What Mexico does wrong points to how we do elections right.
Mexico’s apparent president-elect, Felipe Calderon, says his country’s system of direct voting for president is clearly superior to the American electoral college; under a direct-voting system, “there never would have been any doubt about the triumph of Al Gore.” Meanwhile, his opponent, Andres Manuel Lopez Obrador, is demanding a national recount of all 42-million paper votes in all 130,000 ballot boxes. He and his lawyers are charging bias by the Federal Electoral Institute (IFE) and urging the IFE to annul the entire election, which would necessitate a new election. More than one million of Lopez Obrador’s supporters backed him up in a demonstration in Mexico City, while others transformed an annual teachers’ strike in Oaxaca into a massive people’s movement last weekend.
Suppose Mexico had an electoral college. What difference would it make?
Quite a lot, it turns out. For one thing, Lopez Obrador wouldn’t be demanding a recount, because he would be the winner. He would have 215 electoral votes; Calderon would have 181.
Mexico’s congress includes 300 deputies elected in single-member districts, and 96 senators, three from each of the 31 states and the Federal District. Lopez Obrador and Calderon each won 16 states, but Lopez Obrador carried more of the large ones.
It’s no wonder Calderon prefers a direct vote. With an electoral college, he would be the one demanding a recount. However, he certainly wouldn’t be demanding a complete recount of all 42 million votes; there would be no point. Unlike the U.S. in 2000, there were no really close states in Mexico last week. Virtually the only chance for Calderon would be in Veracruz, with 24 electoral votes, which he lost by about one percent (30,000 out of almost 3 million votes). As a practical matter, that is more like Ohio in 2004 (2.1 percent margin) than Florida in 2000 (less than 0.01 percent). (The closest state was Campeche, which Lopez Obrador carried by about one-half of one percent—less than 1,700 votes out of 312,000. But Campeche would have only five electoral votes: not enough to decide the election.)
Even if this hypothetical analysis isn’t relevant for Mexico, it is for the U.S. American liberals have, since the 2000 election, intensified their calls to abolish our electoral college and replace it with popular-vote plurality system. At first glance, the Mexican outcome would appear to reinforce their argument: two elections with the “wrong” result in six years. But consider what’s happening in Mexico. Lopez Obrador’s vigorous, if unsupported, allegations of widespread voting fraud gain plausibility from the north-south split in Mexican politics. The margin is so close that fraud in some of the conservative party’s strongholds in northern Mexico could have reversed the outcome. So Lopez Obrador’s supporters in the south are protesting, and in the process widening the political divisions in the country.
One of the great virtues of the electoral-college system is that it minimizes the problem of fraud: the incentives, the opportunities, and the consequences. In general it is easiest to steal votes where one party is overwhelmingly dominant—but there is no need to steal votes in those states. Fraud in Guanajuato, Nuevo Leon, or Yucatan, where Lopez Obrador received about 15 percent of the vote, would not matter; fraud in Tabasco, where Calderon received less than four percent, would not matter either. It would only matter in Veracruz, if anywhere.
As a veteran of Chicago vote-counting in 1960, I am familiar with fraud. In many respects it’s easier to commit fraud now than it was then. My wife and I learned this in 2000 when we volunteered to cold-call registered Republicans for George W. Bush in our Washington, D.C., precinct. We certainly didn’t expect it to make much difference in the election, and it didn’t. But again and again we heard that “he (or she) doesn’t live here anymore.” Children who originally voted at home were now college students or even adults; in-laws had moved away or died. Several people that we knew had died in the last few years—a lady in our church, three neighbors—were still on the list. Our precinct includes a military retirement home. Residents in three apartments told us the person we were calling had died—in one case, 30 years ago. Altogether, in the households we contacted, about 25 percent of the listed voters had died or moved. The chief judge in our precinct told us the District had no effective procedure for removing names; it was up to the individual—or the estate of a dead person—to tell the election board that he or she no longer lived in the District.
When I voted on Election Day, the clerk didn’t ask me for identification, just my address. In Chicago, somebody wanting to vote had to have at least a Social Security card, or two witnesses to swear to their identity.
Taken together, the polling list errors and the casual voting procedures made fraud easy. Anybody could have claimed to be one of the individuals who died or moved, and could have voted. It wouldn’t have mattered much even if there was fraud; Gore received 85 percent of the vote in the District of Columbia.
If the U.S. had a system like Mexico’s, however, it could have mattered. Gore’s popular-vote margin was slightly smaller as a share of the total than Calderon’s. Imagine if the U.S. had had to recount every precinct in 2000.
The Mexican election is relevant to other U.S. electoral-reform proposals. Some American reformers have suggested splitting each state’s electoral college vote proportionally, rather than winner-take-all by state. The Mexican congress includes 200 deputies and 32 senators elected by proportional voting, as well as the 300 deputies elected by district and the 96 senators elected by state. If the Mexican electoral college included these additional 232 members, elected proportionally, there would be no winner. Lopez Obrador would have about 288 electoral votes, with 315 needed to win. Calderon would have about 258. The balance would belong to Roberto Madrazo, of the Institutional Revolutionary party (PRI). Despite its diminishing popularity, PRI could pick the winner. Under that system, George Wallace could have selected the U.S. president in 1968, and Harry Byrd could have selected the president in 1960.
Or suppose each congressional district chose an elector, as is now the law in Maine and Nebraska. That would also have left Mexico without an elected president. Madrazo only carried 10 districts out of 300, but that would have been enough to deny an electoral majority to either of the more popular candidates.
Of course, if Mexico had an electoral college, the candidates would have run different campaigns. They would not have devoted much effort to maximizing their votes in their strongholds, any more than Bush did in Texas or Gore in New York. Rather, they would have concentrated on Veracruz, whose residents would have felt rather like the hapless citizens of Florida or New Mexico in 2004, bombarded by commercials and phone calls.
The point is that American liberals should think twice about abolishing the electoral college. Do they really want built-in incentives for fraud, national recounts, or third-party candidates choosing our presidents? The U.S. electoral college minimizes those considerable risks.
And North American liberals must be wondering what it takes to win. They have now lost in the U.S. with an electoral college, in Mexico with a direct popular vote, and in Canada under a parliamentary system. Three countries with three different systems—but with the same outcome.
This article appeared in the National Review Online on July 28, 2006.