For many years, population growth in this Arab and Muslim world was very rapid, due to early and intense nuptiality and very high marital fertility. But globally speaking, the last twenty years have seen the onset of fertility transition in the region which, though beginning later than elsewhere, has generally advanced at a faster pace [4]. This being said, much like in other regions of the world, the speed of change varies from one country or sub-region to another, reflecting the many differences in political regimes, in social and demographic priorities and in economic systems, and the effects of the many “crises” that have hit the region over the last fifty years [5]. In many countries, large segments of the population still face insecurity and poverty.

This chronicle has two objectives. The first is to present readers (demographers or otherwise) with the most reliable, comparable and recent data on changes since 1950 and on the current characteristics of the population of each country. They are summarized in the statistical appendix, in 15 large tables corresponding to each of the topics covered in the text. The second objective is to establish a broad overview and pinpoint key trends in the region. Information is often summarized in small tables and graphs.

The map gives the geographical location of the 20 countries and their grouping into three sub-regions: North Africa (5 countries), the Arabian Peninsula (7 countries) and the Middle East (8 countries) [6]. This is the grouping used in the statistical appendix and in the text.

We will begin with a brief overview of the socioeconomic situation in the region compared with other areas of the world, with a review of the progress achieved in information systems and an outline of population change in North Africa over the last two millennia. We will then consider successively: population growth since 1950; the patterns of demographic transition; nuptiality (age at marriage, celibacy, polygyny and union dissolution); fertility levels and trends; proximate determinants of fertility (contraception, breastfeeding, etc); overall mortality and maternal mortality; child mortality and health from age 0 to 5 (conditions of delivery, vaccination, malnutrition); age structure; urbanization; international migration and displaced populations; and lastly gender inequalities in access to education.

Our approach is essentially descriptive and involves the study of levels and trends in all countries, the discussion of inequality (in terms of education and residential environment) in countries at different stages in the transition or development process. In certain cases we will study the relation between the demographic indices of the countries and their indicators of social, economic, human and health development, which are grouped by country in Appendix Tables A.14 and A.15. It would have been over-ambitious to set our sights any higher.

Our main data sources are the databases of the various United Nations agencies (Population Division, UNDP, WHO, UNESCO, etc.), which provide crucial records of the major changes since 1950 and key specific indicators, and second, the various demographic and health surveys conducted in these countries over the last 15 to 20 years. We also make use of recent studies relating to the region or to a particular sub-region or country.

Table 1
Some economic, demographic and social characteristics of the region compared with other major world regions c. 2002

On average, once again, the region is not the most disadvantaged in economic terms: thanks notably to the high oil and gas revenues of certain countries [8], the per capita gross national income (USD 5,800 in 2002) is double that of South Asia, higher than that of East Asia (USD 4,800), but still below that of Latin America (USD 7,200). This average masks enormous differences between countries in the region (Appendix Table A.14). For example, the per capita GNI is USD 770 in Yemen (one of the world’s lowest), around USD 3,700 in Morocco and Egypt, and around USD 19,000 in Kuwait and Israel. Apart from Yemen, the six other countries of the Arabian Peninsula are very rich.

The overall situation with regard to education in the region is quite poor however (Table 1), with an adult illiteracy rate of 32% in 2002. Excluding Iran (24%), Turkey (15%) and Israel (5%), the 17 Arab countries have an illiteracy rate of 37%, below that of South Asia, but identical to that of sub-Saharan Africa. Here too, large disparities are observed between countries: adult illiteracy ranges from 5% in Israel to more than 60% in Iraq; it is below 15% in five countries (Bahrain, Israel, Jordan, Lebanon and Turkey), and above 40% in four others (Egypt, Morocco, Yemen and Iraq). We will return to this important question of education later on.

Finally, in terms of human development (composite index including income, education and life expectancy), the region is well ahead of South Asia or sub-Saharan Africa, but lags far behind Latin America and East Asia (Table 1). In the UNDP world ranking of 177 countries for 2002 (Appendix Table A.14), Israel is in 22nd place, the oil-producing countries of the Arabian Peninsula are clustered around the 45th place, North Africa is between the 58th (Libya) and 125th place (Morocco) and Yemen is 149th. The region is characterized by a broad diversity of economic, social and health situations.

