Do Anti-Poverty Programs Sway Voters? Experimental Evidence from Uganda

Abstract

A Ugandan government program allowed groups of young people to submit proposals to start skilled enterprises. Among 535 eligible proposals, the government randomly selected 265 to receive grants of nearly $400 per person. Blattman et al. (2014) showed that, after four years, the program raised employment by 17% and earnings 38%. This paper shows that, rather than rewarding the government in elections, beneficiaries increased opposition party membership, campaigning, and voting. Higher incomes are associated with opposition support, and we hypothesize that financial independence frees the poor to express political preferences publicly, being less reliant on patronage and other political transfers.

How incumbent politicians respond to the enactment of a programmatic policy: evidence from snow subsidies

“…Like Blattman et al . (2018), both studies documented declines in support for the incumbent after receiving these goods. Although neither basic services nor infrastructure fit the definition of a programmatic policy, both these studies are relevant because they attributed the declines in support they observed to a similar logic.…”

“…By 2014, more than 50 countries had CCTs (World Bank, 2014). Other (non-CCT) programmatic policies adopted in this period include grants to start small businesses in Uganda (Blattman et al ., 2018), vouchers for the purchase of a computer in Romania (Pop-Eleches and Pop-Eleches, 2012), and universal health care in Mexico (Imai et al ., 2019).…”

“…Blattman et al . (2018) studied a World Bank-funded program in Uganda, the Youth Opportunities Program, which used a lottery to award cash grants for independent trades to groups of poor and unemployed young adults. Four years after the program's implementation, the researchers surveyed winners and losers and found effects in the opposite direction: winners were more likely to support the opposition.…”

“…Blattman et al . (2018) venture that this logic can explain why recipients of the programmatic policy they studied ended up being less likely to support the incumbent.…”

“…Beneficiaries may be engaging in ‘ordinary pocketbook voting’ (rewarding incumbents for improving their livelihoods), inferring pro-redistribution policy preferences or competence on the part of the incumbent, or wanting to help an incumbent who helped them. What is notable about this work, however, is its lack of consensus: in some cases, researchers found support for the PISH (Manacorda et al ., 2011; Pop-Eleches and Pop-Eleches, 2012), but in others, they found no support (and occasionally, evidence of the opposite) (Blattman et al ., 2018; De Kadt and Lieberman, 2020).…”