The announcement of an economic rescue tool often comes with implicit promises of more intense intervention if conditions worsen. In this paper, Tyler Muir and co-authors propose and implement a method to identify conditional policy promises and quantify their impact using data from options markets. When the Federal Reserve introduced corporate bond pur-chases during the COVID-19 crisis, markets expected five times more price support in crash scenarios relative to the median case. This implicit promise to significantly expand the size of the intervention in bad states explains half of the market response to the announcement. Furthermore, they document that the behavior of the price and tail risk of corporate bonds remains substantially distorted even after purchases have ceased. We confirm the pervasive influence of conditional promises across several policy announcements: U.S. quantitative easing, Bank of Japan asset purchases, bank equity injections in 2008, and FOMC releases.