Policy Research Working Paper 10691 The Long-Term and Gender-Equalizing Impacts of the Islamic Republic of Iran-Iraq War in 1980–88 on School Access and Labor Market Performance Trung Xuan Hoang Middle East and North Africa Region Office of the Chief Economist February 2024 Policy Research Working Paper 10691 Abstract This paper uses the context of the Islamic Republic of Iran- men and women. Furthermore, men who were exposed Iraq war in 1980–88 to study the long-term impacts of to the conflict were more likely to work in dangerous jobs exposure to the war during school years on educational or without air conditioning, while no evidence on this is attainment and labor market outcomes in Iraq. The analysis found for women. The paper also shows the impact of the uses an event study and the Iraq Household Socio-Eco- intensity of the Islamic Republic of Iran-Iraq war on educa- nomic Surveys 2006–2007. The findings show that the tional attainment and labor market outcomes. It documents Islamic Republic of Iran-Iraq conflict had a negative impact the education channel through which the war affects labor on the social welfare of men who were exposed to the war, market outcomes, showing that the war decreased the edu- including on social security, pension plan, health care, paid cational levels of men and women born between 1971 and leave, and job permanence, while little impact on women 1981. The findings are robust to a variety of robustness is found. Additionally, the conflict reduced wages for both checks and falsification tests. 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://www.worldbank.org/prwp. The author may be contacted at hoangxuantrung3012@gmail.com. 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 The Long-Term and Gender-Equalizing Impacts of the Islamic Republic of Iran–Iraq War in 1980–88 on School Access and Labor Market Performance1 Trung Xuan Hoang Thuongmai University JEL classification: J1; J3; O1; Keywords: Islamic Republic of Iran-Iraq war 1980-1988, education, event study, and Iraq. 1 This work was supported by the MENA Chief Economist Office under the labor and gender research programs. 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://www.worldbank.org/prwp. The author(s) may be contacted at hoangxuantrung3012@gmail.com. 1. Introduction Wars have serious consequences for a country, including the destruction of physical infrastructure, loss of lives, displacement of people, and reduced economic growth. Several papers have investigated the long-run effects of war from a macro perspective, particularly bombing. Studies examining the persistent impacts of U.S. bombing in Japan (Davis and Weinstein, 2002) and in Germany (Brakman et al., 2004) during World War II find no evidence of long-lasting effects on local population or economic performance. Similarly, Miguel and Roland (2011) show that the massive U.S. bombings in Vietnam during the Vietnam War from 1967 to 1975, which is cited as one of the most severe bombing incidents in human history, did not have long-term impacts on physical infrastructure, population density, literacy, and poverty level, following 27 years after the war ended. However, a growing body of research on the micro-level effects of armed conflicts reveals that armed wars inflict direct and external costs on survivors that continue longer and can be as harmful as the physical consequences. Most of the existing literature focuses on the long-term impacts of individuals who were exposed to war during early childhood on later life. Particularly, children that have been exposed to the conflict have significantly lower height-for-age in Ethiopia and Eritrea (Akresh et al. 2012) and in Burundi (Bundervoet et al, 2009). Gianmarco (2012) shows that the short-term effects of political violence on educational attainment are larger than those in the long-run in Peru. In addition to the effects of exposure to war on education, the existing literature also shows the negative effects of war on health. For instance, exposure to the war related to the German invasion of France during the first 5 years of life has adverse effects on health in adulthood in France (Allais et al. 2021). Individuals exposed to WWII destruction during pregnancy and the early postnatal period have higher BMIs and are more likely to be obese as adults (Yuksel, 2017). Also, Akresh et al. (2023) show that women who were exposed to the 1967-1970 Nigerian Civil War throughout their childhood had shorter adult stature, are more likely to be overweight or obese, and have their first child at a younger age, and men who have been exposed to war marry later and have fewer children. Singhal (2018) finds that early-life exposure to the war in Vietnam has adverse and long-term consequences on mental health in adulthood. The main purpose of this study is to investigate the long-term impact of the Islamic Republic of Iran–Iraq war in 1980-1988 on school-age children between the ages of 6 and 12 in Iraq, during which primary school is of importance. The Islamic Republic of Iran–Iraq war was a military conflict between the Islamic Republic of Iran and Iraq that lasted from September 1980 to August 2 1988. It was one of the bloodiest and longest battles that occurred in the Middle East in the 20th century. On September 22, 1980, Iraqi armed forces launched a full-scale invasion of neighboring Islamic Republic of Iran. The war was ended by a 1988 cease-fire. The conflict caused devastating effects on these two nations and their people, with total casualties of both sides estimated to range from 1 million to 2 million. To quantitatively evaluate the long-lasting effects of the war on children aged 6 to 12 during the war during the 1980-1988 period, this study uses the Iraq Household Socio-Economic Surveys (IHSES) 2006-2007 and the event study method. The treatment event cohort is 1968 because the children would reach the age of 12 in 1980 when the Islamic Republic of Iran–Iraq war started. The treatment cohorts include individuals born between 1969 and 1981. We find that the Islamic Republic of Iran–Iraq war has adverse impacts on labor market performance. The war decreases the social welfare of individuals who were exposed to the war during the school-age years. Specifically, the war decreases the probability of men accessing social security, pension plan, health care, paid leave and job permanence while we find little impact on women. However, we find significant and negative impacts of the war on wages for both men and women. Furthermore men, who were exposed to the war during their school-age years are more likely to work in a hazardous job and without air conditioning; we find no evidence for the women. We document the education channel through which the war affects labor market outcomes. We expect that the war would decrease educational attainment, leading to negative impacts on labor market outcomes. Indeed, the war decreases the educational level of the cohorts of 1971-1981 for both men and women. The Islamic Republic of Iran–Iraq conflict had a greater impact on female education than on male education. The illiterate rate of women increases due to the conflict while we find no evidence of the impact of the war on the illiteracy of men. Our findings are robust to a variety of robustness checks and falsification tests. Our study makes several contributions to the nascent literature on the long-term impacts of war on school-age children. First, our study focuses on Iraq. Existing studies on the effects of war on labor market outcomes and educational attainment in the Middle East and North Africa (MENA) region are rare. This study will fill in this gap by considering the impact of the Islamic Republic of Iran– Iraq war. Iraq presents a remarkable study case. Poverty remains one of Iraq's most pressing social issues, the poverty rate was 22.9 percent in 2007, this rate is still high, and exceeds 40 percent in some governorates, meanwhile 11.3 percent of girls and 5.4 percent of boys aged between 6 and 11 years old did not participate in primary school in 2013 (UNICEF, 2014). Second, most of the existing studies focus on the early exposure to the war. In other words, these studies examine the