Supreme Court Revives Controversial Citizenship Data System

The Customs House at the Polling Station

The system is described as a safeguard: a federal data system that checks voter rolls against citizenship records so that only citizens vote. The mechanism it creates is a filter whose errors cost its operators nothing and cost the filtered everything. That gap between description and mechanism is where this analysis lives, and the Supreme Court of the United States has now reopened it.

I spent part of my life studying customs houses. Suppose an officer is praised for every cask he seizes and never pays for seizing an honest one. He does not become corrupt. He becomes enthusiastic. Any cask with a smudged label looks like contraband to him, because the smudge promises him a commendation and costs him nothing. His zeal comes from how he is paid, not from bad character, and the honest merchant whose wine sours on the quay has no one to invoice.

A citizenship filter on voter rolls has the same structure. When it flags a name, whoever runs it can announce a catch. When the flag is wrong, the cost falls on a voter who must now prove a fact the government already recorded, often at the time and place the government chooses. The operator gets credit for every flag and pays nothing for a wrong one. No rational officer under that arrangement will err toward caution.

The strongest opposing argument deserves a fair statement. The United States has a legitimate interest in confining the franchise to citizens. A central database is arguably more consistent than a county clerk’s guesswork or a neighbour’s accusation. And if non-citizen voting is rare, the defenders say, the filter will simply find little and trouble few. I grant the end completely. My objection is to the mechanism.

Records of citizenship are snapshots taken for other purposes, and people’s circumstances move while the records stand still. A man recorded as a lawful resident in one year may take his oath of naturalization the next. The database remembers the first fact more reliably than the second. So if non-citizen voting is rare, as is widely argued, the filter’s flags will be made up disproportionately of its own mistakes. A net cast into a pond with few fish brings up mostly weeds. The rarer the offence, the larger the share of accused who are innocent.

Now apply the suspicion I keep for merchants. When a trade petitions for protection, it speaks the language of the nation. When officials or campaigns petition for a filter, they speak the language of electoral integrity. The rhetoric may be sincere. The question is who benefits from the mechanism. The claim that non-citizens vote in meaningful numbers is disputed, yet the filter’s value to its promoters does not depend on the claim being true. A flag is a headline whether or not it survives review, and the correction, if one comes, arrives weeks later and draws no press conference.

Then check the competition. An election is a market in which candidates compete for voters. When the officials who oversee that market also control a filter that removes buyers from it, and the filter’s errors are not randomly distributed, you have created a way to shape the market while claiming only to police it. Naturalized citizens, people who moved recently, and people whose names are spelled one way on one form and another way on the next will be over-represented among the flagged. Such a group is not politically neutral, and everyone involved knows it. I would call this the regulation of a competition by one of its competitors.

Consider the woman the system actually meets. She stood in a federal courtroom years ago, raised her hand, and swore the oath. The certificate lives in a plastic sleeve in a kitchen drawer. A letter tells her that her registration is in question, and she takes an unpaid morning off work to carry the certificate to the county office. The clerk, who has no quarrel with her, types her name into a screen that still describes the person she was before that morning in court. The machine has no opinion of her. It has only an old record, and it sends no letter to the officer who ran the search.

A remedy would not abolish the check. It would make the operator pay for the error: independent verification before any removal, a burden of proof that stays with the state, and published counts of flags that turned out to be false. Constrain the officer and the zeal subsides. Leave him unconstrained and he will go on congratulating himself on the quay.

Her paperwork is in order. It was in order the first time.