ASSESSING CREDIT RISK

Abstract

There is provided a computer-implemented method of assessing credit risk, the method comprising the steps of: obtaining non-sufficient funds declines data on past transactions of a plurality of account holders occurring during an outcome window; obtaining transactional data on the past transactions of said plurality of account holders occurring during an observation window that precedes the outcome window; deriving one or more rules based on the occurrence of non-sufficient funds declines during the outcome window and on the preceding transactional data from the observation window; and applying the one or more rules to transactional data on past transactions of an individual account holder occurring during an individual observation window for that individual in order to determine the likelihood of a non-sufficient funds decline occurring during a subsequent individual outcome window for that individual. There is also provided a computer readable medium and a system for assessing credit risk.

Claims

1. A computer-implemented method of assessing credit risk, the method comprising the steps of: obtaining non-sufficient funds declines data on past transactions of a plurality of account holders occurring during an outcome window; obtaining transactional data on the past transactions of said plurality of account holders occurring during an observation window that precedes the outcome window; deriving one or more rules based on the occurrence of non-sufficient funds declines during the outcome window and on the preceding transactional data from the observation window; and applying the one or more rules to transactional data on past transactions of an individual account holder occurring during an individual observation window for that individual in order to determine the likelihood of a non-sufficient funds decline occurring during a subsequent individual outcome window for that individual.

2. A method according to claim 1, wherein the step of applying the one or more rules comprises setting a score for the individual account holder.

3. A method according to claim 1, wherein the past transactions are debit card transactions.

4. A method according to claim 1, wherein each observation window represents a period of twelve months.

5. A method according to claim 1, wherein each outcome window represents a period of six months.

6. A method according to claim 1, wherein the step of deriving the one or more rules is based on an adaptive algorithm.

7. A method according to claim 1, wherein the transactional data comprises information on the geographical location and/or merchant category of the past transaction.

8. A method according to claim 1, wherein the method further comprises a step of assessing whether the likelihood of a non-sufficient funds decline occurring during a subsequent individual outcome window for that individual is over a predetermined threshold.

9. A computer readable medium, the computer readable medium comprising: a memory device comprising a set of program instructions which, when executed by a processor, causes the processor to perform the method steps of claim 1.

10. A system for assessing credit risk, the system comprising: a processor; and a memory; wherein the processor and the memory are in communication with one another, wherein the memory is configured to store program instructions, and wherein the processor is configured to execute the program instructions to perform the method steps of claim 1.

11. A system for assessing credit risk, the system comprising: a processor; and a memory; wherein the processor and the memory are in communication with one another, wherein the memory is configured to store program instructions, and wherein the processor is configured to execute the program instructions to perform the method steps of claim 2.

12. A system for assessing credit risk, the system comprising: a processor; and a memory; wherein the processor and the memory are in communication with one another, wherein the memory is configured to store program instructions, and wherein the processor is configured to execute the program instructions to perform the method steps of claim 3.

13. A system for assessing credit risk, the system comprising: a processor; and a memory; wherein the processor and the memory are in communication with one another, wherein the memory is configured to store program instructions, and wherein the processor is configured to execute the program instructions to perform the method steps of claim 4.

14. A system for assessing credit risk, the system comprising: a processor; and a memory; wherein the processor and the memory are in communication with one another, wherein the memory is configured to store program instructions, and wherein the processor is configured to execute the program instructions to perform the method steps of claim 5.

15. A system for assessing credit risk, the system comprising: a processor; and a memory; wherein the processor and the memory are in communication with one another, wherein the memory is configured to store program instructions, and wherein the processor is configured to execute the program instructions to perform the method steps of claim 6.

16. A system for assessing credit risk, the system comprising: a processor; and a memory; wherein the processor and the memory are in communication with one another, wherein the memory is configured to store program instructions, and wherein the processor is configured to execute the program instructions to perform the method steps of claim 7.

17. A system for assessing credit risk, the system comprising: a processor; and a memory; wherein the processor and the memory are in communication with one another, wherein the memory is configured to store program instructions, and wherein the processor is configured to execute the program instructions to perform the method steps of claim 8.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Certain preferred embodiments of the present invention will now be described by way of example only and contrasted to comparative examples with reference to the accompanying drawings, in which:

[0023] FIG. 1 is a flow diagram showing the steps of a method in accordance with an embodiment of the present invention;

[0024] FIG. 2 is a flow diagram showing the steps of a method in accordance with another embodiment of the present invention;

[0025] FIG. 3 represents exemplary time windows that are applicable to some embodiments of the present invention; and

[0026] FIG. 4 is a schematic diagram of the system for assessing credit risk in accordance with an embodiment of the present invention.

