Poverty and Inequality in Syria ( )

Transcription

1 Arab Development Challenges Report Background Paper 2011/15 Poverty and Inequality in Syria ( ) Khalid AbuIsmail, Ali AbdelGadir and Heba ElLaithy

2 United Nations Development Programme Arab Development Challenges Report Background Paper 2011/15 Poverty and Inequality in Syria ( ) Khalid AbuIsmail, Ali AbdelGadir and Heba ElLaithy Khalid AbuIsmail is Poverty and Macroeconomics Advisor at United Nations Development Programme Regional Centre in Cairo (UNDPRCC). Ali Abdel Gader is Acting Director of the Arab Planning Institute. Heba ElLaithy is Professor of Statistics at Cairo University. Comments should be addressed by to the author(s)

3 Acronyms and Abbreviations CBS CSO FPL GDP HDI HIES HPI IMF ISI LAS LPL MDGs SL UN UNDP UPL USA USSR Central Bureau of Statistics Central Statistical Office Food Poverty Line Gross Domestic Product Human Development Index Household Income and Expenditure Survey Human Poverty Index International Monetary Fund ImportSubstitution Industrialization League of the Arab States Lower Poverty Line Millennium Development Goals Syrian Lira United Nations United Nations Development Programme Upper Poverty Line United States of America Union of Soviet Socialist Republic

4 Introduction Since the early 2000s, in the endeavour to understand the unprecedented levels of rising poverty in Syria in the last decade, the Syrian State Planning Commission, supported by the UNDP, in partnership with the Central Bureau of Statistics, has undertaken a number of studies to support the decision making of policy makers for the achievement of the MDGs and to directly contribute to the poverty reduction objectives stipulated in the National 10 th five year plan of the Government of Syria. Some of these notable efforts have included: Poverty in Syria (2005), which identified the baseline for expenditure poverty and provided evidence of the geographic distribution and concentration of poverty; Macroeconomic Policies for Poverty Reduction (2005), which highlighted the results of the 2004 Household Income and Expenditure Survey (HIES) and provided a menu of policy options at the sectoral and macroeconomic levels to tackle the phenomenon of poverty; Poverty and Income Distribution in Syria (2009,) which sketched a profile of the primary features of poverty in Syria based on the expenditure data from the HIES survey and updated the findings of the first UNDP poverty assessment report by providing a comprehensive overview of the evolution of poverty and income distribution since Two other UNDPled studies in 2009: "Understanding Poverty and Inequality Dynamics in Syria" and "ProPoor in Syria: Diagnosis and Policy Considerations" were undertaken to spur further research and analysis, as well as to inform development practitioners of the structural issues of poverty in Syria. Drawing on the main findings of the abovementioned reports, in this paper, we undertake an inquiry into the imperative question of the extent to which state actions and markets have affected social outcomes, such as the overall spread of poverty and inequality in expenditure and access to social services over the period from 1997 to 2007, for which we have data. This paper, which is a compilation of the findings of the papers Propoor in Syria: Diagnosis and Policy Considerations (2009) by AbdelGadir and AbuIsmail, Poverty and Distribution in Syria (2009) by AbuIsmail and ElLaithy, and Understanding Poverty and Inequality Dynamics in Syria (2009) by AbuIsmail and ElLaithy,is organized in the following manner: First, we explain the main definitions and methodologies applied to achieve a comparable assessment of the trends in poverty in Syria across the two discernible periods and We then present stylized facts taking stock of poverty, growth and inequality trends in Syria over these periods( ), exposing regional and ruralurban discrepancies. In our assessment of the depth of poverty and inequalities in Syria we paint the relationship between income and human poverty over time to shed some light on the possible links between these results and policy stances as well as the impact of extraneous factors such as drought. We then assess the propoorness of the growth processes that took place during these periods and , again examining the nature of this growth process at the national, regional and ruralurban level. We interpret these results of human deprivation with an explanation of the underlying root cause of the shifts in these indicators by taking into account the major shift in macroeconomic policies that took place in Syria during the 1990s and 2000s. Finally, we strive to present Syrian policy makers with a menu of policy interventions to enhance the processes of propoor growth. To start with, we must note that our approach is anchored in the UNDP s human development approach, namely in the conviction that understanding the dynamics of poverty and inequality in any developing country requires the adoption of a broad definition for the "development process". Such a definition, which has gained worldwide acceptance, is that development can be seen as a process of expanding the real freedoms that people enjoy 1. Without getting involved in the philosophical foundations of this approach we need to note that it requires judging the welfare of individuals neither in terms of the utility of goods and services, nor in terms of primary goods, but in terms of substantive capabilities to choose a 1

5 life one has reason to value. A person s capability to achieve functioning s that he or she has reason to value provides a general approach to the evaluation of social arrangements, and this yields a particular way of viewing the assessment of equality and inequality 2. Development as freedom is a much broader approach to understanding what is meant by development compared to other approaches that identify development with increases in per capita income, or with industrialization, or with technological advance, or with social modernization. Being broad, the capability perspective is closely related to the concept of equality of opportunities ; though it is not identical. In a very real sense, a person s capability to achieve does indeed stand for the opportunity to pursue his or her own objectives. But the concept of equality of opportunities is more commonly used in the policy literature in more restrictive ways, defined in terms of the equal availability of some particular means, or with reference to equal applicability of some specific barriers or constraints. Thus characterized, equality of opportunities does not amount to anything like equality of overall freedoms 3 ; (emphasis is in the original) 4. As is well known, the UNDP pioneered the measurement of development as freedom in terms of the by now famous Human Development Index (HDI). Capabilities included in the HDI are: the ability to live longer (as reflected in life expectancy at birth, to proxy health); the ability to read and write and to have access to available information (as reflected in literacy rates, as well as educational enrolment rates); and, the ability to have a decent standard of living (as reflected by real per capita income). The HDI is now very widely used to guide development policy. In addition to the substantive influence of the HDI, being a measure of the broad definition of the development process, on policy, "development as a process of expanding the freedoms that people enjoy" has found international recognition in the Millennium Development Goals (MDGs) adopted by the United Nations in September Over the past nine years, since their formulation, the MDGs and the broader definition of development, on which they are based, influenced the development policy debate around the world (see, for example, the Commission for Africa (2005); Sachs (2005); UN Millennium Project (2005); the World Bank (2006); and, the Commission on and Development (2008)) 5. The IMF, the World Bank, and the UNDP are all involved in helping least developed countries in estimating the feasibility of achieving these goals and the cost of doing so! Furthermore, as is well known under the MDGs, the overarching objective of development in developing countries is the reduction of poverty. The first MDG requires the eradication of extreme poverty and hunger, where poverty is expressed as the proportion of people living below one dollar a day (i.e. the headcount ratio). Four of the remaining MDGs look at poverty from a capability, and achievements, perspective (education, gender equality, health and the environment). Under the dominant moneymetric approach to poverty reduction, over a 25year horizon both economic growth (changes in real per capita consumption expenditure) and changes in the inequality in the distribution of consumption expenditure matter. Thus, changes in poverty over time have a growth component and a distribution component: if a country experiences both economic growth and a decline in inequality it can be assured of reduced poverty. Otherwise, it will all depend on the relative strength of each of the two components. Thus, understanding the dynamics of inequality in any developing country is an important guide to devising a development policy anchored on the overarching objective of development, namely poverty reduction over a long horizon. Indeed, one of the indicators identified to measure progress towards achieving poverty reduction is the share of the 20% of the population in total consumption expenditure. Given the above noted two components 2

