Will Iraq Become a Democracy

Will Iraq Become a Democracy? Will Iraq Become a Democracy? ABSTRACT One of the stated goals of U.S. intervention in Iraq has been to effect “regime change” and, ultimately, to bring about the democratization of that country. We argue that the second of these goals is much less plausible than the first. Given the factors associated with democracy, Iraq is unlikely to consolidate a free system of government. Even with substantial U.S.-imposed reforms, the most likely outcome is an autocratic or semi-autocratic regime. 1 INTRODUCTION The nation of Iraq, with its proud heritage, abundant resources, and skilled and educated people, is fully capable of moving toward democracy and living in freedom. (President George W. Bush, cited in The Economist, 2003) For at least five years, the democratization of Iraq has been a stated goal of U.S. policy. According to the 1998 Iraq Liberation Act: “It should be the policy of the United States to support efforts to remove the regime headed by Saddam Hussein from power in Iraq and to promote the emergence of a democratic government in place of that regime” (LaRocque, 2002). In the run-up to its invasion of Iraq, the administration of George W. Bush made “regime change” a central feature of its policy toward that country. In many respects, the Administration’s hopes for a post-Saddam political system have been as lofty as those expressed by Congress. U.S. Secretary of State Colin Powell recently looked forward “to the day when a democratic, representative government at peace with its neighbors leads Iraq to rejoin the family of nations” (Associated Press, 2003a; see also Jehl and Sanger, 2003; Wolffe and Hirsch, 2003: 25) and President Bush personally expressed the belief that democracy could flourish in Iraq in the wake of a U.S. invasion (The Economist, 2003; Associated Press, 2003b; New York Times, 2003). Even more ambitiously, other Administration officials have suggested that U.S. intervention in Iraq would trigger reform across the Arab and Muslim world, with a newly democratic Iraq serving as a model for other countries in the region (Ford, 2003; Associated Press, 2003a; Tolan and Felch, 2002; Rushdie, 2002; McGeary, 2003; Will, 2002; Mohan, 2003). Arguments about the potential democratization of Iraq have been accompanied by references to the postwar reconstruction of Germany and Japan, which occupation forces effectively remade into liberal-democratic allies of the United States. Skeptics have drawn parallels to recent failures of regime change or nation-building in Afghanistan, Bosnia- 2 Herzegovina, Haiti, Kosovo, Somalia, and elsewhere (McGeary, 2003; Ford 2003). Other analysts have staked out something of a middle ground. For instance, columnist Fareed Zakaria (2003) argued that “no matter what comes after Saddam, it will be better than his totalitarian regime” – a less ambitious claim than the notion that Iraq will rapidly democratize, but one that effectively rules out the possibility of protracted civil conflict or the establishment of a new dictatorship. How realistic is the notion that Iraq will become a democracy following Anglo-American occupation? What sort of regime is likely to emerge in that country? Although predicting the outcome of any one political transition is difficult, scholars now know enough about the causes of democracy to shed some light on these questions. Political scientists too often miss the opportunity to use their expertise to address questions of real world import. Although we have a much better understanding of how democracies emerge and survive than we had only a few decades ago – based on numerous case studies, large-N statistical analyses, and theoretical research on constitutional design – attempts to put this new knowledge to practical use remain relatively rare. This article builds on existing scholarship to construct a model of democracy that can inform our understanding of political development in Iraq. The first section of this article briefly reviews scholarly research on the causes of democracy and develops a simple statistical model. The second section summarizes the predictions of this model for Iraq, discussing how its general results might or might not translate to that particular case. The third section examines recent democratic “success stories” – i.e., countries that appear to lack the requisite conditions for political freedom – in search of hopeful 3 lessons for Iraq. The final section discusses how Anglo-American occupation might be expected to affect Iraq’s democratic prospects. We conclude that Iraq lacks many of the success factors for democracy. Without extensive foreign-imposed reforms, the most likely outcome is an autocratic regime – less brutal than Saddam Hussein’s, but still undemocratic. Given the likely duration of U.S. involvement, the consolidation of a semi-autocratic system seems a more plausible goal than democratization. THE CAUSES OF DEMOCRACY It is possible – just possible – that Iraq could gradually develop into a democracy, but the task is huge and the odds are long against it…[T]he social, economic, and political conditions for establishing democracy in Iraq are far from favorable. (Diamond, 2003) After several decades of intense study, scholars agree that there is no single cause of democratization or path to democracy (Huntington, 1991: 37-8; see also Rustow, 1970: 337; Dahl, 1971; O’Donnell et al., 1986). Some countries have become democracies without many conventional “prerequisites” (e.g., India in 1947 and South Africa in the 1990s). Meanwhile, other countries have repeatedly failed to consolidate democratic systems of government despite apparently favorable circumstances (e.g., Argentina during most of the twentieth century). Nevertheless, most analysts recognize that certain factors significantly enhance or reduce the likelihood that a country will maintain democratic institutions. These include economic development, the nature and type of social cleavages, population, geography, and various cultural or historical factors (such as religion, region, colonial heritage, and past experience with democratic rule). Different combinations of these factors dispose countries to be more or less 4 democratic (Dahl, 1971; O’Donnell et al., 1986; Diamond et al., 1990; Huntington, 1984; Huntington, 1991). In order to assess the likelihood of democracy in post-occupation Iraq, we begin by analyzing the general causes of democracy around the world. Although global patterns do not necessarily apply in the same fashion to every nation, they can offer insights into the challenges that Iraq is likely to face – as well as Iraq’s odds of surmounting these challenges. By specifying and measuring the factors most commonly associated with democracy around the world, we attempt to conduct this exercise in a reasonably systematically fashion. Our first step is to compare levels of democracy across different countries. We do so using the combined