“…On the other hand, we show in Section 4.4.2 that an important insight in Andrews, Roth, and Pakes (2019) allows us to adapt our methodology to conduct sub‐vector inference in a class of conditional moment inequality models. Our analysis is also conceptually related to work on sub‐vector inference in models involving moment inequalities and to a literature on shape restrictions; see, for example, Romano and Shaikh (2008), Bugni, Canay, and Shi (2017), Kaido, Molinari, and Stoye (2019), Gandhi, Lu, and Shi (2019), Chernozhukov, Newey, and Santos (2015), Zhu (2019), and Fang and Seo (2021). While these procedures are designed for general problems that do not possess the specific structure in (1), they are, as a result, less computationally tractable and/or rely on more demanding and high‐level conditions than the ones we employ.…”