DETAILED DESCRIPTION OF THE DRAWINGS

[0027] Issuing financial institutions (or issuers) are institutions that provide financial services for its customers; for example, debit or credit account services. An account holder is therefore a customer of the issuer and each account holder carries a certain amount of credit risk to the issuer. Thus, issuers may assess the credit risk of individual account holders using exemplary methods of the present invention in order to assist in the decision making process of whether to issue credit to a particular individual.

[0028] FIG. 1 shows an exemplary flow diagram 10 representing a method of assessing credit risk, and its steps will be described in further detail below.

[0029] The first step is obtaining non-sufficient funds declines data on past transactions of a plurality of account holders occurring during an outcome window (S11). It is preferable that the non-sufficient funds declines data are collated over a large number past transactions (say, ten thousand transactions) so that global trends may be reliably extracted from the data set. It is also preferable that the past transactions from which the non-sufficient funds declines data are obtained are debit card transactions as it is believed that the occurrence of non-sufficient funds declines on debit card transactions are a good indicator of whether an account holder is experiencing financial distress and therefore whether they represent a high credit risk to issuers.

[0030] The second step is obtaining transactional data on the past transactions of said plurality of account holders occurring during an observation window that precedes the outcome window (S12). Transactional data may comprise information associated with specific past transactions, for example, information on the geographical location or merchant category of the past transaction, or the amount or frequency of transactions of the account holder who performed the past transaction. It will be appreciated that the observation window and the outcome window (and similarly the individual observation window and individual outcome window) may never overlap; however, there may be a period of time between the observation window and the outcome window, for example, one month. It will also be appreciated that the observation window and the individual observation window may overlap, or the individual observation window may occur immediately subsequently to the observation window, or the individual observation window may occur subsequently and at a period of time later than the observation window. It will further be appreciated that the outcome window and the individual outcome window may not overlap, but the individual outcome window may occur immediately subsequently to the outcome window, or the individual outcome window may occur subsequently and at a period of time later than the outcome window

[0031] The third step is deriving one or more rules based on the occurrence of non-sufficient funds declines during the outcome window and on the preceding transactional data from the observation window (S13). As mentioned above, the availability of large data sets may dictate how accurately these rules reflect actual non-sufficient funds declines trends in the current market. In this step, the one or more rules comprise at least one algorithmic model based on the correlation between normalised indices of card activity of the plurality of account holders, and they may be based on trends of particular geographical location, merchant category, or demographic, etc., or alternatively, the one or more rules may be based across a range of geographic locations, merchant categories and/or demographics etc. In one particular example, a rule may be derived in that London-based transactions represent high likelihood of a non-sufficient funds decline occurring during a subsequent individual outcome window for an individual. In another particular example, a rule may be derived in that transactions over 1000 represent high likelihood of a non-sufficient funds decline occurring during a subsequent individual outcome window for an individual. In yet another particular example, a rule may be derived in that a sequence of transactions, for example across a range of different merchant categories, within a short period of time represents high likelihood of a non-sufficient funds decline occurring during a subsequent individual outcome window.

[0032] The fourth step is applying the one or more rules to transactional data on past transactions of an individual account holder occurring during an individual observation window for that individual in order to determine the likelihood of a non-sufficient funds decline occurring during a subsequent individual outcome window for that individual (S14). The one or more rules may be applied at account level (i.e. each individual account holder is assessed separately) or more generally the one or more rules may be applied for categories of account holders (for example, account holders may grouped by geographical locations at postcode level).

[0033] FIG. 2 shows another exemplary flow diagram 20 representing a method of assessing credit risk, and its steps will be described in further detail below.

[0034] The first step is obtaining non-sufficient funds declines data on past transactions of a plurality of account holders occurring during an outcome window (S21). The second step is obtaining transactional data on the past transactions of said plurality of account holders occurring during an observation window that precedes the outcome window (S22). The third step is deriving one or more rules based on the occurrence of non-sufficient funds declines during the outcome window and on the preceding transactional data from the observation window (S23). Details for each of steps S21 to S23 of FIG. 2 may correspond to the above description for steps S11 to S13 of FIG. 1.

