When a bank wants to lend a loan to a customer, the general premise is that the customer and the bank have a long-term settlement relationship and a running account. More importantly, the account manager will chat with corporate finance to understand the business situation and capital needs of the enterprise. Traditionally, he does not deal with strange customers. When the enterprise meets certain conditions, banks begin to intervene in credit lending, including actively selling credit products to customers or actively applying for loans by customers. The borrower makes daily settlement through the loan bank, and the bank can find some information by looking at the account transaction (not all, more information depends on the daily chat and visit between the bank and the enterprise). For example, the borrower recently borrowed 6.5438+million yuan to buy a car with 6.5438+million yuan, so after the payment of 6.5438+million yuan, under normal circumstances, there will be car sales income coming in one after another, for example, if there is no income within one month, then the bank will be very nervous! ! !
There are also borrowers who pay taxes and utilities through banks, and wages are paid through banks. Banks can judge whether there is a major change in the operation of enterprises by observing whether the payment of enterprises is interrupted or obviously reduced.
Analyzing the account transaction process itself is a skill, and the process is closely related to the parameters of the banking system. This is different from a loan company without settlement business. They don't have a settlement network. Although the loan company can require customers to run water, on the one hand, the running water can PS, and the parameters of the running water format of different banks are very different. How does a loan company identify authenticity? Even if it is true, how to identify valid information? Moreover, the banking system is not regularly updated and upgraded, and many of the same subjects are entered in a variety of ways. Running water needs to be analyzed, but not all.
The so-called bank credit risk control is a thorough and meticulous understanding of every detail, not an empty theory like FRM. Therefore, if you want to control the bank's style, you must first be familiar with the bank's settlement system, public and private.
Many internet companies also have ways to control risks through some internet information, similar to human flesh search, using big data, data mining, machine learning, anti-fraud and other calculations to operate in batches. This is a meaningful attempt. Internet companies are burning money at present, and what a mature business model will be is still unknown. Big data is of course important, and as bankers, we often care about small data, small data related to our customers. Settlement data is similar to sampling, which extracts cash flow information that best represents the risk status of customers from thousands of variables of customers. Sometimes the cash flow analysis is done well, and it can be judged that the risk is 80%. Of course, some social network information of customers, such as Weibo and qq information, is sometimes important, so it should be regarded as an early warning information. For those small and micro loans, customers are at the bottom of society, not in the financial system, and there is no account, let alone settlement. They can only use the Internet to catch the blind, and it is better than nothing. It is safer for banks to give up these customers directly.
Familiar with guarantee. The first source of repayment has been discussed earlier. Let me talk about the second source of repayment.
Collateral: be familiar with all kinds of collateral, real estate, the types of real estate, and what are their policy risks? How to apply for mortgage registration? His arrest warrant is also fake. I've experienced it myself. Housing Authority colludes with borrowers to defraud hundreds of millions of loans! ! ! How to deal with the equity pledge, which government department accepts it? What if there are risks? What are the obstacles? How to manage the car mortgage? How to tow a car?
So bank risk control is these trivial things. A small detail is out of control, which is a loophole of several hundred million! ! !
2. Technology and management.
When I was young, I felt that I needed a major and studied many VBA\SAS\CFA\FRM\ risk case models. In fact, I found out later that we still need to be a team, manage and integrate resources. It is another kind of ability. Professional knowledge can be remedied, but the improvement of ability is not easy. He clearly knows what to do, but specific things need to be done. There is something wrong with the quality of his people, and no matter how strong the risk control system is, it will not help. Civil air defense and physical defense technology. At present, there is too much emphasis on technology, such as using big data modeling to screen credit customers and using behavior models for post-loan management. In fact, in banks, more emphasis is placed on the role of personality. People who are too smart are not suitable to be banks.
3. Risk management is essentially about people.
Now that the technology is developed, enterprises have gone to ERP, banks have gone to the credit management system, and with the Internet, big data is flooding. The gap between people has widened. When applying for loans from online mobile phones, banks also use big data to build models to manage loans. From the battle of primitive society to modern hacker attacks, it is similar to an arms race, with sophisticated anti-fraud means and upgraded fraud means. Credit still depends on the feelings between people. Without the cooperation and affection between banks and enterprises, it is hard to say that risk management is very strong. Let enterprises feel that this bank is worthy of respect, flesh and blood, and worth dealing with for a long time, rather than cold data and models. Once the big data system detects that the enterprise data indicators are unqualified, it immediately stops the credit line, draws loans, and breaks loans, forcing the enterprise to die. This kind of big data credit information has prevented temporary risks, but it has hurt enterprises.
