Invisible bonuses: FinTech library with other resources

人大金融科技研究所 view 33 2021-11-3 13:34
share to
Scan QR code with WeChat

The paper found that using anonymous data management tools provided by fintech platforms to examine whether using other data to measure credit can increase access to credit. Comparing the actual results of the fintech platform model with the negative implications of the “traditional model” used for management purposes, the ending resulted in a 60% increase in denial of results and higher interest rates. The borrowers most affected are the “Invisible Primes” - borrowers with low interest rates and a short loan history, but with bad credit defaults. This line shows that granting loans to these lenders will result in better returns for lenders and higher repayments of financial instruments.

introduction

With the rise of online intermediaries and fintech companies, the lending industry is having a big impact. An important characteristic of financial technology companies is the change in face-to-face interactions between lenders and lenders with algorithms and other data. An online business enables a new business to cut costs and realize significant business benefits through its loan products. For example, Quicken Online Loans is the largest lender in the United States and fintech lends monthly to the private lending industry. Despite their growing importance, a clear understanding of how these new middlemen affect their ability to borrow money and the health of the family is still lacking.Traditionally, beneficiaries with high credit scores have more options with low interest rates, they have the same confidence, but the results of fintech loans, if any, indicate that the limited and profitable options available by individuals can change this situation.

The emergence of this new type of mediator has created many problems related to politics.A major problem is the impact on the ability of loan models to use other data and algorithmic arguments.Other titles and multiple ways to raise money can lower start-up costs and lower interest rates. Other writing standards can also identify people currently being ignored by standards such as credit scores. Richard Cordray, former director of the Consumer Financial Protection Agency, said: "Adding this exchange information to our records should give credit to millions more consumers," outside of credit. According to Fair Isaac Corporation, chief credit officer, the profile data of 28 million Americans is not enough to establish a credit score, and 25 million Americans have no credit profile at all. The response of the Consumer Financial Protection Bureau (CFPB) is to encourage lenders to develop new ways of being fair, equitable and impartial in terms of credit, especially for those who do not see credit and for those who have low or low loan history. .

Samples

This information is a process of decensoring information maintained by large financial technology companies operating in the private lending industry that includes information on applicant approvals and rejection. Pin up.

Founded in 2012, Upstart is one of the few open online lending companies and one of the few lenders to lenders across the United States. Upstart's underwriting model differs from that used by traditional borrowers, and its cost algorithm includes other data from non-traditional products. Upstart appealed to the CFPB in 2017 to avoid any conflict with existing rules and received a notice of inaction. The CFPB identifies the Upstart security underwriting model and its features, and compares the benefits (approval and interest rate) with those benefits. It is based on a real model that does not use surrogate data. The CFPB found that the review did not warrant any monitoring or oversight of Upstart. By the end of 2020, Upstart will be approved for an additional three years under the NAL plan.

number of lettersinformation201It begins in the fourth year and ends in the first quarter of 2021. It contains information on the monthly loans of 900,000 loans initiated by Upstart and the characteristics of the loans at the beginning.This free number is limited to 770,523 Financial loans that provide the scores. In recent years, upsstart has increased more of loans and all loans. About 75,000 in 2017 increasing nearly 300,000 in 2020. Less than $ 100 million 2014 increased by 2020 for $ 3.5 billion. The block is directly on the website directly, but most of the loans are made of bonus loans from the cash loan from the cash loan. Traffic can search the platform to choose the needs. Library loans. Credit Karma is exposed to a variety of credit information and reduced cost of the application. The most loans to ensure the card recovery card. Upstart market in every week, but some cities' lenders are of the liver in Washington, California, Nevada and Colorado. Other ways of financial technology has also been found similar models. Cross the bank account without in this state because of the financial the financial an upstart issued IWA and West Virginia. Our administrative rules of the year as a difference between the difference between the difference between the intersection with the intersection with activity.

