Khalid Abu-Ismail, Gihan Abo Taleb, Jennifer Olmsted and Mohamed Mohieddin

Arab Development Challenges Background Paper 2011/05 The ADCR 2011: Employment, Vulnerability, Social Protection and the Crisis of Arab Economic Reforms Khalid Abu-Ismail, Gihan Abo Taleb, Jennifer Olmsted and Mohamed Mohieddin 1

United Nations Development Programme Arab Development Challenges Report Background Paper 2011/05 The ADCR 2011 Employment, Vulnerability, Social Protection and the Crisis of Arab Economic Reforms Khalid Abu-Ismail, Gihan Abo Taleb, Jennifer Olmsted and Mohamed Mohieddin* Khalid Abu-Ismail is Poverty and Macroeconomics Adviser at United Nations Development Programme- Regional Centre in Cairo (UNDP-RCC). E-mail: khalid.abu-ismail@undp.org Gihan Abou Taleb is Research Assistant at UNDP-RCC, E-mail: gihan.ahmed@undp.org Jennifer Olmsted is Associate Professor of Economics at Drew University E-mail: jolmsted@drew.edu Mohammad Mohieddin is Social Policy Consultant at UNDP-RCC. E-mail: mohamed.mohieddin@undp.org Comments should be addressed by email to the author(s) i

Acronyms and Abbreviations AC AHDR ALO AMGDR CIS DR EAP ECA EU GCC GDP GFCF ILO IMF LAC LDC OECD OPT SAS SMEs SSA UAE UN WDI Arab countries Arab Human Development Report Arab Labour Organization Arab Millennium Development Goals Report Commonwealth of Independent States Developing Region East Asia & Pacific(developing countries only) Europe &Central Asia(developing countries only) European Union Gulf Cooperation Council Gross domestic product Gross Fixed Capital Formation International Labour Organization International Monetary Fund Latin America & Caribbean (developing countries only) Least developed countries Organization of Economic Co-operation and Development Occupied Palestinian Territories South Asia (developing countries only) Small and Medium Enterprises Sub-Saharan Africa(developing countries only) United Arab Emirates United Nations World development Indicators ii

Introduction 1 As noted in Abu-Ismail et al. (2011), a central puzzle emerges from the human development and poverty stylized facts: despite being healthier and more educated today than they were in the 1990s, Arab men and women today are unable to utilize their enhanced capabilities in reducing consumption poverty. Grasping the roots of the divergence between the growing level of human capabilities (or the commensurate decline in human poverty) on the one hand and stagnating consumption poverty on the other hand is crucial to understand the nature of the core Arab economic development challenge: structural retardation. In Abu-Ismail et al. (2011) 2, we examined trade and industrial sector performance and policies as one of the root causes and manifestations of structural retardation. In this paper, we examine the ramifications on demand for productive employment and ensuring decent and productive work for all, which is a primary social demand in the region and an essential element for realizing inclusive and sustainable growth. The main objective of this paper is thus to draw policy attention to the need of reconsidering priorities in the region by making employment a central goal of development policy. And the main thesis the paper advances further 3 is that the critical constraint for creating decent jobs has not been on thelabour supply side and lack of available skills but on thelabour demand side. The economic reforms many countries adopted resulted in a few high paying jobs in sectors such as finance, banking, and real estate and many labour-intensive low productivity jobs elsewhere. The paper is structured as follow: the first section reviews the demographic challenges facing the region and labour market conditions through indicators such as labour participation rate, employment-to-population, and unemployment rates based on recent data by the ILO. The second section estimates the projected number and cost for the new jobs to be created under three different scenarios and discusses the macroeconomic constraints to a full-employment policy. Issues pertaining to social protection, vulnerability and the employment poverty nexus are discussed in the third and final part of this paper. Demographic and Employment Stylized Facts The demographic challenge For almost six decades now Arab countries have been experiencing dramatic changes economically, demographically and in turn in theirlabour markets. Demographically, the average life expectancy at birth in all countries in the region is 70 years or higher, except for the five poorest countries, namely Mauritania (55), Comoros (64), Djibouti (54), Sudan (56.7) and Yemen (61.7). With an average of 3.1 children per woman, fertility rates in the region are substantially above replacement level; the highest fertility rate is observed in Yemen (5.9) and the lowest in Tunisia (2.0), followed by Lebanon (2.3), Kuwait and the United Arab Emirates (both 2.4). 4 The Arab region is thus at a relatively early stage of its demographic transition - that is the parallel transition from high to low fertility and from high to low mortality- resulting in an evolution in the age structure of population in favor of the youth and growth of the working age population. Accordingly, The Arab region is still one of the fastest growing and youngest populations in the world despite the decline in the average annual growth rate from 3.1 to 2.3% over the past two decades 5 (with the exception of Sub- Saharan Africa 2.5%); with children under age 15 account for almost one third of the population and young persons (aged 15-24 years) account for one fifth. In other words, approximately half of the population is under the age of 1

