arab econ
at the beginning of July 2002, the United Nations Development Program
released the Arab Human Development Report 2002, the first in a continuing
series. Written by Arab scholars, it was a no-holds-barred indictment of
economic, social, and political backwardness in the Arab world. American and
European audiences received the report with tremendous enthusiasm. Time magazine (December 30, 2002) called it “perhaps the most important volume published
in 2002” (Elliott, 2002). In a review entitled “Self-Doomed to Failure,” The Economist
(July 6, 2002) asked, “What went wrong with the Arab world? Why is it so stuck
behind the times?”
The authors of the Arab Human Development Report 2002 did not claim to be
making a balanced assessment of human development in Arab countries, even
going so far as to construct an “alternative human development index” designed to
show the “deficits” of the Arab world. Nevertheless, their view has become the basis
of conventional wisdom regarding Arab development performance and prospects.
Given the attention currently focused on the Arab world, in part as a result of
adjustments in U.S. foreign policy, a fresh look at Arab socioeconomic performance
is in order.
The Arab world is defined by language rather than ethnicity. The League of
Arab States, formed in 1945, consists of all countries in which (a dialect of) Arabic
is the spoken language of the majority. It is useful to compare the human development diversity of the Arab world to that of Latin America, another vast geographic area defined by language and culture. To be consistent, we use a linguistic
definition of Latin America, including all Western Hemisphere countries in which
a language from the Iberian Peninsula (Portuguese or Spanish) is predominant
and excluding countries such as Haiti. For each indicator used to compute the
Human Development Index (HDI) produced by the United Nations Development
Program, we compare the range of values for Arab League countries to the range
of values for Latin American countries in 2006. The range of Arab values is more
than double the range of Latin American values for life expectancy, more than
25 percent higher for literacy, more than triple for school enrollment, and more
than six times greater for income (adjusted for purchasing power).
From a socioeconomic point of view, then, the Arab world is too diverse to be
a useful aggregate. The source of its tremendous human development diversity is
easily identified. The Arab world includes both destitute sub-Saharan African
countries and countries with fabulous per capita oil and natural gas resources:
Qatar vies with Luxembourg for the title of world’s richest country on the basis of
per capita income. Moreover, both sub-Saharan Africa and fuel-endowed countries
are well known for poor socioeconomic performance. Easterly and Levine (1997),
for example, have written about “Africa’s growth tragedy” and Collier and Gunning
(1999) ask, “Why has Africa grown so slowly?” (By “Africa,” both articles mean
“sub-Saharan Africa.”) The problems of fuel-exporting economies have given rise to
a literature on the “oil curse” (Gelb, 1988), which is a more wicked form of the
“natural resource curse” (Sachs and Warner, 2001). Our strategy in this article is
therefore to disaggregate the Arab world into Arab sub-Saharan Africa, Arab
fuel-endowed economies, and a remainder we call the Arab Mediterranean, and to
compare these three Arab worlds to non-Arab sub-Saharan Africa, non-Arab fuelendowed economies, and the rest of the non-Arab world. We will evaluate Arab
socioeconomic progress from 1970 to as close to the present as the data allow. (We
will note the cases where starting from 1960 would qualitatively change our results;
but starting in 1970 allows us to include data for more countries in our analysis.)
Table 1 shows the 2007 populations of the three Arab worlds and Figure 1
shows their locations on a map. We have omitted the territory labeled “Palestine”
on the Arab League web site and “West Bank and Gaza” by the World Bank since
we will drop all territories that are not United Nations member states from our
analysis below. Figure 1 is potentially misleading in that the Arab Mediterranean
contains nearly half the Arab population, but only 14 percent of Arab land area. We
have grouped Yemen with Arab sub-Saharan Africa: it is at the same latitude and
separated from sub-Saharan Africa only by a strait less than 30 kilometers wide.
How did we identify the fuel-endowed economies? It is standard in the literature to indicate fuel endowment by the ratio of fuel exports to total exports or fuel
exports to GDP. However, either of these criteria tends to select countries for
failure to develop alternative sources of exports, which is tantamount to selecting
them for failure to develop. We instead took a more direct approach to measuring
fuel endowment and used “proven fuel reserves per capita.” In identifying fuelendowed countries, we used fuel reserves data as close to the beginning of our
evaluation period as possible, since it would not make sense to classify a country as fuel-endowed for the purpose of evaluating its socioeconomic performance beginning in 1970 if its reserves were only discovered in 2000. The authoritative data Figure 1 The Three Arab Worlds Table 1 Population of Arab States, 2007 (in thousands; total 324 million) Sub-Saharan Africa Fuel-endowed countries Remainder (“Arab Mediterranean”) Comoros 626 Algeria 33,853 Egypt 75,467 Djibouti 833 Bahrain 753 Jordan 5,719 Mauritania 3,121 Iraqa 28,221 Lebanon 4,097 Somalia 8,696 Kuwait 2,663 Morocco 30,861 Sudan 38,556 Libya 6,156 Syrian Arab Republic 19,891 Yemen 22,383 Oman 2,600 Tunisia 10,248 Qatar 836 Saudi Arabia 24,196 United Arab Emirates 4,365 Total 74,215 Total 103,643 Total 146,283 (23%) (32%) (45%) Source: World Development Indicators Online, 2008, http://go.worldbank.org/6HAYAHG8H0. a The source for the Iraq population is CIA World Factbook (July 2008 est.). James E. Rauch and Scott Kostyshak 167 source is the annual BP Statistical Review of World Energy, which publishes crosscountry comparable estimates of oil and natural gas reserves dating back to 1980. A minor issue arises regarding how to aggregate oil and natural gas reserves. We converted the oil and natural gas into BTUs (British thermal units) using standard conversion factors and computed 1980 BTU reserves per capita. This procedure yielded a ranking with a clean break at Algeria, which has 69 percent more reserves per capita than the next country down and nearly five times more reserves than the next Arab country down. The Arab