Social Mobility through Migration to the Colonies: The Case of Algeria

Many Europeans left their homeland during the age of mass migration to try their luck in the Americas or in the colonies. Most of them intended to improve their economic and social position, especially those who recently migrated to the United States and Europe. But how successful were migrants, and which enjoyed the greatest success? They often entered the labor market in the country of destination at relatively low-status positions. During their lifetime, many indeed improved their social status, paving the way for the generations that followed them to do so as well.

European migrants to the colonies differed from recent migrants. They entered the labor market not at the bottom but at the higher echelons of society; the lowest positions tended to be filled by the native population or, although not the case in Algeria, by unfree laborers from other countries. Since these colonial migrants may also have originated from higher classes than the average recent migrant, we cannot be sure whether migration enabled them to perform better than their parents did. Since the descendants of colonial migrants, unlike the descendants of many recent migrants, did not start at the bottom of society either, they were also less likely to have fared better than their parents; they may not even have been able to retain their parents’ relatively privileged positions. Moreover, whereas recent migrants often do better in their country of destination than they would have done had they not migrated, the same is not necessarily true for European migrants to the colonies, who generally were not moving from a less-developed to a more-developed country.

This study draws inspiration from recent work in the sociology of migration by Güveli et al. who compared the status of Turkish men who migrated to Europe, as well as that of their children and grandchildren, with that of Turkish men who did not migrate. Their research design resembles ours, though the context is different. Güveli et al. study low-skilled Turkish people who relocate to countries where the indigenous inhabitants speak a different language and have a different dominant religion. Our study deals with European migrants, predominantly French, moving to a French colony. For those of French origin, the official language and dominant religion in the colonial society were the same as those in their motherland. We address central questions regarding both present-day and historical migration: Were colonial migrants more successful than non-migrants, both citizens resident in the colony of destination (Algeria) and those who stayed behind (in France)?¹

We answer our research questions using data from French civil marriage registers for both mainland France and the Algerian towns of Algiers, Bône, Constantine, and Oran, comparing the periods 1870–1872 and 1910–1912, when Algeria was officially neither a colony nor a “protectorate” but three départements of France. Algeria at the time, however, was a de facto colony, as indicated by both the French conquest and the failed Algerian war of independence and as signaled by the exclusion of the Muslim population from the registers. Citizens of European descent in Algeria’s départements lived under the same vital registration system as did inhabitants of mainland France. This registration system affords us the opportunity to consider anew the question of whether colonial migration pays dividends. The data permit a comparison of those who stayed in France with those who migrated to Algeria from France and other countries, as well as a further comparison of those migrants with Christians and Jews born on Algerian soil. We look only at men, since the majority of women in our data had no registered occupation from which we could derive a historical measure of social status (see below).²

To derive testable hypotheses from the data about both the sending French population and the out-migrants, we survey two distinct strands of scholarship, that of sociologists of migration and that of economic historians. We compare the average occupational status of men born in France and other European countries who immigrated to Algeria with the average occupational status of two population groups—(1) French citizens born in Algeria (mostly second-generation French immigrants and non-Muslim Algerians who had been granted French citizenship) and (2) French citizens born in mainland France who remained there.

The approach herein—including the categorization of migrants, the theories, and hypotheses derived from the economic and sociological literature about migration, and the research design—contributes to methodological and substantive issues in literatures that are not yet well connected. These general theories and methods can contribute to a better understanding of historical migration patterns. Specific historical case studies, like this one about French settler history in Algeria, can also illuminate or refine more general theories and the historiography of colonial societies. Our article seeks to advance dialogue between historical and contemporary studies about intergenerational social mobility, migration, and the formation and flux of colonial societies. It also injects an interesting Francophone case into streams of literature dominated by the histories of English-speaking settlers.³

The major Algerian cities studied herein are situated in Algeria’s littoral zone, a fertile strip of land—not unlike that of the Mediterranean shores of France, Spain, and Italy—stretching from Tunisia in the east to Morocco in the west. This region had been a powerhouse of grain cultivation since antiquity. Most of the population was comprised of Arabs and Berbers (Arabs or indigènes in French), with a small native Jewish community and a small and sundry mix of Europeans of various Mediterranean nationalities. The French remained a minority among the Europeans until the late nineteenth century, and the Europeans were much in the minority among the population.⁴

In 1830, the French attacked and captured Algiers, expelling its ruler (the dey) and thus ending the Ottoman era. At first, the French tried to populate the country by convict labor, but virtually all the French convicts left Algeria as soon as their prison term had ended. Colonization started with the influx of migrants from Paris and other French cities, and of poor migrants from nearby European states, notably overpopulated Malta. According to Sessions, colonial promotion booklets published by settlement agencies stated, “Without claiming that Algiers is the promised land, we can guarantee that with a little [financial] advance, intelligence, and perseverance, any hard-working man will be able to create an easy life for himself in the colony.” Many beautifully illustrated publications imprinted this rosy image on the inner eye of potential French migrants.⁵

Companies emerged in France to recruit colonial emigrants. In the mid-1830s, the French government sponsored the free passage of migrants from certain provinces to Algeria. Although local French governments initially preferred to send paupers, they soon began accepting only desirable emigrants on board subsidized ships. Sessions writes that as early as 1834, “An administrative network… stretched from new colonial settlements through Algiers and Paris into the smallest villages of rural France.” Algerian policymakers preferred wealthy settlers, but the Parisian authorities reminded them that well-off men were unlikely to migrate. Many relatively poor migrants ended up in poverty in Algeria, too, and had to be repatriated.⁶

By the 1840s, the migration system was working smoothly; the permis d’embarquement issued in Paris were carefully tailored to the quality and quantity of demand in Algeria. Of the c. 1,200 adult, able-bodied men without families (the target group) with permits in 1840, most went to Algiers (c. 700), Philippeville (200), Oran (100), or Bône (100). Almost all of them (c. 1,000) were manual laborers, once called manoeuvres. These navvies were in great demand for building and road construction. Despite the influx of migrants, large landowners still desperately needed labor. The land made available to them had belonged to Turkish beylik or Islamic charities, or they were expropriated common lands thought to be waste. Re-assessed individual property rights usually devolved to Europeans rather than native Algerians.

