How are different governments performing as global citizens? Time for a new index!

February 24, 2017

How introducing electronic votes in Brazil saved lives and increased health spending by a third

February 24, 2017
empty image
empty image

Just came across a paper which overcame even my scepticism about what often seems excessive hype around Brazil electronic voting machinetechnology’s impact on poverty and human rights. Check out ‘Voting Technology, Political Responsiveness and Infant Health: Evidence from Brazil’ by Princeton’s Thomas Fujiwara. He has stumbled across one of those wonderful natural experiments that allow you to try and pin down the causal impact of a particular change.

In this case it was the gradual introduction of electronic voting (EV) systems to replace hyper-complicated forms that meant that large numbers of Brazilians, in particular those who were poor  and/or illiterate, messed up their voting forms and were effectively disenfranchised. EV makes it much easier to get it right (it reduced spoiled ballots by about 10% of total voter turnout). The experimental bit lies in the gradual introduction of EV, which allowed Fujiwara to use a regression continuity design (me neither) to isolate what happened where EV was and wasn’t present. Its introduction was thus effectively a demonstration of what happens when poor people get to vote. Neat, eh? Here’s the findings:

Brazil electronic voting machine 2‘EV caused a large de facto enfranchisement of less educated voters, which led to the election of more left-wing state legislators, increased public health care spending, utilization (prenatal visits), and infant health (birthweight)…… The estimates indicate that the de facto enfranchisement of approximately a tenth of Brazilian voters increased the share of states’ budgets spent on health care by 3.4 percentage points, raising expenditure by 34% in an eight year period. It also boosted the proportion of uneducated mothers with more than seven prenatal visits by 7 percentage points and lowered the prevalence of low-weight births by 0.5 percentage points (respectively, a 19% and −68% change over sample averages).’

Wow. Electronic voting enfranchises 10% of the population, a previously excluded group that then elects authorities that increase health spending by a third over 8 years. And for someone like me who focuses on poverty and power, rather than technical fixes, the nice part is that in this case technology is improving the working of politics, not bypassing it.


  1. Really great story and paper, thanks Duncan. I’m also interested in how exactly the change came about – a case study for HCH edition 2?

  2. I am confused. In Britain there are mutterings among some liberals that if people can’t understand a voting system then they are not competent to vote.

    If they believe that the NHS will get £350 million more per week after Brexit, for example.

    In Brazil, why are the less educated swaying the vote for left-wing politicians (surely there are enough poor people in Brazil anyway)? Why do they not vote for the local equivalent of Brexit or Trump?

    What can we learn in the UK/USA, where less educated people do not seem to be keen to vote for the left wing policies of Jeremy Corbyn or the centre right Hillary Clinton? I read an article at the weekend that blamed it all on a right wing conspiracy that has managed to take over the internet and especially Google searches and Facebook adverts. But that seems too simplistic too. Maybe our left-wing political options just aren’t attractive enough?

  3. Really? electronic vote is easy to change and fraud, obscure to everyone except a small elite, and even for experts difficult to audit. You can’t reach that conclussion without touching the electronic fraud problem and the ignorance about software development in every politician promoting the worst system ever created.

    But of course, nobody can audit a system like that and nobody cares

Leave a comment

Translate »