WPS7488 Policy Research Working Paper 7488 Unhappy Development Dissatisfaction with Life on the Eve of the Arab Spring Efstratia Arampatzi Martijn Burger Elena Ianchovichina Tina Röhricht Ruut Veenhoven Middle East and North Africa Region Office of the Chief Economist November 2015 Policy Research Working Paper 7488 Abstract Despite progress in economic and social development in range of objective and subjective factors with life evalu- the 2000s, there was an increasing dissatisfaction with life ation in the Middle East and North Africa region in the among the population of many developing Arab countries. years immediately preceding the Arab Spring uprisings At the end of the decade, these countries ranked among (2009–10). The findings suggest a significant, negative the least happy economies in the world—a situation that association between life satisfaction levels in the region fits the so-called “unhappy development” paradox. The during this period and each of the main perceived reasons paradox is defined as declining levels of happiness at a for the 2011 uprisings—dissatisfaction with the standard time of moderate-to-rapid economic development. This of living, poor labor market conditions, and corruption. paper empirically tests the strength of association of a This paper is a product of the Office of the Chief Economist, Middle East and North Africa Region. It is part of a larger effort by the World Bank to provide open access to its research and make a contribution to development policy discussions around the world. Policy Research Working Papers are also posted on the Web at http://econ.worldbank.org. The authors may be contacted at eianchovichina@worldbank.org. The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. Produced by the Research Support Team Unhappy Development: Dissatisfaction with Life on the Eve of the Arab Spring Efstratia Arampatzi, Martijn Burger, Elena Ianchovichina, Tina Röhricht, and Ruut Veenhoven JEL Classification: I31, Z13 Keywords: developing Arab countries, Middle East and North Africa; Algeria, Egypt, Iraq, Jordan, Lebanon, Morocco, Syria, Tunisia, Palestine territories, Yemen; grievances; life satisfaction; Arab Spring; uprisings; standards of living; labor market; governance.                                                                We would like to thank the participants in 13th ISQOLS (International Society for Quality-of-Life Studies) Conference in Phoenix, Arizona, the World Bank’s Arab Inequality Puzzle Workshop in Washington DC, and the European launch of the World Happiness Report at Erasmus Univeristy, Rotterdam, the Netherlands. The research was supported by funding through the World Bank’s Strategic Research Program.  Efstratia Arampatzi is junior researcher at the Erasmus Happiness Economics Research Organization, Erasmus University Rotterdam, P.O. Box 1738, 3000 DR Rotterdam, the Netherlands. E-mail: arampatzi@ese.eur.nl.  Martijn Burger is assistant professor at the Department of Applied Economics, Erasmus University, Rotterdam, Tinbergen Institute and academic director at the Erasmus Happiness Economics Research Organization, P.O. Box 1738, 3000 DR Rotterdam, the Netherlands, Tel: +31 (0) 10 4089579, Fax: +31 (0) 10 4089141. E-mail: mburger@ese.eur.nl. URL: http://www.mjburger.net.  Elena Ianchovichina is lead economist at the Chief Economist Office, Middle East and North Africa Region, the World Bank, 1818 H Street NW, Washington, DC 20433, USA, Tel: +1 202 458 8910, E-mail: eianchovichina@worldbank.org.  Tina Röhricht is research assistant at the Erasmus Happiness Economics Research Organization, Erasmus University Rotterdam P.O. Box 1738, 3000 DR Rotterdam, the Netherlands. E-mail: röhricht@ese.eur.nl.  Ruut Veenhoven is emeritus professor at the Erasmus Happiness Economics Research Organization, P.O. Box 1738, 3000 DR Rotterdam, the Netherlands and extraordinary professor at the North-West University, South Africa. E-mail: veenhoven@ese.eur.nl. 1. The ‘Unhappy Development’ Paradox in Developing Arab Countries In the 2000s, many developing countries in the Middle East and North Africa (MENA) did well according to the regularly tracked poverty statistics and human development indicators. Absolute poverty, measured at $1.25 a day, declined in all economies, except the Republic of Yemen, and was low on average. The incomes of the bottom 40 percent, measured as 2005 PPP-adjusted per capita expenditure, grew at higher rates than average expenditures in many developing Arab countries for which information was available (Ianchovichina, Mottaghi, and Devarajan 2015). The Gini inequality indexes were low by international standards and did not worsen in most MENA economies (Ianchovichina, Mottaghi, and Devarajan 2015). Importantly, the region made notable strides in reaching not only the Millennium Development Goals related to poverty and access to infrastructure services (especially drinking water and sanitation and Internet connectivity), but also in terms of reducing hunger and child and maternal mortality, and increasing school enrollment (Iqbal and Kiendrebeogo 2015). Prior to the Arab Spring uprising, most developing MENA countries were seen as relatively stable places. Only two MENA countries—Iraq (7th) and the Republic of Yemen (15th) —made it to the top 25 of the 2010 Failed States Index1 of Foreign Policy. Libya and Tunisia were ranked 111th and 118th of 177 countries, respectively, and so they appeared among the stronger and less fragile countries in the world (Goodwin 2011). With autocratic rulers in power for many years, the cracks in these countries’ models of government remained invisible to most observers, including political scientists (Gause 2011), and some even considered Islam a stabilizing force (Bromley 2014). Thus, the Arab Spring transitions of 2011 took most economists, political scientists, and policy makers by surprise (Gause 2011; Goodwin 2011; Bellin 2012; Bromley 2014). Yet, the emergence of social discontent in the Arab countries could be detected using subjective data. Life satisfaction in many MENA countries was below the average for the group of countries at a similar level of development (figure 1a) and had dropped significantly in the years prior to the Arab Spring events (figure 1b). By the end of the 2000s, people in the developing parts of MENA, especially in the Arab Republic of Egypt, Iraq, the Syrian Arab Republic, Tunisia, and the Republic of Yemen, were among the least happy people in the world (see figure 2 and appendix A).2 In Egypt, for instance, average life-evaluation levels plunged on a 0-10 scale3 from 5.5 in 2007 to 4.4                                                              1 The Failed States Index measures stability based on economic, political, and military indicators. 2 The incidence of depression was also observed to be high in MENA, according to Ferrari et al. (2013). 3 The two extreme ends of the range capture worst possible life (0) and best possible life (10). 2    in 2010—a deep drop in the context of improvements observed in socioeconomic statistics and growth in per capita incomes (see figure 1b). Figure 1a: GDP per Capita and Satisfaction with Life, 2008-10 8 7.5 7 ARE Average Life Evaluation 6.5 QAT SAU 6 BHR 5.5 JOR DZA TUN IRN LBY 5 DJI EGYIRQ LBN SYR 4.5 YEMWBG MAR 4 3.5 3 5 6 7 8 9 10 11 12 Log GDP per Capita (2005 US dollars) Figure 1b: Percentage Growth in GDP and Change in Satisfaction with Life (Weighted Averages) in 106 Countries, 2005-10 1.5 Change in Life Evaluation 2005‐2010 1 0.5 ARE 0 WBG MAR ‐3 ‐1 1 3 YEM 5 7 9 11 13 15 ‐0.5 LBN EGY SAU JOR ‐1 ‐1.5 Average Annual GDP Growth 2005‐2010 Sources: Real GDP per Capita: World Bank Development Indicators; Life Satisfaction: World Database of Happiness. Note: Numbers are weighted averages for 147 countries. Abbreviations: ARE=United Arab Emirates; EGY=Egypt; JOR=Jordan; LBN=Lebanon; MAR=Morocco; SAU=Saudi Arabia; WBG=West Bank and Gaza; YEM=the Republic of Yemen. 3    Figure 2: Average Life Satisfaction in the World, 2006-12 Source: Gallup World Poll data, based on the Question WP16: Please imagine a ladder, with steps numbered from 0 at the bottom to 10 at the top. The top of the ladder represents the best possible life for you and the bottom of the ladder represents the worst possible life for you. On which step of the ladder would you say you personally feel you stand at this time? The phenomenon of rapid economic growth occurring at a time of declining levels of subjective well-being is known as the ‘unhappy growth’ paradox (Graham and Lora 2009). Controlling for per capita incomes, several recent cross-country studies by Deaton (2008), Graham and Lora (2009), and Stevenson and Wolfers (2008) find that people living in countries with higher economic growth levels are on average less happy than those living in countries with less growth, highlighting the importance of taking into account people’s perceptions when attempting to understand a nation’s well-being. In this paper, we focus on the so-called ‘unhappy development’ paradox, defined here as declining levels of happiness at a time of moderate to rapid economic growth and social development. There could be many reasons for this paradox in developing Arab countries. There might have been a rise in people’s expectations and aspirations, particularly those of youth who had acquired better education than their parents and expected to find good jobs after graduation (Campante and Chor 2012). A widening gap between actual and expected welfare may have increased people’s aversion to inequality and social injustice (Verme et al. 2014; Cammett and Diwan 2013) and negatively affected their levels of happiness. This hypothesis is consistent with the findings in Bruni (2004), who argues that more economic wealth does not necessarily transform into higher levels of well-being, since it may negatively affect noneconomic wealth and perceptions. Even in the absence 4    of a shift in expectations, people may have become more frustrated with difficult-to-measure factors related to quality, such as the deterioration in the quality of public services, the ability to get good quality jobs, and institutional and environmental quality. Worsening of other subjective indicators, such as the ability to voice concerns and demand accountability and the incidence of corruption and cronyism, may have also contributed to deterioration in well-being. Motivated by the need to understand the ‘unhappy development’ paradox in developing MENA, this paper empirically tests which factors are associated with life dissatisfaction in MENA countries in the years immediately preceding the Arab Spring uprisings (2009-10), taking into account objective and perceptions data regarding different aspects of life and society. In addition, we compare the extent to which the factors associated with life dissatisfaction are also associated with the Arab Spring social upheaval in developing MENA. The paper adds to the literature in three ways. To our knowledge, we are the first to examine empirically the relative importance of different explanations provided for the declining life satisfaction in developing MENA in the wake of the Arab Spring. In particular, we examine several explanations or hypotheses for the fall in life satisfaction in developing MENA countries, including dissatisfaction with: (1) the political system of autocracy and limited civil freedoms, (2) the standard of living, (3) the high unemployment and poor quality jobs, and (4) corruption and crony capitalism. Second, we investigate systematically the factors behind the decline in life satisfaction by decomposing the decline into two components: a first-order effect associated with changes in the prevalence of dissatisfied individuals and a second-order effect associated with changes in the relative importance of these factors or perception domains for life satisfaction. In other words, this decomposition allows us to determine whether life satisfaction declined because a greater percentage of people became more dissatisfied with certain domain satisfactions or whether the relative importance of the domain satisfaction for subjective well-being increased. Third, we compare the factors related to unhappiness in developing MENA with the perceived reasons for the Arab Spring uprisings. We find that the main perceived reasons for the uprisings are the factors associated significantly and negatively with subjective well-being levels in developing MENA during this period. Our findings suggest that perceptions provide valuable information about public preferences and needs, which are typically not reflected in objective data (Veenhoven 2002). In other words, we make the case that both objective and subjective (or perceptions) data matter for understanding the root causes of political violence (cf. Okulicz-Kozaryn 2011). The remainder of this paper is organized as follows. Section 2 presents the potential root causes of dissatisfaction with life in developing MENA. Section 3 discusses the concepts, methodology, and 5    data used in the empirical exploration. The results of this empirical analysis are presented in section 4. Finally, section 5 concludes with a summary of findings, a discussion of how these results link to the reasons for the Arab Spring uprisings, and a few caveats. 