The registers kept by the Directorate of Algerian Affairs disclose the provenance of 20,000 persons granted free passage between 1841 and 1845, most of whom were urban workers, especially from Paris, and unmarried men from the Rhineland. The German stream dwindled in size before growing significantly after Germany seized the Rhineland in 1870. Many men who voluntarily relocated from Paris to Algeria in 1848 lived in poor parts of the city and made their living as woodworkers or clothiers. Katan suggests that they left to escape poverty, lured by colonial propaganda.⁷

The 1848 French Constitution declared Algeria to be an integral part of France, under French law and French administration, with arrondissements and municipalities (communes). Most laborers working on the land and in the mines, docks, and factories were economically and politically marginalized indigenous Arabs and Berbers. After a failed rebellion, Algeria became, in the words of Ageron, “‘a small French Republic’ in which only the interests of the French settlers counted. The voter’s card became the title of nobility in this feudal system.” The growing number of Jeunes Algériens, Muslim natives in Algeria who had received French-style education, pushed for citizenship after 1870 but to no avail.⁸

Many Europeans, preferably Frenchmen, were needed to populate Algeria—specifically, to administer the colony; to serve in the army; to introduce and supervise new agricultural activities; and to build and man the post offices, railways, and other service sectors. Kateb notes that this demand was fueled by mortality rates that were higher than those in mainland France. The colonial government therefore offered free land in rural areas on condition of residence. The first to receive this offer were Frenchmen driven from Alsace-Lorraine after its German annexation during the Franco-Prussian war of 1870/1, but few of them were interested. The subsequent colonial policy targeting peasants from the southeast of France, as well as the French living in Algerian cities, however, was successful; within ten years, the government had allotted 347,000 hectares to them and created 197 settler villages with 30,000 inhabitants.⁹

We selected the four Algerian cities in our study—Bône, Constantine, Algiers (Alger), and Oran (Figure 1)—to cover a broad stretch of the littoral zone and thus a diversity of colonial life, including settlers from diverse European countries. Our period begins in 1870, the start of French colonial society’s heyday, with ethnic groups that were still highly distinct. Apart from the Berbers and the Arabs, who were not included in the civil-registration records, the population included Jews (close in status to the Arabs) and Frenchmen (the top of the colonial social hierarchy), followed by the Spanish, the Italians, the Maltese, and a host of other smaller groups of European descent. This situation, sketched by Prochaska, was not unlike that in Morocco, as related by Abu-Lughod. By the early twentieth century, a fully formed settler culture encompassing all Europeans, the so-called pieds-noirs, had emerged. According to Prochaska, with reference to Baroli’s work, “The formation of a colonial society recognizably Algerian occurred between 1890 and 1914. It included all the Europeans—French, Italian, Spanish, Maltese—and excluded Jews as well as Muslim Algerians. Contemporary observers perceived a melting-pot effect, a fusion of the European ethnic groups which resulted in a new race.” At this time, the settlers began to refer to themselves as Algériens (as opposed to indigènes) and to express themselves in an Algerian dialect, pataouète, with words borrowed from Arabic, Spanish, Italian, and the patois of southern France.¹⁰

Except in cases of war and pogrom, people generally migrate only to marry or to ameliorate their economic situation. Since migrating involves substantial financial and social costs, the gains also need to be substantial, more so than when just moving to the next town. Hence, the first question to address in this study is, “Do migrants attain a better status than comparable residents who remain in their country of origin or who move within that country?”¹¹

Many of the sociological studies that compare the status of newly arrived migrants with that of residents in the country of destination often provide evidence of an “ethnic penalty” for migration—having to work in a low-status occupation or in an ethnic niche of the labor market, at least upon arrival and for a period afterward. One explanation for the low status of migrants is their disadvantage in language proficiency, education, the cultural codes of the new country, social contacts, and the rich knowledge needed to obtain higher-level employment. Even migrants with a relatively good education in their country of origin may fare worse than residents with fewer years of schooling; the skills learned at school align most closely to the society of the teachers. Nor are school diplomas from the sending country always recognized in the country of destination. Something similar might apply to the skills required to run a farm, which depend on the type of crop cultivated and on soil and climate conditions. Furthermore, migrants generally have less material support from their parents; they certainly cannot inherit a farm or a shop as a means of income. Migrant children might even have to send support to family members at home rather than to receive any themselves.¹²

Migrant groups might be excluded from occupations that residents with no better level of social and human capital generally fill as a matter of course. In most societies, residents tend to prefer their friends, family members, and acquaintances when faced with a choice. Employers often react to the threat of a migrant taking work away from residents even when that threat is ill-founded. Discrimination thus becomes institutionalized in a segmented labor market, in which residents occupy and control entry into stable occupations of high status and high income, leaving only less-desirable work outside this privileged core to less-favored ethnicities. Although the realities of human capital and social capital, as well as discrimination, lead toward a general proposition that migrants have lower social status than residents, the outcomes can vary for different types of migrant, as is especially relevant for this study.¹³

Discrimination theory maintains that certain groups of migrants face more prejudice than others, depending, for example, on the cultural distance between natives and migrants, the size of the migrant group, or the threat level of natives in the labor market. Consideration is hardly ever given to migrants undergoing positive discrimination, but such could have been the case for French migrants to the colony. Because they originated from the same country as those with power in Algeria, French migrants (if not migrants from other European countries) could well have entered Algeria’s labor market with relatively good prospects.