1970 1980 1990 2000 2010 2020 2030 2040 2050 1950 1960 1970 1980 1990 2000 2010 2020 2030 2040 2050 2060 25. This ratio varies across Arab sub-regions but it generally declined since 2000. The decline was sharp in the Maghreb indicating a more rapid pace of demographic transition (Figure 1.A). Moreover, the region is also witnessing a sustained pattern of increase in the working-age population. As illustrated in figure 1.B, the share of working-age population (15-64) in total has increased from 51% in 1970 to 62% by 2010 and is expected to peak at 66% in the year 2040 and decline to 65% by 2050. As the onset of fertility decline is relatively recent, the process of ageing is still one decade away as the share of people aged 65 and above is expected to increase more rapidly by 2020 so that by 2050, the proportion of older persons (aged 65 or over) is projected to quadruple compared to 1980. Figure 1: Age structure of population for Arab countries (A), share of population aged (15-64) to total population by sub-region (B), 1970-2050 (A) (B) aged 0-14 (%) aged 15-24 (%) Arab Countries GCC 100 80 60 40 20 0 aged 25-64 (%) aged 65 or over (%) 80 70 60 50 40 Mashreq LDCs Maghreb Source: Authors estimates based on data from World Population Prospects, DESA Note: Arab countries regional average include the 22 countries The increase in the Arab region youth population serves as both an opportunity and challenge. Countries with a large youth population have a unique demographic window of opportunity, as a large number of young persons would lower the dependency rate. A more economically active young population would not only increase productivity, but also diminish the burden imposed by non-working cohorts (elderly and children). Yet as this group's share will decline in the near future, this window of opportunity will eventually close. Addressing the employment challenges faced by the Arab youth is thus of central importance for Arab economies to reap the benefits of the demographic gift. Labour force participation trends As one would expect from the rapid growth of the working age population (expected to almost double in 2050), the region has one of the highestlabour-force growth rates in the world, exceeding 3% per annum. Thus, Arab countries have witnessed a considerable increase in the share oflabour forces in the total working age population over the period 1995-2009. 6 Thelabour force participation rate varies from 84 per cent in Qatar to 41 per cent in Iraq. This means more than 4 million new entrants to thelabour market in 2009 with more than half belonging to four Arab countries (Egypt, Sudan, Morocco, and Algeria). However, the share of labour force in total population (labour-force participation rate) for the region is still the lowest world-wide (Figure 2.A). This is primarily due to lower rate of participation for women. For example, the highest labour force participation rates are found in GCCs, yet these countries also have male participation rates almost three times higher than the female rates (82% and 26%, respectively). A considerable gap between male and female is found in Mashreq region, while LDCs and Maghreb have a female rate that is approximately half of the male rate (Figure 2.B). This indicates that the Arab women are still facing high barriers tolabour market entry in the region. 2

Figure 2: Labour force participation rates (15+) Arab region versus other developing regions (A), Arab sub-regions (B), 2009 (A) (B) 90 80 70 60 50 40 30 20 10 0 F M MF AC DR EAP ECA LAC SAS SSA 90 80 70 60 50 40 30 20 10 0 F M MF GCC LDC Maghreb Mashreq AC excl. GCC AC Source: KILMnet Note: Arab Countries included are GCC (Bahrain, Kuwait, Oman, Qatar, Saudi Arabia and UAE), LDCs (Comoros, Mauritania, Somalia Sudan and Yemen), Maghreb (Algeria, Libya, Morocco and Tunisia) and Mashreq (Egypt, Iraq, Jordan, Lebanon, Opt and Syria) Young people, who comprise a significant proportion of the population, also have a low participation rate. It is expected that the young population would have a lower participation rate than adults since a sizable proportion will be in schools and universities. However, the Arab youth labour participation rate, at 52%, is much lower than the prevailing rate for developing regions (61%). The gap is even more significant for the female youthlabour participation rate (26.3 compared to 53% in developing regions). As figure 3.A indicates, only three out of every ten Arab young women are economically active. At the country level, Jordan, Saudi Arabia, Occupied Palestine and Iraq have the lowest femalelabour participation rates; well below 10%.Comoros, Somalia, and Mauritania have the highest female youth participations but this mainly because most of them are engaged in agriculture activities. Figure 3: Labour force participation rates (15-24) Arab region versus other developing regions (A), Arab sub-regions (B), 2009 (A) (B) 70 60 50 40 30 20 10 0 F M MF AC DR EAP ECA LAC SAS SSA Source: ibid Note: Arab Countries included are GCC (Bahrain, Kuwait, Oman, Qatar, Saudi Arabia and UAE), LDCs (Comoros, Mauritania, Somalia Sudan and Yemen), Maghreb (Algeria, Libya, Morocco and Tunisia) and Mashreq (Egypt, Iraq, Jordan, Lebanon, Opt and Syria) Employment trends 70 60 50 40 30 20 10 0 F M MF GCC LDC Maghreb Mashreq AC (excl.gcc) Between 1991 and 2009, employment in Arab countries grew at an impressive average annual rate of 3.3%. This rate is considered the highest among developing regions (led by AC 3