countries identified as fuel endowed by this ranking are the same as those identified using the Global Development Network Growth Database criterion. Under this criterion, a country is fuel endowed if fuels accounted for more than 50 percent of total exports of goods and services for the period 1988 –1992. The non-Arab countries identified as fuel endowed by this criterion are Brunei, Iran, Norway, Trinidad and Tobago, and Venezuela. For comparison, using the export-based criterion from the Global Development Network Growth Database would drop Norway from this list and add Angola, Republic of the Congo, Gabon, Nigeria, and Turkmenistan. Not surprisingly, use of the latter group of non-Arab fuel exporters typically results in more favorable comparisons for the Arab fuel-endowed economies. In the next section, we begin our comparisons between the three Arab worlds and their non-Arab counterparts with indicators for health, educational attainment, and income, which are the components of the United Nations Development Program’s Human Development Index. In the following sections we extend our comparisons to population growth, gender gaps, and democracy, three areas in which the Arab countries are widely believed to be among the world’s worst performers.1 We then draw conclusions from our findings. Arab Human Development in Comparative Perspective We measure Arab progress in human development by separately examining each component of the Human Development Index (HDI). The HDI was introduced by the United Nations Development Program in 1990. It gives equal weight to life expectancy at birth, educational attainment and effort, and per capita income. Measurement for the latter two components has changed since 1990. Currently, the education component gives two-thirds weight to literacy and onethird weight to school enrollment and the income component uses the logarithm of per capita income measured in international (purchasing power parity or PPP) dollars. 1 The United Nations Development Programme’s Arab Human Development Report 2002 (and later years) also strongly emphasized the “knowledge deficit” of the Arab world as measured by number of patents and related indicators. This has not been mentioned by economists with whom we have discussed previous drafts of this article, presumably because they are not surprised that countries behind the technological frontier would specialize in imitation rather than innovation. 168 Journal of Economic Perspectives Symmetry would suggest that we consider three pairs of country groups: Arab and non-Arab sub-Saharan Africa, Arab and non-Arab fuel-endowed economies, and the remainders of the Arab and non-Arab worlds. However, the remainder of the non-Arab world is simply too diverse to make a satisfactory comparison group for the Arab Mediterranean on its own. We therefore supplement it with two narrower country groups: Latin America (minus its fuel-endowed economy, Venezuela) and southern Europe, in which we include the United Nations Geographical Region of the same name minus the formerly communist countries. These southern European countries, chiefly Greece, Italy, Portugal, and Spain, represent the aspirations of the Arab Mediterranean. Both country groups were part of the Roman Empire. More importantly, all of the Arab Mediterranean countries have concluded Euro-Mediterranean Association Agreements with the European Union (except Syria whose agreement is only initialed). The clear intention of the “Barcelona Process,” of which these agreements are a part, is to extend the developmental pull of the European Union across to the south side of the Mediterranean in the hopes of repeating the earlier developmental success with the countries on the north side. For each indicator and country group, we compute the population-weighted average in 1970 and in the most recent available year, and the change or growth rate between the two years. Countries are included in the statistics only if they have data for both years. We discuss in the text any cases for which use of the median country instead of the population-weighted average qualitatively changes the comparison between an Arab country group and a corresponding non-Arab country group. We begin with life expectancy. Panel A of Table 2 shows that the three Arab worlds all had lower life expectancies than the comparison country groups in 1970. By 2006, life expectancies in Arab sub-Saharan Africa and the Arab fuel-endowed economies were both higher than in the comparison country groups. Life expectancy in the Arab Mediterranean (rest of Arab world) was also higher than in the remainder of the non-Arab world in 2006, but was still lower than in Latin America or southern Europe. Nevertheless, the Arab Mediterranean had closed the life expectancy gap with Latin America and with southern Europe by seven and ten years, respectively. The fact that by 2006 the population-weighted average life expectancy in Arab sub-Saharan Africa exceeded that in non-Arab sub-Saharan Africa by more than eight years raises the question of whether the Arab countries suffered from the same disease environment as the rest of sub-Saharan Africa. Gallup and Sachs (2001) developed a malaria index that has proved to be a powerful explanatory variable for socioeconomic performance. We examined the values of the index in 1965 to obtain a sense of the disease environment in these countries at the start of our time period. Except for Djibouti, every Arab sub-Saharan country including Yemen was in the highest category for malaria, as were all non-Arab sub-Saharan countries except Lesotho, South Africa, and Swaziland. Next we compare progress in education. Unfortunately, literacy data in 1970
