crisis on employment was less-ened by the mobility available to individualworkers within and between the urban andrural sectors. In the first year of a crisis sig-nificant real exchange rate depreciation usu-ally results that can raise the price of tradablegoods relative to those of nontradables, withimportant implications for real household in-comes and poverty. Crises hit the urban for-mal sector first and, as noted above, can leadto a reallocation of labor from the urban tothe rural sector. But in the absence of any in-centive to increase agricultural production, thisreallocation of human resources may do littlemore than raise rural poverty instead of ur-ban poverty. In principle exchange rate depre-ciation can supply the needed incentive. In theabsence of intervention in domestic markets,it raises the price of export crops relative to


Table 2.2 Employment and real wages in East Asia during the crisis(percentage change)

Indonesia Malaysia Rep. of Korea Thailand1997 1998 1997 1998 1997 1998 1997 1998

Employment, totala 1.8 2.6 4.6 –2.7 1.4 –5.8 1.8 –3.0Agricultural –4.7 13.3 –0.6 –5.3 –3.4 0.0 1.3 –1.8Nonagricultural 6.8 –4.7 5.8 –2.2 2.4 –6.5 2.2 –3.9Manufacturing 4.1 –9.8 7.6 –2.9 –4.3 –13.1 –0.1 –1.9Construction 10.6 –15.9 8.9 –13.4 1.7 –26.4 –5.6 –33.6

Real consumption wage, totalb 8.6 –41.0 — — — — 5.7 –1.5Agricultural 4.1 –35.0 — — — — 10.0 –8.9Nonagricultural 9.9c –42.0c — — 2.6 –10.0 5.0 –0.5Manufacturing 11.1 –44.0 6.0 –2.4 0.7 –10.6 7.1 –4.5Construction 8.5 –42.0 — — 3.3 –14.7 3.8 –2.2

— Not available.a. Figures for the Republic of Korea are from Q4 to Q4. Figures for Thailand are calculations by World Bank staff based on theLabor Force Survey, National Statistical Office, and surveys of national employment for February and August.b. Figures for Indonesia are for August 1998 and average 1997. Figures for the Republic of Korea are seasonally adjusted. Figuresfor Thailand are calculations by World Bank staff based on the Labor Force Survey, National Statistical Office, and average wages(excluding fringe benefits) of February and August surveys.c. Urban areas.Source: Employment – Islam and others 1999; Mansor and others 1999; Shin 1999; World Bank staff calculations. Real wages –Datastream; Islam and others 1999.

58


the prices of other commodities. In rural ar-eas higher food prices spur agricultural pro-duction, and the impact on poverty will dependon the strength of the link between agricul-tural production and the poor. Insofar as smallfarmers benefit, the link to poverty may bestrong. But insofar as export crop productionis concentrated among large farmers, the im-pact depends on whether the demand for ag-ricultural workers increases sufficiently tooffset the growing supply of rural labor.

In Indonesia around 2.5 million workers,or 3 percent of the total work force, were dis-placed by the crisis in the first year. Job lossesoccurred in all sectors of the economy exceptagriculture and the small transportation andcommunication sectors. The manufacturingsector accounted for nearly half of all joblosses, followed by construction. Losses weresomewhat smaller in the mining, trade, andservice sectors. About three-quarters of thejobs lost in these sectors were in rural areas.In urban areas many workers displaced fromthe manufacturing and construction sectors en-tered the trade and other service sectors. Ur-ban employment actually grew from 29.4 to30.3 million persons (Islam and others 1999).In contrast, displaced workers in rural areashad fewer opportunities, and many wereforced to take up agricultural employment.16

Agricultural employment rose in Indonesia in1998 despite severe drought conditions insome areas. While labor reallocation wasgreater in rural than in urban areas, the in-crease in poverty was greater in urban areasdue to the incidence (although limited) ofunemployment, the greater decline in wagesin manufacturing and construction, and thefall in informal sector incomes as more crowd-ing occurred.

In Thailand the crisis greatly affected theflow of labor between urban and rural areas.The principal reason for the increase in pov-erty in the rural areas, particularly in the north-east, was the integration of the rural andBangkok labor markets through migration.The crisis dramatically curtailed the regularflow of workers to Bangkok, particularly from

the northeast, increasing the rural labor sup-ply beyond what it would otherwise have been.The number of recent migrants to Bangkok(those arriving in the past year) had dropped50 percent by February 1998 from the levelsof a year earlier, and the share coming fromthe northeast fell from 68 percent to 38 per-cent. The types of workers who moved toBangkok also changed dramatically. In Feb-ruary 1997, 25 percent of all migrants toBangkok had less than an elementary educa-tion and only 28 percent had a secondary orhigher education. One year later these num-bers were 13.5 percent and over 41 percentrespectively. These results indicate that un-skilled workers stopped going to Bangkok,whereas those with higher skills continued tomigrate. In Indonesia there is evidence of sig-nificant return urban-rural migration. Around1 million urban workers entered the agricul-tural sector, although their families stayed inurban areas.

Other labor market developments. Laborforce mobility and fewer work opportunitieswere accompanied by other labor market de-velopments. First, hours worked per week fellas workers crowded into the urban informaland rural sectors. In Indonesia, the share ofemployees working fewer than 35 hours perweek increased from 30.6 percent in 1997 to34.3 percent in 1998, with the trend towardshorter hours greater in urban areas. In Ko-rea, average hours dropped from 46.6 hoursper week to 46.0, with the shorter workingweek most prevalent in manufacturing. Sec-ond, the composition of employment changed,shifting away from wage employment. In In-donesia, the proportion of workers outsidewage employment rose from 55.1 percent in1997 to 58.9 percent in 1998. Third, in Ko-rea at least, the number of highly skilled work-ers rose as a share of employment. Forinstance, the share of managers and profes-sionals had increased from 17.1 percent at theend of 1997 to 21.1 percent by the end ofMarch 1999.