2. Root Causes of Dissatisfaction with Life in Developing MENA Countries A look at the universal conditions for happiness, as presented in cross-country studies focusing on life satisfaction, provides limited understanding of the root causes of dissatisfaction with life in the Arab world. To understand the factors shaping the subjective well-being in the developing Arab countries prior to the Arab Spring, we must factor in explicitly the social context in these countries during this time period. There is no consensus on the root causes for life dissatisfaction in the Arab world on the eve of the Arab Spring. Several explanations have been put forward: (1) limited freedom and voice in predominantly autocratic states; (2) dissatisfaction with standards of living; (3) unhappiness with persistent unemployment and lack of good jobs due to the growing informality of the private sector; and (4) dissatisfaction with corruption and cronyism, which limits opportunities for those who work hard. Each of these explanations is discussed in greater detail below. Autocracy On the eve of the Arab Spring, most Arab states were longstanding autocracies (Chekir and Diwan 2012; Bromley 2014; and Cammett and Diwan 2013). Power was concentrated in the hands of one person or a small group of elites, backed by the military, who made decisions subject to few legal restraints and mechanisms of popular control. At the same time, the public had few if any channels of safe expression of opinions and grievances and opportunities to develop strong civil society. The longstanding regimes managed to stay in power through a combination of repressive practices and a social contract, which extended benefits such as free public education and health, energy, and food subsidies, and guarantees of public employment in exchange for political support (Bellin 2004; Bromley 2014; Cammett and Diwan 2013). Cammett and Diwan (2013) refer to this social contract as an ‘autocratic bargain,’ in which the middle class was lured with ‘material benefits’ in exchange for ‘political quiescence.’ Thus, despite human development and economic progress after independence, the developing MENA countries scored low in terms of economic and social freedoms and the Freedom House ranked the region as the most repressive in the world (Freedom House 2008). 6    The extent to which people are free to make choices and voice opinions has a major impact on their happiness (Inglehart et al. 2008; Verme 2009). Democracies are, on average, happier than autocracies (Frey and Stutzer 2000), but the effect of democracy on happiness is stronger in countries with established democratic traditions (Dorn et al. 2007). Fereidouni, Najdi, and Amiri (2013) obtained no significant relationship between voice and accountability and happiness in developing MENA countries. Ott (2010) also found that the correlation between happiness and democracy is relatively weak in the MENA region. The ‘autocratic bargain’ may have weakened the direct link between happiness and limited freedom in developing MENA. Individuals who obtain ‘material benefits’ in exchange for political support may express dissatisfaction with living conditions rather than with the system responsible for the deterioration in the authoritarian bargain. They may initially voice mainly their dissatisfaction with living conditions and the factors affecting their quality of life, for instance, poor access to quality services and job market conditions. Dissatisfaction with Standards of Living By the early 2000s, major cracks appeared in the social contract of redistribution without voice in developing MENA. After independence, natural resource rents enabled many Arab countries’ governments to finance redistributive policies without imposing a heavy tax burden on citizens. But in the 1990s and 2000s, fiscal pressures increased, reflecting disappointing growth in the 1980s and growing recurrent expenditures, especially on public wages and subsidies. Governments responded by downsizing the public sector, removing the guarantees of secure public jobs, and initiating reforms of the food and energy subsidy programs.4 During this period, unemployment increased and many households noted deterioration in their standard of living. High dependence on imported food and limited fiscal space meant that the global commodity price increases of the 2000s would transmit to domestic markets despite the presence of food subsidies (Korotayev and Zikina 2011; Ianchovichina, Loening, and Wood 2014).5 For the poor, the increase in food and energy prices meant deterioration in their ability to meet basic needs.6                                                              4 Some governments were more successful than others in cutting subsidies and improving targeting. Most economies made only partial reforms to their subsidy systems and reversed the reforms in response to the Arab Spring events. 5   However, prices for these basic needs are typically not well covered by standard inflation and poverty measures, which would explain why the Arab Spring came as a surprise for many scholars and policy makers. 6 According to Maslow (1943), in the hierarchy of individual demands, a person’s physiological needs for basics such as food, water, and shelter dominate all other needs. In other words, if these basic needs are not supplied, all other human needs are pushed into the background and the individual only seeks to satisfy his or her hunger. Individual anxiety over rising costs of food or shelter can therefore trigger unhappiness and, in some cases, riots (Lagi, Bertrand, and Bar-Yam 2011). The risk of riots is particularly high in lower-income countries where the share of food and other necessities in household expenditure is high (Arezki and Brückner 2011). 7    The global economic crisis of 2008 put additional stress on the MENA economies. In Egypt, the crisis was associated with a steep decline in real earnings growth; in Tunisia, it reinforced the upward trend in unemployment; and in Jordan, it slowed employment growth. Dissatisfaction with basic public services such as health care, housing, schools, and infrastructure also grew in the developing MENA countries, according to Gallup World Poll data, reflecting the erosion in the quality of public services. By the end of the 2000s, this erosion in standards of living was felt not only by the poor, but also by other segments of the population, including the middle class. A gradual shift in government support to the elites became a particular concern (Cammett and Diwan 2013). People were frustrated because they could not get ahead by working hard and share in the prosperity generated by the relatively few large and successful Arab firms that were mostly state-owned or privately owned companies (OECD 2009).7 Reflecting diminishing marginal utility, the widespread system of subsidies could not compensate for the erosion of living standards; food and energy subsidies mattered less for the well-being of the middle class than they did for the well-being of the poor and vulnerable (Ianchovichina, Mottaghi, and Devarajan 2015). Unemployment and Low Quality Jobs Dissatisfaction with job market conditions was particularly strong in developing MENA in the wake of the Arab Spring. In the preceding decade, the MENA region’s average, aggregate and youth unemployment rates were the highest in the world. Without guarantees of secure public jobs, young people, who entered the labor market better prepared than their parents in terms of educational qualifications (Barro and Lee 2010; Campante and Chor 2012), were forced to queue for public sector jobs or take part-time or low-quality jobs in the informal sector (Chamlou 2013).8 Employment in the informal sector offered little protection at old age and limited access to quality health care and benefits, such as paid maternity and annual leave (Angel-Urdinola and Kuddo 2011; World Bank 2014b). The mismatch between educational attainment and economic opportunities created a gap between reality and expectations, lowering youth’s life satisfaction, amplifying perceptions of inequality and unfairness, and potentially contributing to social unrest (Campante and Chor 2012). In the literature, the negative association between happiness and unemployment is well-established and can be explained by a combination of income loss and psychic costs related to psychological                                                              7 According to OECD (2009), very few large Arab firms are publicly traded companies. 8 The informal sector consists of firms, workers, and activities that operate outside the legal and regulatory frameworks. 8    distress and loss of identity and self-respect (Veenhoven 1989; Gallie and Russel 1998). The deterring effect of unemployment on happiness is more severe for the long-term unemployed (Clark and Oswald 1994), which is particularly high in the MENA region, and for people with limited job opportunities (Clark, Knabe, and Rätzel 2010). Crony Capitalism and ‘Wasta’ At a time when public sector employment was contracting, private sector growth was sluggish and few people could find jobs in the formal private sector (Malik and Awadallah 2013). Private sector growth was stifled by ‘cronyism’ and fears that a rise of the ‘nouveau rich’ class would challenge existing power relations.9 Reforms in the 1990s were implemented in an uneven way, benefiting mainly the elites (Chekir and Diwan 2012; Rijkers et al. 2014) who dominated a range of economic sectors (Malik and Awadallah 2013). Perceptions about corruption and crony capitalism also worsened in the wake of the Arab Spring (Cammett and Diwan 2013), as reflected in the retreat of MENA countries’ rankings on the Corruption Perceptions Index of Transparency International between 2000 and 2010. In addition, most MENA countries scored below average on various governance indicator rankings in the 2000s (for example, Kaufmann, Kraay, and Mastruzzi 2011). Corruption and cronyism flourished in developing MENA with detrimental effects not only on aggregate economic and private sector growth, but also on people’s subjective well-being (Ott 2010). There was growing frustration with inequality of opportunity in labor markets and the increased importance of ‘wasta’ or connections with the elites in getting good quality jobs. These feelings were broadly shared and reflected perceptions of citizens that ‘wasta’ matters more than credentials for getting good jobs. In summary, it can be argued that the growing dissatisfaction in the wake of the Arab Spring was fueled by a mix of grievances related to the standards of living, unemployment and low quality jobs, and ‘wasta’ or cronyism. The rest of the paper will test these hypotheses. 