EXTERNAL SHOCKS AND POVERTY: HOW RECESSION IN EUROPE, JAPAN, AND CHINA AFFECTS THE INDONESIAN POOR

Transcription

1 External Socks And Poverty: How Recession In Europe, Japan, and Cina Affects Te Indonesian Poor 207 EXTERNAL SHOCKS AND POVERTY: HOW RECESSION IN EUROPE, JAPAN, AND CHINA AFFECTS THE INDONESIAN POOR Arief Ansory Yusuf 1 Abstract Tis paper analyzes te effect of a recession in Europe, Japan, and Cina on te poverty in Indonesia. We use te GTAP model and te INDONESIA-E3 model to examine te impact of a 2 percent GDP decline in tese tree countries on te poverty in Indonesia. Te results suggest a negative impact on Indonesia s GDP, mainly troug te trade-linkages but wit a small magnitude. Te main reason for tis finding as to do wit te low dependency of Indonesia on international trade. Te sock also sligtly increases te poverty in Indonesia wit a small magnitude. Across te ouseold types, te negative effects of tese recession goes mainly to iger income ouseolds since large part of teir incomes comes from te capital and skill-intensive sectors. Te poor ouseold types are likely to be te first to lose teir jobs in te event of tis recession, since tey are less skilled. Tese findings urge te Indonesian government to lunc employment programs to ensure te employment continuity for tese unskilled laborers in te anticipation of a global recession particularly originating from tese tree countries. Keywords: Recession, GTAP, Computable General Equilibrium, INDONESIA-E3. JEL Classification: 1 Dr. Arief Ansory Yusuf is te Director of te Center for Economics and Development Studies (CEDS) of Universitas Padjadjaran, Indonesia. He is also an Adjunct Fellow at te Australian National University and a Visiting Fellow at King s College London. arief.yusuf@fe.unpad.ac.id

2 208 Buletin Ekonomi Moneter dan Perbankan, Volume 18, Nomor 2, Oktober 2015 I. INTRODUCTION From te start of te New Order government in 1966 up until 1997, broad-based economic development in Indonesia led to an impressive four-fold increase in per capita income. As a result, poverty in te country fell dramatically. Between 1976 and 1996 te number of poor people in Indonesia declined from 54.2 to 22.5 million. However, te Asian Financial Crisis in 1997 brougt a sarp alt to tis economic miracle. Between 1997 and 1998 annual GDP growt in Indonesia fell by 13 percent and te incidence of poverty increased by 7 percent. Te events of were a painful reminder of just ow vulnerable te Indonesian economy was to external socks coming from abroad. For tis reason, te current recession in Europe raises deep concerns about te possible impact of tese events on te Indonesian economy. However, to date tere as not been any quantitative analysis of te possible impact of tese events on economic growt and poverty in Indonesia. Te purpose of tis paper is, terefore, to provide a modeling-based assessment of te possible effects of a recession in Europe, Japan and Cina - EJC countries, for sort -- on Indonesia. Te objective of tis paper is to use general equilibrium models to quantitatively assess te impact of te current recession in Europe, Japan and Cina on Indonesian poverty. It is expected tat te results from suc a general equilibrium analysis will give us important insigts on ow to design appropriate policy instruments for protecting te Indonesian economy from external disturbances. Te next section of tis paper outlines te teory and te structure of te model. Section tree outlines te data and te metodology, including te scenario of simulation. Section four provides result and analysis, wile section five gives conclusion and te policy implications. II. THEORETICAL FRAMEWORK AND EMPIRICAL LITERATURE 2.1. How external sock can affect poverty It sould be clear tat wat it means by external sock is in te context of exogenous cange from te country s persepective. In tis case, tose socks can be a recession in a particular country, a group of countries, or one or some regions. Te socks can also be international price of commodities determined in te world market exogenous to te perspective a a particular country or te assumption of Indonesia being a small open economy. Poverty sould also be defined clearly. Tere are many dimension of poverty, but in tis particular study, te focus is on expenditure poverty wic is quantitatively measured as ead-count poverty incidence, wic can be written as Foster, Greer, Torbecke (FGT) Index of poverty initially developed by Foster et al (1984):

