Gauging Hourly Economic Activity in Your Neighborhood

Imagine gauging hourly economic activity at the local (zip-code) level. We argue that this is now feasible and propose a new granular economic modeling approach (GEM) for measuring, modeling, and predicting economic activity at high levels of granularity. Our approach uses bridging equations and aggregation constraints to link highly granular spatio-temporal data on variables such as foot traffic, traffic flows, electricity usage and credit card spending with aggregate data on payrolls and income. We extract a common dynamic factor that summarizes local, hourly economic activity and is consistent with observable measures observed at far coarser levels of granularity such as city-wide payroll data and GDP.