Human- and social-capital theory predict a rosier situation for second-generation migrants than for first-generation migrants. Although first-generation migrants might lose social capital as they abandon many of their old friends, family, and acquaintances in their country of origin, assimilation theory views them as eventually improving their situation. Thus, the second generation of migrants can expect to reap the rewards of this new social capital and often to have better language proficiency than their parents, especially if they enjoyed the benefit of schooling in the destination country. If the gap between migrants and natives does not narrow in the first generation, it certainly stands to do so with each successive generation—all the more so in colonial Algeria where patronage and clientelism was rife.¹⁴

Because our data do not allow us to distinguish the second generation, we cannot test this notion of improving circumstances directly. We can, however, make a distinction between those who migrated as adults, the first generation of migrants, and those who migrated as children, often termed the 1.5 generation. Human- and social-capital theory predicts the social status of the 1.5 generation to be higher than that of the first generation. Alexander and Ward’s test of this notion on immigrants from Europe to the United States, with data about arrivals linked to the 1940 U.S. census, finds that occupational outcomes depended much on age at arrival—the earlier, the better. The children in the 1.5 generation fared worse than natives with regard to wage and education, especially those from low-income countries with poor educational systems.¹⁵

This argument relies on the assumption of a language gap between the country of origin and the country of destination that has to be bridged. Although this gap is evident in many contemporary migrations, and many historical ones, the motherland and the colonial society in migration studies usually share the same official language. Most of Algeria’s inhabitants had Arabic or Tamazight (Berber) as their first language, but a considerable number of them spoke some French, since French was the official administrative language and thus an asset. Again, we might expect a more propitious situation for migrants from the motherland to the colony than for migrants from countries speaking another mother tongue. Hence, in our case, being born in France was an advantage, whereas being born elsewhere, notably in Italy, and speaking another language was a disadvantage. This difference would have narrowed in the second generation.

In addition to the mostly sociological literature surveyed above, a few important studies in economics and socioeconomic history trace immigrants at the individual level and compare them to stayers and to the host population during the Age of Mass Migration. This literature stresses that the correlation between migration status and occupational status can be the outcome of either a causal mechanism (returns on migration) or a selection effect. Regarding the former, migration may benefit (or harm) migrants by impacting their occupational status for the reasons discussed above. Migrants, however, may also be “negatively” or “positively” selected from the sending population. The brightest and most diligent members of a population generally receive an above-average return on their endeavors, whether they migrate or not. Those who are negatively selected probably fall into occupations of lower social status than do their old or new compatriots. The assumption behind a positive selection effect is that because migration is emotionally and financially costly, people with the resources and the personality traits required to cope with a new situation are usually the ones who take the opportunity to migrate. The assumption behind a negative selection effect is that the mass of “floating proletarians” with limited resources who have to migrate for one reason or another probably would not flourish if they had stayed.¹⁶

Many studies claim that those who migrate for economic reasons do so because they expect higher returns on their skills in their new country than in their old country. Such migrants’ earnings could accrue from the same occupation or a new occupation in the new country. If earnings are generally higher in a new country, a broader segment of the population would be likely to migrate (positive selection) than if earnings are lower (negative selection). Hence, selection effects vary between countries, and over time, according to differences in earning profiles between countries or time periods. Abramitzky and Boustan find that past migrants to the United States could be either positively or negatively selected, depending on their country of origin, whereas today’s migrants tend to be positively selected.¹⁷

Not many historical studies look at selection effects in the period under study here. Wegge studied men and women who migrated from the German principality of Hesse-Cassel in the 1850s predominantly to the United States. Half of them were skilled artisans and half laborers and farmers. She found that most of these migrants were in the middle of Hesse-Cassel’s wealth hierarchy (most types of artisan). Those at the bottom there (general laborers and servants) might not have been able to afford the costly fare for a transatlantic voyage, whereas those at the top of the wealth pyramid had no apparent interest in relocating. In terms of wealth, the emigrant population was, on the whole, negatively selected, but in terms of skills, it was positively selected.¹⁸

Bosma and Mandemakers compared migrants, many of them soldiers, who went from the Netherlands to the Dutch East Indies with a control group drawn from the Historical Sample of the Netherlands, 1812–1922. They found that these migrants tended to come from the highest social class (including officers) and that few of them had a farming background.¹⁹

Abramitzky, Boustan, and Eriksson compared the imputed earnings of Norwegians migrating to the United States in the second half of the nineteenth century with that of their brothers who stayed in Norway. Although this strategy did not eliminate the selection effect completely, it revealed that, on average, siblings from poor households with lower human capital did not do as well in the labor market as did siblings from richer households with more human capital. Abramitzky, Boustan, and Eriksson found consistent evidence of negative selection for Norwegian migrants born in rural areas and mixed evidence for urban migrants during the Age of Mass Migration from Europe to the New World, 1850–1913.²⁰

Given the state of knowledge, we cannot make easy predictions for selective migration to Algeria, but we can distill a few pointers from the literature. Before the start of our research period, French authorities used propaganda and subsidies to persuade its poorer citizens to migrate to Algeria. Although this strategy seems to have worked in the economically difficult years around 1848, the literature does not suggest that it was a strong push factor thereafter. In any event, migration costs from France to Algeria were not high compared with those for transatlantic migration; we have no reason to suspect that traveling to Algeria would necessarily be cost-prohibitive. The period under study, which constitutes the maturation of a settler society, offered a wide range of lower- as well as higher-class occupations.

