1 Introduction
Economic inequality in contemporary Iberian America has become a fashionable topic. And not without good reason, since it is, along with Sub-Saharan Africa, the most unequal region in the world (López and Perry Reference López and Perry2008). For The Economist, «Inequality is as Latin American as good dance music and magical-realist fiction»Footnote 1. Has economic inequality been, as the danzón, which was already danced by late 18th century in the Caribbean, conspicuously Iberian American since colonial times? Or did it appear, as the literary magical realism did, much more recently? Very likely, most economist and economic historians would answer the first question affirmatively.
On the contrary, our provisional answer, based on the evidence presented in this paper, is sceptical. Moreover, it is our contention that the empirical foundations, in particular those of a quantitative character, of the popular idea that Iberian American economic inequality has colonial origins are rather unconvincing. Sometimes they are simply non-existent at allFootnote 2.
The hypothesis that not only high inequality but also low growth in Iberian America are deeply rooted in colonial times has been defended in a series of brilliant, influential and, to a large extent, convergent works by Engerman and Sokoloff (Reference Engerman and Sokoloff1994, Reference Engerman and Sokoloff2002, Reference Engerman and Sokoloff2005) and Acemoglu et al. (Reference Acemoglu, Johnson and Robinson2002).Footnote 3 On the basis of the alleged existence of either «extractive» institutions or institutions producing extreme economic inequality, the Iberian colonial legacy is blamed for the creation of a «reversal of fortune» among European colonies in the Americas — the poorest one ca. 1,500 (i.e. the United States) became richer while the initially richest ones (i.e. Mexican and Inca empires) became poorer — or of an adverse development path that differs sharply from the one followed by the United States. Many authors have been more or less influenced by this neo-institutional interpretation of economic development in Iberian America and other parts of the world since 1500 — that is, Cogneau (Reference Cogneau2003), Angeles (Reference Angeles2007), Baker et al. (Reference Baker, Brunnschweiler and Bulte2008), Bruhn and Gallego (Reference Bruhn and Gallego2008) and Frankema (Reference Frankema2009).
The problem with this successful way of approaching the historical roots of inequality in Iberian America is that it has established a basically anti-empirical way of reasoning. In fact, it has become a commonplace to start with some reference to the colonial origins of the contemporary uneven regional distribution of assets, income and human capital. Just mentioning some words that carry strong negative connotations (mita, encomienda and hacienda, mainly) or referring to those authors who mention them (Engerman and Sokoloff and Acemoglu, Johnson and Robinson) is taken as a valid proof of the centuries-long existence of extractive institutions and of extreme economic inequality. No attempt is made to analyse the real impact across time and space of these institutions.
Thus, it is possible to consider mining as an extractive or extremely unequal institution is spite of the fact that, according to von Humboldt (Reference von Humboldt1822): (1) «The labour of a miner is entirely free throughout the whole kingdom of New Spain; and no Indian or Mestizoe can be forced to dedicate themselves to the working of mines»Footnote 4; and (2) «The Mexican miners is the best paid of all miners»Footnote 5. von Humboldt’s testimony is confirmed by many sources. As to the first point, Ladd (Reference Ladd1992) claims that von Humboldt was right: «ya para las décadas de 1750 y 1760 la mayoría de la fuerza de trabajo de las minas de plata era libre» (p. 54). In the same vein, Brading (Reference Brading1983) states that: «en la minería mexicana en su conjunto, el papel del trabajador forzado indígena disminuyó hasta ser insignificante durante el curso del siglo XVIII» (p. 202). Velasco (Reference Velasco1989) minimizes the importance of coerced labour in the late Bourbon Mexico mines and points at the wage differential as the main factor behind the inflow of free workers into the expanding mining sector. Northern mining centres were populated by free immigrants who felt the attraction of higher living standards (Swann Reference Swann1990). The mobility and the high wages of miners are also highlighted by Brading (Reference Brading1983). The disregard shown for these sources might explain why the early appearance of a genuine market for free, mobile and well-paid labour in most of the New Spain’s mining centres have passed largely unnoticed by the neo-institutionalistsFootnote 6.
It is true that in an unknown number of mining centres some forms of coerced labour were circumstantially permitted (Brading Reference Brading1983; Velasco Reference Velasco1989; Ladd Reference Ladd1992; von Mentz Reference von Mentz1998; Sánchez Santiró Reference Sánchez Santiró2002). However, their effective contribution to the total supply of labour for mining is not generally well-determined. At least in one case, that of Conde de Regla in 1764, it was insignificant (Velasco Reference Velasco1989, p. 580). In this respect, Brading’s (Reference Brading1983) claim that many mining camps in New Spain looked similar «to the British ports of the same period» may turn out relevant to the discussion on the persistence of some forms of compulsion on labour in particular parts and moments of a basically free market.
