IT's role in improving elections

Three of the five primary recommendations of a bipartisan report on election reform are IT-related. This shouldn't be too surprising: like any other business process, voting has a heavy IT component

Former President Jimmy Carter and former Secretary of State James Baker released the final version of a bipartisan election reform report today. The two led a 21-member, privately funded Commission that offered 87 recommendations to strengthen the country's electoral system and build confidence in the political process. Of the five primary recommendations, three have implications for government IT.

The Help America Vote Act (HAVA) required states to consolidate voter registration systems. (Disclosure: I'm a consultant to one of the companies doing this work, Aradyme.) Pre-HAVA, most states allowed voter registration data to be maintained at the county level. Of course, each county used different systems, formats, and even kept different data.

The Carter-Baker report recommends a national voter registration system that ties these varied registration systems together.

[A]dditionally we propose that the U.S. Election Assistance Commission (EAC) develop a mechanism to connect all state lists. These top-down and interoperable registrations will, if implemented successfully, eliminate the vast majority of complaints currently leveled against the decision system. States will retain control over their registration list, but a distributed database can remove interstate duplication and help states to maintain up-to-data, fully accurate registration list. This would mean people would need to register only once in their lifetime, and it would be easy to update their registration information when they move.

I suspect there will be significant pushback on this idea, not from a technical perspective but from a political one. There are no "national" elections in the U.S. There's no federal infrastructure set up to manage elections. It's all done by the states and in most cases, counties within those states. Governors and others will fight this first step toward federalizing the election infrastructure.

A second recommendation that has IT implications is the proposal to use Real ID cards as the sole means of authenticating voters.

[T]o make sure that the person arriving at a polling site is the same one who is named on the list, we propose a uniform system of voter registration based on the "REAL ID card" or a equivalent for people without a driver's license.

While Real ID generally has implications for IT, this proposal just adds to that requirement list. There are, as usual, privacy concerns. Most people don't realize that information about whether they voted or not and in which elections, along with their party affiliation, is a public record in most jurisdictions. This proposal would make that data much more accurate, linkable and transferable between states than it has been in the past.

A final IT-related recommendation concerns voter confidence in the elections system.

[W]e propose ways to give confidence to voters using electronic voting machines that their votes will be counted accurately. We call for auditable backup on paper at this time, but we recognize the possibility of alternative technologies to audit those machines in the future. We encourage independent testing of voting systems (to include voting machines and software source code) under EAC supervision.

This recommendation could lead to federal certification of voting machines, something that has been squarely a state function until now. Having been involved in arguing for auditable voting systems for Utah, I know how sticky this issue can become. Unfortunately, the debate is far too often an emotional one rather than technical. The matter is complicated by the fact that the people ultimately making the decision don't understand the technology and usually don't sufficiently appreciate the subtleties of computer security to feel the need to turn to experts. A federal panel on this issue would probably not fall prey to these same problems.

In all, this report is an interesting look at the U.S. voting system. We shouldn't be surprised that many of the recommendations have IT components--that's true of any business process. The proposals to federalize more of the elections infrastructure will cause considerable distress and, I suspect, be the cause of its undoing if the proposals make their way to Congress.