Every applicant for US citizenship is required to take a test, and every question on the current test is drawn from the same master list of 100 questions. Many are eminently reasonable (“Who is the President of the United States today?”); others are somewhat trivial (“Who said, ‘Give me liberty or give me death’?”). One reads like a discussion question from a fifth-grade history book, the kind with no right answer: “What is the most important right granted to U.S. citizens?”

According to the US Citizenship and Immigration Services, there is a right answer–the “right to vote.” And though the Constitution contains no affirmative right to vote (only barriers to disenfranchisement on the grounds of factors like race or sex), this is hardly an incoherent response. Suffrage is obviously crucial to democracy. However, the question and its presence on the citizenship test implies more than that–a nation with a deep, practical and well-thought-out commitment to suffrage. But the United States is no such nation. Its deficiencies on this front are complex and bewildering; a more honest and interesting exam question would be, “Do all citizens truly enjoy the right to vote?”

In Stealing Democracy, Spencer Overton, a law professor at George Washington University, invokes the metaphor of “the matrix” to convey how the electoral landscape is dominated by a “collection of ever-changing rules and practices employed by various partisans and bureaucrats.” It plays host to trickeries that “often exclude particular voters, enhance the power of certain politicians, and advance specific policy preferences.” Overton’s fervent hope is that we can remake the matrix “so that it more fairly empowers all voters rather than simply privileging the insiders who know how to manipulate it.”

The metaphor (a reference to the sci-fi flick The Matrix) is a touch corny, but overall Stealing Democracy lends argumentative and evidentiary rigor to liberal conventional wisdom on a host of electoral matters: voter ID laws (bad), multilingual ballots (good), electronic voting machines (bad, so far), enfranchisement of released felons (good, as agreed upon by 80 percent of Americans and every democracy except Armenia). This is, of course, valuable work. Of broader value, however, is the instructive mixture of pessimism, pragmatism and hope with which Overton approaches democratic reform. Stealing Democracy is no vague antistate (or antimatrix, or whatever) “power to the people” exhortation. Overton dreams big, but his ultimate recommendations are, given the scope of the problem, quite modest. They inspire without being vague or sweeping–and are all the more compelling for it.

Overton’s pessimism stems from his conviction that wherever elections take place, there is a nearly innate “propensity of self-interested politicians to enact rules that benefit themselves instead of the electorate.” Of course, there is no universal proof for such a premise, and that’s fine. Any intelligent discussion of election systems should proceed as if the premise might be true. Plus, Overton cites a flood of piecemeal evidence suggesting it is true, including California’s 2001 bipartisan redistricting scheme, which reduced its number of competitive US House, State Senate and State Assembly seats from forty-four to seventeen (out of 153); the recent redrawing of House districts nationwide such that 98 percent of incumbents retained their seats in 2004; the admission in 2003 by then-Alabama Republican Party chairman Marty Connors that his party is opposed to restoring released felons’ voting rights “because felons don’t tend to vote Republican.” Perhaps this is why elections in 75 percent of democracies are run or supervised by judges or independent officials, and why almost all countries with single-representative voting districts leave redistricting to independent commissions. Overton wishes we would join the democratic world.

He knows, however, that states’ rights are the foremost impediment to nationwide electoral reform. In most of the United States, duopolistic state legislatures draw voting districts (which invariably become safer for incumbents) and elected partisans oversee elections. States further depute election administration to county, city and town governments. Thus, by Overton’s count “there are 4,600 different election systems” in the country. Every factor–funding, staff, machines, oversight, ballot design, recount mechanisms–varies from state to state and county to county. This gives politicians’ tendency to game the vote plenty of bureaucratic cover, making it more difficult to notice and fight (which partially explains why and how other countries have instituted reforms while America has not). Without positive enshrinement in the Constitution or vigilant protection by nonpartisan enforcers, “an American’s right to vote and have the ballot counted depends on where he or she lives,” Overton says.

He sees few benefits to maximizing state and local control. He doesn’t put it this way, but “local tailoring” smacks much too strongly of “separate but equal.” Obviously, local responsiveness to citizens’ needs is a good thing, as is the ability of states to check federal abuse; Overton cautions against centralizing authority in a single election czar. But a state’s or district’s rights cannot include the right to allow (intentionally or otherwise) some votes to count less than others. Overton would like Congress to recognize this and legislate accordingly.

