Remittance and expenditure patterns ofl low-income immigrants

Samuel Kojo Antobam, University of the Witwatersrand

In the last decade, reports of about the flows of remittance have increased as much as the increase in the amount remitted. National statistical agencies and central banks have recently become more interested and diligent in tracking these flows. In addtion, various econometric models are being used to estimate the impact these flows make on the receiving households. But those who do the remitting, have been left out in these estimations. What are the characteristic differences between remitters and non-remitters? How much are individual migrants willing to send to relatives at home from every Dollar, Euro, Pound or Rand? And how much are migrants willing to vary the expenditure on remittance with change in their income? These are questions that need to be answered to help us understand the impact remittances make in the lives of the migrants. Using the Working-Leser econometric model on dataset from Migration and the New African City project carried out by Forced Migration Program of the University of the Witwatersrand in 2006, this study finds that: 1. Even though a greater percentage (51%) of those who earn more than R35000 annually remit home than those who earn less, the percenatage of inocme that goes into remittance is higher for those who earn less than R35000 annually. 2. When the gender of household head, education level of the household, household size and length of stay in the municipality are controlled, marginal budget share that goes into remittance is much higher for lower income/expenditure group than the higher income groups. 3. The sensitivity of amount of expenditure on remittance to changes in income is measured by remittance expenditure elasticity coefficient. It emerged that all immigrants in various income categories as well as different parts of Africa, have inflexible response to percentage changes in income. However, this response is a bit relaxed for the poorest immigrants.

  See paper

Presented in Poster Session 2