The Gaza Strip and the West Bank have approximately four million inhabitants, of which a large portion are consumers and potential borrowers from the banks, mail-order companies, businesses and mobile-communication providers et cetera. Nevertheless, an established credit inquiry agency with an automatic credit inquiry and a process for assessing financial standing, like Schufa in Germany, thus far does not exist. In light of this desideratum, the Central Bank in the Palestinian Territories, the Palestine Monetary Authority (PMA), in December 2009 explored new ways to optimize the credit lending procedures in the Gaza Strip and the West Bank. With the implementation of a credit risk service, the PMA entered hitherto unknown territory and to introduce “credit scoring” for the first time in this region.

The Creditinfo Schufa GmbH won the bid. The specialists for the development of risk-models Creditinfo Decision from Prague developed the score-card. The Schufa provided that Decision Support System (DSS) and took over the technical aspects of the implementation in cooperation with the allied company insiders and Creditinfo Decision. The decisive factor behind the decision of the PMA was mainly based on consideration of the experience of these institutions with credit-information-services and financial standing data, as well as the premium rate services, e.g. software solutions in different international economic spheres and banking practices.

Independent Scoring Process

The requirements for the credit system stipulated by the PMA were high: The installation of the system has to be integrated smoothly in the existing financial infrastructure, and the core processes of the bank should not be impaired. Beyond this there was an important requirement that credit system be independent of the credit systems of other companies and credit systems. The Central Bank demanded to exercise full control of the scoring-process as well as the calculation of the scoring variables, the score-values and to be able, if necessary, to change eventual risk-groups. Accordingly, any future changes in the calculation would not be implemented only months later, but be realized within a few weeks in order to ensure quick processing of both single-inquiries and bulk-data. To meet this objective it was necessary to make the calculation of the scores directly while producing the credit information. The system should thus be especially expeditious and efficient, easily operable for different departments and not requiring any further programming through IT. In order to reduce the duration of development until market maturity, the variables needed for developing the score-card should not first be calculated by statistic-software, but rather directly in the solution; this requirement was also realized and productively deployed. Another task was the Batch-Scoring of all the client-portfolios of the larger banks, which also should function without impairing core-functionalities of the credit inquiry agency currently in operation.

Coordination among Specialists on the Scene

In a very short time, a project-team of different specialists was put together. Only after a detailed official questioning of the team at the airport in Tel Aviv concerning the upcoming project, it was possible to continue to the operation-theater in Ramallah, where the different specialists started their respective work. The Schufa was responsible, among other matters, for DSS-Training in English. And since the technical departments in Arabic were used s to translate changes on risk-processes, score-cards et cetera from IT, the new means of the Decision Support System had first to be adapted. Six months later and in the course of the evaluation of the score-cards, another employee was trained directly in Prague. The technical aspects of the implementation were again the responsibility of the Schufa’s allied company insiders. They had initially to face the challenge to coordinate the development of the score-card as well as the calculation of the variables to ensure a smooth transition from development-period to the stipulated dead-line.

It was agreed upon that the input of raw data of the complete history of a person be in the domain of DSS. With regard to the format, the XML was relinquished due to considerations of its performance, and a smaller but flexible format was defined. The possible alternative that the DSS dynamically requests the data it needs from the Data Warehouse of the PMA during the processing was rejected; a calculation of the score-variables without external data enabled the development of the configuration of the DSS without direct access to the Data Warehouse of the PMA. Thereby any scenario from simple regression-tests during changes in the Champion Challenge- Process and detailed reports about the portfolio of the prospective barrower can also be transferred to local systems.

With the selected option, former score-decisions can be later reconstructed, on the one hand. The PMA can now, on the other hand, appropriate to the demands, test changes of the score-calculation against large data records, without impairing its Data Warehouse. The flexible integration and installation options of the Decision Support System allow an installation precisely at the necessary locations on the systems and in the system-structures of the PMA.

Score-card-development

In order to secure that the various tasks would be carried out simultaneously, so to speak, parallel to one another, it was decided at a launching meeting not to conduct, as usual, the calculation of the score-variables in a statistic-program and then to translate them into a scoring- system. There are two reasons for that:

Parallelization: While the employment of Creditinfo defined the variable, Schufa and insiders were already able to work together on the implementation of these variables.

This approach made it possible to conclude this phase of the program within two weeks only. As a consequence a DSS-configuration was ready for use, which calculated more than 300 relevant variables for the score-card-development within few milliseconds pro person on a commercially available PC. With the help of these 300 ready variables, the score-card for the PMA was produced in a statistic-program. The final score-card makes use of approximately another 20 of these variables (S. illustration).

Reason Codes complete soft factors

In addition, several Reason Codes were developed by Creditinfo, which, in case of low score-value, provide the official in charge with reasons for the poor evaluation and with that the negative credit-standing. Such an evaluation could be made on the basis, for example, of a score under 300 and in light of the financial transactions of the person in question during the past twelve months. Especially by scoring-inexperienced banks, the Reason Codes proved itself as an important complement to the score-values, which hitherto was regarded as sufficient to determine lending practices.

Upon establishing the criteria of the Reason Codes, as well as that of the risk-groups and the score-card, these criteria were employed within a day in the DSS. For the final development of this Reason Codes, however, the know-how of all the parties, especially those from PMA was required. Thus a meeting of all project-managers convened in March 2010 by Schufa in Wiesbaden. In this meeting, around 100 selected inquiries, inclusive of respective scores and Reason Codes, were examined and further Reason Codes defined and during the meeting in DSS implemented. In this way, the project-manager from PMA was able to adjust and control the future Reason Codes and attained immediately a tangible sense of the new system. The testing during the meeting was only possible through foregoing accesses to the Data Warehouse of the PMA.

Afterwards, the DSS-set of regulations were installed in the test-system of PMA, so that the technical department already could gather the first test-results and feedback.

In view of the total lifetime of a run of only around 50 Milliseconds, the calculation of 300 variables was not removed, but furthermore, besides the score, was calculated and sent to the Data Warehouse of PMA. Here they were temporarily stored for a later evaluation of the score-card.

Conclusion

The first evaluation of the score-card was concluded after six months. Changes in the calculation of the variables or the score-card will be carried out in the future by the technical department of the PMA, and followed by yet another evaluations by Schufa and insiders in an advisory capacity. Yet the cooperation with PMA is not over. The contract has a duration of six years. The first project is thus successfully implemented with mixed teams from three different countries, cultures and economical backgrounds, different in time, budget and quality.

Authors:

Thomas Bluhm is deputy head Technical Solutions of Schufa Holding AG.

Helge Richter is as product-manager DSS responsible for the development coordination and deployment of Decision Support System in insiders GmbH.

To Review the original report, please click here