The per capita income comparisons of Arab to non-Arab country groups need to be discussed cautiously because they are more sensitive than were the comparisons for life expectancy and average years of education to our use of the population-weighted average rather than the median country and to starting in 1970 rather than 1960. Panel C of Table 2 shows that Arab sub-Saharan Africa has grown much faster than non-Arab sub-Saharan Africa, with the result that by 2007 Arab sub-Saharan per capita income has almost caught up to non-Arab sub-Saharan per capita income. For the median country, however, Arab sub-Saharan per capita income is already higher than non-Arab sub-Saharan per capita income in 1970. The Arab and non-Arab fuel economies have grown equally since 1970, with Arab per capita income remaining slightly higher. This changes if we examine median countries, with the median non-Arab fuel economy more than $1000 (in international dollars) poorer than the median Arab fuel economy in 1970 but more than $3,000 richer in 2007. On the other hand, staying with the population-weighted average but starting in 1960 yields a higher growth rate for the Arab fuel economies, 1.5 versus 0.8 percent. Finally, the Arab Mediterranean countries (rest of Arab world) have grown more slowly than the broad comparison group but faster than either of the two narrow comparison groups. Use of the median country cuts the growth rate of the broad comparison group in half, from 4.0 to 2.0 percent, while leaving the Arab Mediterranean growth rate unchanged at 2.7 percent (and the Latin American and southern European growth rates virtually unchanged as well). This comparison mainly shows the huge influence of China: just dropping it from the population-weighted average for the broad comparison group reduces the growth rate of that country group to 2.6 percent. Starting in 1960 still leaves the Arab Mediterranean growth rate at 2.7 percent but raises the southern European growth rate to 2.9 percent. The results starting in 1960 show that there has been a lack of long-run convergence of Arab Mediterranean per capita income to that of its natural target to the north. This is a signal failure of Arab human development and a cause of frustration (and emigration) for Arabs in the region. What are the prospects for convergence in the coming decades? Let us assume that southern Europe has largely completed its own convergence to northern Europe and that its per capita income will grow at an annual rate of 2 percent for the foreseeable future. To achieve noticeable convergence in a generation, per capita income in the Arab Mediterranean would have to grow at an annual rate of, say, 3 percent. Is this feasible? Of the four Arab Mediterranean countries with populations greater than ten million, per capita incomes in Egypt and Tunisia grew at 3.1 and 3.2 percent, respectively, and per capita incomes in Morocco and Syria grew at 2.1 and 2.2 percent, respectively, during the period 1960 –2007. For Egypt and Tunisia to grow The Three Arab Worlds 173 faster would require their growth rates to reach the (South-) East Asian miracle level: the per capita income growth rates of Indonesia, Malaysia, and Thailand from 1960 to 2007 were 3.6 percent, 3.9 percent, and 4.7 percent, respectively. This may be possible for Tunisia, which has the highest level of manufactured exports as a share of GDP of all the Arab countries (it is roughly equal to that of Israel) and a government that has proven its abilities to implement reforms and adapt to shocks (Nabli and Nugent, 2008). However, such a long-term growth acceleration is doubtful for Egypt, which has none of the typical markers of a high-growth economy. Since Morocco and Syria together have five times the population of Tunisia, their per capita income growth will therefore have to increase to nearly 3 percent. A long-term growth acceleration for Syria will probably require a peace deal with Israel and the United States that will remove U.S. sanctions, leading to economic liberalization and a surge of foreign direct investment. Moroccan per capita income growth has already accelerated, to 3.8 percent annually since 2000. True, this was a boom period for the world economy, but the terms of trade for Morocco actually deteriorated by 14 percent from 2000 to 2006 (2007 data are not yet available). The Moroccan economy has also been undergoing a gradual transformation as indicated by the increase in its manufacturing share of merchandise exports from 50 to 68 percent during the decade 1996 –2006. We have sketched a scenario for long-run convergence across the Mediterranean in which Egypt maintains its historical growth performance; Tunisia fulfills its “star” potential, Syria successfully concludes its on-again, off-again negotiations with Israel; and Morocco retains about half of its recent growth acceleration. It is a plausible scenario in our view, but one would have to be a risk lover to bet on it. About to Run Aground? Population Growth and Water Scarcity The absolute and relative progress in human development achieved by the Arab worlds could be threatened by runaway population growth. Is Arab population growth exceptionally fast? Panel A of Table 3 shows the population growth rates for each country group for the period 1970 –2007. Except in sub-Saharan Africa, Arab population growth has been substantially higher than in the comparison country groups. At annual rates in excess of 3 percent, population growth in the Arab fuel-endowed countries would seem to be particularly alarming. However, these figures do not reflect the steep declines in Arab fertility. In 1970, fertility was uniformly high across the Arab worlds: from 6.5 to 7.3 births per woman, consistently higher than in the comparison country groups as seen in Panel B of Table 3. By 2006, Arab fertility was below three births per woman in the Arab Mediterranean and fuel-endowed country groups, and below five births per woman in Arab sub-Saharan Africa, placing it lower than the comparison country group in the latter case and only 0.6 and 0.7 births higher in the Arab fuel-endowed economy and Arab Mediterranean cases, respectively. As a result, the United
theory suggests that they are choosing to spend their time raising higher-quality children and enjoying leisure rather than working. The question, of course, is whether this outcome is indeed their choice or whether it is being forced upon them, either by workplace discrimination or social pressure. Arab countries have laws prohibiting discrimination against women in the workplace that look more or less like laws in other countries (United Nations Development Fund for Women, 2004, p. 227), so discrimination would have to be established de facto rather than de jure, requiring careful studies that to our knowledge have not been done. For the record, no Arab country except Saudi Arabia forbids women to drive, leave their heads uncovered in public, or vote (neither men nor women can vote in the United Arab Emirates, where all government officials are appointed). Not surprisingly, Saudi Arabia has the lowest rate of female labor force participation in our sample. Our sense is that the presence of women in the workplace is not objectionable in most of Arab society, but also that the expectation that married women should be financially supported by their husbands and should stay home to care for their children remains very strong. Investigating the labor force participation of Arab women is an intriguing topic for future