Freedom House scores of civil and political rights, averaged for each country in the world from 1996-2000. To make these scores more intuitive, we rescale them so that they range from 0 to 12, with higher values indicating more democracy.1 The Freedom House scores are appealing because they are of recent vintage and because they comprise various distinct elements of liberal-democratic governance (electoral competition, freedom of the press, etc.). Most importantly, these measures capture the fact that many countries are something other than fully democratic or utterly totalitarian. They thus allow us say not only whether a country is democratic according to some arbitrary cut-off, but also how close to that cut-off it may be.2 By averaging Freedom House scores for each country across several years, we hope to minimize the “noise” in our data in two ways. First, we avoid giving too much weight to coding flukes in particular years. Although the Freedom House scores are quite consistent over time, they are not perfectly so, and using averages helps to cancel out random errors in coding. Second, our five-year average offers a better measure of the medium-run potential for democracy 5 in a given country than data for one particular year. Democratic countries may experience spasms of repression (e.g., England in the wake of terrorist attacks by the Irish Republican Army), and countries destined for autocracy may enjoy a brief period of political liberalization or contestation (e.g., Belarus in the early 1990s). We are less interested in these political ephemera than we are in underlying trends and tendencies. In other words, we do not seek to predict whether Iraq may enjoy one or two years of relative freedom before lapsing back into despotism; rather, we aim to understand how much political liberty is likely to survive in the long run after occupation forces withdraw.3 In analyzing what causes differences in levels of democracy around the world, we focus on the following factors. A. Level of economic development: One of the most robust findings from decades of scholarly research is that wealthier countries tend to be more democratic. Econometric analysis has documented that this relationship reflects impact of economic development on regime type, rather than the other way around (see Przeworski et al., 2000). Economic development tends to increase the size of the middle class, to generate cross-cutting cleavages in the population, to produce multiple independent power centers, to encourage civil society by freeing up time for civic activities, to facilitate popular organization by increasing access to mass communication, to provide the market base for independent media, and to reduce the opportunity cost for rulers of losing power (Lipset, 1981 [1961]; Cutright, 1963; Diamond, 1992; Przeworski and Limongi, 1997; Przeworski et al., 2000; Bates et al., 1998; Lawson, 2002). We consider per capita gross domestic product in dollarized purchasing power parity terms as our principal measure of 6 economic development (World Bank, various; UNDP, various; CIA, various).4 Following other scholars, we use the log of this measure in our multivariate analysis. B. Literacy and education: In general, democracy is more likely to emerge and endure where people have achieved at least a basic level of cognitive sophistication. More educated populations are presumably more informed and tolerant – less susceptible to crude propaganda or extremist ideologies and better prepared to act as citizens. Over and above the impact of per capita income, then, education levels may exercise an additional influence on democracy. Because average education levels and literacy rates track rather closely, either measurement would work for our purposes; we use literacy because it offers slightly better reach and greater comparability across countries (UNDP, various; CIA, various).5 C. Energy exporters: Not all types of economic development are equally likely to lead to democracy (Rueschemeyer et al., 1992). In particular, countries whose economies depend on the exploitation of one or two minerals, especially oil, are thought to face special obstacles in consolidating stable, democratic systems. Petro-states are typically characterized by massive corruption, fiscal profligacy, state weakness, and vicious zero-sum competition for control over oil revenues (Karl, 1997). As a rough measure of these effects, we include a dummy variable for those countries that are net energy exporters (forty-five in all).6 D. Social cleavages: Countries deeply divided along racial, ethnic, linguistic, religious, or class lines are thought to be less likely to establish and maintain democratic rule (Dahl, 1971; Horowitz, 2001; Lijphart, 1968; Lijphart, 1977; Linz and Stepan, 1996). To capture the effects 7 of class cleavages, we include the Gini index of income inequality – a standard measure of income distribution. This measure may also be thought of as a rough indicator of the size of the middle class. For other social cleavages, we use Alesina et al.’s (2003) indices of ethnic, linguistic, and religious fractionalization, which offer the broadest coverage. In our multivariate analysis, we also include the square of each variable; this combination of indicators allows us to take into account the possibility that extreme ethnic heterogeneity, as well as extreme homogeneity, may be associated with democracy. E. Past experience with democracy: It may be more difficult to construct democratic institutions from scratch than to resuscitate previous ones. In countries where democracy has never been put into practice, political actors may be unfamiliar with or even hostile to democratic norms and paradigms. By contrast, countries with a venerable democratic tradition may be loath to abandon it and eager to revive it if democracy has collapsed (Machiavelli, 1979 [1531]: 21828; Stepan, 1986: 64-84; Rohrschneider, 2000). To measure past experience with democracy, we consider the total number of years a given country was democratic from 1900 to 1995.7 F. Colonial legacies: Those countries colonized by England are sometimes thought to be more likely to establish and maintain democratic systems of government (Przeworski et al., 2000). English legal traditions, politically neutral bureaucracies, Westminster-style parliamentary electoral systems, the absence of a centralizing colonial state, lack of intermarriage between colonists and the native population, and a tradition of civilian control over the military may all play a role in the political success of former British colonies. We use a simple binary measure of 8 British colonialism – 1 for countries that were once British colonies, 0 otherwise (Treisman, 2000; CIA, various). G. Religion and culture: Countries that share a Western cultural tradition – including Western Europe, the Anglophone immigrant countries, and to a lesser extent the former Soviet Bloc and Latin America – tend to be more