long-term impacts of war on children who were in utero and in their early childhood years during the war. Little is known about the long-term effects of war on children, who were exposed to the war during school-age years of 6 to 12. Therefore, our study sharpens understanding of the long- term impacts of war on school-age children in their later adulthood. Third, we provide heterogeneous analysis. Specifically, we examine the long-term impacts of the war on male and female labor market outcomes, and male and female educational attainment. Iraqi women now face insufficient educational and health care possibility, limited labor market access, as well as high levels of abuse and discrimination. These situations are frequently worsened by cultural and 3 societal norms, incorrect beliefs and a lack of awareness of women’s rights, and institutional and legal restrictions (Vilardo and Bittar, 2018). Only a handful of studies currently examine the long-term impacts of war on school-age children. We review the most pertinent studies. Olga (2010) examines the impacts of war on school-age children and shows that girls at school age who were exposed to the 1992-1998 armed conflict had lower probability of completing their mandatory schooling than those who were not exposed to the conflict in Tajikistan, but there is no effect of the conflict on education of boys. Similarly, Ichino and Webmer (2004) show that children who were ten years old during World War II received less education and experienced a sizable earnings loss some 40 years later. Blattman and Annan (2010) find adverse impacts on children-abductees who were between 11 and 24 years old during the years they spent with the rebel forces on schooling, skilled employment and earnings in Uganda. 2. Data This study uses the Iraq Household Socio-Economic Surveys (IHSES) 2006-2007. IHSES 2006- 2007 surveyed 18,144 households, including all three layers of the governorates: one layer for the urban area of the center of the governorate, one layer for the rest of the urban areas, and one layer for the rural areas. This survey served as the foundation for revising the Consumer Price Index (CPI) from an out-of-date 1990 one to a revised 2007 index. IHSES 2006-2007 is nationally representative and covers socio-economic data on migration, labor, education, health, anthropometrics, agriculture, nonfarm, expenditure and income. This survey was carried out by the Central Bureau of Statistics and the Kurdistan Regional Statistics Office, in the field for a year, between October 2006 and November 2007. It covered all 18 governorates in Iraq. 3. Methodology The empirical analysis relies on the variation in the timing of when children finished primary school and the war started. Children who finished primary school right after the onset of the Islamic Republic of Iran–Iraq war have little or no impact of the war compared with those children who had not completed primary school at the timing of the war. Children are more likely to drop out of school if they skip the critical years of primary school education. According to the standard human capital theory (Becker, 1993), enrolling a child in school as early as possible is optimal, as this is when the opportunity cost of the child's time is lowest. Psychological research also indicates that enrolling children in school at an earlier age can be beneficial, as brain development related to cognitive skills becomes increasingly less sensitive with age after the age of 5-6 (Shonkoff and Phillips, 2000). Indeed, Chen (2015) shows that delaying enrollment in school for one year results in a 10 percentage point increase in first-grade retention among boys and a 6 percentage point reduction in the likelihood of middle school enrollment for both boys and girls in rural China. Also, Wils (2004) finds that late school enrollment is associated with higher school dropout rates in Mozambique. Similarly, No et al. (2016) indicate that late school entry of students in grades 1– 4 is one of the causes of school dropouts in Cambodia. According to IHSES 2006-2007, the average percentage of completing primary education or above for cohorts of 1963-1969, who finished primary school during the Islamic Republic of Iran–Iraq war, was 88%. In contrast, 4 cohorts of 1974-1979 who did not finish primary school at the time of the war had a completion rate of 81%. The survey also provides statistical information about the completion rates of secondary education or higher for two different cohorts, one that finished primary school during the war (1963-1969) and the other that did not complete primary school during the same time (1974-1979). The first cohort had a higher completion rate of 61%, while the second cohort had a lower completion rate of 49%. We will exploit these significant variations in our identification strategy. Specifically, we conduct an event study along the lines of Jacobson et al. (1993) for the interest outcomes. = 1 + + ∑13 ≥−4,≠0 + 2 + where are the outcome variables including educational attainment, social welfare, and working conditions of an individual i in governorate p at time t. The dummy variables, , jointly represent the treatment event. In particular, we select the year 1968 as the treatment event because the children would be 12 years old in 1980. Those individuals born before 1968 were not affected by Islamic Republic of Iran–Iraq war, and the extent of the impact of the war would increase for those born after 1968 until 1980. We define =1 if t - 1968 = k, and 0 otherwise, k=-4, -3, -2, - 1, 0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13. And, is a set of dummy variables for each governorate. Xipt is a dummy variable for gender. All standard errors are clustered at the governorate level. 4. Long-run impact of the Islamic Republic of Iran–Iraq war 4.1 Long-run impact of the Islamic Republic of Iran–Iraq war on labor market outcomes Table 1 provides the impact of the Islamic Republic of Iran–Iraq war 1980-1988 on the social welfare of workers. This time, we estimate the long-term effect of the war on workers. We use different measures to reflect the social welfare of workers, they are a dummy variable of having social security, a dummy variable for having pension plan, a dummy variable for health care, a dummy variable for paid leave, a dummy variable for job permanence, and log of wage. The findings show that cohorts of 1970-1980, who were exposed to the war, have lower probability of having social security, the results are not statistically significant for cohorts of 1974 and 1975 (Column 1). Also, most of cohorts of 1970-1980 are less likely to have pension plan, health care, paid leave, and job permanence (Columns 2-5). Importantly, workers of cohorts of 1969-1981 who exposed to the war have lower wage. Those findings suggest that the Islamic Republic of Iran– Iraq war had long-term and negative effects on social welfare of workers who were exposed at ages 6-11 years. Table 1. Impact of Islamic Republic of Iran–Iraq war on Social welfare of workers Dummy for Dummy for Dummy for Dummy for Dummy for job Log of social security pension plan health care paid leave permanence wage (1) (2) (3) (4) (5) (6) Year 1964 -0.014 -0.000 -0.030 0.017 0.020 0.117 (0.041) (0.039) (0.030) (0.042) (0.022) (0.133) Year 1965 -0.008 -0.002 -0.012 0.011 -0.030 -0.072 5 (0.034) (0.034) (0.030) (0.032) (0.023) (0.122) Year 1966 -0.018 -0.007 -0.008 0.019 -0.002 -0.052 (0.025) (0.026) (0.029) (0.021) (0.019) (0.115) Year 1967 -0.044 -0.038 -0.054* -0.004 -0.004 -0.098 (0.035) (0.033) (0.029) (0.043) (0.020) (0.111) Year 1969 (Year of -0.057 -0.061 -0.074*** -0.061 -0.019 -0.199* change) (0.036) (0.039) (0.025) (0.038) (0.022) (0.112) Year 1970 -0.060* -0.062* -0.053* -0.039 -0.021 -0.221** (0.029) (0.032) (0.026) (0.029) (0.016) (0.083) Year 1971 -0.043 -0.043 -0.065*** -0.032 0.002 -0.157 (0.028) (0.029) (0.022) (0.027) (0.017) (0.110) Year 1972 -0.065** -0.065* -0.042** -0.048 -0.017 -0.259*** (0.028) (0.031) (0.019) (0.033) (0.014) (0.079) Year 1973 -0.092*** -0.095*** -0.052* -0.072** -0.028* -0.254** (0.029) (0.030) (0.028) (0.033) (0.015) (0.099) Year 1974 -0.032 -0.037 -0.042 -0.016 0.005 -0.100 (0.036) (0.039) (0.025) (0.038) (0.019) (0.119) Year 1975 -0.039 -0.056* -0.039 -0.023 -0.011 -0.160 (0.031) (0.030) (0.033) (0.029) (0.015) (0.107) Year 1976 -0.063* -0.053 -0.043 -0.027 -0.006 -0.200** (0.036) (0.037) (0.028) (0.032) (0.017) (0.093) Year 1977 -0.066* -0.071* -0.049** -0.039 -0.021 -0.355*** (0.032) (0.039) (0.019) (0.027) (0.019) (0.086) Year 1978 -0.069* -0.062 -0.048* -0.044 0.003 -0.339*** (0.036) (0.038) (0.025) (0.035) (0.019) (0.104) Year 1979 -0.093*** -0.092*** -0.051** -0.067** -0.048** -0.439*** (0.028) (0.030) (0.024) (0.026) (0.019) (0.093) Year 1980 -0.092*** -0.097*** -0.074*** -0.063* -0.030 -0.389*** (0.029) (0.031) (0.024) (0.031) (0.023) (0.085) Year 1981 -0.128*** -0.130*** -0.059** -0.081** -0.041** -0.510*** (0.029) (0.032) (0.022) (0.032) (0.018) (0.102) N 10538 10538 10538 10538 10538 10238 adj. R2 0.139 0.151 0.124 0.139 0.037 0.074 Notes: All regressions add birth governorate fixed-effects, household size and gender. Standard errors are clustered at the birth governorate level. *** p<0.01, ** p<0.05, * p<0.1. To measure the working conditions of workers, we use a dummy variable for a worker without a hazardous job, 2 a dummy variable for a worker with air conditioning at the workplace and a dummy variable for health impact. A dummy variable for health impact is defined when the health of workers was adversely affected by one of the following factors: soil and gases, chemicals and radiation, heat, humidity, noise, lighting, insects, and fatigue. The results of Table 2 indicate that most of the cohorts of 1971-1980 work in hazardous conditions. The cohorts of 1969-1981 work without air conditioning. And working conditions had adverse impacts on the health of most of the cohorts of 1970-1981. 2 The question is asked of interviewees in the questionnaires as follows: What is your evaluation of the degree of danger at work? And the responses include: very dangerous, dangerous, relatively, not dangerous. If an interviewee answers “not dangerous”, the dummy variable for a worker without hazardous jobs would be equal to 1, 0 otherwise. 6 Table 2. Working conditions of workers Hazardous job Air conditioning Health impact (1) (2) (3) Year 1964 0.014 0.000 0.129 (0.043) (0.035) (0.187) Year 1965 -0.020 -0.028 -0.090 (0.026) (0.027) (0.131) Year 1966 0.002 -0.000 0.157 (0.026) (0.033) (0.171) Year 1967 0.011 -0.031 0.069 (0.032) (0.034) (0.121) Year 1969 (Year of change) -0.019 -0.066* 0.086 (0.030) (0.032) (0.119) Year 1970 -0.038 -0.083** 0.200* (0.036) (0.038) (0.097) Year 1971 -0.058* -0.037* 0.069 (0.030) (0.019) (0.115) Year 1972 -0.056* -0.084** 0.275** (0.028) (0.030) (0.122) Year 1973 -0.103*** -0.079*** 0.085 (0.032) (0.024) (0.118) Year 1974 -0.098*** -0.054* 0.324** (0.029) (0.030) (0.127) Year 1975 -0.053 -0.077*** 0.081 (0.032) (0.024) (0.109) Year 1976 -0.050 -0.032 0.168 (0.037) (0.026) (0.120) Year 1977 -0.055* -0.057* 0.255* (0.031) (0.031) (0.130) Year 1978 -0.034 -0.077** 0.265* (0.022) (0.033) (0.135) Year 1979 -0.036 -0.085*** 0.382*** (0.031) (0.026) (0.112) Year 1980 -0.066** -0.074** 0.188 (0.024) (0.026) (0.120) Year 1981 -0.070** -0.119*** 0.364*** (0.027) (0.030) (0.102) N 10538 10538 10538 adj. R2 0.125 0.087 0.172 Notes: All regressions add birth governorate fixed-effects, household size and gender. Standard errors are clustered at the birth governorate level. *** p<0.01, ** p<0.05, * p<0.1. 4.2 Channels Education is critical for preparing people to join the labor market and providing them with the skills they need to engage in lifelong learning experiences. Vast research literature provides evidence of the value of investing in education to develop human capital and its contribution to increasing individual earnings (Ashenfelter and Krueger, 1994; Card, 2001; Duflo, 2001; Harmon, 2003; Oreopoulos, 2006; Fiszbein et al. 2007; Clark and Martorell, 2014; Peet et al. 2015; Silliman and Virtanen, 2022). Besides a robust literature examining the impact of schooling on earnings, a large body of literature documents the relationship between educational attainment and labor 7 market outcomes. Oreopoulos and Salvanes (2011) show that individuals with better levels of education have higher work satisfaction, autonomy, occupational prestige, and feelings of achievement, they also have lower unemployment rates. Crystal at al. (1992) find that educational attainment has positive impacts on private pensions and other retirement income sources. Oreopoulos (2007) shows negative effects of schooling on the odds of unemployment. Those findings suggest that education may be a channel through which the Islamic Republic of Iran–Iraq war affects the labor market outcomes of workers in Iraq. Table 3 presents the estimation results of equation (1) for educational attainment. We use a variety of measures of educational attainment, they are a dummy variable for literacy, a dummy variable for primary education completion or above, a dummy variable for secondary education completion or above and years of education. Without controlling for birth governorate fixed-effects, household size and gender, Columns 1- 4 of Table 1 show that the effect on literacy is negative and statistically significant for cohorts of 1971-1981, except for cohort 1974. The results are also negative for cohorts of 1969 and 1970 for literacy, however the results are not statistically significant. The results on primary education completion or above are highly statistically significant at the 1 percent level for cohorts of 1971-1981. For example, children of cohorts of 1971-1981 were between 3.9 and 7.8 percent lower in the probability of completing primary education or above than those of cohort of 1968. We also find that the children of cohorts of 1971- 1981 have lower probability of completing secondary education or above compared with child of cohort of 1968. Similarly, years of education decreased for the cohorts of 1972-1981. The results are quantitatively similar when we control for birth governorate, household size and gender in Columns 5-8. Educational attainment does not show any clear trend in the cohorts before the cohort of 1968. The educational attainment of cohorts of 1964-1967 is statistically significant but positive, or statistically insignificant. Table 3. Impact of Islamic Republic of Iran–Iraq war on educational attainment Literacy Primary Secondary Years of Literacy Primary Secondary Years of education education education education education education completion completion completion completion or above or above or above or above (1) (2) (3) (4) (5) (6) (7) (8) Year 1964 0.037*** 0.024 0.100*** 0.225 0.037*** 0.023 0.100*** 0.107 (0.010) (0.017) (0.019) (0.155) (0.010) (0.017) (0.020) (0.152) Year 1965 0.015** 0.017 0.088*** 0.195 0.015** 0.015 0.086*** 0.117 (0.007) (0.016) (0.022) (0.186) (0.007) (0.015) (0.022) (0.172) Year 1966 0.032*** 0.012 0.072*** 0.128 0.032*** 0.015 0.074*** 0.087 (0.009) (0.014) (0.022) (0.121) (0.009) (0.013) (0.021) (0.130) Year 1967 0.007 -0.002 0.052** 0.193 0.007 -0.001 0.054** 0.187 (0.005) (0.013) (0.021) (0.178) (0.005) (0.014) (0.021) (0.174) Year 1969 -0.004 -0.002 0.013 -0.035 -0.004 0.000 0.013 0.005 (Year of change) (0.006) (0.010) (0.017) (0.127) (0.006) (0.010) (0.017) (0.139) Year 1970 -0.004 -0.027 -0.023 -0.223 -0.004 -0.026 -0.029 -0.181 (0.005) (0.017) (0.021) (0.134) (0.004) (0.015) (0.021) (0.127) Year 1971 -0.013** -0.048*** -0.044* -0.247 -0.013** -0.041*** -0.040* -0.121 (0.005) (0.012) (0.023) (0.145) (0.005) (0.011) (0.023) (0.136) 8 Year 1972 -0.011* -0.057*** -0.058*** -0.477*** -0.011* -0.051*** -0.057*** -0.354** (0.006) (0.018) (0.018) (0.154) (0.006) (0.017) (0.017) (0.131) Year 1973 -0.010* -0.064*** -0.059*** -0.570*** -0.011** -0.059*** -0.057*** -0.439*** (0.005) (0.017) (0.014) (0.105) (0.005) (0.017) (0.015) (0.108) Year 1974 -0.007 -0.059*** -0.054*** -0.483*** -0.007 -0.057*** -0.053*** -0.398*** (0.006) (0.011) (0.015) (0.122) (0.006) (0.011) (0.015) (0.117) Year 1975 -0.013** -0.039*** -0.015 -0.184 -0.013** -0.037** -0.017 -0.134 (0.005) (0.013) (0.021) (0.179) (0.005) (0.013) (0.021) (0.179) Year 1976 -0.011** -0.053*** -0.056*** -0.322** -0.011** -0.051*** -0.055** -0.247* (0.005) (0.014) (0.019) (0.141) (0.005) (0.014) (0.020) (0.139) Year 1977 -0.011** -0.050*** -0.035** -0.228* -0.011** -0.048*** -0.031** -0.135 (0.005) (0.012) (0.015) (0.117) (0.005) (0.012) (0.014) (0.117) Year 1978 -0.012** -0.070*** -0.081*** -0.548*** -0.012** -0.067*** -0.078*** -0.434** (0.006) (0.017) (0.016) (0.177) (0.006) (0.016) (0.017) (0.174) Year 1979 -0.013** -0.059*** -0.058*** -0.274* -0.014*** -0.057*** -0.056*** -0.140 (0.005) (0.012) (0.016) (0.131) (0.005) (0.011) (0.017) (0.123) Year 1980 -0.014*** -0.074*** -0.050** -0.344* -0.015*** -0.070*** -0.046** -0.189 (0.004) (0.011) (0.018) (0.195) (0.004) (0.010) (0.019) (0.197) Year 1981 -0.018*** -0.078*** -0.088*** -0.564*** -0.019*** -0.068*** -0.080*** -0.341** (0.004) (0.012) (0.020) (0.124) (0.004) (0.011) (0.021) (0.144) N 29050 29050 29050 25124 29050 29050 29050 25124 adj. R2 0.010 0.006 0.010 0.003 0.020 0.038 0.023 0.078 Notes: Columns 5-8 add birth governorate fixed-effects, household size and gender. Standard errors are clustered at the birth governorate level. *** p<0.01, ** p<0.05, * p<0.1. 