research. One approach to uncovering social pressure might be to examine whether the Arab gender gaps in labor force participation are smaller relative to comparison groups of countries for single adults than for married adults. Another angle is suggested by Lisa Anderson, Provost at the American University of Cairo, who believes that there is a boom in telecommuting by professional women who want to reconcile their careers with their roles as wives and mothers. Again, we hope that future research will reveal how common this practice might be in Arab countries and whether it biases labor force participation statistics. Democracy There are two widely used measures of democracy, the Freedom House Index and the composite Polity index. The former does not begin until 1972, so we only use the latter, which also allows us to check the robustness of our results against a 1960 start date. The composite Polity index, which we will denote by polity, is from the Polity IV dataset and equals Polity’s Democracy index minus Polity’s Autocracy index. Since each of these indices ranges from zero to ten, polity ranges from ten to minus ten. Table 6 shows the population-weighted average values and the median country values of polity for the three Arab worlds and the comparison country groups in 1970 and 2006. We report the median country values because they differ so often and so greatly from the population-weighted average values, and change the qualitative picture completely for the non-Arab fuel economies. Nevertheless, with regard to comparisons between the Arab and non-Arab country groups, the two sets of figures tell a consistent story. Arab countries have become more democratic (or less autocratic) since 1970, but their gains have been small compared to the waves of democratization that have transformed the polities of non-Arab sub-Saharan Africa, Latin America, and southern Europe. The Arab reputation for lack of democracy is well-founded. As with the gender gap in labor force participation, we must ask whether we have identified an instance of Arab exceptionalism. Again, a “horse race” between percentage Muslim and an Arab indicator seems appropriate. Percentage Muslim has been found to be significantly negative in democracy regressions by Barro (1999) and Acemoglu, Johnson, Robinson, and Yared (2008). Columns 1 and 2 of Table 7 repeat the exercise of Table 5 for polity. The results are the opposite of those in Table 5: the Arab indicator is statistically insignificant when percentage Muslim is present. This finding is robust to the addition of variables normally included in democracy regressions in the literature, including per capita GDP, indicators of former colonial status (such as British colony, French colony, and the like), and an indicator for fuel-oriented economies. The results in Table 7 notwithstanding, in 2007 there existed six Muslim democracies according to Polity IV: Albania, Comoros, Indonesia, Lebanon, Senegal, and Turkey. All made the transition (from polity less than 7 to polity 7 or more)7 in the preceding ten years except Turkey, which has been coded as a democracy since 1983. Can any Arab country join Comoros and Lebanon by making the same
transition in the coming ten years? Some scholars have argued that this can be achieved through the accession to power of “moderate Islamist” political parties (for example, Hamzawy, 2005). The idea is that this formula will reconcile democratic processes with the Arab cultural heritage of exercise of political authority in
the name of Islam. This argument has been bolstered by the apparent success of the Justice and Development Party (AKP) in Turkey since 2003, though it came to power within an already functioning democracy. One could say that, of the Muslim democracies, Turkey is the most appropriate model for the Arab world.8 Through the Ottoman Empire, Turks ruled for four centuries over most of the territory outside of sub-Saharan Africa that became today’s Arab countries. At time of writing, moderate Sunni Islamist parties exist in at least five Arab countries but are legal in only two, Jordan and Morocco. A major obstacle to legalization of the other parties is their past association (or alleged association) with political violence. For example, the most important illegal moderate Sunni Islamist party in the Arab world is the Egyptian Muslim Brotherhood. The first of three incidents (as of 2005) of Islamist political violence in the Arab world large enough to qualify as an armed conflict—to make the threshold of 25 “battle deaths” per year set by the Armed Conflict Dataset—was the uprising led in 1982 by a branch of the Muslim Brotherhood centered in Hama, Syria, in which 5,000 to 25,000 people were killed.9 The Brotherhood does not appear to have been 8 Turkey is a more appropriate comparison than Lebanon, which is sui generis in the Arab world by virtue of having no majority religious group if the Shi’ite and Sunni branches of Islam are counted as different religions. It is possible that the Lebanese experience is of relevance to Iraq (and vice versa!), but Iraq has not been assigned a polity score since the U.S. invasion and falls outside the scope of our analysis. As for Comoros, it seems too small and remote to serve as a model. (Omitting Comoros from the regressions in Table 7 has no qualitative impact.) 9 From the dataset codebook: “An armed conflict is a contested incompatibility that concerns government and/or territory where the use of armed force between two parties, of which at least one is the government of a state, results in at least 25 battle-related deaths. The... dataset defines battle deaths as deaths resulting directly from violence inflicted through the use of armed force by a party to an armed Notes: The “Muslim” variable is defined as a percent. The Arab variable takes the value 1 if a country is considered Arab and 0 otherwise. Notes: Standard errors are in parentheses. ***, **, and * indicate significance at the 1, 5, and 10 percent levels, respectively. 