Comments by von Humboldt on miners’ wages seem also to be reliable. They find supportive evidence in Dobado and García (Reference Dobado and García2009). Their estimates of the purchasing power in terms of grain and meat of the wages of skilled and unskilled Guanajuato miners by early 19th century are surprisingly high by international standards (see section 2).
Free miners were also very important in the case of mining in the Andes, even in Potosí, where mita has stood for centuries as a genuine instance of coerced labour. By early 17th century, almost half of the indigenous labour force employed in silver production was free (Bakewell Reference Bakewell1989)Footnote 7. Still more complex is the picture of the labour market in Potosí shown by Assadourian (Reference Assadourian1987): at the very most, between one-quarter and one-third of the total labour force in the town was coerced (mitayos), while the rest was free. In the mining sector, the shares were roughly the same. By late 18th century, the number of free Indians working in the Potosí’s flourishing mining sector slightly exceeded that of mitayos, while the latter and their families constituted a minority — around a quarter — of the town’s population (Tandeter Reference Tandeter1992). On the other hand, anti-empirical approaches forget that one thing was the quota of men legally established and quite another was that the size of the effective flow of mitayos from the indigenous communities into mining centers. They also disregard the fact that the gap between the former and the later tended to increase over time (Garavaglia and Marchena Reference Garavaglia and Marchena2005). Besides, it also neglects the important fact that mita never existed at all in some main Andean mining centres — that is, Oruro and Lower Peru — and therefore paying wages was the only mechanism in place for attracting labour (Bakewell Reference Bakewell2004; Garavaglia and Marchena Reference Garavaglia and Marchena2005). Free labour played an important role in the gold mining boom of the late colonial period in Colombia, Ecuador and Chile (Garavaglia and Marchena Reference Garavaglia and Marchena2005). More generally, free markets for labour, which ought to be included among those institutions defined by Acemoglu et al. (Reference Acemoglu, Johnson and Robinson2002), as «institutions of private property», did not exist in pre-Columbian America.
Mining is also defined as a «bad» institution by Bruhn and Gallego (Reference Bruhn and Gallego2008)Footnote 8. However, according to Esteva, quoted in Miño Grijalva (Reference Miño Grijalva2001), in 18th century New Spain, one-third of the population of the mining centres was classified as «Spanish». This share is double that of «Spaniards» in the total population by early 19th centuryFootnote 9. Besides, it is simply not true that labour institutions and production technology in colonial mining were similar to those prevailing in agricultural plantations. Furthermore, was mining technology any less «repressive» in contemporary Europe? It is doubtful. In the same vein, the positive effect of mining on economic growth through the division of labour and regional markets integration is overlooked (see Dobado and Marrero, forthcoming).
No less surprising is the superficial treatment by some authors of such a complex issue as that of the changing and diverse agrarian institutions in colonial Iberian America and their varied influence on inequality. Not a single figure on real land distribution in any of the many different territories that formed the several viceroyalties during its centuries-long existence is offered (see Frankema Reference Frankema2009). Neither may it be found any comment on the existence of other, very distinct, not less important indeed, agrarian institutions (community-owned land, small and medium size farms, sharecropping, markets for free labour, etc.) and on the subtleties of their continuous interactions with each other in the long run.
With regard to New Spain, Miño’s challenging view is unequivocalFootnote 10. Similarly, this author claims that conventional opinions on the working conditions prevailing in the haciendas might be mistakenFootnote 11. Quite surprisingly, ca. 1800, almost 4,500 «pueblos de indios» had legal entity status and collectively owned substantial portions of not necessarily unfertile land (Tanck Reference Tanck1999, Reference Tanck2005)Footnote 12. In fact, many «pueblos de indios» seem to be experiencing something not very far from a sort of golden age (Tanck Reference Tanck1999). In any case, they coexisted not only with a few gigantic haciendas but also with other intermediate agrarian institutions (ranchos, small and medium holdings). This is clear in Bulmer-Thomas’s (Reference Bulmer-Thomas1994) description of the «initial conditions» of the agrarian sector after independenceFootnote 13. The importance of collective ownership of land has been more generally emphasized by Coatsworth (Reference Coatsworth2005)Footnote 14.
Baskes’ (Reference Baskes2005) revisionist view on repartimientos de bienes in late colonial Oaxaca has not been taken into account either. From his new approach, this institution, generally judged as «devised to extract wealth from native communities», is presented as an instrument that reduced transaction costs and facilitated the access of Indians to the international markets through the exports of cochineal dye.
In sum, we share Coatsworth’s unequivocal claim on the issue under discussion:
«… what little quantitative evidence there is does not suggest that ownership of land, or other assets for that matter, was more concentrated in Latin America than in the United States» (Coatsworth Reference Coatsworth2008, p. 553).