With the exception of Korea, where theagricultural sector is much smaller than in the

59

other East Asian economies and total employ-ment decreased substantially, the increase inunemployment in 1998 was not particularlylarge (figure 2.2). In Korea, the unemploymentrate peaked at 8.7 percent in February 1999,an increase of 6.4 percentage points over thelow of 2.3 percent in June 1997.17 But the ratehad fallen to 6.2 percent by June of the sameyear. Underutilization of labor was greater thanthe data on unemployment indicated, as for-mal sector working hours fell in all countriesduring the recession. In Thailand total unem-ployment increased by 2.5 percentage pointsduring the crisis and remained high during thefirst half of 1999. In Malaysia open unem-ployment increased by much less than ex-pected, rising from 2.7 to 3.2 percent. It peakedat 4.5 percent in March 1999, both becauseproductivity-based wages allowed real wagecuts and because migrant workers employedin construction, the hardest-hit sector, left thecountry.

It is still too early to assess how the crisisaffected the incidence of unemployment in dif-ferent groups. In Korea unemployment seemsto have risen more among men than amongwomen, possibly because the female partici-

pation rate dropped as the crisis intensified.The number of regular female employees fellby around 20 percent between October 1997and October 1998. Layoffs in the formal sec-tor initially raised unemployment among olderage groups, but unemployment among theyoung is undoubtedly rising in the absence ofjob creation. The less educated and less skilledwere the hardest hit. For those with no highschool diploma, unemployment increased from1.2 percent in June 1997 to 5.8 percent in June1998, and for those with high school diplo-mas it climbed from 2.8 percent to 8.4 per-cent (Na and Moon 1999).

Government safety nets andpoverty alleviationRaising transfers can offset increases in incomepoverty caused by declines in labor demand.For this reason some governments have triedto strengthen safety nets, or income transferprograms. The success of these efforts dependson several factors: the existence of well-func-tioning programs, the institutional and deliv-ery capacity of central and local agencies, thesize of budget allocations and the severity offiscal constraints, and the political economy


Figure 2.2 Unemployment in East Asia during the crisis

8

Percent

7

6

5

4

3

2

1

0

1.6

4.1

Thailand

2.6

7.4

Republic of Korea

4.75.5

Indonesia

2.73.2

Malaysia

Note: Figures for the Republic of Korea are for urban unemployment, and those for 1998 are from the second half of the year. Figures for Indonesia are from August of each year. Figures for Thailand reflect unemployment as a percentage of the current labor force and are an average of February and August figures for each year.Source: Shin 1999; Islam and others 1999; Malaysian Labor Force Survey; World Bank 1999d.

1997

1998

60


affecting redistributive and poverty alleviationefforts.

It is still too early to assess fully the im-pact of efforts to provide safety nets duringthe crisis. However, the available evidencesuggests that even though levels of publicspending on safety nets increased significantly,the impact on poverty was limited. An analy-sis of country experiences suggests that thislimited effect can be traced to a range of fac-tors, including response lags, institutionalproblems, and low levels of spending relativeto the scale of poverty.

The governments of East Asia generallydid increase the budgetary share of incometransfers significantly in response to the cri-sis. However, spending as a share of nationalincome remains low by international standardsand has increased in only two countries(Klugman 1999) (figure 2.3). Korea had thelargest proportionate increase, with spendingon safety nets rising from zero to 5 percent ofthe budget, followed by Indonesia, where thebudgetary share rose from zero to 3.6 percent.In Malaysia the safety net as a share of gov-ernment expenditure held steady at a low 0.16

Figure 2.3 Poverty and spending on social safety nets in East Asia, 1996–98

2.5

2.0

1.5

1.0

0.5

0.0

24

19

14

9

4

-11996

Percent Percent

Percent Percent Percent Percent

Percent Percent

Indonesia

1998 1996 1998

1996 1998 1996 1998

6 20

15

10

5

5

4

3

2

1

0

Republicof Korea

1.0 9

8

7

6

5

4

3

2

1

0

0.9

0.8

0.7

0.6

0.5

0.4

0.3

0.2

0.1

0.0

Malaysia

3.0

2.5

2.0

1.5

1.0

0.5

0.0

14

12

10

8

6

13

11

9

7

5

Thailand

Expenditures (left axis) Headcount index (right axis)

Source: World Bank staff estimates.

61

percent during the period. Spending on safetynets as a share of national income rose quitesteeply in Korea and Indonesia, though expen-ditures were low to begin with. These trendscontrast with those observed in Europe andCentral Asia, where expenditures on safety netsdeclined significantly across all countries(Milanovic 1998).

Korea. The central response of the Ko-rean government was to introduce a publicworks scheme that grew enormously duringthe crisis period, rising to 200,000 participantsin January 1999 and 410,000 in mid-1999.The scheme paid wages lower than the pre-vailing wage for unskilled workers in order toattract only those truly in need of employment.The scheme also was supposed to guarantee ajob for all who wanted one, although therewas significant excess demand for the num-ber of places available by late 1998. (Therewere 700,000 applicants in January 1999.)

The budgetary share of safety nets in Ko-rea did increase during the crisis, but recentanalyses have shown that the incremental bud-get and program coverage in 1998 were inad-equate to meet the country’s needs. Safety netprograms covered only 7 percent of the “new”poor in 1998. Overall coverage of the poor(old and new) dropped from almost one-thirdprior to the crisis to about 17 percent in 1998,and it is expected to fall further (to 16 per-cent) during 1999 (Subbarao 1999).18 Thereis evidence that women have been excludedfrom public works, and there are accusationsof mismanagement and a lack of useful out-put. Still, what has been done reflects twoimportant lessons: keeping the wage sharehigh—around 70 percent—creates more jobsper won spent; and diversifying the jobs menuto include more than engineering works (forinstance, work in libraries) increases job op-portunities. Some other East Asian govern-ments are just beginning to learn these lessons.

