102592 Rising Tempers, Rising Temperatures: A Look At Climate Change, Migration And Conflict And The Implications For Youth In The Sahel Region Erik Alda March 2014 2014 The International Bank for Reconstruction and Development / The World Bank 1818 H Street, NW Washington, DC 20433 www.tdrp.net www.worldbank.org The findings, interpretations, and conclusions expressed herein are those of the author(s) and do not necessarily reflect the views of the Executive Directors of the International Bank for Reconstruction and Development/The World Bank or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denomina- tions, and other information shown on any map in this work do not imply any judgment on the part of the World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Photo Credits: Sergio Pesselano and Thomas Seligman. Cover Design and Text Layout: Duina Reyes Table of Contents Executive Summary........................................................................................ 3 I. Introduction................................................................................................. 4 2. Purpose of the Study................................................................................... 5 2.1 Climate Change: Recent Trends in the Africa and Sahel Regions.................................................. 5 2.2 Changes in Temperature............................................................................................................... 6 2.3 Changes in Rainfall....................................................................................................................... 6 2.4 Consequences of Climate Change: Migration And Conflict. ......................................................... 7 3. Climate Change Implications for Youth...................................................... 14 4. Conclusion and Policy Options . ................................................................ 18 4.1 Climate Change Adaptation Measures:....................................................................................... 18 4.2 Conflict and Crime Prevention Measures.................................................................................... 19 Annex................................................................................................................................................. 20 References.................................................................................................... 21 List of Figures and Tables Figure 1 - Projected climate change in Africa........................................................................................ 5 Figure 2: Rainfall Average (NASA, 2012). ............................................................................................. 6 Figure 3: Population Migration (Total population received in country of asylum). ............................... 7 Figure 4: Average Marginal Effects (Drought)..................................................................................... 11 Figure 5: Average Marginal Effects (Floods)........................................................................................ 11 Figure 6: Marginal Effects of Interaction Term (Sahel*Drought). ........................................................ 12 Figure 7: Marginal Effects of Interaction Term (Sahel*Flood).............................................................. 12 Figure 8: Number of vulnerable people per country (UNOCHA, 2012)............................................... 14 Figure 9: Youth Population Cohort in the Sahel Countries.................................................................. 20 Table 1 - Results of Regression Analyses of Climate Change Related Disasters and Conflict................. 9 Table 2 - Results of Structural Equation Modeling.............................................................................. 13 Table 3 - Direct, Indirect, and Total Effects of Explanatory Variables on Conflict Onset and Migration. ......................................................................................................................... 15 Table 4 - Estimation Results: Probit Regression for the Sahel Region................................................. 16 1 Executive Summary C limate change is perhaps the most important development issue of the twenty-first century. The costs and consequences of climate change already being observed in many countries are severely impacting the livelihoods of millions of people, but that further research is needed in order to better under- stand the mechanisms and paths through which climate change might increase the risk of conflict. Section 3 brings the aspect of youth into the analysis, especially those of the poorest. This paper investigates arguing that youth living in the rural Sahel region will the causal chain between climate change, migration and likely find themselves particularly vulnerable to in- conflict in the context of the Sahel region, with special creased climate change and variability, which will have focus on implications for the youth cohort living in the direct and indirect implications on the security and most vulnerable rural areas, and concludes with policy livelihoods of these individuals. The paper explores the recommendations aimed at mitigation of the effects of past, current and predicted issues faced by rural Sahel climate variability and prevention of youth engagement youth affected by the impacts of climate change, such in climate change effects-related conflict and crime. as chronic unemployment and engagement in the illicit sector due to rural to urban migration from regions Section 1 of the paper provides a brief overview of the affected by drought, desertification and resource relationship between the Sahel region of Africa and scarcity. climate variability trends and predictions, ultimately posing the primary research question of the study: Is The final section of the paper offers policy recommen- the Sahel region more likely to have a higher probability dations for addressing negative climate-change related of conflict and migration as a result of climate change consequences through two key paths of intervention: and climate change-related events? (i) Climate change adaptation measures and (ii) conflict and crime prevention measures. Recommendations for In Section 2, research and analysis aim to identify causal climate change adaptation measures aim at improving paths between climate change, migration, and conflict governmental capacity, building social capital in areas via structural equation modeling focusing on recent such as resilience and empowerment, and increasing climate change trends in the Africa and Sahel regions, asset-based adaptation. Recommendations for conflict particularly changes in temperature and rainfall, and and crime prevention measures center on providing analyzing the consequences of this climate variabil- employment and training programs as well as promot- ity, namely migration and conflict. The results of the ing the preservation of pastoral practices such as crop- analysis yield findings supporting the hypothesis found livestock systems (Izcowizc et. Al, 2012; Shapland et. in the existing literature that climate change, itself, does Al, 2013). Policies and programs should be tailored not cause conflict directly, but rather through a variety to the needs and characteristics of each country in the of channels such as migration, and ultimately concludes region instead of “one size fits all” programs. 