Inspired by Coatsworth’s criticism to Engerman and Sokoloff’s thesis, our study follows a basically empirical approachFootnote 15. We study wages and height in some viceroyalties (New Spain, New Granada, Peru and Río de la Plata) during the late Bourbon period from an international comparative perspective. To the best of our knowledge, this is first time that these two variables, closely interconnected and related to economic inequality, have been jointly analysed for the case of colonial Hispanic American.
Milanovic et al. (Reference Milanovic, Lindert and Williamson2008) have shown some fragmentary evidence about18th and 19th centuries Iberian America as a particular case within their pioneering study on ancient inequality. Based mostly on social tables and tax census data, this original work presents Gini coefficients for twenty-seven pre-industrial societies ranging from early 1st century Rome to British India in 1947 and includes New Spain by late 18th century. As predicted by conventional assumptions, Bourbon Mexico appears to be extremely unequal: its Gini coefficient is highest in the sample (63.5 per cent). This result would confirm mainstream visions while being apparently consistent with von Humboldt’s view: «México es el país de la desigualdad» (von Humboldt 1822:Reference von Humboldt1991, p. 168).
However, there are some problems. Surprisingly, New Spain’s Gini is practically equal to that of Holland in 1732 (63 per cent). Moreover, the interpretation by Milanovic et al. (Reference Milanovic, Lindert and Williamson2008) of Bishop Abad y Queipo’s oversimplified picture of the socio-economic stratification in late Bourbon Mexico is not unquestionable and, in any case, yields an upper bound. Moreover, somewhat disturbing is the fact that New Spain turns out to be so unequal that it is the only case in the sample of pre-industrial societies used by those authors, which lies beyond the curve («inequality possibility frontier») representing the maximum feasible economic inequality. Besides, von Humboldt’s remarks should not be always taken necessarily as prima facie evidence. On the one hand, his reference to inequality is complex since it includes other dimensions as well (geography, population density, urbanization, etc.; von Humboldt 1822:Reference von Humboldt1991, pp. 68-69). Thus, it resists an unequivocal interpretation. On the other hand, his comments are sometimes at odds with the idea of a uniquely unequal New Spain, especially when compared with some parts of Europe (von Humboldt 1822:Reference von Humboldt1991, pp. 66-67). Thus, Bourbon Mexico might well not be the only land of economic inequality after all. In this respect, in his re-examination of inequality in Iberian American over the past five centuries, Williamson (Reference Williamson2009) criticizes the assumption that it has always been relatively unequal by international standards and defends its «normality» since the pre-Columbian era to the «belle époque».
Apart from this introduction, this paper contains four sections. In section 2, evidence on nominal and real wages by early 19th century is presented. Section 3 deals with height in the 18th century. Indexes of economic inequality built on ratios relating GDP per capita to real (grain) wages and height are shown in section 4. Some final remarks appear in section 5. Sources of data and methods of estimation are described in Appendix 1 on wages and Appendix 2 on height.
2 Nominal and real wages
In Dobado and García (Reference Dobado and García2009),Footnote 16 abundant evidence on nominal and real (grain and meat) wages of both skilled and unskilled workers for a number of towns in the Americas, Asia and Europe from early 18th to early 19th centuries is presented. Since baskets of goods properly representing the consumption patterns of Bourbon American workers other than that of Leticia Arroyo for Arequipa (see http://gpih.ucdavis.edu/files/Peru_18th_c_basket.xls) do not exist, we are unable to use appropriate cost-of-living indices. Therefore, we use estimates of wages in terms of grain (an ordinary good) and of meat (a superior good) as proxies for real wages. In this paper, we focus on the two first decades of the 19th century.
Circa 1803, nominal wages of miners in Guanajuato and New Spain were clearly higher than those of other skilled urban workers in our sample (London, Amsterdam, Antwerp, Strasbourg, Istanbul, Gdansk, Leipzig and Milan) and of miners in Spain (see Dobado and García Reference Dobado and García2009). Philadelphia artisans are the only exception. However, it might be objected that the finding of high nominal wages in late Bourbon Mexico was expectable, as it was by far the main world producer of silver. Were they also high in terms of grain? Yes, they were too. Grain wages of New Spain miners were only lower than those of skilled workers in the United StatesFootnote 17. When real wages in terms of a superior good, such as meat, are estimated, differences in favour of Bourbon Mexico become enormous. Meat wages of New Spain miners were not only clearly above those of Western Europe (London, Amsterdam and Antwerp included), but also higher than those in the United States in some casesFootnote 18. The easy access to animal proteins — in contrast with Europe and Asia — in Bourbon Mexico was partially due to the comparatively low prices of beef, which in turn responds to the favourable factor endowments for extensive cattle-raising in the Northern regions of the viceroyalty. The prices of other superior goods (sugar and cocoa) might also be comparatively cheap for late colonial Mexico consumers.