4. Trends.
Will banks become Internet technology companies in the future? I feel that traditional banks, crowd tactics, community finance, banks around us and community banks still have room for survival. Wang next door wants to borrow some money and fill out a bunch of reports. The account manager of the outlet inputs a lot of data online, and the computer automatically searches Wang's activities all over the world (Weibo's speech, qq records, public comments, wearable device database to see when he gets up, where he eats, where he moves, whether he goes in or out of bad places, how much his heart beats, how much his pulse is, his health geometry), and then uses data mining, machine learning technology, and so on. A minute later, the machine said, how much can you batch? This mode is fast, fast and efficient, and machine learning means that people work for machines. Even in the future, we don't need manpower to input information. Self-service loan machine, indeed, we don't even trust the living people around us. On the contrary, we must rely on machines to understand a person. What is Wang's personality? Neighbors don't decide, machines decide, and you can accumulate experience and increase your wisdom. The growth rate of an approver's experience is far behind the accumulation of machine learning wisdom.
The gap between people is getting deeper and deeper.
To borrow a fashionable phrase, "Without trust, there is no finance. Internet has lowered the threshold of financial access, but the threshold of trust is always there. The foundation of financial development is based on "trust", and the threshold of trust has always been there. Only when financial institutions gain the trust of customers through services can they have the opportunity to carry out finance. " As for how to gain trust, it's not just technology.
Some loan companies have found some big data scientists and internet scientists, saying that they can replace banks? But I think, do what you can and earn the money you deserve! ! !
In the short term, we don't understand that cross-border dividends exist, but in the long run, they are definitely balanced.
There will be professional credit reporting companies in the future. They are big data scientists and internet experts who specialize in credit reporting, serving not only credit reporting companies, but also other private surveys. Of course, there will be a charge. There will be many different credit companies, and these scientists will compete with each other, resulting in a balanced price. Because once the model is mature, the marginal cost of these jobs is basically very low, and it is almost zero cost to query one more copy. Therefore, in the future of this industry, once technology companies compete with each other and the price goes to zero, the banks will finally benefit. Banks will not produce ATMs themselves, and ATM manufacturers will compete. ATM can't replace banks, but banks use ATMs. So banks don't have to be chaotic, and professional things should be done by professional institutions. As a financial practitioner, we don't want to be data scientists, but what we can do is to serve real people.
5. Be quite familiar with policies and regulations.
When you take risks, you have to deal with the law, and laws and regulations often change. Sometimes a casual change will lead to many business innovations. For example, on August 6, 2005, the Supreme People's Court issued the Provisions of the Supreme People's Court on Several Issues Concerning the Application of Laws in the Trial of Private Lending Cases, the core of which is the gradual legalization of direct financing channels between enterprises, the abolition of the four-fold interest rate ceiling standard, and the legality of peer-to-peer lending platform guarantee. You may not care much about this thing, but it has a great influence on the credit model. This judicial interpretation clarifies the legality of corporate lending, but sadly, the lending between enterprises has not been included in PBOC credit reporting, and it is technically impossible to be included! ! ! Banks that rely on the credit information system will not be able to grasp the actual liabilities of enterprises. Moreover, the personal lending behavior of an enterprise legal person or person in charge may require the enterprise to bear responsibilities, and these responsibilities cannot be reflected in the financial statements of the enterprise, which increases the difficulty of bank credit investigation.
6. Very familiar with government procedures.
Public security, land, housing management, industry and commerce, taxation, insurance, securities, social security, postal services, finance, telecommunications and other departments.
7. To work in a bank, you must know how to socialize.
Many people will say that they are not sociable, so they hide in the bank logistics to take risks, while the branch's risk control has to deal with customers, so they hide in the branch to do risk control, and the branch has to deal with customers, so they hide in the head office. As for the head office, it needs more management and coordination ability, and it is more cautious than the grassroots to communicate with the CBRC government for instructions, reports, guidance, research, speeches and attendance. Many topics are just introductions, specifically for scientists in scientific research institutes.