Their land areas (Map) range from just 700 sq. km in Bahrain to almost 2.4 million sq. km in Algeria. Four countries (Bahrain, Qatar, Lebanon and Palestine) cover less than 11,000 sq. km, three (Morocco, Yemen and Iraq) total around 500,000 sq. km, and five occupy more than a million sq. km (Algeria, Libya, Egypt, Saudi Arabia and Iran). But in many of these countries, vast regions are covered by deserts and steppes. Historically, populations have settled mainly in the coastal zones (Mediterranean, Red Sea, Persian Gulf) or along rivers (Nile, Tigrus, Euphrates, etc.), in the most fertile or least arid regions. As pointed out by Boustani and Fargues (1990, p. 40), “The Arab world has an exceptional feature: it is inhabited solely around its edges, and the centre of the settlement map is empty. Densely populated areas are separated by vast expanses of practically unoccupied land”.

Population densities vary substantially from one country to another. In 2000, crude densities (Appendix Table A.2) ranged from 3, 11 and 13 inhabitants per sq. km in Libya, Oman and Algeria, to 340 in Lebanon, 525 in Palestine and 960 in the tiny Gulf state of Bahrain (one of the highest densities in the world). Most countries have between 30 and 90 inhabitants per sq. km, though these national densities conceal enormous disparities within countries. For example, the coastal regions with their many large cities commonly exceed 3,000 inhabitants per sq. km, the Nile Valley delta has one of the world’s highest rural population densities, and the Gaza strip counts 3,600 inhabitants per sq. km.

But these crude densities, calculated as the number of inhabitants divided by the total surface area of a country, are always, for this region especially, poor indicators of the relationship between agricultural resources and population pressure (Mathieu and Tabutin, 1996). If, for example, densities are recalculated on the basis of arable and permanent crop land areas, then a very different picture emerges (Appendix Table A.2). Apart from the countries of the Arabian Peninsula which all have predictably high values, density in Algeria for example rises from 13 inhabitants per sq. km to 372, Turkey from 88 to 266, Iran and Iraq from around 50 to more than 400. But in some countries the change is even more radical, notably in Jordan and Egypt, which respectively exceed 1,200 and 2,000 inhabitants per sq. km of arable and permanent crop land [9].

The demographic weight of these 20 countries also varies substantially (Figure 1). The region includes two small countries (Bahrain and Qatar, with around 700,000 inhabitants), eight countries with between 2 and 7 million inhabitants, and six countries with between 20 and 32 million (including Algeria, Morocco and Iraq). The three countries with the largest populations, Turkey, Egypt and Iran, each with more than 70 million inhabitants, alone account for 52% of the region’s total population. The 17 Arab countries make up 65% of the total. By sub-region, North Africa represents 36%, the Arabian Peninsula 14% and the Middle East 50%.

Overall, substantial progress has been achieved over the last 40 years in vital registration. The system exists everywhere [10], though its effectiveness varies, notably for the declaration of deaths and marriages. According to the United Nations, coverage of mortality in 2001 exceeded 90% in only eight of the region’s 20 countries (Libya, Algeria, Tunisia, Egypt, Israel, Kuwait, Qatar and the United Arab Emirates). The situation in Yemen and Morocco is especially poor. Moreover, there is a scarcity of clear and detailed annual publications which means that excepting a few countries, vital records do not yet provide a reliable tool for monitoring the levels and tempo of demographic phenomena.

For censuses, the situation is better, with a total of 84 censuses since 1960 (Table 2). Most states have conducted four or five, at intervals of between 10 and 12 years (Appendix Table A.1). The United Arab Emirates, Kuwait and Turkey have done even better, totalling between six and eight thanks to the adoption of a five-year periodicity in the 1960s, 1970s and 1980s. Three countries have conducted only two censuses however: Oman [11], Palestine (in 1967 and 1997) and Yemen (the country was unified in 1990). Lebanon, for its part, is an exceptional case, with just one census dating back to 1970 [12].

Table 2
Number of censuses and national demographic surveys since 1960 in all 20 countries of the region

With regard to national surveys on fertility, the family and child health, the situation is much more varied (Appendix Table A.1). Most of the surveys [13] since the early 1970s have been conducted under five major international or regional programmes:

Overall, a relatively large number of national surveys have been conducted across the region: a total of 63 since 1970, and 41 since 1990, though there are marked differences between countries. Three countries – Morocco (8 surveys), Egypt (6) and Jordan (8) – are particularly well covered by the different programmes. Many have between three and five surveys, and the least well covered, with one or two surveys only, are the six Gulf States, Palestine, Iraq and Israel.

