This dramatic transformation in Argentina’s fortunes from Maddison (1988) to Maddison (2006) appears to have been due to the addition of the informal sector to the 1990 base year. Maddison (1995, p. 143) wrote:
‘A major problem with the national accounts of Latin American economies is the assessment of activity in the informal sector. Recent official revisions for Argentina have been very substantial. Instead of a GDP totalling 2 830 million australes in 1980, it is. now estimated to have been 3 840 million (nearly 36 per cent higher)’.
Based on this statement, the logical explanation of how Argentina in 1913 became so much richer from one Maddison to the other is that
toward the end of the twentieth century the Argentine government made an adjustment to its official GDP series to take into account the growth of the informal sector. To understand why an adjustment to Argentina’s late twentieth-century GDP would affect Maddison’s estimate for 1913, it is necessary to revisit the steps in his calculations.
1 They are:
- He begins with an estimate of nominal GDP in US dollars in his more or less recent base year.
- He adjusts that estimate for differences in the price level, in order to arrive at an estimate in international purchasing power parity (PPP) dollars.
- He extrapolates backwards from that recent PPP base-year estimate using a volume index of the country’s historical GDP.
Bearing these steps in mind, it seems likely that Maddison adjusted his Argentine GDP series by adding the informal sector to his base-year GDP estimate (step 1), adjusting by the same price level (step 2), then extrapolating back using the same volume index (step 3).
2
Crucially, then, Maddison’s historical GDP statistics will only be as reliable as his underlying figures for the base-year nominal GDP, the base-year price level, and the volume index. The problem with his estimates for Argentina is that the base-year nominal GDP has been adjusted for the growth of the informal sector,
but the volume index has not. Hence, he has extrapolated back from the adjusted base-year estimate
using an identical volume index. This is fatal for Maddison’s calculations
because the informal sector grew disproportionately, which is precisely why the government felt the need to adjust its official GDP estimates toward the end of the twentieth century.
Even as the ‘Argentine paradox’ literature was exploding the 1980s, Adrían Guissarri (1989, ch. 4) was already asking the question of whether his country’s apparent decline was actually an illusion generated by the disproportional growth of the informal sector. Using a methodology developed by Vito Tanzi (1983),
3 Guissarri estimated that the informal sector grew at a trend rate of 5.9 percent per year during 1930-85, whereas the formal sector grew at 3 percent per year; combined, the overall growth rate was 3.6 percent.
In Figure 1 Guissarri’s volume indices of GDP with and without the informal sector have been used to produce Maddison-style historical GDP estimates for 1930-85, this time shown relative to the United States. The result is striking: the unadjusted series produces the standard ‘Argentine paradox’ narrative, as Argentina’s GDP per capita fell from 60 percent the US level in 1930 to 37 percent in 1980; the adjusted series tells a quite different story, however, as
there is no decline over this period, with the Argentine level staying roughly constant vis-a-vis the US level.