Computer-implemented lending analysis systems and methods

• US 7,805,345 B2
• Filed: 08/26/2005
• Issued: 09/28/2010
• Est. Priority Date: 08/26/2005
• Status: Active Grant

• Current Assignee: SAS Institute Incorporated
• Original Assignee: SAS Institute Incorporated
• Inventors: Abrahams, Clark R., Zhang, Mingyuan
• Primary Examiner(s): Cheung; Mary
• Assistant Examiner(s): Holly; John H
• Application Number:
  US11/212,289
• Publication Number:
  US 20070050286A1
• Time in Patent Office: 1,859 Days
• Field of Search: 705/1, 705/35, 705/37, 705/80, 705/51, 705/57, 705/38, 510/301, 510/101, 510/446, 707/101, 707/102, 707/3
• US Class Current: 705/35
• CPC Class Codes: G06Q 40/00 Finance; Insurance; Tax strategies; Processing of corporate or income taxes
• CPC Class Codes: G06Q 40/02 Banking, e.g. interest calculation or account maintenance credit or loans G06Q40/03
• CPC Class Codes: G06Q 40/03 Credit; Loans; Processing thereof
• CPC Class Codes: G06Q 40/04 Trading; Exchange, e.g. stocks, commodities, derivatives or currency exchange

Abstract

In accordance with the teachings described herein, computer-implemented lending analysis systems and methods are provided. A pre-processing module may be used to organize loan applicant data into a plurality of applicant groups based on one or more demographic factors, wherein a protected class is identified from the plurality of applicant groups. An index generation module may be used to calculate a plurality of disparity indices for the protected class by comparing lending-related data for the protected class with lending-related data for one or more control groups selected from the applicant groups. An indicator generation module may be used to calculate one or more singular indicators for the protected class by comparing the disparity indices with one or more reference indices. The indicator generation module may be further used to calculate a global indicator as a function of a plurality of singular indicators.

Claims

1. A computer-implemented method for analyzing lending data, comprising:

processing, using one or more processors, loan applicant data into a plurality of applicant groups based on one or more demographic factors;

selecting one or more control groups from the plurality of applicant groups, each control group having corresponding lending-related data;

identifying a protected class within the plurality of applicant groups, the protected class having corresponding lending-related data;

calculating a plurality of disparity indices for the protected class, wherein calculating includes comparing the lending-related data for the protected class with lending-related data for one or more control groups;

calculating one or more reference indices;

calculating one or more singular indicators for the protected class, wherein calculating includes converting the plurality of disparity indices and the one or more reference indices into coefficient values, and comparing the coefficient values using one or more distance formulas; and

calculating a global indicator using a plurality of singular indicators, wherein a global indicator represents a relative comparison amongst groups of singular indicators.

2. The method of claim 1, wherein the global indictor is calculated using a Euclidean distance formula.

3. The method of claim 1, wherein the global indictor is calculated as a weighted average of the plurality of singular indicators.

4. The method of claim 1, wherein a reference index is calculated for each of the plurality of disparity indices.

5. The method of claim 1, wherein the one or more demographic factors include at least one of race, geographic location, gender, income level, color, ethnicity, religion, national origin, marital status, handicap, disability, familial status, age, or receipt of public assistance income.

6. The method of claim 1, wherein the one or more singular indicators are calculated using a distance analysis technique.

7. The method of claim 6, further comprising:

determining linear coefficients for the plurality of disparity indices and the one or more reference indices; and

calculating the one or more singular indicators as a difference measure between the linear coefficients for the one or more reference indices and the linear coefficients for the plurality of disparity indices.

8. The method of claim 7, wherein the linear coefficients arc calculated using a multivariate analysis to represent the indices with principal components.

9. The method of claim 8, wherein the one or more singular indicators are Mahalanobis as distance measures.

10. The method of claim 1, wherein the one or more reference indices are calculated using pre-defined criteria.

11. The method of claim 1, wherein the one or more reference indices are calculated using peer or industrial composite data.

12. The method of claim 1, wherein the one or more reference indices are calculated by assuming a value of 1.0 for one or more of the disparity indices.