Thailand. In Thailand government safetynets did not fill much of the gap left by infor-mal transfers at least until 1998. Overall safetynet expenditures increased during the crisis,especially during 1999, though some income

transfers appear to have been procyclical—thatis, contracting with the economy rather thanexpanding. This situation was the result of con-flicting pressures and the government’s reluc-tance to undermine the informal safety net.For example, social pensions and family al-lowances in rural Thailand appear to be welltargeted, but they are underfunded (Prescott1999). The benefit value is less than one-thirdof subsistence requirements and reaches onlyone-third of the target group. Further, the realvalue of the benefit transfers in Thailand hasbeen falling over time because of the govern-ment’s failure to adjust them for inflation. Oneprogram that expanded in response to the cri-sis in Thailand was the Ministry of Health’sprogram providing low-income groups withaccess to public health services. Under thestimulus package, job creation for the poorunemployed was a priority.

Indonesia. The major new safety net pro-grams introduced in Indonesia as a result ofthe crisis were a rice distribution scheme(known as OPK) and a public works scheme(Padat Karya). Early evidence suggests thattheir coverage of the poor and their impacton poverty have been limited. The OPK makes10 kilograms of medium-grade rice availableto selected households every month at subsi-dized prices. On average, this amount repre-sents less than 30 percent of the income of asingle individual living at the poverty line andless than 6 percent of the income of a house-hold of five. OPK uses an indicator-based tar-geting system with minimum standards forfood intake, housing, clothing, and medicalexpenditures.

Evaluations of these programs suggest thatjust over one-third of all poor households haveparticipated in the Indonesian public worksprogram. The leakage of benefits to the non-poor has been significant, however. In Jakarta,9 percent of the poor households worked onceon the scheme, compared with 30 percent ofthe middle income households. In Medan only5 out of over 400 poor households partici-pated in Padat Karya, because the contractorused his own workers.


62


Other experiences. Declines in publicspending and a failure to reach the poor haveundermined safety net programs in Russia andCentral Asia. Russia provides a striking ex-ample of the failure of safety net programs toalleviate poverty. Spending on social assistancedeclined throughout the transition, amount-ing to only 4 percent of the poverty gap in1997, while the incidence of poverty more thantripled. This shortfall can be attributed in partto the concurrent collapse of the tax revenuesystem (UNICEF 1998), but it is also the re-sult of an apparent failure to identify andimplement programs that reach the poorand that have sufficient political and electoralsupport.

In addition, the safety net in Russia is notwell targeted. Most Russian households re-ceive some type of government transfer, but asignificant proportion of the very poor (al-most 3 out of 10) and of the poor (1 out of 5)receive no benefits. At the same time almostfour out of five households that are not poordo receive public transfers (Foley andKlugman 1997). Even so, the decline in bud-get allocations for public transfers, coupledwith widespread delays in payments, hasmeant a clear weakening in the impact oftransfers on poverty over time. The reductionin the poverty headcount attributable to pub-lic transfers fell from 29 to 24 percentagepoints between 1994 and 1996 (Klugman andKolev 1999).

Beyond current income effectsof the East Asian crisis

Standard poverty measures based on in-comes and household expenditures captureonly some aspects of the social impact anddistress crises cause. Households use variousmechanisms to cope with shocks from crises.These responses may help mitigate the imme-diate impact, but they may also have impor-tant implications for future poverty andvulnerability to shocks. Crises also create pres-sures on governments for fiscal austerity whichmay exacerbate the negative social impacts,

but appropriate fiscal policy may also help al-leviate some of these effects.

Behavioral responses and assetaccumulation during crisesFinancial crises affect not only current incomesbut also the value of household assets. Infla-tion, for instance, has been found to be one ofthe most significant determining factors ofpoverty (Datt and Ravallion 1997; Agénor1998; Easterly and Fischer 1999). It erodesthe value of fixed-denomination assets suchas money, which is the primary asset of thepoor and near-poor. These groups have littlescope for hedging.

As they do to labor market shocks, house-holds respond in variety of ways to the in-come, wealth, and relative price effects of acrisis. These responses include consumptionsmoothing, changing the composition of theconsumption basket, selling existing physicalassets, and acquiring fewer new ones.

Savings behavior during the East Asiancrisis. Changes in household savings patternsduring the crisis varied significantly acrosscountries. The savings rate changed little inThailand. In Indonesia the savings rate de-clined sharply, falling about 8 percentagepoints of GDP (figure 2.4). The savings re-sponse helped reduce the impact of the crisis

Figure 2.4 Change in gross domesticsavings, 1997–98

6

Percentage of GDP

4

2

0

–2

–4

–6

–8

ThailandIndonesiaMalaysia Republic

of Korea

Source: World Bank staff calculations.

63

on consumption among poor households, butthe reductions in capital accumulation reflectthe severity of the crisis.

In Korea and Malaysia the decline in percapita consumption was much greater than thedecline in per capita GDP between 1997 and1998 (table 2.1). In Korea the savings rate rose,but the increase reflects primarily the behav-ior of high-income groups.19 Total gross do-mestic savings increased by more than 4percentage points of GDP. Private savings in-creased at an even greater rate, especially inlight of the increase in the government deficit.Savings declined from 16.3 percent to 11.6percent among the 20 percent of householdswith the lowest incomes, however (Kakwaniand Prescott 1999). Because of the decline inthe consumption ratio and the varied effectson income distribution in Korea, the incidenceof income poverty increased much less thanpoverty based on consumption (discussedabove, table 2.1), rising from 2.6 percent in1997 to 7.3 percent in 1998.