Thus, Bourbon Mexico miners do not seem to be the epitome of extreme exploitation. High wages are rather incompatible than not with extractive institutions. Why should then mining be considered more «extractive» in New Spain than in England or in other parts of Europe?
In Figure 1, nominal wages of generally urban labourers in 1800-1820 are depictedFootnote 19. Again, the high level of nominal wages in Hispanic America relative to other parts of the world might be considered expectable. More surprising is to realize that they were also high in terms of grain (see Figure 2) and, especially, of meat (see Figure 3)Footnote 20.
Certainly, the early 19th century was not a favourable period for low-income consumers in Europe. However, it was probably even worse in Hispanic America where some of the conflicts surrounding the process of independence were especially violent and long lasting. Consequently, they had important negative effects on the demand for labour and on the supply of basic goods. In any case, the picture shown by available evidence on grain and meat wages from 1800 to 1820 is that of living standards of late Bourbon America labourers being much closer to — or even higher than in some cases – those of the United States and most developed countries in Western Europe than to those of poorer Central and Mediterranean Europe — including Iberian metropolis — and, especially, of Asia. Thus, while New Spain miners clearly belonged to the upper world class of labourers, Bourbon America’s unskilled workers were far from being at the bottom of the international distribution of wage earners’ fortunes. This is all but surprising since, contrary to neo-institutionalist assumptions, what characterizes labour markets in Bourbon America is the relative scarcity of this factor and not the opposite. According to Bulmer-Thomas (Reference Bulmer-Thomas1994), there was «a traditional labour shortage, from which many colonial activities had suffered» (Bulmer-Thomas Reference Bulmer-Thomas1994, p. 30)Footnote 21.
Finding medium-to-high relative real wages in Bourbon America might be interpreted as an indication that labour productivity could not be significantly lower there than in most late pre-industrial European countries. This inference does not seem to be implausible under reasonable suppositions: a similar number of working days per year; similar differences in productivity between wage earners and other segments of the labouring classes — that is, peasants. The proximity in productivity levels might have implications for estimates of GDP per capita. It suggests that an upward revision of estimates by Coatsworth (Reference Coatsworth2008) and Maddison (Reference Maddison2009) would not be unjustified.
Naturally, much more evidence confirming our findings is needed. On the other hand, extending our rather optimistic tentative conclusions about the material welfare of wage earners to other unprivileged sections of the Bourbon American society is a risky business. We should know much more about the living conditions of other segments of the commoners — that is, peasants. Notwithstanding, until further research proves otherwise, we provisionally accept that our findings on real wages in Bourbon America are as representative of the commoners’ living conditions as in other parts of the world.
3 Height in Bourbon Mexico and Venezuela
In this section, we present new quantitative evidence on height in Bourbon Mexico and Venezuela. Studying physical stature from an international comparative perspective plays a double role in our research; it is interesting in itself and serves as a relevant check of our findings on wages. If our sample is representative — nothing suggests the opposite — what we have found is that height in late colonial Hispanic America are comparable to those in Europe in spite of its allegedly lower GDP per capita. These findings are consistent with those obtained through the examination of wages in which they do not confirm the widespread idea of an especially unequal colonial society in Hispanic America (see section 4).
As a result of the growing popularity of anthropometrics after some decades of existence, Iberian America started to appear in a picture in which numerous social groups, countries and periods were already present from much longer (i.e. Komlos and Baten Reference Komlos and Baten2004; Steckel Reference Steckel1995, Reference Steckel2008, Reference Steckel2009). Thus, studies on heights in Argentina, Brazil, Colombia, Mexico and Puerto Rico during the 19th and 20th centuries are availableFootnote 22. However, anthropometric research into colonial Hispanic America is very scarce — that is, Challú (Reference Challú2009) on Bourbon Central Mexico and Salvatore (Reference Salvatore1998) and Salvatore and Baten (Reference Salvatore and Baten1998) on the late viceroyalty of the Río de la Plata). Therefore, it is necessary to widen the time and space dimensions of the colonial Hispanic American sample of height. We try to contribute to that goal by offering new data on 18th century Mexico and Venezuela. Working with military sources (filiaciones and other documents of the conscripts to the colonial militias), we have been able to build a database of almost 6000 observationsFootnote 23. Data include generations born from the 1730s to the 1780s in the Northern and Southern regions of the viceroyalty of New Spain (modern-day Mexico and the South-Western USA) and in Maracaibo (nowadays Venezuela).
The main results of our study of this database are shown in Dobado and García (Reference Dobado and García2009). In Figure 4, an international comparison of height is shownFootnote 24.
