Although the 2018 elections were held last week, the madness continues. As it was in the 2000 presidential election, Florida is once again embroiled in recounts for both its gubernatorial and Senate races, accompanied by allegations of lost and stolen ballots and lawsuits. Meanwhile Maine’s secretary of state is overseeing a complex instant runoff process to decide one of that state’s congressional seats. Other races around the country have been too close to call now for days.

What few people realize is that the election for federal offices, including not only the House and Senate but also the presidency, are actually 51 separate state (and District of Columbia) elections, all with their own rules and procedures. Under the Constitution, the “times, places and manner of holding elections for senators and representatives shall be prescribed in each state by the legislature thereof” (Article I, Section 4), which means there are variations across the country.

In Florida, for example, a state law enacted after the chaos of the 2000 presidential election requires a recount when margins of victory are 0.5 percent or less. If, following a machine ballot recount, there is still a margin of 0.25 percent or less, a manual recount is ordered. Twenty states plus the District of Columbia have such automatic recount provisions, each with its own triggering margin. Most states allow a recount to be requested, though in a few a lawsuit would be required to challenge an election.

While Florida engages in its bruising recounts and lawsuits, Maine is undertaking a more clinical “instant runoff” through its “ranked choice” system of voting. Rather than voting for one candidate, Maine voters rank all the candidates. If no candidate receives more than 50 percent of the votes, the last-place candidate is dropped and the results are then recalculated, which could go on for several rounds. In Maine’s 2nd Congressional District, none of the four candidates received 50 percent of the vote, so the process is underway, with memory sticks and paper ballots all being rounded up. Maine’s secretary of state acknowledged that a “tortured, long journey” is now “going to get a little longer.” While Maine is the only state to use ranked voting, several cities in California and elsewhere use it in local elections.

Another state peculiarity rose up in California again this year when voters in the general election were faced with either two Democrats or two Republicans — for example, Democrat Dianne Feinstein running for re-election to the Senate against Democrat Kevin de Leon. In 2010, California voters approved a proposition establishing a top-two primary in which the two biggest vote-getters in the primary advanced to the general election, regardless of party. In a few conservative strongholds in the state, that may mean two Republicans running, although in blue California it is more often two Democrats. Experts argued that the top-two primary would lead to more moderate and centrist candidates. So far, there has been little evidence of that, but it has reduced voter choice. I call the options in the last two Senate elections there “left and lefter.”

One of the more promising electoral experiments is in Nebraska and Maine where, in the presidential race, they assign electoral votes by congressional district rather than winner-takes-all. This seems to reflect more accurately the purple color of many districts in the country, rather than exaggerate the reds and blues. A more radical approach is the National Popular Vote Bill, adopted in 11 states plus the District of Columbia representing 172 electoral votes. If that bill is passed by enough states to represent the 270 electoral votes needed to elect a president, those states will be obligated to cast all their electoral votes for the winner of the national popular vote, even if that nominee lost the state — an obvious end run around a constitutional amendment to change the electoral system.

The states are still what Supreme Court Justice Louis Brandeis called “laboratories of democracy,” where experiments may be conducted with lower risk. As we are learning again in 2018, some of these voting experiments work better than others.

David Davenport is a contributor to the Washington Examiner’s Beltway Confidential blog. He is a research fellow at the Hoover Institution.