democratic (Huntington, 1991; Diamond, 1999: 161-217; Almond and Verba, 1963; Putnam, 1993; Inglehart, 2002). By contrast, a number of scholars have argued that Islam is antithetical to democracy (Fish, 2002; Norris and Inglehart, 2003). To capture broad cultural and religious influences, we divide the world into eight regions: the OECD countries, non-OECD European countries, Latin America and the Caribbean, Africa, the Middle East, South Asia, and East Asia. We also employ a separate measure of the percentage of the population that is Muslim (La Porta et al., 1999).8 H. Neighborhood effects: Countries in highly democratic regions may face political, economic, and even military pressures to democratize, whereas those in predominantly autocratic regions may face weaker or even countervailing pressures (Whitehead, 1996; Huntington, 1991: 85-108; Diamond, 1999; Child, 2000; Diamond et al., 1990: 31-34; Pridham, 1991). We expect our regional variables to capture many of these influences, in addition to the effects of culture. I. Size and geography: Democracy is thought to be easier in smaller societies, where citizens can better monitor government officials. Following other scholars, we consider population as our principal measure of country size (Dahl and Tufte, 1973; Diamond with Tsalik, 1999). In our multivariate analysis, we use the log of this variable. Small island nations may have an 9 additional advantage over other countries for a number of reasons (Diamond with Tsalik, 1999). We thus also include a simple dichotomous measure of whether a country is an island: one if yes, zero if no. J. War: Violence and warfare may have a destructive impact on political development, corroding civil society and encouraging authoritarian policies (Whitehead, 1996; Walker and Armony, 2000). To capture some of these effects, we consider whether a country experienced a major international war in the most recent decade preceding our 1996-2000 time period (i.e., the 1980s), using a variable coded as one for war during the decade, zero otherwise (Sivard, 1993 from Easterly and Levine, 1997). One quick way to evaluate Iraq’s prospects is to see how it fares on each of these dimensions. With this in mind, the first two columns of Table 1 lists the factors mentioned above and Iraq’s score on each. The third column then presents the set of countries that score closest to Iraq each factor, and the final column presents the average democracy score for that peer group. In most cases, these scores do not augur well for democracy in Iraq. For instance, countries with Iraq’s level of economic development – as measured by per capita income and literacy – are largely undemocratic. So are most other countries in Iraq’s geographic and cultural “neighborhoods.” Countries in Iraq’s peer group range from an average of 2.6 to 6.9 on our democracy scale, depending on which factor is considered. [Table 1 about here] It is possible, of course, to consider the effects of all of these variables operating in concert rather than singly, and thus to assess the likelihood of democracy in a country that fits Iraq’s overall profile. Constructing even a primitive such model necessarily entails a series of 10 methodological decisions. In order to keep matters as simple as possible, we use ordinary least squares (OLS) regression – a standard statistical tool for such analyses. We thus assume that the variables in our model are causal of democracy rather than caused by it, an assumption that we believe is reasonable for the bulk of the variables we consider.9 Because we lack at least some information for many countries, we are faced with a choice of (1) limiting our analysis to a small number of cases for which all data is complete or (2) filling in the missing data in some way. The first alternative would effectively restrict us to developed world – a serious shortcoming in our eyes, given that we are most interested in the democratic prospects of a developing, non-western country. Such a strategy would lend little insight to the problem at hand and might well lead to biased predictions on many dimensions. Consequently, we prefer to impute missing values. In keeping with recent methodological innovations, we employ the multiple imputation method developed by Gary King and his colleagues (King et al., 2001; Honaker et al., 2001).10 Our regression model has a “kitchen sink” quality to it, and one could imagine a more parsimonious approach. Nevertheless, our goal here is not to present the most elegant explanatory framework possible, but rather to make the most accurate point prediction that we can. We thus prefer to err on the side of including a range of plausible variables, even if they do not all turn out to be statistically significant in the multivariate model (e.g., literacy, ethnic cleavages, and British colony). For the interested reader, Table 2 presents both the full model used to generate the predictions in this paper and a restricted model containing only those variables that attained or approached standard thresholds of statistical significance.11 (Our conclusions about Iraq do not change when we employ this model rather than the full one.) [Table 2 about here] 11 Despite the large number of variables that we do include, this model still leaves out certain factors that might well affect the chances of establishing or sustaining a democracy. For instance, we have not taken into account the effects of leadership, patterns of political recruitment, economic policies, particular constitutional arrangements, nuances of political culture, state capacity, specific elements of national development (e.g., the way in which manufacturing classes were incorporated into the political system) or more complicated questions of timing and sequencing (e.g., whether political competition between elites preceded mass enfranchisement). Nor have we measured the degree to which social divisions have actually been rendered politically salient or conflictual. Finally, some of our indicators are at best imperfect proxies of the factors they attempt to measure. Simply having been a British colony for a brief period of time, for instance, may say little about how thorough and transformative that experience was; similarly, the impact of international wars may vary substantially depending on the length of their duration and the damage that they caused. As a consequence of such limitations, we would not expect our model to perfectly predict levels of democracy. That said, the explanatory power of our model is fundamentally an empirical question: to the extent that this model accurately describes the world today, we may feel comfortable that we have not omitted factors which exercise a substantial and systematic impact on democratic development in the late twentieth century.12 Our admittedly inelegant model does a remarkably good job of predicting levels of democracy across all 186 countries included in the analysis.13 In statistical terms, the