ISLAMIC CULTURE AND WOMEN'S RIGHTS: A PATH DEPENDENCE ARGUMENT

Islam as a religion is committed to tolerance and social justice and is not incompatible with gender equality. The Quran does present a clear gender hierarchy but provides relatively favorable conditions for women, its time of writing considered (Mayer Reference Mayer1999, Chap. 6; see also Keddie Reference Keddie1990). For example, the Quran sees Adam and Eve as being “of equal nature” and, contrary to the Bible, it does not hold Eve solely responsible for the original sin (Ishay Reference Ishay2008, 7, 35, 55). Instead, what became crucial for Muslim women's status was the dominant religious interpretation during the formative years of the Islamic civilization and, specifically, the formalization of Islamic law (sharia) two centuries after the death of the Prophet, which led to a progressive loss of women's social and sexual rights (Ruthven Reference Ruthven1984, 87; Syed Reference Syed2010).

After the early Islamic conquests in the seventh and eighth century, Muslims ruled the Middle East, except for Turkey, as well as North Africa and parts of Central Asia (Hodgson Reference Hodgson1974; Ruthven and Nanji Reference Ruthven and Nanji2004). The territorial expansion brought along exposure to foreign culture, and Hellenic influences of philosophy and science inspired a ninth century enlightenment movement in the Muslim world. The movement was led by a group of rationalist theologians, the Mu'tazilites, who emphasized free will and believed that the Quran should be subjected to reason (Ali Reference Ali2007; Hourani Reference Hourani1985; Reilly Reference Reilly2010, Chap. 1). The Mu'tazilites initially gained momentum and were even supported by the Abbasid caliphs, al-Ma'Mum (813–833) and al-Mu'tasim (833–842), but their free thinking challenged the standard orthodox belief that the Quran is uncreated, the work of God, and thus unquestionable. The theological struggle that followed was won by the religiously orthodox who controlled the Islamic education system and eventually won over the political elite when caliph al-Mutawakkil (847–861) withdrew support for the Mu'tazalite doctrine (Ruthven Reference Ruthven1984, 198–200; see also Tibi Reference Tibi2009, Chap. 8).

Over the following 100-year period, the rationalist movement came to a rather abrupt end, and by the 11th century the few remaining Mu'tazalites were subject to repression and their works were destroyed on charges of heresy (Ali Reference Ali2007; Hourani Reference Hourani1985; Reilly Reference Reilly2010). The victorious orthodox doctrine left little room for free will and insisted on the uncreated nature of the Quran and the immutability of the divine laws.

Importantly, the theoretical struggle of the ninth century coincided with the formalization of Islamic law and the consolidation of Islam in everyday life (Ruthven Reference Ruthven1984, 87). It thus marked the end of the formative period of the Islamic civilization, symbolized by the death of one of the most authentic hadith collectors, al-Bukhari (810–870; see Figure 1), whose records of the sayings and acts of the Prophet, which previously had been passed along orally, became a fundamental part of Islamic law (Hourani Reference Hourani1985, Chap. 1; Ruthven Reference Ruthven1984, Chap. 4). As a consequence, the orthodox and not the Mu'tazilite doctrine laid the foundation for the classical period of the Islamic civilization and came to influence Islamic culture for more than a millennium. In Ali's (Reference Ali2007, 446) words, after the defeat of the rationalist movement “it was the anti-philosophic orthodoxy that reigned supreme in the Muslim world.” The cost of altering established arrangements and institutions increases as they become embedded in society (Pierson Reference Pierson2000, 259).Footnote ² This means that even modest historical events can have major long-term consequences if they happen at the right time. The rationalist defeat became the key historical juncture that initiated a path of resistance to social change in the Muslim world, and the durability of this path was strengthened by the orthodox notion of Islamic law as being immutable. The formalization of Islamic law was designed to reject new religious interpretations, as indicated, for example, by the juxtaposition of innovation and heresy in the religious vocabulary (Ruthven Reference Ruthven1984, 152–153; Tibi Reference Tibi2009, 63).

