WPS5286 Policy Research Working Paper 5286 Assessing Poverty and Distributional Impacts of the Global Crisis in the Philippines A Microsimulation Approach Bilal Habib Ambar Narayan Sergio Olivieri Carolina Sanchez-Paramo The World Bank Poverty Reduction and Economic Management Network Poverty Reduction and Equity Unit April 2010 Policy Research Working Paper 5286 Abstract As the financial crisis has spread through the world, that characteristics of people who become poor because the lack of real-time data has made it difficult to track of the crisis are different from those of both chronically its impact in developing countries. This paper uses a poor people and the general population. The findings micro-simulation approach to assess the poverty and can be useful for policy makers wishing to identify distributional effects of the crisis in the Philippines. The leading monitoring indicators to track the impact of authors find increases in both the level and the depth of macroeconomic shocks and to design policies that protect aggregate poverty. Income shocks are relatively large in vulnerable groups. the middle part of the income distribution. They also find This paper--a product of the Poverty Reduction and Equity Unit, Poverty Reduction and Economic Management Network--is part of a larger effort in the department to analyze the poverty and distributional impacts of the 2008­9 financial crisis.. Policy Research Working Papers are also posted on the Web at http://econ.worldbank.org. The authors can be contacted at Carolina Sanchez (csanchezparamo@worldbank.org) and Ambar Narayan (anarayan@worldbank.org). The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. Produced by the Research Support Team Assessing Poverty and Distributional Impacts of the Global Crisis in the Philippines: A Microsimulation Approach Bilal Habib, Ambar Narayan, Sergio Olivieri and Carolina SanchezParamo* * This paper is the second in a series of country notes to be prepared under the work program on Analyzing the Distributional Impact of the Financial Crisis, in the Poverty Reduction Group of the PREM Network. For this exercise we gratefully acknowledge the data, inputs and advice received from the Philippines Country Team of the World Bank ­ in particular Andrew Mason, Manohar Sharma, Laura Pabón, Eric Le Borgne, Jehan Arulpragasam, Ulrich Lachler, Sheryll Namingit, and Rashiel Besana Velarde. 1. Introduction The economic slowdown across the world caused by the financial crisis has taken its toll on poverty in developing countries. Recent World Bank global estimates for poverty suggest that the global slowdown caused by the crisis could add an extra 50 million to the number of people living below $1.25 a day and 64 million to the number below $2 a day, compared to a scenario of uninterrupted growth in the absence of the crisis. By 2010, the crisis is projected to add an estimated 89 million to the number of people living below $1.25 a day and 120 million to the number below $2 a day (Chen and Ravallion, 2009). Along with the impact on poverty numbers, the crisis is also likely to have significant impacts on the distribution of income and consumption among the poor and nonpoor, both within and between countries. It is difficult, however, to predict the distributional impacts of the crisis with a high degree of confidence. Evidence from previous crises suggests that relative inequality falls about as often as it rises during aggregate contractions (Paci et al, 2008). Furthermore, as the crisis rapidly spreads across countries and within countries (through adjustments in domestic credit and labor markets and fiscal policies), its impacts across different groups, sectors or areas have become all the more difficult to track. An analysis of how the crisis has affected poverty and distribution of welfare in a particular country ­ including Philippines, the subject of this paper ­ would need to address a few key questions. These are: how are the impacts going to be shared across the distribution of income or consumption, which sectors and regions are likely to be impacted, and what are the characteristics of those who would become poor as a result of the crisis? In order to be useful to policymakers, these questions would have to be analyzed ex ante, with available data that in most cases predate the crisis, rather than be delayed till the time post(or during) crisis data become available. In the case of Philippines, the latest available household data is from 2006, which would imply that any analysis of how the crisis is likely to affect poverty and income distribution must rely on ex ante simulation methods. Even if some realtime data from households were available, an ex ante approach would still be needed to simulate impacts for hypothetical scenarios that have not yet occurred, or the impacts of other, unanticipated shocks to the economy (in the case of Philippines, typhoons that occurred very recently) . Current approaches to analyze ex ante the impact of a macroeconomic shock, with the limited data available in most countries, are somewhat inadequate in addressing the kind of questions posed above. To improve upon existing approaches given the typical data constraints seen in most developing countries, including Philippines, we use in this paper a microsimulation model to evaluate exante the distributional impacts of the crisis in a country. Section 2 below provides a brief background of poverty trends in Philippines in recent times and the macroeconomic impact of the crisis in the country. Section 3 outlines a rationale for our microsimulation model and presents the model, starting with an overview of the approach and subsequently outlining each step in detail, and ending with the key limitations and caveats to the model. In Section 4, we discuss the results of our simulations on the poverty and distributional impact of the crisis in Philippines. Section 5 concludes the paper summarizing key findings and discussing a few implications of the results for the country. 1 2. Setting the context: The Philippine economy and the crisis1 The Philippines enjoyed high growth throughout the 1990s and into the early 2000s. These years were also witness to a decline in poverty, as the proportion of Filipinos leaving below $1.25/day fell from 34.9 to 22.0 between 1985 and 2003. This is equivalent to a 35.2% reduction in poverty during the period. Despite this general trend, however, estimates using the $1.25 benchmark show that poverty increased in recent years, consistent with official estimates. Between 2003 and 2006, the Philippines' $1.25/day poverty increased from 22.0 to 22.6 percent (or 1.7 million people). Meanwhile official estimates show an even higher increase in poverty, from 30.0 to 32.9 percent. As in many developing countries, there is wide variation in Philippines in the incidence of poverty across regions and between rural and urban areas. Rural poverty rate was 46% in 2006, compared to an urban poverty rate of around 20%. Regional variation in poverty rate was as significant ­ ranging from 10% in NCR to 47% in Mindanao. In addition, poverty rates were around 20% or below for Central Luzon and Calabarzon, and 35% or higher in Other Luzon, Visayas and Mindanao. In addition to income poverty, there are myriad development challenges facing the country. The country has a high Gini coefficient (an index of income inequality) as compared to its Asian neighbors, and many key MDG targets, especially those related to universal primary education and certain health indicators, are unlikely to be achieved by 2015. The financial crisis, coupled with the food and fuel crisis of 2008, is likely to have a significant impact on poverty in the Philippines, which has been suffering from lagging labor market outcomes since the start of the decade. Real wages have been declining since 2001 (especially since the food price shocks of 200708), and unemployment was persistently high, averaging 7.4% in 2008. 50% of the unemployed were in the 1524 age group and there is a high incidence of unemployment even among the college educated. An interesting feature of the Filipino economy is that there has historically been a low (and often slightly negative) output elasticity of employment, which translates to low responsiveness of employment rate to economic growth. Although the exact reasons for this are unknown, it is hypothesized that this is due to either an increase in returning Overseas Filipino Workers (OFWs), increased labor market participation from spouses and children, or increased employment in informal sectors, accompanied by a decline in the wage and salaried employment. Macroeconomic projections of crisis impact The nature of the Filipino economy, as well as realtime evidence that has emerged during the crisis, suggest that the impact of the crisis in Philippines would be transmitted primarily through the following channels. 1 All figures cited in this section, as well as the ensuing discussion, are based on findings from World Bank (2009a) and World Bank (2009b). 2 Firstly, reduced demand for exports, which account for 47.3% of the GDP, is likely to result in reduced labor demand, particularly in the electronics, textile, and garment sectors. As urban, formal sector workers lose their jobs, they could move to the informal sector, which would tend to depress informal sector wages. Secondly, domestic remittances from urban workers to rural households are likely to decrease. This is problematic in light of the fact that 47% of households in 2006 relied on domestic remittances as a source of income. Thirdly, a reduction in international commodity prices (for example, the price of coconut) may have an adverse impact on the income of agricultural workers. Although a reduction in foreign remittances could have potentially occurred due to the impact of the crisis in labor markets countries where OFWs are located, the available macro projections do not contemplate such a decline but rather predict a slightly increase in international remittances (Table 1). OFWs account for an estimated 27.5% of the Table 1: Real output growth projections in Philippines (%) total labor force, and send remittances to about a quarter of Filipino households, in Benchmark Crisis amounts that add up to as much as 10% of 2009 2010 2009 2010 GDP. Total GDP (real) 4.7 5.0 1.4 3.1 Table 1 shows the aggregate and sector Agriculture 3.0 3.0 1.1 2.2 specific growth rate projections for the Manufacturing 4.0 4.2 (6.2) 1.0 Philippines, obtained from the World Bank Other Country team (see Annex 2, Table A.2 for 6.7 6.7 12.1 8.6 Industries more detailed data). Two macroeconomic Services 5.3 5.7 2.9 3.2 scenarios are considered for each of the Remittances years (2009 and 2010): (i) benchmark ­ the (USD billions 17.9 19.5 17.1 18.0 scenario that was expected in the absence of nominal) the crisis; and (ii) crisis ­ the scenario that is Source: World Bank Philippines country team expected (or occurred in the case of 2009) with crisis. Overall GDP growth is expected to be 3.3 and 1.9 percentage points lower with crisis than in the benchmark scenario in 2009 and 2010 respectively. The biggest shock is expected in manufacturing, primarily due to a fall in demand for manufacturing exports. In 2009, the growth of manufacturing output is estimated to fall by 10.2 percentage points, from a 4% expansion