3 I. Introduction C limate change is arguably the most important development and policy challenge of the twen- ty-first century (Verner, 2010). The negative consequences of climate change are and will continue to be a threat to some of the recent progress in terms of The two most pressing negative consequences of climate change are migration and conflict. The dete- riorating conditions resulting from climate change can cause an accelerating increase in migration patterns. Increased migration patterns can exact negative effects poverty reduction, economic growth and health gains. on the already overstretched local capacity of the recipi- In addition, climate change is also likely to unravel ent locations as well as potentially affecting the climate conflict as a result of increasingly limited access to basic of these locations. Conflict is, therefore, perhaps the resources. The intersection between climate change, most serious outcome of the negative effects of climate loss of livelihoods, migration, and conflict are of par- change. In fact, climate change is already being increas- ticular importance in the developing world. ingly considered a security threat in the Africa region (ISS, 2013). The costs and consequences of climate change have already been observed in many countries (NASA, Given all of the above, is the Sahel region more likely 2012). In the coming decades, the forecasts anticipate to have a higher probability of migration and conflict worrisome trends. For example, the Inter-Governmen- as a result of climate change and climate change-related tal Panel on Climate Change predicts that temperature events? More importantly, what are the implications increases in the coming decades from 2 to 4 degrees for the youth population, given that it is a key cohort Celsius are likely to have significant damaging effects for the present and the future of the region? This paper globally (IPCC, 2010). According to estimates from attempts to answer this question by investigating the the UN, countries in Africa are more likely to be more links between climate change and climate change-re- vulnerable to the effects of climate change. In excess lated events, migration, and, ultimately, conflict. It is of 250 million people are likely to be affected by rising important to note, however, that this paper does not temperatures and the negative effects associated with it, attempt to cover or address the vast complexities of which include water and food security and the loss of climate change and its consequences. Rather, it briefly livelihood. Further, countries in West, North and East describes the recent evolution of climate change in Africa are more likely to bear the brunt of the effects of Africa, and, in particular, for the Sahel region, and dis- climate change (UNEP, 2011). cusses how one of the most important consequences of climate change - conflict and crime - can affect one of Although the negative consequences of climate change the most vulnerable population cohorts: youths. The will affect everyone, females, children and youth are remainder of the paper is organized as follows. Section particularly vulnerable (UNESA, 2010).Particularly 2 discusses the effects of climate change with a particu- the youth in fragile regions of Africa like the Sahel are lar focus on Africa, which is likely to experience more likely to be more affected by the rapid and extreme intensely the consequences of climate change. Section 3 change in weather patterns. Extreme weather events discusses the implications that climate change will have and increased climate variability will have direct del- for the youth in the Sahel region. Finally, Section 4 of eterious effects on the health and safety of many youth the paper presents the conclusion and provides broad in this region. For example, it is estimated that between policy recommendations on how governments can best 30 to 50 percent of youth in Africa currently lack access employ adaptation measures to climate change so that to adequate basic services and, thus, are at an increased these youths can adapt to the rapidly changing climate risk of climate change (UNESA, 2010). and its negative consequences while, at the same time, preventing them from engaging in conflict and crime. 4 2. Purpose of the Study 2.1 Climate Change: Recent Trends structure, among others (IPCC, 2007). While there is limited information about the real effects of climate in the Africa and Sahel Regions change in Africa, predictions suggest the following scenario (as it appears in Conway, 2009; ICC, 2010) Available evidence demonstrates that the climate is changing rapidly. In the past century temperatures have ¡¡ Drier subtropical regions will warm more than the increased almost by 1°C, of which more than 0.5°C has moister tropics; taken place since 1979 (IPCC, 2007). Rainfall patterns have also changed and are increasingly erratic through- ¡¡ Southern Africa will become much hotter (as out the continent (Tadesse, 2010). According to IPCC’s much as 4°C or more) and drier (precipitation fourth synthesis report, the increases in climate change falling by 15 percent or more); and climate variability pose a serious threat to long- term development in Africa, which is further aggravat- ¡¡ Wheat production in the north and maize ed by the interplay of other development stresses such production in the south are likely to be adversely as poverty, weak institutional capacity, and poor infra- affected; Figure 1 - Projected climate change in Africa 5 ¡¡ Eastern Africa, including the horn of Africa, and The southern parts of Mali, for example, are expected parts of central Africa average rainfall is likely to to experience a decrease in temperatures whereas in the increase but it will come in the form of intense northern regions of Mali and Mauritania, the increase floods; in temperature can go up to 4°C and 2°C, respectively. Such variation in temperatures coupled with recurrent ¡¡ Vector-borne diseases such as malaria and dengue and more intense droughts is likely to reduce agricul- may spread and become more severe; tural production as well as accelerate the process of ¡¡ Sea levels will rise, perhaps by half a meter, in the desertification. This will reduce the livelihoods of the next fifty years, with serious consequences in the poorest population living in those areas and will trigger Nile delta and certain parts of West Africa. new waves of migrants. In the Sahel region, climate trends over the past century 2.3 Changes in Rainfall have varied by country but, on average, the region has experienced increases in mean temperatures and a In addition to increases in mean temperatures, the decrease in the amount of rainfall (IPCC, 2007). region is also likely to be affected by changes in pre- cipitation patterns. Data on rainfall patterns for the past 2.2 Changes in Temperature century shows an increase in absolute rainfall ranging from 100mm to more than 250mm, annually (UNEP, Specific temperature-change predictions in the Sahel 2011). Such an increase, however, masks tremendous region are significant and vary by country and sub- variation within and between countries in the region regions within countries. While it is predicted that the as well as the intensity of the rainfall. For example, it average increase in temperature in Africa will range is expected that the increase in rainfall could come between 0.2°C and 0.5°C in the coming decades, coun- in very intense forms, such as deluges and flooding. tries in the Sahel are likely to suffer higher temperature Moreover, research shows that too much rain can also increases (Kandji, Verchot, and Mackensen, 2006). For affect the degree to which rain-fed agriculture