Financial Crisis and Migrant Remittances: Estimating the Effects on Growth and Poverty
April 30, 2011
At the macro-level, an empirical analytical framework is developed to estimate the potential impact of a drop in the growth rates of advanced and Middle Eastern economies (which are the major sources of remittances to South Asia) on remittances originating from these economies and the income of the remittance-receiving country. At the micro-level, the effect of remittances on the incidence and depth of poverty is estimated using household-level data.
In Pakistan and Bangladesh, remittances seem to be motivated primarily by consumption, and a slowdown in the growth rate of advanced and Middle Eastern economies is predicted to have a negative effect on remittances growth in the coming years. Using the average growth rate over 2003-07 as a proxy for the growth rate without the crisis, based on the empirical analysis, the growth rate is expected to decline to an average of 7.25% and 17.67% for Pakistan and Bangladesh respectively for the period 2010-12. The impact on flows to Sri Lanka is the smallest where a decline of about 1.75 percentage points is predicted over the period. This is attributable to what appears to be other factors also playing a part in remittance flows rather than just purely altruistic motives of migrants. How will this slowdown in remittances affect these economies? Estimates from the dynamic structural Keynesian model show that the short-run (1 year) yields an estimated decline of 0.4 percentage point in GDP growth for Pakistan, 0.87 for Bangladesh and 0.3 for Sri Lanka. The impact of the economy is also traced through the expected change in unemployment rates through the use of employment elasticity (elasticity of people employed to GDP) for these countries. The resulting impact is in the same order as the impact on GDP and remittances, with the impact on Bangladesh being the largest with a predicted increase in the unemployment rate of 0.7 percentage points, 0.26 in the case of Pakistan and a mere 0.06 for Sri Lanka for the period 2010-12.
As a supplemental exercise, the effect of remittances on household poverty at the micro-level is also estimated using the Multiple Indicator Cluster survey (MICS) data for the province of Punjab. Using the predicted counterfactual income of the remittance-receiving households, poverty and inequality measures such as the headcount ratio, poverty gap, and the Gini inequality coefficient for the remittance and the non- remittance receiving groups are calculated. To overcome the problem of likely self-selection into migration, propensity score matching is also used to better match remittance and non-remittance receiving households on observable characteristics.
Across these two different methodologies, there is robust evidence of a 13-15% decline in the incidence of poverty among households receiving remittances. This translates into a 50-60% reduction in poverty for remittance-receiving households based on the national poverty line estimate of 24%. With the estimated decline in the growth rate of remittances to 7.25% after the crisis, the proportion of households in receipt of remittances in Punjab will be 4.4% as opposed to 4.5%, given the 2007/08 baseline. Given this nominal change, in the overall distribution of households in receipt of remittances, resulting from changes in the growth rate of remittances, any impact of financial crisis on provincial and national level poverty rate is limited and is likely to be statistically indistinguishable from zero in a more structured computational general equilibrium model of the economy.
The final report is now available online at: http://creb.org.pk/sCurrrentResearch.aspx
posted by S A J Shirazi @ 4/30/2011 08:50:00 AM,
104 - C, Gulberg III,
Phones: 92-342-5714936, 38474385
Intersection Main Boulevard Phase VI
Phones: 92-4236560935, 36560939