Method and apparatus for a model assessing debtor behavior

1. A non-transitory computer readable medium containing a program which, when executed by a processor, performs a method comprising:

aggregating at least one of account receivable data and payment data of a debtor (i) from each of a plurality of creditors (j); and

using the aggregated debtor data and debtor data from the plurality of the creditors (j) determining a probability of a credit or collection event (P(CCE ij )) of the debtor (i) with respect to at least one creditor of the plurality of the creditors (j);

storing, in a memory, values corresponding to the (P(CCE ij )), wherein the P(CCE ij ) is determined according to the following model:

P(CCE ij ) is a function of (AR ij , APP ij , OTHER i , BCDI i1 , BCDI i2 . . . BCDI iJ ); where (AR ij ) comprises accounts receivable data for debtor (i) with creditor (j); (APP ij ) comprises optional internal data for debtor (i) with creditor (j); (OTHER j ), comprises optional third party data on debtor (i) selected from at least one of a credit bureau, a demographic data source or a private data source; and (BCDI ij ) comprises Business Credit Data Interchange data elements for debtor (i) across all J creditors in the BCDI; and

making a credit decision utilizing the stored values corresponding to the (P(CCE ij )).

2. A non-transitory computer readable medium containing a program which, when executed by a processor, performs a method comprising:

adapting a debtor score at a first creditor model in response to account receivable data and payment data of the debtor received from a second creditor model, each of the creditor models being generated according to respective Business Credit Data Interchange (BCDI) group, wherein the debtor score comprises a probability of a credit or collection event (P(CCE ij )) of the debtor (i) with respect to at least one creditor of a plurality of the creditors (j), wherein P(CCE ij ) is a function of AR ij , APP ij , OTHER j , BCDI j1 , BCDI j2 . . . , BCDI ij : where AR comprises accounts receivable data for debtor (i) with creditor (APP ij ) comprises optional internal data for debtor (i) with creditor (j); (OTHER j ) comprises optional third party data on debtor (i) selected from at least one of a credit bureau, a demographic data source or a private data source; and (BCDI ij ) comprises Business Credit Data Interchange data elements for debtor (i) across all J creditors in the BCDI;

storing, in a memory, values corresponding to the debtor score; and

making a credit decision utilizing the stored values corresponding to the (P(CCE ij )).

3. The non-transitory computer readable claim 1 , further comprising:

calculating an expected utility loss (EUL) of a debtor with respect to at least one creditor according to the following equation:

EULij=P(CCEij)*ULij; where

utility loss (UL) is a function of one or more of balance outstanding and collection effort costs.

4. The non-transitory computer readable f claim 1 , further comprising:

receiving a creditor inquiry regarding a particular debtor; and

evaluating expected behavior of the particular debtor with respect to the inquiring creditor according to the P(CCE) data.

5. The non-transitory computer readable claim 4 , further comprising:

providing to the inquiring creditor a report regarding the evaluation of the respective debtor.

6. The non-transitory computer readable claim 2 , further comprising:

generating the account receivable data by performing a bivariate statistical analysis upon accounts receivable data elements within the BCDI group.

7. The non-transitory computer readable claim 6 , further comprising:

applying one or more transformations to the account receivable data to create predictive variables, the one or more transformations comprising any of logs, truncations, censoring, variances, averages, measures of volatilities, compound variables, missing variable assignments and the creation of dichotomous variables.

8. The non-transitory computer readable claim 7 , wherein:

said first creditor model is generated by processing candidate predictive variables in association with the dependent variables according to a selection technique.

9. The non-transitory computer readable f claim 8 , wherein the selection technique comprises a stepwise selection technique having associated with it a complexity penalty.

10. The non-transitory computer readable claim 9 , wherein the complexity penalty comprises a Schwartz Bayes Criterion.

11. The non-transitory computer readable claim 1 , wherein the account receivable data further comprises amount 1-30 days past due across creditors.

12. The non-transitory computer readable claim 1 , wherein the account receivable data further comprises total outstanding balance across creditors, amount past due across creditors, amount of write-off (loss) across creditors, amount placed for collection across creditors, and average days past due across creditors.

13. The non-transitory computer readable claim 2 , wherein the respective BCDI groups includes data relating to an amount 1-30 days past due across creditors.

14. The non-transitory computer readable claim 2 , wherein the respective BCDI groups includes further comprises data relating to total outstanding balance across creditors, amount past due across creditors, amount of write-off (loss) across creditors, amount placed for collection across creditors, and average days past due across creditors.