Page 76 - FAO-IPCC Expert meeting on climate change

Appendix 01: Speakers’ summary notes
rice production (hand-tractors), which freed up labour, and (ii) the increased profitability provided funds for investing in cocoa production in the uplands.
Empirical evidence
An early synthesis of case study evidence by Angelsen and Kaimowitz (2001a) provided mixed evidence on the impact of improved technologies on deforestation. The forest impact varies depending on the factors discussed above. They conclude in the following way: “The basic Borlaug hypothesis – that we must increase agricultural yields to meet growing global food demand if we want to avoid further encroachment by agriculture – still holds. Still, that by no means guarantees that specific agricultural technologies that farmers adopt will help conserve forests. The current trend towards more global product, capital and labour markets has probably heightened the potential dangers. Technologies that make agriculture on the forest frontier more profitable and that displace labour present particularly strong risks, while technologies that improve the productivity of traditional agricultural regions and are highly labour-intensive show the most promise” (Angelsen and Kaimowitz, 2001b, 402).
A comprehensive review by Villoria et al. (2014) reach similar conclusions, largely confirming theoretical results, but noting that data and empirical results are lagging behind our theoretical insights. Technological progress at the global level is likely to take pressure off forests, yet low-yield, land-abundant regions are likely to experience further land expansion. Globalization has improved markets access and technology transfer and diffusion. These processes can drive deforestation in new frontiers, exemplified by new soybean and cattle expansion in Southern Africa (Gasparri and Waroux, 2015).
Byerlee et al. (2014) suggests that there is a crucial difference between technology driven and market driven intensification; the former has generally reduced cropland use and deforestation, while the latter has been a major case of agricultural expansion and deforestation. Their definitions are important: “technology-driven intensification occurs when technical change in a crop allows for more output of land for the same level of input”, while market-driven intensification “results from a shift in product mix to higher value crops due to new market opportunities, or a shift in input mix in response to relative price changes” (page 93). Technological change in the presence of favourable market conditions (growing demand and rising prices) does provide strong incentives for land expansion, as exemplified by a series of commodity booms throughout history (e.g., Ruf, 2001).
National and global level analyses are either done by econometric or by simulation models. Ewers et al. (2009) use country-level data for the period 1980–2000 to test whether increases in agricultural yield have serendipitously spared land for nature. If “perfect land-sparing” yield change were occurring (as in the simple global food equation), the land-yield elasticity should be −1. They find a much lower elasticity: −0.152 (t = −1.78) for developing and −0.089
(t = −0.57) for developed countries. Similarly, Rudel et al. (2009) find that “agricultural intensification was not generally accompanied by decline or stasis in cropland area at a national scale during this time period [1970-2005], except in countries with grain imports and conservation set-aside programs” (page 20675).
A central effort among simulation models has been the GTAP (Global Trade Analysis Project) (Baldos and Hertel, 2012b) and the SIMPLE model (Baldos and Hertel, 2012a), and the GLOBIOM (Havlík et al., 2014) . Hertel (2012) gives a theoretical description and numerical illustration, and several more specific applications have been implemented. Stevenson et al. (2013) estimates that Green Revolution research saved 18-27 million ha from being brought into agricultural production (and a significant share of this gain being forest). Their simulation results are, however, “order of magnitude lower than predicted by the simple global food equation that does not take account of feedback
loops through prices of products, consumption demand, and land-use decisions” (page 8365). Villoria et al. (2013) investigated the impact of yield increases in oil palm production, either in only the two dominant producers Indonesia and Malaysia, or globally. If only Indonesia and Malaysia experience technological progress, they observe a modest effect in terms of area expansion locally, but the opposite in other regions. If the change is global, emissions from deforestation are reduced both locally and globally.
Keeping track of indirect land use change (ILUC) is inherently difficult; econometric models cannot capture many of the effects, and simulation models are based on strong assumptions. A causal analysis framework for land use change (Efroymson et al., 2016), whose core is a strength-of-evidence approach, could provide a fruitful route to integrate the multiple sources of evidence that exist in the form of case studies, statistical analysis and simulation models.
The forest outcomes of technological progress in agriculture can be mixed, but the likelihood of win-win outcomes can be enhanced through policies. “Technology-driven intensification by itself is unlikely to arrest deforestation
FAO-IPCC Expert meeting on climate change, land use and food security