News

LIES & STATISTICS

492 words
26 February 2005
Australian Financial Review
First
63
English
© 2005 Copyright John Fairfax Holdings Limited. www.afr.com Not available for re-distribution.

The market in Iraq election turnout 'futures' got it right, says Justin Wolfers.

A lot has hinged on the success of the Iraqi election, both in Iraq and among coalition governments. But was the election a success? Judging by the tone of the news reports over the past few weeks, turnout vastly exceeded expectations. But what do we know about the election today that we didn't know four weeks ago?

Firm numbers about turnout were absent from nearly all of the celebratory reports, and, indeed, turnout numbers were only released subsequent to most news analysis.

Looking for at least one objective measure, I turned to an unusual source an Irish website, tradesports.com, which established a futures market focusing on turnout in the Iraq election. The structure was simple: you could buy or sell a future that would pay $100 if total turnout exceeded 8 million votes.

Before the election, this future was selling for about $65. Following very early reports that turnout had hit 72 per cent, the price of the security rose as high as $95. The price fell sharply as the poor quality of the early information became apparent. It hovered at about $50 while the count proceeded.

What does this tell us?

First, there were no surprises in Iraq, positive or negative. Before the election, markets perceived it more likely than not that turnout would exceed 8 million. And, indeed, when the official count was released recently, turnout was 58 per cent, or 8.56 million people.

Second, most reporting on the "success" of the poll occurred not only before official figures were released but also during a period in which this market suggested commentators actually had little serious indication as to whether or not turnout would exceed expectations. In the end, turnout exceeded 8 million but, given the pre-election trading, the markets suggested this was the sort of success they had expected.

As those who followed the betting markets during the Australian election can attest, election futures can be extremely accurate, aggregating news from multiple formal and informal sources.

Moreover, for the interested citizen, these markets provide a way to cut through the rhetoric and figure out precisely what is known and what is speculation.

Politics has become so much about managing and then beating expectations that it can be hard to tell what is good news and what simply appears to be good news relative to a successful campaign to dampen expectations. These markets provide some insights.

Today's optimistic news headlines on Iraq appear to reflect unrealistically dampened expectations of low voter turnout. So far, it seems that the election went about as well as expected, and much of our optimism simply reflected uninformed speculation.

Justin Wolfers.

Justin Wolfers is an assistant professor of business and public policy at the Wharton School, University of Pennsylvania. His research focuses on predicting markets