Trends and ambiguity in UCR hispanic offender data.
As with the Census or CDC, the FBI UCR collects hispanic origin information separate from race. While reporting race is mandatory, law enforcement agencies are not required to submit hispanic origin. This results in awkward situations where the totals for each category do not match, and it is not possible to get an authoritative disaggregation of hispanics in the UCR summary tables (Though inference can get us most of the way there.)
Within the UCR, there is a difference between "unreported" ethnicity (this jurisdiction did not submit any race/offender information) vs. "unknown" (this jurisdiction reports, but the ethnicity of these offenders is unknown.)
UCR hispanic reporting has improved rapidly. When I first started using UCR data, hispanic reporting coverage was not available for the majority of offenses.
Today, not only do most offenders come from reporting jurisdictions, but the hispanic proportion of known offenders has stablized over time, suggesting there is little bias introduced from extrapolating known hispanic prorportions to cover for the unreporting jurisdictions. This was not the case even a few years ago.