3 External Socks And Poverty: How Recession In Europe, Japan, and Cina Affects Te Indonesian Poor 209 H Σi = 1 P α = 1 ( z - y i ) N z α Were P is te poverty indicator, N is te number of population, H is te number of people wose expenditure per person is below te poverty line, y is a vector of expenditure per person in increasing order, z is te poverty line, α 0 is a sensitivity parameter. Wen α=1 tis become te ead-count poverty incidence, wic is a proportion of population living below poverty line. Tis indicator is wat will be used trougout tis paper. Figure 1 below illustrate ow external sock, like global economic crisis of a particular countries or regions can affect poverty incidence in a national economy. Price of factors Poor ouseold s income Riil Expenditure of poor ouseolds Poverty incidence Production Price of commodities Product market Market of factor of production Houseolds Oter institutions Composite Domestic Import Price of factor of production Price of commodities Poor ouseold s CPI Export +/- Rest of te world Recession in EJC (GDP falls & world price deflation) Import +/- Non poor ouseold s income Non Poor ouseold s CPI Riil Expenditure of Non poor ouseolds Figure 1. Te effect of external sock on poverty incidence Te impact of an EJC recession on te Indonesian economy can work troug various cannels, including trade, foreign direct and portfolio investment, and oter monetary or financial cannels. In tis analysis, te focus is on te impact troug te trade cannel. A recession in te EJC countries will reduce imports and exports from countries like Indonesia. Te impact on te Indonesian economy is likely to be negative, and te magnitude of te impact depends largely on te sare tat tese countries ave in Indonesian exports. Tis is te direct

4 210 Buletin Ekonomi Moneter dan Perbankan, Volume 18, Nomor 2, Oktober 2015 trade effect. Te indirect trade effect is rater unclear. Te reduction of imports from Europe, Cina and Japan will tend to reduce te world prices of commodities. Global deflation like tis may reduce Indonesian exports yet increase Indonesian imports. On te oter and, te decline in te export of goods from te EJC countries may increase te world prices of commodities. To some countries (including Indonesia), tis migt represent an opportunity to increase exports and to raise world market sare. Tis effect tends to work positively for te Indonesian economy. To complicate tings, tese effects work differently among different countries, eac wit a different type of repercussion on te Indonesian economy. Terefore, te overall net impact on te Indonesian economy is difficult to predict and calls for a computable general equilibrium analysis. Te effect on poverty depend on te general equilibrium outcome of prices or new equilibrium prices canged due to te external socks. Tere are two kind of prices tat at te end will affect troug different cannels to ouseolds. First of all is te price of factors of production will affect ouseold income. Te impact on ouseolds will be different even toug all ouseolds face te same price of capital or labor because poor and non-poor ouseolds as different composition of ouseold income. Ric ouseolds as more capital income tan labor, wile poor ouseolds as more labor income. Anoter cannels is its impact on price of commodities. Again, altoug all ouseolds face te same price of commodities, te composition of ouseold ecpenditures may differ. Poor ouseold may buy less of travels but more of rice, wile ric ousolds may buy more of gasoline and oter luxurious items. Te final impact on tese two cannels are on te real expenditure, or expenditure per capita. Poverty incidence ten is used based on tis real expenditure per person before and after te socks and te difference is calculated as te impact of external socks on poverty incidence Empirical literature Tere ave been some studies looking at te of global crisis, particuarly te crisis, wit particular attention on te impact on poverty. In tis review, only studies tat uses economy-wide model, microsimulation model, or combination of bot are reviewed. Wong, (2012) quantifies te macroeconomic and poverty impacts of te world economic crisis on Ecuador, including te effects of te main policy responses of te Ecuadorian Government to face te crisis. Te main ypotesis igligts te magnitude of two transmission cannels: trade and remittances. Te study applies a single-country static computable general equilibrium model for Ecuador combined wit a microsimulation model. From a distributional point of view, te impacts of te crisis were progressive, affecting more negatively ouseolds in te igest income quintile. A key cannel of transmission is te fall

5 External Socks And Poverty: How Recession In Europe, Japan, and Cina Affects Te Indonesian Poor 211 in capital returns and wages of skilled labor. Tese factors are used intensively in te oil sector, a key sector of te Ecuadorian economy, and one of te ardest it by te global crisis. Tere are differentiated poverty impacts: poverty may increase if labor is assumed sector specific, and it may be reduced if labor were mobile. (Habib, Narayan, Olivieri, & Sancez-Paramo, 2010b) develops a simple micro-simulation metod, modifying models from existing economic literature, to measure te poverty and distributional impact of macroeconomic socks by linking macro projections wit pre-crisis ouseold data. Te approac is ten applied to Banglades to assess te potential impact of te slowdown on poverty and income distribution across different groups and regions. Te poverty eadcount rate (based on te upper poverty line) and extreme poverty rate (based on te lower poverty line) are expected to be 0.4 and 0.2 percentage points iger in 2009, respectively, as a result of te crisis. Using a microsimulation analysis, (Ajwad, Francisco, & Azam, 2012) look at te distributional impacts of te 2009 financial crisis on ouseolds in Latvia. Te simulations sow tat Latvia experienced a sarp rise in poverty, widening of te poverty gap, and a rise in income inequality due to te economic contraction in Te 18 percent contraction in gross domestic product (affecting mainly trade otels and restaurants, construction, and manufacturing) likely led te poverty ead count to increase from 14.4 percent in 2008 to 20.2 percent in (Amed & Donogue, 2010) adopt a macro-micro framework in order to evaluate te impact of a global crisis on te Pakistan economy. Tey use a top-down approac to combine a static computable general equilibrium model wit a microsimulation model. Our results suggest tat between 2007 and 2009 te poverty eadcount ratio is likely to ave increased by almost 80 percent, from 22 to 40 percentage points. However, teir results also sow tat tis increase is attributable in part to te fuel and food crisis tat preceded te financial crisis. Tey also indicate a differential impact, wit wage increases for farm workers and a decrease in wages for skilled labour. (Weeks, 2009) using a macroeconomic model finds tat te global financial crisis will cause a fall in export earnings in Sierra Leone of approximately fifteen percent in 2009 compared to He find tat tis decline in exports earnings could result in a fall in national income of almost ten percent and based on te income distribution in te 2003 ouseold survey, a ten percent decline in national income would increase poverty by twelve percent of t population, or about 600,000 people. (Habib, Narayan, Olivieri, & Sancez-Paramo, 2010a) uses a micro-simulation approac to assess te poverty and distributional effects of te crisis in te Pilippines. Tey find increases in bot te level and te dept of aggregate poverty. Income socks are relatively large in te middle part of te income distribution. Tey also find tat caracteristics of people wo become poor because of te crisis are different from tose of bot cronically poor people and te general population.

6 212 Buletin Ekonomi Moneter dan Perbankan, Volume 18, Nomor 2, Oktober 2015 (Bibi et al., 2010) evaluate te potential impacts of te 2008/09 global economic crisis on cild poverty in Cameroon. A dynamic computable general equilibrium (CGE) model is used to simulate various scenarios of te economic crisis togeter wit policies wic respond to te crisis. Te study sows tat te crisis is projected to lower te real GDP growt rate by 1.3 percentage points in 2009, 0.9 in 2010 and 0.8 in Te crisis would also bring about a 1.05% increase in te number of cildren wo were poor in monetary terms in 2008 and a 4% increase in 2009, 2010 and 2011, compared to te situation witout a crisis. In teir analysis, (Estrades & Llambi, 2013) evaluate te impact of te 2008 global economic crisis on Uruguayan economy troug two main cannels: collapse in global trade and drop in capital flows. Tey apply a computable general equilibrium model linked to microsimulations to analyze te distributional impacts of tese policies and assess teir effectiveness. Tey find tat an increase in public investment was te only policy effective in mitigating te negative impact of te crisis on extreme poverty. In a study wic focus on conceptual framework wit te application to Vietnam, (Coxead, Lin, & Tam, 2012) present a framework for understanding te direct and indirect welfare effects of a global market sock. Tey quantify transmission of te sock from global indicator prices to domestic markets. Ten tey use an applied general equilibrium model to simulate te economic effects of te price canges. A recursive mapping to a nationally representative ouseold living standards survey permits tem to identify in detail te ceteris paribus effects of te sock on ouseold incomes and welfare. In tis analysis, interregional and intersectoral labor market adjustments emerge as key cannels transmitting te effects of global price socks across sectors and among ouseolds. (Corong & Taningco, 2011) analyzes ow te global crisis may ave affected te Pilippine economy. To te extent tat te Pilippines is more globally integrated troug trade and labor flow cannels tan te financial sector, it is expected tat impact of te global crisis will weig eavily on te real side of te economy. A dynamic computable general equilibriu m (CGE) model linked to a micro-simulation module used trace effects: from te macro-economic to te microeconomic level; from output and factor supplies and demands to commodity and factor prices; and from ouseold incomes to levels of poverty and income distribution. Simulation results suggest tat all ouseolds experience a significant reduction in real income. Bot inequality and poverty increase, wit urban dwellers experiencing a iger in crease in poverty relative to teir rural counterparts as most export-oriented industries are located in te urban areas and returns to factors intensively used by tese industries fall. To summarize, most of te studies above suggests tat global financial crisis affect te poor negatively. However, te magnitude is different; tere are exceptions were te impact on poverty is negligible because te crisis affect mostly te ric; and policy responses can ave significant impact on te final effect of te crisis. Tere is owever, a metodological gap in most of te studies in ow tey linked global socks into te national model. Te analysis could