to a 6% contraction. The sector is expected to register a small positive growth rate in 2010, but that will still be 3.2 percentage points lower than what manufacturing growth would have been in the absence of the crisis. The spillover effect from the decline in manufacturing ­ including lower domestic demand for goods and services ­ is likely to have an impact on other sectors as well, especially services. Output growth in the services sector is projected to decline due to the crisis by 2.4 and 2.5 percentage points in 2009 and 2010, respectively, compared to the benchmark (nocrisis) scenario. Some impact is expected in the agricultural sector as well due to the spillover effect from other sectors and reduced commodity prices. Agricultural output growth projected to be lower by less than 1.9 percentage points in 2009 and 0.8 percentage points in 2010 due to the crisis. It is important to note that all sectors, with the sole exception of manufacturing in 2009, are expected to register positive growth for both years even with 3 the crisis. The "loss" in growth is in comparison to the scenario of nocrisis for the relevant year, rather than indicating a fall in output from one year to the next. Interestingly, output growth in the sector "other industries" is expected to be higher during the crisis than in the nocrisis scenario (Table 1). In 2009 and 2010, the output growth in other industries is expected to be 5.4 and 1.9 percentage points higher in the crisis scenario than in the benchmark scenario. Most likely, this reflects the impact of the stimulus package put in place by the government in response to the crisis, which includes substantial resources for public investment projects and is expected to boost the construction sector. In addition, the crisis is expected to have a limited impact on remittance flows of $0.8 billion in 2009 and a larger impact of $1.5 billion in 2010 (Table 1). However, even with the crisis, remittances are expected to grow by $1.6 billion between 2008 and 2010, which is about half of the $3.1 billion growth during this period that was projected in the absence of the crisis. In sum, the macroeconomic impacts of the crisis are estimated to be reasonably large in 2009, with a slight recovery in 2010. The impacts are driven mostly by a sharp contraction in the manufacturing sector. Remittances do not appear to be a significant factor in driving changes in household income in 2009, implying that the effects of the crisis on households are likely to be felt mainly through its impact on employment and labor earnings. 3. Proposed approach to simulate poverty and distributional impacts We propose to use a microsimulation model that combines macroeconomic projections with precrisis micro data from household and/or labor force surveys to predict income and consumption at the individual and household level under different scenarios, which can then be compared to measure poverty and distributional impacts. Comparisons will be made between different projected scenarios (most commonly a crisis or low scenario and a benchmark or base scenario) rather than a comparison over time (i.e. a comparison between 2006 and 2009 or 2010 in the case of the Philippines). Figure 1 presents a stylized representation of the methodology. The model focuses on labor markets and migration as transmission mechanisms and allows for two types of shocks: shocks to labor income, modeled as employment shocks, earnings shocks or a combination of both; and shocks to nonlabor income, with special attention paid to international remittances. Shocks can be positive or negative depending on the trends outlined by the macroeconomic projections. In most cases labor income and remittances account for at least 7580% of household income. Minimum assumptions are made about other sources, such as capital and financial income or public transfers, as discussed below. 4 The data requirements can be summarized as Figure 1: Microsimulation methodology follows. At the macro level, information is needed on projected (i) output, employment, remittances and (ideally) labor earnings growth; (ii) population growth and (iii) price changes. At the micro level, information is needed on (i) labor and nonlabor income, and (ii) labor force status and basic job characteristics, including earnings. Needless to say, the reliability and accuracy of the simulation results is a direct function of the quality and level of detail of the information available at the macro and micro levels. Finally, the income and consumption projections from the model can be used to produce a variety of outputs, including aggregate poverty and inequality comparisons across scenarios, poverty and vulnerability profiling of specific groups and/or areas, and various summary measures of distributional impacts, such as growth incidence curves and state transition matrices. We comment on these extensively for the case of the Philippines below. Overview of simulation exercise In this Section we provide a brief overview of the mechanics underpinning the simulation exercise. The exercise can be broken down into three distinct steps: calibration, simulation and assessment of impacts. A description of each step follows and a schematic of the complete model is presented in Figure 2. Figure 2: Schematic of the microsimulation model 5 Calibration. Calibration is the process by which household and individuallevel information is used to model labor market behavior and outcomes and to predict the likelihood of receiving remittances.2 This is done in three steps. First, we model labor force status for all working age individuals (1564) as a function of household and individual characteristics, where labor force status can be out of the labor force, unemployed, and employed in agriculture, manufacturing, other industries or services. Although ideally we would like to work with a more detailed menu of options, the number of labor force states that can be considered is dictated by the level of disaggregation available for the macro projections. We then use a multinomial logit to estimate the parameters of the model, as well as the individuallevel probability of remaining in a particular state and/or changing to a different one, as given by (1). The approach is similar to that used in Ferreira et al (2009). We estimate the model separately for high and lowskill individuals to allow for structural differences in the labor market behavior of the two groups.3 , I iGj s Ind a s zi b s uis a j zi b j uij | j s (1) where s = Labor force status; G = labor skill level (high/low); z = gender, age, education, region, remittances, presence of public sector employees in the household and land ownership. Second, we model labor earnings for all employed individuals ages 15 to 64 as a function of individual and job characteristics and use a standard Mincerian OLS regression to estimate the parameters of the model, as given by (2) (similar to Ferreira et al. 2008). We use a fairly broad definition of labor earnings for the purpose of the exercise that includes wages and salaries, but also income from selfemployment. This is particularly important in the case of agriculture and for economies with large informal sectors, such as Philippines, since wage and salaried workers constitute a limited fraction of those employed in these sectors. It may lead however to a loss in precision and/or predictive power given that the structural relationship between individual and job characteristics and earnings could be different for salaried and nonsalaried workers. To allow for maximum flexibility and (indirectly) account for some of these differences we estimate the model separately for agriculture, manufacturing, other industries and services and for low and highskill workers.4 log wiG sG xi sG vsG,i (2) where x =gender, age, education, region, land ownership, and indicators for export industry, salaried and public employment. The results of the estimation of equations (1) and (2) and a full description of all variables can be found in Annex 1. 2 We estimate a reduced form of the household incomegeneration model which is based on Bourguignon and Ferreira (2005) and Alatas and Bourguignon (2005) 3 For Philippines, low and highskilled refer to individuals with 0 ­ 9 and >10 years of education, respectively. 4 Notice that, although we could estimate separate models for salaried and nonsalaried workers based on the information from the household survey, we would not be able to use these models for the purpose of predicting future earnings since we do not have earnings and employment information disaggregated by salaried/non salaried workers from macro data. 6 Finally, we focus on nonlabor income. For this purpose we design an assignment rule for changes in international remittances and make some minimal assumptions about other sources of nonlabor income. Ideally, we would estimate a probability model to predict how likely a household would be to receive international remittances, given its characteristics. However, if the migrationrelated information in the survey is poor or insufficient and/or the predictive power of probability model is low (as is the case for Philippines), we are betteroff relying on a simple nonparametric assignment rule that is consistent with the existing evidence (the specific rule used for Philippines is discussed in more detail when we describe the simulation process). Simulation. Simulation is the process by which information on aggregate projected changes in output, employment and remittances is used to generate changes in labor and nonlabor income at the micro level using the structural models developed as part of the calibration.5 This is done in four steps.6 First, we use population growth projections to adjust for demographic changes between 2006 (base year) and 20092010. This adjustment is particularly important in the case of Philippines because fertility rates are still high, which implies that the number of labor market entrants rises faster than overall population, and the baseline household survey is relatively old. In practical terms, doing this allows us to explicitly take into account changes in the size of the working age population, and hence to distinguish between employment growth driven (or rather absorbed) by demographic trends and net (or additional) employment growth. Secondly, we use the projections from the labor force status and labor earnings models to replicate projected changes in aggregate total and sectoral output and employment. We start with employment and calculate the total number of individuals that need to be reassigned between employment and non employment and across employment sectors in order to match projected aggregate changes in total and sectoral employment. We then use the estimated probabilities from the multinomial model to select candidates for reassignment.7 The direction and magnitude of flows between employment and non employment and across sectors of employment is given by changes in the relative shares of different status with respect to the reference population. For instance, whether individuals must be reassigned from nonemployment to employment or viceversa depends on whether the employment rate of individuals ages 15 to 64 is increasing or decreasing. Similarly, workers are expelled from sectors whose relative share in total employment are declining and absorbed into sectors whose relative share in total employment is increasing. The sequence in which individuals are reassigned across states and sectors matters for the final simulation results so we briefly describe here the procedure we follow: 5 We do not assure consistency (i.e that absolute aggregate magnitudes are equal) between the data set used at the two modeling stages (see Bourguignon et. al 2008). Additionally, we assume equal changes at macro and micro levels. We cannot run a test if macro changes are similar or not to micro changes because of lack of a panel data at micro level (see Deaton 2001 and Bourguignon et al 2008). 