Unlike for earlier years, the literature does not suggest that the higher social classes in France were completely uninterested in migrating to Algeria, which, after all, was only a short distance across the Mediterranean and had strong ties to France. Because Algeria was a French region and an integrated part of the national labor market, many of the same business and educational selection systems operated there as well, at least by the time when the colonial economy had come of age. Major companies were likely to have recruited in France for employees in their Algerian branches. The highest positions, then as well as now, were accessible primarily to those who had earned top grades from France’s most prestigious schools, where their favored potential incumbents were educated. Those in Algeria had by no means lost human/social capital in either education or employment; they were subject to the same scrutiny and standards as were their counterparts in France. In fact, they brought from Paris, as the metropole of the French colonial empire, novel human and cultural capital, in the form of language proficiency, which carried a mastery of new cultural codes and tastes, as well as fashionable clothes, the belles-lettres, and various leisure activities. The literature, though far from conclusive, suggests that work was available in the colonies for Frenchmen of all social ranks. We have no a priori reason to assume that only the “best” or the “worst” segments of society migrated. We will, however, test this assumption whenever possible.

Based on the literature, the assumptions herein are threefold: (1) that French citizens born in Algeria had the highest occupational status because of their parental networks; (2) that the status of French immigrants was higher than that of other European immigrants; and (3) that immigrants who relocated to Algeria as children achieved a higher status than did those who relocated as adults. The literature also suggests that the status of immigrants surpassed that of stayers. The novelty of this article lies not in its comparison between immigrants and stayers or the host population, using individual-level data, but in its application of theoretical notions from the social sciences to Algeria’s historical context.

To test the theories, we present a typology of colonial migrants (Table 1) that distinguishes between non-migrants and migrants, further subdivided by birth in France or elsewhere and by migration as children with parents or later as adults. Those who migrated as adults are termed first-generation and their colonial-born children second-generation. We cannot distinguish the second from the third and subsequent generations; our data, which consist of unlinked marriage records, do not include parents’ place of birth. We can, however, see which bridegrooms probably migrated as children, because the marriage record states that their parents resided in Algeria. These migrants are called the 1.5 generation.²¹

Although theoretically this classification leads to nine types of bridegroom, some of the combinations are rare; six migration types are observed in our data. According to the assumptions discussed above, non-migrants (those born in Algeria who were not Arab or Berber) likely attained the highest social status at marriage, and those born outside France who migrated to Algeria as adults the lowest social status (see Table 2, Panel A). This table also shows that those who migrated with their parents as children achieved a higher social status than those who migrated as adults—once again, French migrants earning a higher status than those for whom French was not the mother tongue. Table 2, Panel B shows our expectations when comparing migrants from France to Algeria with the French who stayed behind. We expect the migrants to have been the most successful, and internal migrants among the stayers to have achieved higher status than non-migrants.

We work with two data sets, one for Algeria and one relating to mainland France. The French civil-registration system covered both countries. Even before Algeria became incorporated into the French départemental structure, colonial Algeria had copied the vital registration system from the French mainland, as introduced by Napoleon. Hence, the information in the Algerian colonial marriage registers is directly comparable to that in mainland France, though the extent of the population covered in each differs. The marriage registers contain place and date of marriage; names of bridegroom and bride; their places of birth, ages, and places of residence; whether a first marriage; occupations, if any; signatures, if able to sign; and parents’ occupations and countries of residence.

The Algerian data derive from the vital registration system, which covers French citizens. Before 1870, Muslims and Jews in Algeria were considered sujet (subalterns), not citoyens (citizens). The Crémieux Decree (1870) accorded citizenship to Algerian Jews, but not to Algeria’s Muslim majority. Muslims could obtain citizenship only by relinquishing their “status personnel” and renouncing their faith, which not many did. The French marriages analyzed herein thus relate only to the colony’s Christian and Jewish minorities, not to Muslims.²²

The first year of our observation period, 1870, is conventionally regarded as the beginning of Algerian colonial society in its prime and the point at which native Algerian Jews were obliged to become French citizens. Starting earlier would have impeded comparability with data for the later period, and it would have reduced the number of cities for which marriage records are available. Privacy restrictions prevent us from studying the years after 1912. We took a sample of marriage records for four Algerian towns—the capital Algiers and the three provincial centers of Bône, Constantine, and Oran across the coastline—from two periods, 1870–1872 and 1910–1912. The cities selected capture a range of urban experiences across the colony in the period that led to a distinct Algerian colonial culture. Algiers had the most diversified economy. It harbored a large service sector catering to the administration of the entire colony, as well as a sizable production sector, not just for the amenities of life but also for selling luxury products to Algeria’s wealthy colonials.²³

Oran—a harbor city along a riverbed, surrounded by fruit plantations and known for its grain, cattle, leather, and wool trade—was also a minor administrative center with its own ruler until the French rebranded it as a province. In the heyday of Algerian colonialism, it witnessed huge population growth and a construction boom, mainly of Spanish-style buildings, testifying to the Spanish descent of its European inhabitants.