For our study of fertility, nuptiality, child mortality and child health, we will use the available results [15] of these surveys, the most recent ones in particular. We will proceed with all necessary caution when comparing the various indicators, since although the main goals of these different survey programmes were quite similar, the questionnaires are not always strictly identical and data is doubtless of variable quality.

General studies of the history of the world populations are also highly discreet and circumspect about the region. To give a general idea of population trends, the results of a recent overview of changes in the population of North Africa since the year 1000 (Tabutin, Vilquin and Biraben, 2002) are summarized in Table 3. These orders of magnitude were established using data from the most “serious” historical studies based on rigorous scrutiny of sources and information matching, and from recent research by J.-N. Biraben [18].

Table 3
Orders of magnitude of the total population of North Africa (Morocco, Algeria, Tunisia, Libya, Egypt) from 1000 to 1950

The population of North Africa declined substantially between the eleventh and thirteenth centuries following the Hilalian invasions from southern Egypt and the agricultural and political disruption that ensued. This is universally agreed, by both contemporaries and historians of the period (Reinhard et al., 1968, p. 83): “After the eleventh century, the situation worsened, in the countryside especially. With the destruction of plantations and irrigation systems, agriculture was reduced or wiped out, to be replaced by nomadism. There is clear evidence that this transformation drastically reduced the population, for example in North Africa, Egypt and Morocco”. Population decline accelerated in the fourteenth century, this time due to the Black Death of 1348-49 which swept across the entire Mediterranean basin and which, throughout the century, returned to the region every 10 to 15 years. Between 1300 and 1400 (Table 3), North Africa lost around one quarter of its population, which fell from 9.5 to 7.2 million inhabitants [19]. Between the fifteenth and eighteenth centuries, it started rising again, though frequent epidemics, famine and war slowed down the rate of increase [20]. As is the case in other developing regions, the population started increasing more quickly towards the late nineteenth century, from 1870 onwards, when the “colonial” administration introduced the first major measures to promote health, urban hygiene and transport. But it was not until the late 1940s that the major killer diseases such as cholera, typhus and malaria were practically eradicated. The population of North Africa grew by around 60% from 1850 to 1900 (Table 3) and almost doubled between 1900 and 1950 [21]. The pace of growth continued to increase thereafter.

The rate at which population growth is slowing down varies from one country to another. Figure 3, based on the most recent United Nations data (2005), compares pre-transition growth rates of 1970-74 with those of 2000-04 for each of the 20 countries. At one extreme we find two countries, Yemen and Palestine, whose very high population growth rates (3.2 and 3.5% per year respectively), have even increased slightly since the 1970s. At the other extreme, in six or seven countries – Kuwait (which began very high), Algeria, Iran, Bahrain, Lebanon and Tunisia – the growth rate has decreased by almost 50%. Between these two extremes, the growth rates in Iraq [23], Saudi Arabia, Syria and Egypt are declining more slowly. The same applies to Qatar and Israel, which began from much lower growth rates in 1970-74. All in all, there is no clear relationship between the rates of decline and the starting level in 1970-74.

In sum, disparities in the rate of population growth in the region are much larger today than in the past: thirty years ago, most countries had annual growth rates of between 2.4 and 3.5%, while today they range between around 1.2% (Tunisia, Lebanon, Iran) and more than 3.2% (Yemen, Palestine). Out of the twenty countries, twelve have fallen below the 2% threshold, while only three are below 1.5%.

Though the growth rate is indeed slowing down, absolute population growth, and hence demographic pressure on essential goods and services (education, housing, food, employment, etc.) is still very strong. The region’s population has quadrupled since 1950 (see Appendix Table A.2), more than doubled since 1970 and increased by 35% since 1990. The multiplier coefficients [24] of the population in certain countries since 1970, over a period of 35 years, are as follows:

Saudi Arabia: 4.3	Iran: 2.4
Jordan: 3.5	Israel: 2.3
Palestine: 3.4	Morocco: 2.1
Yemen: 3.3	Egypt: 2.1
Bahrain: 3.3	Turkey: 2.0
Syria: 3.0	Tunisia: 2.0
Algeria: 2.4	Lebanon: 1.5

These very high figures are the consequence of recent fertility history, but also, in certain cases, of international migration (refugees, workers). And this population growth is set to continue over the coming decades.