can be example, Burkina Faso is expected to suffer an increase practiced effectively (ECOWAS, 2006). Rainfall varia- in temperatures by 2.5°C; Chad will suffer increases tion will also affect the seasonality of the rains and its of 0.6°C and 1.7°C; Mali could experience increases severity. The occurrence of floods in the region has in temperature as high as 4°C (Kandji et al., 2006). increased in the past 30 years. The data indicate that Further, the increase in temperatures is likely to be un- during the 1985-2009 period, the frequency and area of equally distributed within countries (UNEP, 2011). the flooding increased significantly (UNEP, 2011). The Figure 2: Rainfall Average (NASA, 2012) 6 negative impacts of increased flooding to agricultural on average increasing from one in eight years to one in production and the populations living in those areas are every two or three years. The impact of such a change also significant. For example, the floods in 2010 in the of agricultural production and pastoral lifestyles is sub- lower Niger River valley affected nearly 10,000 Ha of stantial, as these populations, therefore, may be forced rice (Boyd et al., 2013). Similarly, in 2008, the floods in to migrate in order to subsist. Somalia destroyed at least 8,000 Ha (IRIN, 2008). Much of the current concern regarding the security 2.4 Consequences of Climate implications of climate change relates to the possible Change: Migration and Conflict adaptive methods that might be chosen by individuals, communities, and States in the face of large-scale and/ or rapid disruptions to their economies and societies. As climate change is expected to affect rainfall vari- Migration is one such adaptive method. Throughout ability, water resource availability, land degradation history, migration has been a chosen response in the and land availability, one should begin to look for face of poverty, resource scarcity, ethnic tension and links between these causal factors and different types conflict, and local environmental change may simply of conflict, including one-sided violence, intra-State come to serve as yet another push factor. The Horn of violence between groups and unorganized violence. Africa, for example, faces the worst drought it has seen A variety of factors are likely to be determined to con- in the past 60 years. The crisis has resulted in massive dition the results. One mechanism through which displacement within Somalia and across borders into climate changes affect migration and conflict is climate northern Kenya, Yemen, Ethiopia and Djibouti. In fact, change-related natural disasters (Slettebak, 2012). The data collected by UNHCR show a significant increase regional climate trends observed over the last 40 years in the number of migrants in recent years (Figure 3). across the Sahel and the Horn of Africa show that tem- peratures have risen, droughts have been recurrent and The scale of internal and cross-border migration, is, severe, rainfall has generally increased and floods have therefore, expected to rise in the face of drought and occurred more frequently and with greater intensity. desertification. This argument is based on the assump- Furthermore, there has been a general increase in mean temperature across Figure 3: Population Migration the region since 1970, with more than (Total population received in country of asylum) half the population experiencing an increase of between 0.5°C and 1°C. While the mean seasonal rainfall is below the long-term average (1900- 2009, Figure 1), precipitation has in- creased in some parts of the region, although the flooding has increased in frequency and severity. The area has experienced recurrent and severe drought since the 1970s, which has had a very significant impact on live- lihoods, particularly among pastoral peoples, through the gradual and growing unsuitability of traditional grazing grounds and the drying of im- portant water bodies. Cyclical droughts in the Horn of Africa and across the Sahel are increasing Source: Own Analysis in frequency, with the drought cycle 7 tion that increased population growth will increase in heightened water stress in the absence of irrigation. demands for resources, which, in turn, leads to land and The combined effects of these mechanisms could be water degradation and distributional conflicts. Even if expected to reduce African staple crop yields by 10 to one were to assume these conventional causal mecha- 30 percent for every degree of Celsius of warming. This nisms are valid, this analysis does not identify short- is an important finding in the light of the heavy depen- dence of African countries on agriculture. Adverse con- term causes that trigger the outbreak of conflict. sequences of climate change for food security in Africa There have already been cases of mass population move- seems inevitable, as demonstrated by the food situa- ments in response to climate variability throughout the tions in Somalia and the Sahel. Sahel and the Horn of Africa. One study analyzing 38 One particular study specifically predicts the likelihood cases since the 1930s in which environmental factors of conflict emergence in sub-Saharan Africa related to played a role in triggering mass migration concluded environmental degradation (Magadza, 2000). Magadza that in 19 instances some form of conflict resulted foresees conflict arising out of increased frequency and/ (Reuveny, 2007). The study affirms, however, that or severity of droughts. The author based his predic- environmental factors do not work in isolation, but tion, however, on an extrapolation from current condi- rather they contribute to pre-existing migration flows. tions, using no predictive data. Another study suggests an increase of 54% in armed conflict in Africa by 2030 Conflict arises when migrants, particularly those of a compared to the 1980-2000 period (Adano & Daudi, different nationality or ethnicity, move quickly and in 2012). It assumes, however, a link between violent large numbers into neighboring States already suffering conflict and climate-induced conflict. conflict and/or similar environmental degradation. Another study of civil wars between 1945 and 2005 Other research, however, reports contradictory results, found that of 103 ethnic conflicts, 32 included violence delinking climate change from conflict. For example, between members of an ethnic minority resident in a Slettebak (2012) replicated Fearon and Laitin’s (2003) data on conflict, and after including the population particular region and recent migrants from other areas measure in the regression model, the linkages between (Fearon and Laitin, 2003). Similarly, Homer-Dixon climate change-related natural disasters and the onset (1994) argues that large scale climate induced migra- of conflict disappeared. tions might contribute to aggravating existing tensions– ethnic, social, political or economic–that could trigger Although the empirical evidence linking climate change the onset of conflict. to conflict appears to find contradictory arguments, available data increasingly show that climate-related The links between climate change and conflict have been natural disasters are a problem for the Sahel region. tested extensively in the literature although the causality In order to tease out the effects of climate-related disas- mechanisms are not well understood. One such study ters on conflict, this paper’s analysis employs a dataset using geo-referenced data found that water and land (Slettebak, 2012) on climate change-related natural scarcity are only weak predictors of future conflict while disasters