6 of changes in poverty over time (i.e. the growth and inequality components), concern about the share of the poorest fifth of the population gave rise to a huge literature on measuring the propoorness of growth processes. Such literature provides the relevant theoretical, and empirical, framework to understand the dynamics of inequality in a developing country such as Syria. This is the framework that we adopt in this chapter. Methodological Preliminaries For the purpose of an objective assessment of the evolution of poverty, the methodology for measuring income poverty applied in this report is consistent with the other World Bank and UNDPled poverty assessment reports for the Arab Region (see for example, UNDP and LAS (2008) for the Arab Region, UNDP (2006) for Yemen, UNDP (2007) for Lebanon and the earlier UNDP (2005) Syria poverty assessment (Box 1)). The results are presented using three main definitions, which are: 1. Abject poverty : defined as the share of population whose expenditure lies under the food poverty line (FPL); 2. Extreme poverty : defined as the share of population whose expenditure lies under the lower poverty line (LPL); and 3. Overall poverty : defined as the share of population whose expenditure lies under the upper poverty line (UPL). The first Poverty Assessment Report (UNDP 2005) concluded that, in , almost 2 million individuals in Syria (11.4% of the population) were extremely poor. Using a higher expenditure poverty line, overall poverty in Syria was estimated at 30.1%, representing almost 5.3 million individuals. In addition, the 2005 report identified the following major trends and characteristics of poverty: 1. Poverty in 2004 was generally more prevalent in rural than in urban areas of Syria (62% in rural areas). The NorthEastern region (Idleb, Aleppo, Al Raqqa, Deir Ezzor and Hassakeh), had the greatest incidence, depth and severity of poverty; 2. Poverty decreased between and for Syria as a whole, but regional patterns were different. The incidence of poverty declined rapidly in the Central and Southern regions, especially in rural areas. The decline was moderate in urban areas of the NorthEastern and Coastal regions, and poverty actually rose in the rural parts of these regions; 3. At the national level, growth was not propoor. Nonpoor individuals (above the third decile in the expenditure distribution) benefited proportionally more than the poor from economic growth. Between the years , inequality in Syria, as a whole, rose (the Gini index rose from 0.33 to 0.37). In , the bottom 20% of the population consumed only 7% of all expenditure in Syria while the richest 20% consumed 45%; 4. Poverty in Syria is shallow, with most people clustered just below the poverty line; and 5. Education was the single characteristic with the strongest correlation to poverty risk in Syria. More than 18% of the poor population was illiterate, and poverty was highest, deepest and most severe for these individuals. As in the 2005 UNDP poverty report, this report adopts a wellestablished costofbasicneeds methodology that accounts for: (i) differences in consumption patterns and prices across regions, (ii) different basic needs requirements of different household members young versus old, male versus female, (iii) the cost of the actual diet, and (iv) economies of scale within households the fact that nonfood items can be shared among household 3

7 BOX members. The costofbasicneeds methodology thus yieldsan unbiased absolute poverty line, poverty lines that are householdspecific objective, and regionally consistent. As these poverty lines account for regional differences in relative prices, expenditure patterns, activity levels, as well as the size and age composition of poor households, they are expected to vary from one region to another and from one household to another within the same region. However, it is intuitively clear that households with similar age and gender compositions who live in the same region will face similar poverty lines. The first step in constructing any income poverty assessment is to construct the food poverty line. This is completed by identifying a food bundle that yields the predetermined caloric requirements, but with a composition that is consistent with the consumption behaviour of the poor. This bundle was defined for individuals in different age brackets, gender, and activity levels (using tables from the World Health Organization). FPLs were then set using the cost of the required calories based on consumption pattern of the second quintile (using the prices for the food in each region and at each time period). Thus the relative quantities observed in the diet of the poor (proxied by the second quintile) and the prices they face were taken into account while constructing the FPL. As noted above, households whose expenditure is below the FPL are referred to as "abjectly poor". 1 Comparability between the 2004 and 2007 surveys The HIES surveys in 2004 and 2007 are the main source of data for the poverty analyses contained in this report. Based on data in those surveys, we are able to record household income and consumption expenditures on more than 550 goods and services, which is a valuable source of information on the level and distribution of welfare for the Syrian society. The sampling, questionnaire design, and the administration of the two surveys are very similar to allow for comparability. Thus, in both surveys selected households were asked to report their daily expenditures during a tenday period only and other regular expenditures were reported using the same recall period; data on the ownership of assets and imputed rent for owned houses which are very important in the course of poverty assessment were also similarly recorded in both surveys. However, there are two main differences between the surveys, which do not affect their comparability. First, the survey includes more detailed questions on employment characteristics of household members. Another major difference between and surveys is the sample size. Although the sample design is similar in both surveys and about one third of the selected PSU's in were reselected in , the selected households were different. Nonetheless, the sample of the survey is statistically representative at the regional level. The governorates comprising each region and sample size at national, regional and governorate levels for both surveys is listed in the table below. South North East Middle Coastal Region Governorates Sample size ( ) Urban Damascus, Rural Damascus, Deraa, El Suaida and El Quneitra Rural Urban Idleb, Aleppo, Al Raqqa, Deir Ezzor and Hassakeh Rural Urban Homs and Hama Rural Urban Tartous and Latakkia Rural Total Syria Source: HIES 2004 and 2007 Sample size ( ) The second step is estimating the LPL, which is basically the FPL augmented by an allowance for expenditure on essential nonfood goods. The latter is determined based on expenditures by households who have to forego food consumption to allow for nonfood expenditures that are deemed indispensable. The nonfood allowance can be estimated by identifying the share of nonfood expenditure for households whose total expenditure was equivalent to the food poverty line. Any household that spends less than the LPL is 4

8 considered extremely poor. Therefore, the abjectly (food) poor group are just a sub set of the extremely poor. Finally, for a more inclusive poverty measure, we construct the UPL by enlarging the nonfood component to include a more reasonable minimum required level of nonfood expenditures. This UPL is estimated when the FPL is augmented by the nonfood expenditure of households whose food expenditure is equivalent to the food poverty line. The report refers to households whose expenditure lies under the UPL as the "poor" which include the extremely poor and abjectly poor subgroups. Overall poverty in this report thus refers to the share of population whose mean expenditure lies below the UPL. One final observation with respect to the construction of the poverty lines is that all poverty lines for were estimated using the reference food baskets of , but with a different set of regional unit prices derived from the 2007 survey. 6 Household specific food poverty lines were thus derived from multiplying the caloric cost by the required calories of household members. Similar to the poverty lines, the nonfood poverty lines are then estimated based on an Engle s curve regression. Table 1: Average monthly per capita consumption expenditure and poverty lines, 2004 and Expenditure (SL) FPL LPL UPL Expenditure (SL) FPL LPL UPL Southern Region Urban Rural NorthEastern Region Urban Rural Central Region Urban Rural Coastal Region Urban Rural All Syria Urban Rural Total Source: Complied based on UNDP 2005 and CBS Household income and Expenditure Survey, As shown in table 1, the estimated poverty lines (averaged to produce per capita figures) show that the thresholds for different types of poverty vary across the regions. Accordingly, in , on average, individuals who spent less than SL 1439 per month (equivalent to 32% of average per capita expenditure) are considered to be in abject poverty while those who spent less than SL 2183 monthly (equivalent to 48% of average per capita expenditure) are extremely poor (Table 1). Using data in table 1, we observe that the ratios of the value of the poverty lines to total expenditure for 2004 and 2007 were, respectively, as follows: FPL (0.29 and 0.32); LPL (0.41 and 0.48); UPL (0.58 and 0.68). These results indicate that the minimum cost of living in Syria (relative to the average expenditure) has increased. 5

