In a succinct and persuasive paper entitled “If I can shop and bank online, why can’t I vote online?” David Jefferson explains why online voting is unrealistic. A computer scientist by trade—and chair of Verified Voting’s board of directors—Jefferson elaborates technical differences between purchasing textbooks on and registering your vote at a .gov. The paper lends valuable substance to a popular debate.

The Online Voting Debate

Online voting promises policy gains. Voters overseas and in the military would enjoy more access to the “polls,” as would those who work multiple jobs, or lack motility. Allowing citizens to cast ballots also reduces the effort required to vote and eliminates a perennial problem—long lines at polling places. In those respects, online balloting would diminish costs and so increase voter turnout in populations that are only disaffected enough to forego a trip to the polls. These developments would have a disparate impact on disparate demographics.  Still, efficiency gains might be used to ensure that voters without personal computers are given the resources to reach a polling place or a library.

These familiar arguments elicit familiar responses. Most notably, many argue that the “cost” of voting is desirable. Cost ensures that the most informed cast the most ballots. But tabling that assertion—and its assumption that voting rights should be contingent on information, to which access is variable—allows for focus on feasibility.

Jefferson’s paper responds to an intuitive argument about whether online voting is feasible. The argument is that online voting should mimic the mechanisms used to facilitate online commerce. According to Jefferson, there are two logistical obstacles to using the online commerce model in the electoral context: 1) the online commerce model is not secure; and, 2) the logistics of online voting require a structurally distinct network.

Jefferson forcefully advances his first point by referencing ubiquitous fraud and its impact. Online merchants lose billions when third parties capture consumer information. Information can be captured through different mechanisms, like malware. Because malware infects a computer, operating internally, precautions such as secure connections present no obstacle. Notably, the costs associated with such fraud are socialized across the consumer base. One in a thousand transactions may be fraudulent, but resulting losses can be mitigated through other means, including higher prices. Such spreading is not feasible in the electoral context, where results turn on slight margins. Jefferson’s practical considerations can also be supplemented with theoretical varnish: if votes have intrinsic (non-instrumental) value, can efficiency gains and increased participation outweigh the cost of disregarding a voter’s preference? If electoral systems are animated by outcomes instead of civic expression, are the results devalued? Would voters become so disaffected by the process that turnout would remain low?

Even if voters are willing to tolerate the risks posed by fraud, Jefferson establishes that logistical problems leave online voting beyond reach. The logistical problems are numerous. First, the most effective means of addressing potential fraud or manipulation would offset the value of online voting. That means—requiring mail-in paper ballots, which would be subject to random audit—would impose the very costs that online voting is meant to eliminate. A second issue is the discoverability of fraud. Fraud in the commercial setting is ultimately identifiable by means including bookkeeping, receipts, and others. Those procedures are inapplicable in the electoral context, where ballots are secret. To that end, Jefferson says: “a voting transaction is irreversible.

Secrecy also underlies other logistical issues. An online voting system would have to address basic election security concerns. How would eligibility requirements be enforced? How would proxy-voting—when, for example, a spouse votes for both members of the family—be prevented? And what about double voting? The expense of ensuring that online voters do not cast ballots at the polls or by absentee mechanisms might be significant.  Jefferson aptly aggregates the prevailing concern, providing an appropriate way to encapsulate the obstacles to online voting:

As of now there is no reliable infrastructure in place to verify over the Internet the actual identity of a person sitting at a PC or holding a mobile device.