7 External Socks And Poverty: How Recession In Europe, Japan, and Cina Affects Te Indonesian Poor 213 not really explicitly attribute te socks to a recession in a specific countries or regions. Tis is because tey only use a national model. In tis paper, two models will be used, a global or multi-country model as well as national model to andle poverty. Wit a global model, we can be more specific on wic countries experience a recessions. Tis is quite important because te impact of a recessions of different countries will ave a different impact on te export demand of and te price of te commodities faced by te particular affected countries. III. METHODOLOGY To estimate te impact of recession in Europe, Japan, and Cina on Indonesia s economy, we use te Global Trade Analysis Project (GTAP) model, a multi-country, multi-sector general equilibrium model. 2 For tis analysis, te GTAP version7 database is aggregated into 16 regions and 57 sectors. 3 To simulate a recession in te Euro Zone, te European Union (EU), Japan, Cina, and in all combined we reduce te endowment of te primary factors. Our primary focus ere is on te recession in Europe. However, we add te recession in Japan and Cina to te analysis to examine te potential impact of a recession in tese two countries. To reflect a reasonable magnitude of recession, we compare te forecasts of country GDP presented in te IMF s World Economic Outlook database of 2009 and Forecasts of European GDP for 2012, 2013, and 2014 are ten compared assuming tat te forecast in 2009 is business as usual and te forecast in 2011 already reflects te European crisis. On te basis of tis comparison, we calculate tat te European recession can be defined as rougly a 2 percent deviation of GDP below baseline. We use tis 2 percent deviation to simulate te endowment of primary factors in Europe, Japan, and Cina Te structure of GTAP Model Teoretical structure Based on (Hertel, 1997) te teoretical structure of GTAP model can be summarised as follows. Readers wo are interested in detailed exposition of te model can directly refer to (Hertel, 1997) or Brockmeier (2001). Te production of eac sector in eac country is represented by a nested production function wic is a combination of a Leontief, Constant Elasticity of Substitution (CES) specification. Tis kind of a nested or staged production is a convenient way of representing separable, constant returns-to-scale tecnologies. At te bottom of te inverted tree are te 2 For more explanation of te GTAP model, see Hertel (1997). 3 Oceania, Japan, Cina, Rest of East Asia, Indonesia, Singapore, Tailand, Rest of SE Asia, Nort America, Latin America, Euro Zone, Rest of EU, Mid East and Nort Africa, sub-saaran Africa, and te rest of te world.

8 214 Buletin Ekonomi Moneter dan Perbankan, Volume 18, Nomor 2, Oktober 2015 individual inputs demanded by te firm. Tte primary factors of production are: land, labor, and capital. Firms also purcase intermediate inputs, some of wic are produced domestically, and some of wic are imported. For imports, te intermediate inputs must be sourced from particular exporters. Te strong nature of tis kind of multi-country model. Regional ouseold (a representative of ouseold for eac country) beavior is governed by an aggregate utility function, specified over composite private consumption, composite government purcases, and savings. In te GTAP model we employ a special case of te Stone-Geary utility function, wereby all subsistence sares are equal to zero. Government demands follows te Cobb-Douglas utility function. Factors are imperfectly mobile in a GTAP model. Te mobility of tese endowments is described wit a constant elasticity of transformation (CET) revenue function. Database In tis paper te GTAP database version 7 is used. (Hertel, 1997)describe te database as aving 113 regions. Te database was constructed from input-output table from eac of te 113 countries and combined wit a well-developed inter-regional trade flows data. Detailed description of te database including its construction can be found in (Narayanan, Badrie, & Walmsley, 2008) 3.2. Te structure of INDONESIA-E3 model Teoretical structure Since te GTAP model cannot analyze distribution results witin countries, we use te INDONESIA-E3 model to simulate te effect of an EJC recession on poverty in Indonesia. Te strengt of tis multi-sector, multi-ouseold CGE model of te economy is distributional analysis. Most of its structural features are standard, but its capacity for disaggregation of ouseold structure facilitates analysis of ow exogenous socks affect poverty and inequality. Te teoretical structure of INDONESIA-E3 model is conventional for static general equilibrium models. In particular, te equations in INDONESIA-E3 model represent te following economic beaviour: Production sectors minimizing cost of production given a Constant Elasticity of Substitution tecnology. A system of factor demand equation is derived and specified in te model. Tis relates te demand for eac primary factor to industry outputs and prices of eac of te primary factors (labor, capital, land, and intermediate inputs). Tis reflects te assumption

9 External Socks And Poverty: How Recession In Europe, Japan, and Cina Affects Te Indonesian Poor 215 tat factors of production may be substituted for one anoter in ways tat depend on factor prices and on te elasticities of substitution between te factors. Users of commodities wic include industries, ouseolds, investors, government sectors form a system of demand equations. Tis demand system for eac of tese users consists of tree layers (nested demand system). Consumers/users coose te optimal combination of domestically-produced and imported commodities. Te last layer is tey coose te optimal combination of different commodities responding te prices and budget constraints tat tey face. For ouseold, a Linear Expenditure Demand System (LES) is specified. Te ouseold supplies of skilled and unskilled labor as well as capital and land. A distinction between four kinds of labor: agricultural labor, manual/production worker, clerical workers, and managerial workers. Tese are are nested witin te industry production functions. In eac industry, all kind of labor enter a CES production function to produce labor, wic itself enters a furter CES production function for industry output. A set of export demand functions, indicating te elasticities of foreign demand for Indonesia s exports to te rest of te world. Rates of import tariffs and excise taxes across commodities, rates of business taxes, value added taxes and corporate income taxes across industries, and rates of personal income taxes across ouseold types wic reflect te structure of te Indonesian tax system A set of macroeconomic identities wic ensures tat standard macroeconomic accounting conventions are observed. In general, te demand and supply equations for private-sector agents are derived from te solutions to tese agents microeconomic optimization problems (cost minimization for firms and utility maximization for ouseolds). Te agents are assumed to be price-takers, wit producers operating in competitive markets wit zero profit conditions, reflecting te assumption of constant returns to scale. Te unique feature of INDONESIA-E3 model wic is very relevant in tis study is te disaggregation of ouseold by expenditure classes wic allows for precise estimates of te distributional impact and poverty incidence. In te literature of te poverty impact analysis using CGE models, tis class of model is called an integrated CGE model (Bourguignon, Robilliard, & Robinson, 2003). Tis class of model normally as disaggregated ouseolds wic link eac of te ouseolds to bot sources of income (troug market of factors of production) and expenditure (troug market for commodities). Tis sould be distinguised from oter class of model wic is called top-down, were te CGE model is separate from te poverty module, and between tem is only one directional relationsip. In te integrated model, tere is no separation between CGE model and poverty module because all are in one model.