9 Poverty,, and Inequality Stylized Facts Poverty trends, Figure 1 and table 2 present an overall picture of poverty in Syria for 2007 based on the most recent HIES. The former shows that overall poverty (share of the population under the UPL) is relatively widespread in Syria, affecting 33.6% of the population. This implies that 6.7 million Syrian people are considered poor. Of this larger group, almost 2.4 million (12.3% of the population) are estimated to live in extreme poverty. Hence, the average expenditure of approximately 4.3 million Syrian individuals (21.3% of the population) lies between the thresholds of extreme poverty (i.e. under the lower poverty line) and overall poverty (i.e. under the upper poverty line) (Figure 1). The ruralurban discrepancies characterizing poverty in Syria become immediately discernible from the data presented in Figure 1 and table 2. Figure 1: Distribution of Syrian population (%) across poor and nonpoor categories, 2007 Urban Rural National Below FPL Below FPL Below FPL Above FPL and Below LPL Above FPL and Below LPL Above FPL and Below LPL Above LPL and Below UPL Above LPL and Below UPL Above LPL and Below UPL Above UPL Above UPL Above UPL 1% 9% 2% 13% 1% 11% 22% 23% 22% 68% 62% 66% Source: UNDP (2009) These ruralurban discrepancies translate across all regions as well, as demonstrated by table 2, which show cases the strong regional dimension of poverty in Syria. As the table shows, in 2007, 46% of the sampled population in Syria lived in rural areas, but the poor in those areas accounted for about 56% of the extremely poor and 50% of the overall poor population. Rural areas thus recorded higher poverty measures than their urban counterparts in all regions throughout the region. The Coastal region was the least poor with an extreme poverty rate of 7.68%, while the NorthEastern region was the poorest in the country with an extreme poverty rate of 15.4%. The results also show that the poor in Syria were mostly concentrated in the NorthEastern region, particularly in the rural areas of the region, for which the share of the extremely poor and overall poor population (37% and 28.8%, respectively) exceeded by far its population share of 23.1%.When the overall poverty rate is considered, both the NorthEastern and Southern regions had nearly the same incidence of poverty. However, the expenditure levels of the poor in the Southern urban region were, on average, far below the poverty line. Consequently, the poverty gap for the Southern urban region is 25% higher than the poverty gap for urban Syria using the UPL (Table 2). Now turning to the evolution of poverty over the period from 1997 to 2007, table 3 gives a summary of the main changes observed by comparing the results from the three surveys. Accordingly, overall poverty in Syria decreased over the first period (from 33.2 % in 1997 to 30.1 % in 2004) before rising in the second period (to reach 33.6 % in 2007). Thus, the 6

10 deterioration since 2004 has offset the poverty reduction achieved over the period from 1997 to However, not all poverty indicators moved in the same direction from 2004 to 2007: as reported in UNDP (2009), abject (food) poverty decreased (from 2.2 % in 2004 to 1.2 % in 2007), extreme poverty increased slightly (from % to 12.3 %), and overall poverty increased rather more significantly (from 30.1 % to 33.6 %). Table 2: Poverty measurements by region, 2007 Using Lower Poverty Line Using Upper Poverty Line Po P1 P2 % of Poor Po P1 P2 % Poor % of Individuals Southern region Urban Rural Total NorthEastern region Urban Rural Total Central region Urban Rural Total Coastal region Urban Rural Total All Syria Urban Rural Total Source: Complied based on UNDP Poverty Assessment reports (2005) and (2009) and data from HIESs produced by the CBS in 1997, 2004 and Note: P 0 represents the incidence of poverty, P 1 represents the poverty gap index, and P 2 represents the severity index of poverty. Those changes indicate that although poverty most likely increased in Syria over the past few years, there was an improvement in the conditions of the poorest of the poor. Distributional changes over the period from 2004 to 2007 thus affected the poor and the nonpoor differently, with the bottom 5% actually showing an improvement in their living standards. Table 3: Incidence of extreme and overall poverty for rural and urban areas (in per cent), Incidence of Poverty Urban Rural Total Urban Rural Total Urban Rural Total Extreme Poverty Overall Poverty Change in % Change % Change % Change incidence of Poverty Urban Rural Total Urban Rural Total Urban Rural Total Extremely Poor Poor Source: ibid. Changes in poverty varied greatly among Syrian regions. Two consistent patterns of poverty evolution in emerge when the lower poverty line is considered. First, the incidence of poverty increased in the rural areas of the NorthEastern and Coastal regions, especially in the NorthEastern region (from 15.2% to 18%). The poverty gap and severity indices followed a similar pattern. Second, during this first period all other regions 7

11 South U South R NE U NE R Central U Central R Coastal U Coastal R Syria U Syria R Syria South U South R NE U NE R Central U Central R Coastal U Coastal R Syria U Syria R Syria experienced declines in their poverty measurements ( ), with substantial decreases in extreme poverty in the urban and rural areas of the Central region (from 14.8% to 9.02%, for P0 in urban areas and from 22.6 to 11.1%, in rural areas). In the rural Coastal region, although there was a slight increase in extreme poverty, overall poverty substantially declined for the same period. In this region, the 35th60th percentile of expenditure distribution experienced a higher change in their expenditure than in the poverty line. Figure 2: Changes in the incidence of the extreme poverty (A) and overall poverty (B) by region, (A) (B) Source: ibid Regarding the changes in patterns of geographic distribution of poverty over the recent period , clearly, poverty is still more concentrated in the NorthEastern region, especially the rural areas, where 19.7% of the population is extremely poor (i.e. below the lower poverty line). Also, the urban Southern region experienced a considerably large increase in extreme poverty with a poverty incidence in 2007 that is almost twice that of 2004 (Figure 2.A). Accordingly, this region, which had the lowest levels of poverty in 2004, became the second poorest region in In fact, as shown in UNDP (2009), the overall increase in extreme poverty nationwide between 2004 and 2007 was mainly driven by the sharp rise in poverty in the Southern region. Changes in overall poverty (i.e. percentage of population under the upper poverty line) also varied significantly between regions. Four out of the eight regions in Syria followed the overall nationwide trend of increasing poverty (from ). However, the pace of change was not homogeneous across those regions. As in the case of extreme poverty, overall poverty in the urban Southern region showed the largest increase (by 50%). However, contrary to the national trend, overall poverty rates declined in some regions, particularly in the urban areas of the NorthEastern region. They also declined rather insignificantly in the Central region and in urban areas of the Coastal region (Figure 2.B). in per capita consumption expenditure At the outset, it must be noted that striking inconsistencies exist between growth rates obtained from the Syrian national accounts and Household Surveys. The data on growth in per capita consumption expenditure used in this paper, as stated previously, are mainly derived from two UNDPled poverty assessment studies conducted in 2005 and 2008 (El Laithy and AbuIsmail (2005) and ElLaithy and AbuIsmail (2008)), which are based on the Household Income and Expenditure Surveys conducted by the Central Statistical Office in 1997, 2004 and 2007.It should be recalled that in developing countries real per capita consumption expenditure is the most widely used indicator of the standard of living relevant to the analysis of poverty. 8