Before the French conquest, Constantine also had its own ruler, who paid taxes to the dey of Algiers, but it, too, became a French province, alongside the towns of Philippeville and Bône. It lay on a plateau near a river, which flowed through fertile land. Its geography and mild weather allowed for abundant harvests of millet, opium, saffron, and grapes, as well as stock raising in surrounding meadows. Constantine had long since served as a provisioning and trade hub for caravans from sub-Saharan Africa to the coastal cities of Bône and Tunis. During French rule, it also gradually became an industrial city.

Bône (now Annaba)—part of the same province, a little more to the east than Constantine, bordering Tunisia—was a point of entry and exit and a trading post. Surrounded by fertile, marshy ground that, with proper irrigation, became suitable for vineyards and other forms of horticulture. It had connections not only to Constantine and other Algerian cities but also to Jewish merchants in Livorno (Leghorn, Italy) and to French traders in Marseilles, who bought Algerian grain, hides, and other raw materials and sold sugar, silk, and other luxury goods. It had railroad construction and mining activities in the nearby desert.

For Algeria, we aimed to have data for 700 marriages in the first period and 1,000 marriages in the second, and for the provincial towns 300 marriages in the first period and 500 in the second. We divided the total number of marriages in the sources by these numbers (resulting in a specific value x per city) and digitized every xth marriage (N = 4,156). We deleted 13 men younger than 18 at marriage, 695 men with unknown migration status, and 77 men with highly uncommon migration status (see description of typology). Finally, we excluded 199 men for whom we have no information about the dependent variable occupational status at marriage (remaining N = 3,172).²⁴

For mainland France, we used the large tra data set. The starting point for the data collection encompasses all marriages involving men with family names in mainland France that start with the letters tra between 1803 and 1832, plus the marriages of their descendants. From the tra survey we selected those married in the same years for which we have Algerian data, 1870–1872 and 1910–1912 (N = 2,399). We deleted 6 men who married before age 18 and 185 men for whom information on occupational status at marriage is lacking (remaining N = 2,208).²⁵

Both data sets contain missing values for some variables, notably father’s occupational status (50.5 percent missing in the Algerian data and 54.9 percent missing in tra). We used multiple imputation in spss, separately for the two data sets (twenty imputed data sets) to include these cases in the analyses. To improve the imputation, we included all cases from the tra data between 1870 and 1912, deleted the cases with missing values for the dependent variable and migration status only after imputing, and added status of the father-in-law, bride’s age, whether the bride signed the marriage certificate, and the bridegroom’s country of birth in the imputation procedure. Descriptive statistics of the variables after imputation are similar to those after listwise or pairwise deletion.²⁶

To place the migrant typology for the Algerian data in operation, we first categorized the birthplaces of the bridegrooms and the places of residence of the fathers (at the time of their son’s marriage) into “Algeria,” “France,” and “other.” Bridegrooms’ birthplaces are almost always known, but fathers’ places of residence are not (the 688 missing ones are excluded, as indicated in the data description). Non-migrants are bridegrooms born in Algeria with fathers who also lived there. Migrants from France (1.5 generation) were born in France but married in Algeria with fathers also living in Algeria. Migrants from France (first-generation) differ from the 1.5 generation in that their father did not live in Algeria when they married there. The two categories for migrants from elsewhere are constructed in a similar way. In the tra data, we distinguish non-migrants—that is, bridegrooms who marry in the place where they are born—and internal migrants who marry in a different place. The data cover all migration types in the typology presented earlier (Table 3).

For more detail about the migration streams to Algeria, we distinguish the most important countries of origin in the “other” category and describe the origins of the Algerian bridegrooms in Table 4. About half of all marrying men came from Algeria. If not pieds-noirs by birth, they came primarily from mainland France, but about one in six bridegrooms came from Spain and Italy. Oran and Algiers had more Spanish migrants, and Bône had more Italian migrant bridegrooms. Over time, the proportion of bridegrooms born in Algeria increased greatly. These migration streams, well covered by the literature, suggest a labor market for Algerian citizens that was segmented by country of origin. Italians from Livorno, Pisa, and Genoa had settled on the shores of Algeria for centuries prior to the French invasion of 1830 (the excuse for which being the undiplomatic treatment of a French envoy); they continued to come thereafter. Many poor men from Sardinia and Elba came to work as truck gardeners, especially in the region around Bône. The Maltese (the largest group in the “other” category) came to Algeria because it was nearby and because Malta was congested. Miners, masons, and shoemakers came from mainland Italy to flee starvation wages. Whereas the Italian-born were over-represented in the construction and building industry and the Maltese in farming, the native French virtually monopolized colonial administration and the liberal professions. Jews worked mostly in textiles and in the precious metals industry. At the lower end of the labor market were native Algerians, toiling in agriculture, in inexpensive restaurants catering to Algerians, in unskilled day labor, and (together with Jews) in processing tobacco into cigarettes.²⁷

