Internet Voting can Reform Democracy
Senior Fellow, RIETI
A political season seems to have set in once again in Japan with Prime Minister Junichiro Koizumi expected to dissolve the House of Representatives on Oct. 10 and call a snap general election for Nov. 9. Not many people, however, are aware that the existing national election system has a critical defect. The problem I am referring to here is neither about the single-seat constituency system nor about the system of proportional representation. Instead, I am talking about the "handwritten" voting system under which each voter has to write down the name of the candidate of their choice by hand. Voting slips, upon which a candidate's name as been penciled, must be sorted out and counted. Vote counting, which begins immediately after polls close, often continues all night as the presence of numerous questionable or invalid votes tends to cause delays in work. Japan is the only industrialized country to maintain such an old-fashioned voting system. When Florida's punch-card voting system became infamous after thousands of invalid votes were found in the 2000 presidential election, I told my American friend that things are far worse in Japan. Bursting into laughter, the friend replied that Florida might as well export its punch card machines to Japan.
From Electronic Voting to Internet Voting
In fact, the handwritten voting system was abolished once. The 1994 revision of the Public Offices Election Law replaced it with a "select-by-symbol" system, under which voters were to choose the candidate by selecting the corresponding symbol. In the very next year, however, legislators introduced and enacted another revision to the law, which thereby restored the handwritten voting system. Thus, the short-lived select-by-symbol system was abolished without getting a single chance to be tried out. The official reason for the reversion was an expected increase in the number of candidates running for the House of Councilors from proportional representation constituencies. The true reason, however, was probably that the incumbent legislators - whose names were known better by the public than those of non-incumbent candidates - found greater advantage in the handwritten voting system.
Meanwhile, in local elections, six municipalities have hitherto introduced electronic voting. Specifically, voters select the candidate of their choice on a touch panel installed at each polling station, instead of writing down his or her name on a ballot paper. Counting of votes can be completed in 10 to 20 minutes. So far, no big problems have been reported. Such an electronic voting system has been implemented in many other countries and Japan is reportedly considering introducing one in the House of Councilors elections. If realized, it would help streamline vote-counting procedures but there would be little impact in terms of voter turnout and election results.
Many other countries are taking a step beyond that, studying the possibility of allowing voting over the Internet. In the United States, an Internet absentee voting system has been developed under the initiative of the Department of Defense and will be adopted in 10 states for the 2004 presidential elections. With the introduction of the system, not only U.S. servicemen stationed overseas but also the disabled and others who have difficulty in reaching a polling station would be able to cast a ballot. For the time being, Internet voting will remain an auxiliary system that is restricted only to absentee voting. Once the reliability of the system is proved, however, ordinary voters would eventually become able to vote over the Internet.
Let's try it out in Local Elections
Yet, a great deal of opposition to the introduction of an Internet voting system does exist. Concerns have been voiced over the possibility of various acts of fraudulence occurring such as data falsifications, double voting, and spoofing. Some people also question how their anonymity would be ensured and what to do in the event of hacking. Similar arguments were heard in the early stages of electronic commerce, but are no longer. These problems have technical solutions. Indeed, in September this year VeriSign Inc. announced a digital certificate system for the Department of Defense's Internet voting project.
Some people say that an Internet voting system should not be introduced until "absolute security" is guaranteed, because the level of security required for a voting system is far higher than for electronic commerce. But such an argument fails to see another aspect of the problem surrounding the Japanese voting system. Voter turnout at the 2001 House of Councilors elections stood at 56% and therefore the results failed to reflect the will of nearly half of the eligible voters. In contrast, turnout at the 2000 Democrat primary elections in the state of Arizona, which adopted an Internet voting system, was 93%. Internet voting not only rationalizes vote-counting procedures but also increases turnout and may change the results of elections.
When voters are able to cast their ballot via the Internet, the turnout is expected to go up substantially, especially among young and/or urban voters - whose turnout has been characteristically low - and the results of elections would come to reflect the national consensus more fully. There is no network security that is "absolutely safe" and neither is there the need to pursue such perfectness. What is important is to compare and balance the merit of providing several tens of millions of people with an easier opportunity to vote against possible problems that may (or may not) be caused by introducing an Internet voting system. Of course, people who feel uneasy about voting via the Internet can go to the designated polling station and cast a ballot electronically there.
A declining turnout is a phenomenon commonly observed in industrialized countries, and there is a good reason for this. The marginal profit of voting is zero, because a single vote cannot change the results of elections. On the other hand, the cost of going to the polling station and sacrificing one's holiday is to be born by each voter. Therefore, it is quite rational for a voter to abstain from voting. However, if such "free riders" - those who choose not to bear the cost of voting - increase too much, it would distort the results of elections. To solve this paradox, reducing the "cost" of casting a ballot by enabling voters to do so via the Internet would be far more effective than staging a "let's-go-and-vote" campaign.
Administration is not a unique service; and therefore, the idea of providing administrative services online is not fundamentally different from that of providing services for electronic commerce. The essence is to provide high-quality services at a low cost, disclose all possible risks involved, and then let customers (the public) choose by themselves. Internet voting is important not only as a means to have election results reflect better the will of the general public, but also as a tool to turn administration into an "ordinary service." In Japan, since municipalities are free to choose their own system for local elections, it may be a good idea for them to allow, on an experimental basis, absentee voters to cast their ballot over the Internet.
A Security Analysis of the Secure Electronic Registration and Voting Experiment (SERVE)
October 7, 2003
Article(s) by this author
January 13, 2004［Column］
November 21, 2003［RIETI Policy Debate］
October 7, 2003［Column］
June 24, 2003［Column］
May 20, 2003［Keizai Sangyo Journal］