Though recent and updated, the 2004 projections must be seen as one possible future among many, liable to be revised upward or downward over the coming years in line with changing patterns of fertility and, in certain cases, migration. A simple comparison of the two sets of United Nations projections for seven major countries of the region, spaced just four years apart (in 2000 and 2004), illustrates the degree of uncertainty of projections at national level for the horizons 2025 and 2040 (Table 4). For three countries – Turkey, but above all Egypt and Iraq – the projected populations in 2025 and 2040 have been revised substantially upwards, while for four others, including Iran and Saudi Arabia, the opposite is the case. But the diverging trends tend to balance each other out and the total population of these seven countries remains practically unchanged (almost 446 million in 2025, around 510 million in 2040).

Table 4
Comparison of population projections (in millions) for 2025 and 2040 established by the United Nations in 2000 and 2004 for 7 major countries

Appendix Table A.2 shows the population change since 1950 by country and sub-region and the prospects for 2010, 2025 and 2040. Though growth is expected to slow down, it will remain high in most countries (between 1.5% and 2.5% per year), leading to rapid population growth over the next 35 years. Under the United Nations medium scenario (2005), the population of the whole region (Figure 4) will increase from 421 million inhabitants in 2005 to 575 million by 2025 (+37%), then to 667 million in 2040 (+58%). All in all, the region is set to reach 800 to 900 million inhabitants by the end of the twenty-first century, with a doubling of its population over a period of 80 to 90 years. By 2040, the Arab world (17 countries) will have population growth (69%) which is double that of Turkey, Iran and Israel taken together (38%) (Table 5). In 35 years, the region will very probably have a population twice as large that of the 17 major countries of southern and western Europe (Figure 4).

Table 5
Population (millions) and growth (%) in the sub-regions from 2005 to 2040

At sub-regional level (Table 5, Figure 4), the populations of North Africa and the Middle East will grow at a similar pace up to 2040 (by around 50%), but at only half the speed of the Arabian Peninsula (104%) which, driven by Saudi Arabia and Yemen, will double its population in just 35 years due to a slower decline in fertility than elsewhere.

The populations of all the countries will rise over the next 35 years. The overall increase will be large, though the situation will vary from one country to another. According to the United Nations prospects by country, growth rates will be very rapid in certain countries, such as Yemen, Saudi Arabia, Iraq [26] and especially Palestine, leading to a doubling of population size in 35 years, and much more contained in others (Algeria, Morocco, Lebanon, Israel). By 2040, three countries will have close to 100 million inhabitants (Egypt, Iran and Turkey), four will verge on 45 million (Algeria, Morocco, Saudi Arabia and Yemen), and five will have 8 to 10 million (Libya, United Arab Emirates, Israel, Jordan and Palestine).

Under these scenarios, which are practically inevitable despite a certain margin of error, it is not difficult to imagine the scale of human pressure – already posing severe problems in the region today – on ecosystems (oceans, coastal regions and in some cases forests), on land and on a resource as fundamental as water.

We will now examine the two components which, excluding migration, underlie population growth in all geographical entities, namely the birth and death rates [27], whose respective variations over time determine what is commonly known as the transition model.

Overall, the transitional histories of North Africa (5 countries) and the Middle East (8 countries) have followed a similar pattern over the last 50 years, with practically identical initial and final levels (Figure 5, diagrams A and C). The Arabian Peninsula is very different however, with a later and sharper decline in the birth rate (31‰ today), but earlier progress in mortality reduction. As already mentioned, the population grew more rapidly in this sub-region as a result, and continues to do so today.

But the histories of the individual countries comprising the three sub-regions are much more varied. The country data are given in Appendix Table A.3. As an illustration, Figure 6 gives the transition diagrams from 1950 to 2005 for eight countries representative of the region. There are five main models or transitional situations:

Over the next two to three decades, the final transition paths taken by each society, doubtless many and varied, will be determined by their economic, educational and cultural development, their demographic and social policies and their political stability.

We will now look more closely at the major nuptiality, fertility, mortality and migration trends upon which these past, present and future demographic processes so directly depend.