3. Concepts, Methodology, and Data The word ‘happiness’ is used in various ways (Veenhoven 2012). In the broadest sense, it is an umbrella term for all that is good. However, in the social sciences, the word ‘happiness’ is also used                                                              9  The ruling elites controlled large parts of the private sector and profited from monopoly rights and cheap access to land and other resources (Cammett and Diwan 2013).  9    in a more specific way, which refers to an individual’s subjective appreciation of his or her own life. Accordingly, the concept of ‘happiness’ has been defined as ‘the degree to which an individual judges the overall quality of his/her own life-as-a-whole favorably’ (Veenhoven 1984, chapter 2). This is also commonly referred to by terms such as ‘subjective well-being’ and ‘life satisfaction.’ Thus defined, happiness is something on one’s mind that can be measured using surveys. Common survey questions10 read: ‘Taking all together, how happy would you say you are: very happy, quite happy, not very happy, not at all happy?’ (a standard item in the World Value Studies) or ‘Please imagine a ladder, with steps numbered from 0 at the bottom to 10 at the top. The top of the ladder represents the best possible life for you and the bottom of the ladder represents the worst possible life for you. On which step of the ladder would you say you personally feel you stand at this time?’ (a standard item in the Gallup World Poll). Responses to this question from the Gallup World Poll are used in the empirical part of this paper. This question captures predominantly the cognitive component of happiness, also known as contentment. How happy people are depends on objective conditions and subjective factors, including perceptions and expectations. According to Layard (2006), objective factors such as gender, age, marital and education status, financial situation, and health determine to a large extent life satisfaction, but subjective factors associated with perceptions and expectations about family relationships, work, community and friends, personal freedom, institutional quality, and personal values are also imperative to individual happiness. These domains of life reflect the most important human needs as identified by Maslow (1943). The relative importance of the objective and subjective determinants of life satisfaction vary over time and across individuals. To analyze the roots of dissatisfaction with life in developing MENA in the wake of the Arab Spring, we used cross-sectional data from the Gallup World Poll for the years 2009-10 and a simple reduced-form life satisfaction model (see Di Tella, MacCulloch, and Oswald 2003; Arampatzi, Burger, and Veenhoven 2015): LS jit = Θ Individual_Perceptions jit + Σ Personal_Characteristics jit + ε j+ λ t + μ jit. (1)  In this model, LS, the overall life satisfaction of individual j in country i in year t, depends on a vector of Individual_Perceptions about social conditions and domain satisfactions of individual j in country i in year t, a vector of objective Personal_Characteristics of individual j in country i in year t, a vector εi of country dummies to control for time-invariant country-specific characteristics, a vector λt of month-year dummies capturing time-related shocks that are common for all countries in the developing MENA region, and μjit is a residual error. We estimate model 1                                                              10 See Veenhoven (2012) for a discussion of the limitations of direct questioning. 10    using weighted least squares regression (WLS) with robust standard errors and weighting observations using the sampling weights provided by the Gallup World Poll.11 The annual Gallup World Poll includes at least 1,000 randomly selected respondents (adult population of 15 years and older) per country and is representative at the national level. In the Gallup World Poll, individuals report on several aspects of their life, including how satisfied they are with their life as a whole and how satisfied they are with different domains of their life. The common sample we use in this paper comprises in total 25,244 respondents from 10 developing MENA countries, including Algeria, Egypt, Iraq, Jordan, Lebanon, Morocco, Palestine, Syria, Tunisia, and the Republic of Yemen. Figure 3: Distribution of Life Evaluation Scores in Developing MENA (percent by decile) 30% 25% 20% 15% 10% 5% 0% 0 1 2 3 4 5 6 7 8 9 10 Source: Gallup World Poll 2013. Life satisfaction was measured using a single question, known as the ‘Cantril Ladder’ or ‘Self- Anchoring Striving Scale’ (Cantril 1965). This question asks on which step of the ladder, with steps from 0 to 10, a person feels he or she stands at present. The higher the score on the ladder, the closer one’s life is seen to his or her ideal life. Figure 3 shows the distribution of happiness scores in the developing MENA region in the 2009-10 period. The unhappiness in the region is evidenced by the fact that 61 percent of the developing MENA population scores 5 or lower on the Cantril Ladder, while only 10 percent gives his or her life a score of 8 or higher. Within developing MENA, the degree of life satisfaction ranges by country from 4.66 in the Republic of Yemen to 6.23 in Jordan (table 1). It is worth noting that a person with high expectations is more likely to be                                                              11  Following Ferrer-i-Carbonell and Frijters (2004), we treat the dependent variable as cardinal and not as ordinal. 11    dissatisfied with his life than a person with low expectations. Thus, the life satisfaction variable captures indirectly the effect of a gap between expected and real welfare. Table 1: Life Satisfaction in Developing MENA Countries in the Common Sample, (2009- 10) Variable Observations Mean SD Min. Max. Algeria 3,588 5.58 1.65 0 10 Egypt, Arab. Rep. 1,628 4.88 2.14 0 10 Jordan 691 6.23 1.81 0 10 Iraq 2,432 5.07 1.72 0 10 Lebanon 3,382 5.29 2.29 0 10 Morocco 3,144 4.97 1.67 0 10 Palestine 2,942 4.83 2.14 0 10 Syrian Arab Republic 2,169 4.86 2.12 0 10 Tunisia 2,048 5.17 1.69 0 10 Yemen, Rep. 3,184 4.66 2.21 0 10 Source: Gallup World Poll 2013. Our main variables of interest relate to the domain-specific characteristics thought to have a most profound influence on life satisfaction in the wake of the Arab Spring as discussed in section 2. The Gallup World Poll does not have a question on the degree to which people are satisfied with the political system in the MENA countries. Since in autocracies people’s ability to make choices is restricted, we instead turn to the question: “Are you satisfied or dissatisfied with your freedom to choose what you do with your life?” We recognize, however, that this question also reflects how satisfied people are with their freedom to make individual choices about education, marriage, children, and employment. The answer to this question is zero for those who are satisfied and one for those who are dissatisfied with their freedom to make choices.12 We control for objective measures of standards of living by including individual income (given in international dollars). We also include subjective evaluations of living standards based on the answers to the following question: “Are you satisfied or dissatisfied with your standard of living, all the things you can buy and do?” The answers to this question reflect how people value monetary and                                                              12 People answering “don’t know” or who refused to answer this and other questions were omitted from the sample. 12    nonmonetary factors. The latter pertain to the quality of living conditions, including those related to the environment, local institutions, political and economic stability, infrastructure, health and education services, and community safety and cohesion. Other nonmonetary factors are related to the quality of jobs, the variety of choices available to people living in a given area, and the cultural context. Finally, the answers to this question factor in people’s expectations about the future, which may change over time, and people’s own views on what their standard of living should be given the amount of effort they spend at work. The possible answers to this question are zero if satisfied and one if dissatisfied. To examine the effects of unemployment, underemployment, and job market conditions, we include subjective and objective variables related to employment and the education system. With regard to employment status, we distinguish between individuals who are paid employees (reference category), self-employed, underemployed, unemployed, or out of the workforce. The underemployed are respondents who are employed part-time, but who would like to work full- time, while the unemployed respondents are not employed at all and are looking for job opportunities. Respondents who were out of the workforce included homemakers, students, and retirees. In addition, we control for whether people are employed in government positions or not (reference category is “Other”). To reflect on job market conditions and the availability of high-quality jobs, respondents were asked: “Are you satisfied or dissatisfied with efforts to increase the number of quality jobs?” to which they could either reply with a zero if satisfied or one if dissatisfied. The question: “In the city or area where you live are you satisfied or dissatisfied with the education system or the schools?” allows us to capture the effect on life satisfaction of service provision, in particular education services, which determine employment opportunities later in life. The answer to this question can be zero if satisfied or one if dissatisfied. To explore the effect of corruption, cronyism, and ‘wasta’ on life satisfaction, we focus on government corruption as a proxy for perceptions of corruption. The answer to the question: “Is corruption widespread within government?” could be zero, if the level of corruption within government is limited, or one, if government corruption is widespread. When information regarding corruption in government was not available, the question “Is corruption widespread within business?” was used (cf. Helliwell, Layard, and Sachs 2015). In addition, we reflect the extent to which cronyism and inequities affect people’s life satisfaction by incorporating people’s opinions on whether working hard pays off. The answers to the question: “Can people in this country get ahead by working hard or not?” are zero if satisfied and one if dissatisfied. 13    Finally, we control for personal characteristics (demographic characteristics) that may confound the relationship between the designated factors and life satisfaction in developing MENA. These personal characteristics are related to gender, age, marital status and household composition, education level, migration status, and religion. An overview of all variables included in the analysis (including descriptive statistics) and a correlation matrix are provided in appendixes B1, B2, and B3. 