Cristiana Benedetti-FasilGiammario ImpullittiOmar LicandroPetr SedláčekAdam Hal Spencer2026-11-022026-11-022206-0332https://hdl.handle.net/1885/733816712Do firms differ in their innovation response to business cycles? What economic forces shape this heterogeneity? And does it matter for aggregate propagation? We document substantial heterogeneity in firm-level R&D cyclicality, develop a theory to explain this fact, and validate its key mechanism in the data. In a tractable model of business cycles and innovation-driven growth, endogenous selection makes R&D particularly responsive to aggregate conditions for firms close to the exit margin. Consistent with this mechanism, firms closer to exit display more procyclical R&D in the data. Quantitatively, selection and endogenous growth interact powerfully: selection generates heterogeneous R&D cyclicality, while endogenous growth translates these heterogeneous responses into persistent aggregate effects. Ignoring heterogeneous R&D cyclicality misses roughly half of the long-run output cost of a severe recession.engrowth, business cycles, innovation, heterogeneous firms, selection2026-11-02