186 Journal of Economic Perspectives connected with either of the other two incidents, though former Brothers may have been involved: the struggle between the Egyptian government and the Islamic Group and its allies, which lasted from 1992 to 1999 and killed over 1300 people; and the civil war that began in 1992 when the Algerian military canceled elections that looked likely to bring (nonmoderate) Islamist parties to power, which has killed over 91,000 people to date. Whether these violent incidents provoke legitimate fear of moderate Islamist parties by Arab governments or just provide them a convenient excuse to repress a popular opposition is an open question. In either case, we do not foresee democracy coming to any Arab country via moderate Islamist parties in the next five years, and perhaps not even in the next ten. Conclusions In 1970, an observer of the Arab world noting its nearly universal backwardness in socioeconomic indicators might have concluded that there was something fundamentally wrong with Arab “institutions.” Nearly four decades later, we believe it is clear from the evidence presented in this article that this conclusion would have been a mistake. Arab sub-Saharan Africa is surging ahead or rapidly catching up to the rest of sub-Saharan Africa for every indicator we examined except average years of education, despite (according to the most widely used measure) being handicapped by the same tropical disease environment. Arab fuel-endowed economies and the Mediterranean remainder of the Arab world have converged or are rapidly converging to the averages for their non-Arab counterparts, except for income, where they are keeping pace. One important exception to this summary picture is that across the three Arab worlds the gender gaps in labor force participation remain exceptionally large, despite the fact that efforts to raise educational attainment nearly eliminated Arab gender gaps in school enrollment. All three Arab worlds also remain politically backward despite their socioeconomic progress. In this respect the Arab countries mirror the larger Muslim world, whereas with regard to female labor force participation something specific to Arab culture appears to be present. Our analysis begs the question of why the Arab countries were so far behind in 1970. One possible answer is the Ottoman Empire and British, French, and Italian colonialism. Yet by 1970, many Arab countries had experienced one to two decades or more of political independence. Perhaps recognition of the rapid progress that conflict during contested combat.... Contested combat excludes the sustained destruction of soldiers or civilians outside of the context of any reciprocal threat of lethal force . . .” The current conflict in Somalia, which involves Islamist groups, occurs after the most recent year for which data are available. All battle death estimates in the text are from the Armed Conflict Dataset. Thw version-20. James E. Rauch and Scott Kostyshak 187 has occurred since 1970 will spur greater interest in identifying what changed to make that progress possible.
evaluation period as possible, since it would not make sense to classify a country as fuel-endowed for the purpose of evaluating its socioeconomic performance beginning in 1970 if its reserves were only discovered in 2000. The authoritative data Figure 1 The Three Arab Worlds Table 1 Population of Arab States, 2007 (in thousands; total 324 million) Sub-Saharan Africa Fuel-endowed countries Remainder (“Arab Mediterranean”) Comoros 626 Algeria 33,853 Egypt 75,467 Djibouti 833 Bahrain 753 Jordan 5,719 Mauritania 3,121 Iraqa 28,221 Lebanon 4,097 Somalia 8,696 Kuwait 2,663 Morocco 30,861 Sudan 38,556 Libya 6,156 Syrian Arab Republic 19,891 Yemen 22,383 Oman 2,600 Tunisia 10,248 Qatar 836 Saudi Arabia 24,196 United Arab Emirates 4,365 Total 74,215 Total 103,643 Total 146,283 (23%) (32%) (45%) Source: World Development Indicators Online, 2008, http://go.worldbank.org/6HAYAHG8H0. a The source for the Iraq population is CIA World Factbook (July 2008 est.). James E. Rauch and Scott Kostyshak 167 source is the annual BP Statistical Review of World Energy, which publishes crosscountry comparable estimates of oil and natural gas reserves dating back to 1980. A minor issue arises regarding how to aggregate oil and natural gas reserves. We converted the oil and natural gas into BTUs (British thermal units) using standard conversion factors and computed 1980 BTU reserves per capita. This procedure yielded a ranking with a clean break at Algeria, which has 69 percent more reserves per capita than the next country down and nearly five times more reserves than the next Arab country down. The Arab countries identified as fuel endowed by this ranking are the same as those identified using the Global Development Network Growth Database criterion. Under this criterion, a country is fuel endowed if fuels accounted for more than 50 percent of total exports of goods and services for the period 1988 –1992. The non-Arab countries identified as fuel endowed by this criterion are Brunei, Iran, Norway, Trinidad and Tobago, and Venezuela. For comparison, using the export-based criterion from the Global Development Network Growth Database would drop Norway from this list and add Angola, Republic of the Congo, Gabon, Nigeria, and Turkmenistan. Not surprisingly, use of the latter group of non-Arab fuel exporters typically results in more favorable comparisons for the Arab fuel-endowed economies. In the next section, we begin our comparisons between the three Arab worlds and their non-Arab counterparts with indicators for health, educational attainment, and income, which are the components of the United Nations Development Program’s Human Development Index. In the following sections we extend our comparisons to population growth, gender gaps, and democracy, three areas in which the Arab countries are widely believed to be among the world’s worst performers.1 We then draw conclusions from our findings. Arab Human Development in Comparative Perspective We measure Arab progress in human development by separately examining each component of the Human Development Index (HDI). The HDI was introduced by the United Nations Development Program in 1990. It gives equal weight to life expectancy at birth, educational attainment and effort, and per capita income. Measurement for the latter two components has changed since 1990. Currently, the education component gives two-thirds weight to literacy and onethird weight to school enrollment and the income component uses the logarithm of per capita income measured in international (purchasing power parity or PPP) dollars. 1 The United Nations Development Programme’s Arab Human Development Report 2002 (and later years) also strongly emphasized the “knowledge deficit” of the Arab world as measured by number of patents and related indicators. This has not been mentioned by economists with whom we have discussed previous drafts of this article, presumably because they are not surprised that countries behind the technological frontier would specialize in imitation rather than innovation. 