This is a translation-reblog of the post by historian Javier Rodríguez Weber, “Globalisation and Inequality, for a ‘sophisticated’ version of the neoclassical intepretation” (original: “Globalización y Desigualdad. Por una versión ‘sofisticada’ de la interpretación neoclásica”). It shows how Chile’s income distribution in the period 1880-1903 was affected by the combination of international trade, the opening of frontier lands, and (the relative lack of) immigration. There are also interesting comparisons with Australia and New Zealand.
I like this post, but I do not endorse everything in it, and there are some comments which I find a little odd. See note at the end.
Below, anything in brackets [ ] is something I’ve added. I’ve also added hyperlinks to the references mentioned by JRW.
[Begin JRW post]
Resource-intensive economic growth in Chile and other countries of the “periphery” which occurred during the First Globalisation [1870-1914] has been associated with increasing income inequality (O’Rourke and Williamson 1999; Williamson 1998, 2011). According to the Heckscher-Ohlin-Samuelson framework adopted in these works, the greater [international] demand for natural resources caused these economies to integrate with the global market, increasing the returns to the factor they possessed in abundance. In the case of Latin America, this factor was land. Thus, globalisation enhanced the income of the landlord class and thereby exacerbated inequality.
While the substantive contributions of this research programme are not in doubt, nonetheless its reduction of globalisation to basically factor price equalisation is excessively simplified. Therefore, some years ago Harley (2007) argued for a reorientation of research on the First Globalisation, returning to the fore the term “frontier expansion”. In his view, this would address crucial aspects of the process that are outside the neoclassical paradigm, such as the role of technological progress, institutional changes, or the centrality of power relations in the appropriation of natural resources. But even some aspects of the relationship between globalisation, frontier expansion, and inequality which are well suited to the neoclassical approach, have not been addressed by this literature. I refer to the dynamic evolution of factor endowments and their regional dimension. The exploitation of relatively uninhabited areas under Atlantic capitalism requires the movement of labour to these regions. When the dynamic effects of new land on factor endowments are taken into account, its effects on income and inequality may be different than expected.
In the late 1870s Chile embarked on a war with Peru and Bolivia [the War of the Pacific ] over the latter’s taxation of Chilean companies operating in the nitrate region of Antofagasta, then part of Bolivian territory. At the same time, an indigenous uprising broke out in Araucanía, a nominally Chilean territory which was outside of effective state control. Chile came out the winner in both conflicts, with a territorial gain of around a third of its area prior to 1880. This territorial expansion entailed the integration of previously underutilised natural resources in the north; or, in the south, [land that had been] outside the capitalist mode of production. As a result, between 1880 and 1913 Chile went through a period of economic growth known as “el boom” or the “nitrate cycle”.
Table 1 provides a rough estimate of the change in relative factor endowments in Chile due to the expansion of the frontier. The incorporation in the first half of the 1880s of the territories that later formed the provinces of Antofagasta, Tacna, Tarapaca, Malleco, and Cautín, increased the availability of natural resources per capita. From a merely quantitative point of view, the land-labour ratio as measured by the surface area per inhabitant increased by almost a third after 1880. But the new territories were especially rich in natural resources, so that 29% should be taken as the lower bound. Any estimate that would account for the quality of the resources would raise the estimate. That is, natural resources became more abundant relative to labour.
And what were the consequences for incomes and income inequality ?
Economic growth based on frontier expansion resulted in greater demand for labour and an increase in real wages (Matus 2012). Some of this increase in demand came from the mining sector, where relatively high wages were paid. There were also new opportunities for the population of migrant workers in the Central Valley. They went north in search of higher wages; they went south in expectation of new lands; and they went to the cities fed by the mining boom. The emigration of workers out of the centre of the country, peaking around 1895, altered the land-labour ratio in the region, making labour more scarce and raising its compensation. As shown in Figure 1, between 1880 and 1903, the annual income of rural workers not only rose, but increased from 37% to 64% of unskilled workers in the city.
By growing faster than average income, the wages of unskilled workers contributed to the improvement in Chile’s income distribution. The size of this effect is approximated by two counterfactual exercises, whose results are presented in Figure 2. The actual trend of the Gini index is compared with two simulations based on different assumptions about the hypothetical evolution of unskilled wages. In one counterfactual, real wages were assumed to be stable, i.e., nominal wages grew at the same rate as the consumer price index. In this case, inequality remains almost unchanged in 1880-1903. In the second counterfactual, unskilled wages were assumed to grow at the same rate as average income. This produces a smaller increase in real wages, and a smaller improvement in income distribution, than what actually occurred.
What sets the Chilean experience apart from others, is who occupied and exploited the frontier lands. With few exceptions, they did not come from abroad as immigrants, but [as internal migrants] from the central region of Chile. In countries like Australia, New Zealand, Argentina, and Uruguay, the integration of natural resources into the economy was paralleled by an immigration process that radically changed the structure and demographic characteristics of the host societies. There was thus a “race” between resource supply and labour supply. The reason the opening of the frontier did not lead to lower income inequality, as happened in Chile during this period, is that labour supply won the “race” and grew faster than the supply of land. Therefore, although territorial expansion substantially increased the absolute supply of natural resources, available land per capita actually fell (Table 3). This is the crucial difference with the Chilean case (Table 1).
In the Australian colonies of South Australia and Victoria it’s possible to follow more closely the regional dynamics of changes in factor endowments and factor payments (Table 4). In Victoria, which was already well populated by 1870, rising population made land more scarce relative to labour, and the falling wage-rent ratio shows the classic pattern associated with higher inequality. In the frontier region of South Australia, the exploitation of [new] land grew in tandem with population. So there occurred once again that race between land and labour, and the relative factor endowments as well as the wage-rent ratio change over time.
Most research of neoclassical inspiration has noted that those regions which specialised in exporting natural resources during the First Globalisation should show a deterioration in income distribution. But such studies have ignored the dynamic and regional aspects of this process. Once you consider frontier expansion, distributional outcomes may differ from straightforward derivations of the Heckscher-Ohlin-Samuelson approach. But if one adopted a more sophisticated version taking into account changing factor endowments — inspired by Rybczynski (1955) — the neoclassical approach is still useful for understanding what happened to inequality over the period. Thus, if frontier expansion diminished inequality in Chile, but not in Argentina and New Zealand, this was because in the latter cases it was accompanied by a massive immigration that made land relatively scarce. Chile, however, hardly received immigrants; the new lands were occupied by [internal migrants]; and therefore, the changes in relative factor endowments went in the opposite direction. In Australia, the situation was more complex, because the evolution of factor endowments and factor payments are different according to region and period.
But even this “sophisticated” version of the neoclassical approach is not entirely sufficient. It’s necessary to understand why Chile did not receive as many immigrants as Argentina or Uruguay. Here institutional factors must come into play. Part of the explanation, also consistent with the neoclassical approach, is that despite its growth from the late nineteenth century, wages remained lower in Chile than in Argentina or Uruguay (Matus 2013: 438). This was probably because, at least in part, there was never as much a shortage of labour in Chile as there was in other countries when they began receiving immigrants. But there were also institutional factors which, even if they did not impede immigration, did not favour it either.
The absorption of the frontier constituted in all cases a massive violation of the rights of the original inhabitants — an example showing that, contrary to the assertions of the universal claim of the neoinstitutionalist literature, the violation of property rights can favour economic growth. And institutions and existing power relations have been key to determining who gets the new territories. As regards Chile, the opening of the frontier reduced inequality, among other things, through the uncertainty over property rights that prevailed in the region for some years. This uncertainty was due in part to a struggle between those who wanted to distribute properties of [equal] size to settlers and those who proposed auctioning off the southern lands (after seizure from the original [indigenous] occupants) to the highest bidder. In the meanwhile, the southern territories were occupied by thousands of “spontaneous settlers”. By leaving the Central Valley, these migrants reduced labour supply in the traditionally settled areas (as well as the seat of power) and thus raised the wages of those who stayed.
Ironically, large estates [latifundias] eventually spread to the south; restrictions were placed on [internal] migration; and, after 1900, there was a rapid deterioration of inequality. But that’s another story (or rather another post).
PD 2: Apologies for pasting images taken from a pdf for charts and tables, but it was the fastest solution Hand
[End JRW post]
Chapter 6 of JRW’s thesis has much more detail about the above, including the development of Chile’s state capacity, both in terms of its military capability and frontier absoprtion. I almost wish James Belich had written a wider study of settler societies to include the Southern Cone.
Chapter 7 details the post-1903 reversal of the trend described above. It’s full of interesting institutional details about the period. JRW should really have his whole thesis translated into English !
See Jeffrey Willamson’s latest, “Latin American Inequality: Colonial Origins, Commodity Booms, or a Missed 20th Century Leveling?“.
A few cavils:
First, JRW’s comments about the Hecksher-Ohlin model are a little strange. HO is a very flexible general approach toward international trade and its implications for the pattern of production and the distribution of factor payments. To say Rybczynski (1955) is a more “sophisticated” version might have been plausible in the 1950s, but today it is presented as a natural extension, even a tweak, to the basic theory. The Rybczynski Effect is at least alluded to even in undergraduate textbooks on international economics. For example, Krugman & Obstfeld (9th edition) certainly mentions the effect of biased changes in resource endowments. I also see it in the ancient 4th edition of K-O, where an increase in the supply of land is explicitly illustrated (pg 74-75). Moreover, the edited volume, Eli Heckscher, International Trade, and Economic History, contains a chapter by Findlay and Lundahl explicitly modelling frontier expansion (“Demographic Shocks and the Factor Proportions Model: From the Plague of Justinian to the Black Death”). And there are other applications of the HO framework to economic history.
Second, it’s a little strange to hear in 2015 that the “role of technological progress, institutional changes, or the centrality of power relations” are outside the neoclassical paradigm. Perhaps JRW construes the phrase narrowly as the analysis of “purely economic” relations that keep politics, institutions, technology, etc. exogenous. But, really, hasn’t neoclassical economics now endogenised pretty much everything in theory ? And of course neo-institutionalism and public choice theory are part of the neoclassical paradigm. And, finally, the seizure of indigenous land does not contradict the neo-institutionalist emphasis on property rights.