4. Empirical Results This section discusses the results from the baseline and alternative specifications, the results’ sensitivity to changes in variable specifications and data aggregations, as well as endogeneity bias issues. Baseline and Alternative Specifications: Ordinary Least Squares Results Table 2 reports results from different specifications using the Cantril Ladder as dependent variable. In the first specification, we have only control variables for personal characteristics. In specifications 2 to 6, we separately include each of the subjective domain satisfaction variables associated with dissatisfaction in developing MENA, along with related objective factors. In specification 7, all subjective and objective variables are included simultaneously. The final specification in table 2 (model 8) is a replication of model 7 using a reduced sample of countries that experienced uprisings related to the Arab Spring. All the specifications include country and time dummies. The country dummies capture time-invariant, country-specific factors, such as the size of the country, culture, language, distance to markets, and structural features of the political and economic environment. The time dummies control for exogenous factors that changed over the period of interest, controlling herewith for contagion effects in the aftermath of the global financial crisis. In line with the empirical literature on happiness, education and marriage are positively associated with life satisfaction in developing MENA. Against the prevailing perception in the West, Arab women are on average happier than men. Focusing on the main sources of discontent in the wake of the Arab Spring (models 2 to 7), the main findings can be summarized as follows. First, although dissatisfaction with freedom to choose what you do with your life has a negative and significant effect on life satisfaction (table 2, model 2), this effect disappears after controlling for other perceptions (table 2, model 7). This finding supports the view that the social contract has weakened the direct link between authoritarianism (for example, lack of freedom) and life satisfaction. So it is the effect of the authoritarian political system on economic well-being and other domains of life, 14    rather than freedom per se, that initiated unrest in developing MENA countries. It is therefore not surprising that dissatisfaction with standards of living has the largest and strongly significant negative effect on life satisfaction (table 2, models 3 and 7). On average, the life satisfaction score of dissatisfied respondents is 1.24 points lower than the life satisfaction score of respondents who are satisfied with their living standards in the fully specified model in table 2, model 7. Second, poor job market conditions are significantly and negatively related to dissatisfaction in developing MENA countries—a result that retains significance even when we include all other subjective variables (table 2, models 4 and 7). The unemployed and underemployed report life satisfaction scores that are, respectively, 0.34 and 0.11 points lower than people in paid employment. Lack of quality jobs is another reason for the discontent and remains a significant factor even after we control for employment status. On average, respondents who indicate dissatisfaction with the availability of high quality jobs report 0.15 point lower life satisfaction than those who are satisfied with job quality (table 2, model 7). Not surprisingly, people working for the government (considered to be the best kind of jobs in MENA) are, on average, significantly happier than people working in the private sector. Third, we find that dissatisfaction with the education system is associated with life dissatisfaction in developing MENA. Respondents who are dissatisfied with the educational system report 0.17 point lower satisfaction with life than those who are satisfied with the education system (table 2, models 4 and 7). Fourth, perceptions of inequality of opportunities ( or ‘wasta’), corruption, and crony capitalism are significantly and negatively associated with life satisfaction in developing MENA (table 2, model 5 to 7). Respondents who think that people cannot get ahead by working hard report, on average, a 0.22 point lower life satisfaction score than those who are satisfied with this dimension of life satisfaction. Respondents who believe that corruption is widespread in the government are on average 0.27 point less satisfied with life, although this effect is reduced controlling for other perceptions (table 2 , models 5 and 7). Thus, in MENA, the governance problem is perceived to affect life satisfaction not so much through corruption in government, but through practices that affect all aspects of life and prevent people (and those working in the private sector, more generally) from succeeding even when they make great efforts to excel and do a good job. This result is consistent with the findings in Rijkers, Freund, and Nucifora (2014) and World Bank (2014a). 15    Table 2: Determinants of Life Satisfaction in MENA: Ordinary Least Squares Estimates (1) (2) (3) (4) (5) (6) (7) (8) VARIABLES Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7 Model 8 DEV DEV DEV DEV DEV DEV DEV Arab MENA MENA MENA MENA MENA MENA MENA Spring Dissatisfied with freedom to choose life: Yes  -0.351*** -0.033 -0.019 (0.030) (0.031) (0.053) Dissatisfied with standard of living: Yes -1.333*** -1.238*** -1.213*** (0.029) (0.030) (0.053) Income (1,000’s) 0.023*** 0.029*** 0.023*** 0.025*** (0.002) (0.002) (0.002) (0.003) Dissatisfied with efforts to increase high quality -0.361*** -0.154*** -0.139*** jobs: Yes (0.031) (0.032) (0.053) Dissatisfied with the educational system or the -0.340*** -0.166*** -0.158*** schools: Yes (0.030) (0.029) (0.051) (Reference group: Full-time Employed) Self-employed 0.077 0.041 -0.024 (0.064) (0.061) (0.100) Unemployed -0.534*** -0.335*** -0.475*** (0.082) (0.079) (0.145) Out of workforce 0.003 -0.019 -0.028 (0.049) (0.047) (0.076) Underemployed -0.267*** -0.114 -0.242* (0.082) (0.080) (0.133) (Reference group: Other) Working for the government 0.245*** 0.190*** 0.309*** (0.055) (0.052) (0.084) Undetermined -0.011 -0.019 -0.280*** (0.051) (0.049) (0.095) Corruption widespread within government: Yes -0.277*** -0.077** -0.056 (0.036) (0.035) (0.054) People cannot get ahead by working hard: Yes -0.496*** -0.223*** -0.210*** (0.041) (0.039) (0.080) (Reference group: Muslim) Not Muslim/Other religion 0.269*** 0.237*** 0.202*** 0.168** 0.275*** 0.239*** 0.171** 0.176 (0.075) (0.075) (0.069) (0.074) (0.075) (0.074) (0.068) (0.152) 16    (1) (2) (3) (4) (5) (6) (7) (8) (Reference group: Completed elementary education or less) Completed 9-15 years of education 0.452*** 0.438*** 0.295*** 0.356*** 0.448*** 0.447*** 0.282*** 0.393*** (0.033) (0.033) (0.031) (0.033) (0.033) (0.033) (0.031) (0.051) Completed four years of education beyond high 0.917*** 0.894*** 0.544*** 0.672*** 0.918*** 0.902*** 0.538*** 0.533*** school and/or 4-year college degree. (0.053) (0.053) (0.050) (0.054) (0.053) (0.052) (0.051) (0.093) (Reference group: Not a migrant) Migrant -0.145 -0.147 -0.264*** -0.208** -0.142 -0.156 -0.257*** -0.729*** (0.102) (0.102) (0.097) (0.101) (0.102) (0.101) (0.096) (0.182) Female 0.221*** 0.224*** 0.156*** 0.203*** 0.216*** 0.209*** 0.138*** 0.236*** (0.029) (0.029) (0.027) (0.031) (0.029) (0.029) (0.029) (0.052) Age -0.040*** -0.038*** -0.029*** -0.038*** -0.039*** -0.039*** -0.028*** -0.022** (0.006) (0.006) (0.006) (0.006) (0.006) (0.006) (0.006) (0.009) Age ^2 0.000*** 0.000*** 0.000*** 0.000*** 0.000*** 0.000*** 0.000*** 0.000** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (Reference group: Married with children) Married without children 0.092* 0.091* 0.037 0.081* 0.098** 0.088* 0.044 0.015 (0.047) (0.047) (0.044) (0.046) (0.047) (0.047) (0.043) (0.073) Single with children -0.140*** -0.122** -0.101** -0.088* -0.136*** -0.145*** -0.075 0.053 (0.050) (0.050) (0.047) (0.049) (0.050) (0.050) (0.047) (0.079) Single without children -0.086* -0.081 -0.102** -0.067 -0.079 -0.088* -0.077 -0.013 (0.050) (0.050) (0.047) (0.049) (0.050) (0.050) (0.047) (0.080) Separated/Divorced/Widow with children -0.125 -0.098 -0.028 -0.085 -0.124 -0.108 -0.003 0.120 (0.083) (0.082) (0.077) (0.081) (0.083) (0.082) (0.076) (0.119) Separated/Divorced/Widow without children -0.404*** -0.406*** -0.265*** -0.337*** -0.390*** -0.398*** -0.251*** -0.321** (0.099) (0.099) (0.091) (0.095) (0.099) (0.098) (0.090) (0.148) (Reference group: 1 person older than 15 in household) 2 people older than 15 in household 0.018 0.007 0.005 0.008 0.022 0.021 0.010 -0.024 (0.086) (0.086) (0.084) (0.084) (0.086) (0.086) (0.084) (0.112) More than 2 people older than 15 in household 0.030 0.008 -0.033 -0.031 0.030 0.030 -0.030 0.003 (0.081) (0.082) (0.080) (0.080) (0.081) (0.081) (0.079) (0.107) Country fixed effects YES  YES  YES  YES  YES  YES  YES  YES Month and Year of Interview YES  YES  YES  YES  YES  YES  YES  YES Constant 5.560*** 5.686*** 5.708*** 5.747*** 5.768*** 5.628*** 5.824*** 5.588*** (0.172) (0.173) (0.163) (0.178) (0.173) (0.172) (0.173) (0.260) Observations 25,244 25,244 25,244 25,244 25,244 25,244 25,244 9,065 R-squared 0.071 0.078 0.197 0.121 0.074 0.079 0.206 0.192 Note: i. Robust standard errors in parentheses; ***p<0.01; **p<0.05; *p<0.10; ii. Developing MENA includes Algeria, Egypt, Iraq, Jordan, Lebanon, Morocco, Palestine, Syria, Tunisia, and the Republic of Yemen. iii. Employment status includes an additional category (2009) which captures individuals other than employed. 17    Our ordinary least squares (OLS) results largely hold when controlling for interview dates, mood, health (appendix C1), examining heterogeneity with the MENA region (appendix C2), and using alternative variable specifications (appendix C3). Only when we add mood to our OLS baseline regression (model 7), the coefficients for dissatisfaction with availability of high quality jobs and dissatisfaction with the educational system are reduced and become statistically insignificant. Finally, model 8 replicates model 7 with a reduced sample of Arab Spring countries in which all coefficients behave similarly. Therefore, the conclusions based on the full specification for the whole sample of developing MENA countries (table 2, model 7) hold for the reduced sample of Arab Spring countries (table 2, model 8). Dealing with Reverse Causality: Lewbel IV Estimator Our analysis possibly suffers from endogeneity bias. Reverse causality may be a particular problem since life evaluation and domain satisfaction are often jointly determined. Although the usage of conventional instrumental variable (IV) methods would be preferred in a cross-section setting, finding credible instruments is difficult; thus we made use of the Lewbel IV estimator to account for reverse causality. Conventional IVs have to satisfy the following restrictions: the instrument has to be correlated with the independent variables and has to be uncorrelated with the dependent variable and the error term. In our case, a valid instrument should be correlated with the independent variables in our regression, the life domain perceptions, but not with life satisfaction. Given the general unavailability of good instruments with this property, we resort to the implementation of an instrumental variable estimation using heteroskedasticity-based instruments for cross-sectional data, suggested by Lewbel (2012). The Lewbel IV estimator uses internally generated instruments comparable to