This is not a radical suggestion, constitutionally speaking: the Elections Clause, Spending Clause and several amendments (namely the Fourteenth, Fifteenth, Nineteenth and Twenty-fourth) variously empower the government to ensure truly democratic representation nationwide. So, at the very least, Congress can and should quadrennially budget the estimated $4 billion needed to underwrite basic fairness in locally administered elections. (That’s double what is currently spent by districts and the federal government combined, .24 percent of the Defense Department budget for 2006 and twelve days of fighting in Iraq.) Congress could also require states to legislate minimum standards of fairness (like uniform counting procedures) and accountability (like full election audits) before receiving funds. Congress has asserted such authority before: there’s the 1965 Voting Rights Act (VRA), the 1984 Voting Accessibility for the Elderly and Handicapped Act, the 1986 Uniformed and Overseas Citizens Absentee Voting Act, the 1993 National Voter Registration Act and the 2002 Help America Vote Act.

Unfortunately, as Overton sees it, state and local autonomy remain the sacred cows of American election law (see Bush v. Gore and, more recently, Crawford v. Marion County Election Board). Thus, Stealing Democracy expresses little hope for the sort of federal reform and centralization needed to maximize the equality of voting rights. Instead, Overton pragmatically commits much of his energy to insistently defending the Voting Rights Act as the last and most valuable line of federal defense against rampant disenfranchisement in the states. In his eyes the act is invaluable because it codifies, in Section 5, a compliance requirement (known as “preclearance”) designed to prevent state and local governments from sneaking back into exclusionary election practices. Section 5 originally applied to the seven states (Alabama, Alaska, Georgia, Louisiana, Mississippi, South Carolina and Virginia) and four localities (counties in Arizona, Hawaii, Idaho and North Carolina) with low voter turnouts and the most established histories of excluding voters of color via literacy and interpretation tests. These governments were ordered to submit all proposed changes in election practice to the Justice Department or a panel of three federal judges for review.

Since 1965 the VRA has been renewed four times, and Section 5 has been broadened to cover more states, localities and racial groups. The 1970 renewal extended the preclearance requirement to cover several districts in Northern states that had imposed new exclusionary voting tests after 1965, and the 1975 renewal recognized the need to protect Latino, Asian-American, American Indian and Native Alaskan votes. Section 5 review has blocked more than 1,000 discriminatory changes in election law practice. Overton sees it as the most important piece of voting law on the books, and he is perturbed by the increasing popularity of certain attacks on it.

Some politicians affected by Section 5 argue that it is unfair for them to be burdened with the hassle of preclearance as atonement for the long-ago sins of their racist predecessors. “We have repented, and we have reformed,” complained Representative Lynn Westmoreland, a Republican from Georgia, during the debate preceding Section 5’s 2006 renewal. This objection has little practical force, since any electoral district can apply to “bail out” permanently from the preclearance requirement after complying with the VRA for ten years. Anyway, preclearance is no backbreaking burden: Overton calculates that the average application process costs only $458, much less than the multimillion-dollar discrimination suits Section 5 deters.

Still, Overton fully supports moving past potentially “outdated stereotypes,” particularly about the South, when deciding which districts have problems with exclusion and deserve to be covered by Section 5. He proposes eight quantifiable signs that an area deserves federal scrutiny for exclusionary practices, including “Largest racial disparities in voter turnout” and “Most Voting Rights Act objections and claims per capita.” However, since these criteria would generally expand Section 5 coverage, they are unlikely to pacify critics like Westmoreland.

A related objection to Section 5 takes contemporary America to be so “colorblind” that fighting racially motivated voter exclusion is a misguided and impractical use of finite resources. This claim is obviously tendentious (to put it politely). Interestingly, it doesn’t matter much to Overton whether any partisan officials are actually racists. All that matters is the fact, supported with pages of studies and examples, that “physical appearance, socioeconomic factors such as housing segregation, and distinct voting trends make people of color particularly vulnerable targets for exclusion.” Put otherwise: all votes are equally worth protecting, but some remain demonstrably more vulnerable than others. Today’s cutthroat and sophisticated political operatives may or may not be “postrace,” but they are not colorblind. This is why, in the absence of fundamental federal reform, American suffrage needs Section 5.

Lurking behind the question of suffrage is the equally important question of expressiveness: how well voting captures the true will of a people. Imagine that someday all of Overton’s dreams come true: somehow all US citizens are able to register to vote and have their ballots counted. (All right: also imagine away the Electoral College and flawed primary processes.) Could we then stop worrying about the efficacy of voting, supposedly the most basic element of democracy?

We could not, thanks to a fundamental flaw with the very concept of plurality voting. This flaw, described by William Poundstone in Gaming the Vote as the “invisible hand misguiding the whole electoral process,” is familiar to any casual observer of American politics over the past decade: it is the phenomenon of vote-splitting between candidates in a winner-take-all election. In the United States, vote-splitting is manifested most often in the so-called spoiler effect, wherein a third-party candidate takes enough votes from a front-runner to influence an election. Political operatives know all about the spoiler effect and often support hopeless candidates (as Republican operatives supported Ralph Nader in 2004) to siphon votes away from their opponents. Of course, not all third-party candidates are spoilers, and it can be hard to ascertain when spoilage has occurred. Third-party candidates can draw supporters to polls who otherwise wouldn’t have voted.