Height of militiamen, most of them working in mining and cattle raising, from the scarcely populated Northern New Spain regions were similar to those of contemporary Europeans. It implies that they were probably taller than many Asians. «Blancos» (whites) from Maracaibo were even rather taller. In Central Mexico, according either to Challú (Reference Challú2009) or to our less — because of the small size of our sample for that region — significant results, heights would be in the lower range of the available international sample. On the contrary, «blancos» from Southern New Spain were clearly the shortest in Figure 4. However, their average height is not unknown in some European regions and during certain periods of the 18th and 19th centuriesFootnote 25. However, our estimates for Southern Mexico might be somewhat downward biasedFootnote 26. In any case, our findings are not surprising as they are consistent with those from Challú (Reference Challú2009) for 18th century Central New Spain and from López-Alonso and Porras (Reference López-Alonso and Porras2007) and Carson (Reference Carson2005, Reference Carson2007) for México and South-Western United States in the 19th century. On the other hand, differences in height across regions seem to have been very persistent since a North-East stature gradient has also been found in Pre-Hispanic Mesoamerica (Márquez et al. Reference Márquez, Mccaa, Storey and Del Angel2005) and in México during the 19th and 20th centuries (López-Alonso and Porras Reference López-Alonso and Porras2007; Vélez-Grajales Reference Vélez-Grajales2009).
The study of height offers an interesting direct measure of economic inequality: the racial gap. The difference in height between «blancos» and «pardos» is perceptible by mid-18th century in Southern Mexico (less than 3 centimetres) and in Maracaibo (around 1.5 centimetres). However, it tended to decrease in both cases from the 1730s to the 1780s (Dobado and García Reference Dobado and García2009). In Southern Mexico, the gap narrows from about 4 centimetres to practically null. Moreover, the gap we find is significantly smaller than the one observed among the different social classes in some European countries (Komlos Reference Komlos2007). The racial gap in Maracaibo is similar to that existing between black slaves and free whites in the United States by the same period (Margo and Steckel Reference Margo and Steckel1983; Steckel Reference Steckel1986) and higher than that estimated for Brazil and Lima in the 19th century by Baten et al. (Reference Baten, Pelger and Twrdek2009a). To summarize, improvable as they are, our provisional estimates on the average height of Mexicans and Venezuelans of the Bourbon period are basically similar to those of Europeans, while the racial gap is comparatively small and decreasing. These findings may be interpreted as evidence against the idea of an especially unequal Bourbon Hispanic America.
4 Wages, heights and GDP per capita
In this section, we present indexes of economic inequality in late Bourbon America. These indexes consist of the ratio of GDP per capita in 1820 to grain wages in 1800-1820 and of the ratios of GDP per capita in 1700 and 1720 to height by mid-18th century.
In considering the ratio of GDP per capita to grain wages, especially those of unskilled workers as a proxy for economic inequality, we draw on the pioneering work by Williamson (Reference Williamson1999, Reference Williamson2002). In fact, what we do is trying to adapt his methodology, followed by Prados de la Escosura (Reference Prados de la Escosura2007a) as well, to the limited quantitative information existing for the Bourbon period. In our ad hoc Williamson index of inequality, we use Maddison (Reference Maddison2009) estimates of GDP per capita in 1820 and our own estimates of grain wages in 1800-1820. Our rationale is as follows: estimates of, or the educated guesses on, GDP per capita for Hispanic America by early 19th century are lower than in most Western countries; then, finding real wages of unskilled workers in colonial Hispanic America, which are similar to those in Europe indicates that, at the very least, economic inequality in New Spain, New Granada and Upper Peru should not be considered high by Western standards at the end of the Bourbon period (see Figure 5).
Rather, the contrary seems to be true. The Hispanic American values of our ad hoc Williamson inequality index are lower than in Europe. These results are clearly at odds with conventional wisdom on this issue and in particular with neo-instituionalist assumptions. That this measure of inequality turns out to be higher in Asia as well is also in contradiction with Milanovic et al. (Reference Milanovic, Lindert and Williamson2008).
Certainly, using a less crude way of calculating real wages would offer somewhat different results. Missing data for some years might also be altering the ratio corresponding to some countries. However, a shift from low to high levels of Williamson’s economic inequality is rather unlikely than not. Besides, if instead of using grain prices as deflator of nominal wages we use those of meat — or a combination of the two — the ratio of GDP per capita to real wages for Bolivia, Colombia and Mexico would be lower and consequently, by construction, our ad hoc version of the Williamson’s economic inequality would decrease. In any case, whatever the change, if reasonable, in the inputs used for calculation of this measure of economic inequality, a very different picture to that shown in Figure 5 seems rather implausible. Therefore, it is our provisional conclusion that late Bourbon America does not stand at the top of the international ranking of the Williamson’s economic inequality index. If alternative, somewhat lower, estimates of Hispanic American countries GDP per capita in 1820 provided by Coatsworth (Reference Coatsworth2008) and Prados de la Escosura (Reference Prados de la Escosura2007b) are used instead of Maddison’s, our ad hoc version of the Williamson’s economic inequality indexes shows values that are even slightly more favourable for Colombia and Mexico.
