In the eight months since we released our report on post-election auditing, Jim Mueller and I have been working to keep post-election audits on the radar of the members of the Government Accountability Board. Here's where things stand in early April:
We're making enough progress that I'm cautiously optimistic that--with some not-impossible work--we can greatly improve Wisconsin's post-election audit practices in time to better protect the November election results from machine miscounts.
It's boiled down to this: I am convinced that the members of the Government Accountability Board (the judges on the board itself) are interested in and, as I write this, committed to improving Wisconsin's post-election audit practices. I see two roadblocks preventing the Board from exercising leadership:
- The GAB doesn't have statutory authority to exercise effective leadership in election integrity. That sounds harsh, but it's not meant to be. Our elections statutes place much more responsibility on local officials than on GAB, but don't even give local officials clear, explicit responsibility to verify that our machines produced correct totals. (Key point: But statutes don't forbid it, either.) Whoever drafted those laws seems to have been caught up in the seductive fantasy that voting machines are a modern-day Oracle that produces an infallible revelation with the press of a 'print results' button.
- The Board members are not getting much visible support from staff in understanding the possibilities around better post-election audits and their policy options. GAB staff--at least Director Kevin Kennedy and Elections Administrator Mike Haas--have for some reason been providing the board with materials and oral statements that merely repeat the reasons why they believe better post-election audits are too much for Wisconsin to handle. (Sweeping back of hand across forehead; swooning). Their guidance for the Board has not yet included any creative problem-solving about things local clerks or the GAB could be doing under current law to improve post-election verification, and uses the phrase 'legislation is needed' as if it were synonymous with 'it's impossible; end of discussion.' Details and links are below in the chronology.
But we don't really need active, assertive leadership from GAB to improve post-election audit practices. There's no need to criticize either Board or staff for failing to exercise authority that statutes don't specifically give them or to deploy resources they don't have.
It's the local election officials who have the most direct authority (and, in my observation, the strongest sincere desire) for rock-solid accurate results. It's also the local officials who could, if they chose, use citizen volunteers to help conduct post-election audits at a very low cost. So our strategy has evolved to this:
- Get GAB to stop telling the world that Wisconsin's machine-tabulated election results cannot be verified without a change in statutes. Although it's true current statutes do not require local officials to check to make sure the machines tabulated accurately, they are no worse than silent on the matter. And post-election auditing could be among the most cost-effective activities in elections administration. GAB Board member Judge Elsa Lamelas asked me, when I appeared at the Board's March 19 meeting, to tell the Board about some of the options for cost-effective post-election audits. Here's the letter in which I responded, reiterating our request that GAB stop saying statutes need to be amended before local election clerks can verify machine counts, and giving the reasons why we think they should do that.
That's all we're asking of GAB at this point: Not to lead, not to follow; just to get out of the way.
- Work with county clerks and municipal clerks (and through them, local Boards of Canvass) to help them understand the need for post-election verification and their options, and then to help them verify the accuracy of their machines' counts before certifying them. This is going to be more work. We need to build relationships with the County Clerks' Association, the Municipal Clerks' Association, and with individual clerks. Mary Lou Sharpee has started to work on this with me, but we haven't got a lot to update right now; if anyone is interested in helping to develop or carry out a strategy, please contact me at email@example.com
I'm also guardedly optimistic that the Legislative Audit Bureau will release its evaluation of GAB procedures before the November elections. When the Legislature's Joint Audit Committee directed LAB to begin this evaluation, I wrote to the committee encouraging them to make sure LAB looked at post-election audit policies and procedures. If the final report covers the topic of post-election audits, I'm sure it will confirm the findings in our July report. Because LAB knows good audit practices in their sleep, and because they have access to much more information that we did when we wrote our report, I am 100% confident they won't praise GAB's policies regarding post-election audits. I wouldn't be surprised if they find more problems with those policies and practices than we did.
For anyone who wants all the details, read on....
October 2013: We presented our report to the GAB Board. Current GAB procedures do not allow citizens to appear as a regular agenda item at a Board meeting. So I took advantage of their policy of allowing any citizen to make a five-minute comment before every board meeting to encourage them to read our report. Two months later, the staff had--for the first time in their history--publicly reported on the audits.
December 2013: GAB staff reports the results of the 2012 audits. The report did not actually contain the results of the audits. It merely described the procedures and said that no errors were found. (The report starts on page 13 of this document.) The thing is that GAB's procedure defines 'error' as any discrepancy that the municipal clerk can't explain, so no discrepancy is ever counted as an error.
That particular weirdness caught the attention of the members of the GAB Board, and they asked some great questions at the December meeting. In addition, they directed staff to report back on how post-election voting machine audit practices could be strengthened. From their discussion, it was clear that the board members are mindful that November's elections are likely to be contentious, and they expressed a desire to get better post-election processes in place before then.
January 2014: GAB staff redefine the Board's questions, and tell the Board to ask again if they really want to know.
In January, GAB staff provided GAB Board members with a curious memo. (The memo starts on page 31 of this document.) Basically, what staff did was to redefine the issues in a way that made them easy to dismiss without tackling the serious issues that we and the Board members had raised. For example, the memo took pains to argue against Election-Night audits, which I hadn't heard anyone mention at the December meeting. Audits at any time before certification of election results would be a big improvement over current practice; the fact that it would be burdensome to check the machines' accuracy in the wee hours of Wednesday morning is hardly a compelling reason to forgo verification entirely.
The memo concluded by telling the Board that it needs to "determine whether it wishes to receive further information ... and to provide further direction to staff" if they want to know any more.
I wasn't able to listen in to the January teleconference, but the minutes are available here. As I read those minutes, GAB board members remained genuinely concerned that we're not doing enough to make sure voting machines worked correctly, but their discussion was diverted to questions about the 0.025% percent of votes that are ambiguously marked, away from the fact that no one ever checks whether the other 99.975% of the votes were tabulated correctly.
March 2014: We tell the Board to stop saying statutes prevent local verification.
I appeared at the March GAB board meeting to offer a five-minute public comment, hoping to return the Board members' attention to the major issues. In January, the much-anticipated President's Commission on Election Administration had issued its final report, which includes a recommendation that reinforces what we were saying in our July 2013 report:
Recommendation: Audits of voting equipment must be conducted after each election, as part of a comprehensive audit program, and data concerning machine performance must be publicly disclosed in a common data format.
GAB staff told the Board that compliance with this recommendation "would require legislative change as the GAB currently audits only after the General Elections consistent with current statutes. Additional audits would increase costs for both the GAB and municipalities. The common data format would need to be specified."
When it came time for the Board to discuss this item, Judge Lamelas asked Mike Haas, who was presenting the agenda item, about the costs of the audits. Before Haas could answer, Kennedy interrupted to repeat his position that statutes do not allow ballot bags to be opened for verification (they are merely silent); that "our position is a full hand count cannot be performed" (without acknowledging that a full hand count is seldom necessary to verify the outcome); and that audits are 'expensive' (without providing any information about their cost.) Again, staff comments were not the slightest bit oriented toward problem-solving (i.e., "Here are the things clerks could do under current law to move closer to compliance with this standard..."). Instead, their guidance for the Board was oriented entirely toward a conclusion that better post-election verification is too expensive and time-consuming to consider.
I'm not sure right now what more follow-up we can or should do with GAB. I'll continue to attend their meetings and do what I can to keep the Board members' interest alive. As I said above, I am hoping the LAB report will reinforce our message and that of the Presidential Commission, and that we'll find at least a few clerks who are willing to experiment with various methods of verifying the accuracy of machine-counted vote tallies before they certify them as final election results.