But even without spoilers, plurality voting is defective. A 1948 RAND Corporation report by economist Kenneth Arrow presented mathematical proof that no voting system (including plurality) under which voters rank their preferences in elections among more than two candidates can guarantee logically fair and nonparadoxical results. “Arrow’s theorem,” explains Poundstone, “says that election outcomes can be decided by quirks of procedure as much as the voters’ authentic wishes.” The finding (which won Arrow the Nobel Prize in 1972) rattled the American intellectual community by implicitly questioning whether “democracy and individualism might in some deep way be inadequate.” Put otherwise: can voting ever accurately express the will of the people?

Most contemporary voting experts concede that all voting methods contain the possibility of paradoxically unfair outcomes but insist that these possibilities can be minimized. Testy and complex back-and-forths rage over which paradoxes are most troubling and which solutions best minimize them. The method with the strongest combination of academic and popular support is called Instant Runoff Voting. In IRV, voters rank multiple candidates (plurality voting limits a voter to ranking just one candidate as “first-place”). If no candidate wins a majority of first-place votes, a virtual runoff is held in which the candidate with the fewest first-place votes is dropped from the ballot, and his or her supporters’ second choices are distributed appropriately (any potential spoiler’s votes would be transferred to another candidate before spoilage could occur). This continues until a majority winner emerges. Had Florida used IRV in the 2000 presidential election, Al Gore would have won the state and the Electoral College if just 51 percent of Nader supporters placed Gore second. In addition, Nader and other third-party candidates might have scored significantly higher than they did, since IRV seems to eliminate the fear of “wasting your vote” on a probable loser.

IRV should not be confused with the systems of proportional representation at work in most European countries. Poundstone defines proportional representation as “the doctrine that political parties (or other groups) should be represented in a legislature in proportion to their size in the electorate.” Some countries facilitate this doctrine with a voting method called the single transferable vote (STV), in which voters rank every candidate, victory requires a certain number of votes and votes for the most unpopular candidates are transferred to those who remain. Transfer formulas vary, and it gets pretty complicated, but the point is: IRV employs the single transferable vote to elect just one candidate instead of a whole legislature. One can logically support proportional representation (as Poundstone and Overton hint that they do) without supporting IRV, and vice versa.

On the face of it, IRV seems like quite a radical reform. It prompts visions of voters striding to the polls to express themselves honestly, elections unfolding unspoiled, vibrant third parties growing in strength and major parties either courting broader constituencies or losing power. This intuitive appeal explains the support IRV has enjoyed from sources as varied as John McCain, Barack Obama, Howard Dean, the Green Party, the Libertarian Party, the Alaskan Republican Party, at least four state Democratic parties, Ralph Nader and New Yorker election reform guru Hendrik Hertzberg. It is used by Australia, Ireland, Malta, several professional organizations and a handful of American cities and college campuses.

Poundstone, however, is not convinced that IRV is the solution worth gunning for. He is troubled by its ability to produce a “winner-turns-loser” paradox in which ranking a candidate higher actually makes him or her lose by altering the order in which candidates are eliminated. Poundstone demonstrates this complicated truth by reimagining the 1991 Louisiana gubernatorial race under IRV:

Thirty-four percent of the voters are for [corrupt Democrat Edwin] Edwards, 32 percent are for [white supremacist David] Duke, and 27 percent are for [moderate Republican Buddy] Roemer…. Since Edwards does not have a majority, the lowest-ranking candidate will be eliminated. That’s Roemer. Roemer’s votes are transferred to the remaining two candidates, giving Edwards an easy victory over Duke…. Now say that Edwards decided to court the Duke vote just before the election. He gave speeches or ran ads or spread rumors, with the result that a few Duke voters (6 percent of the electorate) switched their votes to Edwards. They ranked Edwards first, ahead of Duke.   The resulting numbers are now 40 percent for Edwards, 26 percent for Duke, and 27 percent for Roemer. See what happens? Duke is now in third place. It’s Duke who is eliminated. The runoff is between Edwards and Roemer. Most of Duke’s archconservatives rank moderate Roemer ahead of liberal Edwards. Therefore, the Duke vote is transferred primarily to Roemer, who beats Edwards by up to 13 points. This outcome is in line with the perception that Roemer would have beaten Edwards in a two-way contest.  For the 6 percent who switched, voting for Edwards instead of Duke caused Edwards to lose. This is through-the-looking-glass politics. It is even crazier than the spoiler effect.