fit of the model is quite good, with adjusted R-squared around 0.68,14 meaning that we have explained slightly more than two-thirds of all the variation in democracy across different countries. Stated another way, the correlation between the level of political freedom in each country that we would 12 predict based on this model and the actual democracy scores for each country is approximately 84% (the square root of the unadjusted R-squared). For comparison’s sake, the correlation between literacy and the log of per capita income is 68%; even the correlation between two different measures of democracy itself – the combined civil rights and political rights scores from the Freedom House index against the combined autocracy and democracy scores from the Polity index over the last twenty years – is only 89%. Our analysis thus offers an extremely accurate picture of the political world today. Most of the findings from the model are as might be expected: several familiar factors exercise a statistically significant impact on political freedom in the expected direction. For instance, per capita income is positively associated with democracy, other factors held constant, as are past experience with democratic government and being an island nation. By contrast, more populous countries, energy exporters, and countries with large Muslim populations score significantly lower. Regionally, countries in East Asia tend to be less democratic, ceteris paribus, while OECD nations are more so. Being in the Middle East appears to be associated with lower levels of democracy, but this effect is not quite statistically significant once Islam and energy exports are taken into account. The impact of religious fractionalization is negative and very close to significant in the full model (p=0.11), while its square has a positive coefficient that is significant at the 0.10 level. This suggests that countries with very low and very high levels of religious fractionalization tend to be more democratic than countries with moderate levels of fractionalization (such as Iraq).15 Remaining variables – other social cleavages, literacy, British colonial history, and recent involvement in international war – do not appear to exercise a significant influence on democracy once other factors are taken into account. 13 Somewhat more surprising than the statistical significance of these coefficients is their relative magnitude. Moving from a per capita income of $2,500 to $8,000 – that is, from Iraq’s level of development now to the level it enjoyed just before the Iran-Iraq war – is worth slightly less than one point on the democracy scale.16 By contrast, being a net energy exporter typically lowers a country’s democracy score by approximately the same amount. Meanwhile, homogenously Moslem countries would be expected to score about three and a half points lower on average, controlling for other factors – an effect equivalent in magnitude (though opposite in direction) to having experienced approximately 100 years of democratic governance. What does this model foresee for Iraq? Unfortunately for Iraqi democrats, the prediction that emerges from our model is indistinguishable from Iraq’s actual score of zero. Using CLARIFY software to generate a series of 1000 simulations based on the full model presented in Table 2, we find an expected value for Iraq of 0.16 (King et al., 2000; Tomz et al., 2001). In other words, Iraq under Saddam Hussein was pretty much where we might have expected it to be given its demographic and economic profile. Left to its own devices, we certainly would not anticipate that Iraq would become a democracy following the removal of the Baathist regime. To further assess prospects for improvement in Iraq’s political development, we conducted a series of statistical simulations employing the results from our full model and different values for a handful of variables.17 Figure 1 highlights three possible scenarios, which we label Current, Realistic, and Optimistic.18 The Current scenario, shown in the solid line, uses Iraq’s actual values on all the variables listed in the full model in Table 2. The expected value for Iraq’s level of democracy, as noted above, is 0.16, with a 95% confidence interval of -1.2 to 1.6.19 If we assume that Iraq can reasonably easily double its per capita income under foreign occupation and assistance, the expected value of democracy rises, but only slightly, to 0.68 on 14 the twelve point scale. (This is shown in the dashed-line density distribution for the Realistic scenario.) In fact, the two confidence intervals for the Current and Realistic scenarios overlap so much that we can infer little or no likely increase in democracy from such a projected rise in living standards. [Figure 1 about here] Our Optimistic scenario assumes that post-war foreign occupation generates a four-fold increase in per capita income, effectively avoids the pitfalls for political development of reliance on oil, and gives Iraq the equivalent of ten years’ worth of experience with democracy. (Such results would exceed even sympathetic prognostications about the likely efficacy of postwar occupation.) Even under this scenario, the expected value of democracy in Iraq rises to just 2.4. Assuming our model and measurements are accurate, we can be 95% certain that the democracy score for a country with Iraq’s profile would fall somewhere in the range of 0.8 to 3.9. Even accounting for large shifts in several important variables, then, it seems unlikely that Iraq would achieve a high level of political freedom. IRAQI EXCEPTIONALISM? The internal life of Egypt is characterized by moderation, that of Syria by tensions, and that of Iraq by extremism. (Eliezer Berri, cited in Miller and Mylroie, 1990: 84) There are, of course, serious problems with extrapolating from a general model to a particular case. Not only are “point estimates” inherently less accurate than claims about statistical averages, such point predictions implicitly assume that each factor operates in the same way across all cases. Rather than mechanically apply general formulae to Iraq, then, we must delve more deeply into how different factors might shape political development in that country. 