Quasi-Real-Time Business Dynamics: Insights from Google Places Data

Thibaut Duprey, Artur Kotlicki, Daniel Rigobon, and Philip Schnattinger

Economists are increasingly focused on real-time tracking of economic conditions, akin to how doctors monitor patients' vital signs, to inform policy decisions concerning business health and recovery. A recent study introduces a novel method of estimating business openings and closures that leverages data from Google Places, the foundational dataset behind Google Maps.

This approach reveals that the lifting of COVID-19 restrictions in Canada was accompanied by a resurgence of temporarily closed businesses, indicating that government support played a crucial role in helping these firms survive during the pandemic. Policymakers can now potentially gauge the health of businesses more effectively in response to rapid economic shifts.

Challenges in Traditional Data Tracking

Existing methodologies for tracking business dynamics often rely on infrequent data proxies, such as tax filings or business registry information, which are typically reported with a significant time lag. This creates a challenge for policymakers aiming to calibrate support measures during crises without risking long-term economic productivity by allowing viable firms to fail or enabling non-productive 'zombie' firms to persist.

The urgency of the COVID-19 crisis spurred innovations in tracking business dynamism. Various analyses have explored alternative data sources, from Google searches to payment processing and foot traffic metrics, to capture the shifting landscape of business operations. For example, Yelp's data illustrated a significant number of temporary closures early in the pandemic while studies by Harvard researchers incorporated metrics from SafeGraph and Facebook to assess the rebound in employment in the US service sector.

Introduction of a New Estimation Method

The methodology introduced in our study aims to enhance the ability of policymakers to monitor business dynamics in quasi-real time. By employing a scraping algorithm that analyzes the appearance and disappearance of business 'pins' on Google Maps, we can construct a continuous record of individual business statuses—either operational or temporarily closed. This process allows for the development of dynamic time series that reflect changes in business conditions.

Our research primarily focuses on the food service and retail sectors, which were disproportionately affected by COVID-19 restrictions. Through this analysis, we observed that the summer of 2021 saw a notable influx of new business entries, largely attributed to the reopening of previously closed establishments. Data indicates that the timing of these openings is closely correlated with the lifting of government restrictions. For instance, businesses in Vancouver experienced one month earlier openings compared to cities like Toronto, which reflects the staggered nature of restrictions across regions.

The Impact of Restrictions on Business Outcomes

During early 2021 lockdowns, approximately 92% of retail businesses remained operational. However, among those that were temporarily closed, around 40% managed to reopen by the end of the summer, while 30% permanently shut down. In the food sector, about 87% of establishments operated through the lockdown, with close to half of those temporarily closed resuming operations once restrictions eased—bars faring particularly well with a 62% reopening rate.

Conversely, cafes experienced the highest permanent closure rates. This juxtaposition highlights the varied resilience across sectors and the need to tailor support mechanisms accordingly.

Implications of Real-Time Business Monitoring

The findings underscore the potential for quasi-real-time business opening and closure rates to inform indices that track the overall health of the business sector. For instance, Statistics Canada is developing a Real-Time Local Business Conditions Index that combines these rates with real-time traffic data to offer insights into both the extent and intensity of business activity. Moreover, integrating job vacancy data could provide further clarity into labor dynamics associated with business openings and closures.

This micro-level data collection could significantly enhance our understanding of small business dynamics over time, allowing for nuanced analyses such as the emergence of home-operated online retailers and comparing business vibrancy across urban and suburban commercial areas.

Method Limitations and Future Considerations

While this methodology presents exciting opportunities, it is not without limitations. For one, the data accessed from Google Places is perpetually updated, making retrospective analysis challenging. This limits benchmarking against pre-pandemic business performance. Additionally, estimating closures can prove difficult as the absence of a business negates confirmation of its closure timing.

Despite these constraints, preliminary data appears to align well with statistics from national sources, indicating the method's reliability for real-time measurements. As economies become increasingly digital, ongoing enhancements in data coverage and accuracy will likely benefit policymakers and researchers alike.

Consequently, entities such as Google Places and others involved in data aggregation may find opportunities in providing comprehensive business health statistics, which would be invaluable for ongoing economic monitoring and response strategies.