Voter Registration: Modernization vs. Technology?

A recent New York Times editorial spoke in favor of changes to U.S. voter registration practices, citing a recent study from the Brennan Center for Justice at New York University School of Law. From a technology perspective, those changes raise the question of impact to existing I.T. systems for voter registration, both those is use and under construction, including the TTV Registrar. Will these systems become obsolete, or at least require a “fork-lift upgrade”?

Here’s the connection. Current VR systems automate processes around the current practice of citizens being required to request to be registered, and re-request every time they change address. From the NYT:

In the United States, the burden of registering falls squarely on voters. In countries where the government does more of the work, according to a new study, registration rates are much higher. … voter registration modernization bill … should follow the lead of nations that are far more serious than the United States about getting eligible voters on the rolls — and have the registration rates to prove it.

The changes under consideration are part of some in-process legislation in the U.S. Senate, being led by Senator Charles Schumer. These changes partly shift the burden to the government. There are many variations and details, but the bottom line is that VR would no longer have a primary focus on processing VR requests from citizens, which would be far less frequent than today.

That’s potentially big change to election administration, but would it be a big change to I.T. systems for election administration, especially a state’s Digital Voter Registration System (DVRS)? I don’t think so. I can only speak from experience in our project to build the TrustTheVote Registrar, an open source software system for states to use to implement a DVRS. The processing of citizen (re)registration requests is an important part of a DVRS, but it’s not actually the technically most significant part. A DVRS is, essentially, a data management application for a voter database, with relatively low complexity of the data, relatively few direct transactions on the VRDB, but several potentially hairy data interchange functions with other systems, such as those of the state DMV, corrections agency, change-of-address and dead persons registries, etc. The DMV, in particular, is an indirect source of citzen VR requests; a DVRS must obtain from a DMV system the notifications of motor-voter transactions, and convert each one into a voter registration request or change-of-address transaction on the VRDB.

What would VR modernization add to a DVRS? More of those potentially hairy interchange functions, with additional external systems that contain information that would help detect VR-relevant changes in a citizen status, such as turning 18, or changing address, or becoming naturalized. So why do I say that’s not a big change? But DVRSs already need to “get it right” for current data interchange requirements. If a DVRS already meets that challenge, then VR modernization would require the addition of more data feeds that can deliver more instances of existing transactions on its VRDB.

So the key is flexibility, and there I must toot the twin horns of open source and agile system development, two of the 3 pillars of the TTV Registrar. Agile development can yield systems that are flexible for extensibility of existing features like data interchange. Open source means that a government IT shop, using an open source DVRS, would not have cost or lock-in barriers to implementing extensions driven by VR modernization. And for VR modernization, the 3rd pillar becomes even more important: operational transparency. With the increased pace of VR (re)requests, increasingly automated, the public may have just as much interest in seeing what a more automated VR process is doing, as it does today in seeing how human operators use or abuse a DVRS. In many areas, some Americans are not so trusting of election officials to conduct voter registration processes with high integrity; with modernization and more automation, the people may not feel any more confident in the correctness of computers’ operation. But with modernized VR processes, and modernized DVRSs, transparency features can help with public trust and accountability on both areas.

— EJS