This paper studies the effect of carbon pricing shocks on the macroeconomy using a highfrequency identification approach. Focusing on the European carbon market, I consider whether carbon policy announcements can be summarised by a single factor, or whether there are additional dimensions that need to be accounted for. By measuring the high-frequency surprise changes of a spectrum of EUA carbon futures around 145 regulatory events, I find that the events can be summarised by two factors rather than just a single factor. A particular rotation of the orthogonal instruments is used to derive two novel instruments, namely, an “action” instrument, which captures changes to the current carbon policy rate, and an “expected path” instrument, which captures changes to the expectations about future carbon policy. I measure the effects of the two factors on a class of asset prices by estimating a daily local-projection model. This is complemented by estimating a Bayesian external instruments VAR model to map out the dynamic macroeconomic effects. I document that a tighter carbon policy successfully reduces emissions, although this is simultaneously met with significantly lower economic activity and higher prices that are persistent over the horizons. More importantly, the results indicate that the "expected path" instrument dominates in its negative implications on macroeconomic aggregates, stressing the importance of capturing the additional dimension of carbon policy announcements, particularly from a policy perspective.