Dr. Goldsmith is professor of management at the University of Massachusetts, Boston.
While delivering the keynote address at a National Endowment for Democracy event in 2005, President George W. Bush singled out Islamic extremism as the number-one foreign policy challenge facing the United States. A primary reason for this threat, he asserted, is the lack of democracy in Muslim countries. “If the peoples of that region [the Greater Middle East] are permitted to choose their own destiny,” Bush argued, “then the extremists will be marginalized, and the flow of violent radicalism to the rest of the world will slow, and eventually end.”1
Bush did not mention it, but he was wading into a longstanding controversy about “Muslim exceptionalism” regarding democracy. For example, the UN Development Programme has identified a “freedom deficit” in the Greater Middle East that places it last among the regions of the world in terms of political freedom.2 The World Bank has also lamented what it calls a “governance gap” in the Middle East and North Africa.3 Among academic observers, the main dispute is over what produces this pattern of embedded authoritarianism, not whether it exists. Some influential writers, such as Samuel Huntington, consider Muslim culture averse to norms of democracy and rule of law.4 Other experts, for instance Larry Diamond, attribute Muslim authoritarianism to entrenched leadership and lack of opportunity for participation by ordinary people, not retrograde values or beliefs.5 Left largely unchallenged is the prior assumption common to all sides of the debate, which holds democracy to be unusually lacking in Muslim countries whatever the root causes might be.
There are 47 countries (including the Palestinian Authority) where over half the population adhere at least nominally to the faith of Islam. This is approximately one quarter of the world’s nation states, a distinctly large share to be impervious to democratic practice, if this is so. For convenience, these 47 nation states will henceforward be identified simply as Muslim or Islamic countries, though there are obviously other countries with substantial Muslim populations. They range in size from Indonesia, with a population of more than 200 million, to the Maldives Islands, with a population under 300,000. We should be wary of brushing such a diverse group with the same broad stroke.
This essay uses independent data to show that Muslim authoritarianism may not be as rife and entrenched as many in the international-relations field seem to take for granted. The existence and dimensions of any modern-day democracy gap are not settled issues but constantly changing empirical questions. Just how exceptional is democracy in the Muslim world today? Is it helpful or valid to think of a generic Islamic democracy gap, or is that too easy an assertion that gives the wrong impression about many specific societies? To be clear, the goal of this paper is modest: to plumb the depths of contemporary Muslim exceptionalism with regard to democracy. The analysis gives context to the excessively polarized debate in international relations about clashing civilizations and suggests it is more fruitful to look at countries in a less monolithic way.
Democratization is an irregular process of replacing authoritarian regimes with rule-bound competitive systems. The pace of democratization has picked up dramatically since the end of the Cold War, with many additional countries undergoing political renaissance that opens them up to greater involvement by citizens and civil society. Democratization is not a universal historical sequence ending in the same types of political systems, even though more and more nation states are adopting similar-looking political institutions and practices based on a common set of Western models. Public-opinion surveys in Muslim countries show broad support for democracy as a system that can work in those societies.6 What local respondents actually have in mind when they respond to questions about democracy, however, may not be quite the same as in the United States or Europe.7
Democracy means different things to different people. At the simplest level, it is a form of government in which citizens pick their rulers, and rulers are therefore responsive to the preferences of citizens. What are the common denominators?
Valerie Bunce puts forward a concise definition that captures two essential aspects of democracy. She suggests we think of the term as “a two-part proposition, having uncertain results (or competition) but also having certain procedures” — in other words, “competition bounded by rules.”8 This will be our minimal definition of democracy in this paper.
Multi-party elections, held at regular intervals, tell an important part of the story about democratization. Elections are not the full story, because elections obviously connote different things in different national settings. Applying Bunce’s two criteria for democracy, electoral competition may be lacking because of weak opposition, or electoral procedures may be uncertain because of lax enforcement, or both. Provided we do not read too much into elections, they are still among the best indicators of where a country is heading along the authoritarian-democratic gradient. Even rigged electoral contests can reveal discontent and may influence authoritarian leaders to bend to popular will or to introduce incremental political reforms as a way to co-opt displeased citizens.
There were 126 “electoral democracies” in place at the end of 2005, according to Freedom House, a nonprofit organization that monitors democracy around the world.9 To qualify as an electoral democracy in Freedom House’s lexicon, a state must have a competitive, multi-party political system, universal adult suffrage, regular elections and open media. Only nine of the 126 countries are from our sample: Albania, Bangladesh, Comoros, Indonesia, Mali, Niger, Nigeria, Senegal and Turkey. Although Muslim countries are clearly underrepresented among the world’s electoral democracies, any representation at all calls into question the simplified mental picture of an Islamic democracy gap. It is also worth noting that these nine nations account for nearly half the total population in the 47 Muslim countries (and about one-third of all Muslims in the world).
