The Third Shelf

Ranked choice voting and the many-sided republic Madison designed

There’s a gas station on a county road I drive on Saturdays, and its cooler tells a story. Two colas own the glass, red on the left and blue on the right, stacked four deep. Down at the bottom, behind a door that sticks, sits a single row of birch beer from a bottler two counties over. The clerk tells me it sells fine when people find it. On a typical day they don’t look.

We’ve run our politics like that cooler for a long time, and now we act puzzled that the country tastes like two flavors of argument. For roughly two hundred years we’ve trusted a head-to-head contest between two teams to drive our growth, the same way the colas drive each other’s ad budgets. That rivalry built a great deal. It also trained us to see a neighbor across the aisle as a rival brand. Then we wonder why the culture feels split down the middle, each half working to beat the other.

It’s tempting to blame Madison. He built friction into the framework on purpose. In Federalist 51 he wrote that “ambition must be made to counteract ambition,” and he trusted competing interests to guard liberty better than any one group’s good intentions could. Read that way, he may have wired a self-destruct button into the Constitution.

Read Federalist 10 again, though, and the design looks different. Madison wanted a great many factions. “Extend the sphere,” he wrote, “and you take in a greater variety of parties and interests.” Farmers and merchants, creditors and debtors, coast and backcountry, a dozen faiths. No single faction could seize the ground because too many others held a piece of it.

Picture a canvas tent staked out with guy-lines on every side. Each line pulls against the tent, and the pulling holds it upright. Washington warned against the spirit of party in his farewell address for this reason. The founders’ fear was a tent with two lines.

Two lines is what we got, and the cause sits more in our election mechanics than in the Constitution. Winner-take-all, single-seat districts reward whoever gathers the largest pile, so citizens learn to stack their votes behind one of two contenders. Political scientists call the pattern Duverger’s law. Madison’s ring of guy-lines got gathered into two ropes, and the tent became a tug-of-war.

A tug-of-war produces a winner. It also produces volatility: power lurching from one side to the other each cycle, with each side spending its turn undoing the other’s harvest. A field plowed under in alternate seasons yields very little.

The cost shows up in the cooler, too. A citizen who likes the birch beer has to weigh whether choosing it hands the election to the cola she likes least. So the third idea goes unchosen, and after a while it goes unstocked. That loss is hard to measure, because it consists of reforms that didn’t get a hearing and coalitions that didn’t form. The value of our possibilities shrinks without anyone casting a vote against them.

Ranked choice voting changes the arithmetic at the ballot box. The citizen ranks her options: birch beer first, blue cola second. If her first choice can’t win, her vote moves to her second. She keeps her voice and gives up no influence. The spoiler problem largely goes away.

Australia has used this method for its House of Representatives since 1918, which gives us a long-planted field to study. Australians still elect governments from two big blocs, and reformers who hope ranked ballots will scatter the parties may find that disappointing. The interesting growth happened inside the blocs.

Preferences flow, and the big parties chase them. Labor knows a large share of Greens voters will rank it second, so it carries some Greens priorities in its own platform to keep those rankings coming. The center-right runs as a standing coalition of two parties, Liberal and National, with town and country bargaining in the open. Labor’s internal factions negotiate in daylight as well. In 2022 a group of independents nicknamed the “teals” took several longtime Liberal seats in well-off city districts, because ranked ballots let unhappy center-right voters leave their party without handing the seat to the other side.

The two-party labels endure, then, while the parties beneath them work like coalitions with more voices at the table. The ropes have more hands on them, and some of those hands pull sideways. Madison would recognize the shape.

The Australians add two more pieces. Their Senate uses multi-member, proportional seats, where minor parties and independents often hold the balance of power. And they require citizens to vote. I’ll admit the libertarian in me has trouble swallowing a turnout mandate. Still, the effect deserves a look: when the whole field shows up, parties win by persuading the middle rather than by whipping their own base to the polls, and some heat leaves the contest.

Here at home, Maine has used ranked ballots in federal races since 2018, and Alaska pairs them with an open top-four primary. The early returns are modest. No revolution followed. A conservative should welcome that: a framework adjustment that lowers volatility and leaves the citizen’s options intact. Ranked choice asks nothing of a voter’s convictions. It removes the penalty for holding a third one.

If we want the fuller restoration of Madison’s many-sided republic, multi-member districts with proportional seats would carry us further. Ranked choice is the on-ramp. Start where the ground is ready.

The colas did nothing wrong by competing. Markets need rivals. A market with two sellers, though, behaves like a duopoly, and Adam Smith would have spotted the trouble at a glance. A republic works the same way.

Give the birch beer a front shelf. Liberty was built to be held from many sides.

Sources: