I’ll keep this short and not-so-sweet as it’s a holiday here today, but I thought I would share an interesting article I read in the Sydney Morning Herald over the weekend. Most of you will probably be aware that there was a general election in Australia last month and what was termed in many quarters the government’s “unwinnable” election was actually won by that same government – not only that but with a working majority.
The Herald has an article on an polling company that “binned” an opinion poll saying the Labor opposition were doing much worse than thought because as the paper quotes the owner of the firm as saying, “no one wants to release a poll that is widely out of step…we didn’t want to be seen as having an inaccurate poll.”
This is not the first time this has happened of course, the same was the case with the 2015 UK election (although not, it has to be said, the Brexit referendum where one poll did, as I recall, predict an exit).
So we have firms not releasing polls because they do not conform to the very same expectations that have been established, presumably by the first half dozen or so polls? They also, as is observed in the aforementioned article, interrogate data that appears “wrong” more vigorously than if it matched expectations.
I think we can all agree this highlights the ridiculous nature of opinion polling in the modern era? On which basis all I can add is, welcome to asset manager benchmarking!