6 This sequence for introducing changes in the model is based on Vos et al (2002) 7 We add error terms which represent the unobserved heterogeneity of agents' labor supply behavior. These lead to some disparateness in responses to a change in the labor demand, capturing the fact that in the real world individuals who are observationally equivalent (i.e. have identical observable characteristics) might still respond differently the same change in labor demand. 7 Step 1 Flows in and out of the labor force: If the labor force participation rate is increasing, nonparticipants with the lowest predicted probability of being out of the labor force will be reassigned. If the labor force participation rate is declining, participants (employed and unemployed) with the highest probability of being inactive will be reassigned. Reassignments will continue up to the point where the change in the labor force participation rate at the micro level matches the change at the macro level. Step 2 Flows out of contracting sectors: For sectors whose share of total employment is declining, those individuals with the lowest predicted probability of being employed in the sector will be selected out and added to the pool of "eligible" workers to be unemployed or employed in growing sectors (notice this pool also contains those who have been reassigned from outside the labor force if the total participation rate is growing). Reassignments out of each sector will continue up to the point where the change in the sector employment share at the micro level matches the change at the macro level. Step 3 ­ Flows into growing sectors: Individuals in the pool of "eligible" workers will be assigned to growing sectors on the basis of their predicted probability of being employed in each sector. Assignments are made sequentially with the sector whose employment share is growing fastest absorbing workers first and the sector whose employment share is growing the slowest absorbing workers last. Reassignments to each sector will continue up to the point where the change in the sector employment share at the micro level matches the change at the macro level. Step 4 ­ Unemployment: Individuals that remain in the pool of "eligible" workers after changes in sectoral employment have been accounted for are classified as unemployed. In other words, unemployment functions as the adjustment variable once predicted changes in the labor force participation and employment rates have been simulated.8 There are a few important features of this process that are worth mentioning. The reassignments described in steps 1 to 3 are calibrated so as to replicate net aggregate flows between employment and nonemployment and across sectors. In reality, movements across these different states are quite significant so that gross flows are usually larger than net flows. The order of proposed steps is such that it allows for nonemployed individuals to become employed and employed individuals to become non employed, but also for individuals to change sectors. In doing this we try to capture the fact that highly "employable" individuals are more likely to remain employed in one sector or another, at times at the expense of less "employable" workers (i.e. highly "employable" workers will crowd others out when employment opportunities become relatively more scarce). We next use the earnings model estimated as part of the calibration to predict earnings for two groups of workers: those with no previous earning history (i.e. nonemployed in 2006) and those who change sector of employment. Because earnings are a function of both observable and unobservable individual 8 In practice this means that projected unemployment figures may appear somewhat out of line with historical data. For this reason, we often present the simulation results in terms of employment and nonemployment, which is a perfectly valid classification from the point of view of examining poverty and distributional impacts given that only those employed receive labor income. 8 and job characteristics, we add a random element to the predicted earnings produced by the model to account for unobserved heterogeneity.9 For all other individuals, we use the earnings information available in the household survey. Once all workers have been assigned positive labor earnings, total earnings in a sector are adjusted to match aggregate projected changes in output. This step relies on the fact that that projected changes in sectoral output can be explained by projected changes in sectoral employment and projected changes in sectoral earnings and profits, and assumes that earnings and profits grow at the same rate. The treatment of public sector workers and those with more than one job differs slightly from what we just described. Total public sector employment is assumed to remain constant (i.e. no individuals are assigned to or out of the public sector) and labor earnings of public sector workers are adjusted in line with their sectoral mapping (agriculture, industry or services). Similarly for those holding more than one job, we assume the sector of employment of their secondary activity remains constant while earnings are adjusted in line with sectoral changes. The third step in the simulation process pertains to changes in international remittances. As mentioned above, in the case of Philippines we simulate these changes following a very simple allocation rule. We calculate the total change in international remittances between 2006 and 20092010, using actual and projected data (this change is positive under both the crisis and the benchmark scenarios) and allocate the dividend as follows: (i) across regions, remittances are allocated proportionally to the 2006 regional distribution; (ii) among households within regions: recipient households are selected at random and given a remittance transfer equivalent in real terms to the average remittance transfer in that region in 2006, with the number of total transfers to be made within each region being equal to the total amount of additional remittances to be distributed divided by the 2006 average value. As a result of this process the overall distribution of remittances across regions remains unchanged; while there is an increase in number of recipient households (both types of households, those who did or did not receive remittances in 2006, can receive new remittances in 200910). Finally we simulate changes in other sources of nonlabor income. For this we assume that capital and financial income grow at the same rate as real GDP, public transfers (primarily safety net and Social Security transfers) remain constant in real terms at 2006 levels, and domestic remittances change at the same rate as labor income. These assumptions appear to be reasonable for Philippines, and can be modified for other countries depending on country circumstances. Assessment of Impacts. Impact assessment is the process by which we use the information on individual employment status and labor income, as well as on nonlabor income at the household level, to generate income distributions and construct various poverty and distributional measures that can then be used to compare the crisis and the benchmark scenarios. This is done in three steps. 9 Specifically, we draw an individual error from the error distribution generated during the estimation of the earnings equation. 9 First, we account for the fact that between 2006 and 20092010, food prices have increased at a much higher rate than other prices. To do this we adjust the 2006 poverty line by the expected food and non food CPI in 2009 (or 2010), using the food and nonfood shares in the NCR poverty line as weights (approximately 57% and 43%, respectively). Since these weights are different from those in the CPI, the adjustment reflects the poverty line required for an individual to be able to afford the same food basket despite the disproportionate rise in food prices. This method yields a 2009 (or 2010) poverty line, which we then readjust to 2006 prices using the appropriate general CPI (see Annex 2, Table C). Second, we calculate total household income by aggregating labor income across all employed individuals and adding nonlabor income, and then use information on household size to construct a measure of percapita household income, as in (3). km * * NL wi I i ym * (3) PCI m * i 1 km Finally, we use information on household and individual income levels to evaluate the poverty and distributional impacts of the crisis by comparing poverty and other outcomes between the benchmark (without crisis) and "with crisis" scenarios. It is important to note that the Philippines poverty line is defined in terms of per capita income, which implies that per capita household income derived from (3) above can be directly compared with the poverty line to obtain poverty measures. Limitations and caveats of the simulation exercise The proposed approach has some appealing features, the primary one being its capacity to generate income counterfactuals at the individual and household levels that can then be used to assess impacts across the entire distribution. However, it also has some important limitations that must be taken into account when interpreting the results presented below. We discuss these below. Firstly, the quality and accuracy of the simulation output is a function of the nature and quality of data underpinning the exercise. The results would depend not only on the micromodels, but also on the macro projections of the crisis and the benchmark or nocrisis scenarios. In a typical exante assessment of this type, building the counterfactual to evaluate impact is especially tricky because the comparison between the situations "with" and "without" the "treatment" (the macro shock) is purely virtual or notional. This is particularly important with regard to the output and employment projections since they are key drivers of the results in the absence of a CGE or similar model. In addition: The ability to account for heterogeneity across sectors, groups, and others depends on the level of disaggregation of the available macro projection. For instance, the behavior of the tradeable and nontradable sectors within industry can only be modeled separately if output growth projections are available for each sector. The ability to accurately predict employment and earning changes depends on the available information and on the assumptions needed to correct for information gaps. For instance, in the absence of projections on total and sectoral earnings growth, we need to assume that earnings 10 and profits grow at same rate within a sector. How realistic this assumption is would depend on the country and sectoral context. The ability to directly model and predict changes in the relative shares of formal and informal employment within sectors is limited by the lack of projections on the growth of each type of employment. Instead, formal and informal status are assigned in proportion to what is observed in 2006 for each sector and consequently all compositional changes in sectoral employment are driven by population growth and individual employment transition, rather than more general labor market cyclical dynamics. The ability to model remittances depends on the quality of the available data on migrants and remittances, particularly for countries with rapid and/or volatile growth of remittances. Secondly, the simulations implicitly assume that the structural relationships estimated as part of the calibration process on the baseline data continue to be valid in the future years for which the projections are made. The more distant in the past the baseline year is, the more questionable this assumption is likely to be. In the case of Philippines, for example, the baseline year is 2006, which is a full 3 to 4 years away from the prediction years (2009 and 2010). This particular caveat, however, links directly back to the constraints imposed by availability of data. In most countries, household survey data that is available for analysis and processed to the extent necessary for the analysis is likely to predate the prediction years (usually 2010) by at least 3 to 4 years. 