The Son Also Rises Read online

  FIGURE 11.1. Shantytown, Valparaiso, Chile.

  In chapter 6 it was argued that the inequality-immobility association is an illusion created by the fact that common measures of social status, such as income, indicate only with some degree of error the true underlying status of families. In societies like Sweden, where ranges of income and wealth are compressed, these measures are particularly poor indicators of underlying social status. Consequently, these societies seem to have rapid rates of social mobility.

  In contrast, in a society of great income and wealth inequality such as Chile, these factors are much better indicators of the true social status of families: income mobility comes closer to measuring the true underlying rate of social mobility. This chapter investigates whether social mobility rates in an inegalitarian society such as Chile truly are lower than in egalitarian Sweden.

  The 2004 Chilean Electoral Register

  Chile provides a nice instrument which can be used to estimate social mobility over the past sixty years: the 2004 electoral register. This data set provides the full name, date of birth, profession, and place of residence of all voters—more than eight million people. Using this register, the population can be divided into two birth generations, 1920–49 and 1950–79, and average social status measured across these generations. The youngest persons under this division were age 25 at the time of the census and so already had an occupation. Alternatively, people can be divided into six birth decades, starting in 1920–29 and ending with those born 1970–79.

  The electoral register does not provide data on income or wealth, but knowing individuals’ occupations, their income can be estimated from the average income for those occupations (this information is available for more than three hundred occupations).2 Knowing average earnings by location, it is also possible to calculate for each person earnings based on their location. These “locational earnings” provide a nice check on occupational earnings, since for some occupational categories (e.g., “employee,” “merchant,” or “farmer”) the occupational-income measure involves much error. Estimating earnings in multiple ways with different sources of error makes the results more robust.

  Evidence that these occupational and locational earnings estimates are informative about social status is shown in figure 11.2. Each comuna (municipality) in Chile has an estimated average measure on the Human Development Index (HDI) calculated for 1990–98 by the United Nations Development Program in its study of the Chilean economy.3 The figure groups the comunas in Chile into five categories with ranks 0.5–0.59 through 0.9–1.0 on this index. The average occupational earnings for each of these comuna groupings is calculated using the 2004 electoral register. These are shown on the vertical axis. Because average earnings by location are also known, average earnings for each of the comuna are also calculable. The figure shows that the two inferred earnings measures are highly correlated and, further, that both these earnings measures are highly correlated with the broader measure of HDI by comuna. Thus there is good information from the electoral census not just on implied earnings, but also on more general implied living conditions.

  FIGURE 11.2. Salaries by community, Human Development Index grouping.

  Underclass and Elite Surname Groups

  As in earlier chapters, social mobility in Chile is estimated by identifying surnames that had high or low status before 1920 and then measuring how quickly the average status of these surnames regresses to the mean using the information from the electoral register.

  One general feature of Chile’s population is an inverse correlation between social class and degree of indigenous ancestry. Studies suggest that for the upper class, 73–91 percent of their genetics is Caucasian, for the middle class 68–70 percent, and for the lower class 41–48 percent.4 So the Mapuche, the major indigenous population of Chile, would be expected to be at the lowest rung of the social ladder.

  When the Spaniards first arrived in Chile, the Mapuche had been fending off the Incas for decades. They occupied a vast amount of land in southern Chile, from Talca to Valdivia. The rich natural terrain had allowed them to maintain a population estimated in the hundreds of thousands while remaining hunter-gatherers. The Spaniards eventually recognized the Bío-Bío River as the border between the areas of Spanish control and independent Mapuche terrain. Under Spanish influence, the Mapuche economy shifted to subsistence agriculture and cattle herding.

  From 1861 to 1883, the Chilean government progressively conquered Mapuche territory and incorporated it, and the Mapuche people, into Chile. By 1900 the Mapuche held titles to only one-tenth of their original land. They lacked the technology and know-how to compete in the new economic environment they had been compelled to enter. They became indebted and a source of low-skill labor for the main landowners.

  Four percent of the contemporary Chilean population is of Mapuche origin. The Mapuche have distinctive surnames, over four hundred of which can be identified.5 This cohort of surnames thus represents the lowest-skilled and most disadvantaged community in Chile. Although they used to be a rural population, confined to southern Chile, today nearly half the Mapuche reside in the largest city, Santiago.

  At the other end of the social scale in colonial times were the encomenderos. These were Europeans supposedly appointed as protectors of the indigenous population, but in reality grandees living off the labor of their charges. The tributes and benefits obtained by the encomenderos varied over time, but they undoubtedly represented the dominant socioeconomic elite of a polarized society.6 The list of encomendero surnames used here consists of rare surnames such as Oyarzun, Ureta, and Iparaguirre, all of which had at least one encomendero holder but fewer than fifty holders in 1853 and 2004.7 These were the original oligarchs of the colonial era.

