On the weekend before Election Day, about 80,000 people statewide were notified that changes to their address or party affiliation made at a Motor Vehicle Administration kiosk were never uploaded to the elections board, and they would have to cast a provisional ballot instead of a regular one. Evidently, that was enough to dissuade a lot of them from showing up at the polls: After June’s primary, state elections chief Linda Lamone testified that those who had been caught up in the snafu voted at about half the rate of those who weren’t.
Anecdotal evidence previously has suggested that people in general don’t fully trust provisional ballots or consider them a hassle and therefore don’t bother to cast them, but there’s not a lot of empirical data to back that up. Under normal circumstances, it would simply be too difficult to conduct surveys outside of a sufficient number of polling places to find the few voters who were told they would have to cast provisional ballots and find out whether they followed through or not. But in this instance, we have evidence suggesting that when voters were notified in advance that they would have to vote provisionally, they didn’t try to vote at all.
Political scientists will no doubt be digging into that phenomenon, but the question is of more than academic interest to candidates who campaigned for months or years only to come up slightly short. And here, I’m thinking of Sen. Jim Brochin, the Democrat who was 17 votes behind former Del. Johnny Olszewski Jr. in the primary for Baltimore County executive.
Candidates who lose close races often wonder if there was something more they could have done to get those last few votes. Having watched Mr. Brochin campaign since his first Senate race in 2002, I’m pretty sure he’s not wishing he had worked harder. Anyone who’s run against him will attest that’s simply not possible. But could something completely out of his control have made the difference, specifically the MVA registration mixup?
If the affected voters were distributed proportionally across the population, it wouldn’t matter — every candidate would have been hurt equally by those who stayed away rather than having to use a provisional ballot. But they weren’t. For whatever reason, certain parts of Maryland’s electorate were more heavily affected by the error than others. Most striking: Republicans make up about 26 percent of registered voters statewide, but according to statistics from the state elections board, they made up almost 58 percent of those affected by the MVA glitch. Those affected weren’t evenly distributed geographically, either. In Baltimore County, the total number of affected voters from all parties ranged from 1,249 in the 2nd council district (which executive candidate Vicki Almond represents) to 2,145 in the 7th district (formerly represented by Mr. Olszewski’s father).
It’s impossible to know, of course, how many of the voters who were notified of the glitch would otherwise have turned out, and I certainly can’t say for whom they would have voted. Still, the state and local election boards have published enough information to provide a reasonable guess at how many “lost” votes there were in the Democratic primary in each precinct.
Precinct by precinct, I compared the overall level of Democratic turnout with the turnout among those affected by the MVA problem to estimate how many Democrats stayed home because of it. I then multiplied each precinct’s “lost” votes by the percentage of the vote each of the candidates received in that precinct to provide a reasonable guess at how many more votes each of them might have gotten absent the turnout differential.
The conclusion? County-wide, about 523 more Democrats might have been expected to vote absent the MVA problem, but if anything, that hurt Mr. Olszewski in Baltimore County. Based on how each candidate performed in each precinct, Kevin Marron might have expected to get 15 more votes; Ms. Almond, 128; Mr. Brochin, 182 and Mr. Olszewski, 198. That is to say, had the error not occurred, Mr. Olszewski’s small margin of victory might have doubled, from 17 to 33 votes.
Using the same margin on two other close races suggest the glitch probably didn’t change the outcome either in the Democratic Montgomery County executive primary or the 1st District Howard County Council race. In the latter, challenger Liz Walsh would have theoretically gained another two votes in her victory over Councilman Jon Weinstein, padding her margin to a whopping eight votes. In Montgomery County, the MVA problem may have cost David Blair, but not enough to make a difference. He lost to Mark Elrich by 79 votes; using this model suggests that absent the MVA glitch he would only have lost by 65.
The problem may not have made the difference in any races in this primary, but it easily could have. At at time when we’re focused on those trying to interfere in our elections, it’s a reminder of how little it can sometimes take to tip the outcome.
--Andrew A. Green