How Google searches can predict economic indicators and give us an accurate picture of today’s economy
Up until recently, Central banks have only used official data when calculating and predicting economic indicators such as unemployment, housing starts and consumer confidence. However official data is published a significant time after it has been collected (usually at least a month) which means that bankers found it difficult to calculate up-to-date predictions. How could bankers determine the present and future state of the economy, using figures that were only a reflection of the past? As Brynjolfsson, a professor at MIT, said “When central bankers were looking at traditional data, they were essentially looking out the rear-view mirror.”
However research has emerged which suggests that by analysing Google searches, we can obtain a clear picture of today’s economy as well as being able to accurately predict future cycles. For example, when trying to estimate unemployment figures, researchers found that by calculating the proportion of Google searches which contained key words such as “JSA” (short for jobseeker’s allowance), “jobs” and even “solitaire”, they could accurately calculate figures that almost exactly matched the official data when it was published weeks later.
Google searches can also be used to determine demand and the willingness to consume in an economy, which is useful if you want to predict GDP or whether our economy will dip back into recession. When you Google search for “blue jeans, Topshop”, analysts can infer that you have the demand to buy a pair of jeans from Topshop.
Over 60% of the adult population use the internet every day, so Google searches cover a large sample. In addition the information obtained is extremely up to date – Google pass on the records of their searches to central banks within three days. This helps governments react quickly to economic concerns and crises. Could the global recession have been avoided, if central banks had an accurate picture of the economy? Would they have been able to use the weeks in which information was delayed, to formulate a crisis aversion plan?
However there are downsides to using Google searches to predict economic factors. Only people of a certain wealth are able to access the internet and younger generations are much more likely to use Google than their older counterparts. Unfortunately this makes the sample of data unrepresentative of the population. However researchers have assured critics that this is a minor issue that will only get smaller because the internet is quickly becoming more accessible to everyone. Another downside is that people may search Google purely for curiosity, for example if someone looks for a new job opening who is already employed, or if you look at jeans on the Topshop website, even though you know you can’t afford a pair this month. ‘Curiosity searches’ such as these distort analysts picture of the economy and increase the chance of inaccuracies in future predictions.
Although using Google searches for this purpose is a relatively new initiative, as you can see from the graphs, the results are astonishingly accurate. From summer 2012 Central Banks have had access to all Google search records – which just shows how much potential banks think this method of economic analysis has. If you still aren’t convinced, why not try it out for yourself? You can do your own analysis using Google search data via this link: http://www.google.com/trends/