Regarding family law, women were placed in a distinctively subordinate relationship with their male relatives (Keddie Reference Keddie1990; Mayer Reference Mayer1999, Chap. 6). Polygamy was accepted, divorce rights were reserved for men, women's inheritance was half of men's, and male guardianship was required. The masculine domination was to a large extent justified in the religious texts. For example, in one of al-Bukhari's hadiths, the Prophet is said to have denigrated both women's intelligence and religiosity: “Is not the witness of a woman equal to half the witness of a man? This is the defect in her intelligence. And when she is ceremonially impure [i.e., menstruating] she neither prays nor fasts. This is the defect in her religion” (Ruthven Reference Ruthven1984, 166). Islamic family law thus relegated women to a life in domesticity, which meant that they came to play an insignificant role in economic, social, and political life (Mayer Reference Mayer1999, 98–99; see also Syed Reference Syed2010).

As already mentioned, historically strong patriarchal traditions are by no means unique to Islamic culture. However, the embeddedness of orthodoxy in religious institutions and in society at large affected the Muslim world's reaction to modernization. Modernization was frequently seen as “Westernization” and social change was resisted, particularly regarding gender equality (Lewis Reference Lewis2002). Hence, when institutions such as banking and cooperation inevitably had to change as a consequence of economic globalization, opposition was maintained to changes involving family life, which were seen as vital to Muslim identities (Kuran Reference Kuran2011, 300). As Huntington (Reference Huntington1996, Chap. 5) argues, modernization and the influence of Western culture brought along an Islamic resurgence where Muslims turned to Islam as a source of identity (cf. Figure 1). Islamist movements became a dominant force because religious organizations held legitimacy and because political authorities legitimated their political power on the basis of conservative Islamist discourses (Huntington Reference Huntington1996, 115; Moghadam Reference Moghadam2003). Resistance to women's empowerment thus still dominates much of the Muslim world (Allawi Reference Allawi2009, 204; Mayer Reference Mayer1999, 99–100), and as Inglehart and Norris (Reference Inglehart and Norris2003, 71) bluntly put it, “an Islamic religious heritage is one of the most powerful barriers to the rising tide of gender equality.”

Figure 1.

Timeline of the Islamic Civilization. Sources: Hodgson Reference Hodgson1974, vol. 1–3; Hourani Reference Hourani1985; Ruthven Reference Ruthven1984; Ruthven and Nanji Reference Ruthven and Nanji2004.

Many Muslim societies have witnessed both significant debate and progress concerning women's rights (e.g., Mir-Hosseini Reference Mir-Hosseini2000; Moghadam Reference Moghadam2003), and a potential caveat to the historical explanation presented here is its inability to account for variation within the Muslim world. A common assumption, which also exists in Ross' (Reference Ross2008; Reference Ross2012) work, is that the lack of development is a particularly Middle Eastern or Arab problem (Arab Development Challenges Report 2011; see also Stepan and Robertson Reference Stepan and Robertson2003).Footnote ³ Such notions would lead us to expect that geography or ethnicity influences Muslim women's status.

An alternative historical proposition would be that Muslim countries differ according to the historical period during which they became a part of the Muslim world. As argued above, the orthodox interpretation of Islam became dominant from the ninth century onward, and one possibility is that it only did so in the countries that were part of the Muslim world at that time. Afterward, during the classical period of the Islamic civilization, Islam spread to parts of Sub-Saharan Africa, Central, South, and Southeast Asia, as well as the Balkan Peninsula via Muslim conquests, merchants, and missionaries (Hodgson Reference Hodgson1974; Ruthven and Nanji Reference Ruthven and Nanji2004; see Figure 1). Conversion was typically a gradual process where Islamic culture and practices had to coexist with non-Islamic traditions. This meant that the Islamic model had to be rather flexible, and Islamic law was often only partly realized in public, commercial, and criminal matters (Ruthven Reference Ruthven1984, 152–162).

Historical processes thus provide a potential source of variation between Muslim countries. But Islamic law was rather consistently applied across the Muslim world in matters of family and personal conduct (Ruthven Reference Ruthven1984, 161). And when European legislation and state law later replaced many sharia principles during the colonial period, family law was left intact (An-Na'im Reference An-Na'im, Witte and Christian Green2012, 62–63; see also Lewis Reference Lewis2002). Accordingly, although social life was far from uniform across the Muslim world, common cultural practices regarding gender relations were created and maintained. If Islamic culture in fact plays a role concerning the empowerment of modern women, we should thus expect it to do so across the Muslim world. However, because important differences may exist, I test for potential variation between Muslim majority countries in the statistical analysis below.