Figure 5 deserves some additional comments. On the one hand, differences between countries in our ad hoc version of Williamson’s economic inequality index are enormous, — that is, roughly a factor of eight between the two at the top and the two at the bottom. Within Hispanic America, the differences are not minor either, even though the three countries in the sample share relatively low degrees of inequality — that is, Mexico and Bolivia vs. Colombia. Are these differences due only to similar disparities in economic inequality? Do other factors intervene? In any case, the high variability in ratio of GDP per capita to grain wages deserves closer scrutiny. On the other hand, it is also striking that small differences in GDP per capita estimates coexist with big differences in nominal (grams of silver) and grain wages — that is, Bolivia vs. Japan. Again, this counterintuitive result seems worth being explored as well.
We have also explored the potential of an additional and complementary approach to the study of economic inequality when direct evidence on income distribution is doubtfully reliable or is non-existent, as it usually happens in early modern societies. As mentioned before, the anthropometric literature has been producing substantial arguments and evidence supporting the notion that height are very sensitive to economic inequality (Steckel Reference Steckel1995, Reference Steckel2005, Reference Steckel2009). On the basis of the above-mentioned literature, which causally links economic equality and average height, our reasoning is similar to the one previously presented with regard to real wages. It is our assumption that, ceteris paribus, for a certain level of GDP per capita, the higher the average height in a given country, the lesser economic inequality might be expected. In other words, finding height in Bourbon Mexico or Venezuela comparable to those in presumably more developed countries would cast serious doubt on the plausibility of mainstream assumptions on the colonial roots of economic inequality in Iberian America.
Thus, we present a first exploration of a methodology that, to the best of our knowledge, has not been used before empirically. This methodological novelty simply consists in calculating ratios of GDP per capita to average height for as many countries as possible. These ratios might constitute an alternative index of economic inequality. The underlying rationale is not only fairly intuitive but also consistent with the currently available evidence on some developed countries (Bilger Reference Bilger2004). In Figure 6, the ratio of GDP per capita in 1700 and 1820 to average height of cohorts of those born in 1750-1760 calculated for a sample of European and American countries is shown.
In 1700 as well as in 1820, the ratios of both Northern and Southern Mexico and, particularly, Venezuela are significantly lower than those of Europe. If this tentative index of economic inequality makes any sense, the inference is clear: those Hispanic American for which we have so far found original sources for height do not seem to be among the most unequal societies of the 18th and early 19th centuries. Again, rather the opposite is true. We find basic consistency between the two indexes of inequality presented here especially reassuring.
5 Final remarks
(1) Much more empirical research is needed to widen the — so far too small — quantitative information on which most claims about viceroyal America’s economic inequality are commonly based. The gap between strong assumptions and weak — or inexistent at all — empirical evidence should urgently be closed. Otherwise, the economic conditions prevailing in the American territories of the Spanish monarchy and their consequences on economic development could not be properly assessed. The importance of the issue goes far beyond academic debate.
(2) In this paper, we present an ad hoc version of the Williamson’s inequality index (the ratio of GDP per capita in 1820 to grain wages in 1800-1820) and our own inequality index (the ratio of GDP per capita in 1700 and 1820 to average height in 1750-1760) for a sample of American, Asian and European countries.
(3) In our interpretation, the limited available evidence does not support the idea that Bourbon America was an especially unequal society from an international comparative perspective.
(4) Those views on the viceroyal period and its economic long-term legacy based on assumptions about extractive, unequal or bad institutions that appeared shortly after 1500 should offer more convincing empirical evidence.