difference Generalized Method of Moments (GMM) and system GMM in a panel data setting to isolate the effect of perceptions on life satisfaction. According to Lewbel (2012), in the absence of conventional IVs, a vector of exogenous variables Z equal to X or a subset of X can be used to generate external instruments [Z- E(Z)]ε, given that there is some heteroskedasticity in the standard errors ε, and E(Xε)=0, and cov(Z,ε)≠0. (2) The validity of these assumptions for our data can be questioned, so we first examine whether the Lewbel requirements are met for regression model (1). First, we test for the presence of heteroskedatisticity. Following Lewbel (2012), we performed a Breusch and Pagan Lagrange Multiplier Test to test for heteroskedasticity. The results show that the test statistic is significantly different from zero in all cases, indicating that there is enough variance in our data to avoid weak instruments. Second, before estimating the second stage of the regressions using the generated 18    instruments, we carefully consider the choice of Z. As indicated by Lewbel (2012), the vector of exogenous variables Z can be a set or subset of X and therefore the obtained estimates could be largely dependent on the specific choice of X’s. Although in general the choice of Z can be random, subject to conditions 2 above, we opted to follow a different strategy to select our instruments. Our strategy for choosing Z is based on the correlation matrix of the generated instruments. The subset of X had to satisfy two basic conditions: (i) it had to be uncorrelated with the dependent variable Y and (ii) it had to be statistically correlated with X in the first place. The generated instruments that did not meet these conditions were excluded from the second-stage regression. After testing whether the conditions were satisfied, we chose a set of instruments and estimated the model using generalized method of moments (GMM). Table 3 provides a replication of table 2 using the Lewbel IV estimator. Several results stand out. First, dissatisfaction with freedom to choose life is not significant in model 10 or in the full specification in model 15, showing that freedom does not explain variation in life satisfaction in developing MENA in the wake of the Arab Spring. Second, in line with the OLS results, dissatisfaction with standards of living, income, and job status remain robust in sign and highly significant predictors across all specifications (models 11, 12, 15, and 16). Third, perceived poor job conditions, reflected in dissatisfaction with the efforts of the government to improve the number of high quality jobs and the educational system, do not have a significant effect on life satisfaction (models 15 and 16). It is highly likely that these domains are jointly determined or are partly reflected by satisfaction with standards of living. Fourth, the effect of cronyism and ‘wasta’ on satisfaction with life remains significant, but the effect of widespread corruption is no longer significant (models 15 and 16). This result supports our initial finding that people are affected not so much by government corruption, but by cronyism and ‘wasta,’ which make it difficult for people to succeed even when working hard. Drivers of Life Satisfaction Changes on the Eve of the Arab Spring Perceptions about living standards, job market conditions, and cronyism have had an important effect on life satisfaction in MENA. This section explores the degree to which each of these factors has contributed to the change in life satisfaction in the period 2009-10. We decompose the change in life satisfaction into the sum of all effects attributed to changes in the incidence of dissatisfaction with each of the domains included in model 1 and another sum of effects, reflecting the change in the importance of each of these domains for people’s life satisfaction between 2009 and 2010. d LS=∑ ∑ (3) 19    Table 3: Determinants of Life Satisfaction in MENA: Lewbel Estimates (10) (11) (12) (13) (14) (15) (16) VARIABLES Model 10 Model 11 Model 12 Model 13 Model 14 Model 15 Model 1 DEV DEV DEV DEV DEV DEV Arab MENA MENA MENA MENA MENA MENA Spring Dissatisfaction with freedom to choose life -0.243 -0.011 0.069 (0.340) (0.789) (0.923) Dissatisfied with standard of living: Yes -1.299*** -1.181*** -1.288*** (0.100) (0.126) (0.186) Income (1,000’s) 0.024*** 0.030*** 0.023*** 0.026*** (0.002) (0.002) (0.002) (0.003) Dissatisfied with efforts to increase high quality -0.353*** -0.085 -0.218 jobs: Yes (0.092) (0.262) (0.333) Dissatisfied with the educational system or the -0.118 -0.076 0.515 schools: Yes (0.245) (0.245) (0.524) (Reference group: Full-time Employed) Self-employed 0.079 0.050 -0.037 (0.064) (0.061) (0.105) Unemployed -0.539*** -0.353*** -0.504*** (0.083) (0.086) (0.163) Out of workforce 0.012 -0.021 -0.038 (0.049) (0.050) (0.078) Underemployed -0.287*** -0.126 -0.251* (0.084) (0.082) (0.135) (Reference group: Other) Working for the government 0.233*** 0.187*** 0.487*** (0.055) (0.054) (0.134) Corruption widespread within government: Yes -0.367*** -0.188 -0.216 (0.128) (0.181) (0.247) People cannot get ahead by working hard: Yes -0.589*** -0.324** -0.512* (0.134) (0.154) (0.269) (Reference group: Muslim) Not Muslim/Other religion 0.245*** 0.180*** 0.165** 0.277*** 0.233*** 0.175* 0.238 (0.081) (0.069) (0.076) (0.075) (0.074) (0.094) (0.169) (Reference group: Completed elementary education or less) Completed 9-15 years of education 0.443*** 0.305*** 0.360*** 0.447*** 0.445*** 0.280*** 0.368*** (0.035) (0.032) (0.033) (0.033) (0.033) (0.037) (0.062) 20    (10) (11) (12) (13) (14) (15) (16) Completed four years of education beyond high 0.901*** 0.569*** 0.679*** 0.918*** 0.899*** 0.539*** 0.486*** school and/or 4-year college degree. (0.057) (0.050) (0.054) (0.053) (0.053) (0.068) (0.136) (Reference group: Not a migrant) Migrant -0.141 -0.200** -0.184* -0.142 -0.155 -0.272*** -0.718*** (0.102) (0.098) (0.102) (0.102) (0.101) (0.096) (0.188) Female 0.223*** 0.157*** 0.203*** 0.215*** 0.206*** 0.140*** 0.226*** (0.029) (0.027) (0.031) (0.029) (0.029) (0.038) (0.072) Age -0.039*** -0.029*** -0.038*** -0.039*** -0.038*** -0.029*** -0.022** (0.006) (0.006) (0.006) (0.006) (0.006) (0.006) (0.009) Age ^2 0.000*** 0.000*** 0.000*** 0.000*** 0.000*** 0.000*** 0.000** (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (0.000) (Reference group: Married with children) Married without children 0.091* 0.038 0.083* 0.100** 0.087* 0.047 0.036 (0.047) (0.044) (0.046) (0.047) (0.047) (0.044) (0.076) Single with children -0.125** -0.096** -0.090* -0.134*** -0.147*** -0.081 0.066 (0.053) (0.047) (0.049) (0.050) (0.050) (0.057) (0.096) Single without children -0.081 -0.095** -0.067 -0.076 -0.089* -0.072 -0.015 (0.050) (0.047) (0.049) (0.050) (0.050) (0.047) (0.082) Separated/Divorced/Widow with children -0.106 -0.032 -0.100 -0.123 -0.105 -0.011 0.111 (0.086) (0.076) (0.082) (0.083) (0.082) (0.089) (0.124) Separated/Divorced/Widow without children -0.403*** -0.273*** -0.343*** -0.385*** -0.399*** -0.252*** -0.290* (0.099) (0.091) (0.095) (0.099) (0.098) (0.093) (0.153) (Reference group: 1 person older than 15 in household) 2 people older than 15 in household 0.010 0.017 0.007 0.023 0.024 0.010 -0.037 (0.087) (0.084) (0.085) (0.086) (0.086) (0.089) (0.125) More than 2 people older than 15 in household 0.015 -0.022 -0.032 0.030 0.030 -0.032 -0.018 (0.085) (0.080) (0.081) (0.081) (0.081) (0.093) (0.135) Constant 5.159*** 5.452*** 5.471*** 5.364*** 5.116*** 5.570*** 5.446*** (0.213) (0.159) (0.182) (0.196) (0.168) (0.198) (0.254) Observations 25,244 25,244 25,244 25,244 25,244 25,244 9,065 R-squared 0.077 0.193 0.117 0.074 0.079 0.204 0.076 Statistics Underidentification test: P-value 83.04 1105.78 194.93 503.81 454.68 29.508 23.872 (0.000) (0.000) (0.000) (0.000) (0.000) (0.013) (0.475) Cragg-Donald Wald F statistic 42.14 563.35 31.23 729.53 425.07 1.826 1.533 Stock-Yogo VC 10% 10.27 10.27 10.89 19.53 10.83 NA NA Hansen statistic 4.25 4.63 5.13 0.133 0.924 8.327 7.516 (0.234) (0.200) (0.953) (0.715) (0.921) (0.871) (0.873) 21    The first sum is the first-order or direct effect. It reflects the contribution attributed to the changes in the percentage of people dissatisfied with domains X in period 2 relative to period 1. If x2 > x1, a higher share of the population has become dissatisfied with certain aspects of individual or social life. The second-order effect shows the part of the negative association attributed to changes in the size of the effect of the obtained coefficients, implying a change in the relative importance of that factor to life satisfaction (LS). In other words, the indirect effect shows evidence that perceptions have changed, making individuals less tolerant of certain social conditions, for instance, cronyism and ‘wasta.’ Table 4: Decomposition of the Change in Life Satisfaction between 2009 and 2010 (Based on Lewbel Estimates of Model (1)) Developing Developing Arab Spring Arab Spring MENA MENA Countries Counties First Order Effect Second Order First Order Effect Second Order Effect Effect Dissatisfaction with standards of -0.031 0.037 -0.084 0.005 living People cannot get ahead by 0.005 0.038 NS NS working hard (Yes) Dissatisfaction with efforts of the government to increase high -0.012 0.021 -0.030 -0.060 quality jobs Dissatisfaction with freedom to -0.014 -0.074 NS NS choose life Corruption widespread within -0.008 -0.088 NS NS government/business (Yes) Unemployed 0.000 0.025 -0.012 0.004 Working for the government NS NS 0.020 0.043 Income (1,000’s) -0.015 0.010 -0.025 -0.077 Notes: (i) Developing MENA includes Algeria, Egypt, Iraq, Jordan (only available for 2009), Lebanon, Morocco, Palestine, Syria, Tunisia, and the Republic of Yemen. (ii) Arab Spring Countries include Egypt, Libya, Syria and the Republic of Yemen. (iii) We only present the coefficients that were significant at least for one out of two years. (iv) The coefficients that are not significant are marked as NS. (v) The full table with results can be found in the appendix. Table 4 shows the decomposition of change in LS into the contributions of direct and indirect effects of domain satisfactions between 2009 and 2010. A more detailed table of this decomposition of effects is provided in appendix D1. In appendix D2, we also provide the results estimated with ordinary least squares. In developing MENA, the largest negative contribution to dissatisfaction with life given by the first-order effects is attributed to the increased share of individuals dissatisfied with their standard of living (-0.031) and decrease in reported income (- 0.015). Similar findings are observed for the Arab Spring countries; in this case, the coefficients are -0.084 and -0.025, respectively. The size of the effects of corruption and limited freedom on life 22    satisfaction rose in the second period, although the coefficients were found to be insignificant in those specifications. In the Arab Spring countries, the largest negative second-order effect on life satisfaction comes from dissatisfaction with the efforts of the government to increase the number of high quality jobs (-0.060). 