10 Thirdly, there is no mobility of factors (labor or capital) across regions and across urban and rural areas under the current model formulation. The same is true about international migration. Consequently all individuals are assumed to remain in their 2006 place of residence, irrespective of whether they experience a change in labor force status or not. Fourthly, our model is also limited in its ability to account for shifts in relative prices between different sectors of the economy as a result of the shock. The model does take into account the impact of shifts in the price of food relative to general inflation on poverty measures, via the simple device of adjusting poverty lines that are anchored to a fixed food basket. It can be argued that changes in the relative price of food represent the most significant source of price impact on poor or nearpoor households in a country like Philippines, as the recent food crisis has shown. At the same time, there are other potential sources of price impacts ­ for example, the effect of a change in the terms of trade between agriculture and other sectors on real household incomes in all sectors ­ our approach does not take into account. Unfortunately, in the absence of a CGE model to link up to, it is nearly impossible to explicitly model for changes in terms of trade between sectors. The final limitation, related to validation of hypothesis, applies to all exante approaches including ours. The only validation or test for our simulation model is to combine exante and expost analysis (see Bourguignon and Ferreira, 2003). Since ex post data will not be available for some time, some 10 An additional caveat, for some countries, relates to our decision to work with income rather than consumption data. This caveat however only applies to countries where poverty is measured against a poverty line defined in terms of consumption, rather than income. In the case of Philippines, since poverty is measured as a function of household income, our approach is entirely appropriate. 11 uncertainty about the simulations generated by the exante evaluations is bound to remain. However, a preliminary validation of the model using historical data suggests that the model performs well in predicting poverty and distributional changes (i.e. projected changes are very close to actual data). 4. Presentation of simulation results Before describing the impact on poverty, it is useful to trace the way the macro impacts are transmitted to the households. In the Philippines, given the nature of the macroeconomic impacts, the main transmission channel is the labor market. What did the Filipino labor market look like in the baseline year before the crisis? In 2006 ­ the last year for which household poverty data are available and therefore the baseline data for our projections ­ 69% of the workingage population (1564) was active in the labor force, of which about 91% was employed. Among those employed, 51% were employed in services, 34% in agriculture, 10% in manufacturing, and 6% in other industries. The employed population includes those working for wages/salaries, the selfemployed and those contributing labor to householdbased enterprises or activities that produce income for the household. On average, 67% of household income comes from labor earnings and 15% comes from international remittances, which form about 22% of nonlabor income (see Annex 2, Table B for detailed figures). The population growth rate between 2006 and 2010 is projected to be 8.2%, with a higher growth rate (10.8%) projected for the working age population. Employment growth projections for 2009 and 2010 are computed using past trends in output and employment and GDP projections shown in Table 1 above.11 Lower GDP growth rates in 2009 and 2010 due to the crisis would translate into employment impacts. Historically, the Table 2: Employment projections output elasticity of labor force participation (individuals aged 1564 yrs) has been slightly positive in the Philippines, % difference in levels between benchmark and crisis which is reflected in our results. Labor force 2009 2010 participation is slightly higher in the crisis scenario than in the benchmark scenario, Inactive 1.2 1.8 Active both in 2009 and 2010. Employment rates Unemployed 20.1 35.6 among the active population, however, Employed 1.0 1.8 show a different pattern. The overall employment rate in the crisis scenario is 1.0 Sectoral employment and 1.8 percent lower than in the Agriculture 1.0 1.4 benchmark scenario in 2009 and 2010 Manufacturing 0.5 0.6 respectively (see Table 2). Interestingly, the Other Industries 0.9 1.3 impact of the crisis on poverty is larger in Services 1.2 2.5 2010 than in 2009 despite a recovery in Source: Own calculations using GDP data and HIES (2000, 2005) economic growth will recover in 2010. This is due to the fact that the 2010 figure reflects the cumulative 11 See Annex 2, Tables A.2 and A.3 for detailed projections. 12 effect of the crisis in 2009 and 2010. Finally there is very little impact on the sectoral shares in employment, although each sector is worse affected in 2010 than in 2009. Aggregate impact on poverty and inequality Average household income is expected to be 2.2% lower in 2009 and 3.6% lower in 2010 as a result of the crisis, compared to the benchmark (nocrisis) scenario. The drop in 2009 income is mainly due to a 3.0% drop in labor income, along with a marginal (0.6%) drop in nonlabor income. In 2010, the larger impact is due to drops in both labor income (4.6%) and nonlabor income (1.9%). The significantly higher loss in nonlabor income in 2010 is due to a larger (2.3%) loss in remittances in 2010 compared to a 0.5% drop in 2009 (see Annex 2, Table B for the detailed estimates on household income). The poverty headcount rate is expected to be 1.45 and 2.07 Table 3: Poverty and inequality with and percentage points higher as a result of the crisis in 2009 without crisis Percentage point and 2010 respectively. The impacts on poverty gap and difference between severity are smaller (between 0.71 and 1.02 percentage crisis and benchmark points) for both years, but again slightly higher for 2010 scenarios than 2009. The impact of the crisis on aggregate measures 2009 2010 of inequality is negligible. Poverty Headcount rate 1.45 2.07 Interestingly, the impact on household income and poverty Poverty gap 0.71 1.02 Severity of is higher in 2010 than in 2009, in spite of the fact that the poverty 0.1 0.59 impact of the crisis on GDP growth is expected to be larger Inequality in 2009 (refer to Table 1, Section 2). There are two main (per capita exp.) Gini 0.001 0.001 explanations for this apparent puzzle. Firstly, remittances Theil 0.002 0.005 flows from abroad ­ which contribute significantly to a Source: Own calculations based on micro household's nonlabor income ­ show no decline in 2009 simulations using macro projections and FIES (2006) due to the crisis but are expected to be lower in 2010 due Box 1: Price adjustments underlying the poverty projections. All projected incomes are at 2006 prices. This would normally imply that no adjustments are necessary for the poverty line (defined at 2006 prices) as well, had it not been for the fact that food prices have risen at a faster rate than the general Consumer Price Index (CPI) since 2006 (by 19.3% as compared to the general CPI rise of 18.8%). Given that the poverty line is anchored to a fixed basket of food items necessary to meet the basic minimum calorie needs of an individual in 2006, a higher appreciation of food prices relative to general inflation would imply that the same poverty line may no longer be sufficient to purchase the basic food basket in 2009 or 2010.1 This would imply that the poverty line for 2009 and 2010 must be adjusted upwards (as described above) to reflect the same level of welfare as in 2006. Food prices ­ most significantly rice prices ­ in Philippines rose sharply during 2008 in response to the worldwide food price shock, which likely led to a temporary spike in poverty rate. Since rice prices have moderated since late 2008, the food CPI projections for 2009 and 2010 are likely to reflect the expected trend of prices rather than the temporary spike in 2008. That said, the poverty projections will change if the financial crisis were to have a further significant dampening (or upward) effect on domestic food prices, beyond what is reflected in the projected price indices used for our simulations. to the crisis (relative to the nocrisis scenario). Secondly, the impact on household income and poverty 13 in 2010, relative to the nocrisis scenario, is the result of the cumulative impact of the crisis on employment, labor and nonlabor incomes in both 2009 and 2010. Both these facts result in the income and poverty impact of the crisis being higher in 2010 than in 2009, and higher in 2010 than what the GDP growth impact in 2010 may suggest. How the poverty impact is spatially distributed ­ across regions and urban/rural areas The impact of the global crisis is not distributed uniformly across the Philippines. The crisis is expected to have a larger effect on urban centers with a large manufacturing base. Thus we expect to find greater impacts in the richer and highly urbanized regions of NCR, Calabarzon, and Central Luzon, which together account for nearly half of employment in the industry and services sectors, including the manufacture of electronic products. These regions also receive the most international remittances. Our simulation results conform to these expectations. The poverty headcount rate is higher by more than 10% due to the crisis (as a percentage of the poverty rate in the nocrisis or benchmark scenario) in NCR, Calabarzon, and Central Luzon (Figure 3). The poverty rate in the highly dense, urban region of NCR, which is home to Manila, is projected to be higher by about 25% due to the crisis. The regions of Visayas and Mindanao, which are more rural and dependent on agricultural employment, are less affected (lower than 7% poverty impact), despite the decline in the prices of copra and coconut oil, which are among the country's largest exports. Box 2: Poverty impact of the October 2009 typhoons The Philippines was hit by four typhoons in the month of October, 2009, two of which made landfall in the National Capital Region of Manila. All together, the typhoons had a devastating cumulative effect, causing heavy rain and winds, and massive floods which washed away shanty towns on the coast. Access to electricity was severely affected, and over 300,000 houses are believed to have been damaged. The typhoons are expected to result in lower output and greater poverty. Our microsimulation model is able to capture some of these effects by treating the typhoon impact as a macro shock. In order to do so, we use macroeconomic projections of the effects of the typhoons on output and employment to determine the householdlevel impacts. The baseline scenarios for 2009 and 2010 are the crisis scenarios as calculated by the model. Thus, we are able to determine the cumulative impact of the typhoons and the financial crisis as compared to 2006. Table 3a shows the poverty impacts of the typhoons according to this methodology. Poverty is expected to increase by 0.5 percentage points in 2009 and 0.7 percentage points in 2010. The increase according to the Poverty gap is larger in 2010 than in 2009. The impact on inequality, however, is negligible. 