1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 the growth rate of employment in GCCs at 3.9%); it explains the upward shift in the employment-to-population ratio 7 from 44.2% to 46%, this rate still stands as the lowest in the world. 8 Women have benefited more from the rise in employment with an average annual growth of 4.4% compared to 3% for men. This has somewhat reduced the gender-gap in the Arab employment-to-population ratio, but the gap still the widest among developing regions. According to ILO data, in 2008 five of the ten lowest employment-to-population ratios were found in the Arab region, namely in Iraq, Jordan, Tunisia, Yemen and the Occupied Palestinian Territories. The gap between male and female employment is also highest in these countries. 9 For Arab sub-regions, this ratio is highest for GCC countries at 57% and the lowest in the Mashreq at 42%. 10 Figure 4.B confirms the significant gender disparities that accompany low employment-to-population ratios. Figure 4: Employment to population ratio Arab sub-regions (A), and by gender (B) 1991-2009 (A) (B) 60 GCC LDC Maghreb Mashreq AC 90 80 F MF M 55 50 45 40 35 Source: ibid. Note: Arab Countries included are GCC (Bahrain, Kuwait, Oman, Qatar, Saudi Arabia and UAE), LDCs (Comoros, Mauritania, Somalia Sudan and Yemen), Maghreb (Algeria, Libya, Morocco and Tunisia) and Mashreq (Egypt, Iraq, Jordan, Lebanon, Opt and Syria) Employment growth for young people has also been quite impressive in Arab countries, which recorded the second largest growth rate (following Sub-Saharan Africa) among developing regions. Yet, as their population growth was significantly above the growth in employment, job creation was not sufficient, particularly for youth as the overall youth employment-to-population ratio for the Arab region fell from 27.2 to 24.4% in the period from 1991-2009. Again, this is the lowest among developing regions (which had an average of 45% in 2009). Furthermore, the gender gap in employment is equally striking for the youth. In 2009, out of every 100 young women, only 11 were employed, compared to 36 out of every 100 young men. Arab youth are not only more numerous, they are also more educated, and marry at a later age than before. This implies an increase in the productive capacity of the population. Over the last few decades, school enrolment rates have risen markedly throughout the region for both young men and women. 11 The numbers, however, have not been translated into higher rates of employment and wages. Some have argued this is mainly due to the low quality of education and the mismatch between educational outcomes and market demand. As argued in this report, we believe the dynamics of oil led growth, which skewed growth away from productive employment, is the principal culprit. Unemployment, gender and education According to the ILO, the Arab region continues to have the highest unemployment rates in the world (9.3% in 2001-2011 versus 6.6% for developing regions), though the rate was lower than during the 1990s (12%). This decrease in the unemployment rate is largely due to 70 60 50 40 30 20 10 0 GCC LDC Maghreb Mashreq AC excl. GCC AC DR 4

the substantial decline witnessed in the Maghreb region (from an average unemployment rate of 21% to 11%) driven mainly by the decline in unemployment in Algeria. In contrast, LDCs experienced an increase in unemployment rates (from 8% to 11%). The unemployment rate in GCC and Mashreq countries also increased slightly (from 4.5% to 4.6% and 9.3% to 9.8% respectively) (Figure 5.A). 12 Figure 5: Overall unemployment rates for Arab sub-regions (A) and versus other developing regions (B), 1990-2011 (A) (B) AC Mashreq Maghreb LDC GCC 1990-2000 2001-2011 4.5 4.6 9.3 9.0 8.8 8.3 12.1 11.9 11.5 22.2 0 5 10 15 20 25 Source: Authors estimates based on ILO Kilmnet dataset and Tunisia National Labour Survey 2011, National Institute of Statistics) Notes: 1) Arab countries included are GCC (Bahrain 1991 and 2001, Kuwait 1995 and 2005, Saudi Arabia 2002 and 2008, UAE 1995 and 2008, LDCs (Yemen 1994 and 1999), Maghreb (Algeria 2001 and 2008, Morocco 1990 and 2009, Tunisia 1997 and 2011 and Mashreq (Egypt 1998 and 2007, Lebanon 2004 and 2007, Syria 2002 and 2010). As expected, unemployment is noticeably concentrated among the youth in the Arab region (15-24 age groups) (Figure 5.B) where youth unemployment based on estimates derived from ILO and UN data reached 24% in 2005-2011 (more than double the world average of 11.9%) accounting for more than 50 per cent of the total Arab unemployed. One of the main reasons for high youth unemployment rate, is the transition from school to work, particularly among the higher educated. It can be seen from figure 6, workers with secondary and tertiary education has the highest unemployment rate. Figure 6: Youth unemployment rates for Arab sub-regions (A) and versus other developing regions (B), 1990-2011 (A) (B) AC Mashreq Maghreb LDC GCC 1990-2000 2001-2011 17.9 18.7 29.2 23.8 24.0 23.7 25.5 22.8 23.3 41.0 Source: ibid. 0 10 20 30 40 50 18 16 14 12 10 8 6 4 2 0 40 35 30 25 20 15 10 5 0 Total Male Female EAP SAS World LAC SSA ECA AC Total Male Female EAP SAS SSA World LAC ECA AC 5

GCC LDC Maghreb Mashreq AC GCC LDC Maghreb Mashreq AC Figure 7: Distribution of unemployment by level of education, 1997-2007 80 primary secondary tertiary 60 40 20 0 Egypt 2006 KSA 2009 Kuwait 2006 Bahrain 2007 Source: WDI and Human Development report, 2010 for Egypt UAE 2005 Tunisia 2005 Morocco 2005 Algeria 2004 As in the case of labour participation, Arab unemployment is gender biased. Figure 8 shows unemployment rates for Arab women are the highest when compared to world regions and the gap between male and female remains high (though it has recently narrowed). Arab women unemployment rate accounted for almost double that for Arab men (18.5% and 7.5 respectively). The problem is particularly acute amongst younger women who are joining the labour force in ever larger numbers due to the spread of higher education and rise in the age of marriage. The average unemployment rate for young Arab women is estimated at 36% according to most recent surveys (2004-2009) compared to 20% for men. Figure 8: Total (A) and youth unemployment rates (B) for Arab sub- regions by gender, 2001-2011 (A) (B) MF M F MF M F MF M F MF M F MF M F 3.2 4.6 6.0 7.6 8.2 9.3 8.8 11.9 11.2 11.5 11.6 12.5 14.1 15.5 18.7 0 5 10 15 20 Source: ibid Note: 1) Arab countries included are same as figure 7, 2) for Tunisia and Algeria same share of male and female to total youth unemployed and total labour force were used for computing youth unemployment figures for 2008 and 2011 respectively. The above data reflects more than the failure of Arab economies to generate sufficient jobs; it points as well to the social biases against women (AHDR 2009). While there seems to be a high correlation between women s educational levels and increased labour force participation rates, even educated women have difficulty finding decent jobs. This is reflected in the unemployment figures as well as the growing phenomenon of discouraged workers 13 in the region. Higher education is not a guarantee for access to quality jobs, especially for women. MF M F MF M F MF M F MF M F MF M F 13.5 17.0 20.4 18.7 20.5 19.4 23.8 23.7 25.5 25.5 25.5 23.3 35.2 35.9 45.9 0 10 20 30 40 50 6