12 On the basis of the survey data, real per capita expenditure in prices are reported in ElLaithy and Abu Ismail (2008). Such estimates are reported for the level of the country as well as for geographical regions, with their residential sectors (urban/rural). For the purposes of our analysis, we use the regional distribution of the population of the country in 2004 population as reported in ElLaithy and Abu Ismail (2005: 42, annex table 2) with a total population for the country of about 17.7 million. It is not surprising that the ruralbias of poverty incidence, previously demonstrated in table 3 and Figure 2, reflects onto the per capita consumption growth process over this period, through a clear urbanbias of per capita consumption growth at the country level (Table 4). Table 4 shows that over the period , real per capita consumption expenditure increased by an annual rate of growth of about 1.6%nationwide, with the urban per capita consumption increasing by an annual rate of about 2.1% compared to a growth rate of only 0.5% for the rural sector. Table 4: Real per capita consumption expenditure, Real Per Real Per Real Per Annual Capita Capita Capita Region Sector Expenditure Expenditure Expenditure Rate (%): 1996 ( S)* 2004 ( S)* 2007( S)* Southern North Eastern Central Coastal Syria Annual Rate (%): Annual Rate (%): Urban Rural Total Urban Rural Total Urban Rural Total Urban Rural Total Urban Rural Total Source: ibid Note: * S denotes Syrian Lira (called Syrian Pound by the IMF) per person per month in 2004 constant prices. This feature of the growth process from 1996 to 2004 is reflected at the level of the regions, except for the Central Region, which posted the highest rate of real per capita consumption expenditure with an overall annual rate of growth of about 2.7% and with both sectors recording a fairly high rate: 3.6% for the urban sector and 3.3% for the rural sector. The Southern Region was the second fastest growing region posting an annual rate of growth of 2%, with the urban sector growing at a rate of 2.6% compared to only 0.4% for the rural sector. The NorthEastern Region was the third fastest growing region posting an annual rate of growth of 0.9% and with the urban sector growing at a rate of 2.8% compared to a negative growth rate of 0.7% for the rural sector. The lowest growing region was the Coastal Region, which posted an overall rate of growth of only 0.5%, with the urban sector reporting a growth rate of 1.3% compared to a negative growth rate of 0.1% for the rural sector. From table 4 it is clear that over the period real per capita consumption expenditure declined rather substantially by an annual rate of decline of about 4% at the level of the country, with the urban per capita consumption declining at an annual rate of about 5.6% compared to a rate of decline of 2.2% for the rural sector. However, it is not clear whether such a per capita consumption growth process could be considered as having been characterized by a clear ruralbias! Though all regions reported negative growth rates for per capita consumption expenditure, they varied in terms of the ruralbias feature. The best 9

13 performing region was the Coastal region, which posted the lowest rate of decline of only 0.2% (with the urban sector posting a positive growth rate of 0.5%, and the rural posting a negative growth rate of 0.8%). The highest rate of decline, of 5.6%, was reported for the NorthEastern region (with both sectors posting a rate of decline of about 5.9%). In the Southern region, which is the second worst performing region with an annual rate of decline of 4.9%, the rural sector posted a positive rate of growth of 1.4% while the urban sector declined at an annual rate of 6.7%. Over the longer period real per capita consumption expenditure as recorded in household budget surveys was stagnant at the country level, declining as it did by an annual rate of only 0.1%. The urban sector reported a positive rate of growth of about 0.4% per annum while the rural sector recorded a negative annual rate of growth of about 0.3%. The Central region posted positive rates of growth for both the urban (about 1% per annum) and the rural (about 2.3%) sectors. As is clear from the table, both the Coastal and the Southern regions recorded positive, but very low, overall rate of growth, while the NorthEastern region recorded a negative overall rate of growth of 0.8%. The urbanbias pattern of the growth process over this longer period was reported for all regions except for that of the Southern region where there was a clear ruralbias. The state of inequality in Syria To appropriately evaluate the state of inequality in the distribution of consumption expenditure in Syria we need to specify an international benchmark. To do so we use the most recent Gini coefficient results on the distribution of consumption expenditure in a sample of 84 countries reported by Ferreira and Ravallion (2008: table 1: 3540) for the period Without loss of generality we consider such results to represent the state of inequality in the world for the 2000 decade. For some of the countries the Gini coefficients were reported for the distribution of income. We adjusted these coefficients by subtracting 6.6% to get the Gini coefficients for expenditure distribution as per the results of Deininger and Squire (1996).. For the above sample of countries we calculated the mean (0.374 or 37.4%), the median (0.371 or 37.1%) and the standard deviation (0.092 or 9.2%). On the basis of these parameters we could classify countries with Gini coefficients falling in the range [ ] as countries that have a medium degree of inequality 8. Countries with low degrees of inequality would be those with Gini coefficients less than 0.328, while those with high degrees of inequality will be those with Gini coefficients of 0.42 and above. With the help of the above benchmarks, we are now in a position to look at the state of inequality in the distribution of consumption expenditure in Syria for the three years for which household budget survey data is available: 1997, 2004 and The information is reported in the following tables as readings of the respective Lorenz curves in terms of decile shares together with the calculated Gini coefficients. The results are reported for the four geographical regions of the country: Southern, NorthEastern, Central and Coastal, and for urban and rural sectors in each region, as well as for the country as a whole with its sectoral composition. We hasten to note that the detailed results reported in this section are based on the original information provided by UNDP based on CSO unit record data. 9 Table 5 provides the evidence on the degree of inequality for the four geographical regions, each split into urban and rural sectors, for Overall, the degree of inequality in Syria in 1997,as per the international benchmark of a Gini of 32.8%, was low, with a Gini coefficient of about 32.6%. The share of the lowest 20% of 10

14 the population was about 8% of expenditure (with a marginal difference between the urban and rural sectors), compared to a share of about 41% for the richest 20% (again with a marginal difference between the two sectors). As is well known, a standard result of the developmentinequality literature is that in developing countries in their early stages of development it is expected that rural sectors would display a more equal distribution compared to urban sectors. Such a result is confirmed for the distribution of consumption expenditure in Syria in 1997 for two geographical regions (NorthEastern and Coastal). For the other two regions the reverse result holds. However, the difference in the Gini coefficients between the two sectors is marginal for all geographical regions except for that of the Coastal region, where inequality in the urban sector is markedly higher than that in the rural sector. The degree of inequality in the distribution of consumption expenditure for the urban sector in the Coastal region (a Gini of about 36%) was also the highest nationwide. For this sector the lowest 20% of the population enjoyed about 7.5% of total expenditure compared to a share of 45% for the richest 20%. The lowest degree of inequality of about 31.76% was recorded for the rural sector of the NorthEastern region. For this sector the lowest 20% of the population received about 8% of total expenditure compared to a share of about 41% for the top 20%. Table 5: Distribution of consumption expenditure by region in Syria (Decile shares in percentages), 1997 Deciles Gini Region Sector Coefficient (%) Urban Southern Rural Total Urban North Rural Eastern Total Urban Central Rural Total Urban Coastal Rural Total Urban Syria Rural Total Source: UNDP calculations based on HIES unit record data. In 2004, there was an increase in the overall degree of inequality in Syria, as summarized by table 6, compared to the levels of inequality in 1997.This degree of inequality is nevertheless medium by international standards, with a Gini coefficient of about 37.5%. The share of the lowest 20% of the population was about 7.8% of expenditure (7.4% for the urban, and 8.3% for the rural sectors), compared to a share of about 44% for the richest 20% of the population (45% for the urban, and 42% for the rural sectors).once again, these results correspond with the inequalitydevelopment result in the sense that the degree of inequality was higher in the urban sector (with a Gini coefficient of about 38%) than that in the rural sector (with a Gini coefficient of about 34%).This ruralurban discrepancy held throughout all regions, which collectively recorded medium degrees of inequality as per the international benchmark of a Gini of 42% (upper bound). 11

