How to reform the Islamic world will be among the topics NATO's leaders will discuss in Turkey next week. Both President Bush and the European Union have proposed bold democratization initiatives in the region. Can such initiatives succeed?

Islam and democracy are frequently presumed to be bitter antagonists. A careful study of the world's 47 Muslim-majority states, however, shows that Islam and democracy can and do co-exist. The real gap is narrower: it is the Arab world, not Islam, which seems at loggerheads with democracy.

This conclusion is based on comparing Muslim countries' "electoral competitiveness." If a government sprang from reasonably fair elections and the elected government is able to fill the most important political offices, the country is deemed "electorally competitive."

Electorally competitive countries are not necessarily democratic: some do not fully control the state's territory; others violate both their constitutions and human rights. But electoral competitiveness is always a necessary condition for democracy - and therefore a central consideration when evaluating a country's prospects for democratization.

Two initiatives - the Polity Project, founded by political scientist Ted Gurr, and Freedom House's annual Freedom in the World surveys - gauge political rights in almost every country. The differences between Arab and non-Arab Muslim countries are striking. Of 29 non-Arab Muslim countries evaluated by Polity IV , 11 enjoyed significant political rights for at least three consecutive years in the period from 1972 to 2000, while eight experienced at least five consecutive years of political rights.

The Freedom House scores are remarkably similar: 12 of 31 non-Arab Muslim countries had relatively high scores for at least three consecutive years, and eight for five consecutive years. Both research teams rate Albania, Bangladesh, the Gambia, Malaysia, Mali, Niger, Nigeria, Pakistan, and Turkey as meeting the three-year criterion, with all but Albania and Niger meeting the five-year criterion.

But on the Arab side, only Lebanon had three consecutive years of relatively strong political rights - before the 15-year civil war that began in 1975 - and no country experienced five consecutive years of strong rights. A non-Arab Muslim country in the period from 1972 to 2000 was almost 20 times more likely to be electorally competitive than an Arab Muslim-majority country.

In the non-Arab Muslim-majority subset, both Polity IV and Freedom House rate nine countries as having experienced at least three consecutive years of substantial political rights. Strikingly, seven of them are clear "overachievers," because they also have low per capita GDP levels. Indeed, five of the nine qualify as "great overachievers" - significant political rights despite annual per capita income of less than $1,500.

By contrast, none of the 16 Arab Muslim-majority countries is an overachiever. Seven had annual per capita income levels exceeding $5,500 in the period from 1972 to 2000 but no significant political rights for three consecutive years. These states are electoral "underachievers" - weak political rights despite relative affluence.

Elections do occur in the Arab world, and they vary in frequency and significance. In the complete autocracies - Saudi Arabia, Libya, Syria, Tunisia, Iraq under Saddam, and the United Arab Emirates - there have been no meaningful elections to fill the most important offices. The UAE, a somewhat decentralized and consensual federation, is not as dictatorial as the others, but only seven voters - the traditional rulers of the federation's seven emirates - elect the president.

Elections have begun to play a greater role in Qatar, Kuwait, Bahrain, and Oman, but none uses free and fair elections to fill the most important and powerful political offices. Traditional monarchs still have broad powers of appointment and decree.

Finally, some Arab countries are not electorally competitive now but once did use free and fair elections to fill the state's most politically powerful offices (Lebanon). Others seemed close to doing so (Yemen, Morocco, and Jordan) or had a possible political opening but are now substantially farther away from being electorally competitive (Egypt and, arguably, Algeria).

It makes intuitive sense that Islam cannot by itself explain such differences. All of the world's religions contain some doctrines and practices that are potentially harmful to the emergence of democracy, and others that are potentially beneficial. A beneficial Islamic doctrine is the Koranic injunction that "there shall be no compulsion in matters of religion." Others include shura (consultation), ijtihad (independent reasoning), and ijma (consensus). Perhaps religious and political leaders in electorally overachieving states such as Senegal, Mali, Bangladesh, and Indonesia draw upon some of these concepts.

Most political cultures can and do change over time, because to some extent they are socially constructed by new opportunities, threats, and contexts generally. Witness Catholicism's pro-democratic transformation. Unless Arab political culture is shown to be uniquely and permanently inimical to electoral competitiveness, it makes more sense to understand the democratic divide in the Muslim world in terms of the political - as opposed to the ethnic or religious - particularities of the Middle East and North Africa.

One such particularity is that many contemporary Arab states were established with relatively new and arbitrary boundaries, and were then occupied and often reconfigured as European colonies. The linguistic dominance of Arabic throughout the Middle East and North Africa, together with pan-Arabism, reinforced weak national identities. Anyone even remotely familiar with the region knows how common is the phrase "the Arab nation" ( watan ).

Furthermore, unlike Latin America, Africa, and elsewhere, Middle East democrats have not benefited from the Cold War's end. The US continues to subsidize authoritarian Arab regimes, such as Egypt (which receives at least $2 billion a year), because it buys peace with Israel and maintains US geopolitical influence in the Arab-Israeli conflict.

Resolving that conflict - itself a key aspect of Arab political identity - might enable Arab political culture to change. But NATO's leaders should bear in mind that, viewed in comparative and historical perspective, such change will be less an imposition from outside and more the result of internal pressures and initiatives.