We use occupations to capture social positions in the marriage registers of both Algeria and the French mainland, first coding them according to the oft-used historical international standard classification of occupations (hisco). This coding grid is the historical extension of isco68, developed by occupational coders worldwide under the guidance of the International Labor Organization (ilo) to ensure comparability between coded censuses and other vital registers across the globe. hisco has about 1,600 groups in its labor classification, as well as auxiliary variables indicating, for example, artisanal or supervisory status. hisco and its derivatives are in use worldwide, not only for past and present European countries but also for historical societies in Brazil, Egypt, South Africa, Taiwan, Uganda, and Uruguay. Once coded in hisco, the occupations can easily be regrouped into the hisclass scheme—in its fullest form, twelve social classes, based on elementary distinctions between economic sector, manual or non-manual work, skill level, and supervisory status.²⁸

Table 5 provides an impression of the social structure in colonial Algeria and France by showing the social classes of bridegrooms and their fathers and, for Algeria, of bridegrooms in the four cities and for both periods. All social layers and economic sectors were represented among the Algeria’s colonial male labor force as reflected in the inflow at marriage of the bridegrooms. Unsurprisingly, few Algerian urban citizens were farmers, but a sizable proportion were farmer’s sons. Algeria had a relatively large number of small and large merchants and retailers. Our finding that Algeria had relatively larger high-status occupational classes than did France is not surprising since France’s native population comprised all individuals and classes, whereas we could include only Algerian citizens, not native Arabs. The differences in the class distribution among the Algerian towns were small, but Algiers had the largest proportion of men working in the lower clerical and sales class, whereas Constantine had relatively few unskilled but many skilled workers. Over time, the top and the bottom of the social pyramid remained stable for the most part, but the middle saw a slight shift from non-manual to manual occupations; that is, the proportion of skilled workers increased and that of lower professionals, clerks, and salespersons decreased.

The dependent variable in our analyses is the occupational status of bridegrooms, as measured via hiscam—a historical version of the contemporary camsis scales based on social-interaction distance, often used in historical stratification research. hiscam presupposes that the frequency of interaction between people from different occupational strata is representative of the overall occupational structure: The more frequent the interaction between individuals from two different occupations, the smaller the distance between these occupations in the overall stratification. The hiscam scale ranges from 1 (low) to 99 (high).²⁹

Of the presumably many factors besides migration status that might have contributed to success on the social ladder in a colony, we take only the most important ones into account. Given that the occupational attainment of a son is greatly influenced by that of his father, we control for father’s status, which we also measured via hiscam. We also control for age and the ability to sign one’s name, since illiteracy (not being able to sign) is a drawback. Although signing is only a rough indicator of human capital, it is a valuable tool for historical research because of the high correlation between the ability to sign and literacy. We also include period and town in the analyses.³⁰

The other obvious characteristic to treat as a control variable under normal circumstances would be religion, since the social status of Jews was generally lower than that of Christians in Algeria. In the case of colonial Algeria, however, the ability to infer religion at all is an anomaly. French marriage records do not normally include religion. In one town from 1870 to 1872, however, we can identify some Jews, though not all. Some records mention that a bride or bridegroom is an “Israelite” (as opposed to a European). Furthermore, in some cases, names were written in Hebrew. But, since only the vital registers of one town in one period contain such information officially, we refrained from adding religion as a control variable.

As Güveli et al. pointed out, most migrants move to improve the prospects that their country of origin could offer them. Whether they succeed is impossible for us to determine at the individual level. But we can compare migrants not only with the population (at least the citizens) in the country of destination but also with similar others in the country of origin. Such data are rare both in contemporary migration research and in historical research.³¹

We can gauge a more complex relationship by estimating the effect of various determinants of status attainment and thus probe further than a bivariate analysis would permit into the relationship between migration and status attainment. For example, we can explore whether differences in age, period, human capital, and social origin underlie the directly observable differences in social status between residents and migrants. We turn first, however, to a comparison of migrant types within Algeria and then compare migrants to Algeria with the population that stayed in France. Finally, as a robustness check, we repeat the main analyses, selecting only Frenchmen marrying in cities, since we can study only the urban population in Algeria.

Table 2 listed our predictions about the effects of migration on status attainment. Our theories led us to believe that, of all the citizens in a colonial society, those born there would have the highest status, followed by migrants from the mother country (France), and that those migrating as children (1.5 generation) would have an advantage over those migrating as adults (first generation). Assuming that our theories are correct, those migrants who did not share the same language and culture as that of the motherland, such as migrants from Italy, would be the most disadvantaged—again, the first generation being worse-off than the 1.5 generation.

Table 6 shows the average status of Algerian bridegrooms per migration type, without controls. The status scores range from 51 (1.5 generation not from France) to 64/5 (first and 1.5 generations from France). Hence, language proficiency appears to be an important factor in determining social status: Among the first-generation migrants, those born in France have an average status more than ten points higher than that of those born elsewhere (64.1 versus 52.4); for migrants from the 1.5 generation, the gap is slightly wider (64.7 versus 51.1). The difference in average status between the first and the 1.5 generations is small, both for those coming from France and those coming from elsewhere, suggesting that migration as a child with parents or as an adult did not much matter. Another clear case in which the data refute the theory is that of native-born Algerian citizens. Far from having the highest social status (as the theories predict), they attained an average of only 58 status points, about the same as the average for all the bridegrooms, lower than the average for those coming from France, and higher than that for those who came from elsewhere.

Table 7 elaborates the occupational differences between the various migrant types. Migrants from countries other than France occupied the lower reaches of the urban labor market, as unskilled workers, much more often than did other types of migrants and non-migrants. At the top of the colonial class system (among managers, professionals, and clerical and sales personnel), those coming from France were over-represented. Non-migrants took a position midway between the two categories. They were less often found among the managerial and professional classes than were migrants from France, and they were far less often found among the unskilled working classes than were migrants from other countries. But they were clearly over-represented among the skilled workers.

