|
|||||||||||
|
|
|||||||||||||
| Solutions > Mortgage Lending |
|
Scorto™ Mortgage Lending Solution 
While creating a credit scoring system for mortgage lending, Scorto™ focused on the specifics of this kind of crediting, which allowed us to invent and implement specialized scoring tools and techniques in addition to the standard scoring solution.
The main questions that arise during the granting of a credit are the following:
Creating Hierarchic (Tree-Like) Models of the Decision-Making Algorithm This tool is a convenient mechanism for expert data formalization and it can be used during any stage of the credit-granting process. This kind of analysis and decision-making strategies is often and successfully used in those fields, where expert data formalization based on mathematical methods is, for different reasons, difficult or entirely impossible. Mortgage lending is one of such fields, as information on credit histories is often missing, the time taken by the change of the borrower's status is quite long, and a large internal database of experimental and expert knowledge is available. Combined Work of Several Scoring Models that Have Different Logical Purposes - Evaluation of the Borrower's Creditworthiness, Evaluation of the Liquidity of the Collateral, the Borrower's Behavioral ModelsThe analysis and evaluation of the borrower for the issuance of mortgage loans is a non-linear, non-homogenous process that, most of the time, requires using a number of evaluations and indicators by the bank. The type and meaning of these evaluations may differ from one another very significantly. That is why the Scorto™ solution for mortgage lending supports several of the more popular types and formats of scoring models: Logistical Regression - a probability model for the evaluation of the probability of the occurrence of an event taking, used for the evaluation of the borrower's creditworthiness, as well as for other purposes, requiring a binary solution (yes/no, grant/not to grant, and so on).An Optimized Mechanism for the Collection of Borrower Information from Different Sources One of the important peculiarities of mortgage lending both during the granting of a credit and during its "life span" is the need to be constantly updated on the related developments with trustworthy information on the borrower and his collateral. All this requires receiving information from a number of external data sources. Specifically for dealing with external (and, often, remote) data sources in mortgage lending, Scorto™ has developed an approach that allows making the receipt of information as fast, transparent and scalable as possible. The Possibility of Long-Term Scoring Analysis Operations Long-term scoring analysis operations presuppose the existence of such stages in the stage-by-stage analysis of the borrower, which are fulfilled not immediately as in express crediting, but may take longer periods of time. For example, for receiving up-to-date information on a collateral and analyzing this information, an independent opinion by several experts may be required, which, of course, takes a certain amount of time. Scorto™ has developed a uniform approach, whereby the process of conducting long-term mortgage lending operations is an adjustable and formalized procedure. Using programming components and mechanisms for the creation and management of this kind of processes allows building a complex decision-making strategy without any specialized expert knowledge. A Readily Implemented Multi-Level Scheme of the Work of Credit Specialists/Security Service Staff In mortgage lending, just like in small-business crediting, decisions based on the opinion of one person are rare. Usually, the decision-making process involves 2 to 7 people - experts of different levels and specialization, from the security service to risk managers, to real estate experts/realtors. For the provision of their cooperation, a specifically developed role-management mechanism is used. The mechanism allows managing the types and levels of credit specialists. Each role (specialist) is provided with an individual interface for working with credit request data. |
    |