Appendix Table A.4 presents the mean ages at first marriage for men and women by country around 1975 and 1996, the age differences between spouses at these two dates, and the recent situation for polygyny and for the proportion of women never married. For the three Maghreb countries plus Iran, more detailed information on the changes between 1960 and 2000 in age at first marriage and the proportions never married at ages 20-29 is shown in Appendix Table A.5.

These changes in the timing of women’s marriages have been very rapid in the three Maghreb countries (Figure 7). The increase in women’s age at marriage began in Tunisia in the 1960s, and ten years later in Algeria and Morocco. Today, mean age at first marriage among women in Algeria and Tunisia currently stands close to 30 years at the national level, one of the highest in the world. A slower pattern of change is observed for Iran (Figure 7), which in broad outline is also that for Iraq, Jordan and Turkey and probably also for Saudi Arabia and Oman.

Ages at first marriage among men, which were already relatively late in the 1970s (commonly around 25), have also risen, though on the whole more slowly. They now stand at around 27 or 28 in a majority of countries (Appendix Table A.4). As is the case for women, these ages are highest (30 or above) in Algeria, Libya and Tunisia. They are lowest (around 25) in Turkey, Saudi Arabia, Iran and Palestine.

Figure 8 clearly illustrates the large differences between countries in age at marriage, for both women and men, which contrasts with the relative uniformity that prevailed in the 1950s and 1960s. It also shows that at the aggregate level the age difference between spouses is not, or only loosely, related to women’s age at marriage [30].

This constitutes an unprecedented change in the marriage regime, and is probably related to the longer time spent in education and to cultural changes, but also to the economic crisis (unemployment among young people) and to the housing shortage. This extended single life is often lived out within the family sphere, in contexts characterized sometimes by low economic activity among women, as is the case in Algeria, but always by a strong social control over the lives and sexuality of young people, particularly women.

Getting married remains the predominant social norm throughout the region, for women as well as for men. In the 1970s, as in the 1990s, permanent celibacy [31] affected barely 1% to 5% of the female population of the countries [32] (Appendix Table A.4.). In other words, practically all the women in these oldest cohorts (aged 40-49) have been married, in many cases at an early age, as mentioned earlier. In all probability, however, the near future will see the end of universal marriage and the appearance, as in other continents, of a “structural” never-married population. On the basis of the current proportions of women (and men) still single at ages 30-39 – ages at which getting married becomes more difficult – it is reasonable to predict a substantial increase in the proportion never marrying in the years 2000-10, particularly in the Maghreb countries.

Contrary to what might be expected in a region that has experienced urbanization and become exposed to external influences, and where education is on the increase, kin marriage remains valued and widespread. In a majority of the ten countries included in Table 6 [37], around 40% of all marriages are between kin members, with Turkey (23%) at one extreme, and Kuwait (53%), Saudi Arabia (58%) and Palestine (66%) at the other. First cousin marriage is universally encouraged.

Table 6
Distribution of marriages by kinship ties between spouses in selected countries (%)

Though endogamy has declined locally, in Tunisia for example (Ben Salem and Locoh, 2001), it has maintained its position in Egypt and in Algeria (Table 6). It remains a key element of marriage systems in the region, even “if it has changed its meaning, form and direction in the new generations, by expanding to include a wider family circle” (Ouadah-Bedidi, 2004, p. 431).

Nevertheless, in six of the ten countries in Table 6, around two in three marriages are between individuals who are not related. The situation is not one of complete freedom of marital choice for young men and women, but families now play a less important role.

As is the case in other world regions, divorce and remarriage are among the least well-documented demographic phenomena, due to the lack of data for both earlier and contemporary periods [41]. The few recent studies have on the whole concluded in favour of a definite decline in divorce in the Arab societies, even though levels remain non-negligible. The divorce rate per 100 marriages [42] in Algeria is believed to have dropped from 38% in 1900 to 13% in the 1960s; in Egypt from 32% in 1941-45 to 14% in 1996-2000 (Fargues, 2002); in Tunisia it remained stable, at around 15%, from the 1960s to the 1980s. In the 1970s, according to the WFS surveys, the probability of divorce after 30 years of marriage was 14% in Egypt, 24% in Morocco, and only 6% in Syria. In Algeria, it dropped from 40% (1970 survey) to 19% (1992 survey). If the data from these few countries are reliable, we are indeed witnessing a relative stabilization of marriage. Divorce, on the other hand, appears to occur as early as ever and would seem to be spreading to all groups in society [43].

