An Economic Model for Bad Ballots

If voters could select the machines, there would be a president-elect by now.

The New York Times, "Economic Scene" , November 30, 2000

For three weeks, the question has been nagging Americans: how can the most technologically advanced nation on earth have such antiquated, unreliable, confusing voting technology? How can the outcome of the presidential race dangle with the chads left on 1975-style punch cards?

Tapping this frustration, two of the Senate's most publicity-savvy members, Arlen Specter, the Pennsylvania Republican, and Charles E. Schumer, the New York Democrat, called for federal studies of voting procedures. In his announcement, Mr. Schumer labeled the current system "antediluvian."

Even Amazon.com jumped into the game, poking fun at the infamous "butterfly ballot" used in Palm Beach County, Fla. The company posted a parody page that claims to provide a new way to navigate the Amazon site. The mock-up lists the company's various shops in the two-column butterfly format, with links in a row of buttons down the middle. (The spoof is no longer on Amazon but can be viewed here.) The links work, so, unlike Palm Beach voters, you know for sure whether you've punched the wrong hole. Amazon's conclusion: "Ease of use -- it matters!"

Ease of use, however, was exactly what Theresa LePore, Palm Beach County's elections supervisor, was trying to get when she created the butterfly ballot. She redesigned the ballot to make the type larger and more readable for her elderly constituents, "so they wouldn't have to squint." The butterfly ballot wasn't antediluvian technology; it was a bad innovation.

The real puzzle is why it's so hard to sort good balloting ideas from bad ones and, consequently, why we don't see the sort of incremental progress we're used to in the rest of society. After all, Americans have been voting for a long time. We've had plenty of opportunities to discover problems. Yet unlike the Amazon site, FedEx mailing forms, A.T.M.'s or state lottery cards, ballots don't become much friendlier or more reliable over time. Bad ballot designs and clunky election processes survive where hard-to-navigate Web sites and poorly run restaurants never could.

One reason is what economists call the principal-agent problem. Voters -- the owners, or principals -- don't choose the voting machinery. Like shareholders of a corporation, voters delegate day-to-day management to professionals, or agents.

But these officials don't necessarily want the same things voters want. Managers may not feel like redesigning vote-counting procedures to adapt to easier-to-use technology. Or they may want to buy the latest whiz-bang gadget because it's cool. They may prefer to spend extra dollars to improve their offices or hire friends. They may ignore complaints, or procrastinate in dealing with problems. This isn't corruption. It's just human nature -- as true of corporate managers as it is of election officials.

There's a vast economic literature on principal-agent issues, much of it devoted to theoretical models of how to align employee compensation with principals' goals. In practice, the carrot for corporate managers is usually a plan that ties their pay to how well the stock price is doing. The stick is the threat of a hostile takeover that will boot them out or, before things get to that, a rebellion by large shareholders.

These real-world incentives depend on competitive financial markets and a "market for corporate control," as Henry Manne, a law and economics scholar, termed the takeover process in a seminal 1965 article. If shareholders don't think their agents are doing a good job, they'll sell their stock, lowering the stock price. If enough shareholders do that, the company will be vulnerable to a takeover. Without such competition, it would be much harder to persuade managers to care about what owners think.

The fundamental problem with the ballot-process "market" is that there isn't enough competition. The agents operate pretty much unchecked by the principals. These officials may be elected, but what they do rarely gets much scrutiny. Except during a rare crisis, voters don't have much information about alternatives. And it's not worth the typical voter's time to investigate. The ballot-process market has no equivalents to institutional investors or hostile takeovers.

There's another vital difference: Since money is fungible, investors generally want the same thing -- higher financial returns. But voters are like customers, with many different preferences. Some want to vote at home, while others enjoy the communal feeling of a polling place. Some like computers, while others fear them. Different people find different ballot formats appealing. Despite this diversity, everyone in a precinct gets the same monopoly service.

Voters do generally agree on a few standards. They want accurate counts, secret ballots and a record that can be audited. They want to make sure all voters are legal and no one votes twice. Those aren't so different from the requirements that bank customers expect from A.T.M.'s. But not all A.T.M.'s use the same interface, and bank customers can also use human tellers, the telephone, the mail or the Internet. There's no "one best way" for all bank customers. Adhering to the same basic public and accounting regulations, banks compete, and even the same bank usually offers a choice.

The result is an open-ended process that pushes banks to find ways to make their procedures more error-proof and customer-friendly. An A.T.M. today is a lot more dependable than an A.T.M. in 1980.

Over the last few weeks, people who believe that progress means standardization have mocked the decentralization of America's election process. But the diversity of our system is the reason Mr. Schumer could cite several promising experiments from around the country.

And it's the reason not every voter was stuck with butterfly ballots or cranky ballot scanners. Picking a single standard can easily mean picking a bad one. It almost certainly means picking a static, inflexible standard. In the 1970's, after all, punch card ballots were cutting-edge technology.

Improving the ballot process doesn't mean making voting machines literally look like A.T.M.'s. It means learning from the competitive process that gave us A.T.M.'s. If we think about the problem that way, we'll seek to regulate only those essential aspects that all voters care about, while allowing diverse alternatives to compete in attracting and pleasing voters.

It may be hard to imagine a world where polling places and polling technologies compete. But a month ago it was hard to imagine the world we have now.