Want free money? An essentially riskless investment with returns guaranteed on November 7th? Or something really close to it? Here’s how.

There are two major “prediction markets”: online trading floors where you can buy and sell financial contracts on political events. On Wednesday afternoon, on InTrade, the most famous for-profit exchange, the contract for “Barack Obama to be reelected” was selling for $5.85; if he wins, it pays out $10. On the Iowa Electronic Markets (IEM), operated as an educational tool by the University of Iowa, the contract for Romney winning the popular vote was selling for 33.9 cents, and will pay out $1 if he wins.

Now, let’s arbitrage. Buy 10 Romney-to-win contracts on IEM, for a total cost of $3.39. Then, buy an Obama-to-win contract on InTrade, for a price of $5.85. You’ve now spent a total of $9.24, but you’re guaranteed $10, no matter who wins. If it’s Romney, your IEM contracts are worth $10. If it’s Obama, your InTrade contract is.

$10 for $9.24: that’s an 8.2% guaranteed return in two months—or 48.4% annualized. I’m going to do this—and maybe you should too.

For financial markets, those are absurd numbers for essentially riskless bets. By contrast, the S&P 500 and Dow 5000 have returned just over 20% year over year. And on average, hedge funds have returned a mere 12.61 percent annually since 2008, according to the FT. Any trading floor would throw a Bollinger-drenched bash upon discovering an arbitrage like InTrade-IEM’s.

But don’t head to Goldman, take out a $10 million loan, and lever up your Romney/Obama bets just yet. IEM imposes a $500 trading limit. That means that you won’t be able to make an unlimited killing on our scheme. At most, you can buy 1,470 IEM Romney contracts ($498 today), and 147 InTrade Obama contracts ($860), for a total initial investment of $1,358, and a guaranteed return of $1,470. Then you have to subtract two rounds of InTrade’s monthly fee, $4.99, as well as IEM’s sign-up fee, $5, for a final return of $1,455.

Minus fees, you make $97 in free money—to get technical, a “net-of-fees” return of 7.1%, 41.4% annualized. Not bad for an essentially riskless investment, which pays out the same amount no matter who wins.

I say “essentially” riskless for one, niggling reason. InTrade’s contracts pay out to the “winner of the presidential election,” while IEM’s pay out to the “winner of the popular vote.” Only three elections in U.S. history have ever handed the presidency to a candidate who did not receive the most popular votes: 1876, 1888, and of course, 2000. But if it happened again in 2012, it could mean either very bad or very good news for our arbitrage.

Let’s say Obama wins the popular vote, and Romney wins the electoral college. Our IEM Romney contracts, based on the popular vote, go to $0. And our InTrade Obama contracts, based on the election winner, also collapse. That’s a 100% loss.Conversely, if Romney gets the most votes, but Obama is reelected, our IEM contracts go to $1 and our InTrade contracts go to $10. If we’re invested up to the maximum in this scenario, our initial investment of $1,358—1,470 IEM Romney and 147 Intrade Obama—becomes worth $2,940, for a staggering return of $1,582, or 116 percent. If you think a popular/electoral vote split is a real possibility, this isn’t so much an arbitrage as a high-stakes coin toss.

Markets aren’t supposed to offer you free money: on a competitive trading floor, there should be no such thing as a free lunch. Here, it appears there’s something close to it. In the real-world capital markets, arbitrages like this are usually closed as soon as they’re discovered, as traders jump in to exploit them. But in the notably illiquid political-prediction markets, assets are events, and the contracts that predict them aren’t fungible: you can’t substitute an IEM contract for an InTrade one, or take IEM’s “Romney-to-win” and sell it to an InTrade investor. So our arbitrage won’t close the gap.

And while that gap exists, so does the free money.