  Another influential and prosperous colonial elite was the Basque community. A significant proportion of governors under the Spanish had Basque last names. By 1800, a quarter of the Chilean population was of Basque origin. The Basques specialized in trade and commerce, employing important connections with their relatives in Europe. They eventually married into the landed colonial elite and formed the Castilian-Basque aristocracy that played a substantial role in Chilean history. To track their social status, a randomly chosen group of Basque last names appearing in the 1853 census is employed. Most of them are rare surnames.8

  Another means of identifying elites is by landholdings. An agricultural yield report was compiled in 1853 to determine land taxes.9 This report includes the records of a large sample of landowners, who together possessed nearly fifteen thousand parcels of land. Owners with rare surnames (those appearing only three to thirty times in the 1853 census) were selected and divided into four wealth groups. These were based on the average annual value in pesos of the land owned: 350 or less (small); 350–1499 (medium); 1500–4999 (large); and 5,000 or more (very large). The last group represents the landowning elite in Chile in 1853.

  In the 1920s, Juvenal Valenzuela created a detailed list of the major agricultural estates in Chile.10 He visited them and estimated the land quality and area irrigated to estimate the value of each estate. He found that in 1920, 10 percent of the landowners held 90 percent of the land, and he created a list of the top one thousand owners. From this list have been selected the surnames that had a frequency of three to thirty in the earliest-born cohorts in the 2004 electoral register. These are the large landowners of 1920.

  One additional wealth-based elite is a set of the rare surnames of the rich of the nineteenth century. After independence, between 1830 and 1930, Chile experienced a period of territorial expansion and economic growth (along with social unrest and recurrent civil war), due in part to the exploitation of mineral resources. The population quadrupled as the economy grew. The benefits of the expansion were perceived as being primarily enjoyed by the elite. From the work of Sergio Villalobos and Ricardo Nazer Ahumada, among other sources, it is possible to create a list of the rare surnames of the nineteenth-century commercial and mining elite.11
/>   Another elite group of the nineteenth century was Italian immigrants. Half of the Italian immigrants who entered Chile in 1853 were involved in commerce or were skilled professionals. More than 80 percent of Italian immigrants could read and write in 1865, compared to 35 percent of the host Chilean population.12 The Chilean electoral register of 2004 lists over six hundred thousand Chileans of Italian descent. A sample of surnames of the more successful Italian families was obtained from the Commercial and Industrial Census of the Italian Colony in Chile for 1926–27, from which were selected surnames held by dentists, doctors, jewelers, factory owners, and the like.13 Most of the surnames are rare.

  The Germans were arguably the most notable of all the more recent immigrant groups in Chile, due to both their numbers relative to other groups and their strong presence in economic, military, and social elites. The most important of the early waves of German immigration came after the enactment of the 1845 Law of Selective Immigration, which gave special incentives to more skilled and wealthier Germans and Austrians. This law aimed to promote economic growth by populating the south-central part of the country. The idea was to create a backbone of Western work ethics and culture and to dilute native influences. Favored immigrants were given five-year loans, livestock, land, and free transportation to Chile for the entire family. Thus 54 percent of all Germans who entered Chile in 1865 were skilled workers, craftsmen, or involved in commerce. More than 82 percent of them could read and write.

  Lists of successful Germans circa 1920 are drawn from the Gazette of the Austral Region of Chile of 1920, which lists commercial and industrial firms and their owners, as well as professionals dedicated to the service industry, passenger listings, and a genealogical dictionary of surname origins.14 Most of the German surnames are rare in Chile, although, again, the surnames were not selected on that basis.

  French immigration to Chile was concentrated in the late nineteenth century. By 1930, well over twenty thousand French immigrants had been recorded. Like their Italian and German counterparts, they were merchants, artisans, and professionals, with above-average literacy rates. The list of French surnames used here is derived from a list of over five hundred significant French businessmen and entrepreneurs recorded in Chile between 1907 and 1920 in consular reports. Since most of these surnames are rare in Chile, this constitutes a list of surnames of prosperous French immigrants circa 1920.15

  Table 11.1 summarizes the numbers of people from each of these surname groups in the 2004 census who were born in the periods 1920–49 and 1950–79. The ratio of the numbers born in the later period and the earlier reveals the relative social status of these groups. For the country as a whole, the ratio is 2.3. For the Mapuche, the lowest-status group, this ratio is higher, at 2.47. The difference reflects a combination of likely higher fertility among poorer groups and lower longevity in the older age cohort.

  Confirmation that the ratio is closely linked to status comes from the surnames of the 1853 landowners. Ranking these by in increasing order by value of their land, the ratios are 2.26, 2.10, 1.91, and 1.89. Thus 150 years later, in 2004, the status of landowners in 1853 continues to influence outcomes, suggesting that there is considerable intergenerational persistence in Chile.

  Table 11.2 shows average occupational income by surname group relative to the average for all persons for those born in the periods 1920–49 and 1950–79. Also shown is the implied intergenerational correlation of average income. The individual estimates vary from 0.70 to 0.95, with an average of 0.84. Estimated correlations using locational estimates of average income are shown in the last column. The individual estimates of persistence vary here, but the average is very close to that for occupational income, at 0.83.

