iPredict got its start in 2005, about six months before the general election that year. The idea was to be up and running in time for that election.
There were two main obstacles to achieving that: one, we didn’t have any software, and two we didn’t have regulatory approval to run a real money market.
As it turns out, by far the easiest of those two problems was to develop software required to run these markets.
iPredict Ltd was formed in January 2008 and development began in earnest at about that time. By April we were ready to launch into the New Zealand market. The only problem was, we hadn’t received regulatory approval to do so.
The regulatory environment in New Zealand was not designed with prediction markets in mind. The idea to bring a prediction market to New Zealand goes all the way back to 1992, when the idea was run past the Financial Markets Authority. By 1995, the view was that it would not be possible to run a real money prediction market in New Zealand that complied with securities regulation, and the idea was shelved.
By 2005, that view had changed, and a way to comply with securities legislation had been formulated. A proposal was submitted in January 2008 and, to our great relief, authorisation to operate as a futures dealer in New Zealand was granted in late August 2008, just 2 months before election day. To this day we count ourselves fortunate to have managed to get the attention and authorisation from the Financial Markets Authority, which at the time was dealing with a spate of finance company collapses.
iPredict launched on 9 September 2008. The timing, as it turns out, could not have been better - on that day Owen Glenn testified to the Privileges Committee that afternoon about his funding of New Zealand First, and our lead stock was whether Winston Peters would survive as Minister of Foreign Affairs through to the election. On our first day in business, we measured the political value of Mr Glenn’s testimony in real time. It was a fun day.
The election went very well, iPredict outperformed all but one polling company throughout the election cycle. 19 polls were published between our launch and election day - compared head to head iPredict was more accurate than 15 of them for a 79% success rate.
iPredict has since built an impressive forecasting record, with nearly 90% of binary (i.e. yes/no questions) predictions being correct (contact us for the analysis showing this).