14 These differences also manifest themselves in the aggregate ruralurban poverty impacts, with the urban poverty rate rising much more than the rural poverty rate relative to their corresponding benchmark poverty rates (11.5% and 5.2% rise for urban and rural poverty rates, respectively) due to the crisis (Figure 3). The income impact in urban areas is also much larger ­ average urban household income with crisis is 4.3% lower than that in the benchmark while average rural household income in crisis is 2.5% lower than in the benchmark (see Annex 2, Table D.1). This difference is driven mostly by a change in labor income, which would decline relatively more in urban areas (5.6% of the benchmark) than in rural areas (3.0% of the benchmark). It is important to note that the absolute change in the poverty Figure 3: Change in poverty headcount due to the crisis in 2010 Poverty rate (upper PL) headcount rate is nearly identical across the regions and across urban and rural areas, at around 2 percentage points (Figure 3). The same absolute change in poverty headcount however translates into much higher percentage change in poverty in the richer, urbanized regions because these regions start with a lower headcount. Moreover, the same absolute increase in poverty rate Source: Own simulations based on FIES (2005). For projected would imply many more people in poverty rates, see Annex 2, Table D.2. poverty in the denser, more urbanized Note: Vertical axis measures the difference between crisis and areas than in the less densely benchmark poverty rates (as % of benchmark rate) populated areas of Other Luzon, Visayas, and Mindanao. However, even after taking into account the significantly higher impact of the crisis in the urban regions, the regional gaps in poverty and income levels would continue to remain large in 2010. This is because the gaps in 2006 ­ the baseline year for all the simulations ­ were large to start with (Figure 4 and Figure 5), and poverty reduction has been slow over the past few years. NCR, Calabarzon, and Central Luzon are projected to remain the (significantly) richer regions of the country, with Other Luzon, Visayas, and Mindanao remaining as the poor regions. 15 Figure 4: Poverty headcount rates for divisions (%) Figure 5: Poverty headcount rates for urban/rural 2006 estimates & 2010 projections 2006 estimates & 2010 projections Source: Estimates and simulations based on FIES (2006) Impact of the crisis by gender According to the simulation results, poverty will Table 4: Crisis Impacts by Gender increase relatively more among femaleheaded (% change between benchmark and crisis) households than among maleheaded 2009 2010 households due to the crisis ­ the impact on the Female Male Female Male Income headcount is 11.9% versus 7.6% in 2010 (Table Labor Income 4.4 2.7 6.1 4.4 4). Poverty rates will continue to be significantly Remittances 0.4 0.6 1.5 3.0 lower among femaleheaded households. These Foreign 0.2 0.5 0.7 2.1 differences can be attributed to slightly higher Domestic 3.2 3.2 4.9 4.9 declines in employment (across all sectors, but especially in services) and average household Earnings labor income among femaleheaded Mean per Worker 2.6 1.9 4.1 2.7 Agriculture 1.1 1.5 1.4 1.8 households. Manufacturing 9.5 10.9 12.8 12.9 Other Industries 4.4 5.0 7.4 7.4 In general, labor income is more negatively Services 1.8 1.2 2.9 2.0 impacted for femaleheaded households ­ in aggregate (household) terms as well as in terms Poverty of average earning per worker (Table 4). In Poverty Headcount 8.3 5.3 11.9 7.6 2009, this is mostly due to a larger loss in Poverty Gap 11.2 8.2 16.3 12.5 manufacturing earnings, whereas in 2010 the Source: Estimates and Simulations based on FIES (2005) impact is distributed more evenly across sectors. Distributional impact of the crisis: Going beyond poverty and inequality indices By generating predicted levels of income and consumption for all households in benchmark and crisis scenarios, our simulation model allows us to examine the type of households that are likely to be affected by the crisis, the primary channels of impact and their relative importance, and the distribution of the impact across different income groups. Here we present the results from three types of analysis 16 that have been selected primarily for illustrative purposes. We also present the results for 2010 only, when the relatively larger impact of the crisis is expected. Firstly, we examine the characteristics of the group we will call "crisisvulnerable", which refers to households that would not have been poor in 2010 had there been no crisis. Secondly, we will use the wellknown analytical device of growth incidence curves to see how change in income, as a result of the crisis, is distributed across income groups and between urban/rural areas and regions. Thirdly, we will construct a few transition matrices to look at upward and downward movement of households as a result of the crisis, compared to the benchmark. These are but examples of what is possible in the way of distributional analysis with the results of our model; the choice of what type of analytics needs to be done for a certain country would depend on the specific country context and policy concerns. A profile of the "crisisvulnerable". Households that are expected to be in poverty in the crisis scenario but not in the benchmark scenario in 2010, termed as the "crisisvulnerable", are projected to suffer large income losses due to the crisis, relative to the benchmark scenario ­ with a 39% drop in average household income, which translates to a 45% loss in per capita income. Most of this loss is due to a massive loss in labor income, amounting to 48% of the benchmark income. Remittances are less important as a component of the overall income loss of the "crisisvulnerable" households. Although there is a 33% loss in remittance income for households conditional on receiving remittances in the first place, the absolute amounts are too small to have a significant impact on poverty. The key characteristics of crisisvulnerable households, relative to the rest of the population, are as follows (see Annex 2, Table E for all results). Comparing the crisisvulnerable to the overall population, the crisisvulnerable are slightly more rural (57.1% as compared to the overall figure of 51.5%), have slightly larger household sizes (5.1 people per household as compared to the national average of 4.8) and have nearly identical dependency ratios. The crisisvulnerable have lower skills on average ­ 67.3% of household heads have education 09 years compared to the national average of 55.9%. Although the employment rates are very similar among the two groups, the sectoral composition is quite different. Crisisvulnerable household heads are more likely to be employed in agriculture and manufacturing, and less likely to be employed in services or other industry. This result further underscores our initial assertion that the shock is likely to manifest itself through lack of demand for exports in the manufacturing sector and commodity price decreases in the agricultural sector. It is also interesting to look at how these crisisvulnerable households compare with the permanently or structurally poor (defined as the households who are poor in both benchmark and crisis scenarios in 2010). In general, structurally poor households are larger, have higher dependency and are lower skilled. They are also far more likely to be employed in agriculture (64.0% of the structurally poor as opposed to 36.9% of the crisisvulnerable), and subsequently less likely to be employed in manufacturing (4.8% as compared to 11.3%) or services (27.4% as compared to 48.5%). This is understandable since those employed in manufacturing tend to be better off in general, even though they are the demographic most likely to be worse off as a result of the crisis. The crisisvulnerable households are also significantly less rural than the structurally poor households (see Annex 2, Table E). 17 The profile of crisisvulnerable households is consistent with our a priori intuition. The reduction in labor income is the primary channel of impact for households that are predicted to become poor as a direct result of the crisis. Compared to the general population, households vulnerable to the crisis are disproportionately employed in manufacturing because output and income losses are expected to occur in these sectors. They are also less skilled than the general population, but quite similar in terms of dependency, household size, and ruralurban composition. The crisisvulnerable or the "newly poor" households as a result of the crisis are also different in a number of key aspects from the more permanently poor. Compared to the structurally poor, they have "better" characteristics (lower dependency, higher skills, living in urban areas and employed outside agriculture) ­ characteristics that our simulations indicate would have taken them out of poverty had there been no crisis. Figure 6: Growth incidence curves (crisis vs. benchmark, 2010) (a) Per capita income: Philippines (b) Per capita income: urban and rural 0.00 0.00 -1.00 -1.00 -2.00 -2.00 -3.00 -3.00 % Change % Change -4.00 -4.00 -5.00 -5.00 -6.00 -6.00 -7.00 -7.00 -8.00 -8.00 0 10 20 30 40 50 60 70 80 90 100 0 10 20 30 40 50 60 70 80 90 100 Percentile Percentile Urban Rural Source: Own simulations based on FIES (2006) Note: 1) GICs have been smoothed to eliminate sharp kinks in the original curves 2) Percentiles of income on the horizontal axis are for the relevant group (all country, urban, rural, etc.) How income losses between benchmark and crisis scenarios are distributed. This analysis is similar to the familiar growth incidence curves (GICs), in examining how the changes in per capita income between benchmark and crisis scenarios in 