5. Discussion and Concluding Remarks How is the declining dissatisfaction prior to the Arab Spring linked to the protests? Unfortunately, the Gallup World Poll does not have information on the reasons for the Arab Spring protests. Therefore, we turn to information from the third wave of the Arab Barometer, in which respondents in developing MENA countries (Algeria, Egypt, Iraq, Lebanon, Morocco, Palestine, Syria, Tunisia, and the Republic of Yemen) were asked to mention the main three reasons that led to the Arab Spring. It appears that the main reasons behind the outburst of social rage during the Arab Spring uprisings are domain satisfactions shaping the level of subjective well-being in developing MENA prior to the Arab Spring (figure 4). Figure 4: Reasons Why the Arab Spring Occurred According to the Developing MENA Population 63.55% 42.40% 28.77% 64.26% 15.74% 57.21% 7.53% 14.56% Betterment of the economic situation Civil and political freedoms, and emancipation from oppression Dignity Fighting corruption Rule of law Social and economic justice Weakening the political and economic relations  with the West Weakening the political and economic relations with Israel   Source: Arab Barometer 2012‐2014.  23    Fighting corruption was mentioned as the most important reason for the Arab Spring by 64.3 percent of respondents, followed by betterment of the economic situation (63.4 percent) and social and economic justice (57.2 percent). These findings are in line with a poll by Zogby in 2005, in which respondents in developing MENA countries indicated that the lack of employment opportunities, corruption, health care, and schooling were seen as the most pertinent problems in developing MENA countries (Zogby 2005). Strikingly, civil and political freedom (42.4 percent) only comes in fourth place and is, hence, neither found associated with dissatisfaction in developing MENA nor regarded as one of the most important factors related to the uprisings. Likewise, relations with the West (7.5 percent) and Israel (14.6 percent) as well as rule of law (15.7 percent) and dignity (28.8 percent) were less often mentioned as important reasons for the Arab Spring, and were not found to be an important determinant of dissatisfaction with life in developing MENA. Hence, standards of living, labor market conditions, and ‘wasta’ are not only strongly related to dissatisfaction with life prior to the Arab Spring, but also mentioned as the main reasons for the Arab Spring uprisings. In sum, it can be concluded that the Arab Spring uprisings in developing MENA countries were preceded by a decline in life satisfaction from already low happiness levels, despite economic and human development progress in the prior two decades. In many developing MENA countries, the so-called “unhappy development” paradox was accompanied by social discontent driven by poor or worsening standards of living, labor market conditions, and crony capitalism. In this light, our study highlights that not only objective conditions count, but also the subjective awareness of shortcomings in these objective conditions. The rising awareness of social ills is partly due to the modernization process in which society is seen to be less of a moral order given by God, and in which an increasing number of educated people call for meritocracy rather than autocracy. Dissatisfaction alone does not bring political action, which typically arises only in combination with perceived chances for change (Klandermans 1997). This paper does not explore the question why some developing MENA countries experienced political violence and fall of regimes, whereas in other developing MENA countries the protests remained rather limited. This question should be addressed in future research. References Angel-Urdinola, D. F., and A. Kuddo. 2011. “Key Characteristics of Employment Regulations in the Middle East and North Africa.” MNA Knowledge and Learning Fast Brief No. 84, World Bank, Washington, DC. 24    Arampatzi, E., M. J. Burger, and R. Veenhoven. 2015. “Financial Distress and Happiness of Employees in Times of Economic Crisis.” Applied Economics Letters 22 (3): 173–79. Arezki, R., and M. Brückner. 2011. “Food Prices and Political Instability.” IMF Working Paper, International Monetary Fund, Washington, DC. Barro, R. J., and J. W. Lee. 2010. “A New Data Set of Educational Attainment in the World, 1950-2010.” NBER Working Paper No. 15902, National Bureau of Economic Research, Cambridge, MA. Bellin, E. 2004. “The Robustness of Authoritarianism in the Middle East: Exceptionalism in Comparative Perspective.” Comparative Politics 36: 139–57. ———. 2012. “Reconsidering the Robustness of Authoritarianism in the Middle East: Lessons from the Arab Spring.” Comparative Politics 44: 127–49. Bromley, R. 2014. “The ‘Arab Spring’ Stress Test: Diagnosing Reasons for the Revolt.” Working Paper, University of Wisconsin-Madison. Bruni, L. 2004. “The ‘Technology of Happiness’' and the Tradition of Economic Science.” Journal of the History of Economic Thought 26(1): 19–43. Cammett, M. C., and I. Diwan. 2013. The Political Economy of the Arab Uprisings. Perseus Books Group. Campante, F. R., and D. Chor. 2012. “Why Was the Arab World Poised for Revolution? Schooling, Economic Opportunities, and the Arab Spring.” Journal of Economic Perspectives 26: 167–88. Cantril, H. 1965. Pattern of Human Concerns. New Brunswick: Rutgers University. Chamlou, N. 2013. “The Economics of Happiness and Anger in North Africa.” Working Paper UNU-Merit. Chekir, H., and I. Diwan. 2012. “Center for International Development. Cingranelli-Richards Human Rights Dataset, 2015.” Clark, A., A. Knabe, and S. Rätzel. 2010. “Boon or Bane? Others’ Unemployment, Well-Being and Job Insecurity.” Labour Economics 17: 52–61. Clark, A. E., and A. J. Oswald. 1994. “Unhappiness and Unemployment.” The Economic Journal 104: 648–59. 25    Deaton, A. 2008. “Income, Health and Well-Being around the World: Evidence from the Gallup World Poll.” Journal of Economic Perspectives 22: 53–72. Di Tella, R., R. J. MacCulloch, and A. J. Oswald. 2003. “The Macroeconomics of Happiness.” Review of Economics and Statistics 85: 809–27. Diener, E. 1984. “Subjective Well-Being.” Psychological Bulletin 95(3): 542-75. Dorn, D., J. A. Fischer, G. Kirchgässner, and A. Sousa-Poza. 2007. “Is It Culture or Democracy? The Impact of Democracy and Culture on Happiness.” Social Indicators Research 82(3): 505–26. Fereidouni, H. G., Y. Najdi, and R. E. Amiri. 2012. Do Governance Factors Matter for Happiness in the MENA Region? International Journal of Social Economics 12: 1028–40. Ferrari, A. J., F. J. Charlson, R. E. Norman, S. B. Patten, G. Freedman, C. J. Murray, et al. 2013. “Burden of Depressive Disorders by Country, Sex, Age, and Year: Findings from the Global Burden of Disease Study 2010.” PLoS Medicine 10(11): e1001547. Ferrer‐i‐Carbonell, A., and P. Frijters. 2004. “How Important Is Methodology for the Estimates of the Determinants of Happiness?” The Economic Journal 114: 641–59. Freedom House. 2008. “Freedom in the World Report: Methodology Section.” Freedom House, Washington, DC. Frey, B. S., and A. Stutzer. 2000. “Happiness Prospers in Democracy.” Journal of Happiness Studies 1: 79–102. Gallie, D., and H. Russell. 1998. “Unemployment and Life Satisfaction: A Cross-Cultural Comparison.” European Journal of Sociology 39: 248–80. Gause III, G. F. 2011. “Why Middle East Studies Missed the Arab Spring.” Foreign Affairs 90: 81– 90. Goodwin, J. 2011. “Why We Were Surprised (Again) by the Arab Spring.” Swiss Political Science Review 17: 452–56. Graham, C., and E. Lora. 2009. “Happiness and Health Satisfaction across Countries.” In Paradox and Perception: Measuring Quality of Life in Latin America, edited by C. Graham and E. Lora. Washington, DC: Brookings Institution Press. 26    Headey, B., R. Veenhoven, and A. Wearing. 1991. “Top-Down versus Bottom-Up Theories of Subjective Well-Being.” Social Indicators Research 24: 81–100. Helliwell, J. F., R. Layard, and J. Sachs (eds.). 2015. World Happiness Report 2015. New York: United Nations. Ianchovichina, E. I., J. L. Loening, and C. A. Wood. 2014. “How Vulnerable Are Arab Countries to Global Food Price Shocks?” The Journal of Development Studies 50 (9): 1302–19. Ianchovichina, E., L. Mottaghi, and S. Devarajan. 2015. Inequality, Uprisings, and Conflict in the Arab World. Washington, DC: World Bank. Inglehart, R., R. Foa, C. Peterson, and C. Welzel. 2008. “Development, Freedom, and Rising Happiness: A Global Perspective (1981-2007).” Perspectives on Psychological Science 3: 264–85. Iqbal, F., and Y. Kiendrebeogo. 2015. “The Reduction of Child Mortality in the Middle East and North Africa: A Success Story.” Policy Research Working Paper No. 7127, World Bank, Washington, DC. Kaufmann, D., A. Kraay, and M. Mastruzzi. 2011. “The Worldwide Governance Indicators: Methodology and Analytical Issues.” Hague Journal on the Rule of Law 3 (02): 220–46. Klandermans, B. 1997. The Social Psychology of Political Protest. Oxford, England: Blackwell. Korotayev, A., and J. V. Zinkina. 2011. “Egyptian Revolution: A Demographic Structural Analysis.” Entelequia: revista interdisciplinar (13): 139–69. Lagi, M., K. Z. Bertrand, and Y. Bar-Yam. 2011. The Food Crisis and Political Instability in North Africa and the Middle East. New England Complex System Institute. Layard, R. 2006. Happiness: Lessons from a New Science. Penguin. Lewbel, A. 2012. “Using Heteroscedasticity to Identify and Estimate Mismeasured and Endogenous Regressor Models.” Journal of Business & Economic Statistics 30. Malik, A., and B. Awadallah. 2013. “The Economics of the Arab Spring.” World Development 45: 296–313. Maslow, A. H. 1943. “A Theory of Human Motivation.” Psychological Review 50: 370–96. OECD (Organisation for Economic Co-operation and Development). 2009. “Ownership Structures in MENA Countries: List Companies, State-Owned, Family Enterprises and 27    Some Policy Implications.” OECD, Paris. http://www.oecd.org/mena/investment/35402110.pdf. Okulicz-Kozaryn, A. 2011. “City Life: Rankings (Livability) versus Perceptions (Satisfaction).” Social Indicators Research 110: 433–51. Ott, J. C. 2010. “Good Governance and Happiness in Nations: Technical Quality Precedes Democracy and Quality Beats Size.” Journal of Happiness Studies 11: 353–68. Rijkers, B., C. Freund, and A. Nucifora. 2014. “All in the Family: State Capture in Tunisia.” Policy Research Working Paper 6810, World Bank, Washington, DC. Stevenson, B., and J. Wolfers. 2008. “Economic Growth and Happiness.” Brookings Papers on Economic Activity 35. Veenhoven, R. 1984 Conditions of Happiness. Dordrecht/Boston: Reidel. ———. 1989. National Wealth and Individual Happiness. Springer Netherlands ———. 2002. “Why Social Policy Needs Subjective Indicators.” Social Indicators Research 58: 33– 46. ———. 2012. “Happiness: Also Known As ‘Life Satisfaction’ and ‘Subjective Well-Being.’” In Handbook of Social Indicators and Quality of Life Research, edited by K. C. Land, a. C. Michalos, and M. J. Sirgy, 63–77. Berlin: Springer. Verme, P. 2009. “Happiness, Freedom and Control.” Journal of Economic Behavior & Organization 71: 146–61. Verme, P., B. Milanovic, S. Al-Shawarby, S. El Tawila, M. Gadallah, and E. A. A. El-Majeed. 2014. Inside Inequality in the Arab Republic of Egypt: Facts and Perceptions across People, Time, and Space. Washington, DC: World Bank. World Bank. 2014a. PovcalNet. World Bank, Washington, DC. http://iresearch.worldbank.org/PovcalNet/. World Bank. 2014b. World Development Indicators. Washington, DC: World Bank. Zogby, J. 2005. Attitudes of Arabs: An In-Depth Look at Social and Political Concerns of Arabs. Arab American Institute. 