How to track housing inventory levels?

Yeah, I need to say that they might not be spurious lest I lead you astray. However, even the best numbers are going to have serious flaws. From what I understand, at least with the statistics that we spend hundreds of thousands of dollars per year to purchase, a lot of information is inferred based upon trends and samples. Just collecting accurate primary data for my projects is so incredibly difficult that it borders on the absurd. I don’t see how mass collection of accurate data is possible (which is the reason I hypothesize that there is so much money to be made in this industry). Let me give you a few examples of data that is far from “hard.” Take CoStar Group for example. Many of their sales present cap rates. However, coming from experienced pros and from what insiders tell me, they infer MANY of these cap rates based upon operating expense comparable information, some rent surveys, vacancy studies, etc. When it comes down to it, the “facts” as presented are far from facts–just an educated estimation. I’ve seen vacancy reports for retail, office and industrial space. NO WAY is this stuff accurate based upon what I know. I’d bet almost anything that vacancy information you get in reports is based upon inferences (there’s that word again) from small sample sizes, interviews with brokers, trends, etc. There simply isn’t the manpower working in this industry to get “accurate” (that is, legitimately correct) information. I can’t tell you how many times LoopNet, CoStar, RCAnalytics, and local assessor information vary. Often times, these VERY GOOD sources of data have very contradictory information or information that exists in some databases and not in others. I know that my company’s database of primary research is MASSIVE and cost a ton to develop. We don’t give this information away cheaply. And I often question our methods of collection. And the information becomes dated very fast. No clue how these companies/organizations can get accurate residential inventory data. Do people have any idea how remote much of the U.S. is? Do people realize how incompetent local governments are that collect much of this information? I have no idea how one organization, for example, can obtain accurate inventory data on residential buildings when there are literally thousands of independent MLSs nationwide that require subscriptions. There simply isn’t the money or the manpower. Basically, what I’m trying to say is that the vast majority of information you obtain in secondary research is going to be accurate within a reasonable variance, but based on what I’ve experienced, I highly doubt the legitimate accuracy of publicly presented information collected en masse. Obviously, the way to ensure the best accuracy is obtaining information directly from your clients, from local governments and any other information obtained in primary, focused research.