Measuring Households' Vulnerability to Idiosyncratic and Covariate Shocks – the case of Bangladesh

The paper examines the level and sources of vulnerability in rural Bangladesh using a household survey. We use a simple two-level random intercept model to estimate expected mean and variance in consumption as well as to decompose the variance into idiosyncratic and covariate components. Our results indicate that both idiosyncratic and covariate shocks have considerable impact on household's vulnerability and idiosyncratic shocks seem to have greater impact on household's consumption vulnerability than the covariate shocks. Furthermore, idiosyncratic shocks appear to have a relatively higher impact on relatively well endowed (i.e. in terms of human capital, land holdings, activity status etc.), well off households and covariate shocks seem to have a relatively higher impact on poorer, less educated, household's vulnerability. Our results also reveal that rural vulnerability in Bangladesh is mainly poverty induced rather than risk induced. Around 78 per cent all who are vulnerable is accounted for by low expected mean consumption and only 22 per cent of them are due to high consumption volatility. Overall vulnerability in rural areas is estimated to be 50 per cent. The categorization of poverty into transient and chronic poverty is even more insightful. The study finds that those without education or agricultural households are likely to be the most vulnerable. The geographical diversity of vulnerability is considerable. It is suggested that ex ante measures to prevent households from becoming poor as well as ex post measures to alleviate those already in poverty should be combined.