13. The method of claim 1, wherein the one or more reference indices are calculated using average values for one or more of the disparity indices.

14. The method of claim 1, wherein the one or more reference indices are determined using average values for one or more of the disparity indices across a plurality of lending institutions.

15. The method of claim 1, wherein the one or more reference indices are calculated by assuming a maximum or minimum value for one or more of the disparity indices.

16. The method of claim 1, wherein the one or more reference indices are calculated using simulations.

17. The method of claim 1, further comprising:

validating the loan applicant data in one or more of the applicant groups to determine if one or more data fields are missing or outside of an expected range.

18. The method of claim 17, further comprising:

replacing data fields that are missing or outside of the expected range with one or more default values.

19. The method of claim 1, wherein the step of processing loan applicant data includes:

defining a handle to represent an applicant group, wherein the handle is defined based on a plurality of lending-related factors regarding the applicant group.

20. The method of claim 19, wherein the handle is a borrower risk based handle, and wherein the plurality of lending-related factors include at least one of debt-to-income range, loan-to-value ratio range, credit bureau score, and credit history.

21. The method of claim 19, wherein the handle is a transaction based handle, and wherein the plurality of lending-related factors include at least one of pricing option, term to maturity, loan amount tier, and down payment amount.

22. The method of claim 19, wherein the handle is a property-based handle, and wherein the plurality of lending-related factors include at least one of age, type of structure on a property, property location, appraisal method, and flood zone.

23. The method of claim 1, further comprising:

generating an interface capable of displaying representations of the one or more singular indicators and the plurality of disparity indices.

24. The method of claim 1, wherein at least one of the disparity indices is an origination disparity index, and wherein the origination disparity index is a rate of loan origination for a control group relative to a rate of loan origination for the protected class.

25. The method of claim 1, wherein at least one of the disparity indices is a fallout disparity index, and wherein the fallout disparity index is a rate of loan fallout for the protected class relative to a rate of loan fallout for a control group.

26. The method of claim 1, wherein at least one of the disparity indices is a market penetration index, wherein a control group represents an entire population of a minority group within a geographically defined area, wherein the protected class represents loan applicants of the minority group within the geographically defined area, and wherein the market penetration index is a population size of the minority group within the geographically defined area relative to a loan application rate of the minority group with the geographically defined area.

27. The method of claim 1, wherein at least one of the disparity indices is a product disparity index, and wherein the product disparity index is a rate of granted loans above a set dollar amount for the protected class relative to a rate of granted loans above the set dollar amount for a control group.

28. The method of claim 1, wherein at least one of the disparity indices is a sub-prime product application mix disparity index, the sub-prime product application mix disparity index being calculated as a first ratio divided by a second ratio, the first ratio being a proportion of sub-prime loan applicants of the protected class relative to a total number of loan applications for the protected class, the second ratio being a proportion of sub-prime loan applicants of a control group relative to a total number of loan applications for the control group.

29. The method of claim 1, wherein at least one of the disparity indices is a sub-prime product organization mix disparity index, which is calculated as a first ratio divided by a second ratio, the first ratio being a proportion of sub-prime loans granted to the protected class relative to a total number of loans originated by the protective class, the second ratio being a proportion of sub-prime loans granted to a control group relative to a total number of loans originated by the control group.

30. The method of claim 1, wherein at least one of the disparity indices is a loan-to-value disparity index, which is a proportion of loan applicants in a control group having a loan-to-value greater than a predetermined threshold relative to a proportion of loan applicants in the protected class having a loan-to-value greater than the predetermined threshold.

31. The method of claim 1, wherein at least one of the disparity indices is a bureau score disparity index, which is a ratio of loan applicants in a control group that have a FICO bureau score below a predetermined threshold relative to loan applicant in the protected class that have a FICO bureau score below the predetermined threshold.

32. The method of claim 1, wherein at least one of the disparity indices is an income disparity index, which is a ratio of loan applicants in a control group having an income below a predetermined threshold relative to loan applicants in the protected class having an income below the predetermined threshold.