15 The urban sectors in the NorthEastern and Central regions recorded the highest degrees of inequality, with respective Gini coefficients of 38.4%, and 39.5% In these higher inequality sectors the lowest 20% of the population enjoyed about 7% of total expenditure compared to a share of about 47% for the richest quintile of the population. The lowest degree of inequality was recorded for the rural sector of the Southern region, where the lowest quintile of the population enjoyed about 9% of total expenditure while the top quintile enjoyed about 40% of total expenditure. Table 6: Distribution of consumption expenditure by region in Syria (decile shares in percentages), 2004 Deciles Gini Region Sector Coefficient (%) Urban Southern Rural Total Urban North Rural Eastern Total Urban Central Rural Total Urban Coastal Rural Total Urban Syria Rural Total Source: Complied based on UNDP Poverty Assessment reports (2005) and (2009) and data from HIESs produced by the CBS in 1997, 2004 and In 2007, Syria recorded a medium degree of inequality in the distribution of consumption expenditure lower than the level in 2004, as summarized by table 7.It is also worthy to note that in 2007, although not on a national level, for all regions, with the exception of the Central region, inequality in the urban sectors was also higher than for the rural sectors, Moreover, it should be noted that for the regions, confirming the observation of the inequalitydevelopment literature, the difference in the degree of inequality between the sectors was very pronounced especially for the NorthEastern region, where the difference between the two sectors wasnearly10%! Further comparison of the geographical regions, reveals that the NorthEastern and Coastal regions recorded a medium degree of inequality compared to the Southern and Central regions, which recorded a low degree of inequality as per the international benchmark of a Gini of 32.8%.The results show that the Central region had the lowest inequality (with a Gini coefficient of about 30%), and the NorthEastern region had the highest inequality (with a Gini coefficient of about 34%). 12

16 Table 7: Distribution of consumption expenditure by region in Syria (Decile shares in percentages), 2007 Region Sector Deciles Gini Coefficient (%) Southern Urban Rural Total Urban North Rural Eastern Total Urban Central Rural Total Urban Coastal Rural Total Urban Syria Rural Total Source: ibid From the table it is also clear that the urban sector in the Coastal region was the highest inequality sector, with a Gini coefficient of about 35%, and where the lowest 20% of the population enjoyed about 8% of total expenditure compared to a share of 43% for the richest quintile of the population. The lowest degree of inequality was recorded for the rural sector of the NorthEastern region, with a Gini coefficient of about 24%, and where the lowest quintile of the population enjoyed about 11% of total expenditure while the top quintile enjoyed about 34% of total expenditure. To summarize, for the three years under study, using the benchmark of a Gini coefficient of about 32.8% or lower to identify countries as having low degrees of inequality, it can safely be concluded that Syria is a medium inequality country. One way of appreciating this is to recall the past history of Syria as an Arab socialist country, prior to its recent adventures into liberalizing its economic system. Low degrees of inequality on a world scale are a feature of fairly advanced Western countries (except for the USA) and exsocialist countries. Moreover, it needs to be recalled that inequality, being a structural characteristic of societies and economic systems, does not change dramatically during short periods of time unless subjected to huge shocks. Human poverty Finally, to supplement the story of poverty revealed by the moneymetric indicators presented above, an assessment of the depth of inequalities in Syria must also take into account the spread of human deprivation. Several human poverty indicators at the national and regional level are reported in the HIESs, which allowed a calculation of our estimation for the HPI, which of course, due to data limitations, differs from the global definition of HPI. The indicators reported in table 8 are: the percentage of households without access to safe water, the illiteracy rate and the percentage of children living in extreme poverty. The table shows that much of the progress in reducing human poverty is related to significant gains in access to safe water and declines in illiteracy rates, but there are discernable regional variations. Consistent with the results from previous indicators of poverty, the NorthEastern region, urban and rural, demonstrates relatively poor performance on each of the components of the HPI presented in table 8. Most notably, the region demonstrates the worst performance in literacy rates, in rural and urban regions, for all three years. Also, it is worthy to highlight that the rural sector of this region is the worst performing sector each year for all components of the HPI, with the exception of 1997 (for the incidence of children living in poor families). 13

17 P0 Table 8: Incidence of human poverty by region, Children living Unsafe Illiteracy Unsafe Year in poor HPI water rate water families Southern Urban Illiteracy rate Children living in poor families Southern Rural NorthEastern Urban NorthEastern Rural Central Urban Central Rural Coastal Urban Coastal Rural All Urban All Rural Source: ibid To shed further light on regional disparities, Figure 3 paints a picture of the relationship between income and human poverty over time by plotting headcount poverty rates (for extreme poverty) against the HPI index for 1997 and 2007 (data points representing 1997 are depicted in blue). The figure is split into eight quadrants (four for each year) separated by the nationwide average HPI and extreme headcount poverty. Thus, areas located in the upper right quadrants are in the worst position with a high headcount poverty rate and HPI score. Conversely, the lower left quadrant represents the best position. Figure 3: Changes in extreme poverty and the HPI by region, National Average HPI 2007 National Average HPI 1997 HPI 22 High P0 and Low HPI CnR High P0 and HPI TCn NER* CnU* SR TNER* CnU NEU SR* TU CsU TS T* TS* SU SU* NEU* TU* TCs CsR CnR* CsR* TCn* TCs* T TR* TR TNER National Average P National Average P NER 6 CsU* Low P0 and HPI High HPI and Low P Source: ibid Notes: NE is an acronym for the NorthEastern region, Cn for the Central region, Cs for the Coastal region, and S for the Southern region. T indicates the total and R or U indicate rural and urban, respectively so that, for example, TNER and SU stand for the total NorthEastern rural region and the Southern urban region, respectively. HPI 14

18 Southern Urban Southern Rural North East Urban North East Rural Central Urban Central Rural Coastal Urban Coastal Rural Southern Urban Southern Rural Nourth East Urban North East Rural Central Urban Central Rural Coastal Urban Coastal Rural Three main conclusions emerge from Figure 3, above. First, the HPI for Syria declined, rather significantly, from 19.7% in 1997 to 14 in 2007 with the rural areas of the Central and NorthEastern regions recording the highest rate of decline. Second, the level of deprivation in living standards is generally commensurate to the level of headcount poverty (as indicated by the upward sloping scattering of the data points). In other words, regions with high (low) income poverty also tend to have high (low) human poverty. Third, the gap between the NorthEastern rural region and other regions is clearly depicted by its situation in the upper right quadrant. Fourth, most regions witnessed a significant improvement in both HPI and extreme poverty rates over the last decade (as represented by the length of the downward sloping green arrows).this is particularly visible for the Central region due to the significant decline in poverty and in access to safe water, as earlier demonstrated in table 8. However, rural areas in the NorthEastern and Southern regions witnessed an improvement (or stagnation) in one poverty measure and deterioration in the other (as indicated by the blue arrows). Only one region (the Southern urban region) showed deterioration in both income and human poverty (albeit a very slight one). Finally, annex table 1 and figure 3 examine the inequality in the distribution of human poverty by deciles (of expenditure) and by region. This is important for investigating the evolution of the gap between the richest and poorest from the angle of human deprivation over the past decade. Figures 3.A and 3.B present the HPI data ranked by decile and the ratio of the HPI for the top to bottom deciles and quintiles, respectively. In annex table 1, deciles which experienced an increase in human poverty are marked in red. Accordingly, the Southern region (both urban and rural) is the only region that saw deterioration for most deciles (which is expected given the earlier reviewed data). Figure 3: Ratio of HPI for top to bottom deciles (A) and quintiles (B) by region, (A) (B) Source: ibid However, the gap between the highest and the lowest deciles appears not to have been altered significantly for the Southern urban and the NorthEastern rural regions compared to other regions, as shown in the above figure. It is also worthy to note that the lowest quintile of the Coastal region (both urban and rural) witnessed a sharp deterioration in its HPI, a pattern which begs further investigation. Propoor growth Given the complexity of the distributional changes presented above, it suffices to state that what matters for poverty reduction is not the average overall growth per se, but rather the growth in the income or consumption of the poor specifically. In other words, if the growth rates for the poor exceeds the growth rate in the mean for the entire population growth is judged to be propoor. Using the methodology outlined in the Annex Technical Notes, we 15