4.3. Robustness checks Our estimation results may be affected by potential channels; therefore, we offer a battery of robustness checks. We expect that individuals who were born in other countries would not be affected by the Islamic Republic of Iran–Iraq war. Those individuals offer us an important falsification test for the impact of the Islamic Republic of Iran–Iraq war, where our hypothesis is that there is no effect on individuals born abroad. We do not have information when the individuals returned to Iraq and those born abroad could return to Iraq to study at the primary school during the Islamic Republic of Iran–Iraq war. However, children born abroad had higher opportunity to send back to foreign countries if they returned to Iraq during the Islamic Republic of Iran–Iraq war. This suggests that those children would not be affected by the Islamic Republic of Iran–Iraq war. Indeed, Columns 1-4 of Table 4 show that cohorts of 1969-1981 are not statistically significant for different measures of educational attainment. The results are similar when we control for birth governorate fixed-effects, household size and gender. The findings generally support our hypothesis that it is the Islamic Republic of Iran–Iraq war 1980-1988 – rather than other war- related factors (e.g., specific political regime) or unobserved region trends – that led to worse long- term educational outcomes for those who were exposed to the war during school age. Table 4. Robustness checks Literacy Primary Secondary Years of Literacy Primary Secondary Years of education education education education education education completion completion completion completion or above or above or above or above (1) (2) (3) (4) (5) (6) (7) (8) 9 Year 1964 0.333** 0.250 0.250 8.500* 0.331** 0.324 0.358 9.465** (0.160) (0.332) (0.386) (4.513) (0.163) (0.325) (0.366) (4.303) Year 1965 0.000 0.250 0.250 0.000 0.001 0.275 0.285 0.000 (0.234) (0.486) (0.564) (.) (0.237) (0.472) (0.532) (.) Year 1966 0.000 -0.083 -0.083 1.000 -0.010 0.024 0.075 0.454 (0.160) (0.332) (0.386) (3.685) (0.163) (0.325) (0.367) (3.517) Year 1967 0.000 0.000 -0.000 -5.500 0.012 -0.156 -0.230 -5.164 (0.148) (0.307) (0.357) (4.513) (0.152) (0.303) (0.342) (4.295) Year 1969 0.000 0.250 -0.083 1.000 -0.010 0.363 0.084 0.475 (Year of change) (0.160) (0.332) (0.386) (3.685) (0.163) (0.325) (0.367) (3.495) Year 1970 0.000 0.250 0.250 3.000 -0.011 0.392 0.459 3.420 (0.181) (0.377) (0.437) (3.685) (0.185) (0.368) (0.416) (3.493) Year 1971 0.000 -0.750 -0.750 -2.500 -0.007 -0.596 -0.524 -1.325 (0.234) (0.486) (0.564) (4.513) (0.239) (0.476) (0.537) (4.308) Year 1972 0.250* 0.250 -0.000 3.833 0.244 0.313 0.094 3.386 (0.148) (0.307) (0.357) (3.364) (0.150) (0.299) (0.337) (3.193) Year 1973 0.000 -0.250 -0.375 -0.500 -0.006 -0.199 -0.299 -1.074 (0.128) (0.266) (0.309) (3.009) (0.130) (0.260) (0.294) (2.867) Year 1974 0.000 0.250 -0.250 2.250 -0.005 0.261 -0.232 0.876 (0.148) (0.307) (0.357) (3.191) (0.151) (0.301) (0.340) (3.059) Year 1975 0.000 -0.062 -0.313 1.100 -0.006 0.006 -0.212 0.644 (0.117) (0.243) (0.282) (2.774) (0.119) (0.237) (0.268) (2.632) Year 1976 0.250* -0.250 -0.375 -2.000 0.245* -0.184 -0.277 -2.343 (0.128) (0.266) (0.309) (3.009) (0.130) (0.259) (0.292) (2.856) Year 1977 0.000 0.250 -0.000 2.750 -0.002 0.282 0.048 2.236 (0.148) (0.307) (0.357) (3.191) (0.149) (0.298) (0.336) (3.036) Year 1978 0.071 0.107 -0.179 2.577 0.064 0.199 -0.043 2.490 (0.119) (0.247) (0.286) (2.799) (0.121) (0.241) (0.272) (2.652) Year 1979 0.000 0.050 -0.050 3.389 -0.009 0.170 0.127 3.613 (0.124) (0.257) (0.299) (2.880) (0.127) (0.253) (0.285) (2.730) Year 1980 0.000 -0.306 -0.417 -0.611 -0.002 -0.285 -0.386 -1.362 (0.126) (0.261) (0.303) (2.880) (0.127) (0.253) (0.286) (2.744) Year 1981 0.000 0.000 -0.125 -0.333 -0.007 0.084 -0.001 -0.830 (0.128) (0.266) (0.309) (3.009) (0.130) (0.260) (0.293) (2.857) N 106 106 106 86 106 106 106 86 adj. R2 0.037 0.036 -0.028 0.101 0.017 0.096 0.089 0.193 Notes: Columns 5-8 add birth governorate fixed-effects, household size and gender. Standard errors are clustered at the birth governorate level. *** p<0.01, ** p<0.05, * p<0.1. There may also be concern that the estimation results may be affected by other factors that are not related to the war including, for example, whether the stated cohorts have disabilities, diseases, or illnesses. To test this hypothesis, we re-run the equation (1) for a variety of measures of health, they are dummy variable for illness or injury in the past 30 days, dummy variable for chronic disease and dummy variable for disability. Table 5 provides the estimation results, which indicate that cohorts of 1969-1981 are not statistically significant for dummy variable for illness or injury in the past 30 days, even when we add birth governorate fixed-effects, household size and gender (Columns 1 and 4). The results are statistically significant for the cohorts of 1972-1981 for dummy variable for chronic disease in both samples with and without birth governorate fixed-effects, household size and gender, however the signs are negative and unexpected (Columns 2 and 5). 10 Similarly, cohorts of 1971 and 1975-1981 are statistically significant for dummy variable for disability, but the signs are negative and unexpected (Column 3), the result is statistically insignificant when we control for birth governorate fixed-effects, household size and gender. Those findings suggest that exposure to the war during early childhood does not affect our estimation results. Table 5. Robustness checks (continue) Dummy Dummy Dummy Dummy Dummy Dummy variable variable for variable variable for variable for variable for for illness chronic for illness or chronic disability or injury in disease disability injury in disease the past 30 the past 30 days days (1) (2) (3) (4) (5) (6) Year 1964 -0.007 0.060*** -0.004 -0.021 0.069** 0.010 (0.016) (0.020) (0.018) (0.019) (0.027) (0.009) Year 1965 -0.015 0.018 -0.002 -0.004 0.019 0.001 (0.012) (0.021) (0.016) (0.017) (0.022) (0.007) Year 1966 -0.002 0.051** 0.002 -0.038** 0.025 0.009 (0.016) (0.023) (0.015) (0.017) (0.029) (0.008) Year 1967 -0.010 0.030 -0.007 -0.029 0.007 -0.008 (0.018) (0.021) (0.013) (0.019) (0.024) (0.005) Year 1969 (Year of change) 0.003 -0.040 -0.014 -0.015 -0.015 -0.004 (0.018) (0.024) (0.011) (0.014) (0.017) (0.007) Year 1970 0.004 -0.016 -0.012 -0.014 -0.020 -0.002 (0.016) (0.023) (0.012) (0.022) (0.019) (0.006) Year 1971 -0.005 -0.016 -0.030** -0.016 -0.013 -0.004 (0.016) (0.020) (0.012) (0.020) (0.023) (0.008) Year 1972 -0.006 -0.042** -0.023 -0.024 -0.032 -0.004 (0.011) (0.019) (0.014) (0.015) (0.021) (0.006) Year 1973 -0.011 -0.061*** -0.017 -0.012 -0.064*** 0.001 (0.015) (0.018) (0.011) (0.019) (0.018) (0.007) Year 1974 0.006 -0.062*** -0.022 0.003 -0.075*** -0.006 (0.012) (0.019) (0.018) (0.013) (0.021) (0.006) Year 1975 0.007 -0.042** -0.027** -0.011 -0.056** -0.000 (0.013) (0.018) (0.013) (0.015) (0.020) (0.010) Year 1976 0.018 -0.066*** -0.024* 0.004 -0.081*** -0.003 (0.014) (0.016) (0.013) (0.016) (0.022) (0.006) Year 1977 0.010 -0.074*** -0.029** -0.012 -0.083*** -0.001 (0.012) (0.016) (0.014) (0.015) (0.017) (0.006) Year 1978 0.019 -0.084*** -0.028** -0.005 -0.097*** -0.010 (0.012) (0.019) (0.012) (0.012) (0.020) (0.006) Year 1979 0.019 -0.082*** -0.039** -0.005 -0.090*** -0.007 (0.015) (0.016) (0.015) (0.015) (0.020) (0.005) Year 1980 0.024 -0.090*** -0.037*** 0.004 -0.090*** 0.001 (0.015) (0.016) (0.012) (0.013) (0.020) (0.006) Year 1981 0.007 -0.095*** -0.040*** 0.001 -0.106*** -0.009 (0.017) (0.016) (0.012) (0.016) (0.018) (0.006) N 14367 14384 14384 14655 14666 14666 adj. R2 0.043 0.043 0.004 0.057 0.055 0.001 11 Notes: Columns 4-6 add birth governorate fixed-effects, household size and gender. Standard errors are clustered at the birth governorate level. *** p<0.01, ** p<0.05, * p<0.1. 