How scoring program "zatsenivaet" potential borrowers.

And so you rush to the bank with confidence that the bank will take a positive decision on granting you credit. It is a call from the bank and was informed that a preliminary decision on your online application for a loan (oh, the miracle of modern Internet!) Positive, ie such as a client you are even fit. What a great disappointment as when, leaving all the documents and answer the questions honestly bank manager, you are denied a loan. Without explanation. This is similar to the denial of visa to the Schengen welcome, but much more painful and confusing. What happens there, to deployed our bank monitors? As banks evaluate the reliability of the potential borrower? Increasingly, we hear the answer: the solution adopted scoring system. What is this - scoring, on which both depend on a positive decision on the loan? The encyclopedia we find: Scoring (born scoring) - a method of classification of all borrowers into different groups to assess the credit risk is a mathematical or statistical model. Scoring is literally translated as earning, counting points. In 1941, when credit analysts in the U.S. mass called to the front, scoring methods for assessing borrowers allowed to compensate for the lack of qualified personnel. David Durand has applied the classification of Fisher "plants: good and bad" credit. Banks, having no capacity to allocate to each individual borrower credit expert to assess the risks of non-returns, have to rely on the scoring system, which allows you to process forms and applications ten times faster. The massive introduction of scoring programs began 10 years later in 50 years of the last century, after the card program "After dinner, subscribe to" system Diners Club. The advanced program not only take positive or negative decision on granting the loan, but also determines the size of the loan. Today, credit scoring can be regarded as the most high-tech, interesting and significant software product to serve retail customers. With it is evaluated not only the risk of loan to a borrower, but also reveal hidden patterns in the behavior of the borrower. In Russia, the emergence of scoring models evaluate the creditworthiness of the borrower was due began several years ago the present credit boom. Express loans - decision in 15 minutes! Half an hour - and the Russians went home happy with their purchases. Illusion that the credits are distributed to all and sundry. In fact scoring model has worked, although still poorly fell on Russia's soil. " Thus, the scoring - a mathematical analysis of the entered information. Potential borrower responds to questions - a bank clerk enters the information into the computer - the system assigns each account a certain ball. Construct mathematical models based on data on the behavior of the borrower in the past in order to determine the "willingness to pay of credit" and "opportunity in the future to repay the debt." The scoring model is based on the following parameters: personal data, financial and related information. Personal data: sex, age, marital status, presence / absence of dependents, education. Financial: The total length of service, overall experience at the last or last place of work, level of earned fees and overall costs. For related information, can be attributed to the presence of additional sources of income, owning a car, garage, villas, land in private ownership. Subject to achieving a certain amount of points, the system gives a positive answer. If the client get the score a little different from the required bars, scoring takes conditionally positive solution. In this case, the bank makes proring on these phones in the questionnaire. If the data match, the client receives a "good". If the confirmation is not received by phone (sometimes even by mistake), the client is denied a loan. But scoring - is not just a calculator points. Virtually any scoring model is a self-learning. It is with the accumulation of certain credit history with the launch of the program in action included the second phase of the analysis of information on potential borrowers. Scoring begins to compare with those clients who have already received credit in the bank and who is suitable for the specified parameters (gender, age, marital status, etc.) the potential borrower. If scoring reveals that out of 10 borrowers, similar to a new customer, only five successfully return a loan, the program assigns it to the problematic and risky group and, consequently, fails, zanesya denial of its base. If within a certain time scoring will continue to refuse to customers in this group, then the bank's specialists have the opportunity to adjust the "behavior" of the program. In addition, the scoring program analyzes and fixes the riskiest lending activities of certain groups of goods in a mode "online". For example, if the percentage of defaults and delays on loans on mobile phones is high, then this group is defined as a hazardous product, the requirements for the number of points increases, and the scoring redistributes loan portfolio in favor of lending to other goods. Thus, scoring analyzes the information received in those criteria, what it interesting, and on which depends the risk of non-repayment of the loan, and in real time. There is a feeling of paradox: a client may refuse a loan to buy a mobile phone, and a week later scoring will give a positive solution to obtain consumer credit at a much higher amount. Therefore, the resulting denial of some bank not need to be perceived as the end of the world and an insult to the end of days. Time does not stand still, scoring the program learns and stores information. It is possible that a system of criteria and values for it will change soon. Scoring - a program based on pure mathematics. For points, it does not see the real person. If the client is registered in the Moscow region in the garden a partnership, he counted the desired point - there is a permanent registration. But any bank manager understands that and then find a client at his place of residence will not be easy. The human factor is of great importance. Much depends on the skills of bank employees and their direct interest to grant credit. Managers, working with the scoring systems and culling data there to get a response, have long learned to enter the data so that the decision was positive, that is actually "cheat" a clever scoring. And under the force of the client to answer questions so as to obtain more points. Therefore, scoring may miss the bank "bad" client, and a good and faithful to reject Scoring models in all, without exception, banks - a mystery which nobody opens. It is known that the parameters of each standard scoring tool to assess the creditworthiness of individuals are always present: negative credit history facts, profession, duration of work in one place, home ownership, number of recent requests to the Bank, during the stay at the current address. There are simplified versions of, say, a small amount of credit a borrower without a specific purpose. In this case the scoring program is a simple form, which introduces passport data, some statements from the statement of earnings, etc. For the issuance of mortgage model is more complicated, to pay for a mortgage loan requires a higher permanent income and the adequacy of collateral. Thus parameters such as sex, age, education are significantly less weight than the same parameters when issuing credit cards. The types of scoring programs: Application-scoring - assessing the creditworthiness of borrowers to obtain credit. This is the most relevant type of scoring for Russia. Collection-scoring - the identification of priority cases and areas of work for borrowers, the state of credit accounts that are classified as "unsatisfactory". Using this type of scoring allows for smooth operation of in arrears until her transfer to the collection agency. For example, according to the results of several studies about 40% of all non-payment accounted for forgetful borrowers who, without any intention of forgetting to make a payment on the loan and "corrected" after the first reminder. Behavioral-scoring (behavioral scoring) - assessment of the dynamics of the credit account of the borrower. Behavioral scoring can predict the change in the solvency of the borrower, to determine the optimal limits for credit cards, etc. In Russia, almost never used because of the lack of scoring systems, capable of solving such problems. Fraud-scoring - evaluation of the likelihood of fraud potential borrower. Its relevance to Russia's market is sufficiently large. According to a number of domestic banks frank fraud is up 10% of all non-payments, and this figure each year continues to slowly but steadily increasing. Retail lending in Russia is slowly but surely be renewed after the year-old period of interdiction rates and the stringent conditions of extradition. It is noticeable that almost without exception, banks began to apply the scoring program. Because of the significant reductions in staff use of scoring important for banks than ever before. It is hoped that self-learning system of scoring will not be counted for each new customer to the risky group because of the increased amount of arrears on payments, and bank employees will see beyond the purely mathematical calculation of points of real bona fide borrower.