Yes, the policy for geographic information in public use microdata is to identify no sub-state entities other than Public Use Microdata Areas (PUMAs) which are specialized units required to have at least 100,000 residents each, helping to protect the identities of respondents. PUMAs subdivide the entire nation, and they can't be designed in such a way to perfectly correspond with all CBSAs or other sub-state geographies. IPUMS USA identifies CBSAs only when the available PUMA information aligns fairly well with CBSAs (within a 15% population mismatch tolerance). This fortunately includes most large metro areas, but it omits many smaller metros and all micro areas. Similarly, IPUMS USA identifies counties only when they correspond exactly to a set of PUMAs. You can click on any variable name in this list to get more info on how IPUMS USA identifies these entities.
Of course, the Census Bureau could use similar policies to identify CBSAs and other areas _when possible_ given the available PUMA info. But it's also true that a large portion of PUMS data users get their data from IPUMS USA anyway--not only for geographic info but for many other added-value features, most notably samples that go back through history with consistent codes across time--so if IPUMS is providing this geographic info already, there's less motivation & need for the Census Bureau to make the effort to do it as well.