Appendix 1: Sources and methods of nominal and real wage estimates
Figure 1: Wages in Amsterdam/Holland, Antwerp/Belgium, Gdansk, Leipzig, London/Southern England, Milan, Strasbourg, Sweden and Vienna from the web page of the International Institute of Social History (http://www.iisg.nl/hpw/data.php). Wages in Bogota, China, Istanbul, Korea, Pennsylvania, Potosi and Pune from the web page of the Global Price and Income History Group (http://gpih.ucdavis.edu/Datafilelist.htm). Wages in Almadén from Dobado (Reference Dobado1989), in Kyoto from Basino and Ma (2005) and in Mexico from Challú (Reference Challú2007). Daily wages in grams of silver:
Almadén: Labourer, average of 1800-1820. Amsterdam/Holland: Opperman, average of 1800-1820. Antwerp/Belgium: mason’s labourer, summer, average of 1800, 1813 and 1816-1820. Bogota: low-skilled worker, average of 1800-1808. China: Beijing rural wages, average of 1807-1808, 1812-1813 and 1816-1820. Gdansk: unskilled worker, 1800-1812. Istanbul: unskilled construction worker, 1800, 1802, 1805, 1807, 1809, 1814 and 1817-1820. Kyoto: unskilled labourers, average of 1800-1820. Leipzig: building labourer, average of 1800-1820. London: building labourer, average of 1800-1820. Mexico: building labourer, average of 1800, 1802-1811, 1814 and 1816-1818. Milan: labourer, average of 1800-1804 and 1808-1820. Pennsylvania: labourer, average of 1800-1819. Porto: agricultural wage, average of 1800-1820. Potosi: unskilled workers, average of 1806, 1808, 1812 and 1813. Pune: average of Indian soldier in the British army and of domestic servant monthly wage divided into 30 days; average of 1805-1820. Strasbourg: labourer, average of 1801-1820. Sweden: labour, male agricultural, average of 1816-1820. Vienna: building labourer, average of 1800-1820.
Figure 2: Grain prices in Amsterdam/Holland, Antwerp/Belgium, Leipzig, London/Southern England, Milan, Porto, Strasbourg, Sweden and Vienna from the International Institute (http://www.iisg.nl/hpw/data.php). Prices of grain in Bogota, China, Istanbul, Korea, Pennsylvania Potosi and Pune from the web page of the Global Price and Income History Group (http://gpih.ucdavis.edu/Datafilelist.htm). Grain prices in Almadén from Dobado (Reference Dobado1989), in Kyoto from Bassino and Ma (Reference Bassino and Ma2005) and in Mexico from Garner (Reference Garner1985). Conversion rates of weight vs. volume units: wheat, 0.772 kilogram per litre; corn, 0.901 kilogram per litre; rice, 0.579 kilogram per litre (http://gpih.ucdavis.edu/Converting.htm). Except for Mexico, where the equivalence used is 55.5 litres per fanega following Florescano (Reference Florescano1986). New Spain reales converted into grams of silver a rate of 3.03 grams per real Burzio (1956-Reference Burzio1958). For nominal wages, see Figure 1. Grain prices in grams of silver per litre:
Almadén: 1800, 1802-1808, 1810-1814 and 1816-1820; grain (wheat) wage average of 1800, 1802-1808, 1810-1814 and 1816-1820. Antwerp/Belgium: 1800-1820; grain (wheat) wage is the average of 1800, 1813 and 1816-1820. Amsterdam/Holland: Utrecht/Groningen, 1800-1801 and 1804-1819; grain (wheat) wage is the average of 1800-1801 and 1804-1819. Bogota: 1801-1804; grain (corn) wage is the average of 1801-1804. China: Averages of 1801-1810 and of 1811-1820; grain (rice) wage is the average of 1807-1808, 1812-1813 and 1816-1820. Gdansk: 1800-1815; grain (wheat) wage is the average of 1800-1812. Istanbul: 1801, 1803-1805, 1807-1808, 1810, 1812-1816 and 1820; grain (wheat) wage is the average of 1805, 1807, 1814 and 1820. Korea: 1800-1817 and 1819-1820; grain wage (rice) is the average of 1800-1801, 1803, 1805, 1809, 1816 and 1819. Kyoto: 1800-1820; grain (rice) wage is the average of 1800-1820. Leipzig: 1800-1810; grain (wheat) wage is the average of 1800-1810. London/Southern England: 1800-1820; grain (wheat) wage is the average of 1800-1820. Mexico: 1800-1817; grain (corn) wage is the average of 1801, 1802-1811, 1814 and 1816-1817. Milan: 1800-1804 and 1808-1820; grain (wheat) wage is the average of 1800-1804 and 1808-1820. Pennsylvania: 1800-1820; grain (wheat) wage is the average of 1800-1820. Porto: 1800-1820; grain (wheat) wage is the average of 1800-1820; Potosi: 1806-1808, 1810 and 1811; grain (wheat) wage is the average of 1806, 1808 and 1812-1813. Pune: 1810/11-1820-1821; grain (rice) wage is the average of 1811-1820. Strasbourg: 1800-1820; grain (wheat) wage is the average of 1801-1820. Sweden: 1816-1820; grain (wheat) wage is the average of 1816-1820. Vienna: 1800-1820; grain (wheat) wage is the average of 1800-1820.
