Standing above today’s proliferation of competing and contradictory political polls, there’s only one type of poll that can claim near-perfect reliability going all the way back to 1884. It’s probably one you aren’t even looking at. And it’s declaring George W. Bush the winner.
The type of poll I’m talking about isn’t the usual public-opinion survey. It’s organized betting on the election. To participate in such a poll, you have to be willing to put your money where your mouth is. I’ll get to the present election in a moment, but first some history.
Since the founding of our nation, citizens have bet money on presidential elections — just as they’ve wagered on sporting events. By the end of the 19th century, election betting had become a formal national institution. Betting — with many millions of dollars at stake — was conducted on Wall Street by specialized brokers called “betting commissioners.” Near elections, the betting odds for each candidate were published daily in the New York Times and other newspapers, just as poll results are published today.
According to a recent paper by Paul W. Rhode and Koleman S. Strumpf, economics professors at the University of North Carolina, Chapel Hill, election betting correctly predicted the popular-vote winner in every presidential race between 1884 and 1940, with only a single exception. They quote the New York Times, which said in 1924 that the “old axiom in the financial district [is] that Wall Street betting odds are ‘never wrong.’”
Here’s the record. From 1884 to 1940 — the heyday of organized election betting — there were thirteen elections. In nine of them, the betting markets strongly favored one candidate by setting the odds at a 60 percent-or-greater probability a month before the election. The favorite won in nine out of ten elections. Also, there were three very close elections, and the betting odds correctly put all three near 50/50. For the remaining election, in 1908, regulatory issues kept the betting markets from functioning until just before Election Day — but the odds did call the winner correctly.
The old Wall Street betting market dried up after 1940 with the advent of stricter anti-gambling regulations and the arrival on the scene of the first public-opinion polls — the same kind we still use today. But thanks to the Internet, election betting is back, and it’s better than ever.
Right now the election betting market makes George W. Bush the favorite to win re-election, with a probability of about 62 percent. This is above the threshold at which — a month out from the election — betting markets have only been wrong once in 116 years.
Today the dominant election betting market is Tradesports.com, a website based in Dublin, Ireland, where you can bet on all manner of sporting, political, and current events. Presidential bets at Tradesports.com take the form of online futures contracts. After the election, the Bush futures will settle at a value of either 100 if he wins or zero if he loses. So the price today is always in-between zero and 100, and indicates the betting market’s estimate of Bush’s probability of winning.
The Bush futures at Tradesports.com trade several thousand contracts a day. They are deep and liquid markets, but small — trading represents a dollar value that is far less than that traded in typical futures markets. Yet similar political futures contracts, traded in even smaller size on the Iowa Electronic Market, a website operated by the business school of the University of Iowa, have correctly predicted every presidential popular-vote winner since 1988. (If you want to track the Bush futures, and a dozen other futures contracts on the election and current events, I’ve set up a feed on my website where they are all together in one convenient place.)
Okay, let’s look at the risk that the Bush futures are wrong in making the president the 62 percent favorite.
The one time the presidential betting markets were wrong was the election of 1916. The election-year dynamics back then were similar to the ones we are experiencing today. In 1916 the nation was involved in a divisive debate over America’s involvement in the European conflict that would become known as World War I. Today it’s Iraq, and the broader war against terror. Like Bush, Democrat Woodrow Wilson was the incumbent in that race. But Wilson ran on an anti-war platform of strict neutrality (though he would bring America in the war the following year). The challenger, Republican Charles Evan Hughes, wanted more mobilization and preparedness.
Hughes was heavily favored in the betting odds, and was thought to be the winner on election night. But the next morning he ended up losing, in a squeaker decided by a handful of votes in rural California. It goes to show that voter sentiment is volatile and difficult to predict when it comes to matters of war and peace, and of America’s role on the world stage.
Another risk with the Bush futures is that they could be manipulated by unscrupulous political operatives seeking to send a false signal before the election. Trading in the Bush futures seems to be quite orderly, in general, but on September 14 there was a very wide trading swing in response to a sudden surge of selling volume: Bush futures fell in one hour from the low 60s (where they are right now) to almost 49. Then, a half hour later, when the selling abated, they popped right back up to the low 60s.
Likely this was the effect of a larger-than-usual order hitting a relatively small market. But who’s to say that it wasn’t someone mounting a “speculative attack” deliberately designed to move the market? It would be no different than the kind of thing George Soros did in 1992 against the British pound (he’s not called “the man who broke the Bank of England” for nothing). And come to think of it, Soros is certainly on record as being quite opposed to Bush’s reelection.
Today’s election betting markets are no crystal ball — after all, what is? But I think they are the most reliable polls going because they harness the power of markets to make the best-informed and least-biased forecasts. As the New York Times explained in 1924,
The Wall Street odds represent the consensus of a large body of extremely impartial opinion that talks with money and approaches Coolidge and Davis as dispassionately as it pronounces judgment on Anaconda and Bethlehem Steel.
I’ll take that over one of the polls run by the liberal media any day.
In my previous column for NRO, I reported on a distortion by Paul Krugman of George Bush’s strategy for North Korea. I quoted New York Times “public editor” Dan Okrent’s response to the distortion, in which he said, “I think this is what columnists do, and I’m not going to hold Krugman to a standard any different from the one I hold the other columnists to.”
The sentence I quoted was from a somewhat longer statement, and Dan complained to me that “I believe your deletion of the first sentence significantly altered the meaning of what I said.” I disagree, as the sentence I quoted fully reflected the standard-setting process involved in considering corrections. Moreover, in the very same sentence in my column, I provided a link to his entire statement as it had been published on my blog.
That said, his full statement is of interest in its own right, because it reveals something of how Dan reads Paul Krugman’s columns. For that reason, and as a courtesy to Dan, with whom I have enjoyed a longstanding correspondence, here is his statement with the two preceding sentences restored.
It’s incomplete, it’s tilted, it’s possibly unfair. But I don’t think this is a correctable error. I think this is what columnists do, and I’m not going to hold Krugman to a standard any different from the one I hold the other columnists to.