168 Journal of Economic Perspectives Symmetry would suggest that we consider three pairs of country groups: Arab and non-Arab sub-Saharan Africa, Arab and non-Arab fuel-endowed economies, and the remainders of the Arab and non-Arab worlds. However, the remainder of the non-Arab world is simply too diverse to make a satisfactory comparison group for the Arab Mediterranean on its own. We therefore supplement it with two narrower country groups: Latin America (minus its fuel-endowed economy, Venezuela) and southern Europe, in which we include the United Nations Geographical Region of the same name minus the formerly communist countries. These southern European countries, chiefly Greece, Italy, Portugal, and Spain, represent the aspirations of the Arab Mediterranean. Both country groups were part of the Roman Empire. More importantly, all of the Arab Mediterranean countries have concluded Euro-Mediterranean Association Agreements with the European Union (except Syria whose agreement is only initialed). The clear intention of the “Barcelona Process,” of which these agreements are a part, is to extend the developmental pull of the European Union across to the south side of the Mediterranean in the hopes of repeating the earlier developmental success with the countries on the north side. For each indicator and country group, we compute the population-weighted average in 1970 and in the most recent available year, and the change or growth rate between the two years. Countries are included in the statistics only if they have data for both years. We discuss in the text any cases for which use of the median country instead of the population-weighted average qualitatively changes the comparison between an Arab country group and a corresponding non-Arab country group. We begin with life expectancy. Panel A of Table 2 shows that the three Arab worlds all had lower life expectancies than the comparison country groups in 1970. By 2006, life expectancies in Arab sub-Saharan Africa and the Arab fuel-endowed economies were both higher than in the comparison country groups. Life expectancy in the Arab Mediterranean (rest of Arab world) was also higher than in the remainder of the non-Arab world in 2006, but was still lower than in Latin America or southern Europe. Nevertheless, the Arab Mediterranean had closed the life expectancy gap with Latin America and with southern Europe by seven and ten years, respectively. The fact that by 2006 the population-weighted average life expectancy in Arab sub-Saharan Africa exceeded that in non-Arab sub-Saharan Africa by more than eight years raises the question of whether the Arab countries suffered from the same disease environment as the rest of sub-Saharan Africa. Gallup and Sachs (2001) developed a malaria index that has proved to be a powerful explanatory variable for socioeconomic performance. We examined the values of the index in 1965 to obtain a sense of the disease environment in these countries at the start of our time period. Except for Djibouti, every Arab sub-Saharan country including Yemen was in the highest category for malaria, as were all non-Arab sub-Saharan countries except Lesotho, South Africa, and Swaziland. Next we compare progress in education. Unfortunately, literacy data in 1970
The per capita income comparisons of Arab to non-Arab country groups need to be discussed cautiously because they are more sensitive than were the comparisons for life expectancy and average years of education to our use of the population-weighted average rather than the median country and to starting in 1970 rather than 1960. Panel C of Table 2 shows that Arab sub-Saharan Africa has grown much faster than non-Arab sub-Saharan Africa, with the result that by 2007 Arab sub-Saharan per capita income has almost caught up to non-Arab sub-Saharan per capita income. For the median country, however, Arab sub-Saharan per capita income is already higher than non-Arab sub-Saharan per capita income in 1970. The Arab and non-Arab fuel economies have grown equally since 1970, with Arab per capita income remaining slightly higher. This changes if we examine median countries, with the median non-Arab fuel economy more than $1000 (in international dollars) poorer than the median Arab fuel economy in 1970 but more than $3,000 richer in 2007. On the other hand, staying with the population-weighted average but starting in 1960 yields a higher growth rate for the Arab fuel economies, 1.5 versus 0.8 percent. Finally, the Arab Mediterranean countries (rest of Arab world) have grown more slowly than the broad comparison group but faster than either of the two narrow comparison groups. Use of the median country cuts the growth rate of the broad comparison group in half, from 4.0 to 2.0 percent, while leaving the Arab Mediterranean growth rate unchanged at 2.7 percent (and the Latin American and southern European growth rates virtually unchanged as well). This comparison mainly shows the huge influence of China: just dropping it from the population-weighted average for the broad comparison group reduces the growth rate of that country group to 2.6 percent. Starting in 1960 still leaves the Arab Mediterranean growth rate at 2.7 percent but raises the southern European growth rate to 2.9 percent. The results starting in 1960 show that there has been a lack of long-run convergence of Arab Mediterranean per capita income to that of its natural target to the north. This is a signal failure of Arab human development and a cause of frustration (and emigration) for Arabs in the region. What are the prospects for convergence in the coming decades? Let us assume that southern Europe has largely completed its own convergence to northern Europe and that its per capita income will grow at an annual rate of 2 percent for the foreseeable future. To achieve noticeable convergence in a generation, per capita income in the Arab Mediterranean would have to grow at an annual rate of, say, 3 percent. Is this feasible? Of the four Arab Mediterranean countries with populations greater than ten million, per capita incomes in Egypt and Tunisia grew at 3.1 and 3.2 percent, respectively, and per capita incomes in Morocco and Syria grew at 2.1 and 2.2 percent, respectively, during the period 1960 –2007. For Egypt and Tunisia to grow The Three Arab Worlds 173 faster would require their growth rates to reach the (South-) East Asian miracle level: the per capita income growth rates of Indonesia, Malaysia, and Thailand from 1960 to 2007 were 3.6 percent, 3.9 percent, and 4.7 percent, respectively. This may be possible for Tunisia, which has the highest level of manufactured exports as a share of GDP of all the Arab countries (it is roughly