Idaho Election Costs: Fun With Denominators

[Image courtesy of Eye on The Hill]

A recent election in Ada County (Boise), Idaho is getting some attention after a local television station obtained some cost data and decided to do the math. From KTVB.com:

The May 21 election in Ada County saw extremely low voter turnout, with only 5.4% of registered voters going to the polls. With low turnout, 7 Investigates looked at how much each vote was “worth” in terms of the cost to the county taxpayers for each vote cast.
On the May ballot in Ada County were issues from funding the Eagle City Hall to electing the board of the Greater Auditorium District and Kuna schools. Overall, 9,457 people voted.

“I would say the voter turn-out was lower than we anticipated. I thought it would be below 10% for this election because it was still a very small district election, but I thought it would be much closer to that 10% mark as opposed to down to 5% where we actually ended up,” Ada County Chief Deputy Clerk Phil McGrane said.

The Ada County Elections office preliminarily tells KTVB the May election cost $149,452.71. That number will go up because figures have not yet been calculated for things like staff wages, postage, notification cards and poll book printing.

Using that cost and the number of voters who showed up at the polls, each vote is calculated to have cost $15.80.

For context, the 2012 November election had such high turn out (184,000) that even with a total cost of $740,000, the county only spent $4.02 per vote.

Neither one of those numbers is terribly surprising; Presidential elections are always going to be higher-turnout affairs, with the result that cost per vote cast is going to be smaller. But is that “cost per vote cast” really the best figure to use? I took a look at that issue way back in September 2011 in the context of some North Dakota data and found that “cost per eligible voter”, using voting age population as the denominator, provides the basis for an apples to apples comparison and one which “washes out the unpredictable effects of turnout.”

Consequently, using the Census Bureau’s 2012 estimate of 310,886 for voting age population (total population estimate times 76% voting age) the new cost figures are about $0.48 for the 2013 election and $2.38 for last November. Those numbers not only eliminate the effects of turnout, they also better capture the increased costs to a jurisdiction for a longer ballot like a presidential race.

Maybe a better measure of the impact of turnout is to calculate the percentage of total election costs borne by actual voters as opposed to non-voters. That figure would be 3% (9,457 votes /310,886 voting age population) for May 2013 (or only a little over $4,480 of the total cost) and just over 59% (or about $436,000) last November.

In that context, the cost to non-voters – many of whom are taxpayers, whether or not they are registered or actually vote – is much clearer.

All of this is very simple math but it nonetheless gives a preliminary picture of how election costs are borne across a jurisdiction. Of course, none of this fun with denominators is possible without a numerator – the very important ingredient of cost reporting. Kudos to Ada County for sharing and here’s hoping many (many!) more jurisdictions follow their lead …