Was Argentina Really Better Off Than the United States in 1800?

Argentina’s economic history provides yet another example of the problem of Mickey Mouse numbers.

When a prominent economic historian provides a new estimate of something, it is likely that the estimate will be taken at face value. Other economic historians will cite it, so it becomes reified, until it is treated as fact, even when it is little more than fancy. John Coatsworth’s estimate of Argentina’s GDP in 1800 provides an example of this.

Coatsworth (1998, p. 26, Table 1.1) claims that Argentina’s GDP per capita in 1800 was 2 per cent higher than the United States’ GDP per capita – a finding that other economic historians have repeated (for a recent example, Irigoin 2015, p. 6). Nevertheless, when Coatsworth’s original article is examined, it is impossible to see how he arrived at his figure. What is more, a major error can be clearly seen in his methodology.

Coatsworth provides two nuggets of methodological detail on how he constructed his GDP estimate for Argentina in 1800. First, in an Appendix he states:

The Argentine […] figures are based on more fragmentary evidence. [T]he GDP figures in the table are really income estimates based on extrapolations from wage data. […] Lyman Johnson’s Buenos Aires study cites an average monthly wage of 17 pesos or 204 pesos per year for urban unskilled construction laborers in the first decade of the nineteenth century, while various sources put rural wage levels at 6 pesos per month plus food rations, for a total of 76.5 pesos per year. This implies a per capita income of roughly 94 pesos for the province of Buenos Aires; using the same wage rates for the remaining provinces brings down the colonywide per capita income to 82 pesos. (Coatsworth 1998, p. 45)

Then, in a footnote, he adds:

For the predominantly urban population of Buenos Aires province, this figure multiplies the urban wage rate ($204) times the Buenos Aires urban labor force of 25,600 (assumed to be 64% of the Buenos Aires population of 40,000) and applies the rural wage rate ($76.50) to the rural labor force of32,168 (64% of the remaining population of 50,262). According to Maeder’s account (1969, chap. 1), roughly a third of the total Argentine population (including the province of Buenos Aires) lived in towns and urban areas. Again assuming a labor force participation rate of 64% for both rural and urban populations, the same wage rates yield a per capita income estimate of $81.50. This colonywide estimate excludes the Chaco, Misiones, and the areas of the pampas and Patagonia outside European control. The per capita income of the colony without Buenos Aires province comes to $69. (Coatsworth 1998, p. 49, fn. 30)

Coatsworth provides, then, more data for Buenos Aires (population, labour force, and wage rates) than for the rest of the country. Yet, it is impossible to replicate his income estimate even the former, as will be demonstrated.

In Table 1 the numbers in bold are taken from Coatsworth’s passages reproduced above, while the other numbers are calculations. As can be seen, the numbers provided by Coatsworth lead to a per capita income of 85 pesos in Buenos Aires and it is impossible to see how he arrived at a figure of 94 pesos. His estimate for Buenos Aires is therefore inexplicable. As for the rest of the country, Coatsworth does not supply enough methodological detail to even attempt to replicate his findings.

Table 1: Coatsworth’s data for Buenos Aires

Urban Rural Total
Population 40,000 50,262 90,262
Labour forcea 25,600 32,168 57,768
Wage rates (per year) 204 76.5 133
Income (total)b 5,222,400 2,460,828 7,683,228
Income (per capita)c 131 49 85

64 per cent of the population.
b Labour force multiplied by the wage rate.
c Total income divided by the population.

More worryingly still, a crucial part of Coatsworth’s underlying data appears to be obviously incorrect. In the article cited by Coatsworth, Lyman Johnson (1990a) never provides an estimate of the annual income of an unskilled labourer in late colonial Buenos Aires. Moreover, in another of Johnson’s (1990b) essays that was published in the same year there was a table, reproduced below, that suggests a far lower wage rate of just 61-70 pesos for urban unskilled labourers.

Johnson_tableSource: Johnson 1990b, p. 145, Cuadro 2.

If Johnson’s 61-70 pesos are used instead of Coatsworth’s 204 pesos, it makes a major difference to the overall income estimate. This is shown in Table 2, where rather than the 85 pesos suggested by Coatsworth’s data, Johnson’s numbers lead to an overall per capita income of at most 47 pesos.

Table 2: Coatsworth’s corrected data for Buenos Aires

Urban Rural Total
Population 40,000 50,262 90,262
Labour forcea 25,600 32,168 57,768
Wage rates (per year) 70 76.5 74
Income (total)b 1,792,000 2,460,828 4,252,828
Income (per capita)c 45 49 47

a 64 per cent of the population.
b Labour force multiplied by the wage rate.
c Total income divided by the population.

Coatsworth’s GDP estimate for Argentina in 1800 thus appears to be completely insubstantial. Not only is it impossible to replicate his findings using the data that he provides, but a crucial piece of that data is refuted by the scholar who Coatsworth cites as its source. That economic historians are willing to repeat Coatsworth’s estimate as if it were fact unfortunately says much about their craft.

 

References

Coatsworth, J.H. (1998) ‘Economic and Institutional Trajectories in Nineteenth-Century Latin America’, in idem and A.M. Taylor, eds. (1998) Latin America and the World Economy since 1800, Cambridge, MA, pp. 23-54.

Irigoin, A. (2015) ‘Representation Without Taxation, Taxation Without Consent: The Legacy of Spanish Colonialism in America’, Economic History Working Papers No 227, LSE.

Johnson, L.M. (1990b) ‘The Price History of Buenos Aires During the Viceregal Period’, in L.L. Johnson and E. Tandeter, eds., Essays on the Price History of Latin America, Albuquerque, 1990, pp. 137-71.

_____, ‘Salarios, precios y costo de vida en el Buenos Aires colonial tardio’, Boletín del Instituto de Historía Argentina y Americana ‘Dr. E. Ravignani’, 2:7, pp. 133-157.

Maeder, E. (1969) Evolución demográfica argentina de 1810 a 1869, Buenos Aires.

Facebooktwittergoogle_plusredditpinterestlinkedinmail

2 thoughts on “Was Argentina Really Better Off Than the United States in 1800?

  1. Peter Sims

    Do we really know so little about wages in late colonial Buenos Aires? Surely all the various estimates don’t all just mis-cite Lyman Johnson and call it a day? I’ll have to look into this more closely, but I find the implications here a bit implausible – I just don’t think Buenos Aires was that poor. Not as rich as the US? Maybe not. But surely not as poor as you are suggesting. (Though, in fairness, my view is shaped by the extant literature, almost all of which suggests a relatively high wage BA.)

    What is the difference among the last 3 categories in Johnson’s table? They first column is obviously skilled labour, but the second is just a day’s wage, no? (Jornal?) Who receives 10 reales, 4, and 2, and for what work? None of these figures correspond directly to the figures Coatsworth cites, and I do not know the derivation of it (though I certainly can’t say it’s wrong without knowing its provenance…) Surely it can’t be the case that the vast majority of workers were in general all earning 2 reales a day, unless the skill premium in Buenos Aires was absurdly high, at around 6x the wage of an unskilled worker. The 2 real figure also means that wages were higher in the countryside than the city, at least if the rural wage estimate used by Coatsworth still stands. Is this plausible? It seems very unlikely to me.

    Reply
    1. Joe Francis Post author

      Johnson (1990b) provides an abundance of wage data, but it has been largely ignored. The recent Arroyo Abad et al paper, for instance, cites Coatsworth as its source and uses his wage rate of 204 pesos per year. Coatsworth ascribes that number to Johnson, but it never appears in Johnson’s work – indeed, Johnson’s work directly contradicts it.

      The figure of 2 reals per day for an unskilled urban labour is plausible. Coatsworth (1998, p. 46) himself notes that in Chile ‘fragmentary data suggest urban unskilled wage levels at about two reales (0.25 pesos) or so per day’ – why would it be so much higher in Buenos Aires?

      Johnson gives the typical rural wage at roughly 6.5 pesos per month, which was equivalent to around 1.7 reals per day, although the rural workers were also given food and possibly board.

      Reply

Leave a Reply

Your email address will not be published. Required fields are marked *