28    Appendixes Appendix A: Average Life Satisfaction (ALS) across Countries, 2006-12 Rank Country ALS Rank Country ALS Rank Country ALS 1 Denmark 7.80 54 Poland 5.79 107 China 4.84 2 Switzerland 7.59 55 Sint Maarten 5.79 108 Djibouti 4.84 3 Norway 7.58 56 El Salvador 5.78 109 Zambia 4.81 4 Netherlands 7.51 57 Bolivia 5.71 110 India 4.79 5 Finland 7.50 58 Croatia 5.65 111 Bangladesh 4.78 6 Canada 7.47 59 Kazakhstan 5.64 112 Iraq 4.78 7 Sweden 7.41 60 Lithuania 5.59 113 Mozambique 4.76 8 Iceland 7.36 61 Jordan 5.57 114 Mongolia 4.72 9 Australia 7.32 62 Belarus 5.55 115 Serbia 4.72 10 New Zealand 7.31 63 Ecuador 5.55 116 Angola 4.68 11 Austria 7.30 64 Paraguay 5.50 117 Azerbaijan 4.64 12 Costa Rica 7.25 65 Mauritius 5.48 118 Mauritania 4.58 13 Israel 7.22 66 Moldova 5.47 119 Palestine 4.58 14 United States 7.19 67 Hong Kong SAR, China 5.45 120 Tajikistan 4.55 15 Ireland 7.18 68 Uzbekistan 5.45 121 Egypt, Arab Rep. 4.53 16 Belgium 7.08 69 Vietnam 5.45 122 Macedonia, FYR 4.53 17 Luxembourg 7.04 70 Bahrain 5.43 123 Armenia 4.42 18 United Arab Emirates 7.04 71 Peru 5.43 124 Botswana 4.42 19 Panama 6.92 72 Algeria 5.42 125 Malawi 4.42 20 Mexico 6.91 73 Cuba 5.42 126 Nepal 4.42 21 United Kingdom 6.89 74 Estonia 5.37 127 Sudan 4.42 22 Venezuela, RB 6.89 75 Libya 5.37 128 Uganda 4.39 23 Oman 6.85 76 Albania 5.36 129 Congo, Dem. Rep. 4.38 24 Brazil 6.80 77 Kosovo 5.36 130 Cameroon 4.36 25 France 6.75 78 Russian Federation 5.35 131 Syrian Arab Republic 4.32 26 Germany 6.64 79 Honduras 5.34 132 Senegal 4.31 27 Spain 6.61 80 Turkey 5.26 133 Yemen, Rep. 4.27 28 Puerto Rico 6.59 81 Portugal 5.25 134 Kenya 4.26 29 Qatar 6.58 82 Indonesia 5.23 135 Sri Lanka 4.25 30 Saudi Arabia 6.58 83 Nicaragua 5.20 136 Côte d'Ivoire 4.20 31 Singapore 6.55 84 Montenegro 5.18 137 Madagascar 4.14 32 Kuwait 6.48 85 Romania 5.15 138 Mali 4.14 33 Cyprus 6.46 86 Pakistan 5.14 139 Niger 4.14 34 Belize 6.45 87 South Africa 5.09 140 Haiti 4.13 35 Argentina 6.35 88 Ukraine 5.08 141 Congo, Rep. 4.12 36 Czech Republic 6.35 89 Dominican Republic 5.05 142 Zimbabwe 4.12 37 Trinidad and Tobago 6.35 90 Nigeria 5.04 143 Gabon 4.11 38 Italy 6.33 91 Lao PDR 5.01 144 Afghanistan 4.09 39 Suriname 6.27 92 Lebanon 4.98 145 Burkina Faso 4.08 40 Colombia 6.26 93 Tunisia 4.98 146 Cambodia 4.07 41 Chile 6.25 94 Iran, Islamic Rep. 4.91 147 Liberia 4.04 42 Guatemala 6.14 95 Hungary 4.90 148 Rwanda 4.03 43 Uruguay 6.07 96 Kyrgyz Republic 4.90 149 Chad 4.00 44 Japan 6.06 97 Lesotho 4.90 150 Guinea 4.00 45 Malta 6.02 98 Ghana 4.89 151 Georgia 3.99 46 Thailand 6.02 99 Myanmar 4.89 152 Bulgaria 3.95 47 Guinea-Bissau 5.99 100 Namibia 4.89 153 Central African Rep 3.87 48 Slovak Republic 5.98 101 Philippines 4.89 154 Tanzania 3.87 49 Turkmenistan 5.94 102 Somalia 4.89 155 Sierra Leone 3.77 50 Korea, Rep. 5.89 103 Bosnia and Herzegovina 4.87 156 Comoros 3.74 51 Greece 5.83 104 Latvia 4.87 157 Burundi 3.69 52 Malaysia 5.83 105 Morocco 4.87 158 Benin 3.51 53 Jamaica 5.81 106 Swaziland 4.87 159 Togo 2.98 Note: Developing MENA countries are highlighted 29    Appendix B1: Description of Variables Category: Independent Variable code Exact question Answer categories perception variables Satisfaction with Standard of Wp30 Are you satisfied or dissatisfied 1 Yes Living with your standard of living, all the 2 No things you can buy and do? Satisfaction with Standard of Wp40 Have there been times in the past 1 Yes Living twelve months when you did not 2 No (Index construction) have enough money to buy food that you or your family needed? Satisfaction with Standard of Wp43 Have there been times in the past 1 Yes Living twelve months when you did not 2 No (Index construction) have enough money to provide adequate shelter or housing for you and your family? Satisfaction with Standard of Index_fs Construction of variable wp40 and Not applicable Living wp43 (Alternative specification) Satisfaction with freedom TO Wp134 Are you satisfied or dissatisfied 1 Yes CHOOSE LIFE with your freedom to choose what 2 No you do with your life? Satisfaction with civil freedom Wp143 Do you have confidence in the 1 Yes (Alternative specification) Quality and Integrity of the 2 No Media? Perceptions about Corruption Wp145 Is corruption widespread within 1 Yes business? 2 No Perceptions about Corruption Wp146 Is corruption widespread within 1 Yes government? 2 No Perceptions about Corruption Wp6267 Do you think the level of 1 Same or lower corruption in this country is lower, 2 Higher about the same or higher than it was 5 years ago? Cronyism Wp128 Can people in this country get 1 Yes ahead by working hard or not? 2 No Quality of jobs Wp133 Are you satisfied or dissatisfied 1 Yes with efforts to increase the 2 No number of quality jobs? Quality of jobs Wp89 Thinking about the job situation in 1 Good time (Alternative specification) the city or area where you live 2 Bad time today, would you say that it is now a good time or a bad time to find a job? Satisfaction with education Wp93 In the city or area where you live, 1 Approve are you satisfied or dissatisfied 2 Disapprove with the education system or the schools? 30    Category: Other control Personal variables information Gender Wp1219 1 Male 2 Female Age Wp1220 Until 99 Marital children = Computed Marital_children Combination to Wp 1223 and Wp 1 Married with from marital status and number of 1230 children children 2 Married without children 3 Single with children 4 Single without children 5 S/D/W with children 6 S/D/W without children Marital status Wp1223 What is your current marital 1 Single/never been (Index construction) status? married 2 Married 3 Separated/ divorced/ widowed Number of children Wp1230 How many children under 15 (Index construction) years of age are now living in your household? Religion religion 1 Muslim 2 Non-Muslim/other religion Migration status Wp4657 Were you born in this country, or 1 Born in this country not? 2 Born in another country Level of education wp3117 1 Completed elementary education or less 2 Secondary - 3 year tertiary secondary 3 Completed four years of education beyond high school and/or received a 4- year college degree Employment status emp_2010 1 Employed full time for an employer/ Employed part time/ do not want full time 2 Employed full time for self 3 Unemployed 4 Out of workforce 5 Underemployed 6 Other Government employee Wp1227 Are you a government worker or 1 Other not? 2 Yes 3 Undetermined Household composition Adults Wp12 Including yourself, how many 1 One people who are residents of age 15 2 Two or over currently live in this 3 More than two household? Household income (US$, inc_001 Expressed in thousands) international dollars Month and year of Interview m_year Appendix B2: Descriptive Statistics 31    Variable Observations Mean SD Min. Max. Life evaluation 25,244 5.09 2.00 0 10 Dissatisfied with standard of living: 25,244 0.37 .48 0  1 Yes People cannot get ahead by working 25,244 0.17 .37 0  1  hard: Yes Dissatisfied with efforts to increase 25,244 0.66 .47 0  1  with high quality jobs: Yes Dissatisfied with freedom to choose 25,244 0.38 .48 0  1  life: Yes Dissatisfied with the educational 25,244 0.37 .48 0  1  system or the schools: Yes Corruption widespread within 25,244 0.78 .41 0  1  government*: Yes Self-employed 25,244  0.08 0.27 0  1  Unemployed 25,244  0.04 0.20 0  1  Out of workforce 25,244 0.31 0.46 0 1 Underemployed 25,244 0.03 0.18 0 1 Other** 25,244 0.25 0.43 0 1 Government worker 25,244 0.09 0.29 0 1 Undetermined 25,244 0.28 0.45 0 1 Not Muslim 25,244 0.07 0.26 0 1 Completed 9-15 years of education 25,244 0.48 0.49 0 1 Completed four years of education 25,244 0.11 0.31 0 1 beyond high school and/or 4-year college degree Migrant 25,244 0.02 0.15 0 1 Income (1,000s) 25,244 10.16 12.20 0 229.99 Female 25,244 0.48 0.49 0 1 Age 25,244 35.23 14.54 15 99 Age squared 25,244 1,453.04 1,210.98 15 99 Married without children 25,244 0.15 0.35 225 9,801 Single with children 25,244 0.20 0.40 0 1 Single without children 25,244 0.17 0.37 0 1 Separated/divorced/widow with 25,244 0.03 0.17 0 1 children Separated/divorced/widow without 25,244 0.26 0.16 0 1 children 2 people older than 15 years in 25,244 0.23 0.42 0 1 household More than 2 people older than 15 25,244 0.73 0.45 0 1 years in household Alternative Measures Index_fs 21,376 0.41 0.67 0 2 Bad time to find a job: Yes 23,592 0.71 0.46 0 1 Are levels of corruption higher: Yes 10,926 0.55 0.65 0 1 Index positive affect 12,582 64.11 29.09 0 100 Index negative affect 4,739 33.13 29.96 0 100 Dissatisfaction with health: Yes 11,016 0.16 0.36 0 1 32    Appendix B3: Correlation Matrix (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) Life evaluation 1.00 Dissatisfaction with standard of living: Yes -0.36 1.00 People cannot get ahead by working hard: -0.09 0.16 1.00 Yes Dissatisfied with efforts to increase high -0.13 0.19 0.14 1.00 quality jobs: Yes Dissatisfied with freedom to choose life: -0.10 0.16 0.16 0.33 1.00 Yes Dissatisfied with educational system -0.13 0.16 0.09 0.17 0.17 1.00 Corruption widespread within government: -0.07 0.10 0.07 0.19 0.11 0.13 1.00 Yes Income (1,000’s) 0.22 -0.16 -0.01 -0.07 -0.08 -0.09 0.00 1.00 Self-employed 0.00 0.00 0.00 0.01 -0.02 -0.01 0.02 0.08 1.00 Unemployed -0.07 0.09 0.04 0.05 0.04 0.03 0.00 -0.05 -0.06 1.00 Out of workforce -0.03 0.01 -0.01 0.03 0.04 -0.02 -0.02 -0.07 -0.20 -0.14 1.00 Underemployed -0.04 0.07 0.02 0.03 0.02 0.04 0.01 -0.03 -0.06 -0.04 -0.13 1.00 Government employee 0.04 -0.04 -0.03 -0.03 -0.01 0.02 -0.03 0.01 -0.04 -0.07 -0.22 0.04 1.00 Undetermined 0.05 -0.02 0.03 -0.04 0.00 -0.01 0.00 -0.04 -0.12 -0.13 -0.42 0.02 -0.20 1.00     33    Appendix C1: Robustness Analysis: Omitted Variable Bias Our analysis possibly suffers from simultaneity and omitted variable biases. It is well known that in survey research happier respondents, or those who are in a better mood during an interview, have a tendency to report more positively about different aspects of their life. For example, the amount of negative feelings one experiences during the day could possibly predispose people to lower their life satisfaction rating. When evaluating their satisfaction, respondents can reason as follows: “I am generally dissatisfied with my life, so apparently I am dissatisfied with my government” or “I feel sad now, so apparently I am dissatisfied with my government” (see also Diener 1984; Headey, Veerhoven, and Wearing 1991). We use different strategies to cope with these problems in our baseline OLS analysis in model 7. First, we control for mood during the interview, by including (1) interview date dummies, assuming that turbulent time indicators of satisfaction can be subject to daily developments; (2) affect indices by Gallup related to very recent positive13 and negative experiences14 measured on a 0-100 scale; and (3) satisfaction with health measured on a 0-10 scale. This way we are able to capture the daily mood of individuals, which may affect the responses related to satisfaction. Table C1 shows the results of these additional robustness checks. The effects of dissatisfaction with the standard of living, income, and employment status remain significant and robust to controlling for interview dates, mood, and health satisfaction. Likewise, perceptions of inequality of opportunity, corruption, and crony capitalism remain an important source of dissatisfaction in developing MENA, where the dissatisfaction with not being able to get ahead by working hard and feelings about corruption in government are negative and statistically significant in most specifications. In general, the inclusion of interview dates (column 1) or satisfaction with health (column 4) does not affect the main conclusions drawn from the results presented in table 2. However, when we add a negative experience index to our baseline regression, the coefficient for unemployed is reduced and becomes statistically insignificant (column 3). To some extent, this also reflects the                                                              13  The Gallup positive experience index is based on the following five questions: (1) “Did you feel well-rested yesterday?” (2) “Were you treated with respect all day yesterday?” (3) “Did you smile or laugh a lot yesterday?” (4) “Did you learn or do something interesting yesterday?” and (5) “Did you experience the following feelings during a lot of the day yesterday? How about enjoyment?” 