Can you Google the Economy?

With over a billion searches a day, Google is arguably one the wonders of the modern world.  Search for almost anything and you will be returned tens of thousands, if not millions of results in a fraction of a second.  Recent studies are now suggesting that search patterns can tells us more than just the cheapest source of the must have gadget or who was the lead character in ‘Bonanza’.

Internet advertisers have long understood the power of search to establish trends in usage and demand for products and services to produce instantaneous views of economic activity known as ‘nowcasting’, but recent research is considering whether these trends in web activity can actually provide an accurate forecast of what is happening in the economy.

If you think that your random search for a suitable Christmas gift for your partner is of little interest to anyone other than yourself, think again!  Last year no lesser august body than the Bank of England reviewed all the recent studies on Google search activity to understand whether there was any validity in the proposition that economic forecasts could be based on our search behaviour.

In particular, the Bank homed in on two particular areas of interest in order to consider search as a suitable allegory for an accurate economic forecast.  They focused on unemployment and the housing market, both areas which are frequently quoted a real world factors related to economic activity.  They found that Internet search data have a number of appealing properties as economic indicators. They are extremely timely and cover a potentially vast sample of respondents (approximately 50% of the adult population in the United Kingdom now use the internet every day). In contrast to most traditional survey methods, they are collected as a by-product of normal activity, rather than requiring individuals or firms to respond to survey questions after the event. This can avoid problems associated with non-response or inaccurate responses. And it also means that information is continually collected on a wider range of issues, rather than just on a few pre-determined questions. As a result, search data can help analyse issues that arise unexpectedly.

Although it remains early days, this analysis suggests that internet search data contain valuable information for analysis of unemployment and house prices. These applications treated the search data in a similar manner to existing surveys in conducting standard regression analysis. However, internet search data also has the potential to answer different sorts of questions to existing indicators. Such data has the particular advantage that it can help analyse issues that arise unexpectedly at short notice. Whereas survey data must be consciously collected based on pre-determined questions whereas internet data is collected based on behaviour at the a particular time and trend data going back into the recent past helps to flesh out our understanding.  Such data is generally generated without engendering any form of bias that studies and surveys often find it difficult to avoid.  Equally the idle musings of millions of internet users may well provide a great deal of ‘noise’ which obscures the key trends.

However, without doubt the internet and associated social media reflect the pulse of today’s society and as such reflect our collective social mood. People learn what others are thinking and express what they themselves are thinking by posting and promoting their thoughts and opinions on websites like Twitter, Facebook, and on blogs. People reveal what is on their minds by querying topics on search engines and together create a learning process and a continuous feedback loop, whereby people form, express, and reveal expectations in real-time.

If the revealed expectations derived from Google Trends suggest genuine predictive power for future outcomes, then the symbiotic nature of social media and internet search behaviour suggests that these channels may facilitate self-fulfilling prophecies and if the predicted outcomes relate to economic events, then analysis of Google’s search trends can become important tool for the Bank of England to gauge and manage inflation expectations.