How do firms cope with economic shocks in real time?

In a turbulent world, effective policy requires granular estimates of firm responses to shocks as they unfold. We develop a replicable framework that combines high-frequency administrative and survey microdata with a pre-registered shift-share design to estimate multi-margin firm responses in real time. We show that the resulting coefficient vectors can be interpreted as first-order perturbations of firms’ optimal-response functions. We apply the framework to the 2021–23 UK energy price shock and find that energy-intensive firms pass on cost increases, build cash reserves and shift towards homeworking. We find little evidence of aggregate employment losses or firm exits. Responses are highly heterogeneous: small firms drive price pass-through; large firms, capital investment. We show that our real-time estimates are consistent with the UK’s structural business survey, released two years later, and across survey instruments. We apply the framework out of sample and in real time to the energy price shock triggered by the 2026 US–Israeli strikes on Iran and discuss implications for the design of energy support schemes and environmental policy.