to test whether this relationship holds for the high population density is a consistently strong predic- Sahel region. tor of armed conflict. The same study also found that The data include measures related to the onset of conflict, the interaction of population growth and water scarcity population and economic measures, and climate related proved to be significant (Raleigh & Urdal, 2007). disaster measures. In addition, the analysis includes Burke et al used 1981 to 2002 panel data on climate a variable that captures the migration patterns in the variation and conflict events to examine the historical Sahel countries by country of asylum. risk of conflict in Africa. The results revealed that tem- In order to understand the relationship between climate perature can affect agricultural yields through increases change, migration, and conflict, this paper employs two in crop and surface water evapotranspiration, resulting different approaches. The first approach uses a similar 8 Table 1 - Results of Regression Analyses of Climate Change Related Disasters and Conflict VARIABLES Probit:Sahel Probit:Sahel Probit:Africa Probit:Africa Conflict previous year -0.907** -1.101*** -0.439* -0.441* [0.45] [0.41] [0.25] [0.25] F&L per capita income (lagged) -0.718 -0.89 -0.172** -0.172** [0.59] [0.61] [0.07] [0.07] F&L log population (lagged) 0.362*** 0.354** 0.141** 0.138** [0.12] [0.17] [0.07] [0.07] F&L log (% mountainous) 0.138 0.114 0.0435 0.0363 [0.15] [0.13] [0.06] [0.06] F&L instability (lagged) -0.0889 -0.00187 0.0629 0.0578 [0.48] [0.39] [0.15] [0.15] F&L democracy (lagged) 0.0369* 0.0375* 0.0280** 0.0283** [0.02] [0.02] [0.01] [0.01] F&L ethnic fractionalization 0.248 0.378 0.138 0.0923 [0.42] [0.53] [0.31] [0.31] F&L religious fractionalization -0.83 -1.274** 0.0819 0.119 [0.58] [0.50] [0.39] [0.38] Climate-related disaster, binary -0.0801 -0.375 -0.520* [0.30] [0.24] [0.28] Drought 0.767** -0.19 0.319 [0.33] [0.38] [0.22] Flood 0.0831 0.0465 0.173 [0.18] [0.22] [0.12] F&L oil exporter 0.573** 0.582*** [0.22] [0.22] Sahel -0.834 -0.793 [1.20] [1.28] Sahel* F&L per capita income (lagged) -0.0936 -0.0603 [0.46] [0.45] Sahel* F&L log population (lagged) 0.11 0.106 [0.16] [0.17] 9 Sahel* F&L log (% mountainous) 0.103 0.104 [0.12] [0.12] Sahel* F&L oil exporter -1.354*** -1.406*** [0.38] [0.38] Sahel* F&L instability (lagged) 0.313 0.292 [0.24] [0.24] F&L democracy (lagged) -0.0207 -0.0213 [0.02] [0.02] Sahel* F&L ethnic fractionalization 0.106 0.112 [0.57] [0.58] Sahel* F&L religious fractionalization -0.557 -0.522 [0.67] [0.66] Sahel* Drought 0.828** [0.47] Sahel* Conflict previous year -0.0947 -0.0455 [0.38] [0.38] Sahel*Flood 0.129 [0.23] Sahel* Climate-related disaster, binary 0.211 [0.30] Constant -3.927*** -3.898*** -2.939*** -2.909*** [1.10] [1.46] [0.55] [0.56] Observations 172 172 1671 1671 Standard errors in brackets; * p < 0.10, ** p < 0.05, *** p < 0.01 Source: Own Analysis; Dependent Variable: Conflict Onset (0/1) methodology as the one employed by Slettebak (2012). The results of the regression analysis presented in Table This study, however, isolates the Sahel region and then 1 reveal that proxy measures for climate change-relat- compares it with the rest of Africa. The models are ed disasters such as the occurrence of droughts have a then estimated using probit regression with the onset positive and statistically significant association with the of conflict as the dependent variable and then a variety onset of conflict in the Sahel region. The floods are also of explanatory variables related to climate change and positively correlated with conflict, but the estimate is conflict. The second approach is a structural equation statistically insignificant. Interestingly, in the first model model to discern the direct and indirect effects of specification, the binary measure for climate related di- climate related disasters on migration and on conflict.1 sasters is in the opposite expected sign and it is insignifi- cant, controlling for other factors. 1   For more information on the dataset and the regression models The variable that captures whether a country has expe- used, see Slettebak (2012). rienced conflict in the previous year was significant and 10 negative. This is a somewhat unexpected finding since Figures 4 and 5 present the average marginal effects of it is expected that the prior conflict would have some conflict by levels of drought and flood. The plot suggests degree of inertia into subsequent years. This finding, that the relationship between the average increase in however, is in line with previous research on this topic probability of outbreak of conflict and the presence or (see Fearon and Laitin, 2003). absence of drought is quite significant for countries like Mali, Ethiopia, and Niger (see Figure 4 and 5). The results of the second model specification, however, mask variance between countries in the Sahel region. After presenting the results for the countries in the Sahel region, it is important to look at whether they are different from the rest of Figure 4: Average Marginal Effects (Drought) Africa. Columns 3 and 4 in Table 1 present the regression for the whole African con- tinent and include interaction terms of the explanatory variables with the dummy variable for the Sahel. Given the way in which the regression model is specified, findings reveal whether impacts for the Sahel are different from the region as a whole. Overall, the analysis shows that the Sahel, on average, is not different from the rest of the region as a whole. It is interesting to note that the interaction term between the drought dummy and the Sahel dummy is statistically significant and positive, sug- gesting that conflict in the Sahel is condi- tioned by the presence of drought. Figures 6 and 7 present the plot of the Source: Own Analysis marginal probabilities of the interaction term. While the dummy for the Sahel and Figure 5: Average Marginal Effects (Floods) drought are not statistically significant in the main analysis, their interaction was sig- nificant. Figure 5 shows that the marginal effects for interaction term were larger for the Sahel countries than for the rest of the region. In Figure 6, the variables “Sahel” and “Flood” did not yield any statistical result in the main analysis nor did the in- teraction term show any statistical signifi- cance. However, the plot shows significant variation in the interaction term depending upon the level of flooding. The remaining variables in the regression are similar to those found in columns 1 and 2 of Table 1. The results presented above show that climate change-related disasters appear to have a significant correlation with the onset Source: Own Analysis of conflict in the Sahel region. While these 11 through climate change related disasters– Figure 6: Marginal Effects of Interaction Term (Sahel*Drought) and conflict. However, as mentioned previously, conflict can be a direct con- sequence of climate change and also an indirect consequence as a result of an increase in climate change-related migra- tion (Rubin, 2010). Understanding the direct and indirect effects of climate change on migration and conflict are important since the climate-security literature iden- tifies - albeit without much empirical support -climate change induced migration as a potential cause for conflict (Raleigh et. al, 2008). Therefore, this paper employs a structural equation modeling approach to ascertain whether climate change and climate change-related disasters are cor- related with migration and, migration, in Source: Own Analysis turn, is correlated with conflict. An