4.4. Heterogeneity analysis Female versus male In this section, we further examine the heterogeneous effects of the Islamic Republic of Iran–Iraq war on gender. Table 6 provides the estimation results on male and female educational attainment. Column 1 shows no evidence of the impact of the Islamic Republic of Iran–Iraq war on literacy for men. The result is negative and statistically significant for male primary education completion or above for cohorts of 1972-1981 (Column 2). Also, secondary education completion or above and years of education decreased for men for the cohort of 1972-1981. In terms of the impact of war on female education, we find that literacy, primary education completion or above, secondary education completion or above decreased for women, however, the result is not statistically significant for years of education (Columns 5-8). We find that the magnitude of the impact of the Islamic Republic of Iran–Iraq war is larger for female education than for male education, for example, the Islamic Republic of Iran–Iraq war decreased the probability of male cohort of 1972 completing primary education or above by 3.1% while this figure is 7.1% female cohort of 1972. Overall, female education was more severely affected by the Islamic Republic of Iran–Iraq war than male education. Table 6. Educational attainment between males and females Male Female Literacy Primary Secondary Years of Literacy Primary Secondary Years of education education education education education education completion completion completion completion or above or above or above or above (1) (2) (3) (4) (5) (6) (7) (8) Year 1964 0.035*** 0.027 0.063** -0.036 0.038** 0.019 0.131*** 0.274 (0.011) (0.021) (0.029) (0.194) (0.015) (0.028) (0.039) (0.319) Year 1965 0.011 0.024 0.090*** 0.131 0.018 0.005 0.081** 0.082 (0.006) (0.019) (0.027) (0.244) (0.013) (0.021) (0.029) (0.234) Year 1966 0.016** 0.022 0.064 0.061 0.047*** 0.009 0.087*** 0.117 (0.008) (0.014) (0.037) (0.219) (0.015) (0.018) (0.024) (0.141) Year 1967 0.007* -0.000 0.060** 0.150 0.009 -0.000 0.048 0.198 (0.004) (0.023) (0.026) (0.272) (0.009) (0.024) (0.029) (0.247) Year 1969 (Year 0.006 0.004 0.021 -0.169 -0.014 -0.004 0.006 0.226 of change) (0.005) (0.018) (0.026) (0.208) (0.013) (0.015) (0.025) (0.213) Year 1970 0.003 -0.005 -0.025 -0.391 -0.011 -0.046** -0.029 0.085 (0.004) (0.017) (0.030) (0.238) (0.007) (0.019) (0.025) (0.171) Year 1971 0.003 0.001 -0.014 -0.288 -0.029** -0.084*** -0.066** 0.082 (0.004) (0.018) (0.028) (0.199) (0.010) (0.020) (0.025) (0.176) Year 1972 0.004 -0.031* -0.037 -0.473** -0.026** -0.071** -0.077*** -0.193 (0.004) (0.018) (0.023) (0.205) (0.010) (0.026) (0.026) (0.161) Year 1973 -0.002 -0.053*** -0.053** -0.551*** -0.019* -0.065** -0.061** -0.315 12 (0.002) (0.017) (0.024) (0.186) (0.010) (0.028) (0.025) (0.193) Year 1974 0.003 -0.056*** -0.033 -0.465** -0.016 -0.058*** -0.073*** -0.335 (0.005) (0.019) (0.025) (0.191) (0.010) (0.018) (0.021) (0.195) Year 1975 0.001 -0.032** -0.026 -0.453* -0.027*** -0.043** -0.009 0.236 (0.002) (0.012) (0.032) (0.246) (0.009) (0.020) (0.023) (0.237) Year 1976 0.003 -0.055** -0.055* -0.435* -0.024*** -0.047** -0.054** -0.028 (0.003) (0.023) (0.027) (0.234) (0.008) (0.020) (0.024) (0.203) Year 1977 0.004 -0.052*** -0.018 -0.322 -0.025** -0.045*** -0.044* 0.080 (0.003) (0.015) (0.029) (0.212) (0.009) (0.015) (0.022) (0.252) Year 1978 0.003 -0.078*** -0.073** -0.704*** -0.027** -0.056*** -0.083*** -0.151 (0.003) (0.021) (0.028) (0.234) (0.010) (0.019) (0.022) (0.209) Year 1979 0.002 -0.066*** -0.049** -0.453*** -0.029*** -0.047** -0.063** 0.230 (0.003) (0.016) (0.021) (0.152) (0.008) (0.019) (0.023) (0.191) Year 1980 -0.001 -0.087*** -0.032 -0.449 -0.027*** -0.054*** -0.058*** 0.131 (0.003) (0.018) (0.029) (0.263) (0.008) (0.013) (0.020) (0.240) Year 1981 0.002 -0.080*** -0.054* -0.536** -0.038*** -0.057*** -0.105*** -0.102 (0.003) (0.018) (0.029) (0.221) (0.009) (0.016) (0.023) (0.225) N 14384 14384 14384 13452 14666 14666 14666 11672 adj. R2 0.013 0.045 0.029 0.042 0.023 0.032 0.031 0.062 Notes: All regressions add birth governorate fixed-effects, household size and gender. Standard errors are clustered at the birth governorate level. *** p<0.01, ** p<0.05, * p<0.1. Now, we consider the impact of the Islamic Republic of Iran–Iraq war on female and male labor market outcomes. The results show that most of the male cohorts of 1970-1981 are statistically significant for dummy variables for social security or pension plan, however, the female cohorts of 1970-1981 are not statistically significant. Similarly, we find that the Islamic Republic of Iran– Iraq war decreased the probability of men having health care, paid leave and job permanence for most of cohorts of 1969-1981. While we find little evidence of the impact of the Islamic Republic of Iran–Iraq war on women for the cohorts of 1969-1981. In terms of wage, the results are negative and highly statistically significant for most of the male cohorts of 1969-1981. The Islamic Republic of Iran–Iraq war also has negative impacts on female wages for the cohorts of 1973 - 1981. Female labor force participation in Iraq is limited due to lack of government backing and social norms. Women's labor force participation in Iraq is low: only 12.3% of working-age women in Iraq were employed or searching for employment in 2018 (Reach, 2019). Table 7 also shows that only 1,715 women participated in the labor market while 8,823 men did in the labor market. In other hand, women participating in the labor market accounted for only 16.3 percent in 2006 in Iraq. Furthermore, of those workers who were exposed to the Islamic Republic of Iran–Iraq war during childhood, about 72 percent of workers with social security, pension, or paid leave were male, 74 percent of workers with health care were male, and 85 percent of workers with job permanence were male in 2006. This may suggest why the effect of the Islamic Republic of Iran– Iraq war on female labor market outcomes is smaller compared with male market outcomes. 13 Table 7. Social welfare of workers between males and females Male Female Dummy Dummy Dummy Dummy Dummy Log of Dummy Dummy Dummy Dummy Dummy Log of variable variable variable for variable variable for wage variable variable variable variable variable for wage for social for health care for paid job for social for for health for paid job security pension leave permanence security pension care leave permanence plan plan (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) Year 1964 -0.010 0.003 -0.011 0.028 0.029 0.120 0.014 0.037 -0.098* 0.028 0.017 0.156* (0.049) (0.047) (0.035) (0.051) (0.027) (0.165) (0.031) (0.025) (0.049) (0.037) (0.030) (0.075) Year 1965 -0.021 -0.013 0.002 -0.003 -0.040 -0.106 0.058 0.058* -0.083* 0.079* 0.029 0.064 (0.040) (0.040) (0.036) (0.036) (0.027) (0.145) (0.038) (0.030) (0.045) (0.042) (0.025) (0.076) Year 1966 -0.010 0.004 0.023 0.030 -0.007 -0.016 -0.022 -0.024 -0.131 -0.001 0.036 -0.126 (0.031) (0.034) (0.028) (0.028) (0.023) (0.140) (0.034) (0.040) (0.084) (0.041) (0.033) (0.121) Year 1967 -0.048 -0.041 -0.039 -0.008 -0.013 -0.081 -0.027 -0.023 -0.142** 0.017 0.040 -0.167** (0.041) (0.039) (0.034) (0.047) (0.022) (0.128) (0.048) (0.032) (0.052) (0.049) (0.043) (0.062) Year 1969 -0.069 -0.073 -0.078** -0.083* -0.025 -0.224* -0.008 -0.009 -0.090* 0.035 0.012 -0.061 (Year of change) (0.040) (0.044) (0.032) (0.044) (0.025) (0.127) (0.045) (0.044) (0.050) (0.053) (0.034) (0.105) Year 1970 -0.066* -0.065 -0.039 -0.056 -0.036* -0.216** -0.008 -0.029 -0.120** 0.053 0.047 -0.180* (0.037) (0.039) (0.030) (0.036) (0.020) (0.095) (0.044) (0.047) (0.051) (0.054) (0.036) (0.091) Year 1971 -0.048 -0.049 -0.048* -0.042 -0.002 -0.164 -0.006 0.005 -0.136* 0.029 0.031 -0.107 (0.034) (0.035) (0.026) (0.033) (0.019) (0.131) (0.051) (0.042) (0.065) (0.052) (0.036) (0.078) Year 1972 -0.085** -0.085** -0.040 -0.072* -0.032* -0.295*** 0.041 0.047 -0.053 0.078* 0.059** -0.062 (0.032) (0.035) (0.025) (0.038) (0.017) (0.090) (0.037) (0.028) (0.039) (0.039) (0.025) (0.069) Year 1973 -0.113*** -0.116*** -0.045 -0.094** -0.036** -0.262** 0.024 0.021 -0.087* 0.038 0.021 -0.214** (0.035) (0.036) (0.033) (0.040) (0.017) (0.116) (0.047) (0.036) (0.042) (0.053) (0.034) (0.085) Year 1974 -0.047 -0.057 -0.047 -0.034 -0.005 -0.115 0.035 0.061* -0.034 0.076 0.056* -0.037 (0.042) (0.044) (0.032) (0.041) (0.022) (0.135) (0.044) (0.031) (0.052) (0.046) (0.030) (0.077) Year 1975 -0.050 -0.069* -0.045 -0.040 -0.018 -0.134 0.008 0.003 -0.040 0.051 0.030 -0.300** (0.036) (0.034) (0.033) (0.034) (0.016) (0.119) (0.039) (0.035) (0.071) (0.046) (0.036) (0.116) Year 1976 -0.072 -0.060 -0.037 -0.037 -0.012 -0.195* -0.017 -0.017 -0.083 0.021 0.029 -0.237*** (0.044) (0.044) (0.033) (0.039) (0.020) (0.106) (0.038) (0.038) (0.058) (0.040) (0.024) (0.064) Year 1977 -0.077* -0.086* -0.051** -0.058* -0.041* -0.377*** -0.006 0.012 -0.058 0.065 0.086** -0.185** (0.039) (0.044) (0.023) (0.030) (0.022) (0.099) (0.052) (0.045) (0.076) (0.052) (0.034) (0.071) Year 1978 -0.087** -0.079* -0.027 -0.066 -0.008 -0.360*** 0.049 0.046 -0.172*** 0.086* 0.063** -0.178* (0.041) (0.043) (0.031) (0.039) (0.023) (0.116) (0.044) (0.035) (0.053) (0.049) (0.026) (0.096) Year 1979 -0.113*** -0.111*** -0.058** -0.091** -0.065*** -0.466*** 0.026 0.020 -0.031 0.061 0.046* -0.256** (0.032) (0.034) (0.027) (0.032) (0.019) (0.108) (0.050) (0.040) (0.048) (0.045) (0.025) (0.091) 14 Year 1980 -0.102*** -0.110*** -0.070** -0.082** -0.046* -0.403*** -0.024 -0.006 -0.113** 0.049 0.068 -0.259*** (0.035) (0.037) (0.028) (0.037) (0.024) (0.095) (0.050) (0.040) (0.051) (0.044) (0.051) (0.079) Year 1981 -0.152*** -0.157*** -0.063** -0.109** -0.056** -0.523*** 0.021 0.040 -0.054 0.078 0.040 -0.370*** (0.034) (0.038) (0.028) (0.038) (0.020) (0.115) (0.038) (0.029) (0.045) (0.046) (0.024) (0.098) N 8823 8823 8823 8823 8823 8577 1715 1715 1715 1715 1715 1661 adj. R2 0.035 0.038 0.083 0.032 0.020 0.047 0.011 0.029 0.179 0.028 0.066 0.033 Notes: All regressions add birth governorate fixed-effects, household size and gender. Standard errors are clustered at the birth governorate level. *** p<0.01, ** p<0.05, * p<0.1. 15 Table 8 provides heterogeneous analysis on working conditions of workers. The results show men of cohorts of 1969 to 1981 are more likely to work in dangerous conditions and work without air conditioning while we find no evidence of the impact of the Islamic Republic of Iran–Iraq war on working conditions of female workers like hazardous job and air conditioning. Descriptive statistics shows that female workers work in hazardous conditions or without air conditioning accounted for only 26 percent in the survey. This explains why we find no impacts of the Islamic Republic of Iran– Iraq war on working conditions, including hazardous job and air conditioning. Regarding health impact, health of men of the cohorts of 1972, 1979, and 1980 was adversely affected at the working environment, while we find the evidence of the negative impact for health of women for all cohorts of 1970-1981. Table 8. Working conditions of workers between males and females Male Female Hazardous Air Health Hazardous Air Health job conditioning impact job conditioning impact (1) (2) (3) (4) (5) (6) Year 1964 0.002 -0.009 0.039 0.048 0.034 0.603** (0.052) (0.046) (0.216) (0.045) (0.065) (0.262) Year 1965 -0.026 -0.043 -0.182 0.028 0.029 0.491** (0.033) (0.031) (0.150) (0.046) (0.059) (0.228) Year 1966 0.006 -0.012 -0.010 0.002 0.059 0.903** (0.029) (0.038) (0.158) (0.056) (0.062) (0.357) Year 1967 0.011 -0.042 0.020 0.001 -0.003 0.413** (0.038) (0.039) (0.135) (0.047) (0.065) (0.171) Year 1969 (Year of change) -0.018 -0.095** 0.054 -0.018 0.045 0.376 (0.036) (0.037) (0.122) (0.051) (0.049) (0.277) Year 1970 -0.048 -0.097** 0.172 0.004 -0.028 0.451* (0.037) (0.045) (0.120) (0.051) (0.052) (0.215) Year 1971 -0.061 -0.053** -0.056 -0.036 0.031 0.769** (0.035) (0.019) (0.126) (0.076) (0.057) (0.315) Year 1972 -0.068* -0.111*** 0.265* 0.006 0.029 0.452* (0.034) (0.031) (0.153) (0.034) (0.055) (0.231) Year 1973 -0.102** -0.089*** 0.006 -0.100* -0.034 0.536** (0.036) (0.026) (0.143) (0.050) (0.068) (0.227) Year 1974 -0.109*** -0.069** 0.270 -0.015 0.024 0.703*** (0.031) (0.031) (0.164) (0.060) (0.083) (0.212) Year 1975 -0.049 -0.082*** 0.030 -0.059 -0.055 0.415** (0.039) (0.025) (0.136) (0.038) (0.067) (0.159) Year 1976 -0.058 -0.049 0.126 -0.003 0.050 0.473* (0.043) (0.031) (0.111) (0.044) (0.062) (0.253) Year 1977 -0.065* -0.072** 0.188 -0.008 0.004 0.616** (0.034) (0.032) (0.142) (0.056) (0.075) (0.288) Year 1978 -0.044* -0.085** 0.214 0.052 -0.051 0.608* (0.024) (0.035) (0.142) (0.059) (0.064) (0.301) Year 1979 -0.047 -0.115*** 0.303** 0.024 0.052 0.876*** (0.038) (0.025) (0.114) (0.042) (0.066) (0.221) Year 1980 -0.067** -0.089*** 0.133 -0.042 -0.005 0.553** (0.025) (0.028) (0.137) (0.051) (0.065) (0.218) Year 1981 -0.078** -0.149*** 0.290** -0.012 0.036 0.833*** 16 (0.032) (0.032) (0.111) (0.039) (0.043) (0.287) N 8823 8823 8823 1715 1715 1715 adj. R2 0.047 0.032 0.148 0.078 0.052 0.195 Notes: All regressions add birth governorate fixed-effects, household size and gender. Standard errors are clustered at the birth governorate level. *** p<0.01, ** p<0.05, * p<0.1. Severity of the Islamic Republic of Iran–Iraq war Educational attainment and labor market outcomes may be different between governorates. The children living in the areas with more intense fighting should do worse in school and later in the labor markets. The conflict concentrated along the countries’ joint border. Therefore, we expect that educational attainment of children born in the governorates near the Islamic Republic of Iran’s border would be worse than educational attainment of children born in other governorates in Iraq. We divide the sample into two sub-sample: a sub-sample with children born in the governorates near the Islamic Republic of Iran’s border and other sub-sample with children born in other governorates in Iraq. Table 9 reports the heterogeneous results on educational attainment between governorates near the Islamic Republic of Iran’s border and other Iraqi governorates. The Islamic Republic of Iran–Iraq war has larger impacts on literacy of cohorts 1975-1980 who born in governorates near the Islamic Republic of Iran’s border compared with literacy of those born in other governorates in Iraq. The impacts of the Islamic Republic of Iran–Iraq war on primary education completion or above is lower in governorates near the Islamic Republic of Iran’s border than in other governorates in Iraq. However, the impacts of the Islamic Republic of Iran–Iraq war on secondary education completion or above is greater in governorates near the Islamic Republic of Iran’s border than in other governorates in Iraq. Overall, we find that cohorts of 1970, 1972, 1974 who attain lower years of education in governorates near the Islamic Republic of Iran’s border than in other governorates in Iraq. Table 9. Educational attainment between border governorates and other Iraqi governorates Governorates near the Islamic Republic of Iran’s Other governorates in Iraq border Literacy Primary Secondary Years of Literacy Primary Secondary Years of education education education education education education completion completion completion completion or above or above or above or above (1) (2) (3) (4) (5) (6) (7) (8) Year 1964 0.027 0.043 0.047** -0.107 0.040** 0.013 0.122*** 0.170 (0.013) (0.026) (0.014) (0.210) (0.013) (0.022) (0.027) (0.200) Year 1965 0.023* 0.050 0.054 0.154 0.012 -0.000 0.101*** 0.110 (0.011) (0.026) (0.030) (0.394) (0.009) (0.019) (0.030) (0.197) Year 1966 0.024* 0.039 -0.018 -0.239 0.035** 0.005 0.115*** 0.218 (0.010) (0.035) (0.027) (0.160) (0.012) (0.013) (0.023) (0.161) Year 1967 0.003 0.023 0.009 -0.073 