Table 7 does not offer all the reasons. Compositional differences—for example, social background (father’s status) or the ability to sign—may bear some responsibility. Table 8 shows these compositional differences, supporting the notion that migrants from France performed well because of superior parental resources and human capital. In the two groups of migrants from France, most bridegrooms had fathers who were higher managers or higher professionals. Moreover, migrants from France were more likely to be able to sign their marriage certificates. In both respects, they are most unlike the migrants from other countries; the non-migrants occupy a position midway between the two categories. Just under half of the migrants (44 percent to 47 percent) who came to Algeria from countries other than France did not sign their marriage certificates; they were probably illiterate. The corresponding figure for the migrants from France was just between 3 and 4 percent. Furthermore, not only did migrants from other countries often take unskilled work in Algeria, they were also likely to originate from unskilled working classes, as implied by their fathers’ occupations.

Finally, we test whether the differences in occupational status between the migrant groups are indeed the result of these compositional differences by regressing the status of bridegrooms in Algeria on migrant type and the control variables. Although the general pattern remains the same, the differences decline by c. 50 percent (Table 9). For example, the difference in status between non-migrants and the 1.5 generation coming from France was 6.3 status points in the bivariate analysis and is now 3.6. The difference between non-migrants and the 1.5 generation from other countries declined from 7.3 to 2.8. A father’s status was a great driver of a son’s status in colonial Algeria. One standard deviation in fathers’ status (10.5 points) led to a 5.3-point increase in sons’. This strong effect is similar to that of being literate (6.5 points). Time and place show no significant differences. Hence, the multivariate analysis reveals that the previously observed bivariate effects of migration type do not disappear if we control for social background (father’s status), human capital (being literate), age, and place and period of marriage.

To see whether an average Frenchman would have done better by staying in France or by migrating to Algeria, we compare the position attained by the French in France, as indicated by our second data set, with that of our colonial Algerians in the first data set. Table 10, which describes the occupational status of bridegrooms in France in the two periods, shows that those who moved within France had, on average, a slightly higher status (55.3) than those who stayed (55.0). This difference is miniscule, however, compared with the average status of those who migrated to Algeria as children (64.7) or as adults (64.1). Hence, Frenchmen who migrated to the colonies, at least to Algeria, attained a higher status than those who stayed in France. The reason could involve the characteristics of the migrants—a positive selection of the population—or the characteristics of the colony, notably more favorable labor and land markets. Those migrants who came to the colony from outside France were even less successful than the French non-migrants (but possibly more successful than their unobserved compatriots who did not migrate): As shown above, their average status was 51.1 (for the 1.5 generation) and 52.4 (for the first generation). Finally, on average, non-migrants in France attained a slightly lower status than non-migrant citizens in Algeria (58.4).

As before, we move from a bivariate to a multivariate analysis. When we control for father’s status and the age and literacy of the bridegroom, as well as for the period, the pattern of differences in attained social status between migrant types that we saw in the bivariate analyses remains the same, although again the differences become smaller (Table 11). Staying in France led to a six- to seven-point lower status, both for resident Frenchmen and for internal migrants, compared with migrating to Algeria. On average, a man gained more status by moving to the colony than by remaining in mainland France. This benefit of colonial migration was not the result of fathers’ economic or social resources, such as money, information, or other intercession, since we controlled for fathers’ status. Although not the main focus of our study, the difference between men who stayed in France and men who migrated to Algeria from other countries disappears after taking the control variables into account. The very low status of the migrants from elsewhere in Europe is the result of a lack of resources, whereas the very high status of migrants from France to Algeria is the result not of selection but the colonial relationship between France and Algeria at the time.

Finally, we repeat the main analyses, this time including only French bridegrooms who marry in a city because they provide a better comparison with those in our data who married in Algerian cities. The non-migrants and internal migrants in the cities have a higher status than those in the whole of France. The status score for non-migrants is 57.3 versus 55.3 and for internal migrants 56.8 versus 55.0—only slightly lower than the status of non-migrants in Algeria (58.4). If we control for the resources of the bridegrooms, both categories of bridegroom marrying in France become indistinguishable from the non-migrant bridegrooms in Algeria (Table 12). Our main result, however, does not change. The first- and 1.5-generation migrants from France to Algeria still achieve a considerably higher status than those who stay in France. Although not the main focus of this study, the worse outcomes for those who migrated from other countries to Algerian cities relative to those for French non-migrants is certainly intriguing. It suggests that migration to a country without a colonial tie to the country of origin and with a different spoken language harmed a migrant’s initial position in the labor market.

This article advances the discussion of historical and contemporary studies of intergenerational social mobility, migration, and the nature of colonial societies, while introducing an interesting Francophone case into a literature dominated by attention to English-speaking settlers. The status attainment of migrants to colonies has not seen much quantitative study. We asked how the process of migration influences status by comparing migrants to Algeria with stayers in both France and Algeria. Do migrants pay a penalty compared to natives? If so, which migrants suffered the most? To answer, we compared natives with migrants in Algeria, distinguishing between French migrants who shared the dominant language in the destination country and other migrants who did not. We also looked at differences between the first and 1.5 generations of migrants in the years 1870–1872 and 1910–1912.