Changes in asset holdings and the com-position of the consumption basket. Informa-tion about changes in asset holdings in thecrisis countries is limited. Some evidence is

available for Indonesia showing that peoplein the most affected regions, such as Java, soldsome of their assets (Poppele, Sumarto, andPritchett 1999). But more complete evidenceis available on changes in the consumptionbasket. Rising food prices are especially im-portant in determining changes in the compo-sition of consumption, because higher pricesreduce the real incomes of households withrelatively high food expenditures—mainly theurban poor. Food prices rose relative to othercommodities in all countries after the crisis(figure 2.5). However, the effect was smallexcept in Indonesia, where relative food pricesrose by 40 percent between mid-1997 and mid-1998. The effect is reflected in the dramaticchanges in the composition of expenditures:the share of staple foods increased from 23.1to 31.7 percent, while that of meats and non-food items (including health and education)declined (Poppele, Sumarto, and Pritchett1999).

Between 1996 and 1998 households inThailand, particularly those with low incomes,increased essential real expenditures such asfood, fuel, medical supplies, shelter, and edu-cation but reduced other expenditures (World


Figure 2.5 Relative price of foods during the crisisIndex of food prices relative to total consumer price index (1997=100)

140

130

120

110

100

90

80Q3

1999Q2

1999Q1

1999Q4

1998Q3

1998Q2

1998Q1

1998Q4

1997Q3

1997Q2

1997Q1

1997Q4

1996Q3

1996Q2

1996Q1

1996Q4

1995Q3

1995Q2

1995Q1

1995

Source: Datastream.

Indonesia

Malaysia

Thailand

Republic of Korea

64


Bank 1999d). Korean households had a dif-ferent response. The shares of food, clothing,and furniture in total expenditures actually de-clined, but spending for education and healthincreased (Kakwani and Prescott 1999).

Fiscal austerity and household demandfor health and educationThe East Asian economies eventually widenedtheir fiscal deficit targets to counter the re-cessionary effects of the crisis. Yet real gov-ernment consumption expenditures fell in allcountries except Thailand in 1998. In Indone-sia the decline outpaced the fall in GDP. As aproportion of GDP, health expenditures re-mained relatively unchanged during the last halfof the 1990s, including the first year of the re-cession (table 2.3). Education expenditures fellrelative to GDP when the crisis struck in Ma-laysia and Korea, but rose in Thailand.

Changes in public spending on educationand health affects both the availability of theseservices and, because the services may becomemore expensive, households’ decisions to usethem. The full impact of the crisis on publicservices remains unclear because of the differ-ences in fiscal policy responses and the lim-ited information available on changes in thecomposition of public expenditures. Some dataare available for Korea, Thailand, and Indo-nesia, however.

Korea. In 1998 households spent less onitems such as clothing and recreation butmaintained spending levels on education andhealth. The rates of decline for education (9percent) and health (14 percent) expenditureswere lower than the decline in total expendi-tures (18 percent) (Kakwani and Prescott1999).

Thailand. Families and government pro-grams acted to cushion the impact of the cri-sis on education and health (World Bank1999c). So far the crisis has not had a nega-tive effect on education. In the year followingthe onset of the crisis, total gross enrollmentsin primary and general upper secondaryschools increased from 74.8 percent to 75.5percent. The dropout ratio—children of schoolage not attending school—continued to declinebetween 1996 and 1998. Families employeda variety of strategies to keep their children inschool. First, households spent less on nones-sential consumables such as alcohol, tobacco,clothing, footwear, household goods, transpor-tation, and communications. Second, familiesused savings or borrowed from informalsources to finance education, so that realspending on education rose for both the poorand nonpoor. Third, for nontertiary educationfamilies shifted their children from private topublic schools as the relative cost of attend-ing private schools increased. Finally, families

Table 2.3 Public spending on health and education(percentage of GDP)

1994–95 1995–96 1996–97 1997–98 1998–99

HealthIndonesia 0.7 0.6 0.6 0.6 0.6Korea, Rep. of 0.5 0.5 0.5 0.6 0.6Malaysia 1.3 1.3 1.4 1.4 1.3Thailand 1.1 1.1 1.2 1.4 1.3

EducationIndonesia 1.4 1.2 1.4 0.7 0.7Korea, Rep. of 5.0 5.0 5.1 4.3 4.0Malaysia 5.3 4.9 4.0 4.7 4.3Thailand 3.4 3.3 3.1 3.4 4.2

Note: Public expenditures include national and local government.Source: Baptist 1999.

65

made greater use of government scholarshipand loan programs.

Government policies and programs alsosupported continued investment in education.Real expenditures on education remained con-stant between 1997 and 1998, and the shareof education increased in total expenditures.A number of programs and measures were in-troduced to protect educational opportunitiesfor the vulnerable. These included allowingparents to pay tuition fees in installments, per-mitting schools to waive tuition fees on a case-by-case basis, introducing scholarships,expanding the education loan program, en-couraging private schools to extend paymentdeadlines, and providing vouchers to privateschool children (in the Bangkok metropolitanarea). These achievements are remarkable butmay be difficult to sustain, as the recoveryremains weak and the effects of the crisis con-tinue to be severe. Households, particularlythe poor, may be less able to shift more re-sources to education and sustain higher debts.

There is also no evidence of a negativeeffect on national health outcomes. For in-stance, the number of reported cases of mal-nutrition continued on a downward trend in1998. Households’ real expenditures on bothprivate and public health services declined sig-nificantly. Out-of-pocket real expenditures onmedical and institutional care were 36 percentlower in 1998 than in 1996, whereas spend-ing on self-medication increased by 12 per-cent. The decline in expenditures was lowerfor the poor, who undoubtedly tried to sus-tain essential health expenditures and who alsobenefited from public health services. Thegovernment maintained its level of investmentin health, with real expenditures on healthdown by 5 percent in 1998 from 1997 levelsbut still 11 percent higher than they were in1996. The decline mainly affected investmentexpenditures. The government enlarged itshealth safety net by increasing the coverageof public health insurance. Use of public healthservices increased between 1996 and 1998,with the number of outpatient visits rising by22 percent.

