http://www.ft.com/intl/cms/s/0/4c912406-a5b1-11e5-a91e-162b86790c58.html#axzz3vAottJQ2Look beyond the polls for the real odds on Donald Trump
A canny Scots acquaintance observed the other day, “you can lie to a pollster any time you want but you don’t lie to your bookmaker.” Anybody can lay off or hedge bets, but his fundamental point has validity, never more so as the 2016 presidential election gets closer.
Public opinion polls are a dime a dozen but some have shown, as the Republican Donald Trump repeatedly asserts, that he could beat Hillary Clinton, the Democratic frontrunner, next November. Others have suggested that Marco Rubio and Ted Cruz, the two Republican senators apparently on the rise, could run her close.
The betting shops, however, paint a different picture. The higher messrs Trump, Rubio and Cruz rise in the GOP, the heavier favourite Mrs Clinton becomes to win it all. It probably helps her that she has established a dominant lead over her only real challenger, Senator Bernie Sanders, who seemed a bit of a threat in the summer. As it stands a repetition of 2008, when she was sandbagged by Barack Obama en route to the nomination, appears unlikely.
She has also been campaigning pretty much under the radar, so much attention having been devoted to the Republican contest, which in turn, as Mr Trump says with every waking moment, is all about the polls — where he sits on top.
The problem is with the polls themselves, both generically and specifically. Everybody, not least the reporters who now use them as the crutch without which they could not write or say a word, should read Professor Jill Lepore’s damning dissection of their reliability in the New Yorker magazine last month.
Initially, in the days where Thomas Dewey “beat” Harry Truman — as the Chicago Tribune so famously miscalled the 1948 presidential election on its front page — there was a fascination with polling. The very notion that ordinary people could be asked for their opinions was seductive, with upwards of 90 per cent of those approached responding. Today, that percentage is in single-digits — and not only because half the country now uses only mobile phones, difficult for pollsters to reach, and not landlines.
Polling’s demographic models look increasingly suspect and their samples smaller and smaller, often as few as 300-400 people. Of those, some respondents surely lie about their opinions or indeed whether they intend to vote at all (Mr Trump’s supporters are believed not to be regular voters, which may be true, but the evidence is not exactly persuasive).
Also, polls have focused almost exclusively on the Republican contest and Republican voters in the first two states where actual ballots will be cast, Iowa and New Hampshire. Attempts have been made to extrapolate such findings to the general election, but, again, they are inherently suspect. They have tended to show even the rank outsiders in the Republican field within spitting distance of Mrs Clinton, which defies credulity.
A cruise through the betting websites (PaddyPower, Bovada, Intrade, the University of Iowa’s futures contracts and so on) paints a much less volatile — though not static — picture, in which Mrs Clinton has strengthened her standing. PaddyPower at the moment makes her the 8/11 favourite to win it all in November, against Mr Rubio at 5/1, Mr Trump at 6/1 and Mr Cruz at 6/1. That translates into giving her a 55-58 per cent chance of becoming the next president. The University of Iowa market puts it slightly higher at just over 60 per cent.
Using the US method of expressing odds, in which the number 100 means evens and a minus number reflects odds on — in other words, that it is more likely to happen — Bovada has Mrs Clinton at -130, up from -110 in August.
She is trailed by Mr Rubio at 400, Mr Trump at 600 and Mr Cruz at 1000*. It may be noted that the oddsmakers rank Mr Trump only second favourite to win the GOP nomination, in sharp contrast to the opinion polls that have him ahead by streets. They must have some faith that the party establishment — which loathes both Messrs Trump and Cruz — will find a way of asserting itself.
Of course stuff does happen and change the odds, as when people actually cast a vote rather than express opinions. Before the Iowa caucuses in 2008 Mrs Clinton was given a 70 per chance of winning the Democratic nomination against Mr Obama with 25 per cent. Even after the caucuses, which he won, she was given a 52-44 per cent edge. By April, however, with several big state primaries remaining in her strongholds, he was north of 80 per cent while she was logging in at barely 12 per cent.
Stuff could happen again. The populist anti-establishment wave sweeping over the Republicans could engulf the wider electorate. The Clinton name could prove a liability, just as the Bush one has to Jeb Bush, the former Florida governor whose campaign is flailing, especially if the mantra of change again takes hold. Terrorists could strike again at home and the Middle East deteriorate further, if that is possible. The economy could turn sour, although the Federal Reserve does not think so. All the demographic factors favouring Mrs Clinton (with women, minorities of all colours and sexual persuasions) will be worth less if people do not vote. Pigs might even fly.
But the bookmakers, who generally win, calculate all this in laying their odds. If only the pollsters, proven so wrong in too many elections around the world to count, could do the same