33. The method of claim 1, further comprising:

displaying representations of the one or more singular indicators, the global indicator, and the plurality of disparity indices using a graphical interface, wherein drilling down upon a global indicator reveals one or more single indicators, and wherein drilling down upon a single indicator reveals one or more disparity indices in the plurality of disparity indices.

34. The method of claim 1, wherein singular indicators are grouped by geographical information such as Metropolitan Statistical Area (MSA), state, or channel.

35. The method of claim 1, wherein singular indicators are grouped by loan type or property type.

36. The method of claim 1, wherein singular indicators are grouped by geographical information such as Metropolitan Statistical Area (MSA), state, county, congressional district, neighborhood, zip code, census tract, or census tract grouping including high, middle, moderate, or low income tiers within state or MSA boundaries.

37. The method of claim 1, wherein singular indicators are grouped by loan purpose, loan type, owner occupancy, channel, or property type.

38. The method of claim 1, wherein singular indicators are grouped by organizational information corresponding to an institution, a legal entity/subsidiary, a line of business, a processing center, a branch, a sales office, a channel, an operating unit, or a loan officer.

39. The method of claim 1, wherein singular indicators are grouped by borrower, collateral, or transaction data categorizations including loan-to-value ratio, debt-to-income ratio, amount of liquid assets, savings rate, collateral price volatility, credit history, credit score, handle segment identifier, type of loan pricing, or loan term.

40. A system for analyzing lending data, comprising:

a processor;

a computer-readable storage medium containing software instructions executable on the processor to cause the processor to perform operations including;

processing loan applicant data into a plurality of applicant groups based on one or more demographic factors;

selecting one or more control groups from the plurality of applicant groups, each control group having corresponding lending-related data;

identifying a protected class within the plurality of applicant groups, the protected class having corresponding lending-related data;

calculating a plurality of disparity indices for the protected class, wherein calculating includes comparing the lending-related data for the protected class with lending-related data for one or more control groups;

calculating one or more reference indices;

calculating one or more singular indicators for the protected class, wherein calculating includes converting the plurality of disparity indices and the one or more reference indices into coefficient values, and comparing the coefficient values using one or more distance formulas; and

calculating a global indicator using a plurality of singular indicators wherein a global indicator represents a relative comparison amongst groups of singular indicators.

41. A computer-readable storage medium encoded with instructions that when executed, cause a computer to perform a method for analyzing lending data, comprising:

processing loan applicant data into a plurality of applicant groups based on one or more demographic factors;

selecting one or more control groups from the plurality of applicant groups, each control group having corresponding lending-related data;

identifying a protected class within the plurality of applicant groups, the protected class having corresponding lending-related data;

calculating a plurality of disparity indices for the protected class, wherein calculating includes comparing the lending-related data for the protected class with lending-related data for one or more control groups;

calculating one or more reference indices;

calculating one or more singular indicators for the protected class, wherein calculating includes converting the plurality of disparity indices and the one or more reference indices into coefficient values, and comparing the coefficient values using one or more distance formulas; and

calculating a global indicator using a plurality of singular indicators, wherein a global indicator represents a relative comparison amongst groups of singular indicators.

42. A computer-implemented method for analyzing lending data, comprising:

processing, using one or more processors, loan applicant data into a plurality of applicant groups based on one or more demographic factors;

selecting one or more control groups from the plurality of applicant groups, each control group having corresponding lending-related data;

identifying a protected class within the plurality of applicant groups, the protected class having corresponding lending-related data;

calculating a plurality of disparity indices for the protected class, wherein calculating includes comparing the lending-related data for the protected class with lending-related data for one or more control groups;

calculating one or more reference indices;

calculating one or more singular indicators for the protected class, wherein calculating includes convening the plurality of disparity indices and the with one or more reference indices into coefficient values, and comparing the coefficient values using one or more distance formulas;

calculating a global indicator using a plurality of singular indicators, wherein a global indicator represents a relative comparison amongst groups of singular indicators; and

displaying representations of the one or more singular indicators, the global indicator, and the plurality of disparity indices using a graphical interface, wherein drilling down upon a global indicator reveals one or more single indicators, and wherein drilling down upon a single indicator reveals one or more disparity indices in the plurality of disparity indices.