Neither Bunce’s idea of bounded competition nor Freedom House’s electoralism says much about constitutional liberalization, whereby citizens gain freedom from interference by government in addition to the right to vote and run for office. Electoral participation and constitutional liberalization are not entirely compatible. Political philosophy has long recognized a tension between giving voice to the “general will” of society and protecting individuals against the “tyranny of the majority.” Many fledgling democracies in the developing world are “illiberal” populist systems, not self-restraining liberal democracies. They embrace mass participation but are indifferent about safeguarding individual freedoms or minority group rights.
The Islamic Republic of Iran is a good example of such a system among Muslim nations. It is a religious state that abridges freedom of worship and does not allow people to speak and write freely, yet elections are held on a reliable schedule. Within the bounds set by spiritual leaders, there is lively political participation and vigorous competition for office. The opposite blending of liberalism and autocracy is also possible and quite common in the Middle East.10 These are countries that steer clear of multi-party national elections but do allow at least some of the individual rights emphasized by Enlightenment-era liberal philosophers. Many autocratic regimes tolerate or even encourage some liberal freedoms as a strategy to stave off pressures to open the leadership selection process and make it more inclusive. Qatar is an illustration, having freed its press and allowed broadcasting of the free-wheeling satellite television station al-Jazeera, while making few concessions to open the political realm to new participants. Political systems are neither black nor white, suggesting care about drawing too bright a line concerning regime type. Democratization moves in starts and stops, and real countries often exhibit unexpected combinations of democratic and authoritarian features.
FREEDOM HOUSE INDEX
One way of roughly marking the line between regimes is to start with one of the available quantitative indicators of democratization in different countries. Freedom House publishes an annual Index of Political Freedom that is widely used for such purposes in comparative empirical research.11 The organization codes all countries for both political rights and civil liberties. It ranks countries on a pair of declining seven-point scales. These scales are combined into a composite political-freedom score that can be interpreted as a snapshot of the degree of democracy obtained by any country in a given year.
The lower the composite score (two points is the lowest possible mark), the closer the nation approaches liberal democratic norms. Results go back to 1972, making it possible to witness long-term trends in democratization around the world. Freedom House establishes three cutoff points and categorizes countries as “not free,” “partly free” and “free.” To make more subtle comparisons, however, this paper uses the entire 13-point composite scale.
A very elementary approach toward identifying an Islamic democracy gap is to plot the average indexes for Muslim nations versus non-Muslim nations through 2006. The results are shown in Figure 1 (the y-axis is reversed to make the slope of average composite Freedom House rating registered a slight improvement starting in the 1990s, but the upswing was too little to narrow the difference with non-Muslim countries’ mean annual ratings, which continued along an improving path as well. It took until 2003 for the Islamic countries’ mean score to get back to what it was in 1972. This certainly looks like a democracy gap.
But the divergent trends graphically portrayed in Figure 1 may be an unfair comparison. First, we should be wary about making any inference that what is true of a whole must also be true of its constituents. The average Freedom House score conceals considerable variation; specific Islamic countries perform considerably better or worse than the group average. Second, the democracy gap that we see in Figure 1 may partly be an “income gap.” A well-established correlation exists between national income and democratic governance, which makes the raw freedom-index scores a questionable basis for cross-national comparisons. Poor countries tend to be less democratic than rich countries, and Islamic countries on the whole have low incomes. In 2000, the median GDP (gross domestic product) per capita in the 47 Muslim nations was only $3,567, compared to $6,496 for the non-Muslim nations.12 All other things being equal, poorer average democratic performance is to be expected.
It is more instructive, therefore, to make comparisons of regime type that control for the influence of national income in recent years, and to look at individual country achievement, not just averages.
This dual task can be accomplished with a worldwide regression model that finds the extent to which you can predict a country’s 2000-06 average Freedom House rating by knowing the logarithm of its GDP per capita in 2000. The logarithmic transformation of GDP per capita is used because democratization probably does not respond linearly to income. Gains in democracy are likely to diminish as countries get richer.
The resulting estimate indicates that the logarithm of GDP per capita “explains” one-quarter of the international variation on the index (the adjusted r-squared is 0.25). Knowing a country’s recent individual income level thus tells you a good deal about the average level of its democracy since 2000, as indicated by the Freedom House composite index.