Indonesia. The severity of the shock andfalling living standards led to a decline inschool enrollment rates (Frankenberg, Tho-mas, and Beegle 1999). This decline was muchlarger at the secondary level—some 4 to 5 per-centage points of enrollment rates20 —than atthe primary level. Consistent with the urbanbias in the effects of the crisis on incomes, thedecline in school enrollment was largest in ur-ban areas, particularly Jakarta. The popula-tion with the lowest per capita expenditureshad the highest rates of decline in school en-rollment: more than 5 percentage points forthe 13–19 age group for the lowest twoquartiles, and more than 6 percentage pointsfor the 7–12 age group for the lowest quartile.Because they had to increase food expendi-tures during the crisis, families had a difficulttime maintaining expenditures on education,and its share in total expenditures declinedfrom 3.5 percent in 1997 to 2.9 percent in1998.

The effects of the crisis on health werecomplex and heterogeneous but clearly nega-tive. The share of household expenditures onhealth declined from 1.4 percent in 1997 to 1percent in 1998. The use of public health ser-vices following the crisis declined by 1.8 per-centage points (from 7.2 to 5.4 percent) foradults and by more than 7.1 percentage pointsfor children. The proportion of visits made totraditional practitioners nearly doubled. In1997 nearly one-half (46.7 percent) of all chil-dren under the age of five had visited a com-munity health post in the month before thesurvey, but this rate declined to about one-quarter (27.7 percent) in 1998.

Fostering sustained growth andreducing the social costs ofvolatility and crises

Financial crises have large social costs andtend to retard or even reverse gains in pov-erty reduction for significant periods of time,even in the most successful countries. Policiesand institutions that reduce these risks, helpprevent financial crises, and minimize their ef-


66


fects when they do occur can help smooth thegrowth process and maximize the positive ef-fects of growth on poverty alleviation (WorldBank 1999a; Ferreira, Prennushi, andRavallion 1999; Lustig 1999). Realizing thelong-term benefits of openness, reducing pov-erty over the long run and avoiding tempo-rary setbacks in poverty reduction requiresappropriate national and international poli-cies. These policies must minimize the risksof external volatility and improve the capac-ity to manage it at both levels. Safety nets arepart of any broad-based strategy for limitingthe impact of crises and negative shocks onpoverty.

Preventing crises. Avoiding crises is clearlythe most effective way to achieve stable andsustainable growth. Macroeconomic and fi-nancial policies that avoid profligate fiscal andmonetary policies, seriously overvalued ex-change rates, and unsustainable current ac-count deficits are necessary to prevent crises(World Bank 1999a; Lustig 1999). More flex-ible exchange rates, greater reliance on fiscalpolicy, and better and tighter domestic finan-cial regulation are often needed to reduce ex-cessive capital inflows and domestic lendingbooms (World Bank 1999a). Financial sectorliberalization must proceed carefully and instep with the capacity of countries to enforcetighter regulation and supervision. Efforts toimprove prudential safeguards and bankingoperations need to be accelerated in most de-veloping countries. The opening of the capi-tal account, particularly to the more volatilecapital flows, needs to be carefully orches-trated to match the capacity of countries tomanage risk (World Bank 1999a; Stiglitz andBhattacharya 1999). Other important policiesmust focus on improving corporate gover-nance, increasing transparency, and buildingsupportive institutions.

Socially sensitive crisis management. Mac-roeconomic policy responses to crises need tobe designed to minimize the social costs andavoid large declines in aggregate demand andemployment (World Bank 1999a; Lustig1999). Monetary policy should always avoid

high inflation as well as excessive increases ininterest rates, both of which worsen any con-traction in aggregate demand. Policy responsesaimed at reducing the social impact of crisesshould make fiscal policy countercyclical inorder to reduce the extent of contraction.However, developing countries typically haveprocyclical fiscal policies that tend to aggra-vate the impact of downturns (Easterly, Islam,and Stiglitz 1999). This phenomenon is theresult of the high sensitivity of tax receipts tochanges in incomes, underdeveloped domes-tic financial markets, limited access to foreigncapital markets, and the risk of losing inves-tor confidence. These factors make pursuingcountercyclical fiscal policies difficult. Estab-lishing effective countercyclical fiscal policiesrequires that public finances be managed wellduring good times, so that there is room forexpansionary policies during negative shocks.The adequate well-institutionalized use of sta-bilization funds may also be helpful (Lustig1999). But even East Asian countries that hadresponsible fiscal policies before the crisis havefound it difficult to achieve the looser fiscalobjectives.

Fiscal adjustments should also protect theexpenditures that are most important for thepoor, such as employment and human devel-opment programs and targeted subsidies.Where crises result in high unemployment, thefiscal stimulus needs to be directed to labor-intensive activities.

Managing volatility. In the long run de-veloping countries stand to make gains ingrowth and to reduce poverty through opentrade policies and integration into the worldeconomy. But external shocks, such as capitalflow reversals and collapses in commodityprices, may cause temporary increases in pov-erty that are difficult to reverse. Developingcountries must develop the capacity to man-age increased external and internal volatilitythrough better economic management, morerobust institutions for managing risks (suchas banks), and improved safety nets.

Safety nets. Before a crisis, safety nets canspur productivity and growth by providing the

67

insurance necessary for households to makerisky choices with higher potential returns.Safety nets also help ensure that crises do nothalt development. They help maintain essen-tial household investments in education andhealth and eliminate the need for the poor todivest themselves of physical capital. Settingup safety nets during times of economic growthmay be the only effective way to protect thepoor during crises (Ferreira, Prennushi, andRavallion 1999). The need for redistributioninevitably increases during crises, even if mostpeople’s incomes fall. Although the newly poorgenerate most of the increase in need, policiesneed not distinguish between the old and newpoor.