15 One way to do so is to review the most important variables discussed above in the specific context of Iraq. A. Per capita income: In contrast to most countries, living standards in Iraq have declined precipitously over the last two decades. If past high levels of development exercise an influence on how Iraqis think or act politically today, then present levels of development may be misleading indicators of Iraq’s democratic potential. In addition, national income is likely to increase rapidly once sanctions are lifted and oil production rises; as discussed above, this change should shift the odds in favor of democracy. However, as the scenarios described above indicate, these effects are not likely to be particularly large. One factor that these scenarios do not consider is the potentially legitimating impact of rapid economic recovery. If a new democratic regime were to preside over a period of rising living standards, Iraqis might become more supportive of democracy for purely instrumental reasons. When we consider Iraq’s level of economic development and its growth prospects more closely, therefore, we find reason to suspect that Iraq’s chances of democracy may be somewhat greater than we might have initially suspected. B. Social cleavages: Social cleavages are poor predictors of democracy because the relationship between demographic divisions and political outcomes is complex. To begin with, demographic divisions are not automatically seen as socially relevant; history and state policies matter tremendously in determining which differences among people will emerge as salient (Anderson, 183; Weber 1976; Laitin 1998; Nobles, 2000; Chandra, 2002). Second, and related, not all differences that are recognized as socially relevant become the basis for political mobilization or 16 conflict; historical legacies, electoral institutions, and government actions matter a great deal in determining how perceived social differences affect political attitudes and behavior (Fearon and Laitin, 1994; Laitin, 1998; Petersen, 2001; Horowitz, 2001; Varshney, 2002; Chandra, 2003; Petersen, forthcoming). Third, even when particular divisions become politicized, significant cross-cutting cleavages may mitigate the impact of any one division: many different constituencies may exist, such that no single bloc constitutes a permanent majority or minority (Lipset, 1981 [1961]; Dahl 1971). Fourth, even when a single immutable cleavage has become the focus of political competition, creating a permanent majority and minority (or minorities), clever constitutional engineers can often craft institutions that dampen or contain the resulting tensions (Lijphart, 1968; Lijphart, 1977; Horowitz, 1991; Lijphart and Waisman, 1996; Ordeshook and Shvetsova, 1996). Finally, international actors concerned about violence or democratic breakdown can sometimes offer rival ethnic groups guarantees that each might be unable to provide on their own. On average, therefore, the simple fact of having a particular demographic profile tells us relatively little about a country’s prospects for stable democracy. At the same time, it is easy to find countries where racial, religious, and ethnic divisions have led contributed to democratic breakdown or undermined attempts at political reform: Bosnia-Herzegovina, Cyprus, Fiji, India, Lebanon, Macedonia, Northern Ireland, Russia/Chechnya, Rwanda, South Africa, Sri Lanka, and the U.S. South for the century following the Civil War all come to mind. In all these cases, ethnic cleavages have exercised a much greater effect on prospects for stable, democratic governance than statistical averages would lead us to predict. One logical question, therefore, is whether the types of religious and linguistic divisions that characterize Iraq are more innocuous or more likely to result in political violence or 17 domination. On the negative side of the ledger, ethnic cleavages in Iraq have historically been the locus of massive bloodletting. It is well known that the Baathist regime, dominated by Sunni Arabs from one Tikrit-based tribe, slaughtered thousands of Shia Arabs and Kurds. But statesponsored ethnic conflict has long been an intimate part of Iraq’s history, including: the massacre of Assyrians in 1933, a pogrom against Iraqi Jews in 1941, communal violence in Mosul and Kirkuk in 1959, Baathist show trials of Jews and others in 1969, forced deportation of the Faylis (Shia Kurds) in 1971-72, forced deportations of “Iranian” Shiites in 1980, large-scale Shia rebellions in the 1990s, and several campaigns of varying brutality by Arab-led governments against the Kurds. In short, a range of social divisions have erupted into violence. As one expert described the turmoil in Mosul in 1959: For four days and four nights Kurds and Yezidis stood against Arabs; Assyrians and Aramean Christians against Arab Moslems; the Arab tribe of Albu Mutaiwat against the Arab tribe of Shammar; the Kurdish tribe of al-Gargariyyah against Arab Albu Mutaiwat; the peasants of Mosul against their landlords; the soldiers of the Fifth brigade against their officers; the periphery of the city of Mosul against its center; the plebeians of the Arab quarters of al-Makkawi and Wadi Hajar against the aristocrats of the Arab quarter of ad-Dawwash; and within the quarter of Bab al-Baid, the family of al-Rajabu against its traditional rivals, the Aghawat. It seemed as if all social cement dissolved and all political authority vanished. (Batatu, 1978: 866, cited in Makiya, 1998: 237-238) During his rule, Saddam Hussein made a systematic effort to play off different clan, tribal, and ethnic groups against each other, exacerbating such divisions. Consequently, a range of communal divisions in Iraq have become politically relevant and conflictual. In addition, Sunni Arabs in Iraq have recently experienced a status reversal (Petersen, forthcoming): they have gone from being the historically dominant ethnic group to a minority in a country that will likely be dominated by Shia Arabs. Finally, many households in Iraq today possess at least one firearm, increasing the potential for communal violence. None of this bodes well for democratic consolidation, or indeed for political stability of any kind. 18 On the other hand, Iraq’s communal divisions may be less likely to destabilize the country than conventional wisdom would have it. First, none of Iraq’s three major groups (Kurds, Sunni Arabs, and Shia Arabs) currently demands a separate national homeland. In this sense, divisions within Iraq are less problematic than in Kosovo, Sri Lanka, Kashmir, or Chechnya. Moreover, the group most likely to favor secession (the Kurds) has no external lobby state willing to support irredentist demands. There is thus no equivalent of a Serbia for Bosnian Serbs, nor even (as there was in the past) an Iranian regime ready to provide Iraqi Kurds with support for organized separatist rebellion. Second, clever constitutional design could substantially reduce the potential for ethnic conflict in Iraq. For instance, large multi-member districts whose boundaries matched those of the three main ethnic communities could encourage crosscutting cleavages exist within each ethnic group (Lawson, 2003; Chandra 2003; Ordeshook and Shvetsova, 1994). Both the numerically dominant Shia Arab population and the minority Sunni Arab population, for instance, are divided between fundamentalists and secularists, as well as along class, tribal, or regional lines. If institutions accentuate these divisions and moderate those of primary ethnic attachments, the prospects of any one group permanently dominating the political process would diminish. The management of communal divisions would be easier still if Iraq