Assessing the Institutions of Agriculture Governance

Several variables have been selected to help us to characterize the diversity of agricultural systems. Our dataset comes from two main sources: the World Bank 2008 World Development Indicators, and the CEPII 2009 Institutional Profiles Database (IPD).[1]

The first set of variables is linked to the characterization of agricultural public policies and transfer policies. Special attention is given to the weight of agriculture in the national economy and to the existence of an urban bias. Concerning the share of agriculture in the national economy, two variables have been selected: the share of agricultural GDP in national GDP, and the percentage of agricultural workers in the active population. In line with Bezemer and Hedeay (2008), urban bias is measured by the difference between urban and rural areas of access to safe water.

The second set of variables deals with the multiple purposes of agricultural production (food crops vs. cash crops). We retain the share of agriculture in exportation, which is expected to be higher in less developed, more agriculture-dependent countries. On the contrary, the agroindustrial share of GDP may reflect greater integration between national industry and national agriculture and, thus, a lesser dependence on agriculture. Food security is measured by three variables: the malnutrition prevalence height for age—measured by the percentage of children under five, the malnutrition prevalence weight for age—the percentage of children under five, and the undernourishment and Global Hunger Index.[2]

Types of farm organization are described by three indicators: the use of fertilizer per hectare, the number of tractors per hectare, and the productivity of a worker in agriculture (measured by the GDP per worker in agriculture). Whereas these three variables are rather good at describing modern agriculture, they need to be complemented in order to depict such peasant economy specificities as the small size of land assets, for which the land Gini, measuring the inequalities of land distribution, provides a good measurement.

As for strictly institutional aspects, these essentially focus on property rights. The six selected variables are provided by the CEPII 2009 IPD. They respectively characterize: (i) the diversity of land tenure rights systems (traditional, customary, collective, religious, “modern” rights, etc.); (ii) government recognition of this diversity; (iii) the significance of public land tenure policies,[3] (iv) the security of land tenure rights;

  • (v) land pressure, measured by the strength of the demand for land; and
  • (vi) the “Land tenure and large investors” variable measures the extent of large investment (national or international) in land property.

For all these variables, the reference year is 2005, with missing values, whenever possible, being completed by the nearest year for which a value is available. We have cut down the initial sample of 154 countries by eliminating those for which less than 50% of variables were known,[4] and then controlled for the representativeness of the remaining sample.[5] The PCA has thus been conducted for a sample of 145 countries for the year 2005. In the entire analysis, the role of the remaining missing data has been cancelled out by using the corresponding mean values. After sample adjustment, only 12 active variables have been retained for the empirical analysis: the percentage of agricultural workers in the active population, the share of agricultural GDP in national GDP, the urban bias indicator, the share of agriculture in exportation, the share of agro-industry in GDP, the number of tractors per hectare, the use of fertilizer per hectare, the productivity of a worker in agriculture, the diversity of land tenure right systems, the government recognition of the diversity of land tenure right systems, the significance of public land tenure policies, and the security of land tenure rights.[6] The data summary statistics and simple correlations between considered variables are shown in Tables 9.6 and 9.7 in the Appendix.

  • [1] The sources are presented in Table 9.6 in the Appendix. The CEPII 2009 IPD is available on:http://www.cepii.fr/francgraph/bdd/instit_form/login2009.asp
  • [2] The global hunger index is calculated on the basis of: (i) the proportion of undernourished peoplein the total population (in percentage); (ii) the prevalence of underweight in children under five (inpercentage); (iii) the under-five mortality rate (per 1000 live births). See Wiesmann et al. (2006)for a more detailed presentation.
  • [3] This variable is a synthesis of three elements: (i) the public arrangements available for formalisa-tion/registration of land rights in urban, suburban and rural areas; (ii) the policy fostering access toland for certain disadvantaged groups (minorities, natives, indigenous peoples, immigrants, etc.);(iii) eviction operations over the last three years (excluding conflicts, civil wars, etc.).
  • [4] Afghanistan, Bosnia and Herzegovina, Chad, Cuba, Ireland, Liberia, Libya, Somalia and VirginIslands have thus been excluded from the analysis. Moreover, Iceland and Singapore have also beenexcluded because they are extreme outliers.
  • [5] Note that complete information is available for 45.5% of the individuals and that 23.1% of themonly suffer one single missing variable.
  • [6] Six variables have been excluded from the PCA because they are misrepresented on the first twocomponents, and because they do not significantly contribute to the axis orientation. These variables are the malnutrition prevalence height for age, the malnutrition prevalence weight for age, theundernourishment index, the Global Hunger Index, the land Gini, the demand for land, and the“land tenure and large investors” variable. Nevertheless, these six variables will be reintroduced inthe second step of the analysis (cluster analysis) as supplementary variables in order to refine thecharacterization of the different country groups.
 
Source
< Prev   CONTENTS   Source   Next >