2010 vary across the income distribution. GICs are commonly used to look at the distribution of growth over time; however the same device is useful to see changes between two "states" of the world as well, in this case the benchmark (nocrisis) and crisis states for the same point in time. A GIC in this case will plot the percentage change in per capita income between 2010 benchmark and crisis scenarios for every centile (1% of the population) in the benchmark distribution of per capita income (when the households are ordered by per capita income). For the country overall, the GIC of loss in per capita income (as a percentage of the benchmark income) shows significant losses (ranging from 3% to just above 4%) for households the 10th and 90th percentiles of the income distribution (Figure 6a). The largest losses (47%) occur for the bottom 10 percent of the distribution, but this finding is of less significance since the errors in measuring income can be high for 18 the lowest and highest ranges of the distribution. More significant is the fact that the losses are higher for 50th and 90th percentiles than for the 10th to 50th percentiles, indicating that the middle and upper middle classes suffer somewhat higher losses than poorer groups (with the exception of the bottom decile). Looking at how the impact is distributed within urban and rural areas, there are some important differences between the two areas (Figure 6b). The urban incidence curve shows higher impacts for the bottom 10% of the distribution and lower impacts above the 70th percentile, while the rural incidence curve is quite flat, with marginally higher impacts seen only for some of the poorest (below 10th percentile) and wealthy (70th to 90th percentile) groups. It is important to note that the percentiles of expenditures are defined with reference to each area (urban or rural). Since urban areas are better off on the average than rural areas, the kth percentile of urban households is better off than (and thus not strictly comparable with) the kth percentile of rural households. These results imply that the urban poor, nearpoor and middleclass (up to the 70th percentile of the urban distribution) are more adversely affected than richer urban households, while the impacts are more evenly distributed among rural households. Notably, the impact is significantly higher among urban households than among rural households for the entire range of the distribution, consistent with earlier results showing that the poverty impact is relatively higher in urban areas (refer to Figure 3). Figure 7: Regional growth incidence curve (crisis vs. benchmark, 2010) Per capita income: Philippines -2.00 -3.00 -4.00 -5.00 % Change -6.00 -7.00 -8.00 -9.00 -11.00 -10.00 0 10 20 30 40 50 60 70 80 90 100 Percentile NCR Central Luzon Calabarzon Others Source: Own simulations based on FIES (2006) Note: 1) GICs have been smoothed to eliminate sharp kinks in the original curves 2) Percentiles of income on the horizontal axis are applicable to the relevant group (in this case, region) The regional GIC (Figure 7) reinforces our earlier result about urban poverty in the NCR, Central Luzon, and Calabarzon. Especially in NCR, the poorest are by far the worst affected, with the effects sharply declining as one moves to the top of the income distribution. Calabarzon shows a similar trajectory, although there the middle class and the poor appear to face similar impacts, and in Central Luzon, it is primarily the middle class who bear the brunt of the crisis. 19 Movements of households up or down the distribution due to the crisis. The GICs described above provide a clear picture of the incidence of the losses in income due to the crisis across the entire distribution. They also lead to some important followup questions from a distributional perspective: where would the households suffering these losses end up relative to other households as a result of the crisis, and what is the extent of "churning" that occurs in the distribution as a result of the crisis? Transition matrices are useful to address such questions, by indicating movements up or down the income distribution between the crisis and benchmark scenarios. These matrices are constructed for deciles of per capita income and labor income, keeping the upper and lower limits of each decile fixed at the 2010 benchmark income levels. This implies that movements up or down by households in these matrices reflect the shifts that occur as a result of the crisis, relative to the benchmark (nocrisis) income distribution (see Annex 2, Table F). Figure 8 shows the share of households in each decile that (i) remain in the same decile, (ii) move up to a higher decile or (iii) move down to a lower decile as a result of the crisis. Summing across all deciles, 84.5% of households are found to remain in the same decile by per capita income, while 83.4% do so for deciles by per capita labor income. Given that the deciles are fixed at the benchmark level and the crisis yields an overall loss of income, movements down are much more common than movements up. The most significant downward movements occur for households in the middle of the distribution, namely in the 5th8th deciles by per capita total income, with the largest movement occurring for those in the 7th decile. The movements between deciles of per capita labor income are smaller, with the largest movements occurring for the 7th and 8th deciles. Upward movements are extremely rare. Figure 8: Movements up and down the distribution due to crisis compared to the benchmark in 2010 Source: Own simulations based on FIES (2006) Thus the highest amount of "churning" occurs in the middle part of the distribution (5th to 8th deciles), with downward movements being more common for the middle and uppermiddle class relative to the poor and the richest. Between 15.5% and 17.5% of households in the 5th to 8th deciles of per capita income move down to a lower decile as a result of the crisis. Labor income losses (that are significant enough to lead to downward movements) show a very similar outcome to the total income losses, because, as discussed earlier, they are the largest drivers of overall income loss. 20 5. Conclusion: Implications of our results Even with the caveats and limitations discussed in Section 3, our model appears to yield reasonable and intuitively appealing results, based on a transparent and flexible approach using macro projections and household survey data that are typically available in a developing country. The value that our model adds to the simpler models (e.g. POVSTAT or elasticity based methods that are most commonly used in the World Bank) is the ability to analyze potential distributional impacts in detail, linking these changes to the channels through which the impact of the crisis is likely to flow. The application of our model to Philippines yields some useful insights relevant to the country, in three main areas discussed below: monitoring of impacts of the crisis, providing deeper insights into who are likely to be affected by the crisis, and how that in turn can inform the design of policies to mitigate impacts. Firstly, our results suggest that in the Philippines, a select few indicators are prime candidates to monitor as "leading" or realtime indicators for the poverty and distributional impact of the financial crisis, or of future shocks with macroeconomic impacts similar to this crisis. To be useful for rapid monitoring of poverty and distributional impacts, the indicators must be easily obtained and measurable, sensitive to changes in economic conditions, and correlated with changes in poverty and distribution. By these criteria, movement of wages and employment in manufacturing and services, changes in remittance flows from abroad and the movement of prices (particularly food prices) can serve as useful indicators for rapid monitoring of economic shocks in Philippines. It is important to notice, however, that data on wages is only collected in the October round of the Labor Force survey and made available with a large time gap. Inclusion of monitoring questions on wages in other LFS quarters, together with timelier processing and publication of the information could significantly enhance the government's capacity to monitor potential poverty and distributional impacts of macroeconomic shocks. Why these indicators? From our analysis, the labor market emerges as the major source of impact in Philippines, with the effects being transmitted through loss of employment in certain sectors (primarily manufacturing) and lower labor earnings as output growth slows. Another channel of impact is remittance flows, since remittance growth in 2010 is expected to fall below what was expected before the crisis had occurred. Any rapid, unexpected change in food (mainly rice) prices could also have a significant impact on poverty ­ at least in the shortrun when wages would lag behind food prices. These indicators also have the advantage of being relatively easy to monitor ­ with administrative or financial data (for aggregate remittance flows), and quick market or household surveys (prices and wages). Tracking wages as a part of rapid monitoring would be particularly important. First, growth slowdown in a country like Philippines is more likely to be manifested in lower earnings rather than open unemployment ­ a hypothesis supported by our results as well. Second, while wages represent only a part of the labor earnings from a sector, wages are much easier to track (e.g. through quarterly labor force or quick enterprise surveys) than income from household enterprises and selfemployment. Wages are also likely to be reasonably correlated with broader measures of labor income, particularly in the manufacturing sector where most of the labor market impact is expected. 21 Secondly, our results indicate that just focusing on poverty numbers would provide a partial view of the distributional impact of the crisis in Philippines. It is useful here to revisit some of the key distributional results. The adverse impact on income, while occurring throughout the income distribution, is somewhat larger for those in the 50th to 90th percentile of the national distribution ­ a group that can be characterized as the middleclass. Also, underlying the national results are important urbanrural and spatial differences in the magnitude and distribution of the impact. Urban households are projected to suffer relatively larger losses than rural households, with the result that income losses and poverty impact would be higher in regions that are more urbanized (NCR, Central Luzon and Calabarzon). Within urban areas the incidence of loss would be higher for the poor and the middleclass (the bottom 70 percent of the urban distribution), while rural losses are more evenly distributed. Thus the urban poor and nearpoor would appear to be key groups of concern from a public policy perspective, as the groups most likely to slip into (or deeper into) poverty. While public programs would have little role to play in directly mitigating the losses suffered by the middle and uppermiddle classes, these losses may have important political economy implications. The fact that the impact is relatively high for the urban middleclass (40th 70th percentile of the urban distribution) could be particularly significant from a political economy perspective. Finally, our results provide useful insights on the types of households who are likely to be poor as a direct result of the crisis. Such households are found to have characteristics somewhat different from those of the structurally poor ­ more likely than the structurally poor to be urban and employed in non agricultural sectors, and with higher skills and fewer dependents in households. This would imply that any intervention to mitigate the impact of the crisis on vulnerable households would need to be modified (or designed differently) from that of the traditional safety net programs. Also, the income and poverty impacts are somewhat higher for households headed by women than for those headed by men. While more analysis needs to be done to fully understand the source of this gender difference, the seemingly higher vulnerability of femaleheaded households to the crisis is important to take into account in gauging the impact of the crisis and considering policy responses. Apart from these somewhat general statements, a discussion of what kinds of public policies and programs are best suited to mitigate the impact of such a macroeconomic shock in Philippines is beyond the scope of this paper. The results from our simulation exercise, however, can be useful in informing such discussions, at least till the time when actual survey data become available on who are being impacted by the macroeconomic shock and to what extent. 