Given that national income is not the only factor that accounts for the Freedom House scores, the observed value for most countries falls above or below the value predicted by the regression model. If we look at the unexplained (or residual) variation after fitting the regression model, we see Islamic nation-states either “over-perform” or “underperform” their economic status. In other words, some countries get better marks for political and civil rights than they “should,” given their level of income; others get lower marks.
Figure 2 displays the results in graphic form. Eleven Islamic nation-states are at or above the world trend line and thus meet or exceed expectations for democratic performance; the remaining 36 countries are below the trend line and fall short of expectations.
Looked at this way, Turkey has disappointing democratic attainment. It is among the top-rated Islamic countries for civil and political freedom since 2000, but you would expect somewhat better performance of a country with Turkey’s income level. Niger, on the other hand, has superior political freedom positive and therefore more intuitively readable). A persistent noticeable difference appears in the Muslim-country data line. Mean freedom scores were about two points worse on the composite scale than non-Muslim countries’ mean scores in the 1970s. As the “third wave” of democratization gathered force in the 1980s, the disparity in scores increased to about five points, with Islamic countries bucking the international trend by becoming even less democratic. Their attainment in governance relative to its meager resources. Niger’s democracy rating is almost the same as Turkey’s for 2000-06, but that is a surprisingly high achievement level given Niger’s low per capita GDP. As these two examples suggest, the distribution of income-adjusted performance is quite different from the distribution of unadjusted scores. “Over-performing” Muslim countries are largely African (Burkina Faso, Comoros, Gambia, Mali, Nigeria, Senegal and Sierra Leone in addition to Niger). However, the list of “over-performers” also includes two very large Asian countries, Bangladesh and Indonesia, plus one European country, Albania.
Which countries “underperformed” for the 2000-06 period? In addition to Turkey, they include Arab Mediterranean countries (Algeria, Egypt, Jordan, Lebanon, Libya, Morocco, Syria and Tunisia) and the Persian Gulf oil states (Bahrain, Iran, Iraq, Kuwait, Oman, Qatar, Saudi Arabia, the United Arab Emirates, and modest oil exporter Yemen). Eight additional “underperformers” are nations in sub-Saharan Africa (Chad, Djibouti, Eritrea, Guinea, Maldives, Mauritania, Somalia and Sudan); five are former republics of the Soviet Union in Central Asia (Azerbaijan, Kyrgyzstan, Tajikistan, Turkmenistan and Uzbekistan); and the remaining “underperformers” are scattered in Asia (Afghanistan, Brunei, Malaysia and Pakistan).
This is a long list of Muslim countries that have less democratization than expected according to their per capita GDP. However, the regression model probably overstates the likelihood of open governance in several oil-exporting states.
Because oil wealth inhibits democratization, high national income is a misleading indicator in these countries. According to Michael Ross, an abundance of oil hinders democracy in several ways, including its delaying effect on economic diversification, which, in turn, forestalls development of an indigenous professional and commercial class that would push for liberal democratic reforms.13 Oil-exporting countries jump out as outliers in Figure 2, but the actual extent of their underperformance is possibly less striking than appears at first. The former Soviet states of Central Asia are also outliers. There may be special historical reasons for their lagging performance in democratization that have more to do with the legacy of communist rule than with either national income or religion.
In any case, the widely scattered distribution of countries in Figure 2 suggests that blanket charges about a uniquely Islamic democracy deficiency are misleading, especially if income level is taken into consideration. Muslim countries run the gamut. Enough of them do so poorly on the democracy index that it is easy to see how perceptions of Muslim exceptionalism have arisen. However, the relatively positive performance of other Muslim countries serves to refute any strong claim that Islam and democracy are irreconcilable, at least as Freedom House measures regime type. What is fair to say is that over half the Muslim countries do worse than expected on the index, which could be interpreted as support for a weak or probabilistic version of the democracy gap thesis.14
Because the Freedom House index is highly aggregated and subject to many possible biases, it is useful to approach Muslim exceptionalism in an entirely different way and see if results are analogous. An obvious method is to turn to election campaigns and election results.
Rather than trying to rank or label countries using a vague or disputable index of political freedom, we can simply count the number of electoral contests and catalog their outcomes. Table 1 shows comparative election data for Muslim countries from 1990 through 2006. Voting seems far more common than might be assumed based on strong assumptions about a democracy gap.