Figure 9: Firms training (%) 60 57 50 40 30 20 10 0 Source: Tzannatos (2011) based on surveys covering 48,000 firms in 99 developing countries 2002-2007 quoted in Almeida and Aterido (202010). The Arab countries includes Algeria, Egypt, Jordan, Lebanon, Morocco, occupied Palestinian Territories, Oman and Syria The limited labour demand for educated workers and skills can be inferred from many different symptoms in the labour market. First, educated workers get a relatively low wage premium when employed and the Arab region has one of the highest skilled emigration rates in the world 14 suggesting they have skills that are in demand elsewhere, often in more technologically advanced countries (e.g. OECD economies or better jobs in the GCC economies).second, contrary to what would expect if firms faced skills shortages (a common complain of employers pointing to the low quality and irrelevance of the education system) the percentage of firms providing training is the lowest in the Arab region. A related comment on the incidence of unemployment is that unemployment in the Arab region is present both in low income and high income households. 15 This is not an unexpected finding as it correlates to the high unemployment rates of more educated workers who come predominantly from better off families. And lends further support to the views, on the one hand, that the economy does not generate decent jobs acceptable to most workers, let them be youth or adults or youth and men or women and, on the other hand, the generalized resentment of the citizens that led to the explosion of social unrest. Figure 10: Percentage of unemployed with unemployment benefits (contributory and noncontributory), 2011 Contributory schemes Non-contributory schemes World Western Europe Source: Tzannatos, Haq and Schmidt, 2011. Note: Regional estimates are weighted by labour force, 2009 53 In addition to the limited concerns about the effects of the Arab economic reforms about the quality of jobs that were being created, there was little protection for those who were unemployed. Only a few countries, such as Algeria and Tunisia, have unemployment insurance schemes. Bahrain also extends unemployment benefits to first-time jobseekers. Kuwait introduced a social assistance scheme for first-time jobseekers in 2003 whereby nationals who declared they wanted to work but were unemployed received a monthly allowance for a period of up to one year. Jordan, Syria and Saudi Arabia are considering the 40 26 25 EAP LAC EAC SSA AC North America Latin America CIS Central and Eastern Europe Asia Arab Countries Africa -20 0 20 40 60 80 7

GCC LDC Maghreb Mashreq Arab Countries GCC LDC Maghreb Mashreq Arab Countries introduction of various forms of unemployment insurance/assistance. Still, the coverage of unemployment insurance in the region remains the lowest in the world (Figure 10). There are also sharp gender-related demarcations in sectoral employment. Figure 11 depicts the gender distribution of employment by sectors for the Arab region. It shows significant gender disparities. Women are mainly concentrated in social, personal and community services especially in GCC and Mashreq region; while most women work in agriculture in LDC and Maghreb countries. It can also be seen from the figures that the distribution for men is more diversified than women. As noted in the first Development Challenges report, such patterns of employment reinforce to a great extent the traditional division of labour. More importantly, it deprives Arab societies from productive and innovative capacities that could be gained in labour markets. Figure 11: Distribution of employment by economic sector for males (A) and females (B), 2006-2008 (A) (B) 100% Not Classified Not Classified 90% 100% Service Service 90% 80% Activities Activities 80% 70% Financial Financial Services 70% Services 60% Tranport 60% Tranport 50% 50% 40% Trade Trade 40% 30% 20% 10% 0% Construction Utilities Manufacturing 30% 20% 10% 0% Construction Utilities Manufacturing Mining and Quarrying Agriculture Mining and Quarrying Agriculture Source: Arab Labour Organization (ALO) Notes: All Arab countries are included except United Arab Emirates, Oman, Palestine, Djibouti and Comoros. Macroeconomics Policies: the case for a Renewed Commitment to full Employment Number of jobs and investments required to reach full and productive employment by 2030 In light of the afore-mentioned stylized facts about the relatively high unemployment rate in the Arab region especially among women and youth and the lowlabour participation rates, it is obvious that Arab countries as a group face the challenge of creating employment opportunities in a sustained manner. In this section, three different employment scenarios were analyzed. The first business-as-usual scenario projects the number of jobs required if the Arab Region is to keep the unemployment rate constant with no change in the labour force participation rate. In Scenario 2 we estimate the number of jobs required to reach full employment. Scenario 3, which is the best case scenario, calculates the number of jobs required to reach full employment and raise the women labour force participation rate to reach 35% (which is still 15% below the average for developing regions). 8

