Can Google Search Data Predict Economic Activity? Google Chief Economist Delivers Okun Lecture

Can Google search data predict the economy?

In a talk this week hosted by the Department of Economics, Google’s Chief Economist Hal Varian presented on how Google Trends can be used to forecast economic activity, including GDP, unemployment, and where you’ll buy your next coffee.

Varian visited the Department to deliver the Okun Lecture, an annual series that “seeks to recognize and encourage professional economists to search for policies that will contribute to the betterment of life and living.” His talk, titled “Nowcasting with Google Trends,” explored how Google search queries can be used to predict economic variables in real time. “Nowcasting in economics is the prediction of the very recent past, the present, and the very near future state of an economic indicator,” he said.

Hal’s work and presentation builds on research he co-authored with Hyunyoung Choi in 2010, titled Predicting the Present with Google Trends, which has since been cited over 3000 times. The full lecture can be viewed below or on YouTube here. 

Data is often created continuously, but reported monthly or quarterly. For example, much of the economic data that governments release have a lag: the data for a given month is generally released a month later, and are typically revised several months after that.

Google Trends—which provides data on what people are searching for in real time—provides daily or weekly reports on the search volume on any topic. His talk provided an overview, and several examples, of how search data correlates with economic activity in given industries, and how it can predict the more infrequent and periodic government data.

Varian highlighted Google as a “database of intentions,” citing that many searches lead to quick actions. For example, people who search for “coffee near me” are very likely to purchase coffee very soon thereafter. For some economic activity, however, intentions occur months or even years in advance of the actions. Searches related to renting a car, going on vacation, or buying a house can predict aggregate economic activity that occurs months later. “Lots and lots of tourism issues boil down to this behavior. Google shows possible places where people are considering vacationing, and then what the actual outcome is,” he said.

Taken together, these types of variables can add up to predict what’s happening in the economy as a whole. While official gross domestic product (GDP) data are released quarterly in the US, researchers have devised ways of using real time data to predict economic swings in the economy in the shorter term. The OECD’s weekly tracker of GDP Growth provides a real-time high-frequency indicator of economic activity using Google Trends. It aggregates data about search behavior from 46 countries related to consumption, labor markets, housing, trade, industrial activity, and more, all to help the OECD forecast GDP and inform economic policy decisions.

Varian highlighted several examples of how search data, all publicly available for free through Google Trends, could be used to track important economic indicators.

For example, initial claims for unemployment are an important metric for predicting the overall unemployment rate, which is used for macroeconomic policy decisions. While unemployment claims are reported weekly, and the unemployment rate is reported monthly, Varian presented examples of how the behavior of people who are recently unemployed can be used to forecast labor outcomes. In the immediate aftermath of a layoff, many people use Google to look for information about unemployment benefits. On the Monday after a Friday layoff, many people search for: “Where is the unemployment office?” or “What do I have to bring to the unemployment office?” These searches correlate with actual unemployment benefits given days or weeks later, and thus could be used as one input to assess the overall health of the labor market, in real time.

Varian also gave examples of how search and Google Maps data were used at the onset of the COVID-19 pandemic to help gauge economic activity. In early 2020, the Federal Reserve was interested in how much economic activity had slowed due to lockdowns. Working with Google Maps data, they used the daily average user base as a general measure of activity (e.g., mobility of users predicts economic activity in an area), and observed a more than 50 percent drop in daily active users for Google Maps. This helped inform the Fed’s monetary policy decisions early on in the pandemic, and Google continued to make this daily user data available until very recently.

The second part of his talk focused on the econometric models and techniques his team uses to build tools for prediction using Trends data. These tools include the Kalman filter, spike and slab regression, Markov chain Monte Carlo methods, Harvey’s basic structural model, and Baysian structural time series. Putting this all together, his team focuses on providing practical tools for using economically relevant terms as predictors for what’s happening in the economy now.

About Hal Varian: Hal R. Varian is the Chief Economist at Google. He started in May 2002 as a consultant and has been involved in many aspects of the company, including auction design, econometric, finance, corporate strategy and public policy.

He is also an emeritus professor at the University of California, Berkeley in three departments: business, economics, and information management. He received his S.B. degree from MIT in 1969 and his MA and Ph.D. from UC Berkeley in 1973. Professor Varian has published numerous papers in economic theory, econometrics, industrial organization, public finance, and the economics of information technology.

Varian joins a long list of acclaimed economists who have given the Okun Lecture, including Claudia Goldin, Nicholas Kaldor, Charles L. Schultze, Robert E. Hall, Jeffrey D. Sachs, Joseph E. Stiglitz, Alan Blinder, Richard Freeman, Kenneth Rogoff, Lawrence Summers, Robert Rubin, Raghuram Rajan, Daniel Kahneman, Timothy Geithner, and Ben Bernanke.

About the Okun Lecture: The lecture and conversation series honors the memory of Arthur M. Okun (1928–1980). The donor, a Yale alumnus who was a long-time associate, friend, and admirer of Okun, stated the reasons for the series in these words:

Arthur Okun combined his special gifts as an analytical and theoretical economist with his great concern for the well-being of his fellow citizens into a thoughtful, pragmatic, and sustaining contribution to his nation’s public policy.

Extraordinarily modest personally, he was a delightful and trenchant activist on behalf of others — both as members of our whole society and as individuals. He touched many, many people in ways they will always cherish.

Offered in affectionate appreciation of Art’s gifts, this lecture series seeks to recognize and encourage professional economists to search for policies that will contribute to the betterment of life and living.