In the future, our study could extend into more of the twentieth century by including the non-public civil marriage records or the ecclesiastical marriage records of the Roman Catholic Church, the originals of which for the dioceses of Algiers, Oran, and Constantine are now in France. A study of the status attained at a later age would also be worthwhile. Linking marriage records would allow us to follow the career patterns of these men between their own marriage and that of their children. Future studies could also investigate the extent to which data are available outside the main cities in this study. Some migrants settled on land outside the cities. Including those who remained farmers could affect our results, as could including migrants from mainland France to Algeria who soon returned to France. Future research could also look at aspects unstudied herein—such as return migration to France, migrants who did not marry in Algeria, Algerian citizens who temporarily went to France to marry—as well as status attainment outside our four cities. A detailed case study of migration from a location in mainland France to Algeria through an internal labor market, a branch of government, or a company would also be illuminating. Though identifying the right sources might prove difficult, an examination of the subaltern, non-citizen populations might help to unravel the internal mechanisms of colonial domination.

Dealing with the questions at hand in this article, our theoretical framework, drawn from the sociological and economic-history literature, is abstract enough to have a wide reach but concrete enough to be tested. At this point, we cannot surmise the extent to which our findings can be generalized, but we offer them as null hypotheses for further research.

In line with our theoretical expectations, migrants to Algeria from outside France were observably less successful than those from France. Not only did they originate from lower social classes, and likely a negative selection from Italy and other mostly Mediterranean countries, they also lacked the language proficiency of French migrants and the understanding of how French businesses and colonial administration worked—not to mention the valuable asset of having friends and connections in high places.

Contrary to expectations, migrants to Algeria from France were more successful than native-born (non-Muslim) Algerian citizens. One reason might be their better resources. On average, French migrants were more literate than Algerian non-migrants and their fathers more professionally elevated than were those of Algerian non-migrants. Another possible explanation might be that migrants from France had closer ties to the colonial administration and business than did people already residing in Algeria. After all, Algeria, which comprised three départements of France, was just a short distance away by boat. Some of the French migrants may well have been brought to Algeria with the express purpose of filling high-status positions, or they may have had pre-existing connections that enabled them to obtain such employment. Many of these newcomers may have arrived with the latest skills in, say, winemaking or construction, giving them a leg up on those born in Algeria. Arguably, because Algeria functioned as a French region and an integrated part of the national labor market, the same business- and educational-selection systems operated there. In the absence of information, one might assume that major companies would recruit men in France to fill important positions in their Algerian branches.³²

The most prestigious positions, then as well as now, were accessible predominantly to those who attended the top schools in France and earned the highest grades. French migrants to Algeria at this level were not likely to lose human capital when they moved. Those who migrated from Paris, or even from the provincial universities and technical schools, brought new knowledge and new human and cultural capital to Algeria that were redolent of success, even though they were not born into local networks of friends and patrons.³³

This potential explanation is also in accord with our second unexpected finding. The 1.5-generation migrants were no more successful than were those of the first generation. We expected that children who came to Algeria would be at an advantage, because they had the benefit of parents to support them in their careers. But this process might have been offset by their parents’ direct ties to the colonial administration and their pre-ordained, nontransferable positions high in the colonial hierarchy. The French educational- and business-selection procedures might have made such ties especially favorable for new migrants. If this interpretation is correct, sociological theories of migration need a little tweaking when applied to French colonies, especially to nearby Algeria, which had important votes in French parliament that could not be ignored.³⁴

Future research could focus on case studies of the metropole exporting human and cultural capital to the colonies, favoring French immigrants whose strengths could offset any losses in local social capital, even in a colonial setting rife with clientelism. Frenchmen who stayed in mainland France, both internal movers and stayers, achieved a lower status than those who migrated to Algeria from France. Did the most talented migrate, or were the opportunities in Algeria clearly better than in the homeland? We cannot rule out the possibility that colonial migrants had unobserved traits that set them above non-migrants, such as industriousness or the ability to adapt quickly to new circumstances. The evidence herein, however, does not point in that direction: Neither differences in social background nor in basic human capital can fully explain the status differences between migrants and residents.

Frenchmen born in mainland France had more to gain than mere status by migrating to the colonies. The perks that befell them in colonial society were greater than those that they were likely to enjoy in France. As Abu-Lughod observed for migrants to the Maghreb, “[They] aspired to a life not at the level to which they had been accustomed, but at the level to which they wished to become accustomed. For, as was the case with Englishmen who went to India, most of the immigrants who flocked to Morocco were working-class or petit-bourgeois at best. They migrated chiefly to improve their positions, and the standards they wished to establish for themselves in Morocco were definitely at a level far above their lifestyles at home; they were commensurate with their new positions as part of the ruling caste.”³⁵

The dual reward associated with colonial migration—gaining social status and enjoying the advantages of being colonizers—was apparently widespread and consequential. It helped to create the kind of deep bond between colonizers and colony that ultimately made leaving more problematical when native Algerians launched an armed campaign for independence. In Sessions’ words, “French Algeria was a textbook case of this pattern. Despite their often-humble origins, the Français d’Algérie came to enjoy tremendous economic privilege that was defended by their elected representatives and a powerful lobby in Paris, and their sense of familial affinity with the metropole was strengthened by the assimilationist fiction that Algeria was part of France itself. Hence, it took eight years of fighting for Algerian nationalists to uproot the settlers in the twentieth century, and the horrific violence of that struggle left scars that long endure after … Algerian independence.”³⁶