As a response to a temporary shock, in-creases in spending on social safety nets areideally financed over time, both past and fu-ture. Despite this logic (or indeed, because ofit) countries usually have to adopt policies withthe lowest budgetary costs. Targeting benefitsis one desirable means of keeping costs low.Self-selection mechanisms such as public worksare an important means of reducing the bud-getary costs of redistributive policies and ofestablishing institutions to deliver transfers.Establishing well-functioning institutions maybe one of the largest setup costs countries in-cur in response to deep crises. Ideally, ofcourse, such investments precede a crisis. Re-cent international experience confirms thatopen and transparent institutions operating ina noncorrupt way are as important to the es-tablishment of safety nets as they are to otherareas of public action.

The possibility of making a guarantee oflow-wage work on community-initiatedprojects the central element of a safety net maybe limited. If there is an institutional basis forsignificantly expanding workfare programs,with central and local agencies operating in atransparent and noncorrupt fashion, then theseschemes could play a significant role in allevi-ating poverty. The contribution of publicworks to poverty reduction tends to be largerin countries such as Korea, where the socialcosts of crises have primarily taken the form


of high unemployment. Public works are moreeffective at reducing poverty in these coun-tries because the opportunity costs of partici-pating in public works projects are lower forthe jobless than for the working poor. Publicworks are an important and useful option forreducing poverty during crises, especially indeveloping countries in need of infrastructureinvestments. However, they are best imple-mented alongside other programs for both theable-bodied population and those unable towork.

Notes1. Easterly and others (1993); Lutz (1994);

Mendoza (1994); Hausmann and Gavin (1995);Spatafora and Warner (1995); Deaton and Miller(1995); Collier and Gunning (1996); Guillaumont,Jeanneney, and Brun (1999); Lundberg and Squire(1999). The evidence concerns GDP and output growthand does not account for the direct real income growththat results from increased purchasing power on in-ternational markets.

2. Deaton and Miller (1995) suggest that in coun-tries that are marginal producers of commodities,where labor accounts for a significant share of costs,poverty itself may explain why real commodity pricesdo not increase in the long run. These prices cannotrise as long as there are unlimited supplies of labor atthe subsistence wage. Long-term marginal costs areset by poverty in tropical countries, and commoditytrade cannot contribute to reducing it. Incentives fortechnical progress are weak, and even when progressoccurs it tends to make prices fall while real wagesremain at the subsistence level.

3. The headcount index is the proportion of in-dividuals in the population whose income or consump-tion expenditures fall below the poverty line.

4. The elasticity for other measures of poverty,such as the poverty gap, are usually higher. See Liptonand Ravallion (1995).

5. For a lower-end Gini index of 0.25, the elas-ticity would be –3.3. It is –1.82 for a higher index of0.59 (Ravallion 1997).

6. Aizenman and Marion (1999) find similar re-sults.

7. This result holds under risk neutrality, but italso requires some degree of imperfect competition(Caballero 1991).

8. The authors also find that foreign aid and goodpolicies can offset this vulnerability.

68


9. This finding is contrary to earlier findings byMcBean (1966) that export instability has no effecton growth.

10. The asymmetry notion is also supported byfindings about the relationship between the headcountpoverty index and mean country income (Ravallion1997). More generally the poverty headcount will,for a given absolute decrease in income, increase morefor a poor than for a wealthy country (Milanovic1998).

11. Poverty figures for the rural areas are notavailable for Korea.

12. The figures used for 1997 are estimates basedon the declining poverty trends before the crisis.

13. Rural areas gained if they were not primarilyrice producing and were not affected by drought (pricesremained low in these areas until August 1998). Someareas also had natural disasters during the crisis. Thedrought of 1997–98 was not as bad on Java as hadbeen feared, but it did hit hard in the Eastern Islands,on the west coast of Sumatra, and in parts of Sulawesi.East Kalimantan suffered an ecological disaster whenthe drought interacted with wildfires.

14. The measured Gini coefficients are eitherbased on consumption expenditures or incomes. Inthe latter case they also may be biased, as they donot reflect adequately all incomes and changes of as-set values.

15. This result holds true even in industrial coun-tries. See Farber (1993), and Layard, Nickell, andJackman (1994).

16. There was a shift from rural nonagriculturalto agricultural jobs. Rural employment increased from56.05 to 57.37 million, while agricultural employmentincreased from 34.8 to 39. 4 million.

17. Among the crisis-hit countries, Korea is theonly one that had an unemployment insurance scheme,and had extended its potential coverage.

18. Budgetary allocations to support the unem-ployed in Korea were to double in April 1999 to in-crease the program’s coverage of the poor.

19. The savings response in Korea and Malay-sia, which have higher incomes than the other crisis-affected countries, may reflect greater wealth effectsfrom the crisis.

20. The average enrollment rate for the 13–19age group was around 60 percent in 1997.

ReferencesAgénor, Pierre-Richard. 1998. “Stabilization Policies,

Poverty and the Labor Market: Analytical Issuesand Empirical Evidence.” EDI, World Bank,Washington, D.C. December.

Aizenman, Joshua, and Nancy Marion. 1999. “Vola-tility and Investment: Interpreting the Evidencefrom Developing Countries.” Economica 66(262): 157–79. London. May.

Baptist, Jacqueline. 1999. “Public Expenditures onHealth and Education during the Crisis and Pre-Crisis Period in East Asia.” World Bank. August.

Bourguigon, Francois, Jaime de Melo, and AkikoSuwa. 1991. “Distributional Effects of Adjust-ment Policies: Simulations for Archetype Econo-mies in Africa and Latin America.” World BankEconomic Review 5: 339–66. May.

Caballero, Ricardo J. 1991. “On the Sign of the In-vestment-Uncertainty Relationship.” AmericanEconomic Review 81 (1): 279–88. March.