During those years, 33 countries in the Islamic world held at least one multi-party leadership election (usually a presidential election, but included are legislative elections under constitutional monarchies or in systems where presidents are indirectly selected by legislatures). Multi-party election means a contest where opposition parties can compete even if sometimes under duress. (Since these are emerging democracies, it seems reasonable to allow for a somewhat less-level playing field than if U.S. or European standards were applied.) Thus defined, there were a surprisingly high number of these multiparty leadership elections over the 17-year period: 83 in all without double-counting runoffs.
Nonpartisan legislative elections have also occurred or been scheduled in five additional Persian Gulf states (Bahrain, Oman, Kuwait, Qatar and the UAE).
These particular elections are not leadership contests that affect the national executive directly, but they do give some voice to popular sentiment. After the balloting takes place in these countries, de facto parties often do emerge in the national parliament among the elected representatives, so labeling the elections nonpartisan may not be entirely accurate. In other recent developments, Saudi Arabia is experimenting with local government elections, and Maldives has legalized political parties in anticipation of national elections. Only eight Muslim countries appear to have had no competitive campaigns for political office whatsoever since 1990 (Brunei, Eritrea, Libya, Somalia, Syria, Tajikistan, Turkmenistan and Uzbekistan). Even with these holdouts, the number and extent of contested elections in Islamic countries seems far higher than the Muslim-exceptionalism thesis would predict.
Recurring elections are also more common than might be expected. Over one-third of the sample (or 17 countries) held three or more such contests. Three elections in 10 or 15 years would probably meet Bunce’s test for regular electoral competition and would qualify those countries as democratic under that narrow criterion.
What about the uncertainty of results, also emphasized by Bunce as the second important dimension of democracy? Here the evidence is more in line with a modest version of the democracy-gap thesis. The third column in Table 1 shows the top vote getters’ share of ballots cast in the most recent leadership elections. The winners’ median vote total was 65 percent of votes cast, a lopsided victory margin that suggests many of the contests failed to meet the democratic criterion of having uncertain results. Hosni Mubarak’s reelection in Egypt in 2005, for example, was never in doubt. In other countries, boycotts by major opposition parties assured the outcome, as when Lansane Conté ran unopposed in Guinea in 2003. These sorts of predetermined election results suggest the limits to democratization in many Muslim countries.
The best measure of uncertain outcomes (or the degree of competition) in an electoral system is how often it produces changes in leadership or simply revalidates the existing power structure. Incumbents always have an insider edge, and they will invariably use those advantages to extend their own power.15 Rather than democracy, this sort of system is better described as “competitive authoritarianism,” with duly elected leaders misusing their office to manipulate the electoral system and hold onto power. It is opposition victories that usually catch us unaware and indicate the sort of electoral unpredictability associated with democracy. Column 4 in Table 1 shows that such unexpected outcomes happen more frequently in the Muslim world than might be assumed.
Ten Muslim countries had one or more opposition victories between 1990 and 2006, with the losing incumbent (or the incumbent’s designated successor) accepting defeat and agreeing to step aside. The countries range from Albania in Europe to Iran, Lebanon and Turkey in the Middle East; to Bangladesh, Indonesia, Kyrgyzstan and Pakistan in Asia and to Mali and Senegal in Africa. These ten nation states would appear to satisfy both of Bunce’s minimal conditions for democracy— certain procedures plus uncertain results. The ranks of modestly democratic systems may well grow in the future if elections keep recurring in other Muslim countries at the rate they have for the past decade. The law of averages suggests additional incumbent losses are inevitable.
Even a relatively peaceful transfer of power through the ballot box is no proof that democracy has been consolidated — that it has become institutionalized and unlikely to break down or be reversed. A simple check of democratic consolidation is the “two turnover test.” The first time an opposition leader replaces an incumbent power holder does not necessarily establish a tradition of peaceful political change. It is only after the former opposition leader is defeated and leaves office that we can begin to be confident democratic procedures have taken root.16 Second turnovers are still rare in the Muslim world. They occurred in only four countries (Albania, Bangladesh, Mali and Iran) during the 1990-2006 period, but additional turnovers can be anticipated as time passes and more ruling parties wear out their welcome with voters.
Muslim exceptionalism is not exactly a mirage, but the closer you look at the latest data, the harder it is to see a powerful, unique pattern of democratic deficiency in these countries. This paper has looked at two different types of political data: the Freedom House index and the results in raw numbers of multi-party national leadership elections. Each set of figures suggests that to talk about a generic Muslim democracy gap is to simplify matters to the point of misrepresentation.
