Bank's latest Quarterly Bulletin found a correlation between Google searches for 'estate agents' and changes in house prices
They may not know it, but web users who type "unemployment", "estate agents" or "VAT" into Google are helping the Bank of England to gather information on the state of the British economy.
The central bank is turning to internet search data as it seeks to widen its understanding of unemployment, consumer spending and the housing market.
Research has shown evidence of a correlation between the volume of certain search terms on Google and economic data from traditional surveys. Although the Bank admits that the approach has its limitations, it believes online search data can be an increasingly useful source of intelligence about the state of the UK.
"Internet search data have the potential to be useful for economic policy making," wrote Nick McLaren and Rachana Shanbhogue, in an article for the Bank's latest Quarterly Bulletin. "As further developments are made in this area, and the backrun of the data increases, these data are likely to become an increasingly useful source of information about economic behaviour."
McLaren and Shanbhogue's article showed three examples, each using search volume patterns sourced from Google's Insights for Search. The most striking results came from the "estate agents" search term, which appeared to closely track the change in average house prices over the last eight years. According to the pair, the Google search data, when correctly handled, gave a more accurate picture than other housing surveys.
Monitoring the popularity of searches related to unemployment also appeared to give an insight into the state of the job market. The volume of searches for "JSA" (jobseeker's allowance) and "unemployment" rose in correlation with an increase in the number of people out of work as the recession bit - although the two search terms diverged in 2010.
The quarterly labour force survey (LFS) - the best measure for joblessness - has a time-lag of several weeks, such that data for January-March is published in mid-May, for example. The Bank's research showed that "JSA" search data was more accurate than a consumer survey asking people how they thought unemployment would change in the coming months. However, it was less accurate than the official claimant count data which is published the following month.
Because Google search data is much more current than either the claimant count or the LFS data, it could improve the quality of the "nowcast" data used to give policymakers an up-to-date view of the economy.
The Bank admits that Google search data has its limitations.
"There is only a short backrun, there is no information on the actual volume of searches, and as the index is based on a subsample the backrun of data can change," Shanbhogue and McLaren warned.
Previously, the Bank has used surveys of business leaders and consumers, and reports from its regional agents, to paint a picture of the UK economy. The move to include internet data has been welcomed in the City, although the Bank has been cautioned against putting too much faith in it.
"Any initiative to try and get increased and more timely evidence on the state of the economy is useful," said Howard Archer, chief UK economist at IHS Insight. "Use of the internet, in the way that the Bank is talking about, may be useful for identifying potential turning points or developments in trend, but I would be very careful in interpreting too much from them."
"I would want to see sustained evidence on how the information gathered from the internet shapes up with subsequent actual hard data before putting major store on the data and using it to significantly influence policy," Archer added.
McLaren and Shanbhogue cited a third piece of research, which tracked searches for "VAT" since 2003 against the GfK consumer confidence survey asking whether shoppers feel happy to make a major purchase. The resulting graph showed some correlation, but did not accurately explain consumer behaviour.
Source: http://www.guardian.co.uk/business/2011/jun/13/bank-of-england-google-searches
No comments:
Post a Comment