Currently notations (identifiers) that are taken from source data are used as is without changing the case sensitivity. This means that, URIs end up something like:

http://example.org/code/CL_FREQ/Q
http://example.org/dataset/BIS_M_CIBL_UR/Q/M/U/B/S/5A/KR/2012-Q2

The rationale for keeping them the same was to perhaps allow external consumers to easily match (even though this is trivial) these notations.

From the Linked Data perspective, it doesn't make much difference one way or another. Sometimes lower-case looks nicer and not like some stats are SHOUTING AT YOU!!!