Can and should we count ballots faster?
Our national elections are complex and enormous, with as many as 160 million ballots, each with multiple pages and multiple races. As a result, there have never been final results on a national election night, ever. At no point in American history did we snap our fingers, count our ballots, and know the results of every race, no matter how narrow, on election night, or even the following day. In every single state, it takes days or weeks to certify and confirm the results and to make them official.
You may have seen a race called by the news media on election night, or the next day. But when that happens, ballots are still being counted. The confirmed results are still days or weeks away.
So can we count ballots faster—and should we?
If anything, ballots are counted faster and more efficiently now than ever before. Spurred by the Help America Vote Act of 2002, a number of upgrades over a 20-year period have improved the process of ballot counting, including
- Optical scanners and other systems to count ballots more quickly and accurately than the punch card and lever machines they replaced
- Electronic voting systems, most often with auditable paper ballots
- An expansion of early voting and mail-in ballots, allowing processing before Election Day, which makes for faster final tallies
- An increase in centralized counting locations for mail-in ballots, allowing for streamlined processes and quicker tallying
- Real-time reporting systems so counted results can be reported more quickly and accurately
But one recent development that helps improve counting efficiency, the expansion of mail-in voting, can have the paradoxical effect of slowing things down when it reaches a certain scale. As of October 2023, eight states (California, Oregon, Washington, Hawaii, Colorado, Nevada, Utah, and Vermont) now conduct their elections entirely or almost entirely by mail, which requires more time for sorting, verifying, and counting than in-person ballots.
In any large, complex system, there are always improvements to be made. But given the size and complexity of our national election system, and thanks to the great gains made in the speed, accuracy, and efficiency of vote counting in recent years, we are approaching reasonable limits in how quickly our national elections can be accurately counted and confirmed.
In fact, spurred by conspiracy theories about voting machines and optical scanners, a movement is now underway to reverse progress, ditching machine counts and returning to hand counts of all ballots.
If accurate elections are the goal, it’s hard to fathom a worse idea. Many studies have confirmed that humans, prone as we are to error and fatigue, make far more mistakes than machines. And this doesn’t even begin to address the question of speed. If waiting a day or a week for results is unbearable, laborious error-prone hand counting will virtually guarantee elections that go uncalled for weeks or months.
So can we count ballots faster? Marginally, perhaps. But should we? Only if it can be done without negatively impacting accuracy.