19 now look at the nature of the growth process that has taken place in Syria over three periods of time: , and We hasten to note that only the results for the two relatively longer periods of and should be considered as credible in view of the fact that changes in inequality over short periods of time must be reflecting great shocks as we noted earlier. Recall that in order to judge the propoor nature of the growth process we need to compare the average of the growth rates of per capita consumption expenditure of the deciles [γ*], which will be called the effective growth rate, with the overall average growth rate of per capita consumption expenditure, [G(μ)] for the whole society, which will be called the actual growth rate. When the effective growth rate is greater than the actual growth rate there is said to be a gain in growth and the growth process will be designated as propoor, and vice versa. At the national level: We start with the first period at the level of the country. According to official information in 1996, average real per capita consumption expenditure ( prices) amounted to SL 3224 per person per month. By 2004 this real average consumption per capita level increased to SL 3611 per person per month. Thus, over this 7years period the actual annual growth rate was 1.63%. The growth rate of real per capita consumption expenditure for the deciles is reported in table 9. Table 9: Propoor growth in Syria, Deciles Per Capita Expenditure (SL per month) 2004 Per Capita Expenditure (SL per month) 1996 Annual Effective Rate (%) [Actual Rate (%)] Average [1.63] Source: Abdel Gadir and AbuIsmail (2009) According to the results in the table, the effective annual per capita growth rate (i.e. the average of the growth rates of deciles) over the period was about 1.29% compared with the actual growth rate of 1.64%, representing a loss in per capita growth. Hence the growth process of was antipoor, meaning that the growth process during this period was accompanied by deterioration in the distribution of consumption expenditure. Indeed our earlier results on the state of inequality in Syria confirm this as evidenced by the increase in the Gini coefficient of the distribution of consumption expenditure between the two years. Thus the gains of the poor from the growth process that took place in the country over the period were comparatively less than those of the rich. We now move to the second period at the level of the country. The growth rate of real per capita consumption expenditure for the various deciles is reported in table 10. As noted above average real per capita consumption expenditure amounted to SL 3611 per month in 2004, declining to SL 3194 per person per month by Thus, over this 3years 16

20 period the actual annual growth rate of real per capita consumption expenditure was negative 4%. Table 10: Propoor growth in Syria, Deciles Per Capita Expenditure (SL per month) 2007 Per Capita Expenditure (SL per month) 2004 Annual Effective Rate (%) [Actual Rate (%)] Average [4.01] Source: ibid Comparing the effective growth rate of real per capita consumption expenditure of negative 2.38% over the period , with the actual growth rate of negative 4% we see that there was a gain in per capita growth. Hence the growth process of was propoor, meaning that the growth process during this period was accompanied by an improvement in the distribution of consumption expenditure. Indeed our earlier results on the state of inequality in Syria confirm this as evidenced by the decrease in the Gini coefficient of the distribution of consumption expenditure. Thus the gains of the poor from the growth process that took place in the country over the period were comparatively more than those of the rich. Last, we consider the whole period at the level of the country, as reported in table 11. In 1996, average real per capita consumption expenditure amounted to SL 3222 per month and SL 3194 per month in Thus, over this 11years period the actual annual growth rate of real per capita consumption expenditure was negative and amounted to about 0.08%. This, we hasten to note, is tantamount to saying that over the period under consideration real per capita consumption expenditure in the country stagnated. Table 11: Propoor growth in Syria, Deciles Per Capita Expenditure Per Capita Expenditure Annual Effective Rate (%) (SL per month) 2007 (SL per month) 1996 [Actual Rate (%)] Average [0.08] Source: ibid Comparing this actual negative growth rate of about 0.1% with the effective annual per capita growth rate (i.e. the average of the growth rates of deciles)of about 1.5 %, we can see that there was a gain in per capita growth. Hence the growth process over the period

21 Coastal Central North Eastern Southern Region 2007 was propoor, meaning that the growth process during this period was accompanied by an improvement in the distribution of consumption expenditure. Thus the gains of the poor from the growth process that took place in the country over the period were comparatively more than those of the rich. At the level of the regions Using the information in tables 12, 13, 14 and 15 we can have a look at the propoorness of the Syrian growth processes at the level of the four regions. In this respect it will be recalled that real per capita consumption expenditure per person per month in 1997 amounted to SL 3533 for the Southern region; SL 2917 for the NorthEastern region; SL 2858 for the Central region; and SL 3862 for the Coastal region. By 2004 all regions witnessed an increase in their standard of living, recording an annual rate of actual growth rate (in squared brackets) of 2.29% for the Southern region, 0.99%for the NorthEastern region, 3.05%for the Central region, and 6.1%for the Coastal region. Table 12: Propoor growth in Syrian regions, Real Per Capita Consumption Expenditure by Decile (L. S. per month) Year Real Per Capita Cons. Exp. (LS. / month) [2.29] (%) [0.99] (%) [3.05] (%) (%) [0.61] Source: ibid Comparing these actual growth rates with the effective growth rate (i.e. the average of the growth rates of deciles) over the period ,it is clear that for all regions the effective growth rate was less than the actual growth rate: about 2% compared to about 2.3% for the Southern region; about 0.8% compared to about 1% for the NorthEastern region; about 2.4% compared to about 3.1% for the Central region; and, about 0.4% compared to about 0.6% for the Coastal region. These results mean that the growth process during this period was antipoor in the sense that it was accompanied by deterioration in the distribution of consumption expenditure in each of the four regions. Thus, the earlier result pertaining to the level of the country is confirmed at the level of the regions, namely that the gains of the poor from the growth process that took place in the country over the period were comparatively less than those of the rich at the level of each region. Similar to the reading of the results in table 12, the last column of table 13 gives the effective annual per capita growth rate (i.e. the average of the growth rates of deciles) over the period 18