GFCC/GDP growth rate (%) 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 Total employment growth rate (%) To do this exercise we relied on the projected rates of growth of the working-age population by the UN and unemployment rates, labour force participation rates and employment to population ratios reported by the ILO. The number of jobs required under the three scenarios is estimated over three periods: 2011-2015, 2011-2020 and 2011-2030. The order of magnitude of the results indicates an overwhelming task ahead of Arab countries. According to the first scenario, the region (represented by Algeria, Egypt, Jordan, Lebanon, Mauritania, Morocco, Syria, Tunisia and Yemen) needs to generate 6.1, 13.6 and 28.6 million jobs by 2015, 2020 and 2030, respectively. For the second scenario, the number of jobs required increases to 15.4, 24.1 and 40.8 million, while for the third scenario; it reaches a staggering 24.7, 47.1 and 91.8 million for the periods up to 2015, 2020 and 2030, respectively. 16 Figure 12: Growth of employment and gross fixed capital formation in Arab countries, 1992-2009 GFCF/GDP Total Employment Linear (GFCF/GDP) Linear (Total Employment) 30% 25% 20% 15% 10% 5% 0% -5% -10% -15% -20% -25% -30% Source: Authors estimates based on KILMnet and WDI Employment creation is often closely associated with output growth and fixed capital formation. This means that unsatisfactorily labour market outcomes are primarily due to unfavorable macroeconomic conditions that inhibit investment in fixed capital and productivity growth, as well as to inadequate growth of labour income, which constitutes the most important source of domestic demand. Figure 12 shows the growth in gross fixed capital formation as a share of GDP (constant 2005 prices) versus the total employment for the Arab region. It shows a constant trend for employment growth accompanied by a slightly positive improvement in gross fixed capital formation. This suggests a declining employment generating capacity of investment. In addition, Figure 13 shows the other side of the coin: Productivity gains over time have been lowest in the Middle East region and not significantly higher in North Africa Figure 13: Annual rate (%) of productivity growth by region, 1991-2010 Contributory schemes Non-contributory schemes World Western Europe North America Latin America CIS Central and Eastern Europe Asia Arab States Africa 0 10 20 30 40 50 60 70 Source: Tzannatos (2011) based on Trends Econometric Models, Geneva, ILO October 2010 Using the above data on employment, we can easily estimate the order of magnitude of investment needed to generate the required job opportunities. By dividing the sum of total gross fixed capital formation (constant 2005 prices) during the time period of 2004-2009 by the total number of new jobs in the same period, we observe a relatively high average 8% 6% 4% 2% 0% 9

1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 investment cost per employment opportunity created at approximately $100,000 (but still significantly lower than Latin America, Europe and East Asia). Excluding the GCC, where the investment required per job exceeds $400,000, the regional average drops to approximately $48,000. Simple calculations show that the total investment bill for the region (excluding the GCC) required to reach full and productive employment by 2030 (scenario three) will reach $4.4 trillion (in constant 2005 prices). This implies an average annual investment bill of $220 billion, which is 50% of the GDP of these countries in 2009 (against an average investment to GDP ratio of 27.8% for 2004-2009). The required rate of investment is substantially higher for LDCs at around 100% of GDP, clearly requiring significant external support for its realization. For middle income countries, however, the desired rate of investment is well within reach at around 40% of GDP. 17 Macroeconomic policies and the crisis of employment 18 A recent ILO working paper by M. Muqtada (2010) empirically anchors the evolution of macroeconomic thinking over the past quarter century from its neo-liberal origins to the present day s search for a more employment-centred macroeconomic framework. The paper assesses evidence on the relationships observed between a number of macroeconomic stability variables (inflation, budget deficit, investment and current account deficit) and GDP growth based on panel data for 80 developing countries. The five main stylized facts which the paper identifies for developing regions as a whole based on cross-country evidence are as follows: 1. nearly two decades of stabilization reforms have in fact reduced the inflation levels in most countries of the world; 2. with the exception of a few countries, growth of GDP is observed to be low or inadequate during the 1980s and much of the 1990s; 3. despite substantial declines in inflation rate and budget deficit, hence arguably better macroeconomic stability, investment-to-gdp ratio (I/GDP) has failed to gain momentum; 4. investment, and more specifically the investment to GDP ratio as a single variable tends to explain growth better than the other stability variables; and 5. Current account deficit is negatively related to growth but not robustly. Figure 14: Public investment as a share of GDP (A-B), budget surplus and GDP (c-d) and current account balance and GDP (E-F) in Arab countries and developing countries, respectively, 1980-2009 (A) (B) 35 Mean Median 35 Mean Median 30 25 20 15 30 25 20 15 10