22 References Alatas, V., and F. Bourguignon (2005). "The evolution of Income distribution during Indonesia's fast growth, 198096" in The microeconomics of income distribution dynamics in East Asia and Latin America, eds. Bourguignon, F., Ferreira, F., and Lustig, N., Washington DC: Oxford University Press and The World Bank. Beach, Ch. (1989). "Dollars and Dreams: A reduced middle class? Alternative Explanations", Journal of Human Resources, 24, 162193 Blackburn, M. and D. Bloom (1987). "Earnings and income inequality in the United States", Population Development Review, pp. 575609 Bourguignon, F. and F. Ferreira (2005). "Decomposing Changes in the Distribution of Household Incomes: Methodological Aspects" in The microeconomics of income distribution dynamics in East Asia and Latin America, eds. Bourguignon, F., Ferreira, F., and Lustig, N., Washington DC: Oxford University Press and The World Bank. Bourguignon, F. and F. Ferreira (2003). "Exante evaluation of policy reforms using behavioral models" in The Impact of Economic Policies on Poverty and Income Distribution, Bourguignon, F. and Pereira da Silva, L. (eds), World BankOxford University Press. Bourguignon, F., M. Bussolo and L. Pereira da Silva (2008). "Introduction: Evaluating the impact of Macroeconomic Policies on Poverty and Income Distribution", in The Impact of Macroeconomic Policies on poverty and Income Distribution, eds. Bourguinon, F., Bussolo, M. and Pereira da Silva, L. Washington DC, The World Bank. Burkhauser, R., A. Crews, A. Daly and S. Jenkins (1999). "Testing the significance of income distribution changes over the 1980s business cycle: a crossnational comparison", Journal of Applied Econometrics, 14, pp 253272 Chen, S. and M. Ravallion (2009). "The Impact of the Global Financial Crisis on the World's Poorest". World Bank Development Research Group. Deaton, A. (2001). Intervention in a panel discussion on the International Monetary Fund ­ sponsored conference on "Macroeconomic Policies and Poverty Reduction" April, 13, 2001, Washington D.C; transcript at http://www.imf.org/external/np/tr/2001/tr0104143.htm Ferreira, F., P. Leite, L. Pereira da Silva and P. Picchetti (2008). "Can the Distributional Impacts of Macroeconomic Shocks Be Predicted? A Comparison on TopDown MacroMicro Models with Historical Data for Brazil" in The Impact of Macroeconomic Policies on poverty and Income Distribution, eds. Bourguinon, F., Bussolo, M. and Pereira da Silva, L. Washington DC, The World Bank. Jenkins, S. (1995). "Did the middle class shrink during the 1980s? UK evidence from kernel density estimates", Economics letters, 49, 407413 23 Levy, F. and R. Murnane (1992). "U.S. Earnings levels and earnings inequality: A review of recent trends and proposed explanations", Journal of Economic Literature, Vol XXX Paci, P. and M. Sasin (2009). "Making Work Pay in Philippines: Employment, Growth, and Poverty Reduction". World Bank Directions in Development: Poverty. Paci, P., B. Rijkers, and A. Seinaert (2008). "Crunch Time in the Developing World Too? Policy options for dealing with the potential poverty and distributional impact of the financial crisis". World Bank Poverty Reduction Group, PRMVP. Ravallion, M. (2009). "The Developing World's Bulging (But Vulnerable) Middle Class". Policy Research Working Paper 4816. World Bank. Washington DC Vos, R., L. Taylor, L and R. Paes de Barro eds. (2002) Economic liberalization, distribution and poverty. Latin America in the 1990's. UNDP World Bank (2007). Philippines: Strategy for Sustained Growth. Philippines Development Series. Paper No. 18. Dhaka. World Bank (2008). Poverty Assessment for Philippines: Creating Opportunities and Bridging the East West Divide. Report No. 44321BD. Washington DC. World Bank (2009a). Sailing Through Stormy Waters. Philippines Quarterly Update. Washington DC. World Bank (2009b). "Global Economic Crisis and Household Vulnerability in the Philippines: Potential Impacts and Policy Responses". East Asia Human Development Group. Washington DC. World Bank (2009). Welfare Impact of Rising Food Prices in South Asian Countries. Policy Note, South Asia Economic Policy and Poverty Unit. Washington DC. World Bank PovertyNet website. "PovStat" (Accessed August 12, 2009). http://web.worldbank.org/WBSITE/EXTERNAL/TOPICS/EXTPOVERTY/EXTPSIA/0,,contentMDK:20483029 ~isCURL:Y~menuPK:1108016~pagePK:148956~piPK:216618~theSitePK:490130,00.html 24 Annex 1 OCCUPATIONAL DECISION MODEL FOR THE PHILIPPINES Individuals between 15 and 64 years old Low skills (*) High skills (**) Unemployed Agric Manuf. Other Ind. (1) Services Unemployed Agric Manuf. Other Ind. (1) Services male 1.000*** 2.132*** 1.073*** 4.200*** 0.187** 1.095*** 2.444*** 0.789*** 4.713*** 0.669*** (0.122) (0.0846) (0.148) (0.512) (0.0915) (0.0933) (0.101) (0.103) (0.319) (0.0773) hhead 1.351*** 0.664*** 0.584*** 0.0696 0.587*** 1.898*** 0.817*** 1.213*** 0.131 0.885*** (0.294) (0.145) (0.224) (1.116) (0.135) (0.381) (0.224) (0.194) (0.800) (0.146) age 0.0737*** 0.190*** 0.235*** 0.322*** 0.231*** 0.106*** 0.304*** 0.360*** 0.426*** 0.339*** (0.0112) (0.00795) (0.0132) (0.0149) (0.00823) (0.0119) (0.0117) (0.0137) (0.0181) (0.00895) age2 0.000855*** 0.00218*** 0.00285*** 0.00401*** 0.00274*** 0.00163*** 0.00360*** 0.00479*** 0.00531*** 0.00424*** (0.000142) (9.77e05) (0.000163) (0.000185) (0.000101) (0.000157) (0.000149) (0.000179) (0.000229) (0.000114) pric 0.171* 0.201*** 0.175* 0.688 0.224*** (0.0941) (0.0464) (0.0898) (0.478) (0.0479) seci 0.0173 0.422*** 0.292*** 0.0555 0.286*** (0.0912) (0.0507) (0.0925) (0.405) (0.0493) gen_pric 0.234* 0.249*** 0.155 1.089** 0.152* (0.132) (0.0852) (0.133) (0.486) (0.0915) gen_seci 0.00931 0.0579 0.0581 0.210 0.162* (0.121) (0.0795) (0.128) (0.413) (0.0833) married 2.215*** 0.728*** 1.044*** 1.332*** 1.294*** 2.356*** 1.235*** 1.752*** 1.723*** 1.670*** (0.0897) (0.0643) (0.108) (0.468) (0.0595) (0.0693) (0.0807) (0.0741) (0.238) (0.0495) remitt 0.231*** 0.0230 0.0600 0.275 0.109*** 0.168*** 0.00462 0.102* 0.445** 0.0476 (0.0809) (0.0414) (0.0769) (0.389) (0.0405) (0.0601) (0.0621) (0.0617) (0.226) (0.0370) depen 1.091*** 0.334*** 0.0128 0.429*** 0.343*** 1.186*** 0.234** 0.217** 0.347** 0.145** (0.114) (0.0710) (0.119) (0.128) (0.0737) (0.101) (0.0945) (0.107) (0.136) (0.0717) perce 0.00566 3.626*** 4.658*** 4.446*** 4.439*** 0.000885 4.497*** 5.714*** 5.711*** 5.379*** (0.0856) (0.0640) (0.106) (0.112) (0.0677) (0.0784) (0.0861) (0.0980) (0.121) (0.0684) oth_pub 0.0682 1.318*** 1.051*** 1.126*** 0.113* 0.0260 1.241*** 1.105*** 1.139*** 0.0563 (0.0963) (0.0709) (0.126) (0.133) (0.0594) (0.0635) (0.0677) (0.0817) (0.102) (0.0427) enrolled 4.237*** 2.628*** 3.274*** 4.512*** 2.868*** 4.641*** 3.002*** 3.787*** 3.952*** 3.129*** (0.140) (0.0638) (0.170) (0.343) (0.0755) (0.132) (0.101) (0.177) (0.312) (0.0686) gen_hhd 1.012*** 0.748*** 1.335*** 1.923* 1.506*** 0.702 0.0457 0.469 1.576* 0.811*** (0.381) (0.241) (0.330) (1.138) (0.242) (0.444) (0.301) (0.292) (0.832) (0.232) gen_rem 0.193* 0.0914 0.186 0.103 0.119 0.151 0.124 0.0242 0.475* 0.105 (0.113) (0.0717) (0.127) (0.400) (0.0809) (0.0960) (0.0986) (0.110) (0.246) (0.0785) hhd_rem 0.0286 0.0705 0.249 1.103 0.0491 0.728* 0.199 0.461** 0.204 0.134 (0.321) (0.156) (0.237) (1.194) (0.144) (0.405) (0.251) (0.225) (0.917) (0.156) gen_hhd_rem 0.123 0.487** 0.885*** 1.589 0.528*** 0.407 0.171 0.0302 0.566 0.282 (0.366) (0.203) (0.290) (1.204) (0.199) (0.437) (0.284) (0.269) (0.929) (0.199) hhd_married 1.602*** 0.192 0.626** 0.632 0.638*** 0.735*** 0.610*** 0.463** 1.365 0.757*** (0.347) (0.191) (0.269) (1.225) (0.164) (0.277) (0.228) (0.221) (0.843) (0.133) gen_married 2.587*** 1.373*** 1.941*** 2.364*** 2.399*** 2.540*** 1.367*** 2.447*** 2.390*** 2.385*** (0.176) (0.143) (0.192) (0.490) (0.144) (0.120) (0.127) (0.131) (0.262) (0.103) hhd_marr_gen 1.797*** 0.695** 1.636*** 1.775 1.872*** 0.664* 0.691** 1.294*** 2.095** 1.677*** (0.442) (0.294) (0.390) (1.252) (0.284) (0.362) (0.314) (0.324) (0.878) (0.235) urban 0.115** 1.381*** 0.00443 0.0800 0.283*** 0.0968** 1.406*** 0.179*** 0.139** 0.0680** (0.0556) (0.0377) (0.0556) (0.0594) (0.0353) (0.0477) (0.0446) (0.0514) (0.0628) (0.0338) supi 0.199*** 0.293*** 0.178** 0.851** 0.128*** (0.0669) (0.0693) (0.0708) (0.336) (0.0412) supc 0.835*** 0.567*** 0.199*** 2.328*** 1.190*** (0.0708) (0.0929) (0.0768) (0.282) (0.0448) gen_supi 0.444*** 0.697*** 0.576*** 1.839*** 0.582*** (0.0969) (0.0955) (0.105) (0.345) (0.0725) gen_supc 0.563*** 0.621*** 0.581*** 3.347*** 1.073*** (0.114) (0.132) (0.125) (0.301) (0.0921) Constant 1.834*** 7.109*** 7.243*** 11.61*** 5.484*** 1.543*** 9.935*** 8.016*** 13.51*** 6.493*** (0.220) (0.225) (0.269) (0.575) (0.161) (0.203) (0.274) (0.242) (0.446) (0.160) Regional Dummies Yes Yes Yes Yes Yes Yes Yes Yes Yes Yes Observations 55315 55315 55315 55315 55315 54842 54842 54842 54842 54842 Pseudo R2 0.314 0.314 0.314 0.314 0.314 0.316 0.316 0.316 0.316 0.316 Notes: National Capital Region (NCR) is the base region; Inactive is the base category (1) Other industries include Mining & quarrying; Electricity, Gas & Water and Construction (*) Low skills includes individuals with completed secondary school; (**) High skills includes individuals without a secondary school education Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 Source: Own estimations based on FIES 2006 i Variable definitions for Occupational Decision Model Variable Definition male =1 if male hhead =1 if household head age Age of the individual age2 Age squared pric =1 if primary education completed seci =1 if secondary education incomplete married =1 if married remitt =1 if household receives remittances (foreign or domestic) depen dependency ratio between number of members <15 or >64 and total number of members in the household perce ratio between number of income preceptors in the household 1 and the number of potential income preceptors within the hhd oth_pub =1 if there is another member