  TABLE 11.1. Surname samples in 2004 Chilean electoral register

  TABLE 11.2. Intergenerational correlation of occupational and locational income by surname group from 2004 electoral register

  As in the other countries, the social mobility rates estimated here are much slower than those estimated by conventional methods. Núñez and Miranda, for example, derive estimates of only 0.52–0.67 for income and educational persistence using conventional methods.16

  The estimates of the intergenerational correlation for the Mapuche, the one poor group in the table, are lower in each case than the average: 0.79 for occupational income and 0.63 for locational income. These results may reflect higher social mobility by the Mapuche than in the rest of the society: as noted, the Mapuche have relocated in large numbers from rural areas to Santiago. But these figures may also reflect the difficulty in estimating occupational and locational earnings for this still heavily rural group.

  There is a distinct possibility that these correlation estimates are systematically biased upward. Earnings vary substantially within each occupation. An elite group, such as the descendants of the landholders of 1920, is likely to produce attorneys who earn more than the typical attorney and doctors who earn more than the average doctor. As this elite group regresses toward the mean, so will the salaries of group members in high-income occupations. Thus there is a second type of earnings regression for elite and underclass groups that is not reflected in the average figure for occupational earnings. The intergenerational correlations reported represent upper bounds for the intergenerational correlation of earnings and of underlying status.

  The analysis so far shows that the underlying rate of social mobility in modern Chile is indeed low. Surnames that were high status in the colonial era, such as those of the encomenderos and the Basques, have remained high status in the generation born in the years 1950–79. The way that elite surname groups were identified does not affect the measured rates of social mobility. If they are defined as ethnic groups—Italians, French, Germans, Mapuche—then they show the average rates of social mobility. If the elite surnames are defined purely based on their earlier wealth, such as landholdings in 1853 or 1920, or wealth in the nineteenth century, then mobility rates are very similar.

  Thus the low mobility rates observed using surnames in Chile are not the result of ethnic groups walling themselves off and marrying only endogamously. In the surname groups defined by wealth, only rare surnames were used, so the mobility being measured here is for individual families of landowners or industrialists, which do not experience much faster rates of social mobility. It is not that there is plenty of individual family mobility within ethnic groups such as the German community, but a tendency within that group for everyone to regress to a higher German mean occupational status. Mobility is generally low.

  However, there is no evidence that Chile’s social mobility rates are any different from those of the United States, the United Kingdom, Sweden, or even Communist China. Chile serves instead to confirm the hypothesis that social mobility rates are mainly determined within families and are mostly independent of social institutions.

  Mobility by Decade

  Above, mobility is measured by generation. But it is also interesting to consider mobility by decade, even though with this approach random fluctuations appear because of the smaller amounts of data. This approach is of interest for Chile because of its relatively short-term but dramatic social disruptions in the later twentieth century. The years 1964–73, under the presidencies of Eduardo Frei and Salvador Allende, saw significant attempts to improve the educational outcomes for the poorer sections of Chilean society. That social experiment came to a calamitous end with Augusto Pinochet’s coup d’état of 1973 (figure 11.3) and the ensuing years of repression under military rule. During Pinochet’s dictatorship (1973–1990), there was reduced expenditure on public education and programs targeted at the poor and an expansion of the private sector in higher education. If public support had any effects on social mobility, then we would expect that the people born in the decade 1950–59, many of whom enjoyed the benefits of the Frei and Allende social programs as they attended secondary school and college, would experience enhanced mobility. In contrast, the people born in the decade 1970–79, most of who
m were educated in the Pinochet years, should experience more restricted mobility.

  FIGURE 11.3. The assault on the presidential palace during the 1973 Chilean coup d’état.

  FIGURE 11.4. Earnings variation by birth decade by surname type, colonial elites and underclasses, 1920–79.

  Figure 11.4 shows occupational incomes by decade for a set of surname groups from the colonial era, including the encomenderos and the Mapuche. The incomes are shown essentially as their variation above and below the mean occupational income for people of that decade.17 The scaling of the figure means that social mobility rates are portrayed by the slope of the line. It is evident that the political regimes of 1964–73 and 1973–90 had little perceptible impact on social mobility rates. The 1950–59 birth cohort did not see any systematic increase of mobility rates, despite the enhancement of educational opportunities for the poor under Frei and Allende. The 1970–79 cohort did not see any systematic decline in social mobility. There was indeed an apparent increase in occupational status for the encomendero surnames. But that is based on small numbers and may be subject to sampling error.

  Figure 11.5 shows the variation in occupational incomes by decade for other elite-surname groups: French, German, Italian, and the rare surnames of medium and large landowners in 1853. Again there is no sign of any systematic enhanced mobility for the 1950–59 birth cohort. There is a modest sign, however, of a decline in downward occupational mobility for those educated in the Pinochet years.