y = 0.2127x + 5.9535 R² = 0.3485 LBN (C) 16 14 12 10 JOR 6 EGY OMN TUN MAR 4 (E) 8 BHR DZA Source: Authors estimates based on WDI 8 2 QAT KWT 0-15 -10-5 0 5 10 15 y = 0.0614x + 4.1617 R² = 0.1486 WBG SDN BHR 6 JOR OMN LBN TUN MAR EGY 4 SYR DJI YEM 2 0-2 IRQ LBY KSA KWT -4-40 -20 0 20 40 y = 0.0858x + 5.2008 R² = 0.0132 LBN (D) 18 16 14 12 10 8 JOR 6 EGY OMN TUN MAR 4 2 (F) BHR DZA QAT KWT 0-15 -10-5 0 5 10 15 y = 0.0466x + 4.6956 R² = 0.0403 8 SDN KWT 6 JOR BHR MAR OMN LBN TUN EGY 4 SYR LBY DJI YEM SAU 2 0 IRQ -60-40 WBG -20-2 0 20 40 18 16 14 12 10-4 Based on trends observed in figure 14, we can easily conclude that many of these stylized facts equally apply to Arab countries albeit with a slightly higher correlation with growth for the current account and budget surplus. The decline in public investment for the Arab region, however, was sharp since 1980 before rebounding again along with the trend for developing regions, by 2000. The limited explanatory power of the individual stability indicators further confirms the significant role of investment in explaining growth, which opens up a host of issues with respect to what, in fact, explains the growth in investment, and why despite all efforts at stabilization, investment failed to pick up. The experiences of the East Asian growth success have shown us that macroeconomic stability was only one of the objectives of their macropolicy; financial policies and incentives towards resource mobilization as well as focused public interventions helped boost saving and investment in these economies, an issue which is picked up in further detail in Roy et al. (2011). 19 Disenchantment with the stabilization-focused policy reforms produced debates on the appropriate design of macroeconomic strategy. However, as noted by Muqtada (2010), the debates within mainstream circles were diverted, away from the core objectives of macroeconomic policy, to why stabilization pursuits at the country-level were not yielding the expected output and outcomes. Attention was thus deflected toward the pace and sequence of stabilization; to rigid labour markets; and, not least, to the lack of good governance. Meanwhile, the long record of low and inadequate growth, and widespread unemployment and poverty, did not receive commensurate attention. The consequences of the manner in which stability was achieved were, as many have noted, eventually led to disenchantment with policy reforms. The instruments that were used to attain stability had clear trade-offs; the consequences of these trade-offs were not properly 11

assessed or anticipated. Restrictive monetary policies often led to hikes in interest rates which adversely affected investment; the elimination of import restrictions, apart from depriving governments of substantial revenues, affected industries and a country s exportcapacity; reduction in government expenditures badly affected development and various social transfer programmes, thus inhibiting growth of physical and social expenditures. 20 A steady decline in investment, a distinct fall in private consumption and public expenditures, and sluggish oil export performance these coalesced to produce a fall in aggregate demand. Thus as Muqtada (2010) argues, dynamic efficiency has been traded off with static gains. Stabilization reforms in developing regions thus led to a crisis of a different kind, a crisis of employment. In Arab countries, this trend was accentuated by a virtual freeze on public employment as a vent for demand in almost all non-gcc countries. In a scenario where growth and employment are badly affected, pro-cyclical policies, to attain strict stabilization targets, can only aggravate the situation. Unfortunately, the Washington consensus and the conservative macroeconomic policy stance continue to feature strongly in the Bretton Woods advocacy even during this present economic catastrophe. At the G20 London Jobs Summit in 2009, the IMF has sent out a strong caution to countries to adhere to macroeconomic stabilization, while at the same time acknowledging the necessity of massive, coordinated fiscal stimulus packages pronounced by the advanced countries. For the latter, these anti-cyclical measures are seen as temporary, time-bound and targeted. Nevertheless, it would appear as double standards since the developing world, facing similar growth and job crisis, is de facto denied access to such measures. 21 A recent staff position paper by Blanchard et al. (2010) provides possibly the first serious introspection by IMF on what has been a long and rigid IMF policy position with respect to macroeconomic strategy. The paper laments that we (the IMF) thought of monetary policy as having one target, inflation, and one instrument, the policy rate (p.3), thereby acknowledging (as argued in this paper and by the ILO) that there can be multiple targets in a macroeconomic strategy, and multiple instruments to attain these targets. In particular, the paper relents on the singular focus on inflation. Following the Keynesian principle, the paper also recognizes the role of fiscal policy as a significant tool, especially in times of crisis when countercyclical measures are warranted, and when monetary policy often reaches its limits, as observed in Japan during the 1990s, and now, in the advanced economies during the 2008-09 crisis. As argued by Muqtada, the paper s call for exploring the consequences of raising the target of policy rate from existing 2 per cent to 4 per cent may appear sensitive to the developed countries but, such inflation targeting will still remain insensitive to the conditions prevailing in the developing countries. In fact, as argued by the IMF there is an inflation threshold of 11-12 per cent below which growth will not be hampered by inflation. 22 The IMF stocktaking thus fails to acknowledge a distinction between inflation targeting and targeting a stable inflation level. The arguments above, clearly suggest that there needs to be a serious rethinking on the goals of macroeconomic policy, as well as on the instruments through which the goals can be attained. Adapting from Muqtada (2010), five policy considerations can be identified for Arab countries: First, a foremost imperative of macroeconomic policy design is a renewed commitment to the goal of full employment. In the case of Arab countries, where job-generation towards poverty reduction will continue to dominate the development agenda for quite some time, macroeconomic policy design, without an unambiguous weight to employment objective, would be a highly risky method of policymaking. 12