were to adopt a parliamentary form of government, rather than a winner-take-all presidentialist system (Dawisha and Dawisha, 2003). Finally, the relatively geographically segregated nature of Iraq’s three main populations means that federalist institutions could be used to mitigate inter-group tensions. In other words, Kurds in the north and even Shia Arabs in the south could be granted a much greater degree of self-rule than they have been accorded by previous regimes (Dawisha and 19 Dawisha, 2003). Although Iraq’s social cleavages are probably more explosive than similar divisions in most other countries, they may prove manageable. C. Islam: Related to the question of cleavages is the issue of religion in general. Despite the revival of certain elements of Islamic law over the last decade, Iraq boasts a well-established secular tradition. Until recently, fundamentalists never constituted a major political force in Iraq; secular pan-Arabists (including the Baath), Communists, Iraqi nationalists, and even liberal democrats commanded greater support than Islamists in the pre-Baath era. Today, the social status of women – one benchmark of traditionalist Islamic influence (Lewis, 2003) – is higher in Iraq than in many other Muslim countries.20 It may thus be that Iraq is more amenable to democracy than other Muslim countries (see Grant, 2002). On the other hand, the fact that Saddam Hussein’s regime was largely secular may have given fundamentalists greater legitimacy than they would enjoy otherwise. Indeed, Islamist movements in Iraq today seem relatively well organized compared to other Iraqi groups, and they may well become a potent political force. Likewise, there is little reason to believe that the putative negative impact of Islam on the creation of an autonomous civic sphere in Iraq will be radically different than elsewhere. Finally, Iraqis do not appear to see themselves as part of a Western cultural tradition, for whom democracy would be a logical political choice. It is thus not clear that, in political terms, Iraq should be though of as a non-Islamic society – i.e., that it is immune from the influences that affect other predominantly Muslim nations. D. Region: Similar issues emerge with regard to region. If Iraq’s political legacy and traditions have set it apart from the rest of the region, then this relatively modest neighborhood “penalty” 20 may not apply. Unfortunately, Iraq’s history of limited colonial rule, foreign-imposed monarchy, elite-dominated parliamentarism, various military coups and coup attempts, and protracted single-party rule cum personalistic dictatorship sounds depressingly familiar. Although Baathism has presumably been delegitimized among most sectors of the population, other forms of autocracy may not have been. It is thus far from clear that Iraq has already exhausted undemocratic political options and is poised to follow a radically different course than its neighbors. One potentially important difference from other countries in the region is the relatively large number of citizens – perhaps as many as one million – who have resided for years outside of Iraq. Many of these expatriates have lived in Iran or Arab states and are thus unlikely to have learned much about the practice of democracy.21 A number of exiles, however, have lived in Western democracies. If large numbers of skilled, democratic-minded expatriates return in force, they might diminish some of the typical influences of region. On the other hand, the new Iraq also confronts several negative regional influences that we have not taken into account. These include potential predation by Turkey in the north and support for anti-democratic forces by Iran in the Shia areas. In addition, events in Palestine could prove a source of mobilization by anti-democratic forces among Arab Iraqis – as they have in the past (Batatu, 1978; Makiya, 1998). Neighborhood effects that go beyond culture or tradition may thus exercise a negative influence on democratic consolidation. E. Experience with democracy: Unlike postwar Austria, Germany, Italy, or Japan, Iraq lacks any tradition of democracy. Taking into account this simple fact, however, does not fully capture the effects of Iraq’s political history. If past regimes were relatively liberal and open, 21 despite being autocratic, then democracy might have a better chance of taking hold. On the other hand, if past regimes were especially vicious and repressive, the odds of democratic consolidation would presumably be lower. Unfortunately for democrats, Baathist rule in Iraq was particularly brutal, even by the standards of modern dictatorship. Displays of dissent were viciously punished; large numbers of informers were recruited throughout the country; torture was used systematically as a means of political control; fear was pervasive; and civil society was dismantled. Even exiled dissidents were assassinated and their families in Iraq harassed or murdered. Countries with such horrific pasts face special challenges when they attempt to democratize (see Rohrschneider, 2000). As Iraqi exile Kanan Makiya put it in his aptly named study of the Baathist regime: For a quarter of a century, the polity has been built on distrust, suspicion, conspiritorialism, and betrayal, values with which it in turn has infected everyone. Every Iraqi today, whether in the opposition or outside it, carries the marks of that victimhood deep inside…Almost any post-Baathist future in Iraq is going to be like walking a tightrope, balancing the legitimate grievances of all those who have suffered against the knowledge that if everyone is held accountable who is in fact guilty, the country will also be torn apart. Iraq after Saddam is going to be a country in which justice is both the first thing that everybody wants and the most difficult thing for anyone to deliver. (1998: xxx-xxxii) As in eastern Germany, Romania, and elsewhere, the legacy of totalitarian rule could well undermine Iraq’s prospects for democracy to a degree that we have not yet taken into account. All told, closer examination of how our general model might apply to Iraq does not suggest that we have unfairly assessed that country’s prospects for democratic consolidation. When we consider Iraq’s level of development, its secular traditions, and the potential role of Iraqi expatriates, we find that we have probably underestimated those prospects. On the other 22 hand, when we consider Iraq’s legacy of ethnic conflict and totalitarianism, as well as various regional tensions, we are tempted to believe that we have overestimated them. The odds of Iraq becoming a democracy on its own thus continue to look poor. UNLIKELY SUCCESSES Even if general explanations for regime type would seem to apply to Iraq, there are always exceptions to these general trends. Countervailing factors that we did not consider in our quantitative analysis could allow countries to overcome what appear to be highly adverse circumstances. To