advan- tage of structural equation models is that it permits one to estimate the direct, indirect Figure 7: Marginal Effects of Interaction Term (Sahel*Flood) and total effects of the independent vari- ables on the dependent variable. For the purposes of this analysis, the same dependent variables are employed as in the analyses before - the onset of conflict (0/1). Because it is unlikely that conflict occurs in the immediate aftermath of a climate change-related disaster, the dependent variable is advanced by one year so that the explanatory variables are lagged.2 The mi- gration variable employed for this analysis is the log of the total number of migrants by country of asylum (UNHCR, 2013). The remaining variables in the analysis are the same as those employed in previous analyses above. Source: Own Analysis The results of the structural equation model are presented in Table 2. The estimates yield results are consonant with previous research on climate interesting results.3 First, migration is statistically signif- change and conflict, these findings should be inter- preted as mere correlations and do not imply causation. More research is needed to be able to better identify the causal mechanism/s through which climate change 2   I also re-estimated the model forwarding the dependent variable by 2, 3, 4, and five years. However, after the second year forward the model causes conflict. became really unstable and the sample became really small to draw any meaningful inferences. The preliminary analysis presented above only offers 3   Only those measure related to climate change and conflict are estimates on the correlation between climate change– discussed. 12 icant and positively associated with higher Table 2 - Results of Structural Equation Modeling likelihood of conflict, albeit the effect is small (only at the 10% level of significance). Second, Conflict Onset Migration the measures for floods and drought generate VARIABLES (Forward 1 year) (Logged) somewhat surprising results as they carry the Migration (Logged) 0.0293* opposite expected sign, however they are not statistically significant. These results might be (0.0168) capturing the fact that when intense flooding GDP per capita (lagged) -0.0545 occurs, people are forced to migrate, which is (0.0580) supported by the positive and statistically sig- Population (Logged) 0.0316** nificant coefficient of flood on migration. As expected, flooding does increase the likelihood (0.0143) of migration, controlling for other factors. The % Mountainous (Logged) 0.0199 positive correlation of the flooding variable in (0.0142) migration can be interpreted as both a direct Oil Exporter -0.139† impact of flooding on migration and an indirect impact of flooding on battle deaths via migra- (0.0386) tion. Political Instability -0.0165 (0.0187) The measure that captures whether the country had conflict in the previous year is not directly Polity2 0.00439 associated with a higher number of battle (0.00369) deaths but its effect is mediated by the migra- Ethnic Fractionalization 0.0218 tion variable. Moreover, the effect of previous (0.0751) conflict on migration is actually quite large and statistically very strong. Religious Fractionalization -0.00781 0.272** (0.0986) (0.132) In sum, structural equation modeling allowed Flood -0.00900 -0.380 to identify the causal paths between climate change, migration, and conflict.4 The results of (0.0157) (0.235) the analysis yielded interesting findings sup- Drought -0.00450 2.595† porting the hypothesis found in the literature (0.0243) (0.713) that climate change, per se, does not cause Conflict Previous Year -0.162† 8.767† conflict directly, but rather through a variety of channels such as migration (Rubin, 2010). (0.0578) (0.621) More research, however, is needed in order to Constant -0.417* 8.767† better understand the mechanisms and paths (0.216) (0.621) through which climate change might increase Observations 694 694 the risk of conflict. Robust standard errors in parentheses; † p<0.01, ** p<0.05, * p<0.1 Source: Own Analysis 4   The term causal path does not indicate a causality Dependent Variable: Conflict Onset (0/1) mechanism. Dependent Variable: Total Migration by country of asylum (logged) 13 3. Climate Change Implications for Youth A s mentioned in the previous section, the effects of climate change and climate-related disaster will likely affect the poorest and most vulner- able. Increasing climate variability and climate change- livestock. Furthermore, increasing population growth will compound challenges associated with diminish- ing natural resources in this region. Population data on youth 15 to 29 years of age showed that, between related events will further erode their already stretched 2000 and 2012, the average size of the youth cohort capacity to cope in addition to having insufficient assets was 26.8% in the Sahel region, compared to 28.5% in for adapting to the rapidly changing circumstances the rest of Africa. By 2020, the Sahel’s size of the youth (Verner, 2010). cohort is expected to increase to 28.1%, compared to a decline to 28.2% for the rest of Africa. One of the population cohorts that will be affected the most is the youth. In particular, youths who live in rural Around 70% of youths in Africa live in rural areas and areas will be hard hit by the progressive depletion of their livelihood depends entirely on agricultural pro- natural resources as a result of climate change, as their duction (UNESA, 2010; World Bank, 2009). In the livelihood is dependent on agricultural production and Sahel region, youth live in rural areas as part of pastoral- Figure 8: Number of vulnerable people per country (UNOCHA, 2012) Source: UN OCHA, 2012 14 Table 3 - Direct, Indirect, and Total Effects of Explanatory Variables on Conflict Onset and Migration Direct Effects Indirect Effects Total Effects Onset of Conflict Onset of Onset of Conflict Migration Migration Migration Conflict (Forward 1 year) (Logged) (Forward 1 year) (Logged) (Forward 1 year) (Logged) Migration (Logged) 0.029** 0.029* GDP per capita (lagged) -0.05 -0.05 Population (Logged) 0.031** 0.031** % Mountainous (Logged) 0.02 0.02 Oil Exporter -0.138† -0.138† Political Instability -0.02 -0.02 Polity2 0.00 0.00 Ethnic Fractionalization 0.02 0.02 Religious Fractionalization -0.01 0.272** 0.01 -0.01 Flood -0.01 0.272** 0.01 0.00 0.272** Drought 0.00 -0.38 -0.01 -0.02 -0.38 Conflict Previous Year -0.162† 2.595† 0.075** -0.086† 2.595† Source: Own Analysis †<0.01 ** <0.05 * <0.10 levels of statistical significance ist communities5. These traditional communities pass of extreme events further isolated such populations on their knowledge from one generation to the next. (UNESA, 2010). Further exacerbating the situation is Pastoralists generally have extensive historical knowl- the issue of weak governance, which is rampant in this edge about climate variability and how to cope with it. region.6 For example, the World Bank Governance Indi- However, the intensity, frequency, and erratic nature of cators show that the countries of the Sahel region have climate change events pose a threat to this knowledge. some of the lowest average institutional governance For example, research in Latin America suggests that scores in the world (World Bank, 2013). indigenous peoples tend to find fault in their shortcom- ings in the way they dealt with land and resources for the The growing youth bulge in Africa, and in the Sahel current climate change and climate variability related region in particular, presents a serious challenge as issues (Verner, 2010). Thus, the increasing pressure of well as an opportunity for policymakers in the region. climate change is compromising the future of pastoral- Governments in the region should, therefore, devise ist communities if agriculture and livestock continue policies to ensure that the growing youth cohort can be to be the sole sources of employment and livelihood in a productive asset for the countries in the region instead this region (UNESA, 2010). In addition to dealing with the stressors related to climate change, lack of govern- ment resources to assist these communities in periods 6   For example, if we analyze at the World Bank Governance Indicators countries that form the Sahel region have some of the lowest average institutional governance scores in the world. For example, the world’s average for the governance indicators for the 1996 to 2011 period stands at 5   20% of the whole population in the Sahel are pastoralists (www. 0 whereas the average for the countries in the Sahel region is -2.1 on a scale sahel.org.uk) in which the lower bound of the indicator is -2.5. 