Collier, Paul, and Jan Willem Gunning. 1996. “PolicyTowards Commodity Shocks in DevelopingCountries.” IMF Working Paper 96/84. Interna-tional Monetary Fund, Research Department,Washington, D.C. August.

Cuddington, John. 1988. “Fiscal Policy in Commod-ity-Exporting LDCs.” Policy, Planning, and Re-search Working Paper (WPS) 33. World Bank,Washington, D.C. July.

Datt, Gaurav, and Martin Ravallion. 1997. “Macro-economic Crises and Poverty Monitoring: A CaseStudy of India.” Review of Development Eco-nomics 1 (2): 135–52.

Deaton, Angus S., and Ronald I. Miller. 1995. “Inter-national Commodity Prices, Macroeconomic Per-formance, and Politics in Sub-Saharan Africa.”Princeton Studies in International Finance 79.Department of Economics, International FinanceSection. Princeton, N.J.: Princeton University Press.

Dehn, Jan, and Christopher L. Gilbert. 1999. “Com-modity Price Uncertainty and Economic Growthand Poverty.” Processed. August.

De Janvry, Alain, and Elisabeth Sadoulet. 1998.Growth, Poverty and Inequality in Latin America:a Causal Analysis, 1970–94. Univeristy of Cali-fornia at Berkeley. September. (Presented at theInter-American Development Bank Conference onSocial Protection and Poverty. February 1999.)

Easterly, William, Roumeen Islam, and Joseph E.Stiglitz. 1999. “Shaken and Stirred: Volatility andMacroeconomic Paradigms for Rich and PoorCountries.” Michael Bruno Lecture, XIIth WorldCongress of the International Economic Associa-tion, Buenos Aires, August 27.

Easterly, William, and Art Kraay. 1999. “Small States,Small Problems.” Policy Working Paper 2139.World Bank, Washington D.C.

Easterly, William, and Stanley Fischer. 1999. “Infla-tion and the Poor.” Presented at the Annual Bank

69


Conference on Development Economics. WorldBank, Washington, D.C. April 28–30.

Easterly, William, Michael Kremer, Lant Pritchett, andLawrence H. Summers. 1993. “Good Policy orGood Luck? Country Growth Performance andTemporary Shocks.” Journal of Monetary Eco-nomics 32: 459–83.

Fallon, Peter, and Robert E. B. Lucas. 1999. “Losersand Winners During Economic Crises.” WorldBank, Washington, D.C. May.

Farber, Henry S. 1993. “ The Incidence and Costs ofJob Loss: 1982–91.” Brookings Papers on Eco-nomic Activity 1: 73–132.

Ferreira, Francisco H. G., Giovanna Prennushi, andMartin Ravallion. 1999. “Protecting the Poorfrom Macroeconomic Shocks: An Agenda forAction in a Crisis and Beyond.” Policy ResearchWorking Paper 2160. World Bank, Washington,D.C.

Foley, Mark, and Jeni Klugman. 1997. “The Impactof Social Support: Errors of Leakage and Exclu-sion. ” In Jeni Klugman, ed. Poverty in Russia.Public Policy and Private Responses. EDI Devel-opment Studies, Economic Development Institute.World Bank, Washington D.C.

Fox, M. Louise, Edward Amadeo, José MarcioCamargo. 1994. “Brazil.” In Susan Horton, RaviKanbur, and Dipak Mazumdar, eds. Labor Mar-kets in an Era of Adjustment Vol. 2. EDI Devel-opment Studies, World Bank, Washington, D.C.

Frankel, Jeffrey A., and David Romer. 1999. “DoesTrade Cause Growth?” American Economic Re-view 89(3): 379–99.

Frankenberg, Elizabeth, Duncan Thomas, andKathleen Beegle. 1999. “The Real Costs ofIndonesia’s Economic Crisis: Preliminary Findingsfrom the Indonesia Family Life Surveys.” Laborand Population Program Working Paper Series99–04, Rand, Santa Monica, Calif. March.

Gelb, Alan H. and Associates. 1988. Oil Windfalls:Blessing or Curse? New York: Oxford Univer-sity Press. (A World Bank Research Publication.)

Guillaumont, Patrick, Sylviane Guillaumont Jeanneney,and Jean-François Brun. 1999. “How InstabilityLowers African Growth.” Journal of AfricanEconomies 8 (1): 87–107.

Hausmann, Ricardo, and Michael Gavin. 1995. “Over-coming Volatility in Latin America.” InternationalMonetary Fund Seminar Series (International);No [1995-34]:1–86. August.

Islam, Rizwanul, and others. 1999. “Indonesia Coun-try Paper.” Paper presented at Seminar on Eco-nomic Crisis, Tokyo, 13–15 October 1999. ILO.Geneva. August.

Kakwani, Nanak. 1998. “Impact of Economic Crisison Employment, Unemployment and Real In-come.” Development Evaluation Division. Na-tional Economic and Social Development Board,Bangkok.

———. 1999. “Poverty and Inequality During the Eco-nomic Crisis in Thailand.” Office of the NationalEconomic and Social Development Board,Bangkok. NESB Newsletter. Volume 3, No. 1.

Kakwani, Nanak, and Nicholas Prescott. 1999. “Impactof Economic Crisis on Poverty and Inequality inKorea.” Processed. World Bank, Washington, D.C.

Klugman, Jeni. 1999. Social Safety Nets and Crises.PREMPOV. World Bank, Washington, D.C.

Klugman, Jeni, and Alexandre Kolev. 1999. “TheWelfare Repercussions of Single Parenthood inRussia in Transition. ” In Jeni Klugman andAlbert Motivans, eds. Family Structure in the NewRussia. London: Macmillan, forthcoming.

