From the AFR this morning comes the news that Treasury forecasts are to be against a MacroBusiness benchmark (well…kinda):
The next time you tweet or make a post on Facebook it could help influence the national economy, according to Treasury’s head of technology, Peter Alexander. From the first half of 2013 the department will begin trials using so-called big data to create economic indicators and models from social media, and to blog posts on the internet. Big data involves collecting and organising vast quantities of seemingly disparate information. By using smart algorithms and high-speed technology, organisations gain new insights that could be missed by human analysis.
…The testing of big data comes as Treasury completes an external review of its economic forecasts, which began in June 2012 amid criticism that some of its predictions had been overly optimistic. “I think it’s a possibility there could be a range of new indicators that come out based on big data, better analytics and more speedy capabilities,” Mr Alexander said. “For example, over time we could see if negative sentiment over social media is a good predictor of a downturn in retail sales.
I can’t claim expertise in this area but it sounds plausibly useful. Then again, maybe not. In the end, getting economic forecasts right is surely about who is reading the data? And for that to work you need a culture of serious, forever skeptical individuals. MB has not been ahead of the curve for two years because it has access to any magical data. It has a culture of folks that are uneasy with cliches and easy or conventional answers. That’s what Treasury needs.