As in other regions of the world, remarriage is less frequent among women than among men, and takes place after a slightly longer interval. In Algeria, twice as many men remarry after a divorce than women (Ouadah-Bedidi, 2004). In Iran, the frequency of female remarriage may even have fallen between 1977 and 1991 (Ladier-Fouladi, 2003).

Table 7
Median age at first marriage of women^(a) by residence and educational level in five countries (years)

An urban-rural contrast in median age at first marriage [44] is observed everywhere but to varying degrees depending on the country. The differences between median ages are still negligible in Yemen where half of women were already married at age 16, small in Jordan and in Turkey, and large (around three years) in Morocco and Egypt. The educational level of women is shown – a classic result – to have an enormous influence on their age at first marriage. Between illiterate women and those with a primary education, the increase is already perceptible (including in Yemen), and those with a secondary education or above marry at a much later age, ranging from 23 years in Egypt to nearly 28 years in Morocco.

The same pattern is observed for other components of nuptiality. The period of single life is much longer in the cities, notably the largest ones, and of course among highly educated men and women, for whom permanent celibacy is likely to become a real prospect in the near future. Polygyny seems to be more present in rural than in urban areas, and among the least educated groups. This also appears to be the case for family endogamy which, though universally present, is tending to decline, in Algeria at least, as the educational level of men and women rises. Similarly, divorce is often more frequent in the cities than in rural areas (Libya, Algeria).

These combined factors mean that several nuptiality regimes co-exist today in the same country, with different models by social group, living environment and region. This is scarcely surprising insofar as modernity, cultural openness, the relations between men and women, the effects of economic crisis and of development all vary from one sub-population to another, in North Africa and the Middle East, as elsewhere in the world. These diverse and, in some cases, rapid changes in marriage patterns have an appreciable influence on fertility levels and trends in these countries.

We can illustrate the diversity and specificity of the transition histories using some well-documented individual cases.

Figure 11 shows clearly that the fertility transition follows diverse paths in North Africa and the Middle East, as in the rest of the world. So as we have seen, overall fertility depends partly on changes in age at marriage [49], and these changes have been substantial in some countries. We will now move on to look at marital fertility.

Table 8
Total and marital^(a) fertility rates for selected countries in the 1990s (number of children per woman)

Marital fertility remains relatively high in the region. Excluding Turkey (3.1 children) and probably Israel, a woman who marries at age 15 and stays married until age 50 has around 6 children (from 5.1 in Egypt to 8.0 in Yemen and Saudi Arabia). In a country like Algeria, where overall fertility is now very low (2.2 children), marital fertility still stands at 6 children (Oudah-Bedidi, 2004). But it has fallen considerably, from 11 children in 1970 to 8.3 in 1992 and then to 6.0, as a result of increased contraceptive practice among married couples.

The indicator of demand for children varies between countries. Among married women aged 25-34 years, the ideal number of children ranges from 2.5 in Turkey to 4.4 in Palestine and 4.6 in Yemen (Appendix Table A.7). Ideal family size is close to observed fertility levels (TFR) in Jordan, Turkey and Morocco, markedly higher in Algeria (by almost one child), and substantially lower in the two high fertility countries (Yemen and Palestine).

The indicators of nuptiality, examined earlier, and fertility all suggest that profound changes in the traditionally dominant family model are under way in the region. Overall, there is a shift from the large family model to that of the smaller nuclear unit, doubtless partly in response to the difficult living conditions in many countries (for employment, housing, cost of living, etc.), but also because of the environment (urbanization) and cultural change. At the national level, the model is not (yet) that of two children per couple, since reproductive norms and behaviour still vary substantially, even within countries.

Table 9
Total fertility rates and mean ideal number of children^(a) by women’s residence and educational level, and for the extreme regions of the country (five countries)

Urban and rural areas are everywhere quite distinct as regards both actual and ideal numbers of children. Urban fertility levels are appreciably lower in all the countries, even in Yemen where the transition has only just started [51], but on the whole these urban-rural differences decreased substantially during the 1990s. Likewise, the desired number of children is consistently lower in the cities. These contrasts between urban and rural areas have varied over time, like those observed between social groups and educational levels. The case of Algeria (Table 10) provides an example. From a low level in the 1970s with universally high fertility, the disparities widened in the 1980s when changes in norms and behaviour began to spread across the cities. They then narrowed again in the 1990s due to the rapid “diffusion” of these same changes in the rural areas.

