Wednesday, May 27, 2009

Open Letter to the New York State Board of Elections -- Election Audits Must Be Simple But Effective

A fellow election integrity advocate suggested I post the following as an open letter. It was sent to the NY State Board of Elections (SBoE) on April 20, 2009, a week before they were to meet with the county boards at a conference in Albany. I was told that auditing of elections was going to be on the agenda. Having been involved in the drafting of the relevant regulations, I was prompted to write this note. However, I have heard reports from conference attendees that election auditing was not actually discussed.

Note that the phrase "firm Constitutional ground" in this letter means that Equal Protection derives from an equal chance of detecting incorrect electoral outcomes of contests for the same elected offices in different parts of the State. I did not mean to imply that the use of concealed vote counting would necessarily be constitutional in the first place.

Election Audits Must Be Simple But Effective

Dear Co-Executive Dirs., Co-Chair Kellner and Commissioner Peterson

Thank you for correctly stating in your Apr. 8th letter to the counties that "under some circumstances ... safeguards will possibly escalate what will be required to include a full hand count."

The problem is, without an initial risk-limiting audit sample, there is no way to know whether such an escalation, or any escalation, should or should not occur. The solution can be relatively simple, but NOT as simple as the 3% "spot check" of machines or systems required by EL § 9-211.

In fact, a one-size-fits-all, percentage-based audit would violate Equal Protection by resulting in unequal chances of discovering material discrepancies that could change the outcomes of contests for the same office (e.g., State Senator) in different legislative districts throughout the State. Since the math underlying this is unequivocal, fixed-percentage audits should not be allowed to stand for this reason alone, except as an arbitrary minimum or "floor."

I believe you will be on firm Constitutional ground by implementing a risk-limiting approach, now endorsed by national groups such as LWV and NY good government groups such as NYPIRG.

The only remaining problem is: how to make it simple.

I've spent a great deal of time on this problem, pro bono, with the help of Dr. Mark Lindeman at Bard College and Dr. Ron Rivest and his colleagues at MIT and Northeastern U. The process of simplification has taken years, but I believe we are at a point now where we can offer NY a very simple but effective method, based on the maximum size of EDs, which I hope will remain at 1,150 active registered voters (but could be adjusted if necessary).

Frankly, I was hoping that by now, we would have received some feedback and possibly an endorsement of this approach from the SBoE. Time may be running out for NY and there is no reason for additional delays in light of the new method I have proposed to Commissioner Kellner, Bob Brehm and Kim Galvin, to simplify the risk-limiting audits.

Please advise when you are ready to discuss this further. You have to present something tenable to the counties in only a few days, and I'd hate to see such an opportunity squandered.

Best regards and thanks for doing your best to protect the franchise,

Howard Stanislevic
Founder, E-Voter Education Project

Friday, May 22, 2009

New York Rolls Out Uncertified Voting Systems for 2009 Elections

ALBANY -- At a May 12th Commissioners' meeting, after collaborating with the US Dept. of Justice, the New York State Board of Elections cavalierly decided to risk the disenfranchisement of nearly a million of the state's voters, by allowing what one commissioner called a "huge pilot" of uncertified software-driven electronic vote-counting systems around the state in 45 of its 62 counties.

Here are the links to the Commissioners' resolution, and other documents containing the details of the plan:

Over 900,000 voters (read: guinea pigs) could be affected by these irresponsible tests, which one county election commissioner, perhaps unwittingly, compared to filling out lottery tickets. Gambling with the votes of a million New Yorkers is hardly a way to instill public confidence.

The plan contains almost no provisions for manual recounts of the paper ballots to check the computer tallies, other than those that might be obtained through the courts. The only exceptions are for contests with a margin of victory of 1% or less. Full recounts of those contests will be conducted, but we bristle at the suggestion that the victory margin reported by the uncertified voting system will be the one used to determine whether or not the hand count to check the system will take place.

Commissioner Douglas Kellner made a motion at the May 12th meeting to allow any candidate to ask for and obtain a full hand recount. His fellow commissioners defeated it by a bipartisan 3 to 1 vote.

Apparently Kellner's colleagues believe that:
  • any candidate can convince a judge that a voting machine didn't count her votes -- even without evidence to support such a claim;

  • the judge will also believe that the paper ballots have been preserved inviolate and thereby allow them to be hand counted to find out who really won an election (contrary to a number of previous decisions by the highest court in the State -- not to mention the highest court in the land); and

  • the court would gladly spend taxpayers' money for such high-minded purposes as convincing losers of elections, and their supporters, that they really lost fair and square -- even given the amount of money already spent on the new voting systems.
But the voters of New York deserve more than just naive speculation about the ease of obtaining hand counts from a potentially partisan and cost-conscious judiciary. They deserve the actual hand counts if and when they are needed.

So what if the margin of victory happens to be slightly more than 1% (say 1% + 1 vote for example), and the courts deny the recount request? In that case the hand count reverts to only a 3% spot check, per Election Law § 9-211 -- part of the Election Reform and Modernization Act of 2005 that brought us this mess in the first place. If there are no discrepancies found in the spot check, the election could be certified -- which is more than can be said for the voting systems that actually produced the election results.

Unfortunately, the math is unequivocal: in many elections, a 3% audit can reveal absolutely NO discrepancies, and the outcome of the elections can still be absolutely wrong. If that happens, no one will be the wiser.

One other potential safeguard remains for the million voters who will be subjected to this foolish experiment: the long-awaited and yet to be promulgated State Board of Elections auditing regulations known as Part 6210.18. For well over a year now, we have been involved in the drafting of these regulations. They offer the only hope for anything better than the ill-considered 3% spot checks in the Election Law.

A year ago, many of New York's good government groups wrote to the Board, asking for these regulations to reflect best practices. But so far, progress has been slow to non-existent in this area, even as the mad rush to run real elections using potentially fake voting systems continues.

All that said, while the value of certification has been greatly exaggerated, we think it might be fair to say that if done properly, certification can prove that a voting system can work -- not that it actually will work. This weak assurance is of course not sufficient, but it's better than no assurance at all. The only way to be sure to prevent the disenfranchisement of New York's voters by untrustworthy computers, is to hand count 100%, at least until the systems are certified.

We'll be following this story and reporting on efforts to fix this latest debacle and avoid the Floridization of New York's elections. We don't think this is what New Yorkers signed up for when the State accepted $50-million in federal funds to replace its lever voting machines under the Help America Vote Act of 2002 (HAVA). The State has not been able to certify a voting system to replace the levers since then, and as always, it's important to read the law first.

In New York, the law has been decidedly anti-recount. So paper ballots or no paper ballots, the software counts will rule -- just as they did in Florida's 2000 election which brought us Bush v. Gore and ultimately, HAVA itself. Ironic, huh?

The Board did agree to present the matter to the State's Citizens' Election Modernization Advisory Committee. While their opinions are as yet unknown, and probably not binding, at least one member of the group has gone on record as favoring 100% hand counts of ALL votes counted by ANY uncertified voting system. This is in direct conflict with 3 out of the 4 State Board of Elections Commissioners who represent our two major political parties, but unfortunately may no longer be faithfully representing the voters of New York.