Figure 3: Meat prices and wages in Amsterdam/Holland, Antwerp/Belgium, Leipzig, London/Southern England, Istanbul, Milan, Strasbourg and Vienna from the web page of the International Institute of Social History (http://www.iisg.nl/hpw/data.php). Meat prices in Bogota and Pennsylvania from the web page of the Global Price and Income History Group (http://gpih.ucdavis.edu/Datafilelist.htm) and in Mexico from Quiroz (Reference Quiroz2005). For nominal wages, see Figure 1. Prices in grams of silver per kilo:
Amsterdam/Holland: 1800 and 1811-1820; meat wage is the average of 1800 and 1811-1820. Antwerp/Belgium: 1817-1819; meat (beef) wage is the average of 1817-1819. Bogota: 1801-1804, 1806 and 1808-1809; meat (beef) wage is the average of 1801-1804, 1806 and 1808. Gdansk: 1801-1815; meat (beef) wage is the average of 1801-1812. Istanbul: 1814-1820; meat (mutton) wage is the average of 1814 and 1817-1820. Leipzig: 1800-1810; meat (beef) wage is the average of 1800-1810. London/Southern England: 1800-1820; meat (beef) wage is the average of 1800-1820. Mexico: 1800-1805 and 1807-1811; meat (beef) wage is the average of 1800, 1802-1805 and 1807-1811. Milan: 1800-1804 and 1808-1820; meat (veal) wage is the average of 1800-1804 and 1808-1820. Pennsylvania: 1800-1820; meat (beef) wage is the average of 1800-1819. Strasbourg: 1800-1820; meat (beef) wage is the average of 1801-1820. Vienna: 1800-1820; meat (beef) wage is the average of 1800-1820.
Figure 5: Sources and methods for estimating grain wages as in Figure 2; GDP per capita in 1820 from Maddison’s web page (http://www.ggdc.net/maddison/). GDP per capita in 1990 international dollars for 1820 of the countries considered divided by the respective average grain wage of 1800-1820 in the towns, regions or countries for which data exit. Thus, Amsterdam/Holland grain wage is associated to GDP per capita of the Netherlands, Kyoto to Japan, Potosi to Bolivia, etc. GDP per capita of Poland in 1820 is the Eastern European average. GDP per capita of Bolivia and Colombia are the Latin American average.
Appendix 2: Sources and methods of physical statures estimates
Figure 4: Our estimation of the average height in Bourbon Mexico and Venezuela has followed the methodology suggested by Komlos (Reference Komlos2004). Thus, we first exclude all individuals whose age lies outside the range of 23-50 years, since they might either still continue growing — those under 23 years — or start to lose height because of aging — those over 50 years. Second, we have drawn the histograms using the original measures (see Dobado and García (Reference Dobado and García2009) for further details) in order to verify whether the height distributions of the military units approach a normal distribution, which it is needed for a proper selection of the truncation points. Third, we have applied the method proposed by Komlos and Kim (Reference Komlos and Kim1990) to estimate the average height of the whole distribution. Although Komlos and Kim’s method is intended to obtain the trend of series, it may also be used to estimate the average height of a population by assuming a constant standard deviation of 6.86 centimetres (Komlos and Kim Reference Komlos and Kim1990, p. 120). This method yields results similar to those of the RTML (restricted truncated maximum likelihood) by A’Hearn and Komlos (Reference A’Hearn and Komlos2003).
Sources of height are: The United States: Sokoloff and Villaflor (Reference Sokoloff and Villaflor1982); Great Britain: Floud et al. (Reference Floud, Wachter and Gregory1990), Komlos (Reference Komlos1993) and Cinnirella (Reference Cinnirella2008b); Sweden: Heintel et al. (Reference Heintel, Sandberg and Steckel1998), Bavaria: Baten (Reference Baten2001); Saxony: Cinnirella (Reference Cinnirella2008a); France: Komlos et al. (Reference Komlos, Hau and Bourguinat2003); Lombardy: A’Hearn (Reference A’Hearn2003); Austria — Hungary: Komlos (Reference Komlos1989); Russia: Mironov (Reference Mironov2005); Interior Spain: García Montero (2009); Bourbon Mexico and Venezuela: Archivo General de Simancas, Simancas, Valladolid, Spain, Secretaría del Despacho de la Guerra, files 7299-2, 7299-3, 7299-4, 7299-5, 7299-6, 7198-18, 7198-2, 7028-7, 7029-1, 7034-1, 7027-12, 7026-1, 7048-5, 6991-2, 7025-2, 7047-9, 7047-10, 7047-11, 7047-13, 7047-14, 7047-15, 7047-16, 7047-17, 7047-18, 7047-19, 7047-20, 7047-21, 7047-22, 7047-23, 7047-24, 7047-25, 7047-27, 7047-29, 7048-2, 7048-3, 7048-4, 7048-6, 7048-7.
Figure 6: GDP per capita in 1700 and 1820 from Maddison’s web page (http://www.ggdc.net/maddison/). Sources of height may be seen in Figure 4. For all countries in the sample, GDP per capita in 1700 and 1820 is divided by the average height of cohorts born in 1750-1760, except Interior Spain (1767-1770).