Table 10
Total fertility rate in Algeria, 1970-2002, by residence (number of children per woman)

Fertility in some capitals or major cities now stands at particularly low levels. In cities like Algiers, Rabat, Casablanca and Istanbul, the fertility transition is complete, with slightly fewer than two children per woman. In these contexts, demand for children is low, marriage is late for both men and women, and contraception is widely practised.

Similarly, the education of women has an increasingly important role in the region, as elsewhere in the world. Except in Jordan [52], receiving a primary education is enough in itself to modify behaviour appreciably; and this effect is greatly accentuated by access to secondary or higher education (Table 9). The largest inequalities associated with access to education are observed in Turkey and Morocco. As we have seen for age at marriage – and as we shall see for contraception and breastfeeding – education considerably modifies the role of the intermediate fertility variables.

Similarly, the countries still exhibit marked regional diversity, with differences in fertility ranging from 0.5 to 2.2 children per woman between the extreme regions of the five countries (Table 9). These differences remain particularly large in Turkey (western region/eastern region), in Morocco (Grand Casablanca/Tangiers-Tetouan), and in Egypt (Urban Governorates/Upper Egypt).

All in all, the comparative social and spatial homogeneity observed during the 1970s and 1980s clearly belongs to the past, even if the countries are at different stages of transition.

The correlations are statistically significant (R² = 0.29 for HDI, R² = 0.24 for infant mortality), but the points are widely dispersed. For example, at an identical level of human development (HDI around 0.73), fertility values range from 2.0 children in Tunisia to 4.1 in Saudi Arabia and even to 5.6 in Palestine. Conversely, for the same fertility level (around 3 children), the human development index ranges from 0.62 in Morocco to 0.91 in Israel. The same is true for infant mortality: Tunisia with 2.0 children, Saudi Arabia with 4.1 children and Palestine with 5.6 children are at a low level (around 22‰); at 3 children or thereabouts, mortality ranges from 5‰ in Israel to 40‰ in Egypt. All in all, in many countries of the region, notably in the Middle East and Arabian Peninsula, current fertility is substantially higher than it would be elsewhere for the same level of development and health. The absence of a clear relationship also confirms the distinctiveness of national histories and the diversity of the paths followed by the fertility transition.

In the Arab and Muslim societies of the region, the proximate determinants [56] of fertility followed a traditional model comprising early marriage for women, absence of contraception within marriage, relatively short post-partum abstinence, and widespread and prolonged breastfeeding. As already noted, this pattern gave very high levels of total and marital fertility, of around 8 and 10 children respectively, and quite short birth intervals, of the order of 24 months. What is the situation at present? The most recent data by country, covering the period 1995-2003 [57], are given in Appendix Table A.8.

The limited recent and comparable data point to an appreciable lengthening of birth intervals (Appendix Table A.7). Calculated for all women aged 15-49, the median last birth interval is 34 months in Egypt, 37 months in Turkey and 42 months in Algeria and Morocco. This tendency indicates a growing use of birth spacing practices. By contrast, in Palestine and Yemen intervals remain of traditional length, at 22 and 28 months respectively.

Post-partum abstinence is traditionally short in the Arabo-Muslim world [58], where reference is often made to a norm of 40 days [59]. To judge from the results for the last 30 years, it remains close to this duration, with a median value of 1.7 to 1.9 months in the latest DHS surveys (Appendix Table A.8). It has changed little over time; it is simply a little shorter in the cities and among educated and younger couples [60].

Breastfeeding [61] continues to be widely practised in the region and generally for fairly long periods, with median durations of 11 or 12 months in Palestine, Turkey and Libya, between 13 and 15 months in Algeria, Morocco, Jordan and Tunisia, and even 18 months in Egypt and Yemen (Appendix Table A.8). On the whole, no major changes have occurred in the last fifteen years. This cross-country variability is nevertheless accompanied by substantial diversification in norms and practices within countries. In Egypt (DHS 2000), for example, the very long breastfeeding durations of the national average are being maintained in the cities as well as in rural areas, and vary little with the educational level or residence of the mother. In Turkey (DHS 1998), on the other hand, the practices are more differentiated between places of residence and social groups, although overall breastfeeding duration remains at around 12 months. The changes in behaviour for certain groups are clearer still in Morocco. Relatively weak at the start of the 1980s (1979-80 survey), the urban-rural differentials have grown considerably since then. In 2003 the median duration of breastfeeding was only 11 months in urban areas versus 16 in the countryside; it even fell to 9 and 8 months among women with the most education and of the highest social group, versus 16 months among those who were illiterate and in the poorest group. This quite radical break with tradition seems to have been less abrupt in Tunisia (Ayad and Jemai, 2001). No significant differences in breastfeeding between boys and girls are observed anywhere.

