Learn to live with an imperfect election system

Tired of the endless arguments over the counting, recounting and re-recounting in the Washington governor’s race? Me, too.

It’s time to acknowledge that no one can count votes with perfect accuracy. Once we’re over that hump, we can sensibly discuss what to do in this race and what sort of changes we’d like to see in the future.

King County election officials have been quoted as saying that their vote counting in the governor’s race was 99.98 percent accurate – implying an error rate of 2 one-hundredths of a percent.

By looking at the kind of voting systems we use, UW Professor Phillip Howard estimates Washington’s vote counting error rate to be 1.1 percent, which is among the lowest in the nation.

Sounds great, right? However, the gap in the Washington governor’s race – at least the last time I looked was 129 votes.

Let’s put all these numbers together. There were just over 2.8 million votes cast. If the error rate was 1.1 percent, then we’re talking about 31,000 counting mistakes.

Even if the more optimistic King County error rate is right, and if it applies statewide, then something like 560 ballots would have been miscounted. Either way, the error rate swamps the margin of victory.

It’s not that this election was especially messed-up. No one has provided any evidence that the counting process was worse than usual. It’s just that in most elections we don’t notice the warts in the system.

In this election, first the Democrats and then the Republicans have had something to gain by squawking loudly about how messed-up the election was. This is to be expected, but it’s still not very classy.

We’ve counted three times. The counts have been almost identical. If we made a fourth count, it would also be about the same. When an election is as close as the one we just had, there isn’t any way to reduce counting errors enough to get a guaranteed accurate result.

All we can reasonably ask is that we follow the rules that we agreed to before the election. Those rules do include provision for making judgment calls on whether to count particular ballots: they place the responsibility with the county canvassing boards, and we should stick to that agreement for this election.

Looking forward, shouldn’t we reform the election process so that the whole operation is more accurate? The answer isn’t as obvious as it might seem.

Let’s divide reform options into three levels.

First, there are some changes that are easy and inexpensive. Let’s ask whether it would be more accurate, and maybe even cheaper, to do recounts by machine rather than by hand. And surely it would be both sensible and cheap to make procedures for checking signatures uniform across the state.

Next come some options that are doable, but aren’t so cheap. We could have additional election personnel, better training, more equipment, etc. But additional personnel, better training … all these cost money.

The third level of reform would be to commit to perfect counting. As political rhetoric, this is appealing. But perfect counting isn’t a feasible option because nothing people build is perfect. When it comes to vote counting, a commitment to perfection would amount to sending a large amount of tax money chasing after very little benefit.

The truth is that whenever the voters of the state are as evenly divided as we were in the recent election, all the vote counting schemes we can think about come down to taking our best guess as to who squeaked out a few more votes.

If we can make that guess more accurate in a cost effective way, great. But beyond the easy fixes, we might be better off learning to live with an imperfect system. When it comes to getting the count correct to the last decimal place, there are more pressing needs for our limited tax money.

We should insist that the folks running elections be honest and competent – which in Washington is already pretty much the norm. We should not insist that they be perfect.

Dick Startz is Castor Professor of Economics and Davis Distinguished Scholar at the University of Washington. He can be reached at econcol@u.washington.edu.