IRV has many flaws, including an inherent bias against a moderate running in a field of extremists, but the winner-turns-loser possibility (referred to by voting geeks as a nonmonotonic outcome) is the most unpalatable. It becomes more complicated and more unsettling when combined with the potential for voters to attempt to exploit the paradox by pretending to prefer candidates they really hate. As you might imagine, this can really foul things up.

According to the Center for Voting and Democracy (also known as FairVote), which campaigns heavily for IRV and other electoral reforms (and counts Hertzberg among its board members), winner-turns-loser outcomes are so unlikely that they are not worth worrying about. This is less than reassuring. It is also representative of FairVote’s off-putting tendency to act (at least on the subject of IRV) more like a strategically message-centric and obfuscation-prone political campaign than as a project of free inquiry.

A search of the FairVote website for “monotonicity” returns just two unique results (“winner-turns-loser” returns zero). One is a long-archived page explaining that “one statistical study found that if STV elections were to be held through the United Kingdom, a nonmonotonic result would occur less than once a century.” The one current page similarly insists that “Potential non-monotonicity…is irrelevant in practice and will not affect voter strategy.” To support this claim, FairVote cites one paper in the Journal of Politics. I read the paper; its conclusions are highly tentative, and the author writes that his argument “does not establish [IRV’s] general superiority.” Some open discussion with critics takes place in the comments section of the fairvote.org blog, but cherry-picking and repetitious obscurantism are common there too. As an institution, FairVote is obviously terrified of admitting possible holes in IRV’s armor.

FairVote’s position is further exemplified by its response to range voting, the method Poundstone ends up tentatively advocating. In range voting, voters don’t rank all candidates; they rate them on a scale (such as 1 to 10, with an X for “no opinion”), and the candidate with the highest average rating wins. (Almost all advocates of range voting support requiring a candidate to receive ratings from at least half of voters to win.) By giving up on ranking, range voting appears to avoid any troublingly paradoxical outcomes. FairVote’s website notes disparagingly that range voting “could cause the defeat of a candidate who was the favorite candidate of 51% of voters.” But this could happen only if another candidate received strong approval from a majority of voters. Rare outcomes like this would be no blow to democracy, only to the arguably antidemocratic practice of victory by minimal coalition.

Nonetheless, range voting has received next to no support in academia or the real world. The Center for Range Voting is little more than the pet project of former Temple University mathematician Warren Smith, without whom range voting would probably be unknown (even to Poundstone). The CRV website is ugly and hard to navigate but populated by sophisticated statistical analyses and thoughtful, often-damning responses to almost every claim made by FairVote and advocates of other voting methods. It feels like the work of a man who isn’t out to win but to get a complicated problem right. FairVote would serve democracy well by following that man’s lead. Of course, then it might end up having to agree with Poundstone that range voting certainly seems worth trying.

In the midst of pondering how many voters might be expected to vote honestly in a range-voting election, Poundstone (following Smith) offers the following synopsis of the textbook “voter paradox”: “If there is a one-in-a-billion chance of your vote’s deciding the election, and if it costs you one dollar in lost time or transportation expense to vote, then voting is irrational unless you place the value of your candidate’s winning at one billion dollars.”

He goes on to speculate briefly as to why “more-or-less rational people” vote anyway. Some, he suggests, have surely been “programmed to feel guilty when they don’t vote.” For the nonprogrammed, “voting is a symbolic or quasispiritual act. They vote to feel part of the political process, to be connected to something bigger than their own petty concerns. For whatever reason we vote, narrow self-interest plays little role.” Here Poundstone misses a valuable chance to point out that rational-choice theory sometimes fails to offer the richest framework for explaining human actions. He realizes that the purpose of voting probably isn’t to hope against odds that your vote will tip the race and thereby increase your personal utility, but he doesn’t offer any alternative.

A suggestion: voting is a radical act in which democratic free agents join a collective process; in doing so, they simultaneously surrender to that process and find richer definition as free individuals. If this is right, the purpose of Section 5 and voting system reform is to honor this radical surrender by making the relevant collective process more honest, expressive and, well, actually collective.

To the extent that this sounds hokey, I think it’s because we are unaccustomed to thinking this way about life in a democracy. Nonetheless, the possibility of doing so remains enticing. I think Spencer Overton realizes this and ultimately draws hope from it. In the conclusion of Stealing Democracy, he recalls one of his students asking him, “Why should we care?” Overton responds by sharing the stories of people who, despite their recognition that broad reform is generally impossible without a crisis (the Voting Rights Act was passed only in the outraged aftermath of Selma’s “Bloody Sunday”) and the fact that “life is overwhelming,” devote time to working toward incremental election reform, mostly at the state and local levels. As a result, they feel less “detached, alone, and shortchanged” than before. They realize that living in a democracy means more than having one vote, and they start defining their vote against a lived constellation of commitments. They feel, perhaps for the first time, actively American.