FRANCISCO R. RODRÍGUEZ, Francisco.Rodriguez4@du.edu, @frrodriguezc
JAKE JOHNSTON, johnston@cepr.net, @jakobjohnston
Johnston is a senior research associate at the Center for Economic and Policy Research. He quotes Rodríguez, who is the Rice family professor of the practice of international and public affairs at the University of Denver’s Korbel School of International Studies, in a recent piece: “Ahead of Venezuela’s Election, What Do the Polls Really Show?”
Johnston writes: “This Sunday, July 28, Venezuelans will head to the polls in what has been framed as the opposition’s best chance to unseat President Nicolás Maduro, who has been in office since 2013. From the media coverage, one comes away with the impression of only two options: an opposition win, or massive pro-Maduro fraud.
“There are many pollsters in Venezuela, and indeed most have shown opposition candidate Edmundo González Urrutia with a clear advantage. There are eight other candidates running, though none appear to have significant support. González is ‘leading the race with support from more than 50 percent of respondents in several polls,’ the New York Times reported last week. Polls show he ‘would easily win the vote,’ according to the Wall Street Journal.
“They aren’t wrong. But it’s possible the pollsters are.
“As Venezuelan economist (and CEPR Senior Research Fellow) Francisco Rodríguez recently noted on X (formerly Twitter), an average of the polls conducted ahead of the July 28 vote show González with a 28.9 percentage point advantage over Maduro. But, he continued, after correcting for polling bias, that lead largely evaporates. Since 2015, pollsters have overestimated opposition support by an average of 29.5 percentage points, he found.
“Rodríguez corrected for polling bias in three different ways — based on the average bias from 2004 to 2021, the average bias from 2017 to 2021, and based on an overall average that gave greater weight to polls in more recent years. He also adjusted the numbers by giving a relatively greater weight to pollsters with greater historical precision. The results show a range of outcomes, from a 9.5 percentage point advantage for González to an 8.1 percentage point advantage for Maduro.
“’Rather than providing a precise projection, what these estimates show is the profound uncertainty we may have today about the electoral outcome given how inadequate polls have been as predictors of results in the recent past,’ Rodríguez wrote.”