0.009 -0.011 0.074*** 0.302 (0.011) (0.015) (0.044) (0.353) (0.006) (0.019) (0.020) (0.198) Year 1969 -0.016 0.013 -0.015 -0.214 0.001 -0.006 0.027 0.086 (Year of 17 change) (0.011) (0.028) (0.032) (0.233) (0.007) (0.009) (0.020) (0.166) Year 1970 -0.005 0.001 -0.110*** -0.521** -0.004 -0.038 0.008 -0.043 (0.005) (0.012) (0.023) (0.168) (0.006) (0.021) (0.020) (0.141) Year 1971 -0.018 -0.025 -0.089*** -0.124 -0.011* -0.047** -0.017 -0.124 (0.011) (0.014) (0.021) (0.221) (0.005) (0.015) (0.029) (0.177) Year 1972 -0.021 -0.029 -0.106** -0.558* -0.008 -0.061** -0.034* -0.285* (0.014) (0.026) (0.029) (0.268) (0.006) (0.023) (0.017) (0.147) Year 1973 -0.012 -0.020 -0.094** -0.433 -0.010* -0.075*** -0.039** -0.439*** (0.012) (0.032) (0.027) (0.267) (0.005) (0.019) (0.016) (0.122) Year 1974 -0.010 -0.042* -0.093** -0.605* -0.005 -0.064*** -0.035** -0.322** (0.008) (0.020) (0.032) (0.261) (0.008) (0.014) (0.013) (0.124) Year 1975 -0.017* -0.032 -0.044 -0.185 -0.011* -0.038** -0.003 -0.119 (0.007) (0.027) (0.041) (0.529) (0.006) (0.016) (0.023) (0.155) Year 1976 -0.017* -0.030** -0.123*** -0.612 -0.010 -0.059** -0.023 -0.088 (0.008) (0.009) (0.028) (0.309) (0.005) (0.020) (0.020) (0.133) Year 1977 -0.016* -0.017 -0.069 -0.135 -0.009 -0.062*** -0.014 -0.142 (0.007) (0.021) (0.035) (0.233) (0.007) (0.012) (0.012) (0.140) Year 1978 -0.014 -0.045** -0.085* -0.451 -0.012* -0.078*** -0.073*** -0.445** (0.014) (0.015) (0.036) (0.501) (0.006) (0.022) (0.019) (0.163) Year 1979 -0.019** -0.026 -0.088 0.027 -0.011 -0.070*** -0.042*** -0.233** (0.007) (0.013) (0.045) (0.365) (0.006) (0.015) (0.012) (0.104) Year 1980 -0.020* -0.052*** -0.099* -0.265 -0.012** -0.076*** -0.020 -0.151 (0.008) (0.012) (0.040) (0.604) (0.005) (0.013) (0.015) (0.148) Year 1981 -0.027*** -0.062** -0.140*** -0.556 -0.015** -0.071*** -0.054** -0.255 (0.005) (0.021) (0.033) (0.310) (0.005) (0.013) (0.023) (0.162) N 8802 8802 8802 7262 20142 20142 20142 17776 adj. R2 0.017 0.051 0.027 0.086 0.020 0.032 0.023 0.076 Notes: All regressions add birth governorate fixed-effects, household size and gender. Standard errors are clustered at the birth governorate level. *** p<0.01, ** p<0.05, * p<0.1. Table 10 provides the results on social welfare disparities between border governorates and other Iraqi governorates. Workers born in the governorates near the Islamic Republic of Iran’s border have lower opportunities of accessing to social security and pension plan than those born in other Iraqi governorates. 18 Table 10. Social welfare disparities between border governorates and other Iraqi governorates Governorates near the Islamic Republic of Iran’s border Other Iraqi governorates Dummy Dummy Dummy Dummy Dummy Log of Dummy Dummy Dummy Dummy Dummy Log of variable variable variable variable variable for wage variable variable variable variable variable for wage for social for for health for paid job for social for for health for paid job security pension care leave permanence security pension care leave permanence plan plan (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) Year 1964 -0.087* -0.066 -0.013 -0.088* 0.017 -0.054 0.021 0.032 -0.036 0.068 0.021 0.198 (0.037) (0.044) (0.043) (0.035) (0.054) (0.148) (0.057) (0.053) (0.041) (0.053) (0.021) (0.185) Year 1965 0.032 0.064 0.065 0.042 -0.014 0.067 -0.025 -0.030 -0.044 -0.002 -0.036 -0.124 (0.050) (0.049) (0.054) (0.047) (0.046) (0.168) (0.045) (0.044) (0.034) (0.042) (0.027) (0.160) Year 1966 -0.040 -0.025 0.058 -0.006 -0.013 -0.259 -0.009 0.001 -0.036 0.030 0.003 0.033 (0.059) (0.058) (0.044) (0.051) (0.036) (0.247) (0.028) (0.031) (0.035) (0.022) (0.024) (0.128) Year 1967 -0.097* -0.087 -0.078* -0.084 -0.042 -0.375*** -0.017 -0.011 -0.039 0.038 0.017 0.040 (0.042) (0.045) (0.032) (0.073) (0.036) (0.082) (0.044) (0.041) (0.039) (0.048) (0.022) (0.139) Year 1969 (Year -0.061 -0.069** -0.041 -0.063* -0.023 -0.250** -0.054 -0.055 -0.088** -0.059 -0.017 -0.176 of change) (0.036) (0.025) (0.037) (0.028) (0.049) (0.078) (0.050) (0.056) (0.032) (0.054) (0.025) (0.158) Year 1970 -0.091** -0.089* -0.048 -0.050 -0.034 -0.216* -0.040 -0.045 -0.053 -0.031 -0.011 -0.215* (0.023) (0.038) (0.038) (0.050) (0.022) (0.092) (0.040) (0.043) (0.035) (0.037) (0.022) (0.115) Year 1971 -0.019 -0.022 -0.024 -0.015 -0.007 -0.264 -0.056 -0.052 -0.084*** -0.040 0.006 -0.107 (0.040) (0.033) (0.045) (0.042) (0.027) (0.155) (0.037) (0.041) (0.023) (0.036) (0.022) (0.142) Year 1972 -0.036 -0.026 -0.001 -0.047 -0.028 -0.229* -0.075* -0.079* -0.059** -0.044 -0.010 -0.269** (0.038) (0.050) (0.028) (0.068) (0.027) (0.098) (0.037) (0.039) (0.024) (0.037) (0.016) (0.106) Year 1973 -0.096* -0.092* -0.044 -0.087 -0.022 -0.235 -0.085* -0.092** -0.054 -0.060 -0.029 -0.254* (0.040) (0.041) (0.033) (0.056) (0.032) (0.124) (0.040) (0.040) (0.040) (0.042) (0.017) (0.135) Year 1974 -0.018 0.000 -0.007 0.027 0.015* 0.043 -0.038 -0.053 -0.060* -0.034 0.000 -0.163 (0.045) (0.054) (0.035) (0.052) (0.006) (0.137) (0.050) (0.053) (0.033) (0.050) (0.028) (0.158) Year 1975 -0.053*** -0.061* -0.044 -0.060** 0.009 -0.202 -0.027 -0.050 -0.034 -0.001 -0.021 -0.145 (0.013) (0.024) (0.076) (0.019) (0.030) (0.191) (0.045) (0.043) (0.035) (0.040) (0.017) (0.134) Year 1976 -0.128*** -0.092** -0.048 -0.076 0.011 -0.286** -0.033 -0.036 -0.040 -0.006 -0.011 -0.166 (0.024) (0.027) (0.028) (0.047) (0.030) (0.086) (0.049) (0.052) (0.040) (0.040) (0.022) (0.128) Year 1977 -0.089*** -0.090 -0.021 -0.006 -0.001 -0.352** -0.055 -0.062 -0.061** -0.051 -0.029 -0.355*** (0.022) (0.054) (0.022) (0.057) (0.037) (0.124) (0.044) (0.052) (0.025) (0.032) (0.022) (0.114) Year 1978 -0.070** -0.028 -0.006 -0.037 -0.021 -0.289* -0.068 -0.077 -0.068** -0.046 0.013 -0.360** (0.021) (0.048) (0.048) (0.055) (0.045) (0.118) (0.053) (0.053) (0.030) (0.047) (0.020) (0.144) Year 1979 -0.068* -0.068 0.002 -0.074 -0.036 -0.459*** -0.108** -0.107** -0.081** -0.067** -0.053* -0.437*** (0.030) (0.039) (0.027) (0.056) (0.032) (0.095) (0.039) (0.040) (0.031) (0.029) (0.025) (0.129) Year 1980 -0.124** -0.135** -0.086** -0.103* 0.010 -0.464*** -0.075* -0.076* -0.065* -0.041 -0.050 -0.350** 19 (0.033) (0.036) (0.032) (0.044) (0.040) (0.059) (0.038) (0.040) (0.033) (0.039) (0.028) (0.118) Year 1981 -0.127** -0.117** -0.062 -0.072 -0.077 -0.595*** -0.128*** -0.135** -0.060** -0.084* -0.027 -0.478*** (0.037) (0.035) (0.046) (0.037) (0.040) (0.101) (0.039) (0.045) (0.027) (0.044) (0.018) (0.141) N 3351 3351 3351 3351 3351 3169 7159 7159 7159 7159 7159 7043 adj. R2 0.125 0.141 0.110 0.130 0.033 0.074 0.145 0.156 0.131 0.145 0.036 0.075 Notes: All regressions add birth governorate fixed-effects, household size and gender. Standard errors are clustered at the birth governorate level. *** p<0.01, ** p<0.05, * p<0.1. 20 Table 11 shows that workers born in the governorates near the Islamic Republic of Iran’s border work in more hazardous conditions and have lower probability of accessing to air conditioning than those born in other governorates in Iraq. Also, we find that workers born in the governorates near the Islamic Republic of Iran’s border are more adversely affected at the working environment than those born in other governorates in Iraq. Table 11. Disparities in working conditions of workers between border governorates and other Iraqi governorates Governorates near the Islamic Republic of Other governorates in Iraq Iran’s border Hazardous Air Health impact Hazardous Air Health impact job conditioning job conditioning (1) (2) (3) (4) (5) (6) Year 1964 -0.012 -0.076 0.411 0.027 0.038 -0.007 (0.054) (0.042) (0.354) (0.059) (0.046) (0.227) Year 1965 -0.015 -0.052 -0.061 -0.019 -0.014 -0.113 (0.049) (0.052) (0.366) (0.033) (0.033) (0.122) Year 1966 -0.012 -0.009 0.386 0.010 0.007 0.051 (0.057) (0.039) (0.272) (0.030) (0.045) (0.228) Year 1967 -0.007 -0.092 0.164 0.019 0.003 0.030 (0.053) (0.060) (0.274) (0.043) (0.038) (0.132) Year 1969 -0.007 -0.106 0.049 -0.021 -0.046 0.093 (Year of change) (0.043) (0.074) (0.193) (0.040) (0.033) (0.149) Year 1970 -0.133* -0.114* 0.364* 0.008 -0.068 0.127 (0.063) (0.052) (0.157) (0.035) (0.051) (0.118) Year 1971 -0.052 -0.045 0.157 -0.061 -0.033 0.031 (0.051) (0.042) (0.136) (0.039) (0.021) (0.160) Year 1972 -0.076 -0.040 0.298 -0.046 -0.107** 0.270 (0.049) (0.056) (0.160) (0.036) (0.036) (0.159) Year 1973 -0.172** -0.094 0.169 -0.071* -0.071** 0.062 (0.052) (0.050) (0.210) (0.038) (0.027) (0.147) Year 1974 -0.164** -0.116 0.468* -0.066* -0.025 0.254 (0.045) (0.062) (0.206) (0.035) (0.030) (0.162) Year 1975 -0.116 -0.138** 0.254 -0.020 -0.047* 0.002 (0.058) (0.044) (0.221) (0.037) (0.026) (0.119) Year 1976 -0.105 -0.093 0.534** -0.020 -0.002 -0.019 (0.089) (0.051) (0.200) (0.037) (0.026) (0.125) Year 1977 -0.109 -0.093 0.512* -0.030 -0.039 0.138 (0.073) (0.050) (0.206) (0.032) (0.039) (0.166) Year 1978 -0.088 -0.120 0.636* -0.007 -0.056 0.099 (0.051) (0.066) (0.277) (0.021) (0.038) (0.143) Year 1979 -0.076 -0.071 0.672** -0.017 -0.094*** 0.251* (0.058) (0.061) (0.180) (0.038) (0.028) (0.125) Year 1980 -0.115* -0.142** 0.470* -0.041 -0.038 0.050 (0.048) (0.040) (0.206) (0.025) (0.028) (0.142) Year 1981 -0.111*** -0.143* 0.628** -0.049 -0.105** 0.242** (0.024) (0.064) (0.230) (0.037) (0.034) (0.104) N 3351 3351 3351 7159 7159 7159 adj. R2 0.112 0.083 0.166 0.130 0.089 0.127 21 Notes: All regressions add birth governorate fixed-effects, household size and gender. Standard errors are clustered at the birth governorate level. *** p<0.01, ** p<0.05, * p<0.1. 