22 Coastal Central North Eastern Southern Regio n for each of the four regions. It also gives the actual growth rate in squared brackets. Comparing these two growth rates it is clear that for all regions the effective growth rate was greater than the actual growth rate albeit that the effective growth rates were negative for all regions except for the Coastal region: about (3.8)% compared to about ( 4.8)% for the Southern region; about (3.4)% compared to about (5.6)% for the North Eastern region; about (0.4)% compared to about (3.2)% for the Central region; and, about 0.3% compared to about (0.3) % for the Coastal region. These results mean that the growth process during this period was propoor in the sense that it was accompanied by an improvement in the distribution of consumption expenditure in each of the four regions. Table 13: Propoor growth in Syrian regions, Real Per Capita Consumption Expenditure by Decile (Ls. per month) Year Real Per Capita Cons. Exp. (LS. / month) (%) [4.82] (%) [5.57] (%) [3.17] (%) Source: ibid [0.25] Similar to the reading of the results in tables 12 and 13, the last column of the table 14 gives the effective annual per capita growth rate (i.e. the average of the growth rates of deciles) over the period for each of the four regions. It also gives the actual growth rate in squared brackets. Comparing these two growth rates it is clear that for all regions the effective growth rate was greater than the actual growth rate except for the Coastal region where the two growth rates were almost equal: about 0.2% compared to about 0.1% for the Southern region; about (0.5)% compared to about (0.9)% for the NorthEastern region; about 1.2% compared to about 0.8% for the Central region; and, about 0.3% compared to about 0.32% for the Coastal region. These results mean that the growth process over the period was propoor in the sense that it was accompanied by an improvement in the distribution of consumption expenditure in each of the four regions. Thus, the earlier result pertaining to the level of the country is confirmed at the level of the regions namely that the gains of the poor from the growth process that took place in the country over this period were comparatively more than those of the rich at the level of each region, albeit the gains were relatively small over this longer time period. 19

23 Coastal Central North Eastern Southern Region Table 14: Propoor growth in Syrian regions, Real per capita consumption expenditure by decile (LS. per month) Year Real Per Capita Cons. Exp (LS. / month) (%) [0.10] (%) [0.93] (%) (%) Source: ibid At the rural and urban levels 1.20 [0.85] [0.32] Using the information in tables 12 and 14 we can also have a look at the propoorness of the Syrian growth processes at the level of the two sectors of the economy. We start with looking at the urban sector reported in table 15. In this respect it will be recalled that real per capita consumption expenditure per person per month in 1996 in the urban sector amounted to SL 3558; by 2004 the sector witnessed an increase in its standard of living to SL 4192, recording an actual annual rate of growth of about 2.4%. By 2007, however, the sectors saw its welfare level decline to SL 3528, recording an annual rate of decline of about 5.6% over the period Over the longer period the urban sector also recorded an annual rate of decline of about 0.1%, implying a period of stagnation. Table 15: Propoor growth in the Syrian urban areas, Annual Effective Rate (%) [Effective Per Capita Expenditure (SL per month) Decile Rate (%)] Average Source: ibid 1.64 [2.37] 3.67 [5.59] 0.03 [0.08] The results in the last row of the above table enable us to compare the above noted actual annual rates of growth (in squared brackets) with the effective rates of growth (the average rate of growth of the deciles). For the period the table shows that the urban 20

24 effective rate of growth of about 1.6% was less than the actual rate of growth (about 2.4%). This means that the growth process during this period was antipoor in the urban sector: meaning that the poor in urban areas gained less than the urban rich during the period. For the period , which saw a decline in the standard of living of the urban sector, the table shows that the growth process was propoor in view of the fact that the effective rate of decline of per capita consumption expenditure (about 3.7%) was lower than the actual rate of decline (of about 5.6%). This means that the poor suffered less than the rich during the process of decline over this period. A similar result obtains for the longer period , which also recorded a decline in the welfare of the urban sector. The table shows that the growth process was propoor in view of the fact that the effective rate of decline of per capita consumption expenditure (about 0.03%) was lower than the actual rate of decline (of about 0.08%). While this means that the poor suffered less than the rich during the process of decline over this period, the extremely low value of two rates suggests that indeed the urban sector stagnated over the period without any significant movements in the welfare of urban residents. Having noted the above regarding the urban sector we now move to consider the propoor nature of the Syrian growth processes in the rural sector. In this respect it will be recalled that real per capita consumption expenditure per person per month in 1996 in the rural sector amounted to SL 2804; by 2004 the sector witnessed an increase in its standard of living to SL 2922, recording an actual annual rate of growth of about 0.6%. By 2007, however, the sectors saw its welfare level decline to SL 2728, recording an annual rate of decline of about 2.3% over the period Over the longer period the rural sector also recorded an annual rate of decline of about 0.3%, implying a period of stagnation. Table 16: Propoor growth in the Syrian rural areas, Decile Per Capita Expenditure (SL per month) Annual Effective Rate (%) [Effective Rate (%)] Average [0.59] 0.99 [2.26] 0.16 [0.25] Source: ibid The results in the last row of table 16 enables us to compare the above noted actual annual rates of growth (in squared brackets) with the effective rates of growth (the average rate of growth of the deciles). For the period the table shows that the rural effective rate of growth of about 0.68% was more than the actual rate of growth (about 0.59%). This means that the growth process during this period was propoor in the rural sector: meaning that the poor in urban areas gained more than the rural rich during the period. For the period , which saw a decline in the standard of living of the urban sector, the table shows that the growth process was propoor in view of the fact that the effective rate of decline of per capita consumption expenditure (about 1.0%) was lower than the actual rate of decline (of about 2.3%). This means that the poor suffered less than the rich during process of decline over this period. A similar result obtains for the longer period , which also recorded a decline in the welfare of the rural sector. The table shows that the growth process was propoor in view of the fact that the effective rate of decline of per capita consumption 21

25 expenditure (about 0.12%) was lower than the actual rate of decline (of about 0.25%). While this means that the rural poor suffered less than the rural rich during the process of decline over this period, the relatively low value of the two rates suggests that indeed the rural sector stagnated over the period without any significant movements in the welfare of rural residents. Historical Views and Prospective Analyses Any explanation of the underlying root cause of the shifts in inequality over the period since the mid1990s must take into account the major shift in macroeconomic policies which took place in Syria during that period. Historically, Syria has been long characterized as a highly centralized stateled economy, with an extensive importsubstitution industrialization (ISI) program, with pervasive restrictions covering foreign trade, foreign investment and with a limited level of export orientation (unlike East Asian Tigers). By virtue of the special relations between Syria and the former soviet bloc, a segment of the Syrian industry managed to break out and successfully penetrate East European and former USSR markets; however, the dismantling of the former USSR broke this peculiar economic relation. The political and economic ramifications of the collapse of the Soviet Union, coupled by the global onslaught of economic neoliberalism, provided a strong rationale for opening up the Syrian economy, but stronger pressures for economic liberalization grew out of a domestic balance of payments constraint during the mid1980s. In Syria, as elsewhere, ISI tended to foster mass consumption over capital accumulation, establishing the building of a national market as its main raison d être. This created dependency on imported capital goods and high tech goods, without fostering the export capacity needed to earn foreign exchange, thereby precipitating balance of payments difficulties. When aid from Arab countries and oil revenues declined, along with the fall of global oil prices in the 1980s, Syria s trade imbalance deteriorated sharply. The resulting fiscal and foreign exchange crises forced austerity cutbacks in public spending, and the state began to turn to the private sector to finance investment and create new jobs. In return, private business was awarded concessions, leading to the further opening of the economy in the early 1990s with legislation that allowed and encouraged private investment. Subsequently, the fiscal situation eased up considerably as oil exports grew to replace these traditional sources of foreign exchange. Against this backdrop, the first objective of this section is to explain the rather substantial shifts in distribution witnessed over the period from 1997 to 2007, first by giving an account for the first period ( ) during which the poor benefited proportionately less from growth, second by providing some possible explanations for the opposite trend witnessed in the second period ( ) wherein growth was propoor, and last, by explaining the long term inequality dynamics. The period from 1997 to 2004 As shown earlier, at the national level, from 1997 to 2004, growth was not propoor. Nonpoor individuals (above the third decile in the expenditure distribution) benefited disproportionally more than the poor from economic growth), thus inequality rose. However, large increases in per capita expenditure outweighed the effect of this worsening distribution as the fall in income poverty over the period was driven by a growth of per capita real expenditure of 2.0% per annum (Table 17). Yet, data from national accounts reviewed in UNDP (2005) suggest that per capita GDP growth between 1997 and 2004 was less than 1% per annum. The increase in average salaries, after adjusting for inflation, was also a 22