Second, the need to incorporate employment in a macroeconomic policy framework would warrant the appropriate introduction of multiple targets, the relative weights of which would depend on country-specific circumstances. The contours of such a multi-dimensional perception can be seen in the cases of East Asian economies, as well as Chile, where macro-fundamentals were achieved in tandem with macroeconomic support toward investment and savings; further, such a policy framework also ensured an equity-sensitive growth pattern that helped support a faster pace of poverty reduction. Third, regarding the setting of target(s), there must be convincing basis of how the threshold of the target(s) is defined. As we have observed earlier, a major difficulty with IMF`s stringent position on stabilization was with respect to their adherence to inflation targeting, rather that on trying to achieve a stable and moderate level of inflation. 23 In the case of the latter, various instruments would be at play, depending on the accuracy of analysis of the root causes of inflation, and how it may affect other macroeconomic variables. Fourth, given that the dominant target of orthodox macro-policy was price stability, there was an over-riding role of monetary policy in their advocacy, through the policy rate. It is now well-known, especially in the aftermath of the recent crises that fiscal policy needs to play an equally significant role, not just as a counter-cyclical instrument in short-term business cycles, but also in longer term a locative functions toward growth and equity. Fiscal discipline must not be conflated with fiscal policy, and the proactive role that this instrument can play in both stabilization and growth processes. This theme is discussed in more analytical detail with respect to Arab countries in Roy et al. (2011). 24 Fifth, another imperative of macroeconomic policy, as crucially highlighted by the stylized facts in the previous Arnim et al. (2011), 25 is on the need to incorporate measures to reduce volatility of Arab growth, and to offset jobs and income insecurity that it entails. This would warrant adopting a range of social protection and safety net measures. Though initially, these may be rudimentary in scope and coverage, such social protection systems when internalized in a macroeconomic framework would, apart from reducing insecurity, tend to act as important stabilizers of the economy, especially through sustaining consumption and aggregate demand during recessionary periods. This, in fact, is at the heart of the automatic stabilizers in the macroeconomic policy design of the advanced welfare economies. The poorer developing countries, which during crises usually resort to (often with donor support) discretionary interventions and extensions of existing safety net programmes, need to make their first assessments at what it would take to install a basic social protection system, one that their macroeconomic policy design could incorporate as an approximate automatic stabilizer. In what follows we discuss this issue further. Informality and Social Protection The concept of Informality is used to describe the collection of firms, workers, and activities that operates outside the legal and regulatory frameworks. 26 However, in other definitions it also includes unreported activities and workforce of registered firms. Several forms of informality can be distinguished including labour informality, product informality and land informality. The World Bank states that by some estimates, more than 30% of the developing world s GDP and 70% of its workers are outside the official economy. 27 Informality is a fundamental characteristic of underdeveloped countries and is closely associated with poverty. Empirical evidence shows that the level of informallabour (measured by lack of pension coverage) tends to be higher in countries with lower GDP per capita. The terms social security and social protection are not always clear and deserve some more explanation: The notion of social security adopted here covers all measures providing 13

benefits, whether in cash or in kind, to secure protection, inter alia, from; (a) lack of workrelated income (or insufficient income) caused by sickness, disability, maternity, employment injury, unemployment, old age or death of a family member; (b) lack of access or unaffordable access to health care; (c) insufficient family support, particularly for children and adult dependants; and (d) general poverty and social exclusion. The term social protection is used across the world and institutions with an even greater variety of meanings than social security. It is often interpreted as having a broader character than social security (including, in particular, protection provided between members of the family or members of a local community), but is also used in some contexts with a narrower meaning than social security (understood as comprising only measures addressed to the poorest, most vulnerable or excluded members of the society). Thus, in many contexts the terminology of social security and social protection may be largely interchangeable, and the ILO (following the European tradition) certainly uses both in discourse with, and the provision of relevant advice to, its constituents. A third term is sometimes confounded with the term social security: social insurance describes one particular type of social security programme, which is partially or fully funded from contributions from workers and employers. 28 The conventional economic wisdom with regard to the relationship between social protection and economic performance is that: (1) Social expenditure is perceived to be exclusively consumptive expenditure and does not have an investment character. (2) Economic growth is believed to reduce poverty automatically (often described as the trickle-down effect ) and thus in an environment of growth the need for redistributive (social protection) policies may remain very limited. (3) There is a trade-off between social expenditure and economic efficiency or growth and hence high levels of redistribution are detrimental to growth. 29 In direct opposition to this conventional wisdom the ILO views social protection as a system that reduces poverty and is an investment in productivity. This is so because poverty is a risk to security and lack of security is a hindrance to the investment climate. Furthermore, only people who enjoy a minimum of material security can afford to take entrepreneurial risks. In addition, The empirical and statistical evidence of the last decade shows clearly that economic growth does not automatically reduce poverty without employment promotion and income redistributive mechanisms (such as social security systems) being put in place, otherwise countries with the same levels of GDP per capita would not experience a wide range of different levels of poverty and inequality. Finally, the famous trade-off between growth and equity does not hold true. There is ample evidence that countries with identical levels of social spending experience a wide range of different levels of GDP, contradicting the hypothesis that there is an automatic negative correlation between economic performance and levels of redistribution. The latter implies that there is no hard and fast rule as to what countries can afford. There is, however, a fairly strong positive correlation between per hour productivity and per capita expenditure on social protection in OECD countries. Thus, superior economic performance and high social expenditure do coexist and social expenditure and economic performance support each other. The famous growth equity trade-off is a myth rather than fact. 30 Vulnerability and the working poor Informal employment has probably grown in most countries of the Arab region during the last decades. Data are incomplete but, available data show that the share of formal economy employees at the totallabour force (which is here used as a proxy for formalization) hardly exceeds 60 per cent. Furthermore, the degree of informality in Arab countries is much larger than in the typical developed country but lower than in the majority of other developing regions. 14

