Venezuela: Academics' Study Backs Fraud Claim In Chavez Election
David Luhnow And Jose De Cordoba, The Wall Street Journal
2004 September 07 - Two Venezuelan academics claim to have found statistical evidence of fraud in last month's referendum on President Hugo Chavez, fueling the opposition's claims of a rigged vote and raising the possibility that despite Mr. Chavez's victory, the country's tense standoff will continue.
The claims were made Sunday by Ricardo Hausmann, a professor at Harvard University's John F. Kennedy School of Government and former chief economist at the Inter-American Development Bank, and Roberto Rigobon, a professor of applied economics at the Massachusetts Institute of Technology's Sloan School of Management.
The pair issued a report that tried to measure the possibility that the vote was clean using two separate analyses of the official results. In both cases, they said, the chances of a clean vote were less than one in 100.
Members of a civic group called Sumate that organized the referendum, which Mr. Chavez won by a 59% to 41% margin, seized on the study to suggest Mr. Chavez had won by tampering with the electronic-voting machines used in the contest. "We don't think the truth about the referendum has been revealed yet," Alejandro Plaz, a spokesman for Sumate, told reporters in presenting Mr. Hausmann's study Sunday. Sumate requested help from the academics in analyzing the referendum data but didn't pay for the study.
Mr. Chavez's government reacted with disbelief to the claims, saying the opposition's previous claims of fraud had so far proved incorrect. Vice President Jose Vicente Rangel said members of the Atlanta-based Carter Center and the Organization of American States had already validated the result. "No one believes in their theories anymore because three weeks have gone by and they haven't been able to prove anything," Mr. Rangel said.
Members of the Carter Center and the OAS were unreachable for comment yesterday. But both organizations have consistently stood by their findings in the past weeks and watched as other theories of fraud fell short under scrutiny.
The results of the study, however, prompted some independent experts on computer voting to call on the Venezuelan government to open up all aspects of the election -- including electronic codes from voting machines -- to public scrutiny.
"The Hausmann/Rigobon study is more credible than many of the other allegations being thrown around," said Aviel Rubin, a computer-science professor at Johns Hopkins University who has warned about security flaws with electronic voting. Mr. Rubin recently conducted a study of opposition claims that machines were rigged to limit the number of votes against Mr. Chavez and concluded the claims were highly unlikely.
"I would encourage the Venezuelan government to open up all aspects of the election to public inspection, not just to selected observers. That includes all of the paper ballots, the source code in the voting machines, the random generators ... that were used to pick the sites to audit," he said in an e-mail interview.
The study by Messrs. Hausmann and Rigobon suggested the government may have tampered with only some of the machines, leaving others clean for observers to audit. They said the sample used for the audit, which was carried out days after the election, wasn't randomly chosen and limited to the "clean" machines.
The study says the computer that determined which ballot boxes were to be subjected to a recount belonged to Venezuelan election officials. However, the Carter Center's Jennifer McCoy has said the group tested and verified the computer program used to select the sample.
The study compared the votes obtained by the opposition during the recall vote with the signatures gathered in November 2003 requesting the referendum. For the recounted votes, the correlation between the number of "yes" votes matched the 2003 petition numbers at a rate that was 10% higher than in the ballot boxes that weren't recounted. They calculate the probability of this taking place by chance at less than 1%.
The government's sample recount "was not a random sample, and I can say that with 99% confidence," Mr. Hausmann said in a telephone interview.
The academics used another technique to look for suspicious patterns in the results, using the 2003 petition and an exit poll on the day of the vote as a vague measure of a voter's intention. Because both measures are imperfect for different reasons, the academics argued, the measures should make different mistakes in predicting the final result.
But the academics found that each method had similar margins of error when compared with the official results, something that would happen only one in 100 times without fraud, they argued.
send this article to a friend >>