The heavier line is High-Propensity Business Applications (FRED series HBUSAPPWNSAUS, US Census Bureau) -- applications with real characteristics suggesting they'll become an actual employer business, not every EIN filing. Weekly, right axis. This is the closest thing a government agency tracks as an ongoing time series for "how much real entrepreneurship is happening" -- GEM's own Total Entrepreneurial Activity rate, the one that shows nascent/pre-revenue founders directly, is survey-based and annual, not something chartable here. The lighter line is total nonfarm payroll employment (the Employment Trend), left axis, one real reading a year against the formation data's weekly resolution -- same honest mismatch as the Fed Funds Rate page.

Same pandemic-window-removed treatment as the Fed Funds Rate page's claims panel (March 2020 through December 2021 cut out and labeled, not just visually compressed -- see that page for the true-scale version showing the actual size of the 2020 spike). Claims data only goes back to 2016, so this panel is scoped to 2016 forward even though business applications go back to 2006. Worth naming what this panel can plausibly support and what it can't: rising claims alongside rising applications in the same period is consistent with the necessity-driven story (more people starting businesses because they lost a job), but this alone can't distinguish that from people starting businesses for unrelated reasons at the same time claims happened to rise. The actual necessity-vs-opportunity motive data lives in a separate GEM survey, not in these two series together.