Share

The World Bank Policy Research Working Paper No. 5604, March, Jose Cuesta, Svetlana Edmeades and Lucia Madrigal (2011)

The paper examines the reduction of food insecurity in Bolivia by adopting a supply side approach to analyze the role of agricultural spending on vulnerability. The paper indicates the importance of the content of public agricultural spending in determining the vulnerability to food insecurity.

Food security has been pursued as a key policy objective by many developing countries after the onset of the food crisis. Bolivia is no exception to this and the government has established several programs to address the issue of food security.  Food security, as identified by the authors, is a phenomenon determined by various factors. In the case of Bolivia, declining food security could be the result of some major factors such as: a) reduced cultivated land area and increased land division, b) increased food imports and dependence on the international market, c) a shift of labour away from rural to urban areas and d) environmental shocks like droughts and floods. Furthermore, other structural factors must also be taken into account such as the geography of the country and limited road coverage which hinders domestic market integration.

The researchers studied the relationship between agricultural spending and food security by taking a supply-side approach. The authors pointed out that agricultural spending in Bolivia totaled around 13 percent of the GDP in 2008, and as such, is a good indicator of public interventions aimed at tackling food security issues. They stated that by showing linkages between public expenditure and a measure of vulnerability, which is known as the Vulnerability Analysis and Mapping or VAM, policy recommendations can be made based on this guide.

The authors measured vulnerability by using the VAM, an index which measures the degree of food insecurity and vulnerability at the municipal level. It is assigned a value from 1 to 5 where 1 refers to a very low level of food insecurity and 5 indicates a very high level. It is widely accepted in Bolivia as a measurement of food insecurity and is available for 327 municipalities across Bolivia. The VAM consists of a couple of components – namely rural population density, schooling years, log of per capita consumption, under-five malnutrition rate, rainfall and flood propensity categorical variable consisting of four values. The VAM for 2003 includes life expectancy, agricultural and forestry potential which is a categorical four point scale capturing soil capacity, road density and per capita expenditures etc. The authors established that the average vulnerability status decreased over time which means that municipalities have reduced their vulnerability to food insecurity. They further found out that at the municipal level, 62 percent of the municipalities were categorized in 2007 as having a VAM rating of moderate to very low vulnerability (1 to 3) to food insecurity while 38 percent had ratings of high to very high vulnerability (4 to 5). However the authors pointed out that mobility across vulnerability levels was very limited; 80 percent of municipalities did not change their statuses of low/moderate or high/very high between 2003 and 2007.

The study showed that public spending in agriculture has increased somewhat over the last five years. Public expenditure used in the study included data on agriculture and rural development disaggregated by function, which contained spending on research, extension, irrigation, rural roads, etc. and also disaggregated by current and capital spending, for the period 1996-2008. Spending is further separated into restricted and extended spending.

The authors ascertained that agricultural spending and each category of vulnerability are only weakly correlated with each other and based on this they put forward several explanations. One explanation was that the impact of each component in agriculture spending may not be captured through this one aggregate measure of public spending in agriculture – the composition of spending may be more important for reducing vulnerability to insecurity. Another reasoning that was put forward was that higher agricultural spending may only be a substitute for lower development levels, rather than a variable that picks up impacts on vulnerability.

The researchers further analyzed the effects of agricultural spending on high and very high vulnerability at a municipal level.  They found that for the year 2003, per capita agricultural spending is positively associated with a high or very high vulnerability to food insecurity. Investment spending and infrastructure spending also had positive associations with higher vulnerability to food insecurity. This suggests that there is more investment and spending on infrastructure in areas of Bolivia where there is a greater risk of food insecurity. For the year 2007 the association is the same as in 2003, but the magnitude is less.

However, the authors established that public spending levels could have been dictated by past vulnerability and categories of public spending such as infrastructure might be determined by changes in vulnerability. To verify this issue the researchers used other estimation techniques. They test for the longer term inter-temporal effects between the years 2003 and 2007. The authors found that the effect of agricultural spending in 2006 has a significant impact on the vulnerability to food in 2007 – spending in 2006 has a positive impact on VAM in 2007. Investment spending and current spending also has a bearing on the vulnerability in 2007. In case of spending in 2003, the results were similar.

Research spending and extension spending and their effect in reducing vulnerability were also significant, thus implying that such spending is effective in addressing vulnerabilities when a longer lag is considered.

The authors conclude the paper by stating that the relationship between spending and vulnerability needs to be examined more thoroughly and that components to VAM, such as weather shocks and political economy, have proven to be poor measures and as such other components need to be looked at.