January, 2014. MERA has developed a detailed legislative proposal for post-election audits.
The Michigan proposal was guided by the work of the National Audit Summit's State Audit Working Group, which refined principles and best practices for post-election audits from expert opinion and experience across the country.
Central to the MERA proposal are statistical or "risk-based" based audits, which are the most efficient and effective way to confirm election night results. The statistical audit would begin right after the election and if its findings showed a different winner of a contest, then the audit results would bind election officials to certify a corrected result.
In the proposed audits, precincts are selected randomly and the paper ballots are counted manually (hand to eye) and then compared to machine tallies. To insure independence, audit teams are selected for each county by a central audit board, under the authority of the Michigan State Treasurer.
A novel feature of the MERA proposal is an election night audit in each precinct. One contest is selected at random and hand counted to detect significant errors in the performance of the precinct's tabulator.
Following a Minnesota law (see Report Appendix 1, Subd. 4 and 8), the proposal creates a strong market incentive for vendors of electronic voting equipment to make sure their equipment functions correctly and to eliminate security vulnerabilities. The vendor of any brand of electronic equipment that fails an audit would be penalized significantly.
Strong transparency provisions ensure public oversight of the entire audit process, from setting standards and designing the statistical methods to final reporting. Other provisions permit challenger groups to conduct hand count audits under the supervision of election officials, and encourage candidates, parties, issue committees and others to initiate selective and targeted audits.
There is a striking final provision in the proposal. A new election would be mandated for any audited contest if the State Vote Audit Board cannot determine an outcome of the contest with, in its judgment, a reasonably high level of probability.
Taken as a full package, the proposal would provide Michigan with the highest level of assurance of election accuracy of any state in the country.
MERA's proposed bill is here: Post-Election Audits in Michigan.
The actual voting arrangements managed by Michigan election officials consist of many elements -- such as policies, roles, voting machines, staff training, safeguards and procedures, budgets and supplies, and so on. Traditionally such elements were integrated into a functional system that had developed gradually over many years. The traditional system produced reasonably reliable, efficient, and accurate vote counts that were open to public scrutiny and trustworthy. Now under the Help America Vote Act new voting systems based on electronic vote tabulation have been put in place.
In Michigan use of voter verified paper ballots is required by law. During voting, the paper ballots are fed into and read by optical scan electronic tabulators. When the polls have closed, the results are passed on in the form of memory cards or tapes to a central location in each county where the votes are aggregated and results declared. For each voting precinct, tabulators are programmed in advance of the election to count the votes, and the tabulators are then pretested before election day to verify the accuracy of the programming. However, this system was established in a relatively short time and, particularly with respect to security from electronic tampering, has not been adequately understood, tested, or refined. Traditional practices to safeguard an honest, accurate vote count are no longer effective with new electronic voting machines.
In 2006 the outspoken republican Oakland County Clerk Ruth Johnson (who has now replaced Land as Secretary of State) sent then Michigan Secretary of State Terri Lynn Land an important letter which documented a shocking list of problems with the operation of the new electronic voting machines in Oakland County.
Operational reliability, however, is only part of the problem. The trustworthiness of electronic vote tabulations is no better that the security for the tabulating machines themselves, including their programming and the memory cards & tapes. Security was never a criterion for certification! So now election officials are faced with the disturbing possibility that what appears to be correct in pre-testing can be altered through either direct or remote access: machine programs can have hidden instructions activated by the machine's internal clock, memory cards can be switched (perhaps in the guise of vendor "upgrades"), internal machine records of illicit manipulations can be erased or never recorded in the first place, and so forth. For a full account please see the Brennan Center report: Analysis of security vulnerabilities in the three most commonly purchased electronic voting systems.
Now we are in a period of danger. Comprehensive, integrated, functional security policies and procedures appropriate to the new technology, and to the officials and voters who use it, have not been created. The current transitional arrangements are inadequate and vulnerable to breakdown and tampering. The goal of fair, efficient, and accurate vote counts that are transparent to a concerned public is still far away.
How can voters know their votes will count and be counted accurately? If nothing is done to address this concern, then we can't and won't be able to know. We continue to face the possibility of extensive breakdowns and malfunctions on election day. Many completely legitimate voters could be prevented from voting except on "provisional ballots," which may or may not be counted. With the design and operating system of machines held as proprietary secrets by Vendors, even an independent, expert computer programmer can not assure the public that the machines will be secure and reliable on election day. In short, without concerted action, the outcome of any given election could be spoiled or hijacked and we would have no way to know and little or no recourse.
MERA has formulated an extensive
Legislative Plan and a fully vetted and detailed plan for post-election audits in Michigan. These measures provide the basis for overcoming the security problems of Michigan's electronic
voting systems, while increasing access to voting. If we are to have a system that is secure
and transparent to the concerned public, then it is urgent that we make these reforms soon as possible.