Or in other words, do governments know best? Best, that is, for poor people. If so, then adding conditions to a cash transfer alters household behaviour in ways that might help them escape poverty faster. Linking cash transfers to school participation is an example. Conditional cash transfers (CTTs) are on the rise (see this new book by BWPI’s Armando Barrientos and David Hulme, as well as the CPRC Kampala conference on social protection next week).
One view is that households know what’s best for themselves. So adding conditions to cash transfers to poor households is redundant. But collecting information is costly (in both time and money). Therefore households are unlikely to have full information. If so, then conditionality could steer them in a direction that they would go if they were fully informed. For example: Governments are likely to have a better understanding of the benefits of immunization than households, so conditioning the transfer on immunization will help.
Knowing whether conditionality does deliver tangible gains is vital. One reason is that conditionality by its nature implies monitoring, and monitoring has a cost. If conditionality does not deliver a gain then the money spent on monitoring might better go to increasing the size of the cash transfer itself (and even if there is a gain, it needs to be one big enough to justify the administrative cost of imposing conditionality).
So how do we capture the behavioural effect? It’s tricky, because behaviour can’t be directly observed. A new IPC one pager reports on the latest work by Alan de Brauw and John Hoddinott which aims to get around this problem (go to IFPRI for the full paper). They found that some households in Mexico’s PROGRESA (now called Oportunidades) did not receive the forms necessary to monitor their children at school, so their cash transfers were in effect unconditional. Their school enrollment is compared to the (majority of) households who were monitored.
Result? If the household was not monitored then its children were less likely to attend school on average. This effect is significant but modest. However, the absence of conditionality really kicks in when children should be moving from primary to secondary school. School attendance was severely reduced when children were making the transition to lower secondary school. These effects are even stronger when the household head is illiterate. De Brauw and Hoddinott conclude that:
“… debates over “to condition or not to condition” are overly simplistic. In the case considered here, there is clearly little benefit to conditioning transfers based on enrollment in primary school. However, in terms of increased school enrollment, there are large benefits associated with conditioning transfers at entry into lower secondary school”.
Hence, you can get a lot more bang from your CTT buck by focusing in on situations where altering household behaviour has the biggest effects — in this case encouraging more households to send their kids to secondary school.