14  The Gallup negative experience index is based on the following five feelings, which respondents had to reflect on based on the question: “Did you experience the following feelings during a lot of the day yesterday?: physical pain, worry, sadness, stress, and anger.”  34    fact that when we include the experience index, the sample size is reduced from 25,244 to 6,221 respondents. Table C1: Determinants of Life Satisfaction in Developing MENA in Alternative Models (OLS) (1) (2) (3) (4) VARIABLES +Interview + Positive + Negative + Satisfaction Dates Experience Experience with Health Index Index Dissatisfied with freedom to choose -0.039 -0.019 -0.022 -0.046 life: Yes (0.030) (0.036) (0.044) (0.048) Dissatisfaction with standard of living: -1.242*** -1.103*** -1.124*** -1.055*** Yes (0.030) (0.037) (0.044) (0.046) Income (1,000’s) 0.023*** 0.020*** 0.022*** 0.019*** (0.002) (0.002) (0.002) (0.002) Dissatisfied with efforts to increase -0.155*** -0.101*** -0.089* -0.108** high quality jobs: Yes (0.032) (0.037) (0.046) (0.050) Dissatisfied with the educational -0.169*** -0.115*** -0.099** -0.142*** system or the schools: Yes (0.029) (0.035) (0.043) (0.045) Corruption widespread within -0.083** -0.104** -0.127** -0.080 government: Yes (0.035) (0.042) (0.050) (0.056) People cannot get ahead by working -0.238*** -0.228*** -0.199*** -0.340*** hard: Yes (0.039) (0.047) (0.055) (0.061) Positive experience index 0.007*** (0.001) Negative experience index -0.007*** (0.001) Dissatisfied with personal health: Yes -0.369*** (0.060) Self-employed 0.054 0.113 -0.012 0.141* (0.061) (0.079) (0.117) (0.081) Unemployed -0.352*** -0.234** -0.112 -0.291*** (0.078) (0.103) (0.137) (0.105) Out of workforce -0.027 -0.005 -0.053 0.017 (0.047) (0.058) (0.088) (0.065) Underemployed -0.138* -0.087 -0.141 -0.180* (0.080) (0.097) (0.166) (0.101) Individual characteristics YES YES YES YES Country fixed effects YES YESA YESB YESA Month and year of interview NO YES YES YES Constant 5.839*** 5.420*** 6.221*** 6.013*** (0.198) (0.207) (0.250) (0.246) Observations 25,244 18,442 12,582 11,016 R-squared 0.230 0.201 0.198 0.191 Note: Robust standard errors are in parentheses. *** p < .01, ** p < .05, * p < .1. A Morocco missing. B Morocco and Tunisia missing. 35    Appendix C2: Robustness Analysis: Heterogeneity within Developing MENA The developing MENA region encompasses a wide variety of Arab countries. Hence, the correlates of dissatisfaction with life might differ across countries. In our robustness analysis, we distinguish between (1) North Africa, (2) Middle East, (3) Levant (including and excluding Iraq), and (4) Iraq. Table C2: Determinants of Life Satisfaction in Developing MENA by Subregion (OLS) (1) (2) (3) (4) (5) VARIABLES North Middle Levant Area 1C Levant Area 2D Iraq AfricaA EastB Dissatisfied with freedom to 0.017 -0.077* -0.115** -0.078* -0.052 choose life: Yes (0.041) (0.043) (0.055) (0.044) (0.079) Dissatisfaction with standard of -1.211*** -1.246*** -1.295*** -1.204*** -0.657*** living: Yes (0.042) (0.040) (0.052) (0.043) (0.081) Income (1,000s) 0.023*** 0.022*** 0.018*** 0.021*** 0.064*** (0.002) (0.003) (0.003) (0.003) (0.015) Dissatisfied with efforts to increase -0.117*** -0.192*** -0.096 -0.190*** -0.334*** with high quality jobs: Yes (0.039) (0.048) (0.061) (0.050) (0.098) Dissatisfied with the educational -0.218*** -0.144*** -0.148*** -0.181*** -0.285*** system or the schools: Yes (0.039) (0.041) (0.054) (0.043) (0.080) Corruption widespread within -0.079* -0.078 0.029 -0.082 -0.472*** government: Yes (0.044) (0.052) (0.062) (0.052) (0.114) People cannot get ahead by -0.209*** -0.233*** -0.283*** -0.245*** -0.172** working hard: Yes (0.060) (0.048) (0.061) (0.049) (0.084) Self-employed 0.380*** -0.143* -0.209* -0.115 0.105 (0.085) (0.083) (0.107) (0.087) (0.145) Unemployed -0.168 -0.437*** -0.576*** -0.368*** -0.220 (0.106) (0.111) (0.156) (0.114) (0.184) Out of workforce -0.011 -0.048 0.017 -0.009 -0.193 (0.061) (0.071) (0.087) (0.073) (0.154) Underemployed -0.098 -0.167 -0.185 -0.158 -0.148 (0.131) (0.102) (0.148) (0.116) (0.175) Other 0.170*** 0.269*** 0.226** 0.202** 0.114 (0.065) (0.088) (0.103) (0.081) (0.181) Individual characteristics YES YES YES YES YES Country fixed effects YES YES YES YES YES Month and year of interview YES YES YES YES YES Constant 6.142*** 5.956*** 6.226*** 5.561*** 5.291*** (0.223) (0.243) (0.297) (0.246) (0.911) Observations 10,444 14,800 9,184 13,244 2,432 R-squared 0.249 0.188 0.206 0.180 0.174 Note: Robust standard errors are in parentheses. ***p < .01; **p < .05; *p < .10. A  North Africa includes Morocco, Algeria, Tunisia, Libya, and Egypt.  B  Middle East includes Syria, Palestine, Jordan, Lebanon, the Republic of Yemen, and Iraq.  C  Levant 1 includes Syria, Palestine, Jordan, and Lebanon.  D  Levant 2 includes Syria, Palestine, Jordan, Lebanon, Egypt, and Iraq. Table C2 shows the results of the subsample analyses, where three findings stand out. First, the socioeconomic correlates of satisfaction with life are fairly consistent across different groupings of countries in the developing MENA region. Second, satisfaction with freedom to choose life is not 36    equally important for determining life evaluation. In North Africa (column 1) and Iraq (column 5), the effect of freedom has no significant value, while the most significant effect of satisfaction with freedom can be found in the Levant area (column 3). Third, the association between widespread corruption and life satisfaction is very sensitive to the selection of countries. The effect of widespread corruption is only negative and statistically significant for North Africa and Iraq. Appendix C3: Robustness Analysis: Alternative Variable Specifications In addition, we performed several robustness controls to verify the significance of our findings. Table C3 shows five alternative specifications. In specification 1, satisfaction with the standard of living is measured by the Gallup Food and Shelter Index, which is based on the question whether individuals experienced a shortage of money to provide food and shelter for their family. In specification 2, dissatisfaction with efforts to increase the number of high quality jobs is replaced by job expectations measured based on answers to the question: “Thinking about the job situation in the city or area where you live today, would you say that it is now a good time or a bad time to find a job?” In specification 3, autocracy and lack of democracy are captured by a variable related to freedom and integrity of the media based on the question: “Do you have confidence in the quality and integrity of the media?” Corruption was alternatively measured in specification 4 by perceptions about changes in the levels of corruption over the past years (“Do you think the level of corruption in this country is lower, about the same or higher than it was 5 years ago?”). Table C3 shows the results for the regressions using the alternative variable definitions. The results are not directly comparable with the results in table 2, since the alternative variables are not available for some countries and/or waves. Still, the results in table C3 show that our conclusions regarding dissatisfaction with the standard of living and job opportunities as important drivers of life dissatisfaction in developing MENA generally hold, while freedom is again found not to be important for explaining life dissatisfaction in developing MENA. Although perceptions of increased corruption seem to be associated with life satisfaction in developing MENA, its effect is smaller compared with the rest of the alternative measures reported in table C3. The effect of feelings of not being able to get ahead by working hard and dissatisfaction with the education system remains statistically significant across all specifications. 37    Table C3: Determinants of Life Satisfaction in developing MENA: Alternative Variable Specifications (OLS) (1) (2) (3) (4) VARIABLES Alternative Alternative Alternative Alternative standards of job civil freedom widespread living opportunities corruption Food and Shelter Index -0.976*** (0.059) Would you say that it is now a good -0.141*** time or a bad time to find a job: Bad (0.033) time Do you have confidence in each of the -0.029 following? How about the quality and (0.042) integrity of the media: No Level of corruption is higher -0.081*** People cannot get ahead by working -0.305*** -0.195*** -0.225*** -0.200*** hard: Yes (0.042) (0.040) (0.057) (0.039) Dissatisfied with the educational -0.252*** -0.150*** -0.166*** -0.169*** system or the schools: Yes (0.034) (0.031) (0.043) (0.030) Individual characteristics YES  YES  YES YES  Country fixed effects YESA YESB  YES YES  Month and year of interview YES  YES  YES YES  Constant 6.470*** 5.865*** 5.902*** 5.676*** (0.200) (0.180) (0.267) (0.177) Observations 21,376 23,592 10,926 24,012 R-squared 0.162 0.207 0.220 0.210 Note: Robust standard errors are in parentheses. ***p < .01; **p < .05; *p < .10. A Morocco and Syria are missing. B Morocco is missing. 38    Appendix D1: Changes in Averages and Decomposition of Effects, Lewbel Estimator Change in Change in the obtained the obtained Change in Change in Developing Developing Arab Spring Arab Spring coefficient coefficient the averages the averages MENA MENA countries counties (2009-10) (2009-10) (2009-10) (2009-10) first-order second-order first-order second -order DEV MENA Arab Spring DEV MENA Arab Spring effect effect effect effect Dissatisfaction with 0.101 0.015 0.028 0.091 -0.031 0.037 -0.084 0.005 Standards of living People cannot get ahead by 0.213 0.120 -0.012 0.004 0.005 0.038 -0.001 0.012 working hard (Yes) Dissatisfaction with efforts 0.033 -0.098 0.033 0.070 -0.012 0.021 -0.030 -0.060 of the government to increase high quality jobs Dissatisfaction with -0.199 -0.058 0.038 0.052 -0.014 -0.074 -0.015 -0.018 freedom to choose life Dissatisfaction with 1.000 0.910 -0.015 0.076 -0.007 0.382 0.029 0.331 educational system/schools Corruption widespread -0.113 -0.387 0.016 0.111 -0.008 -0.088 -0.040 -0.262 within government/ business (Yes) 0.393 0.179 0.001 0.030 0.000 0.025 -0.012 0.004 Unemployed Working for the 0.339 0.481 0.038 0.034 0.013 0.022 0.020 0.043 government 0.001 -0.009 -0.600 -1.306 -0.015 0.010 -0.025 -0.077 Income (1,000’s) 39    Appendix D2: Changes in Averages and Decomposition of Effects, OLS Change in Change in Change in the obtained the obtained the averages Change in Developing Developing Arab Spring Arab Spring coefficient coefficient (2009-10) the averages MENAA MENA countries counties (4) (2009-10) (2009-10) DEV (2009-10) first-order second-order first-order second -order DEV MENA Arab Spring MENA Arab Spring effect effect effect effect Dissatisfaction with 0,064 0,072 0,028 0,091 -0,033 0,024 -0,106 0,023 Standards of living People cannot get ahead by -0,033 -0,027 -0,012 0,004 0,003 -0,006 -0,001 -0,003 working hard (Yes) Dissatisfaction with efforts -0,010 -0,024 -0,005 -0,007 -0,011 -0,015 of the government to 0,033 0,070 increase high quality jobs Dissatisfaction with freedom 0,050 0,265 0,038 0,052 -0,001 0,019 0,006 0,083 to choose life Dissatisfaction with -0,006 0,014 -0,015 0,076 0,002 -0,002 -0,012 0,005 Educational system/schools Corruption widespread 0,043 -0,317 0,016 0,111 -0,001 0,033 -0,027 -0,215 within government/business (Yes) 0,258 0,339 0,001 0,030 0,000 0,017 -0,009 0,008 Unemployed 0,298 0,542 0,038 0,034 0,013 0,019 0,020 0,049 Working for the government 0,001 -0,011 -0,600 -1,306 -0,014 0,010 -0,025 -0,094 Income (1,000’s) 40