Today, knowledge of contraceptive methods is widespread in the region. If the results from the latest surveys can be relied upon, between 90% and 98% of married women in most countries report knowing at least one modern method [62]. Contraceptive practice now depends essentially on the demand from couples and on the availability of services (geographical or cultural accessibility, cost and quality).

Around the year 2000 [63], modern contraceptive prevalence among married women aged 15-49 was over 40% in 9 countries of the region, between 20% and 40% in 9 others, and remained at particularly low levels only in Yemen (10%) and Oman (18%). The three Maghreb countries, and Egypt, Israel and Iran, stand out especially, with current prevalence levels of around 55% (Figure 14). By contrast, with the exception of Kuwait, levels in the countries of the Arabian Peninsula are clearly lagging behind (around 30% in 1996).

The three Maghreb countries, Iran and, to a lesser degree, Lebanon have experienced a contraceptive revolution that has surprised by its speed. Such a rapid change could not have been imagined barely 20 years ago. In Morocco, for example, modern contraceptive prevalence has climbed from 19% to 55% in 22 years, and in Algeria from 22% to 52% in 20 years.

Overall, and consistent with what is observed in other regions, a strong inverse relationship exists between modern contraceptive practice and total fertility (Figure 14), with a coefficient of determination of 0.73. The phenomenon is especially clear in the extreme situations, somewhat less so in the intermediate countries: thus a contraceptive prevalence of around 30% gives fertility ranging from 3 to 6 children per woman. The relationship with marital fertility, which we have not presented due to too few data, would have been even stronger.

As in other regions of the world, contraceptive use has spread – and continues to do so – from cities to rural areas, and from the most educated and affluent couples to the most disadvantaged couples. The fact that not all countries are yet at the same level results in clear differences between geographical and social groups by country. In the most advanced countries in terms of family planning provision (the examples of Egypt and Morocco in Table 11), social and geographical disparities are now relatively small, though educated and urban women still have a slight advantage. The disparities are much larger in countries where contraceptive use is at intermediate levels (Turkey and Jordan in Table 11). And they are enormous in Yemen, where policy on contraception is in its very early stages.

Table 11
Proportion of married women (15-49) using a modern contraceptive method, by residence and educational level (%)

The modern methods of contraception used for birth spacing or birth stopping vary markedly between countries [64]. The pill is still the most widely used method in Morocco and Algeria where it has been adopted respectively by 73% and 90% of users of modern methods, whereas the IUD has become by far the most popular in Egypt, Tunisia, Palestine and Jordan, with between 55% and 65% of users. On the whole, the condom remains unpopular, except in Turkey (22%). Female sterilization, which for long was viewed negatively and seldom used, is on the increase in countries such as Morocco, Jordan and Turkey where between 5% and 10% of married women say they use a modern method; it has even become the second method in Tunisia, after the IUD. The traditional methods reported are mainly withdrawal and periodic abstinence and in some cases prolonged breastfeeding [65].

But to what extent can the substantial reductions in total fertility described earlier be attributed to the increased age at marriage of women and to what extent to the adoption of contraception among married couples?

Table 12 ranks the three main fertility-reducing factors for each country in the 1970s, 1980s and 1990s. During the 1970s, the reduction in fertility (relative to the theoretical maximum) was due primarily to post-partum sterility, but the effect of nuptiality was already perceptible whereas that of contraception was weak [68]. The latter began to develop in the three countries from the mid-1980s and increased its contribution to fertility reduction in the 1990s, as did nuptiality, through a large rise in women’s age at marriage (see Section VII). In the future, the effect of contraception is likely to increase further and that of nuptiality to stagnate, given the already late ages at marriage. Finally, the effect of breastfeeding (which has hardly altered over 25 years) will be determined by the intensity of changes in behaviour.

Table 12
The three main factors in fertility reduction (ranked by order of importance) in the Maghreb in the 1970s, 1980s and 1990s