multiple hypothesis testing

I have a large dataset with 6 variables in time-series format (total, including time-series, is 2.1 million records, csv format). In addition, there is a small dataset that matches along one column, with a few other variables. Objective: describe what are the primary predictors for the binary variable "Bad." For details, please see the excel sample, which includes comments for each field. The objective isn't to build a predictive model going forward, but to explain the key factors driving loan default during the period 2008-2012. Default defined as "bad"=1. There are 3 hypotheses to be tested: 1) higher risk lending caused high defaults (higher risk = larger loan sizes, longer loan terms, different loan type (e.g. the Grp indicator) 2) over-concentration in certain districts caused defaults (measured by the penetration rate variable) 3) collapse of 4th lender (2nd tab) undermined performance for the other three, which would show up as a factor in the districts where it was most active (as borrowers realized that their neighbors weren't having to pay back). Summary stats per district are included for this lender in 2nd tab. The question is which of these three factors were most influential in causing subsequent default? Also, for the time-series, please test the above hypotheses separately, since influence of factors (esp. hypothesis 3) may have changed over time. For the initial analysis, please test the following time periods: Q1-08 Q3-08 Q1-09 Q3-09 Q1-10 Q1-11 Q1-12 Sample file is attached. In addition, analysis examining hypothesis 1 (impact of loan size, type & maturity on defaults) is also attached for your reference. This project is time-limited. I would like to have the bulk of the findings done by March 4 or March 5 latest, so I'll prioritize bids that can meet this deadline. If you can work on this full-time or near full-time starting monday, that would be ideal.