Researchers are using Google to forecast economic activity
- For some years now, Google's search data has been used to monitor or forecast economic activity.
- This data, available on a weekly basis responds to a need for speed as it takes longer for traditional indicators such as GDP to become available.
- Google searches are an interesting indicator of economic health as they provide real-time information on Google users' perception of the economy and their willingness to consume.
- Indicators from Google are particularly relevant in times of crisis, as they react quickly to changes in the economy.
Why have researchers and institutions, such as the OECD, turned to Google data to forecast countries’ economic activity? What needs does it address?
Usually, to make macroeconomic forecasts, we use data from central banks or statistical institutes such as INSEE. This data is very informative, but it is not immediately available. That’s why people are interested in other data sources that can provide information in real-time.
If an economic policymaker needs to boost the economy, for example, they need to know what the current economic situation is. This is not possible using only the official data. GDP data is published quarterly, on average one and a half months after the end of the quarter in question. It is therefore impossible to adjust economic policies instantly. The idea of using alternative sources, of which Google is one, is really to address this problem of delayed official data.
What Google tools are used in this macroeconomic forecasting work?
There are two types of data from Google: Google Trends and Google Search. The primary source of both databases is the same, however: words typed into the Google search engine. Most researchers use Google Trends: a web page that everyone has access to. The data corresponds to search trends by country and by category (entertainment, business, health, science, sports). Google assigns the search keyword to a category.
Google searches can be seen as a summary of how people perceive the economy.
Google Search, on the other hand, provides data sets from internet searches, made available by Google and given to the European Central Bank. The two databases are constructed differently, Google Trends looks at the volume of searches while Google Search gives information on the change in volume. In my study, we worked with the Google Search data.
So, the idea is to predict economic health by analysing what Google users type into the search engine. Where did this idea come from and why is this data relevant?
The first papers on the subject were published by Hal Varian, chief economist at Google. This data is fairly new: it has been available since 2004, but I started my project using Google data in 2016. The intuition and presupposition behind using this data for macroeconomic forecasting is that Google searches can be seen as a summary of how people perceive the economy. If the economy is healthy, people tend to search for culture, travel, etc. On the other hand, if there are problems with unemployment, there will be more job-related searches.
How effective is this data in predicting a country’s economic activity? Is it useful for predicting periods of growth and recession?
What I have observed in my research is that these tools are particularly useful in times of crisis. During the crisis of 2008–2009, for example, Google data anticipated economic activity well because it is more responsive to change, compared to official data.
However, the data from Google has very little correlation with GDP. Except in times of crisis, official information is still more informative. It is also essential to make a pre-selection, as there are about 300 categories per country. If you use them all, it can make the estimation process less clear. Before making a forecast, it is therefore necessary to select the most correlated Google categories to predict GDP. If this is done, the results can be very informative, even for stable periods, and when no official information is available.
Which research categories are most useful for forecasting economic activity?
The most correlated categories are often consumer-related, such as leisure and entertainment. This is easily explained: during times of economic stability, people are more inclined to buy. We should also consider categories related to social networks. People may be more or less active in their use of social networks, depending on the state of the economy: getting information on platforms or consulting sites like LinkedIn to find job offers, for example.
What are the advantages and limitations of Google data compared to official data?
The main advantage in comparison to official data is the issue of speed. In fact, the overview of the economy is almost instantaneous: we listen to the news, we see that there is a war or a political crisis, we react immediately, we adapt our behaviour. However, industries take much longer to adapt to an economic crisis, it doesn’t happen overnight. Most economic actors are slower to react.
The main limitation is that this data is difficult to use. In my study we tried a number of methods, and some of them did not work at all. For example, the method of pre-selecting research categories only works during periods of stability: in times of crisis, you should not pre-select.
How do you see the future of economic forecasting, in terms of data sources? Do you think that the use of Google will continue to grow?
I don’t imagine that any one data source will become better than others: we will continue to use several sources and several models. Depending on the economic context, we will have better forecasts with certain data. What we need to do now, in addition to further automating the methods we have put in place, is to compare the performance of Google data in terms of forecasting against other alternative sources of data, such as newspaper articles that follow economic and financial news: this is what I am trying to apply in my research.
