Long-Term Land Use, Poverty Dynamics and Emission Trade-Offs (part of CRC 990)

Apl. Prof. Dr. Jann Lay
2012 - 2015
German Research Foundation, SFB 990 (DFG)
Cooperation partners

Prof. Stephan Klasen, Ph.D. and Rivayani Darmawan, M.Sc., Department of Economics, University of Göttingen

Prof. Marhawati Mappatoba, Department of Economics, Tadulaku University

Nunung Nurartono, Faculty of Economics and Management, Bogor Agricultural University



Research Questions

  • What are the long-term drivers of the poverty and vulnerability of rural households in transformed forested landscapes in Sulawesi and Jambi?
  • To what extent are particular income packages able to sustainably reduce the poverty and vulnerability of households?
  • What impact do production decisions (forest conversion and use intensification, crop and production choices, income portfolios, and technologies) as well as consumption decisions have on households’ GHG emissions?
  • How can we best understand the drivers of GHG emissions and the heterogeneity among households to identify trade-offs and win-win situations between poverty reduction and emission reductions?

Contribution to International Research
The transformation of forested landscapes, poverty reduction, and GHG emissions are closelyinterrelated. Understanding the drivers of land-use change is key to understanding these interrelationships. However, due to shortcomings in data availability and methodological approaches, many questions remain open. Few studies have been able to analyse the dynamics of resource use in the medium to long term as this requires panel data and associated econometric techniques that control for household-specific effects, endogeneity, and initial conditions. In addition, the link between the dynamics of land and forest use and vulnerability has not been sufficiently examined. This vulnerability is also likely to differ by region and according to the crop type and the associated technologies. Therefore, this study’s comparative perspective, which assesses these issues using data from smallholder cocoa farmers in Sulawesi and rubber and palm oil producers in Jambi, will be particularly useful, as will further comparisons with national cross-sectional and panel data that will assess the relevance of the findings from the study areas at a higher level of aggregation.

Research Design and Methods
To analyse the drivers of long-term poverty and vulnerability in transformed forested landscapes, this project will combine long-term panel analyses, comparative surveys from two study regions, and national data. An existing panel survey was already extended by a fourth wave of the income and expenditure panel surveys in Sulawesi beginning of 2013. Using this panel survey, we will assess these long-term drivers applying advanced panel econometric methods. In a second group of analyses, the researchers will assess the drivers of (static) income poverty using an explicitly comparative framework. This step will draw on the data from Sulawesi, the household surveys generated by subprojects of CRC 990 in Jambi, and national cross-section (SUSENAS) and panel (RAND) household surveys.

Preliminary Results
In an empirical study on land-use change at rainforest frontiers in Sulawesi Indonesia, we analysed the drivers of land-use change with a specific focus on their impact on economic development. Using a three wave household panel survey from 2001 to 2013 we applied household fixed effects to estimate first the micro level drivers of households’ land use patterns, and second the determinants of household’s’ income portfolios. Preliminary results suggest that land-use change in the Lore Lindu region is indeed determined by price signals. Households adopting the relevant cash crops within the region are significantly better-off comparing to households cultivating staple crops. Further we found that more land is put to cash crop cultivation in regions with higher immigration.

Research Programme


For further information about the project, please visit the website of the University of Göttingen