To sum up the theoretical section, the study does not argue that Islam as a religion is extraordinarily repressive or that Muslim women are destined to live in servitude. Rather, opposition to women's empowerment is among the expected consequences of a historical path characterized by orthodoxy and anti-rationalism. It might be soothing to shy away from cultural explanations and see natural resources as the primary cause of the repression of women, but the probable “resource curse” comes on top of a 1000 years of frozen social hierarchy and resistance to gender equality (Allawi Reference Allawi2009; Hourani Reference Hourani1985; Keddie Reference Keddie1990). In the counterfactual situation where Muslim countries were not overrepresented among oil-based economies, orthodox traditions and institutions would still be expected to impede women's empowerment. In the next section, I discuss the empirical strategy of the study including the operationalization of women's rights. Are women's economic, social, and political status in the 21st century influenced by Islamic culture, despite Ross’ (Reference Ross2008; Reference Ross2012) alternative explanation?

DATA AND METHODS

In answering, “How can we be sure that women in the Middle East are impeded by oil, not by Islam, or the region's distinctive culture and history?”, Ross (Reference Ross2012, 124) shows that women's participation clearly covaries with oil income in the Middle East, that is, between countries with an approximately equal Islamic culture. Yet the association between oil and women's empowerment within the Middle East does not say much about Muslim women's relative rights level. Even in Tunisia, which Ross highlights as an example of the prospects of women's empowerment in the absence of extensive oil wealth, women's labor participation lags behind the average developing country (Arab Development Challenges Report 2011, 6).

Ross (Reference Ross2008; Reference Ross2012) further examines the effect of oil rents and Muslim population (as a proportion of a country's total population) on female labor force participation and the proportion of female seats in parliament in a global sample of countries. The two outcome variables are used as indicators of women's economic rights and political influence. However, by purely focusing on proportions, these indicators might not capture de facto discrimination and repression of women. To exemplify, female labor force participation was 47.9% in Nigeria in 2013, roughly the same as in Belgium (47.7%), and substantially higher than in Italy (37.9%) (Human Development Report 2013). Yet few would disagree that Nigerian women, particularly in the Muslim north, experience more systematic repression and labor discrimination than women in Western Europe (Gender in Nigeria Report 2012; World Economic Forum 2014). In the same way, female seats in parliament do not equal female access to political power. In non-democracies, legislatures often have no real power, and if there is a general resistance to women's empowerment in society, a powerless legislature might even make it easier to place women in parliament (Paxton Reference Paxton1997, 445). Iran, for instance, does have female members of parliament, but their lack of political influence is rather obvious as they are prohibited from running for president (Moghadam Reference Moghadam2003, 74). Islamic culture might thus be of consequence despite Ross’ non-findings.

To assess women's de facto rights, we thus need broader measures that take into account how countries’ laws and practices actually protect women. CIRI's women's rights indicators do just this by measuring the extensiveness of laws and government practices that protect women against discrimination (Cingranelli and Richards Reference Cingranelli and Richards2014). The indicators take a large number of internationally recognized rights into account and are coded annually based on US State Department Reports. The first indicator, women's economic rights, covers factors such as equality in pay, promotion, job security and hiring, including the right to work in the military and the police force. It also covers free choice of profession without a husband's or male relative's consent (Cingranelli and Richards Reference Cingranelli and Richards2014, 77). The second indicator, women's social rights, includes education, travel, and participation in social and cultural life as well as equality in inheritance, marriage, and divorce. It does not, however, account for domestic violence, rape, and honor killings.Footnote ⁴ The last indicator, women's political rights, includes the right to vote, join political parties, run for office, and hold government positions. It also takes women's proportion of the national legislature into account, and it does not assume that women lack political rights because a given country has no formal democratic institutions. Each measure is quantified on a four-point ordinal scale (0–3) where higher scores indicate greater respect for women's rights (see also Chaturvedi and Montoya Reference Chaturvedi and Montoya2013).

