Consumer Lending Statistical Analysis

We have a database of less than 500 consumer loans. A regulatory agency has conducted a regression analysis of these loans and has determined that there appears to be a disparity in the pricing of loans between minority and non-minority borrowers, after controlling for a number of variables (such as loan amount, past delinquencies, age of collateral, loan term, etc.). The regulatory agency does not share the specifics of its analysis, only the end findings (e.g., an XX.X basis point interest rate disparity between loans to minority vs. non-minority borrowers). We are looking for someone who can take the data and: 1) replicate (essentially reverse engineer) as closely as possible how the regulatory agency may have come up with its end findings; 2) identify any weaknesses in the regulatory agency's end findings (for example, perhaps the regulatory agency's analysis failed to take into account medical-related delinquencies, which the lender does not factor into its pricing decision) that may exist, and if so, and if controlled for, would help minimize the pricing disparity. We have these types of projects come up from time to time and are looking to establish a potentially long-term relationship with an individual or a company that is skilled at regression analysis, innovative in helping to find other variables that could help explain variances in pricing disparities, and has a good command of English.