15 of a potential risk factor for conflict and crime. with formal means of employment, they will likely find themselves increasingly recruited by organized crime As a result, areas vulnerable to climate change in the groups to make a living (Ohlsson, 2000). Sahel region lack even the most basic governmental presence in terms of infrastructure, education, and Table 4 presents the results of the regression analysis social protection programs, which will impact the de- after controlling for the share of youths 15 to 29 years velopment–both economic and social–of these areas old over the total population. In the Sahel countries, (Gubbels, 2011). the analysis indicates that the higher the percentage of youths in the region the more likely countries are to In light of the situation described above, it is important engage in conflict, controlling for other factors. In other to look at how climate change can affect the youth. As words, an increase in one percent in the share of the it has been argued previously, migration might be an in- youth population increases the risk of conflict, albeit creasingly used mechanism to adapt to climate change marginally, by 29 percent, all else equal. This result is (Brown, 2007). Data show that the lack of infrastructure, in line with other studies where the size of the youth employment, and social programs has led to massive mi- population is associated with a higher probability of gration in the region (ibid.). In fact, the traditional tem- conflict (Beehner, 2007). This result also suggests that porary migration southward into main urban centers given the climate change and youth population projec- has been replaced by a more permanent migration into tions for the region, the risk of conflict and onset in the urban centers. A compounding, and more serious issue, region is likely to increase in the future. is the risk of conflict and criminality that increases as a result of youth migration into urban areas. Whether As in the previous regression analyses, drought is statis- or not conflict and crime occur will depend upon a complex interplay of socio-economic factors at the des- Table 4 - Estimation Results: Probit Regression tination location. But, these socio-economic risk factors for the Sahel Region are already present. Chief among them, (i) youth unem- ployment is already among the highest in the world in VARIABLES Coeffcient (Std. Err.) many countries of the Sahel (United Nations, 2013); Conflict previous year -1.025* (0.473) (ii) climate change and induced migration may exacer- bate existing social and ethnic inequities; and (iii) weak F&L per capita income -1.293* (0.606) (lagged) infrastructure in urban areas in these countries lacks the capacity to cope with the rapid population growth, F&L log population -0.070 (0.158) (lagged) scarce resources, and limited employment opportuni- ties will increase the number and frequency of disputes F&L log (% mountainous) 0.308* (0.145) over them (Boano, Zetter and Morris, 2008). F&L oil exporter 0.122 (0.492) F&L instability (lagged) 0.082 (0.322) Evidently, this complex array of risk factors will likely trigger conflict before long and is already pushing many F&L democracy (lagged) 0.018 (0.019) youths into organized crime. There is increasingly con- F&L ethnic fractionaliza- 1.843* (0.769) vincing evidence that suggest many of these unem- tion ployed youth are targets of the increasing activities of F&L religious fractionaliza- -1.511** (0.325) Al-Qaeda in the Maghreb ( Jamestown Foundation, tion 2011). The Sahel serves as a transshipment corridor for Drought 0.641† (0.339) drug trafficking, human smuggling and other organized Flood 0.011 (0.183) crime activities (Werz and Conley, 2012). As terrorism and organized crime activities in the region continue to Youth ( %15-29) 0.292* (0.138) proliferate7 and disaffected youths continue to struggle Constant -6.450** (2.327) Observations 694 694 Source: Own Analysis; 7   Since 9/11, the number of terrorist attacks in this region have increased ten-fold from 21 in 2001 to more than 200 in 2009 (Alexander, †<0.01 ** <0.05 * <0.10 levels of statistical significance 2012). 16 tically significant and positively associated with a higher ity beyond the life of the project and youths are not able likelihood of conflict, controlling for other factors. Con- to capitalize on the skills learned to secure longer-term versely, floods do not show any statistically significant jobs, which, in turn, might increase the likelihood of effect on the onset of conflict. The results of this analysis, turning back to crime. therefore, clearly suggest that the size of the youth pop- ulation matters for the onset of conflict in this region. Most evaluations of these programs have been con- ducted in developed countries and very few in low and Figure 8 presents the graphs of the size of the youth middle-income countries. Evaluations in developed population cohort in the Sahel, which is very significant countries show limited impact: only 15 percent of youth in countries like Mali or Eritrea (see Annex A). employment programs in low and middle-income Heightened risk of conflict and crime for youths in the countries have evaluated the effects of these programs’ Sahel will be elevated directly through climate change interventions on beneficiaries’ outcomes.8 and its deleterious effects on natural resources; and in- The available evidence on actual program impact shows directly through the compounded effects of migrating that learning technical skills alone does not appear to to urban areas. The onset of conflict and crime will be have a significant effect in maintaining a job over the first mitigated through effective climate change adapta- long term. However, one aspect that, coupled with tion policies. However, doing so may prove challenging learning technical skills, seems to work in the long-term due, first, to weak governmental capacity to provide the is life skills. Programs that use a three-prong approach tools and mechanisms to the most vulnerable commu- that includes life skills, vocational skills, and job attach- nities to be able to increase their resilience; and second, ment appear to have a more sustained impact even in the due to permanent migration flows into urban areas, absence of employment. For example, a recent impact which are already placing additional stresses on the evaluation conducted by the Inter-American Devel- already scarce resources and weak institutional capacity opment Bank (IADB) on an employment program of local governments. In