Obviously, rights such as the freedom to take a job without male consent and equality in marriage and divorce challenge the interpretation of Islamic family law presented above. If this interpretation was embedded and has survived in Muslim institutions and practices, countries with large Muslim populations must be expected to underperform on these women's rights indicators. To accommodate potential biases or measurement error in the CIRI indicators,Footnote ⁵ I also include several alternative outcome variables in the analysis: from the World Bank Database, the proportion of women in the total labor force, the proportion of women in parliament, and the proportion of girls enrolled in primary and secondary school;Footnote ⁶ from the Organisation for Economic Co-operation and Development Gender, Institutions and Development Database, the female proportion of professional and technical workers, the ratio of female to male earned income, and the proportion of women in top positions such as legislators, senior officials, and managers.Footnote ⁷

Following Ross (Reference Ross2008; Reference Ross2012), I use cross-national regression to assess the relationship because religious heritage, as described above, is considered to be constant (time-invariant) within countries given the rather limited time span for which human rights data are available. This type of analysis can show whether variables are statistically significantly correlated while potential confounding factors are controlled for. I use the mean value of each time-varying variable over a 10-year period, 1999–2008, to reduce measurement error by diminishing the influence of singular country-year scores. Although the CIRI indicators are ordinal, their 10-year means demonstrate approximate linearity, and I use ordinary least squares regression. The CIRI indicators are clearly interrelated but they are far from identical. The correlation between economic and political rights is 0.50 in the sample, between economic and social rights it is 0.81, and between political and social rights it is 0.57. This means that the indicators do touch upon different aspects of women's rights and thus might have different associations with the independent variables (cf. Drury and Peksen Reference Drury and Peksen2012, 9).

As for the first independent variables, I also follow Ross and proxy Islamic culture with the Muslim proportion of each country's population.Footnote ⁸ I use 1990 Muslim population data from the Pew Research Center.Footnote ⁹ As it seems plausible that a potential effect would stem from Muslim majority culture, I retest the models with a binary measure of Muslim majority (see Fish Reference Fish2002). Of the 166 countries in the sample, 42 have a Muslim majority population. Next, oil rents are measured as the 10-year mean of the total value of oil and gas production in thousand US dollars per capita (from Ross Reference Ross2012).Footnote ¹⁰ Concerning control variables, I include the 10-year mean of variables that are correlated with oil rents and the Muslim proportion of population and at the same time are likely to be associated with women's rights:

1. 1. Income, per capita gross domestic product (logged, in constant prices, from the Penn World Tables). Aside from oil economies, Muslim countries are relatively poor, and as a proxy for modernization, income is seen as an important driver of women's empowerment.

2. 2. MENA, a binary indicator of whether a country is situated in the Middle East and North Africa (including Turkey and Israel). Both oil economies and Muslim countries are heavily over-represented in the region. It might be a specific regional factor rather than oil or Islamic culture that impedes women's empowerment.

3. 3. Democracy, an indicator of whether governmental offices are filled through contested elections (Cheibub et al. Reference Cheibub, Gandhi and Vreeland2010). There are few democracies among oil economies and Muslim countries, and democratic institutions are expected to safeguard women's rights.

4. 4. Socialist law, an indicator of whether a country is or used be a communist state (La Porta et al. Reference La Porta, Lopez-de-Silanes, Schleifer and Vishny1999). Socialist law is underrepresented in the Middle East and gender equality was one of the legitimating principles of socialist states (Marcus Reference Marcus2014).Footnote ¹¹

The statistical results are presented in the next section. I report several model specifications for each of the main dependent variables. First, I report the baseline model where oil rents are not included and, next, the full model with oil. Then, the full model is reported, but the main predictor, Muslim proportion, is substituted for a dummy variable indicating whether a country has a Muslim majority population. In a fourth set of models, I drop all Western countries from the sample to make sure that the positive women's rights record of the West, in which the proportion of Muslims is relatively small, does not distort the findings.

As mentioned above, I also want to test whether a small number of oil-rich countries with particularly negative women's rights records could be what drives the effect of oil wealth. In a fifth set of models, I thus drop the 11 OAPEC members (see Appendix A), which are characterized by their oil wealth and Islamic culture and are among the usual suspects when it comes to repression of women. If Ross is correct in stating that oil abundance in the Middle East is the main culprit of women's lack of empowerment then Islamic culture should have no effect when these countries are dropped from the model. Finally, I present models that test the potential variation between Muslim majority countries. As discussed above, countries that were part of the Muslim world at the end of the formative period of the Islamic civilization, when Islamic law was formalized, might be particularly prone to orthodox tendencies. Of the 42 Muslim majority countries in the sample, 25 were influenced by “Early Islam” (–900 CE). The last 17 became part of the Muslim world during the classical period of the Islamic civilization (authors coding, based on Hodgson Reference Hodgson1974; Ruthven and Nanji Reference Ruthven and Nanji2004).