Greenhouse gas mitigation by agriculture and livestock intensification in Brazil

1. Greenhouse gas mitigation by agriculture and livestock intensification in Brazil Joaquim Bento de Souza Ferreira Filho University of São Paulo, Brazil. Mark Horridge Centre Of Policy Studies, Monash University, Australia. Jonathan Gonçalves da Silva University of São Paulo, Brazil.

2. Background • Brazilian agriculture has been growing at average year rates of 2.1% since 1990: • Area of crops: from 53 Mha in 1990 to 69.1 Mha in 2012 • Pastures area: 160 Mha in 2006. • This increase generated concerns about deforestation: about 47.8% of all humid tropical forest clearing between 2000 to 2004 occurred in Brazil (Hansen et al. 2008). • Agriculture and change in use of land and forests accounted for about 58% of total GHG emissions in Brazil in 2005

3. Annual crops area (1,000 ha) and rate of deforestation (1,000ha/year) evolution in Brazil

4. Background • Productivity gains also important in agricultural growth

5. Fall in deforestation in Brazil • Effects on GHG emissions: 38.7% fall in total emissions in 2005-2010, mostly due to fall in deforestation. • Brazil is close to meet the climate goals of Copenhagen (2009): emissions could increase by 66% from 2010 levels and still hit its 2020 target. • As seen before, the fall in deforestation has no apparent effect on agricultural supply (so far) in Brazil. • However, the land expansion effect will be limited by this process in the future. • How important are productivity gains in the process?

6. Objective • Analyze the potential for land sparing and GHG mitigation in Brazil through intensification of agriculture and livestock production, in order to meet future scenarios of food supply. • Productivity increases in livestockandagriculture: twooppositeffects in theliterature: • “The “Borlaug hypothesis”: conditions under which the intensification of agriculture and livestock production in Brazil would allow a significant reduction in deforestation rates. • The “JevonsParadox”: relates increases in productivity to expansion in land areas.

7. Methodology: CGE model calibrated for year 2005 • Annual recursive dynamic, inter-regional, bottom-up: • Stock-flow relation between investment and capital stock, which assumes a 1-year gestation lag; • Positive relation between investment and the rate of profit; • Relation between wage growth and regional labor supply. • 15 aggregated regions inside Brazil • 38 production sectors. • 10 types of workers (wage classes) • 10 household types (income classes) • ILUC module: tracks land use change by state and by biome.

8. Land use module • Four land types: Crops, Pastures, Forestry, Unused. • Crop land fixed by state in each year: relocation among industries by CET rule. • Transitionmatrix drives land use betweenyears • PRIORS: Natural forest->pastures->crops • TRANSITIONS AND EMISSIONS: themodeltracks GHG emissions in transitionfrom natural foreststoother uses. • TRANSITIONS: from 6 differentbiomes, with particular emissions rates ontransitions.

9. Land use transitions module: GHG emissions • The model captures theemissionsassociatedtotransitionsfromsixbiomes: • Amazonia • Cerrado • Caatinga • Atlantic Forest • Pampa • Pantanal • Emissionscalibratedfromsatelliteimagery for the1994-2002 period.

10. Natural forests: 6 biomes Cerrado Caatinga Atl. Forest Pampa Amazonia Pantanal Observedfromsatelliteimagery EMISSIONS Transitionmatrix in land uses Crops Pastures Elevencrops

11. Emissionsassociatedwithtransitions in Brazil. Gg CO2 eq. 1994-2002 • For example, the total amount of land under the "Unused" category (native vegetation, parks, reserves and other areas) in 1994 converted into pastures generates in the final period the amount of 9,388,356 Gg net CO2eq, aggregated over all biomes

12. Total emissionsin transitionstoCropsandLivestock in Brazil, bybiome. Gg, 1994-2002. Observedvalues.

13. Baseline: moderategrowth, followstrend. % change, accumulated in 2030.

14. Simulation: 2015-2030 Frontierstates • Policyshocks: • Livestock: 1% per yearproductivityincreseabovetrend for fiveyears (2015-2019) in theagriculturalfrontierstates. • Agriculture: 0.5% increaseabovetrend in thesameperiod

15. Table 4. Agricultural output and land area variation, cumulative percent change 2005-2030 (base scenario)

16. Emissions in land use transitions, bybiome. Gg of CO2 equivalents. Baseline, accumulated.

17. Policy/Base deviations, macro variables: total growth 2005-2030 and terminal annualgrowth rates. • Greatercontribution of agricultureproductivitygrowthto GDP increase.Agriculture: 2.7% of Total productionvalue, Livestock 0.7% of TPV.

18. Intensification of livestockandagriculture: simulationresults • Crops area would be 240,000 ha less: • 270,000 ha increase due to the livestock shock and a 500,000 ha reduction due to the agriculture shock • Pastures area would be reduced by 170,000ha in 2030 due to the shocks Borlaug hypothesis?

19. Productivity shocks • Livestock: • Transition matrix commands a large transfer of pastures to agriculture, allowing agriculture to expand. • Agriculture: • The land release effect for livestock is relatively smaller, since the observed transition from crops back to livestock is smaller than the other way out, a stylized fact in land use transitions in Brazil brought into the model by the transition matrix

20. Resultssupport • The “Borlaug hypothesis” for productivityincreases in livestock. • The “Jevonsparadox” for productivityincreases in agriculture. • The assimetrictransition in land use givesupporttodifferenteffects of productivityincreases, dependingonthe sector where it appears.

21. Model results. Land use variation due to the productivity shocks. Policy scenario, % change, accumulated.

22. Modelresults. Land use variation, bybiome in Brazil. Mha. Accumulated in 2030

23. Total emissions and emissions by source in the initial and terminal period.

24. Final remarks • Mixedevidence in the Borlaug x Jevons debate. • Borlaug hypothesis: livestock • Jevonsparadox: agriculture. • Focus onlivestockproductivityincreases for land sparing. • Amazonbiomewouldbenefitthemost, followedby Cerrado and Caatinga. • Productivity gains: are permanent, andremainevenaftertheshockends.

25. Thank you. • Email: jbsferre@usp.br