addition, government capacity for youths found that the life skills component of the must be improved to respond quickly to mitigate the in- program had a significant impact on the beneficiaries’ fluence of climate change factors and socio-economic employability and ability to maintain a job (Ibarrarán factors. et al., 2012). Thus, it appears that while youth employ- Conflict and crime prevention strategies that target ment programs in isolation have limited impact, the youths, in particular, should be prioritized by policy- inclusion of life skills modules into these programs might have positive effects in both the short term and makers. Strategies for youth can take many forms, but long term. More research, however, is needed in order some of the most commonly utilized programs are those to understand the mechanisms through which life skills related to employment since unemployment appears to training help youths enter the labor market. be one of the major risk factors for youths engaging in conflict and/or crime (Ihlanfeldt, 2006). In fact, for the Finally, new and innovative approaches, such as youth majority of youth who lack a university degree, finding micro-finance programs are being implemented in an employment in the formal sector is virtually impossible. effort to address issues of unemployment for disaffect- The vast majority of youths that come from pastoralist ed youth. For example, Yemen, with the assistance of communities can barely read or write, thus, making it Qatar, is currently implementing a youth-finance facility even more difficult to enter the formal labor market if that provides financial services (micro-loans) to young they migrate to urban centers. Employment and public adults ages 18 to 30. The main objective of this facility– works programs are, thus, devised to provide a mean- Al-Amal microfinance Bank9 – is to provide youths with ingful source of income. However, these programs are access to fixed income loans from US$50 to US$100 to generally limited to several months/years and tied to a start a business or to expand a current business. specific project. Youths learn a set of skills in order to increase their employability and might find short-term employment during the life of the project. One of the 8   http://www.youth-employment-inventory.org major limitations encountered is the lack of sustainabil- 9   http://www.alamalbank.com/index.php?lang=en 17 4. Conclusion and Policy Options C limate change is perhaps the most important development challenge of the twenty-first century. The implications of climate change for the Sahel region are significant and likely to play out to ensure that this cohort acquires tools and mecha- nisms to adapt to a changing climate and prevent them from migrating to urban areas and increase the risk of engaging in conflict and crime? Because the con- in different ways. Two of the most important effects of sequences of climate change are transferred through climate change for the future of the Sahel are migration multiple channels, policies will have to be designed and, potentially, conflict (UNEP, 2011). This study has and implemented to address these issues on multiple attempted to establish the link between climate change, fronts. Based on the findings, this study has identified migration, and conflict with a particular regional focus - two potential key avenues through which policies can the Sahel region - and population focus - the youth. At be devised: (i) climate change adaptation measures and the regional level the results of the preliminary analyses (ii) conflict and crime prevention measures. These two reveal that climate change-related disasters have a strong areas are not exclusive and can benefit from each other. correlation with behavioral responses such as migration and conflict. At a population level, while the poorest 4.1 Climate Change Adaptation will be affected the most by the negative consequences of a changing climate, youths are the cohort that is likely Measures to be affected the most by a complex interplay of com- Building resilience to climate change by adapting to it is pounding factors related to climate change and socio- considered a good development policy (Verner, 2010). economic conditions (UNESA, 2010). It requires both national and local governments to The findings of this study, however, should be inter- devise coordinated strategies to effectively respond to preted with caution since they do not identify causality the needs of local communities. There are several policy between climate change and migration and/or conflict. areas through which climate change adaptation can be While the estimates from the regression analyses are implemented (Verner, 2010).10 These options target robust in the sense that climate change and climate three key areas of development: (a) Good governance; change-related disasters pose a serious risk for conflict (b) Social capital; and (c) Asset-based adaptation. in many vulnerable areas, they call for further research. i. Good Governance: Improving the institutional Most research on this topic has not concluded that capacity at the national and local government climate change alone causes conflict. The mechanisms levels is a key pillar for effective climate change through which conflict takes place are complex and adaptation policies. Government entities should varied and climate change is increasingly becoming an have adequate human, financial, and infrastructure important one. Additional research on this area should capacity to identify and respond to issues related aim at identifying the areas within the countries of this to climate change and make services accessible region that are most vulnerable to the negative con- to the most vulnerable population. Improving sequences of climate change. Also, given that climate government capacity, however, poses a serious change poses a serious risk to the onset of conflict, it challenge for many of the countries in the Sahel is also likely to put at risk gains made in post-conflict region. As mentioned above, institutions in the reconstruction/stabilization programs. In light of the serious situation that young people in the 10   For an extensive review of these areas, please refer to Verner Sahel are facing, what are the potential policy options (2010). 18 Sahel are weak and responses to climate-related policies to address the risks of conflict and crime require crises have already reduced governmental capacity adapting them to the local needs instead of employing (Kandji et. al, 2006). a “one size fits all” approach. There is significant het- erogeneity between countries in the Sahel that calls for ii. Social Capital: Social capital encompasses a tailored approaches. Potential policy areas could target broad range of behaviors and interactions between communities and individuals that ultimately (a) urban youth and (b) rural youth. Both are related influence the degree of their vulnerability to to providing employment for this cohort since it is one climate change. In the face of a rapidly changing of the main risk factors for conflict and crime (World climate and the increasing scarcity of resources, Bank, 2011). cooperation between communities is likely Providing skills and employment opportunities for to decrease. Programs at the community level unemployed urban youth is instrumental to prevent should aim at building resilience while restoring conflict and crime both in the short and long term. cooperative relations so that communities can Youths in urban areas are more likely to suffer from dis- be equipped to respond to