Layard, P. Richard G., Stephen Nickell, and RichardJackman. 1994. The Unemployment Crisis. Ox-ford; New York: Oxford University Press.

Lipton, Michael, and Martin Ravallion. 1995. “Pov-erty and Policy.” Chapter 41 in J. Behrman andT. N. Srinivasan, eds. Handbook of DevelopmentEconomics Vol. III, Elsevier Science.

Londoño, Juan Luis, and Miguel Székely. 1997a. “Dis-tributional Surprises After a Decade of Reforms:Latin America in the Nineties.” Working Paper352. Office of the Chief Economist. Inter-Ameri-can Development Bank, Washington, D.C. August.

———. 1997b. “Persistent Poverty and Excess Inequal-ity: Latin America 1970–1995.” Working Paper357. Office of the Chief Economist. Inter-Ameri-can Development Bank, Washington, D.C. Sep-tember.

Lundberg, Mattias, and Lyn Squire. 1999. “Growthand Inequality: Extracting the Lessons for Policy-makers.” World Bank, Washington, D.C. May.

Lustig, Nora. 1998. Mexico: the Remaking of anEconomy, Second edition. Washington, D.C.:Brookings Institution Press.

———. 1999. “Crises and the Poor: Socially Respon-sible Macroeconomics.” Sustainable Develop-ment Department, Poverty and InequalityAdvisory Unit. Inter-American DevelopmentBank, Washington, D.C. September.

Lustig, Nora, and Ruthanne Deutsch. 1998. “The In-ter-American Development Bank and PovertyReduction: an Overview.” Sustainable Develop-ment Department, Poverty and Inequality Advi-sory Unit. Technical Study No. Pov-101-R.Inter-American Development Bank, Washington,D.C. May.

70


Lustig, Nora, and Miguel Székely. 1998. “EconomicTrends, Poverty and Inequality in Mexico.” Sus-tainable Development Department, Poverty andInequality Advisory Unit. Technical Study No.Pov-103. Inter-American Development Bank,Washington, D.C. December.

Lutz, Matthias. 1994. “The Effects of Volatility in theTerms of Trade on Output Growth: New Evi-dence.” World Development 22 (12): 1959–75.

Mansor, Norma, Tan Eu Chye, Ali Boerhanordin,Fatima Said, and Saad Said. 1999. “MalaysiaCountry Paper.” Paper presented at Seminar onEconomic Crisis, Tokyo, October 13–15.

Matusz, Steven J., and David Tarr. 1999. “Adjusting toTrade Policy Reform.” Policy Research WorkingPaper 2142. World Bank, Washington, D.C. June.

McBean, Alasdair. 1966. Export Instability and Eco-nomic Development. London: Allen and Unwin.

Mendoza, Enrique G. 1994. “Terms of Trade Uncer-tainty and Economic Growth: Are Risk Indica-tors Significant in Growth Regressions?” Boardof Governors of the Federal Reserve System. In-ternational Finance Discussion Paper 491. De-cember.

———. 1995. “The Terms of Trade, the Real ExchangeRate, and Economic Fluctuations.” InternationalEconomic Review 36 (1): 101–37. February.

Milanovic, Branko. 1998. Income, Inequality, andPoverty during the Transition from Planned toMarket Economy. Regional and Sectoral StudiesSeries. Washington, D.C.: World Bank

Morley, Samuel. 1994. Poverty and Inequality in LatinAmerica: Past Evidence, Future Prospects. Over-seas Development Committee, Washington, D.C.

Na, Seong-Lin, and Hyung-Pyo Moon. 1999. “SocialImpact of Current Economic Crisis in Korea.”World Bank: World Development Report Con-sultative Meeting in Kuala Lumpur, May 10–12.

Pindyck, Robert S. 1991. “Irreversibility, Uncertaintyand Investment.” Journal of Economic Literature29 (3): 1110–48. September.

Poppele, Jessica, Sudarno Sumarto, and Lant Pritchett.1999. “Social Impact of the Indonesian Crisis:New data and Policy Implications.” World Bank,Washington, D.C. February.

Prescott, Nicholas. 1999. “Social Risk Management—Lessons from the Crisis.” Presented at the Insti-tute of Policy Studies on Asian EconomicRecovery: Policy Options for Growth and Stabil-ity, June 21–22.

Pritchett, Lant. 1998. “Patterns of Economic Growth:Hills, Plateaus, Mountains and Plains.” WorldBank Policy Research Paper 1947. Development

Research Group, World Bank, Washington, D.C.July.

Ramey, Garey, and Valerie A. Ramey. 1995. “Cross-Country Evidence on the Link Between Volatil-ity and Growth.” American Economic Review 85(5): 1138–51.

Ravallion, Martin. 1997. “Can High-Inequality De-veloping Countries Escape Absolute Poverty?”Economics Letters 56: 51–7.

Riveros, Luis A. 1994. “Chile.” In Susan Horton, RaviKanbur, and Dipak Mazumdar, eds. Labor Mar-kets in an Era of Adjustment Vol. 2, EDI Devel-opment Studies, World Bank, Washington, D.C.

Riveros, Luis A., and Carlos E. Sanchez. 1994. “Ar-gentina.” In Susan Horton, Ravi Kanbur, andDipak Mazumdar, eds. Labor Markets in an Eraof Adjustment Vol. 2, EDI Development Studies,World Bank, Washington DC.

Rodrik, Dani. 1998. “Where Did All the Growth Go?External Shocks, Social Conflict and GrowthCollapses.” NBER Working Paper 6350. NationalBureau of Economic Research, Cambridge, Mass.

Shin, Donggyun. 1999. “Statistical Overview of Cur-rent Labor Market Conditions and Trends.” Ko-rea Labor Institute. July.

Spatafo

Documents

External Shocks, Financial Crises, and Poverty in Developing ......external shocks explain 30 percent of cross-country variation in GDP volatility in Latin America. When terms of trade,