of the household with a public job enrolled =1 if individual is currently attending school urban =1 if living in an urban area supi =1 if teritary education incomplete supc =1 if tertiary education complete gen_pric; gen_seci; gen_hhd; Interaction terms gen_rem; hhd_rem; gen_hhd_rem; hhd_married; gen_married; hhd_marr_gen; gen_supi; gen_supc ii LOG EARNING EQUATIONS (Individuals between 15 and 64 years old) Low skills (*) High skills (**) Agric Manuf. Other Ind. (1) Services Agric Manuf. Other Ind. (1) Services male 0.497*** 0.356*** 0.399** 0.322*** 0.425*** 0.128*** 0.394*** 0.120*** (0.0272) (0.0646) (0.174) (0.0297) (0.0471) (0.0368) (0.104) (0.0174) hhead 0.439*** 0.182** 0.949** 0.210*** 0.131 0.0560 0.156 0.0618** (0.0413) (0.0887) (0.388) (0.0346) (0.0916) (0.0704) (0.319) (0.0261) age 0.0810*** 0.0884*** 0.0727*** 0.0725*** 0.106*** 0.0780*** 0.0758*** 0.0726*** (0.00368) (0.0102) (0.0102) (0.00450) (0.00821) (0.00857) (0.0123) (0.00371) age2 0.00102*** 0.00108*** 0.000887*** 0.000832*** 0.00129*** 0.000914*** 0.000878*** 0.000760*** (4.37e05) (0.000125) (0.000126) (5.49e05) (0.000102) (0.000114) (0.000157) (4.69e05) pric 0.0816*** 0.248*** 0.104** 0.104*** (0.0165) (0.0523) (0.0429) (0.0229) seci 0.0974*** 0.325*** 0.222*** 0.183*** (0.0192) (0.0521) (0.0435) (0.0228) supi 0.209*** 0.144*** 0.136*** 0.223*** (0.0345) (0.0346) (0.0452) (0.0159) supc 0.535*** 0.736*** 0.953*** 0.916*** (0.0512) (0.0380) (0.0581) (0.0151) urban 0.0228 0.217*** 0.0518 0.0273 0.00687 0.169*** 0.0772* 0.00941 (0.0211) (0.0466) (0.0381) (0.0198) (0.0353) (0.0359) (0.0417) (0.0146) pub_job 0.144*** 0.0608*** (0.0391) (0.0195) sala 0.279*** 0.315*** 0.662*** 0.107*** 0.280*** 0.479*** 0.395*** 0.306*** (0.0165) (0.0497) (0.0574) (0.0195) (0.0335) (0.0445) (0.0658) (0.0150) gen_hhd 0.0925** 0.403*** 0.463 0.207*** 0.274*** 0.361*** 0.726** 0.334*** (0.0463) (0.106) (0.389) (0.0449) (0.0974) (0.0789) (0.322) (0.0314) Constant 5.927*** 5.668*** 5.845*** 6.484*** 6.035*** 6.463*** 6.849*** 6.651*** (0.173) (0.220) (0.262) (0.0975) (0.251) (0.164) (0.259) (0.0737) Regional Dummies Yes Yes Yes Yes Yes Yes Yes Yes Observations 15757 2076 2120 10075 5057 3530 1945 20940 Rsquared 0.278 0.318 0.262 0.198 0.240 0.289 0.327 0.301 Adj Rsquared 0.277 0.310 0.253 0.195 0.236 0.284 0.318 0.300 Notes: National Capital Region (NCR) is the base region (1) Other industries include Mining & quarrying; Electricity, Gas & Water and Construction (*) Low skills includes individuals with completed secondary school; (**) High skills includes individuals without a secondary school education For lowskilled workers, primary incomplete is the base level; For highskilled workers, secondary complete is the base level Standard errors in parentheses *** p<0.01, ** p<0.05, * p<0.1 Source: Own estimations based on FIES 2006 iii Variable definitions for Earning Equation Variable Definition male =1 if male hhead =1 if household head age Age of the individual age2 Age squared pric =1 if primary education completed seci =1 if secondary education incomplete supi =1 if teritary education incomplete supc =1 if tertiary education complete urban =1 if living in an urban area pub_job =1 if individual works in the public sector (govt. org., public factory, or local govt.) sala =1 if individual is a salaried worker gen_hhd =1 if male and household head iv Annex 2 Table A.1: POPULATION GROWTH CHANGE (in millions) 014 1564 +65 Total Total 30.15 53.18 3.60 86.93 2006 Male 15.40 26.71 1.60 43.72 Female 14.74 26.46 2.00 43.20 Total 30.90 57.33 3.94 92.17 2009 Male 15.79 28.80 1.75 46.34 Female 15.12 28.53 2.18 45.83 Total 31.16 58.80 4.06 94.01 2010 Male 15.91 29.54 1.81 47.26 Female 15.24 29.26 2.25 46.75 Total 3.35 10.56 12.79 8.15 % Male 3.32 10.59 12.94 8.11 Female 3.38 10.56 12.76 8.21 Source:National Statistics Office. 2000 Censusbased Population Projection collaboration with the InterAgenWorking Group on Population Projections Table A.2: SECTORAL OUTPUT GROWTH CHANGE (Base: 1985 = 100; Pesos in Millions) Benchmark Crisis 2006 2009 2010 2009 2010 TOTAL 1,276,156 1,485,526 1,559,412 1,438,056 1,482,592 Agriculture 239,777 267,215 275,274 262,312 268,136 GDP in constant Manufacturing 305,663 342,490 356,874 308,899 311,988 prices 109,152 Other Industries 144,746 154,507 152,108 165,122 Services 621,564 731,075 772,757 714,736 737,345 Total 5.40 4.69 4.97 1.35 3.10 Agriculture 3.70 3.01 3.02 1.12 2.22 Annual Growth Manufacturing 4.60 4.00 4.20 1.00 (6.20) rate 5.41 Other Industries 6.67 6.74 12.09 8.56 Services 6.47 5.26 5.70 2.91 3.16 Remittances (USD Billions nominal) 12.7 17.9 19.5 17.1 18.0 Source: NSO GDP of Philippines and Projections Table A.3: EMPLOYMENT SECTORAL GROWTH CHANGE Individuals >15 & < 64 years old Benchmark Crisis 2006 2009 2010 2009 2010 Millions % Millions % Millions % Millions % Millions % Total 51.1 55.1 56.5 55.1 56.5 Inactive 16.1 31.5 17.9 32.4 18.6 33.0 17.7 32.0 18.3 32.4 Unemploye 3.2 9.1 3.2 8.6 3.2 8.4 3.9 10.3 4.3 11.2 Employed 31.4 90.9 34.0 91.4 34.7 91.6 33.7 89.7 34.1 88.8 10.6 Agriculture 33.8 11.3 33.2 11.5 33.1 11.2 33.2 11.3 33.2 Manufactur 3.0 9.5 2.9 8.5 2.9 8.4 2.9 8.7 2.9 8.6 Other Indus 1.9 6.1 2.1 6.3 2.1 6.1 2.1 6.3 2.1 6.1 Services 15.9 50.6 17.7 52.1 18.2 52.6 17.5 51.9 17.8 52.2 Source: NSO GDP of Philippines and Projections v Table B DEMOGRAPHIC, HOUSEHOLD INCOME & LABOR MARKET OVERVIEW 2009 2010 2006 Benchmark Crisis Benchmark Crisis Female Male Female Male Female Male Female Male Female Male Population (million) Total 41.4 42.0 44.0 44.6 44.0 44.6 44.9 45.4 44.9 45.4 Urban 20.5 20.1 21.8 21.3 21.8 21.3 22.2 21.7 22.2 21.7 (1) Workingage 25.4 25.7 27.4 27.7 27.4 27.7 28.1 28.5 28.1 28.5 (2) Household Income ($/month) Total 19,441.0 16,803.4 19,975.6 17,808.0 19,504.0 17,431.2 20,272.8 18,130.9 19,558.2 17,475.9 (3) Labor income 9,207.0 12,132.8 9,740.3 13,038.7 9,307.8 12,682.2 9,929.9 13,317.6 9,328.9 12,732.1 (4) Nonlabor income 8,145.4 3,153.5 8,146.8 3,244.9 8,107.6 3,224.6 8,254.3 3,286.4 8,140.8 3,216.8 Remittances 6,158.5 1,743.6 6,128.1 1,807.8 6,103.6 1,797.1 6,225.8 1,840.7 6,135.3 1,786.1 Abroad 5,156.2 1,243.3 5,071.2 1,274.7 5,080.4 1,281.1 5,152.6 1,297.5 5,114.9 1,269.7 Domestic 1,002.2 500.3 2,085.9 1,172.1 2,019.2 1,134.6 2,117.4 1,194.1 2,013.1 1,135.2 Implicit rent 2,088.6 1,517.1 2,088.5 1,524.5 2,088.5 1,524.5 2,088.5 1,526.9 2,088.5 1,526.9 (1) Workingage (million) I. Labor Force 24.5 24.9 26.5 26.8 26.5 26.8 27.1 27.5 27.1 27.5 Inactive 10.9 3.9 12.2 4.4 12.0 4.3 12.8 4.6 12.5 4.5 Active 13.7 20.9 14.3 22.4 14.5 22.5 14.4 22.9 14.7 23.0 Unemployed 1.3 1.9 1.3 1.9 1.5 2.3 1.3 1.9 1.7 2.6 Employed 12.6 19.3 13.0 20.5 13.0 20.2 13.1 21.0 13.0 20.4 Economic sectors 12.4 19.1 13.0 20.5 13.0 20.2 13.1 21.0 13.0 20.4 Agriculture 2.7 7.9 2.8 8.3 2.8 8.2 2.8 8.5 2.8 8.3 Manufacturing 1.3 1.7 1.2 1.7 1.2 1.7 1.2 1.7 1.2 1.7 Other industries 0.1 1.9 0.1 2.0 0.1 2.0 0.1 2.0 0.1 2.0 Services 8.2 7.7 8.9 8.5 8.9 8.3 9.1 8.8 8.9 8.5 II. Earnings ($/month) (5) Mean per worker 5,876.2 6,457.7 6,353.0 7,116.9 6,188.1 6,978.9 6,497.4 7,304.2 6,233.2 7,108.1 Economic sectors Agriculture 1,934.4 3,743.4 2,016.2 3,935.6 1,993.3 3,875.5 2,032.0 3,962.8 2,004.1 3,893.5 Manufacturing 6,182.5 8,139.5 6,897.3 9,557.8 6,239.0 8,514.5 7,165.5 9,889.2 6,245.5 8,609.0 Other industries 12,327.1 6,863.7 15,675.0 8,552.1 16,365.4 8,975.6 16,366.4 8,978.4 17,572.3 9,641.6 Services 7,233.9 8,855.2 7,733.1 9,470.6 7,594.3 9,356.5 7,928.5 9,713.9 7,698.2 9,518.0 Notes: (1) Individuals in 1564 age category (2) Refers to female vs male household heads (3) Labor earnings from all activities (4) Includes capital(rent land, property, profits, etc), remittances, social(insurances, charity, etc) and other nonlabor incomes (5) Refers to female vs male total workers Source: Own estimations based on FIES 2006 and projections vi Table C FOOD PRICES COMPARED TO OVERALL CPI (Base 2006 = 100) Weight of food in CPI and Poverty Line (PL) CPI CPI Food 50.03 General Food NonFood Nonfood 49.97 (1) (2) (3) PL NCR Food 57.41 2006 100.00 100.00 100.00 Nonfood 42.59 Benchmark Benchmark variation in PL '09 0.07 2009 113.14 113.65 112.63 variation in PL '10 0.07 2010 118.80 119.33 118.26 Crisis Crisis variation in PL '09 1.08 2009 116.03 122.70 109.99 variation in PL '10 1.08 2010 121.25 128.22 114.94 Source: NSO Consumer Price Index (CPI) and Inflation Rate several reports TABLE D POVERTY IMPACT BY AREA & REGION Table D.1: Household income (mean Pesos/month at constant 2006 prices) Household Income (mean Pesos/month) AREA REGION TOTAL Rural Urban NCR Central Luzon Calabarzon Other Luzon Visayas Mindanao Benchmark Total income 18,508 13,575 23,089 27,122 20,179 20,861 15,312 16,700 14,610 2010 Labor income 12,721 9,413 15,698 18,985 13,241 14,304 10,028 10,955 10,867 Remittances 2,612 1,799 3,380 3,239 3,768 3,453 2,300 2,654 1,317 Crisis Total income 17,795 13,773 23,538 27,616 20,659 21,244 15,537 16,990 14,837 2010 Labor income 12,090 9,573 16,066 19,393 13,637 14,614 10,197 11,192 11,063 Remittances 2,553 1,829 3,445 3,300 3,842 3,517 2,346 2,698 1,337 Loss of income with crisis (% of benchmark) 3.9 1.5 1.9 1.8 2.4 1.8 1.5 1.7 1.5 Loss of labor income (% of benchmark) 5.0 1.7 2.3 2.1 3.0 2.2 1.7 2.2 1.8 Loss of remittance (% of benchmark) 2.3 1.7 1.9 1.9 2.0 1.9 2.0 1.7 1.6 vii TABLE E PEOPLE WHO ARE POOR "DUE" TO CRISIS 2010 Households characteristics CrisisVulnerable Total Structurally poor Area Urban Rural (%) 57.1 51.5 73.1 Members 5.12 4.82 5.87 Dependency 0.39 0.38 0.49 Employed (individuals Benchmark) 59.3 60.4 0.6 Agriculture 36.9 33.1 64.0 Manufacturing 11.3 8.4 4.8 Other industries 3.3 6.1 3.9 Services 48.5 52.4 27.4 Household income (CrisisVulnerable) Benchmark Crisis Income (mean) Total Income 12,051.0 7,317.2 Labor income 9,594.3 5,017.5 Nonlabor income 1,650.0 1,493.0 Remittances from abroad % of household receiving 15.1 13.7 Mean conditional on receiving 380.2 253.8 Implicit rent 806.6 806.6 Percapita Income (mean) 2,569.4 1,402.0 Household head CrisisVulnerable Total Structurally poor Age (mean) 46.10 47.69 45.65 Male (%) 84.5 82.4 88.6 Education level Low (%) 67.3 55.9 81.0 TABLE F INCOME STATUS "CHANGES" Table E.1: Transition matrix: Benchmark Crisis Percapita income Crisis 1 2 3 4 5 6 7 8 9 10 1 96.04 1.84 0.55 0.48 0.21 0.34 0.29 0.14 0.12 2 8.04 88.32 2.10 0.76 0.29 0.28 0.12 0.04 0.05 3 1.76 10.60 84.09 2.28 0.41 0.34 0.30 0.06 0.16 4 1.19 1.25 11.69 81.69 3.12 0.43 0.25 0.17 0.12 0.08 Benchmark 5 1.16 0.80 0.86 13.15 79.34 3.39 0.60 0.43 0.25 0.01 6 0.63 0.67 0.69 1.22 14.19 78.58 3.26 0.52 0.15 0.09 7 0.59 0.29 0.60 0.90 1.71 12.47 79.87 2.78 0.75 0.04 8 0.73 0.21 0.45 0.64 0.86 1.40 12.29 81.50 1.80 0.13 9 0.40 0.25 0.19 0.21 0.42 1.10 1.60 10.23 84.13 1.47 10 0.29 0.09 0.08 0.11 0.07 0.13 0.32 0.74 7.12 91.04 Table E.2: Transition matrix: Benchmark Crisis Percapita Labor income Crisis 1 2 3 4 5 6 7 8 9 10 1 96.39 1.89 0.43 0.18 0.31 0.18 0.29 0.15 0.06 0.13 2 7.75 88.61 2.06 0.49 0.45 0.15 0.14 0.23 0.07 0.05 3 1.41 10.31 84.04 2.45 0.68 0.54 0.31 0.14 0.11 4 1.08 0.64 12.71 80.54 3.46 0.55 0.39 0.47 0.15 Benchmark 5 0.84 0.84 1.43 13.52 78.2 4.05 0.54 0.27 0.22 0.11 6 0.48 0.60 0.77 0.72 14.29 77.6 3.88 0.76 0.65 0.22 7 0.58 0.39 0.51 0.55 1.12 15.80 76.8 3.77 0.42 0.09 8 0.55 0.27 0.27 0.44 0.97 0.98 13.16 80.54 2.48 0.34 9 0.48 0.12 0.27 0.33 0.54 1.09 1.72 12.60 81.3 1.53 10 0.30 0.09 0.14 0.17 0.21 0.17 0.58 1.00 7.79 89.5 viii