10 216 Buletin Ekonomi Moneter dan Perbankan, Volume 18, Nomor 2, Oktober 2015 INDONESIA-E3 as been used in various researc for example to analyze te distributional impact of fuel pricing reform (Yusuf & Resosudarmo, 2008); te poverty and distributional impact of carbon tax (Ministry of Finance Republic of Indonesia, 2009); greenouse gasses emission from land use cange (Warr & Yusuf, 2011). More detailed exposition of te model can be found in (Yusuf, 2008) Database construction Due to space limitation detailed construction of INDONESIA-E3 s database including all te steps and assumptions can be found in (Yusuf, 2006). INDONESIA-E3 model of te current version use a Social Accounting Matrix (SAM) for its database representing Indonesian economy for te year Te integration of igly disaggregated ouseolds adequate for accurate distributional analysis is made possible by constructing an Indonesian Social Accounting Matrix (SAM) wic serves as te core database to te CGE model. Te SAM consists of up to 175 industries, 175 commodities, and 200 ouseolds (100 urban and 100 rural ouseolds grouped by percentile of real expenditure per capita). Te data used for constructing te SAM include Indonesian Input-Output Table, official SAM, and most importantly ouseold level survey data (SUSENAS). Using a general equilibrium model wit a disaggregated ouseold sector feature makes it possible to conduct controlled experiments tat focus on te effects of different economic socks on ouseold income, expenditure, poverty, and inequality. Te model identifies two categories of ouseolds, rural and urban, eac of wic is divided into 100 subcategories of equal population size, wit te subcategories arranged by expenditures per capita. 4 To link and transmit te result of GTAP simulations into te INDONESIA-E3 model, we (1) aggregate te sectors in INDONESIA-E3 into 57 to matc te sectors in GTAP and ten (2) introduce te cange in Indonesian exports by commodities, in te world price of imports, and in te world price of exports resulting from GTAP simulations as socks in te INDONESIA-E3 model. IV. RESULTS AND ANALYSIS Results from our GTAP simulation suggest tat Indonesia is not among te countries most eavily affected by a recession in Europe. A 2 percent decline in Euro zone countries GDP reduces Indonesia s GDP by only percent relative to baseline. A 2 percent decline in all 27 countries members of te EU still reduces Indonesia s GDP by only percent relative to baseline (Table 1). 4 For more on te INDONESIA-E3 model, please see Yusuf (2008).

11 External Socks And Poverty: How Recession In Europe, Japan, and Cina Affects Te Indonesian Poor 217 Table 1 sows tat te countries most negatively affected by a European recession are tose in te Middle East and Nort African (MENA) region. Wit a 2 percent fall in te EU s GDP, MENA s GDP falls by almost 0.6 percent, an elasticity of 0.3. Europe is te biggest importer of tis region s energy and MENA countries are eavily dependent on energy exports. In Asia, Singapore is te country most affected by a European recession because of its dependence on international trade. Japan and Cina, owever, experience a sligtly positive increase in GDP. If Cina and Japan experience a recession of similar magnitude te impact on Indonesia s GDP will not be stronger tan te impact of te European crisis. But te impact is disproportionate because Japan and Cina are single countries wile te EU consists of 27. A 2 percent decline in Cina s GDP causes a 0.07 percent decline in Indonesia s GDP and a similar decline in Japan s GDP causes a 0.03 percent decline in Indonesia s GDP. In oter words, Cina s potential impact on Indonesia is almost twice tat of Japan s. Te impact of recession in Japan will be felt most in Singapore, were GDP will fall 0.24 percent, and te impact of recession in Cina will be felt most in Tailand, were GDP will fall 0.33 percent relative to baseline. Singapore and Tailand are among te most open economies in Asia.

