1. Farm Typologies in Tanzania and Malawi Africa RISING (@ IFPRI) ESA Africa RISING writeshop, Dar es Salaam July 1, 2016

2. Why do we need typologies? Identify suitable farms to target innovations (ex-ante): not all innovations are suitable for all farms, and classifying farmers into groups would support the identification of technology-specific suitable farming systems. Scale out innovations: given population heterogeneity we can formulate extension messages, policies and incentive schemes to further spread the use of designed innovations. Assess agro-economic effects (ex-post): Explaining trends and farmer ‘behavior’ (functional characteristics, including SI indicators) and assessment of agro-economic effects of the interventions across different farm types.

3. Statistical Typologies: Methodology 1)We divide the variables in TARBES and MWARBES into the five SI domains: productivity, economic, environment, social and human. 4) With the parsimonious set of socio-economic variables that explain most of the variation we perform a cluster analysis and divide the farms into 4 groups. 3) We use scree plots to define the number of factors to look at and, within each of the selected factors, we consider the variables with the highest absolute values of factor loads. 2)We perform separate factor analysis on each domain to select the variables explaining the largest portion of the variation in the data.

4. Factor Analysis: Tanzania

5. Cluster Analysis: Tanzania T1 T2 T3T4

6. Results: Tanzania

7. Regional distribution

8. SI Performance: Tanzania NOTE: The following variables are used to measure each domain: cereals yield (Productivity), asset-based wealth index (Economic), soil conservation index composed of crop rotation, alternative tillage, experience of soil erosion without measures for mitigating it and share of parcels with incrusted soils (Environment); gender equality index composed by female responsibility in managing certain plots and livestock (Social), and average education in the household (Human).

9. SI Performance (cont’d) Type 3 and 4 are the most productive groups, while type 2 shows mid-levels of productivity and type 1 is lagging far behind. In terms of economic endowments, type 4 shows a very strong performance, while the other groups being fairly close to each other at a lower level. In terms of human endowments, measured by educational attainments, types 2, 3 and 4 are very similar, while type 1 shows very low levels of performance. Finally, type 1 and type 4, despite their wide differences in productivity, economic and human endowments, perform similarly in terms of soil conservation practices and gender equality.

10. Suggestions for Tanzania  AR can focus on increasing productive capacity and economic endowments for group 1 and 2 through introduction of superior agricultural technologies. In addition, through nutrition trainings the project can improve food insecurity of type 1 and, to a minor extent, of type 2.  AR can focus on sharing information about the importance of preserving soil fertility and improving gender equality. This will improve the performance of type 2 and 3 in these two domains and prevent the decrease of the scores of type 1, which may result from the improvement of its economic and productivity conditions.

11. Cluster Analysis: Malawi T1 T2 T3 T4

12. Results: Malawi

13. District level distribution

14. SI Performance NOTE: The following variables are used to measure each domain: cereals yield (Productivity), asset-based wealth index (Economic), soil conservation index composed of crop rotation, alternative tillage, experience of soil erosion without measures for mitigating it and share of parcels with incrusted soils (Environment); gender equality index composed by female responsibility in managing certain plots and livestock (Social), and average education in the household (Human).

15. SI Performance (cont’d)  The largest differences are observed in the productivity and human domains for group 3 and, especially, group 4.  In terms of economic endowments, type 4 shows a very strong performance, while the other groups are fairly close to each other at a lower level.  Endowments in the social domain, measured by gender equality, are rather equally distributed across groups and on average relatively high, especially in terms of wage gaps.  Type 1 lags behind in terms of soil conservation practices, while type 4 performs the best. The groups with the least problems of soil quality are group 2 and 3.

16. Suggestions for Malawi  Farmers in type 1 need to increase their productive capacity, improve their endowment level and be trained about soil conservation practices.  Farmers in type 2 need a targeted intervention aimed at improving their productive capacity.  Farmers in type 3 and 4 are already performing quite well across all the domains and can facilitate adoption. AR should also support farmers in type 4 in mitigating their soil degradation problems.

17. Africa Research in Sustainable Intensification for the Next Generation africa-rising.net The presentation has a Creative Commons licence. You are free to re-use or distribute this work, provided credit is given to ILRI. Thank you!