Tunisia Jordan Algeria Syria Lebanon Sudan Yemen Egypt Morocco Syria Tunisia Algeria OpT. Egypt Yemen Lebanon Morocco 2007 Using the measure based on lack of pension coverage, the typical Arab country haslabour informality larger than that in Eastern Europe and Central Asia (ECA), comparable to Latin America and the Caribbean (LAC), and lower than East Asia and Pacific (EAP), South Asia (SAS) and Sub- Saharan Africa (SSA). 31 However, there is a wide variation between the Gulf States on the one hand, where formalization reaches more than 90 per cent on average and other Arab countries on the other hand, with formalization rates of around 50%. These data can also serve as proxies for the level and variation of formal social security coverage in the region. 32 As noted in the previous Arnim et al. (2011) 33, much of thelabour absorption took place low value-added in the service sector which leads us to suspect that the share of workers in vulnerable employment has increased. Figure 15 summarizes two keylabour market indicators, which are commonly used to measure vulnerability and their percentage change since the 1990s. The figure shows the share of self-employed among total employment in non-agriculture sectors has increased in the three largest countries (Egypt, Morocco and Algeria) since the 1990s. The exception was Tunisia, where the share is significantly lower in the 2000s and has declined since the 1990s. Due to data limitations, we were not able to establish the trends for Sudan and Yemen, which held the highest shares of self employed in non-agricultural sectors during the 1990s. Figure 15.B reveals the share of informal employment in total non-agricultural employment reached nearly half for the majority of Arab countries, but it is highest in Morocco and lowest in Syria and Tunisia. Figure 15: Self-employed in non-agricultural employment (1990s-2000s) (A) and share of informal employment in total non-agricultural employment (B), 2000-2007 (A) `(B) 60 50 40 30 20 10 0 1990s 2000s 80 70 60 50 40 30 20 10 0 Source: Authors estimates based on household data reported by OECD. In order for policy makers to be able to address the problem of economic vulnerability, they need to have an adequate understanding of the links between employment and poverty. This link is complicated and multi-faceted. At the micro level, both individual and household characteristics can shape the likelihood that a particular individual will suffer from poverty, with macro level economic conditions also playing an important part. One important question to consider when examining poverty patterns is the role that employment plays. Individuals may be at risk of poverty either because they are unable to participate inlabour market, because they are unable to find employment, or because the type of employment they find does not provide them with sufficient hours or a living wage. The poor s relationship to thelabour market can thus be lumped into four broad categories: (1) Individuals who are unable to participate in thelabour market, either due to disability, old age, or householdlabour demands often related to prevailing gender norms, particularly in the case of women (labour market access); (2) Individuals who are seeking employment, but unable to find it (unemployment); (3) Individuals who are employed, but working very low numbers of hours (underemployment); and (4) Individuals who are employed and working many hours (over-employment). A careful examination of these four categories can help 15

policy makers gain a better sense of the various ways that the poor may or may not be incorporated intolabour markets and because all four may have macroeconomic implications. Studying each of these categories is particularly important, since the policy solutions needed to address these various problems will differ. While general macroeconomic stimulus, combined with unemployment insurance, will provide protection to those who find themselves in economic difficulties due to temporary unemployment, policies that improve the types of jobs, or the conditions under which work is performed may be needed to help those who either work very many or just a few hours, while another form of entirely different safety net is needed to assist those who are incapable of participating in the labour market. For this section, Household Income Expenditure surveys (HIES) for Egypt (1998, 2005 and 2009), Lebanon (2004), Syria (1997, 2004, 2007) and Yemen (1998, 2006) were analyzed, with a focus on how poverty status was related to various employment factors. In particular, regression analysis was carried out, to determine whether the poor were more or less likely to be under, over or unemployed than the non-poor. 34 Most of the HIES surveys included questions about hours worked, although the ways the question was asked varied a bit across the surveys. For example in the case of Egypt, data on days of work were collected, while in the other three countries questions were asked about hours. Rather than choosing a definition of under 35 and over-employment ex-ante, we examined the data to see where breaks occurred, and defined under and over-employment accordingly. In the case of Egypt, a clear break occurred between five and again at six days, so underemployment was defined as working fewer than five days, while over-employment was defined as working the entire week (seven days). For Lebanon, breaks were observed at 35 and 48 hours. In this case underemployment was defined as working fewer than 35 hours, while over-employment was defined as working more than 48. In the case of Yemen the upper break was at 49 hours, where the lower break was again at 35. Before turning to the regression results, it is important to point out that the four countries are quite different in terms of the level and type of development. Lebanon for example is the most developed in terms of both per capita income and the structure of the economy. The majority of the population lives in urban areas, and the economy is highly concentrated in the service sector, with a relatively small agricultural sector. Only 10% live in extended family households, with most individuals living in nuclear family setting, and a growing number also living alone or with non-family members. By contrast Yemen is the poorest and remains the most rural and agriculturally based of the four economies. Yemen also has the highest rate of extended family households, with a third of the total households being made up of three or more generations. Rates of female headship also vary considerably across countries, with Egypt having the highest proportion of female headed households, at 16.6% in 2009, followed by Lebanon, with about 14%. In Syria the rates of female headship remain lower (9%). In Egypt, Syria and Yemen the majority of female headed households are run by widows, while in Lebanon there are more divorced women heading households. Also worth noting is that while they remain a small proportion of all households, the% of one person households is growing and is already quite high in Lebanon (5% compared to 1% elsewhere). While Lebanon and Yemen represent the extremes, Syria and Egypt have fairly similar per capita income rates and are in some ways quite similar, while in other ways they are quite different. For example, despite similar levels of per capita income, far more Syrians live in extended family households, compared to Egyptians. This may explain to some extent the lower incidence of poverty in the former. An analysis of the HIES data suggest that indeed all four of the factors identified above are at play when it comes to poverty in the Arab world. As illustrated in figure 16, the poor 16