I'm not so sure there should be any delay between luma and Pr, Pb in any source, except for the effects of subsampled Pr, Pb sample siting. I would expect any delay due to Pr, Pb filtering to be corrected in the source.
Here's a paper on the possible variations. In any mpeg-2 source, I would expect to see the slight (one sample) luma delay shown (MPEG-2 4:2:0 sampling, far right at top of p.2), not a luma lead.
https://poynton.ca/PDFs/Chroma_subsampling_notation.pdf