[0035] The fourth step is applying the one or more rules to transactional data on past transactions of an individual account holder occurring during an individual observation window for that individual and setting a score for the individual account holder in order to determine the likelihood of a non-sufficient funds decline occurring during a subsequent individual outcome window for that individual (S24). The one or more rules define at least one algorithmic model which attaches a score to represent propensity of an individual account holder to experience non-sufficient funds declines. The score could be a number, a percentage or a grade. In one example, the score could be a percentage (with 0% representing low credit risk and 100% representing high credit risk). In another example, the score could be a grade between A and D (with A representing low credit risk and D representing high credit risk). The predetermined threshold could, for example, be 60% or a grade B.

[0036] The fifth step is assessing whether the likelihood of a non-sufficient funds decline occurring during a subsequent individual outcome window for that individual is over a predetermined threshold (S25) i.e. a binary outcome is produced.

[0037] In a first scenario, the likelihood of a non-sufficient funds decline occurring during a subsequent individual outcome window for that individual is over the predetermined threshold and so the issuer does not issue a credit card (S26A). Although not shown in FIG. 2, the issuer may alternatively choose to impose a different measure such as reducing credit limit on an existing account.

[0038] In a second scenario, the likelihood of a non-sufficient funds decline occurring during a subsequent individual outcome window for that individual is not over the predetermined threshold and so the issuer issues a credit card (S26B). Although not shown in FIG. 2, the issuer may alternatively choose to impose a different measure such as increasing credit limit on an existing account.

[0039] FIG. 3 represents exemplary time windows that are applicable to some embodiments of the present invention, for example, the embodiments of FIG. 1 or FIG. 2, on a timeline 30. In FIG. 3, the timeline 30 shows intervals of time (in months) between T=18 and T=12, and the present day is indicated at T=0. Non-sufficient funds declines data is obtained during outcome window 31 and transactional data is obtained during observation window 32. After one or more rules are derived based on the occurrence of non-sufficient funds declines during the outcome window 31 and on the preceding transactional data from the observation window 32, said one or more rules are applied to transactional data on past transactions of an individual account holder occurring during an individual observation window 33 in order to determine the likelihood of a non-sufficient funds decline occurring during a subsequent individual outcome window 34 for that individual.

[0040] In the example of FIG. 3, outcome window 31 runs from T=6 to T=0, individual outcome window 34 runs from T=0 to T=6, and each the outcome window and the individual outcome window 34 represent a period of six months. Observation window 32 runs from T=18 to T=6, individual observation window 33 runs from T=12 to T=0, and each the outcome window 32 and the individual outcome window 33 represent a period of twelve months.

[0041] The one or more rules may be derived based on an adaptive algorithm or based on an algorithm with adaptive variants. In other words, the algorithm may change its behaviour based on updated or real-time information or the algorithm parameters are automatically adjusted as more data is collated as time progresses past T=0. As time progresses, the outcome window 31, observation window 32, the individual observation window and the individual outcome window may shift correspondingly in a way such that their relationship with one another remain constant, i.e. outcome window 31 begins when observation window 32 ends, individual outcome window 34 begins when individual observation window 33 ends, and outcome window 31 and individual observation window 33 end simultaneously.

[0042] It is preferable that the outcome window 31 and the individual outcome window 34 are of equal time periods and also that the observation window 32 and the individual observation window 33 are of equal time periods. This makes it easier for the one or more rules that have been derived from data obtained during the outcome window 31 and observation window 32 to be applied to the individual observation window 33 for determining an outcome in individual outcome window 34. Consequently, time periods a (the aggregate time period of the outcome window 31 and the observation window 32) and 13 (the aggregate time period of the individual outcome window 34 and the individual observation window 33) are also preferably equal.

[0043] FIG. 4 shows an exemplary system 40 for assessing credit risk in accordance with an embodiment of the present invention, the system 40 comprising a processor 41 and a memory 42, wherein the processor 41 and the memory 42 are in communication with one another, wherein the memory 42 is configured to store program instructions 43, and wherein the processor is configured to execute the program instructions 43 to perform the method steps as described above with reference to FIGS. 1 and 2.

[0044] It will be appreciated that the flowcharts described above should not be understood to prescribe a fixed order in which the method steps must be performed, rather, the method steps may be performed in any order that is practicable. It will also be appreciated that features that are described in the context of separate embodiments may be provided in combination in a single embodiment, and conversely, features that are described in the context of a single embodiment may also be provided separately or in any suitable sub-combination.