12 218 Buletin Ekonomi Moneter dan Perbankan, Volume 18, Nomor 2, Oktober 2015 In sum, a European recession of a sensible magnitude (2 percent GDP deviation relative to business as usual) working troug full trade linkage (taking into account indirect trade effects troug oter countries as well) will not ave a very large negative impact on Indonesia s GDP. Te reason for tis is tat Indonesia s economy depends little on external trade. Tis does not mean tat European recession will not affect te world economy to a considerable extent; oter countries, like Singapore and MENA countries, will be greatly affected. Wit respect to poverty in Indonesia, te impact of a recession in Europe, Japan, and Cina is also small. A combined 2 percent decline in GDP in Europe, Japan, and Cina will increase Indonesia s national poverty ead count by only 0.19 percent (Table 2). Te biggest impact will be if te recession is in eiter Cina, or all of te EU. Te impact on a recession on poverty in Indonesia is small because (1) te impact on overall mean real consumption is small and (2) te impact is most strongly felt among iger income ouseolds. As suggested in Figure 1, were percentage cange in real expenditure is plotted on te percentile of expenditure per capita, te impact of recession in Europe, Japan, or Cina is similar. Te impact increases wit income status. For example, te effect of a recession in te Euro zone on real consumption of te ricest 1 percent of urban ouseolds is 3 times bigger tan on te poorest 1 percent. Similarly, te ricest 1 percent of rural ouseolds experiences a decline in real consumption 4 times bigger tan te poorest 1 percent. Te result is similar in all simulations (Euro zone, EU, Cina, or Japan recession). Te range of te ratio of te impact is between 2.3 to 2.9 for urban ouseolds, and between 3.3 to 4.4 for rural ouseolds. Tis suggests tat te simulations ave an inequality-reducing tendency. Te distributive effect discussed earlier can be explained by ow recession affects (1) ouseold income, particularly return on factors of production; and (2) te pattern of cange in te ouseold-specific consumption price index (CPI). Te final effect on ouseold real consumption is a function of tose two factors. Global recession tends to decrease te global prices of commodities. Transmitted to te Indonesian economy, tis deflationary effect will tend to reduce commodity prices for domestic consumers. A careful look at ow tis affects ouseolds of different income status suggests tat te effect is only sligtly progressive if not neutral. Across all simulations, te ratio of te impact on ouseold-specific CPI between te ricest 1 percent and te poorest 1 percent is less tan 1. Tis suggests tat te progressivity of impact arises largely from te income side. Ricer ouseolds tend to experience income falls far greater tan poorer ouseolds. Figure 2, wic sows te simulated impact of recession in Europe, Japan and Cina on real income of different factors of production, explains tis. Almost all factors of production experience real decline in income; owever, te biggest decline is were ownersip is concentrated among iger income ouseolds. Capital income declines more tan labor income, and skilled labor income declines more tan unskilled labor income. Tis pattern of impact favors poorer ouseolds wose incomes largely depend on unskilled labor.

13 219 External Socks And Poverty: How Recession In Europe, Japan, and Cina Affects Te Indonesian Poor

14 220 Buletin Ekonomi Moneter dan Perbankan, Volume 18, Nomor 2, Oktober 2015 Source: Autor s calculation based on simulation wit INDONESIA-E3 model. Grapic 1. Simulated Impact on Real Consumption by Houseold Groups

15 External Socks And Poverty: How Recession In Europe, Japan, and Cina Affects Te Indonesian Poor 221 Figure 2. Simulated Impact on Real Income of Factors of Production Figure 3. Simulated Impact on Output of Aggregated Sectors (% cange from baseline)

16 222 Buletin Ekonomi Moneter dan Perbankan, Volume 18, Nomor 2, Oktober 2015 Te pattern of impact on production factors income is largely driven by ow industries respond to trading partners recession. As sown in Figure 3, te simulated recession in Europe, Japan, and Cina negatively affects output in certain sectors, most notably oil, gas, and oter extractive sectors. Tese sectors are capital- and skill-intensive. Figure 4 sows ow te contraction and expansion of output is correlated wit canges in exports of tose sectors. A closer look at te impact on poverty suggests tat rural areas would be arder it tan urban areas (Table 2). For example, a recession in Europe, Japan, and Cina combined would increase te incidence of poverty in urban areas by 0.10 percent and in rural areas by 0.28 percent. To understand tis, we examine wat appens to real consumption in marginally poor urban and rural ouseolds. A marginally poor ouseold is one wose expenditure per capita is closest to te poverty line. Knowing wat appens to tese ouseolds sows ow te simulation affects poverty incidence. As explained in Warr, Menon, and Yusuf (2012), Te basis for te decomposition is as follows. We focus on te sources of canges in te real expenditure of a particular ouseold, say ouseold, arising from some external sock. Upper case Roman letters, like Z, will denote levels of variables and lower case Roman letters, like z, will denote teir proportional cange, so tat z = dz /Z. Te levels of nominal income and nominal expenditure of ouseold will be denoted Y and E, respectively. Let te proportional cange in te nominal expenditure of ouseold, be e = e ~ + p, were I i e is te proportional cange in te ouseold s real i expenditure and p = i = ε p is te proportional cange in a consumer price index specific 1 i i to ouseold, wit ε = E / E denoting tat ouseold s expenditure sare on commodity i, E i denoting its nominal expenditure on commodity i and p i denoting te proportional cange in te consumer price of commodity i. Te absolute cange in tis ouseold s nominal expenditure is now de = E e = E ( e ~ + p ~ ) = de + E p ~ = de + I i= 1 E i p i Tat is, te cange in nominal expenditure of te ouseold is given by te cange in its real expenditure plus te cange in its true cost of living, te latter an expenditure weigted sum of te canges in te consumer prices tat ouseold actually faces, were te expenditure weigts pertain to tat particular ouseold. 5 Te cange in nominal expenditure is also equal to te cange in nominal income minus te cange in saving, so tat de = dy ds. Disregarding any canges in transfer income or direct taxes, for simplicity, te cange in nominal income is equal to te cange in nominal factor income, dy = dy f. 5 It sould be noted tat real expenditures means expenditures measured at constant prices, defined ere to mean base period prices. Tus, te levels of nominal and real expenditures in te base period are identical, meaning E = E.

