...
Currently we have finite population, and statistical population is a child of that. We could break that down further, for example with respect to how the estimation is done and how non-sampling errors are calculated, but that detail is likely not important for FIBO, whereas coverage is. For housing starts, that is:
COVERAGE
New, privately-owned residential buildings currently authorized by a building permit or started in areas not requiring a building permit.
Understanding how one collects detail on housing starts in areas that don't require a building permit is done, or how to calculate sampling errors related to that, is fairly arcane, and likely not impactful with respect to what we are doing in FIBO with economic indicators.