the extent that political scientists have the obligation to be “possibilitists” as well as “probabilists” – as Philippe C. Schmitter has argued – it is worth considering these unlikely successes. One way to approach this issue is to examine those countries whose level of political freedom is substantially higher than one might expect. If these cases have something in common – and if Iraq shares this something – we can feel more optimistic about its prospects. Likewise, if each unexpected success appears to have exceeded expectations in its own fashion, this diversity would suggest that there are multiple ways to overcome apparent structural constraints. At first glance, our examination of outliers does not offer much ground for optimism. To begin with, no country’s democracy score is under-predicted by as much as six points. The world today thus affords not a single case of democracy (eleven or twelve on the scale) in a country predicted to score below six; there are no “semi-democracies” (nine or ten on the scale) with predicted scores below four. If a country with Iraq’s profile were to achieve even the level of political freedom that prevails in Georgia or Guatemala, it would constitute the biggest outlier in the entire world. 23 The most under-predicted countries in our model are shown in Table 3. As the table shows, in only ten countries did we under-predict democracy by more than three points (Benin, Latvia, Lithuania, Mali, Malawi, São Tomé and Príncipe, South Africa, South Korea, Slovenia, and Thailand), and in only four cases did we under-predict democracy by more than four points (Benin, Lithuania, Mali, and South Africa). These countries are undeniably heterogeneous – five are from sub-Saharan Africa; three are from Eastern Europe, and two are from East Asia – which suggests some grounds for optimism. In several of these cases, however, the principal lesson appears to be the impact of international pressure for political liberalization on small, poor, weak states that depend heavily on foreign loans and assistance.22 For better or worse, Iraq will be in a much stronger position vis-à-vis foreign lenders than Benin, Mali, Malawi, or São Tomé and Príncipe. Awash in foreign exchange from the sale of oil, it will be less susceptible to any sort of political conditionality imposed by the IMF, World Bank, the regional development banks, the U.S., or the European Union. [Table 3 about here] A different set of factors is presumably responsible for the unexpectedly positive experiences of Slovenia and the Baltic states.23 All of these countries had much stronger cultural and historical affinities with Western Europe than the Orthodox Christian and Muslim nations of the rest of the former Communist Bloc. In part for this reason, all were considered leading candidates for membership in democratic clubs (like the European Union and NATO) after the collapse of European Communism. Presumably, this combination of cultural legacies and favorable external influences contributed to their higher-than-expected levels of democracy. If Iraq proved equally distinct from the rest of the Arab Muslim world, and if international actors 24 reached out to it in a substantial way, it might well be able to achieve a higher level of democracy. Instead of scoring between zero and two (e.g., Laos or Brunei), it might instead fall somewhere between three and five (Haiti or Pakistan). The two Asian cases – South Korea and Thailand – also seem to have escaped their regional penalty. That is, the amount by which their democracy scores are under-predicted roughly equals the depressive effect of being in East Asia (which in turn is about the same as the cost of having an overwhelmingly Muslim population). To the extent that Iraq is also able to jettison its heritage, it could do several points better than our model predicts. Again, this would leave it substantially better off but still quite undemocratic. The one remaining outlier suggests different grounds for optimism. Given what political scientists know about democracy, South Africa should be a loose authoritarian regime with a score of around six – comparable to Mexico under the Institutional Revolutionary Party (PRI). Instead, it scores an eleven – comparable to Italy and Costa Rica. South Africa’s long history of electoral competition, however racially restricted, may be partly responsible for this outcome, as might intense international pressure for reform. But inspired political leadership presumably also played an important role in establishing and consolidating South African democracy. Were an individual like Nelson Mandela to emerge and take office in Iraq, the country’s politics would presumably take a less autocratic course (at least as long as he held office). Unfortunately, there is little evidence that such a figure exists in Iraq today. Neither returning exiles, internal dissidents, nor former officials of the old regime command national admiration or legitimacy. Despite the fact that many leaders of the Iraqi National Congress and the Iraqi Forum for Democracy have lived in the West, the democratic credentials of even some of these leaders remains uncertain. Indeed, one U.S. official recently likened the Iraqi National 25 Congress to the leadership of the Miami Cuban community – hardly an auspicious comparison (Wright, 2003). The Iraqi equivalent of Nelson Mandela, if he indeed exists, has yet to emerge. All told, an examination of unlikely democratic success stories offers both encouraging and cautionary signs. Some countries have been able to partially overcome obstacles to democratic consolidation through external intervention, domestic leadership, Western influences, or a combination of the three. This fact clearly offers hope for Iraq’s political development. On the other hand, the extent to which structural barriers constrain democracy is also striking; even the biggest outliers are not radically different from their predicted scores. Were Iraq to have the good fortune of these countries – and thus to score several points higher than one might expect – it would still remain quite undemocratic. IMAGINING DEMOCRATIZATION Force-feeding democracy will lead not to reform but to radicalization…In Iraq, it [the United States] cannot behave as an occupying power if it wants to send the right signals to Iraqis that they can freely run their country. The stay of American and British forces should be short (Muasher, 2003). We have to get rid of this naïve notion that by turning on the lights and fixing the hospitals we are going to be able to build a moderate, representative government in Iraq. We’re going to have to play the old imperial game of divide and rule, and the stakes could not be higher (Former U.S. Assistant Secretary of State Martin Indyk, cited in Wright, 2003). A different way to approach the issue of “possibilism” is to ask whether Iraq might benefit from any special conditions that place it outside the “normal” framework