climate change. affection as they might have already gone through the Well-designed community-driven development process of migrating to urban areas due seeking better programs (CDD) can contribute to empowering ways to earn a living. Programs should be available to communities and strengthening cooperative ensure that they earn basic skills and can enter the labor relationships to effectively adapt to climate change. market. However, in addition to providing basic voca- Some CDD programs have shown promising tional skills, employment and training programs should results and best practices should be adapted to the also incorporate life-skills training. local characteristics of the countries in the Sahel region (Verner, 2010). Rural youth might be inclined or forced to migrate as a iii. Asset-based Adaptation: The household assets result of climate change. In order to prevent this cohort of the poor are far more vulnerable to the negative from migrating to urban areas, it is important to provide consequences of climate change (Verner, 2010). It proper mechanisms for pastoralist activities to be pre- is in this context devising mechanisms to prevent served. Pastoralist communities have devised strategies an accelerated depletion of household assets - to adapt to climatic change for centuries. As the climate livestock, education, health, income - is crucial changes rapidly, governments in the region should to increase the resilience and adaptive capacity of ensure that policies and programs enhance the capacity vulnerable households to climate change. of these communities to adapt and possibly expand their activities. There is evidence that multi-level gov- 4.2 Conflict and Crime Prevention ernment coordination - from national to local - coupled Measures with land-use management and land-use regulations and investment in infrastructure could contribute for Preventing conflict and crime is also good development the sustainability of pastoralist communities and ex- policy. There is ample evidence that crime and conflict pansion crop-livestock systems (Izcowizc et Al, 2012). disrupt economic development, social resilience, and As a result, youths would have access to employment democratic stability (World Bank, 2011). Devising opportunities within their communities. 19 Annex Figure 9: Youth Population Cohort in the Sahel Countries Source: Own Analysis 20 References Adano, W. R., and Daudi, F. (2012). Links between climate change, conflict and governance in Africa. Pretoria, South Africa: Institute for Security Studies. Alexander, Y. (2011). The Consequences of Terrorism–An Update on al-Qaeda and other Terrorist Threats in the Sahel and Maghreb. Potomac Institute for Terrorism Studies. Barnett, J. and W Neil Adger. 2007.”Climate Change, Human Security and Violent Conflict.” Political geography 26(6):639-55. Boano, C., Zetter, R. and T.Morris. 2008.”Environmentally Displaced People.” Forced Migration Policy Briefing 1. Boyd, E., Cornforth, R. J., Lamb, P. J., Tarhule, A., Lélé, M. I., & Brouder, A. (2013). Building resilience to face recur- ring environmental crisis in African Sahel. Nature climate change, 3(7), 631-637. Brown, Oli. 2007. “Climate Change and Forced Migration: Observations, Projections and Implications.” Vol.: Human Development Report Office (HDRO), United Nations Development Programme (UNDP). ECOWAS, Sahel and West Africa Club & Organization for Economic Co-operation and Development. 2006. “The ecologically vulnerable zones of the Sahelian countries.” Atlas on Regional Integration in West Africa. Economic Community of West African States. Fearon, James D and David D Laitin. 2003.”Ethnicity, Insurgency, and Civil War.” American political science review 97(1):75-90. Gubbels, Peter. 2011. “Escaping the Hunger Cycle: Pathways to Resilience in the Sahel.” Homer-Dixon, Thomas F. 1994.”Environmental Scarcities and Violent Conflict: Evidence from Cases.” Internation- al security 19(1):5-40. Ibarraran, Pablo, Laura Ripani, Bibiana Taboada, Juan Miguel Villa and Brigida Gar-cia. 2012. “Life Skills, Employ- ability and Training for Disadvantaged Youth: Evidence from a Randomized Evaluation Design.” Ickowicz, A., Ancey, V., Corniaux, C., Duteurtre, G., Poccard-Chappuis, R., Touré, I., and A. Wane (2012, April). Crop–livestock production systems in the Sahel–increasing resilience for adaptation to climate change and preserving food security. In Building resilience for adaptation to climate change in the agriculture sector. Pro- ceedings of a Joint FAO/OECD Workshop, Rome, Italy, 23-24 April 2012. (pp. 261-294). Food and Agricul- ture Organization of the United Nations (FAO). Ihlanfeldt, K. R. 2006. Neighborhood Crime and Young Males’ Job Opportunity. Journal of Law and Economics, 49(1), 249-283. IPCC. 2007. “Climate Change 2007 – The Fourth Assessment Report of the Intergovernmental Panel on Climate Change,” Cambridge University Press. Cambridge, UK. Kandji, Serigne Tacko, Louis Verchot and Jens Mackensen. 2006. Climate Change and Variability in the Sahel Region: Impacts and Adaptation Strategies in the Agricultural Sector: World Agroforestry Centre. 21 Magadza, CHD. 2000.”Climate Change Impacts and Human Settlements in Africa: Prospects for Adaptation.” Envi- ronmental Monitoring and Assessment 61(1):193-205. NASA 2012. The Current and Future Consequences of Global Change. Ohlsson, L. 2000. Livelihood conflicts: Linking poverty and environment as causes of conflict. Stockholm: Environmental Policy Unit, Swedish Inter- national Development Cooperation Agency. Raleigh, Clionadh and Henrik Urdal. 2007. “Climate Change, Environmental Degradation and Armed Conflict.” Political Geography 26(6):674-94. Raleigh, Clionadh, Lisa Jordan and Idean Salehyan. 2008. “Assessing the Impact of Climate Change on Migration and Conflict,” in World Bank Seminar on Exploring the Social Dimensions of Climate Change. Reuveny, Rafael. 2007. “Climate Change-Induced Migration and Violent Conflict.” Political Geography 26(6):656-73. Shapland, P., Prihodko, L., and, & Hanan, N. 2013. Climate Change, Pastoral Resources and Livestock in the Sahel. Re- trieved from http://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=3&ved=0CD8QFjAC &url=http%3A%2F%2Flcccrsp.org%2Fwp-content%2Fuploads%2F2013%2F06%2FRB-12-2013.pdf&ei=E djMUbiXEuz_4APM9IHQCw&usg=AFQjCNHGwka1VZP3cqDsAra5KFg5-b27ig&bvm=bv.48572450,d. dmg&cad=rja Slettebak, Rune T. 2012. ”Don’t Blame the Weather! Climate-Related Natural Disasters and Civil Conflict.” Journal of Peace Research 49(1):163-76. Tadesse, D. (2010). The impact of climate change in Africa: Institute for Security Studies. UN. 2013. Sahel Regional Strategy 2013. United Nations. UNEP. 2011. “Livelihood Security: Climate Change, Migration and Conflict in the Sahel,” United Nations Environ- ment Programme. UNESA. 2010. “World Youth Report: Youth and Climate Change.” United Nations Economic and Social Affairs. Uppsala, University 2013. “Uppsala Conflict Data Program,” Uppsala, Sweden: Uppsala University. Verner, Dorte. 2010. Reducing Poverty, Protecting Livelihoods, and Building Assets in a Changing Climate: Social Implications of Climate Change in Latin America and the Caribbean: World Bank Publications. World Bank. 2009. Africa Development Indicators 2008/2009: Youth and Employment in Africa: The Potential, the Problem, the Promise. Washington, D.C.: The World Bank. World Bank. 2013. World Bank Governance Indicators. Available at: http://info.worldbank.org/governance/wgi/ sc_country.asp 22 Transitional Demobilization and Reintegration Program World Bank 1818 H Street NW Washington, DC 20433 USA www.TDRP.net 23