26 rather moderate 0.8% per annum. Moreover, a breakdown of per capita GDP by expenditure reveals private (real) consumption grew at only 0.3% per annum. 10 Turning to the sources of growth, annex tables 3 and 4 reveal that oil exports (as indicated by mining and manufacturing) and private consumption contributed the most to growth during the period. On the supply side, oilled growth fuelled a major expansion in services. However, over the period from 2000 to 2004, the contribution of the mining and manufacturing sector was negative due to declining oil production (6.37%). As argued in AbuIsmail et al (2005), the expansion in trade and services was essentially due to economic liberalization policies, in particular trade and capital account liberalization, reduction ofagricultural and public sector subsidies, deregulation of investment and financial sector reform 11. Table 17: Average and annual percentage change in per capita expenditure by region, in 2004 prices Average expenditure (SL) per capita in Region Gini coefficient Annual % growth Actual change Urban Southern NorthEastern Central Coastal Rural Southern NorthEastern Central Coastal Total Source: UNDP (2005) In this context of economic liberalization and urbancentred serviceled growth, the major beneficiaries were urban areas, which were best situated to take advantage of the shift in economic policy regime. This is clearly confirmed in table 17 above which shows that urban areas experienced significantly larger growth in real per capita expenditure and rise in inequality compared to rural areas. The period from 2004 to 2007 distribution dynamics over the second period indicate an opposite trend. Distribution improved significantly particularly for the poorest of the poor. At the same time, however, growth in expenditure faltered. Thus, monthly real per capita expenditure (2004 prices), at the national level, declined from SL3541 in 2004 to SL3141 in This represents an annual decrease in the average real per capita expenditure (expressed as a welfare ratio to the value of poverty line) of 3.92% (Table 18). As in the earlier period, this does not sit well with the story coming from national accounts data which shows GDP growth has picked up since 2004 despite the rapid depletion of oil reserves mainly due to demand stimulus, which in turn is partially explained by the large presence of Iraqi refugees. Iraqis presence in Syria is therefore considered as a favourable factor for economic growth since the majority live on savings and/or transfers from a third country. This de facto amounts to a kind of export of services (tourism) to the Syrian economy. Hence, it is conventionally assumed 12 that the Iraqi refugees had significant positive impact on growth via private consumption which mitigated the negative shock from declining oil revenues 23

27 Table 18: Average and annual percentage change in per capita expenditure by region, in 2004 prices Average expenditure (SL) per capita in Region Gini coefficient Annual % growth Actual change Urban Southern NorthEastern Central Coastal Rural Southern NorthEastern Central Coastal Total Source: UNDP (2009). Notwithstanding the discrepancy between national accounts and survey data, it is clear that any explanation for the distributional shifts over the period must take into account the fact that not all regions experienced the same rate of decline in expenditure and inequality. In this regard, three questions warrant further attention. First, what caused the significant decline in inequality in the NorthEastern rural region? Second, what caused the significant decline in growth in the NorthEastern region (both urban and rural) and Southern urban regions? Third, what has been the impact of the Iraqi refugees on poverty and inequality? Regarding the first question, there are two possible explanations. One is that the government interventions have been effective in reaching the poorest of the poor in that region and thus decreased the gap between the rich and poor. Although this explanation is backed by official policy, which has since 2005 identified the North Eastern rural region as the primary target of state intervention in the 9th five year plan, it is less probable, as the midterm review of the plan (2006 to 2010) has raised several questions regarding its effectiveness in reaching targeted beneficiaries in that region. In addition, as noted by the IMF (2007), over the period from 2004 to 2007, fiscal policy had to contend with a cumulative loss of oil revenues in excess of 10% of GDP, owing both to a sharp contraction in oil production and rapid growth in the consumption of subsidized petroleum products (reflecting, in part, increased smuggling and the impact of the Iraqi refugees). A cumulative improvement of 7% of GDP over narrowed the nonoil budget deficit to about 10% of GDP in However, the bulk of the adjustment was achieved through rationalization of investment spending and taxation reform. This implies that the fiscal adjustment has been mainly antipoor. A more plausible explanation for the reduction in inequality should take into account internal migration dynamics which took place from the North East rural region as a result of the successive droughts. If, as one would expect, the migration pattern was characterized by an overwhelming majority of the poorest strata, then this would in effect play a major role in reducing inequality in that region, while simultaneously reducing per capita expenditure growth at the destination region. Given that the majority of these poor migrants are more likely to have targeted the Southern urban region (as discussions with Syrian officials suggest), this would go a long way in answering the second question. 24

28 Figure 4: Distribution of Iraqis by governorate as of July 2007 Rural Damascus Aleppo Homs Other Damascus 23% 6% 4% 62% 5% Source: AbuIsmail et al (2008). The third question is perhaps the most difficult to address. At the outset, as shown below in Figure 4, there is little dispute among Syrian officials that the influx of Iraqis has been concentrated in the Southern urban region and that it has increased demand for state subsidized goods and services 13, which may explain the fact that this region is the only one which witnessed a deterioration in human poverty index over the period from 1997 to One possible hypothesis is that higher demand for consumption goods and the housing sector by Iraqis was the primary cause of the unprecedented inflation since 2003, which was the root cause behind the decline in real expenditure. Such a conviction is not unfounded, especially if we take into account that the Syrian economy was under a tight price control regime due to high levels of direct and indirect government subsidies, and it experienced relatively few episodes of doubledigit inflation during the past few decades. It is important, therefore, to underline the actual factors that have been driving the rate of inflation up. However, this theory is challenged by the composition and distribution of inflation. Data suggests inflationary pressures started to build up in mid2005, and the average inflation increased to 10% in According to the CBS, half of the increase in the latter is attributed to higher food pricing due to weatherrelated shortages in domestic production and higher demand in food prices in the international markets. In addition, although it is true that from 2001 to 2003 a big jump took place on the rent index due to a legislative reform which liberalized rents and eased rent control, housing inflation in the Southern urban region does not appear to be significantly higher than in other regions 14. In short, there have been many internal and external factors that contributed to the high rates of inflation in Syria since 2003.One of the reasons can be the influx of Iraqis, but it is neither the only one nor necessarily the most important one at the aggregate or subnational levels. A more plausible theory is that the large influx of Iraqi refugees and internal migrants from the NorthEastern rural region may have exerted significant pressure on wages in the informal sector of the Southern urban. This may have been an important determining factor for growth and inequality, in decreasing the growth rate in the urban sector of the Southern region and in decreasing the Gini coefficient of the NorthEastern region. In this regard, one of the important findings in the UNDP (2009) poverty assessment report is that the percentage of unpaid workers increased by almost 50% among poor individuals in urban areas from 2004 to On the other hand, the percentage of unpaid workers declined for both poor and non poor groups in rural areas. Another important feature that could explain the significant increase of poverty in the urban Southern region is the disproportionate rise in irregular work (seasonal, temporary or occasional work) amongst the poor (and non poor) in that region. As shown in Figure 5, the irregular workers among poor persons in urban Southern region increased from 13.6% in 2004 to 20.2% in The corresponding figures 25