for understanding democracy. Can we imagine any unique, pro-democratic influence that would fundamentally alter Iraq’s political trajectory? Is there any factor that we have not yet considered which could dramatically enhance the prospects of democracy in Iraq? 26 The obvious answer, of course, is long-term occupation by pro-democratic foreign powers. In theory, thoroughgoing reconstruction of Iraq by Anglo-American forces could help overcome the structural impediments to democracy that Iraq faces. For instance, occupation forces could forestall the eruption of ethnic violence, train new security forces, offer justice for victims of human rights abuses, ensure economic reconstruction, prevent the establishment of a “petro-state”, oversee electoral competition, encourage the revitalization of civil society, and otherwise promote political reform (Lawson, 2003; Dawisha and Dawisha , 2003). These steps might ultimately prove insufficient to create a durable democratic system in Iraq, but their impact on that country’s political development could still be substantial. In order to assess the degree to which U.S. occupation might reshape Iraq’s political prospects, we attempted to measure directly the impact of U.S. efforts at military-led regime change.24 With this goal, we coded all countries in the world into three categories: those which were occupied or administered at least in part by the U.S. for several years in the twentieth century, and in which the U.S. attempted to impose democratic institutions (Austria, Cuba, the Dominican Republic, Germany, Haiti, Japan, the Marshall Islands, Micronesia, Nicaragua, Palau, Panama, the Philippines, and Vietnam); those which were occupied at least in part by the U.S. for a shorter period of time during the twentieth century and in which occupying powers attempted to establish democratic institutions (Bosnia-Herzegovina, Cambodia, Grenada, Somalia, South Korea); and all other countries.25 In some cases, these codings are not perfectly clear. For instance, the “occupations” of Bosnia-Herzegovina and Somalia were done under United Nations auspices; the Dominican intervention in 1965-66 was technically conducted by the OAS; and the invasion of Grenada was undertaken with the assistance of six neighboring states. Likewise, only portions of Vietnam, Germany, and Austria were ever under American 27 control. Measuring duration is also not as straightforward as it might at first seem: some countries (such as Cuba, the Dominican Republic, Haiti, the Philippines) were occupied more than once; South Korea was only truly occupied during 1945-48 (though U.S. forces later returned in large numbers); and postwar West Germany and Austria remained under Allied control for much longer than foreign forces actually administered the government. Finally, U.S. commitment to democratization in the first half of the twentieth century was somewhat less consistent than its political vision for defeated Axis powers after World War II. Nevertheless, we believe that our coding captures the basic outlines of U.S. efforts to leave behind democratic systems in the wake of occupation. When dummy variables measuring U.S. occupation were included in our multivariate analysis, they did not approach statistical significance and in fact had the “wrong” sign. They also failed to attain statistical significance in simple bivariate regressions, and even the magnitude of this non-significant effect was not particularly impressive. 26 Countries that had been occupied or administered for a long time by the U.S., for instance, scored only 1.4 points higher on average than countries that were never occupied.27 The bottom line seems to be that protracted U.S. occupation exercises only an indirect effect on a country’s prospects for democracy, working through variables that are already included in our model, and on average even this effect is relatively modest (see Lawson and Thacker 2003).28 In this context, it is worth considering the nature of America’s commitment to democratize Iraq. Although an intelligently conceived, well-executed, and thoroughgoing occupation might increase the odds of democratization, the United States faces substantial pressures to leave much sooner. Even a relatively brief stay – i.e., just a few years – could shorten further if the costs of occupation grow or if military conflicts elsewhere demand U.S. 28 attention.29 Meanwhile, lengthy occupation itself could create hostility among Iraqis. Ultimately, popular disaffection with foreign occupation could become so thorough than any democratic institutions put in place under Anglo-American auspices were regarded as tainted. Even more problematic is the possibility that American intervention might not actually be aimed at promoting democracy. U.S. authorities presumably prefer a democratic system, but they also clearly fear the establishment of a Shia-dominated, fundamentalist regime. Some American officials might even favor an undemocratic but pro-western client state over an Islamist semi-democracy. If these fears lead the U.S. to impose restrictions on political competition, the prospects for democracy diminish still further. CONCLUSION It may be that the establishment and maintenance of political systems are inherently impossible to predict. In the language of “transitology”, democratization may be too dependent on enlightened leadership (virtu) and serendipitous events (fortuna) for academics to offer meaningful insights (O’Donnell, et al. 1986). The pretense of political science, however, is that many important political outcomes are amenable to scholarly research and analysis. From this perspective, several decades of intensive investigation into the causes of democracy ought to be able to offer something in the way of guidance on Iraq’s prospects. All told, U.S. intervention aimed at removing Saddam Hussein and establishing a more participatory system of government is likely improve the level of political freedom in Iraq. Per capita income may well increase, laying the social foundation for more representative government. In addition, occupying powers could take specific, concrete steps that would 29 somewhat improve the prospects for democracy in Iraq. Such reforms might well convert Iraq into a decidedly less authoritarian country than it was under the Baathist regime. Nevertheless, many key features of Iraqi society are unlikely to change with foreign occupation. Even with substantial increases in national income and extensive U.S.-imposed reforms, Iraq will remain a relatively poor, overwhelmingly Muslim country with little history of political freedom, located in a particularly rough neighborhood of the world. Collectively, these factors substantially limit the prospects for democracy in Iraq. Although there is little reason to believe that Iraq’s next regime will be as bad as the one that Anglo-American forces replaced, there is equally little reason to think that it will be democratic. 30