Needs-based targeting without knowing household incomes: How would it workin Russia?

Proxy means tests are often suggested for use in countries where tighter ta rgeting of benefits of social programmes are desired but verification of ho usehold incomes is difficult. This paper reports on simulations of the use of two types of proxy means test in determining eligibility for Russia's ho using allowance programme and the payments which eligible households would receive. Based on 1995 and 1996 survey data, tests are reported for three c ities: Vladimir, Moscow and Gorodetz. The results consistently show that th e proxy procedures introduce substantial errors into who is admitted to the programme. Most disturbing is the large share of eligible households who w ould be denied benefits using the first proxy (predicted income used for el igibility determination)-10-30 per cent. The average benefit loss for the p oorest quintile of households falsely denied benefits is equivalent to abou t 20 per cent of their incomes. Under the alternative procedure based on de mographic characteristics, errors are smaller than under the first procedur e for pensioners and large families but much greater for lone mothers. But even for the first two groups, undercoverage runs at about 15 per cent. On balance, these results suggest that very great caution should be used in pr oposing proxy income tests of the type analysed here in the countries of th e former Soviet bloc for means-tested programmes in general.