We estimate the impact of the Green Revolution in the developing world by exploiting exogenous heterogeneity in the timing and extent of the benefits derived from high-yielding crop varieties (HYVs).
We find that HYVs increased yields by 44% 1965–45201014ya, with further gains coming through reallocation of inputs. Higher yields increased income and reduced population growth. A 10-year delay of the Green Revolution would in 2010 have cost 17% of GDP (gross domestic product) per capita and added 223 million people to the developing-world population. The cumulative GDP loss over 45 years would have been US$118.37$832010 trillion, corresponding to ~1 year of current global GDP.
…The IARCs targeted developing countries, so all European countries, all former Soviet republics, Australia, Canada, Israel, Japan, New Zealand, and the United States are excluded from the sample…Our shift-share variable indicates that HYVs increased yields of food crops by 44% 1965–45201014ya. The total effect on yields is even higher because of substitution toward crops for which HYVs were available and because of reallocation of land and labor. Beyond agriculture, our baseline estimates show strong, positive, and robust impacts of the Green Revolution on different measures of economic development. Most striking is the impact on GDP (gross domestic product) per capita. Our estimates imply that delaying the Green Revolution for 10 years would have reduced GDP per capita in 2010 by US$1,815.52$1,2732010 (adjusted for PPP [purchasing power parity]), or 17%, across our full sample of countries. The dollar amount is large, in part because some of the countries grew relatively rich during the period we study: the comparable loss in today’s least developed countries is US$559.06$3922010. By 2010, the cumulative global loss of GDP of delaying the Green Revolution 10 years would have been about US$118.37$832010 trillion—roughly a year of present-day global GDP. Needless to say, this surpasses the amount of resources that went into developing HYVs by several orders of magnitude. The income loss would have been much greater had the Green Revolution never happened, perhaps reducing GDP per capita in the developing world to 50% of its current level, if our estimates are taken at face value—although we stress that this number is subject to considerable uncertainty and depends on a somewhat implausible counterfactual. Despite these reservations, the results of this paper clearly place the Green Revolution among the most important economic events in the 20th century.
We find no evidence that the gains from increased agricultural productivity were offset by any Malthusian effects; the increased availability of food does not appear to have been eroded by population increases. Instead, we find a negative effect of the Green Revolution on fertility. Our estimates suggest that the world would have contained more than 200 million additional people in 2010 if the onset of the Green Revolution had been delayed for 10 years. Lower population growth increased the relative size of the working-age population, leading to a demographic dividend that accounts for roughly one-fifth of our estimated effect on GDP per capita. Our paper also sheds light on a concern, often expressed in the literature, that agricultural productivity improvements would pull additional land into agriculture at the expense of forests and other environmentally valuable land uses. We find evidence to the contrary: in keeping with the “Borlaug hypothesis”, the Green Revolution tended to reduce the amount of land devoted to agriculture.
…The start of the Green Revolution can be dated quite precisely. As noted above, the first high-yielding rice varieties were crossed in 1962 at IRRI, and after several generations of selection, they were initially released in 1965 to national research programs in rice-growing countries around the world. For wheat, it is similarly possible to identify a zero date for the Green Revolution: the first successful crosses from the Rockefeller wheat program took place in the 1950s, but they were not released to farmers in other developing countries until 1965. Maize followed soon after. For each crop, we can identify with reasonable precision the date at which the research institution first released a variety based on breeding work that took place within the institution.
The Mexican case is unique in the sense that the first HYVs were developed in a research program that did not yet have standing as an international institution. As a result, the diffusion of the wheat semi-dwarf varieties took place within Mexico slightly before the varieties became available in other countries. For all our other crops, HYVs developed at the international research centers became available to all countries at effectively the same moment—either upon a formal initial release from the international center or through the inclusion of the material in “nurseries” of promising experimental material that were shared with researchers across the developing world.
…Converting our estimates from logarithms to levels, we find that relative yields are on average 9% higher 10 years after a HYV release (β10 = 0.09) and 75% higher after 40 years (β40 = 0.56). The gradual increase in yields happens both because adoption is gradual, along an extensive margin, and because successive vintages of HYVs of a crop increase yields beyond what the first HYV could achieve. Our estimated magnitudes are consistent with the micro-level literature, surveyed in 2003b [Crop Variety Improvement and Its Effect on Productivity], which shows that HYVs typically have at least 50% higher yields than traditional varieties for a given set of inputs. Inputs are not fixed, however. Many HYVs respond better to fertilizer and other inputs than traditional varieties, raising yields still further; gains of the magnitude observed in Figure 2 are not unexpected, in cases when HYV adoption is widespread.
…The event study for GDP per capita in Figure 4A shows that 10 years after the onset of the Green Revolution in 1965, countries specialized in wheat, rice, and maize begin to have faster income growth than other countries.
…To put our estimated effect sizes into perspective, the effect of delaying the Green Revolution by 10 years is of a magnitude comparable (with opposite sign) to the income effect of democratizing, which et al 2019 estimate to be about 20% after 25 years, and to the effect of railroad access in 19th-century India, which 2018 puts at 16%. The population effect we find is substantially smaller than the effect of medical innovations, which, according to et al 2020, has increased the population by 45% 1940–40198044ya in their sample of countries and by even more in low-income and middle-income countries.
…The Green Revolution is often associated with the 1960s and 1970s, but rather than slowing down, the rate of adoption and the number of new HYVs increased in the 1980s, 1990s, and 2000s. Scattered evidence from sub-Saharan Africa suggests that the HYV adoption rate has increased by as much in the 2000s as in the 4 preceding decades.27 One reason is that, compared to that in other parts of the world, especially Southeast Asia, African agriculture is specialized in cassava, sorghum, millet, and other crops for which HYVs became available relatively late. Our results consequently shed light on the divergence between Southeast Asia and Africa during the second half of the 20th century.