5. Conclusions A large body of literature examines the long-lasting impacts of early exposure to the war on educational attainment and labor market outcomes in adulthood. Nonetheless, there is little known about the long-term impact on children, who were exposed to the war at the ages of 6 to 12 in Iraq. The Islamic Republic of Iran–Iraq war in 1980-1988 provides an excellent setting to study the long- term impact of war. Apart from all other—human, financial, and emotional—costs, the Islamic Republic of Iran–Iraq war led to a significant drop in educational attainment for both men and women who were of elementary school age during the conflict. The Islamic Republic of Iran–Iraq war had a large impact on men’s education level and women’s education level. In addition to this significant educational loss, the Islamic Republic of Iran–Iraq war appears to have caused an adverse impact on social welfare of men, such as accessing social security, pension plan, health care, paid leave and job permanence. This impact is not statistically significant for women. We also find that men who were exposed to the war during elementary school age have higher probability of working in hazardous jobs and without air conditioning; however, we find no evidence for women. Similarly, we find evidence of the impact of war intensity on educational attainment and labor market outcomes. We conduct a variety of robustness checks and falsification tests, and our estimation results are still robust. We believe that the most likely channel through which the Islamic Republic of Iran–Iraq war affects the labor market outcomes is the educational loss induced by the war. The educational loss of men is larger than the educational loss of women, and the Islamic Republic of Iran–Iraq war has more adverse impacts on labor market outcomes for men than for women. 22 References Akresh, R., Caruso, G, D., Harsha, T. 2022. Detailed geographic information, conflict exposure, and health impacts. World Development, 2022, 155. Akresh, R., Bhalotra, S., Leone, M., Osili, U.O. 2023. First and Second Generation Impacts of the Biafran War, Journal of Human Resources, 58(2), 488-531. Allais, O., Fagherazzi, G, Mink, J. 2021. The long-run effects of war on health: Evidence from World War II in France. Social Science & Medicine, 276, Ashenfelter, O. and A. Krueger. 1994. Estimates of the economic return to schooling from a new sample of twins. American Economic Review, 84(5): 1157-1173. Becker, G.S., 1993. Human Capital, third ed. The University of Chicago Press, Chicago. Blattman, C and Annan, J. 2010. The Consequences of Child Soldiering, Review of Economics and Statistics, 92(4): 882–898. Brakman, S., Harry G and Marc, S. 2004. The Strategic Bombing of Cities in Germany in World War II and its Impact on City Growth,” Journal of Economic Geography, 4, 1-18. Bundervoet ,T., Verwimp, P and Akresh, R., 2009, Health and Civil War in Rural Burundi, Journal of Human Resources, 44(2). Card, D. 2001. Estimating the return to schooling: Progress on some persistent econometric problems. Econometrica 69(5): 1127-1160. Chen, Qihui. 2015. Ready for School? Impacts of Delayed Primary School Enrollment on Children's Educational Outcomes in Rural China. International Journal of Educational Development, 45:112- 128. Clark, D. and P. Martorell. 2014. The signaling value of a high school diploma. Journal of Political Economy, 122:2: 282-318. Crystal S., Shea D., Krishnaswami S. (1992). Educational attainment, occupational history, and stratification: Determinants of later-life economic outcomes. Journal of Gerontology, 47, S213–S221. Davis, D and David, W., 2002, Bones, Bombs, and Break Points: The Geography of Economic Activity, American Economic Review, 92, pp. 1269-1289. Duflo, E. 2001. Schooling and labor market consequences of school construction in Indonesia: Evidence from an unusual policy experiment. American Economic Review 91(4): 795-813. 23 Fiszbein, A., P.I. Giovagnoli H.A. Patrinos. 2007. Estimating the returns to education in Argentina using quantile regression analysis: 1992-2002. Económica 53. Gianmarco, L., 2012, Civil Conflict and Human Capital Accumulation: The Long-term Effects of Political Violence in Perú, Journal of Human Resources, 47(4), pp. 991-1022. Harmon, C., H. Oosterbeek and I. Walker. “The returns to education: Microeconomics.” Journal of Economic Surveys 17(2): 115-156. Ichino, A., and R. Winter-Ebmer. 2004. “The Long-Run Educational Cost of World War II,” Journal of Labor Economics, 22(1), 57–86. Miguel, E and Roland, G., 2011, The Long-run Impact of Bombing Vietnam, Journal of Development Economics, 96, 1-15. No, Fata, Kyoko Taniguchi, and Yukiko Hirakawa. 2016. School Dropout at the Basic Education Level in Rural Cambodia: Identifying its Causes through Longitudinal Survival Analysis. International Journal of Educational Development, 49: 215-224. Olga, S, 2011, The Effect of Armed Conflict on Accumulation of Schooling: Results from Tajikistan, Journal of Development Economics, 95, 1986-200. Oreopoulos, P. 2006. Estimating average and local average treatment effects of education when compulsory schooling laws really matter. American Economic Review, 96(1): 152-175. Oreopoulos P. 2007. Do dropouts drop out too soon? Wealth, health and happiness from compulsory schooling. Journal of Public Economics, 91, 2213–2229. Oreopoulos P., Salvanes K. G. 2011. Priceless: The nonpecuniary benefits of schooling. Journal of Economic Perspectives, 25(1), 159–184. Peet, E.D., G. Fink and W. Fawzi. 2015. Returns to education in developing countries: Evidence from the living standards and measurement study surveys. Economics of Education Review, 49: 69-90. Reach, 2019, Iraq: Assessment on Employment and Working Conditions of Conflict-affected Women Across Key Sectors, available at https://reliefweb.int/report/iraq/iraq-assessment- employment-and-working-conditions-conflict-affected-women-across-key Shonkoff, J.P., Phillips, D.A., 2000. From Neurons to Neighborhoods. The Science of Early Child Development. National Academic Press, Washington, DC. Silliman, M, and Virtanen, H. 2022. Labor Market Returns to Vocational Secondary Education. American Economic Journal: Applied Economics, 14 (1): 197-224. 24 Singhal, S. (2019). Early life shocks and mental health: The long-term effect of war in Vietnam, Journal of Development Economics, 141. UNICEF. 2014. Iraq Country Report on Out-Of-School Children. Baghdad: UNICEF. UNESCO Institute for Statistics (UIS). 2005. Children Out of School: Measuring Exclusion from Primary Education. Montreal: Canada. Yuksel, A, M. 2017. War during childhood: The long run effects of warfare on health, Journal of Health Economics, 53, 117-130. Vilardo, V and Bittar, S. 2018. Gender Profile – Iraq: A situation analysis on gender equality and women’s empowerment in Iraq, Research report, Oxfarm , UN Women. Wils, Annababette. 2004. Late Entrants Leave School Earlier: Evidence from Mozambique. International Review of Education 50, no 1:17-37. 25