The tables you sent to us were reviewed by OCFO staff with experience in statistical analysis. The following are their findings:
The Washington Post researchers focus on the order of bid winners in the tax sale auctions. They observe that for a given number of wins, the observed order of wins do not appear to be random. They assert that certain bidders emerge as consecutive bidders more often than they would if these bids were won in a completely random order. This finding, according to Post researchers, is evidence that there is collusion among certain bidders. We reject the notion that the order of the winning bids is sufficient to determine that there is collusion among bidders. This information, on its own, does not inform us about any collusion activities, nor is it sufficient to reject the existence of such collusion.
a. The order of the calling of the bids is not random. Bids are called in the order of square and lot numbers, and therefore properties that are bid consecutively share similar risk profiles.
b. While the Buyers’ Book data, on which the analysis is based, gives us information about the winners of the bids, it reveals no information about the losers.
c. The Post suggests that consecutive bids are quid-pro-quos: the analysis implies that if company A and B are colluding, trading bid 1 and 2 are of similar value. This assertion is very weak. The value of the bid depends on the amount of the tax lien (tax sale buyers generally purchase the properties for the return in interest over the tax obligation. They would bid beyond the tax obligation because the 18 percent offered over the tax lien could still mean a higher than market return on their total bids which is generally larger than the tax obligation). Two consecutive bids could have different value propositions.
The bottom line – the data from the Buyers’ Book reveal little about the return on the bid and the motivations of the buyers, the bid orders cluster similar properties with similar risk profiles, but not similar return profiles, and the structure of the auction makes pre-commitment among bidders almost impossible. Therefore the data cannot be used as definitive evidence that the companies who participate in the bid have colluded.
We welcome the analysis from the Post, and encourage you to submit any evidence of criminal activity you may have in your possession to law enforcement authorities.