On average, Upstart loans are around $ 11,700. A different model of US $ 10,000 indicates that there is a difference between borrowers, and some older borrowers. The average contract has an interest rate of 22% per year and a term of 4 years. The borrower's average startup score is typically 653. The highest score is also slightly above 680, indicating that the Upstart goal isn't always for the most trusted people. Most borrowers are between 28 and 46 years old, with an average age of 37. The letter also deals with information on the income of borrowers. The average annual income is $ 67,000 (the standard difference is $ 173,000) and the debt-to-income ratio is approximately 18%. About 44% of borrowers have a university degree and work an average of five years. As an indicator of entry into the mortgage industry, lenders have an average of 18 savings accounts.

One of the main advantages of the information in this sentence is that it contains all the information requested by the requester and does not receive payment.On average, applicants earn 70 more points than non-applicants and earn $ 14,000 per year. The total debt and credit balance of beneficiaries (US $ 118,000) is also greater than bad debts (US $ 65,000). Mortgage lenders can also have access to a college education, less work, and often use computers and mortgages to pay off debts.

This information also provides access to the applicant's credit report for the first month. Upstart regularly provides information on outstanding loans and monthly information on non-performing loans. Upstart reserves the right to withdraw loan information for ineligible applicants within 12 months from the date of application, typically multiple times during this period, and finally approximately 12 months. The credit reports in the form on this line are all from the same credit union (TransUnion).

To provide evidence of external use, this article adds analysis with loan performance data from Moody's Analytics and Freddie Mac, as well as loan demand data from the House Law Act (HMDA). Moody's Analytics Data provides initial loan and monthly performance data for non-mortgage lenders.This sentence limits the model to a 30-year fixed mortgage and the sample duration is after 2000.

This first paragraph describes how your current credit score affects the lender. Your credit rating depends on your major expenses, your payments, the length of your credit history, and the type of credit you currently have. The higher the score, the lower the risk. Next, this article examines the strength prediction scores of the Upstart loan model. Specifically, Upstart loan data was used to estimate the balance using two key metrics (clearing and past due) based on different outcomes. Amortization is recorded on the loan balance after deducting the loss, and the deduction is defined as a non-payment of more than 90 days. The results showed that credit scores were not as good as traditional loans for estimating fintech defaults, especially for people with low credit scores. This line uses the RFE-RF process for a 3 year loan initiated by Upstart using 38 routine and non-routine variations. School level also remains one of the key predictors of Upstart's performance.The results show that the Upstart results below represent a significant improvement in credit score, and these improvements cannot be fully explained by the data in the report. The use of surrogate data is crucial.In this analysis, this sentence focuses on the 98,671 loans granted by the Standard Model. We analyzed the percentage of non-performing loans in the old model and compared the difference in interest rates expressed by the two models. Differences in structure affect the extent and variability of the structure.

Therefore

With the importance of fintech loans in the unsecured lending industry, it has become important to understand how loans work and the implications of participating. The lack of detailed control data regarding the functioning of fintech libraries has affected this understanding.To clarify this area, this sentence uses special data for large loans. This sentence shows that the evidence used by Upstart is more of an estimate than the credit score based on the negative result, the metric standard is still used to determine the mortgage lender's money.This sentence provides more information,Their excellent ability to estimate start-up costs leads to higher mortgage rates, especially for low credit lenders. These benefits are detected in general terms (e.g. whether the person is able to obtain credit) and the loan interest rate (i.e. the interest rate on which the lender is satisfied. of his loan). The beneficiaries of the fintech underwriting model are the borrowers with the lowest scores.Otherwise, you could be denied credit or face high interest rates on a regular basis.Giving credit to these people results in higher returns on Upstart, indicating that there may be a personal boost to adopting the new lending model. People with low credit scores also have less ability to deal with other debts (such as debit cards) and are more likely to improve their credit scores later.

The conclusion concludes the argument on the use of other documents by providing evidence of a positive impact on the loan approval process. While they do not address the controversy surrounding potential issues of confidentiality and statistical exclusion, the results of this article suggest that other data models offer benefits for both lenders and mortgage lenders.

btcfans公众号

Scan QR code with WeChat

Link
Disclaimer:

Previous: 央行数研所副所长狄刚:区块链在数字金融领域面临的六大技术挑战 Next: What do digital advantages mean for emerging markets and nation-building?

Related