Livestock and Private Tree Holdings in Rural Ethiopia: The Effects of Collective Action Institutions, Tenure Security and Market Access

Abstract

This article uses househld panel data spanning the period 2000–2007 to test hypotheses from the literature that secure land tenure, market access and collective action promote accumulation of private capital assets in rural highland Ethiopia. The three natural capital assets analysed in the article, livestock, eucalyptus trees and non-eucalyptus trees on households’ farm plots, make up virtually 100 per cent of privately held disposable assets. Incomes and capital stocks are extremely low and constant and tree assets are at least as important as livestock. We find that collective action and secure land tenure have strong positive effects on accumulation of livestock and other trees, but not eucalyptus. We also find evidence that market access promotes eucalyptus holdings and that other types of wealth tend to be positively associated with private natural capital stocks.

Acknowledgements

The authors would like to thank two anonymous reviewers for comments and suggestions that greatly improved the quality of the article. We also gratefully acknowledge Sida and Gothenburg University for financial support and the Department of Economics at Addis Ababa University for making the data available. We also acknowledge participants in EDRI seminars for their especially helpful insights, and Bluffstone thanks Ecotrust Northwest where he was on sabbatical during the revision. Access to data and code used in this article are available upon request.

Notes

1. Natural capital investments are tools that can be used to enhance livelihoods that happen to be provided by nature. Examples relevant to this article include animals and trees.

2. Collective action occurs when ‘more than one individual is required to contribute to an effort in order to achieve an outcome’ (Ostrom, 2004, p. 1). It is related to notions of ‘cooperation’, ‘trust’ and ‘property rights’, because of the important incentive problems associated with collective action (Olson, 1965).

3. See http://www.exchangerate.com. These are official exchange rates fixed by the government. Changes in exchange rates therefore do not provide sufficient adjustment for changes in the overall price level. Production includes agricultural output and biomass collections, which are primarily used for household consumption, as well as products for sale, sale (for example, honey, tree and animal products) wages and remittances. All products are valued at average local prices.

4. These estimates are well in line with MOFED (2002), which reports that average income per poor adult in Amhara was $97, which is less than our estimate. Dercon and Krishnan (1996) estimate that in 1989/1990 per capita incomes averaged only $52.62. We are therefore confident our estimates are reasonable.

5. Ali et al. (2011) also found significant perceived insecurity even though involuntary reallocations were very limited.

6. In 2006 Amhara Regional State passed a land use proclamation that granted land users legal right to use land. The effects of such changes, which since 2006 have spread to other regions, are a subject of active research (for example, see Tsegaye et al. [2012]). At the time of this study, land certification had not been actively launched in Amhara.

7. We use an LM test to test for homoscedasticity and reject the null hypotheses (Χ²(2) = 4.529e–36). Robust standard errors are therefore estimated. We also test for endogeneity of COLLECTIVE in the pooled models using Durban Χ² and Wu-Hausman tests. Identification variables include district (woreda) and local forest management dummies, community forest type and years since households acquired their land. These variables were effective instruments in Mekonnen and Bluffstone (2008) and Bluffstone et al. (2008). We find that for neither the livestock nor private tree stocks are the test statistics significant. We therefore are unable to reject exogeneity and proceed with uninstrumented models.

8. OWNSLAND and EXPLNSZ5YEARS Spearman’s ρ = −0.15.

9. Grazing is generally the second or third most important use of labour in rural areas of low-income countries.

10. Zone fixed effects in levels are included through the household level fixed effects. We nevertheless include the dummies to explicitly see whether and how location affects asset holdings.