Putting aside any preconceptions, individual variation dwarfs many of the aggregate similarities among Islamic polities. Policy makers should be cautious not to overgeneralize and stereotype or to speculate strongly about any particular Muslim country’s democratic prospects based on political attributes of Muslim countries as a group. The average Muslim country is not very democratic, but this is a dubious abstraction. Depending on how we define our terms, perhaps ten or so (the list varies) mildly democratic systems already exist in the Muslim world today. More may be on the way.
That said, policy makers also need to be careful about projecting the experience of these less authoritarian countries onto all the rest. The strong version of the Muslim exceptionalism thesis is an obvious oversimplification, but that does not mean President Bush is on firm ground when he says that despite “skeptics of democracy [who] assert that the traditions of Islam are inhospitable to representative government.…It should be clear to all that Islam — the faith of one-fifth of humanity — is consistent with democratic rule.”17 Every country has unique characteristics, and sweeping claims about Muslim societies’ capacity for democracy are just as suspect as broad claims to the contrary. In considering the role of democratization in foreign policy toward the Greater Middle East, it is important to try to set aside prior judgments and look at the nuances of each case.
1 George W. Bush, “President Discusses War on Terror at National Endowment for Democracy,” The White House (October 6, 2005).
2 Arab Human Development Report: Freedom and Good Governance (UNDP, 2004).
3 Better Governance for Development in the Middle East and North Africa: Enhancing Inclusiveness and Accountability (World Bank, 2003).
4 Samuel P. Huntington, The Clash of Civilizations and the Remaking of World Order (Simon and Schuster, 1996), p. 29.
5 Larry Diamond, “Can the Whole World Become Democratic? Democracy, Development, and International Policies” (Center for the Study of Democracy, Paper 03-05, April 17, 2003), p. 13. For additional observations regarding this debate, see Muqtedar Khan, “Prospects for Muslim Democracy: The Role of U.S. Policy,” Middle East Policy, Vol. 10 (September 2003), pp. 79-89; Sanford A. Lakoff, “The Reality of Muslim Exceptionalism,” Journal of Democracy, Vol. 15 (October 2004), pp. 133-139; and Marcus Noland, “Explaining Middle Eastern Authoritarianism” (Institute for International Economics, Working Paper No. 05-5, 2005): Available at Social Science Research Network: http://ssrn.com/abstract=759652.
6 Pew Research Center, “Iraqi Vote Mirrors Desire for Democracy in Muslim World” (A Pew Global Attitudes Project Commentary, February 3, 2005): Available at <http://peoplepress.org/commentary/display.php3? AnalysisID=107>.
7 Salwa Ismail, “Democracy in Contemporary Arab Discourse,” in Political Liberalization and Democratization in the Arab World: Vol. 1, Theoretical Perspectives, ed. Rex Brynen, Bahjgat Korany, and Paul Noble (Lynne Rienner, 1995).
8 Valerie Bunce, “Comparative Democratization: Big and Bounded Generalizations,” Comparative Political Studies, Vol. 33 (August-September 2000), p. 714.
9 Freedom House, Global Survey 2006: Middle East Progress amid Global Gains in Freedom (New York, 2005).
10 Daniel Brumberg, “Democratization Versus Liberalization in the Arab World: Dilemmas and Challenges for U.S. Foreign Policy” (Strategic Studies Institute of the U.S. Army War College, July 2005).
11 Several other indexes of political-regime type exist, most of which correlate highly despite using different methodologies, according to Adam Przeworski and his colleagues in Democracy and Development: Political Institutions and Material Well-Being in the World, 1950-1990 (Cambridge University Press, 2000), pp. 56-57.
12 Calculated from Alan Heston, Robert Summers, and Bettina Aten, Penn World Table Version 6.2 (Center for International Comparisons of Production, Income and Prices, University of Pennsylvania, September 2006).
13 Michael L. Ross, “Does Oil Hinder Democracy?” World Politics, Vol. 53 (April 2001), p. 356.
14 For similar conclusions, see Manus Midlarsky, “Democracy and Islam: Implications for Civilizational Conflict and the Democratic Peace,” International Studies Quarterly, Vol. 42 (September 1998), p. 504; and Steven M. Fish, “Islam and Authoritarianism,” World Politics, Vol. 55 (October 2002), p. 16.
15 Marc Howard and Philip Roessler, “Liberalizing Electoral Outcomes in Competitive Authoritarian Regimes,” American Journal of Political Science, Vol. 50 (April 2006), pp. 365-366.
16 Samuel P. Huntington, The Third Wave: Democratization in the Late Twentieth Century (University of Oklahoma Press, 1991), pp. 266-267.
17 George W. Bush, “Remarks by the President at the 20th Anniversary of the National Endowment for Democracy,” The White House, November 6, 2003.