17 External Socks And Poverty: How Recession In Europe, Japan, and Cina Affects Te Indonesian Poor 223 Tus, rearranging terms, d E = dy f ds I i=1 E i p i Te cange in te ouseold s real income is decomposable into tree components: (i) te cange in its nominal factor income minus (ii) te cange in its savings minus (iii) te cange in its true cost of living. Importantly, te cange in nominal factor income is itself additively decomposable into its factor components, as identified in te model. Tis decomposition is applied to tose ouseolds are tose wo are in te border of poverty line. Tey are percentile 8 ouseold in urban area and te percentile 15 in rural area. Wat appens to tese two representative ouseolds may explain wat appens to poverty incidence. As sown in Table 2, te percentage cange in te real expenditure of marginally poor ouseolds in urban and rural areas does not differ muc. For te combined simulations, for Figure 4. Simulated Impact on Export of Aggregated Sectors (% cange from baseline)

18 224 Buletin Ekonomi Moneter dan Perbankan, Volume 18, Nomor 2, Oktober 2015 example, marginally poor ouseolds in urban areas experience a 0.34 percent decline in real consumption wile tose in rural areas experience a 0.37 percent decline. Tis muc larger impact on poverty incidence in rural areas suggests tat te elasticity of poverty incidence wit respect to te cange in te real consumption of marginally poor ouseolds is relatively larger. Te slope of te cumulative distribution function (CDF) of te real consumption of te rural ouseolds around te region of te poverty line must be iger tan tat of urban ouseolds. Wit te same magnitude of cange in teir consumption, te rural poor are more vulnerable to external socks tan urban ouseolds. It is also useful to understand to wat extent cange in te real expenditure of te marginally poor ouseold is driven by its decomposable drivers. Te decline in te real expenditure of te poor can be generally decomposed into income effect (decline in income) and consumer price effect (increase in te price of te commodity basket of te poor or living cost). Te income effect can also be decomposed (e.g., labor income, capital income). Table 2 sows ow it is te case for urban and rural marginally poor ouseolds. Several conclusions can be drawn from tis decomposition exercise. First, te poverty-increasing effects come only from te income side. Poverty increases because te income of te poor decreases, not because of a rise in inflation or te cost of living. Global recession, ceteris paribus, is deflationary and as sown in Table 2, te cost of living for te poor actually falls. Second, te fall in income of te poor is largely te result of a fall in unskilled labor income. As our earlier discussion suggests, income from unskilled labor falls te least as compared to oter factor income in aggregate, but tis income makes up te dominant sare of te poor s income. Tird, despite a decline in te cost of living, te real consumption of te poor falls because te decline in income cannot be compensated for. As a result, poverty increases in urban and rural areas. V. CONCLUSION Te effect of a recession in Europe, Japan and Cina on oter economies troug trade linkage depends on many factors. Tese factors include te importance of bilateral trade between te affected countries or regions, te openness of te economy, and te country s level of diversification in export commodities and trading partners. Te impact on poverty in te affected countries depends even more on various country specific aspects suc as ow te poor earn its income and ow te spend teir earning. If bigger part of its source of income are affected by te global crisis and te price of te commodities tey intensively consumed are eavily affected, ten te poor will be vulnerable to external socks. Tis is empirical and tis paper touc tis important issue. Tis paper sows tat te impact of a recession in Europe, Japan, or Cina on Indonesia is relatively small simply because Indonesia is a relatively less open economy and depends