equal to that of Israel) and a government that has proven its abilities to implement reforms and adapt to shocks (Nabli and Nugent, 2008). However, such a long-term growth acceleration is doubtful for Egypt, which has none of the typical markers of a high-growth economy. Since Morocco and Syria together have five times the population of Tunisia, their per capita income growth will therefore have to increase to nearly 3 percent. A long-term growth acceleration for Syria will probably require a peace deal with Israel and the United States that will remove U.S. sanctions, leading to economic liberalization and a surge of foreign direct investment. Moroccan per capita income growth has already accelerated, to 3.8 percent annually since 2000. True, this was a boom period for the world economy, but the terms of trade for Morocco actually deteriorated by 14 percent from 2000 to 2006 (2007 data are not yet available). The Moroccan economy has also been undergoing a gradual transformation as indicated by the increase in its manufacturing share of merchandise exports from 50 to 68 percent during the decade 1996 –2006. We have sketched a scenario for long-run convergence across the Mediterranean in which Egypt maintains its historical growth performance; Tunisia fulfills its “star” potential, Syria successfully concludes its on-again, off-again negotiations with Israel; and Morocco retains about half of its recent growth acceleration. It is a plausible scenario in our view, but one would have to be a risk lover to bet on it. About to Run Aground? Population Growth and Water Scarcity The absolute and relative progress in human development achieved by the Arab worlds could be threatened by runaway population growth. Is Arab population growth exceptionally fast? Panel A of Table 3 shows the population growth rates for each country group for the period 1970 –2007. Except in sub-Saharan Africa, Arab population growth has been substantially higher than in the comparison country groups. At annual rates in excess of 3 percent, population growth in the Arab fuel-endowed countries would seem to be particularly alarming. However, these figures do not reflect the steep declines in Arab fertility. In 1970, fertility was uniformly high across the Arab worlds: from 6.5 to 7.3 births per woman, consistently higher than in the comparison country groups as seen in Panel B of Table 3. By 2006, Arab fertility was below three births per woman in the Arab Mediterranean and fuel-endowed country groups, and below five births per woman in Arab sub-Saharan Africa, placing it lower than the comparison country group in the latter case and only 0.6 and 0.7 births higher in the Arab fuel-endowed economy and Arab Mediterranean cases, respectively. As a result, the United
theory suggests that they are choosing to spend their time raising higher-quality children and enjoying leisure rather than working. The question, of course, is whether this outcome is indeed their choice or whether it is being forced upon them, either by workplace discrimination or social pressure. Arab countries have laws prohibiting discrimination against women in the workplace that look more or less like laws in other countries (United Nations Development Fund for Women, 2004, p. 227), so discrimination would have to be established de facto rather than de jure, requiring careful studies that to our knowledge have not been done. For the record, no Arab country except Saudi Arabia forbids women to drive, leave their heads uncovered in public, or vote (neither men nor women can vote in the United Arab Emirates, where all government officials are appointed). Not surprisingly, Saudi Arabia has the lowest rate of female labor force participation in our sample. Our sense is that the presence of women in the workplace is not objectionable in most of Arab society, but also that the expectation that married women should be financially supported by their husbands and should stay home to care for their children remains very strong. Investigating the labor force participation of Arab women is an intriguing topic for future research. One approach to uncovering social pressure might be to examine whether the Arab gender gaps in labor force participation are smaller relative to comparison groups of countries for single adults than for married adults. Another angle is suggested by Lisa Anderson, Provost at the American University of Cairo, who believes that there is a boom in telecommuting by professional women who want to reconcile their careers with their roles as wives and mothers. Again, we hope that future research will reveal how common this practice might be in Arab countries and whether it biases labor force participation statistics. Democracy There are two widely used measures of democracy, the Freedom House Index and the composite Polity index. The former does not begin until 1972, so we only use the latter, which also allows us to check the robustness of our results against a 1960 start date. The composite Polity index, which we will denote by polity, is from the Polity IV dataset and equals Polity’s Democracy index minus Polity’s Autocracy index. Since each of these indices ranges from zero to ten, polity ranges from ten to minus ten. Table 6 shows the population-weighted average values and the median country values of polity for the three Arab worlds and the comparison country groups in 1970 and 2006. We report the median country values because they differ so often and so greatly from the population-weighted average values, and change the qualitative picture completely for the non-Arab fuel economies. Nevertheless, with regard to comparisons between the Arab and non-Arab country groups, the two sets of figures tell a consistent story. Arab countries have become more democratic (or less autocratic) since 1970, but their gains have been small compared to the waves of democratization that have transformed the polities of non-Arab sub-Saharan Africa, Latin America, and southern Europe. The Arab reputation for lack of democracy is well-founded. As with the gender gap in labor force participation, we must ask whether we have identified an instance of Arab exceptionalism. Again, a “horse race” between percentage Muslim and an Arab indicator seems appropriate. Percentage Muslim has been found to be significantly negative in democracy regressions by Barro (1999) and Acemoglu, Johnson, Robinson, and Yared (2008). Columns 1 and 2 of Table 7 repeat the exercise of Table 5 for polity. The results are the opposite of those in Table 5: the Arab indicator is statistically insignificant when percentage Muslim is present. This finding is robust to the addition of variables normally included in democracy regressions in the literature, including per capita GDP, indicators of former colonial status (such as British colony, French colony, and the like), and an indicator for fuel-oriented economies. The results in Table 7 notwithstanding, in 2007 there existed six Muslim democracies according to Polity IV: Albania, Comoros, Indonesia, Lebanon, Senegal, and Turkey. All made the transition (from polity less than 7 to polity 7 or more)7 in the preceding ten years except Turkey, which has been coded as a democracy since 1983. Can any Arab country join Comoros and Lebanon by making the same
transition in the coming ten years? Some scholars have argued that this can be achieved through the accession to power of “moderate Islamist” political parties (for example, Hamzawy, 2005). The idea is that this formula will reconcile democratic processes with the Arab cultural heritage of exercise of political authority in
the name of Islam. This argument has been bolstered by the apparent success of the Justice and Development Party (AKP) in Turkey since 2003, though it came to power within an already functioning democracy. One could say that, of the Muslim democracies, Turkey is the most appropriate model for the Arab world.8 Through the Ottoman Empire, Turks ruled for four centuries over most of the territory outside of sub-Saharan Africa that became today’s Arab countries. At time of writing, moderate Sunni Islamist parties exist in at least five Arab countries but are legal in only two, Jordan and Morocco. A major obstacle to legalization of the other parties is their past association (or alleged association) with political violence. For example, the most important illegal moderate Sunni Islamist party in the Arab world is the Egyptian Muslim Brotherhood. The first of three incidents (as of 2005) of Islamist political violence in the Arab world large enough to qualify as an armed conflict—to make the threshold of 25 “battle deaths” per year set by the Armed Conflict Dataset—was the uprising led in 1982 by a branch of the Muslim Brotherhood centered in Hama, Syria, in which 5,000 to 25,000 people were killed.9 The Brotherhood does not appear to have been 8 Turkey is a more appropriate comparison than Lebanon, which is sui generis in the Arab world by virtue of having no majority religious group if the Shi’ite and Sunni branches of Islam are counted as different religions. It is possible that the Lebanese experience is of relevance to Iraq (and vice versa!), but Iraq has not been assigned a polity score since the U.S. invasion and falls outside the scope of our analysis. As for Comoros, it seems too small and remote to serve as a model. (Omitting Comoros from the regressions in Table 7 has no qualitative impact.) 9 From the dataset codebook: “An armed conflict is a contested incompatibility that concerns government and/or territory where the use of armed force between two parties, of which at least one is the government of a state, results in at least 25 battle-related deaths. The... dataset defines battle deaths as deaths resulting directly from violence inflicted through the use of armed force by a party to an armed Notes: The “Muslim” variable is defined as a percent. The Arab variable takes the value 1 if a country is considered Arab and 0 otherwise. Notes: Standard errors are in parentheses. ***, **, and * indicate significance at the 1, 5, and 10 percent levels, respectively. 186 Journal of Economic Perspectives connected with either of the other two incidents, though former Brothers may have been involved: the struggle between the Egyptian government and the Islamic Group and its allies, which lasted from 1992 to 1999 and killed over 1300 people; and the civil war that began in 1992 when the Algerian military canceled elections that looked likely to bring (nonmoderate) Islamist parties to power, which has killed over 91,000 people to date. Whether these violent incidents provoke legitimate fear of moderate Islamist parties by Arab governments or just provide them a convenient excuse to repress a popular opposition is an open question. In either case, we do not foresee democracy coming to any Arab country via moderate Islamist parties in the next five years, and perhaps not even in the next ten. Conclusions In 1970, an observer of the Arab world noting its nearly universal backwardness in socioeconomic indicators might have concluded that there was something fundamentally wrong with Arab “institutions.” Nearly four decades later, we believe it is clear from the evidence presented in this article that this conclusion would have been a mistake. Arab sub-Saharan Africa is surging ahead or rapidly catching up to the rest of sub-Saharan Africa for every indicator we examined except average years of education, despite (according to the most widely used measure) being handicapped by the same tropical disease environment. Arab fuel-endowed economies and the Mediterranean remainder of the Arab world have converged or are rapidly converging to the averages for their non-Arab counterparts, except for income, where they are keeping pace. One important exception to this summary picture is that across the three Arab worlds the gender gaps in labor force participation remain exceptionally large, despite the fact that efforts to raise educational attainment nearly eliminated Arab gender gaps in school enrollment. All three Arab worlds also remain politically backward despite their socioeconomic progress. In this respect the Arab countries mirror the larger Muslim world, whereas with regard to female labor force participation something specific to Arab culture appears to be present. Our analysis begs the question of why the Arab countries were so far behind in 1970. One possible answer is the Ottoman Empire and British, French, and Italian colonialism. Yet by 1970, many Arab countries had experienced one to two decades or more of political independence. Perhaps recognition of the rapid progress that conflict during contested combat.... Contested combat excludes the sustained destruction of soldiers or civilians outside of the context of any reciprocal threat of lethal force . . .” The current conflict in Somalia, which involves Islamist groups, occurs after the most recent year for which data are available. All battle death estimates in the text are from the Armed Conflict Dataset. Thw version-20. James E. Rauch and Scott Kostyshak 187 has occurred since 1970 will spur greater interest in identifying what changed to make that progress possible.
Comments
Post a Comment