19 External Socks And Poverty: How Recession In Europe, Japan, and Cina Affects Te Indonesian Poor 225 predominantly on its domestic market. According to our analysis, te current recession in Europe, Japan and Cina will ave only a small negative effect on Indonesia s GDP and poverty rate. However, results suggest tat it would be wise to pay attention to Cina. According to our simulation, a recession in Cina could reduce Indonesia s GDP two times more tan a recession of similar magnitude in Japan. Tis analysis sows tat a recession in Europe, Japan, and Cina would ave its largest negative impact on ricer ouseolds in Indonesia. Ricer ouseolds would suffer te most because teir incomes are largely dependent on te capital- and skill-intensive sectors of te economy. Wit respect to te poor, our analysis suggests tat a recession in Europe, Japan or Cina will not ave muc of an impact on prices in Indonesia. In oter words, te cost of te living of te poor will not be muc affected. Policymakers in Indonesia sould terefore concentrate less on pricing policies and more on employment policies to protect te poor from loss of employment. Since tey are less skilled, te poor are likely to be te first to lose teir jobs in te event of a recession. Finally, our analysis does not cover te effects of recession in Europe, Japan and Cina on economic cannels oter tan trade, suc as foreign direct investment and te financial sector. Te general equilibrium model used in our analysis does not cover te financial sector, so oter studies focusing on tis cannel are warranted. In addition, in te longer term, poverty is affected by various factors oter tan income and prices. Human capital accummulation, sosial assitance policies and equality of opportunity are important in poverty alleviation. Tis effects are not accounted for in tis analysis.

20 226 Buletin Ekonomi Moneter dan Perbankan, Volume 18, Nomor 2, Oktober 2015 REFERENCES Amed, V., & Donogue, C. O. (2010). Global Economic Crisis and Poverty in Pakistan. International Journal of Microsimulation, 3(1), Retrieved from ttp:// microsimulation.org/ijm/v3_1/ijm_36.pdf Ajwad, M. I., Francisco, H., & Azam, M. (2012). Simulating Te Impact of te 2009 Financial Crisis on Welfare in Latvia. Bibi, S., Cockburn, J., Emini, C. A., Fofana, I., Ningaye, P., & Tiberti, L. (2010). Impact of te Global Economic Crisis on Cild Poverty in Cameroon and Options for a Policy Response. Italy. Bourguignon, F., Robilliard, A.-S., & Robinson, S. (2003). Representative Versus Real Houseolds in te Macro-Economic Modelling of Inequality. Paris. Brockmeier, Martina (2001). A Grapical Exposition of te GTAP Model. GTAP Tecnical Papers (No. 8). Corong, E., & Taningco, A. (2011). Global Economic Crisis and te Pilippine Economy: A Quantitative Assessment. Coxead, I., Lin, V. H., & Tam, L. D. (2012). Global Market Socks and Poverty in Vietnam: Te Case of Rice. Agricultural and Applied Economics, 43(5), Estrades, C., & Llambi, C. (2013). Lessons from te 2008 Financial Crisis: Policy Responses to External Socks in Uruguay. Te Developong Economies, 51(3), Habib, B., Narayan, A., Olivieri, S., & Sancez-Paramo, C. (2010a). Assessing Ex Ante te Poverty and Distributional Impact of te Global Crisis in a Developing Country (No. 5238). Wasington DC. Habib, B., Narayan, A., Olivieri, S., & Sancez-Paramo, C. (2010b). Assessing Poverty and Distributional Impacts of te Global Crisis in te Pilippines (No. 5286). Hertel, T. (1997). Global Trade Analysis: Modeling and Applications. Cambridge and New York: Cambridge University Press. Ministry of Finance Republic of Indonesia, M. o F. (2009). Economic and Fiscal Policy Strategies on Climate Cange: Indonesian Green Paper. Jakarta Narayanan, Badrie, G., & Walmsley, T. L. (2008). Global Trade, Assistance, and Production: Te GTAP 7 Data Base (8t ed.). Center for Global Trade Analysis, Purdue University. Retrieved from ttp://

21 External Socks And Poverty: How Recession In Europe, Japan, and Cina Affects Te Indonesian Poor 227 Warr, P., & Yusuf, A. A. (2011). Reducing Indonesia s Deforestation-Based Greenouse Gas Emissions. Australian Journal of Agricultural and Resource Economics, 55(3), ttp://doi.org/ /j x Weeks, J. (2009). Te Impact of te Global Financial Crisis on te Economy of Sierra Leone. International Policy Centre for Inclusive Growt, 18, Wong, S. (2012). Sort-term Macroeconomic and Poverty Impacts of te Global Economic Crisis on te Ecuadorian Economy. Yusuf, A. A. (2006). Constructing Indonesian Social Accounting Matrix for Distributional Analysis in te CGE Modelling Framework (No ). Yusuf, A. A. (2008). An Indonesian Applied General Equilibrium Model for Analyzing te Economy, Equity, and te Environment (No ). Yusuf, A. A., & Resosudarmo, B. P. (2008). Mitigating Distributional Impact of Fuel Pricing Reform: Indonesian Experience. ASEAN Economic Bulletin, 25(1), Retrieved from ttp://

22 228 Buletin Ekonomi Moneter dan Perbankan, Volume 18, Nomor 2, Oktober 2015 Halaman ini sengaja dikosongkan