Today's Headlines Are:
Alternative data gets a thumbs up
Machine Learning Applications in Assessing Credit Risk
The pros and cons of using machine learning to prevent fraud
P2P lending hits a speed bump
Popularity of P2P lending high in Asia
As digital lending moves from frill to necessity – where are we at?
Barclays begins deploying funding through MarketInvoice
New finance council launched to help SMEs ahead of Brexit.
It’s been five years since the Financial Conduct Authority (FCA) took over regulation of the P2P lending sector, but in some ways it feels like it is just getting started. P2P lenders are bracing themselves for a number of regulatory changes later this year.
The new rules require platforms to introduce appropriateness tests and restrict their marketing to sophisticated and high-net-worth investors, people receiving regulated investment advice, or those who certify that they will not invest more than 10 per cent of their portfolio in P2P. Other key rules include ramped-up disclosure requirements and more clarity around wind-down plans in the event that a platform goes under.
However, there is still some way to go. For instance, the concept of a loan can be easily misrepresented in the current working of the rules. In P2P, loans can sometimes be called loan parts, or contracts, or investments. From the regulator’s point of view, they are treating the lender as a loan provider, but this may not always be true – the lending platform could be trading in loan parts, or acting as an aggregate or intermediary between borrowers and lenders.
Alternative Data Gets a Thumbs Up
Alternative data has the ability to increase access to credit and decrease its cost. But, given how important this is for lenders and borrowers alike, regulators have trodden carefully when approaching the use of alternative data for credit decisions.
In 2017, the Consumer Financial Protection Board (CFBP) requested information from the industry to learn more about how lenders are using alternative data. This signalled to the industry that the CFPB might consider future activity to encourage responsible use of alternative data and lower unnecessary barriers to its use.
The CFPB recently issued an update on credit access. The results show that alternative data and machine learning models to issue credit decisions approves 27 percent more applicants than the traditional model, and yields 16 percent lower average APRs for approved loans. This reported expansion of credit access occurs across all tested race, ethnicity, and sex segments resulting in the tested model increasing acceptance rates by 23 percent to 29 percent and decreasing average APRs by 15 percent to 17 percent.
Machine Learning Applications in Assessing Credit Risk
Recent advances in digital technology and big data have allowed digital lending to emerge as a potentially promising solution to reduce the cost of credit and increase financial inclusion. However, machine learning methods that lie at the heart of digital lending have remained largely a black box for the nontechnical audience.
The IMF’s Working Paper from May 2019 discusses the potential strengths and weaknesses of machine learning-based credit assessment through (1) presenting core ideas and the most common techniques in machine learning for the nontechnical audience; and (2) discussing the fundamental challenges in credit risk analysis.
The paper argues that FinTech credit has the potential to enhance financial inclusion and outperform traditional credit scoring by (1) leveraging non-traditional data sources to improve the assessment of the borrower’s track record; (2) appraising collateral value; (3) forecasting income prospects; and (4) predicting changes in general conditions.
The Pros And Cons of Using Machine Learning to Prevent Fraud
Long ago, racketeers had to physically steal credit cards to make fraudulent purchases. As shopping evolved, so did the scammers. With the expertise to harvest personal information from the Internet, they can now plunder and sell thousands of identities easily with just a few clicks.
This is where machine learning can swoop in to save the day. Machine learning, a subdivision of AI, enables machines to independently process and learn from past information to predict future patterns. The ability to process information almost instantly also allows the machine learning system to render an analysis in the middle of a transaction: a critical function for preventing fraud on the spot.
To build a machine learning system in-house, companies will have to hire not just one, but a team of data scientists to design and continually update the system. However, to build a viable machine learning model with long-term functionality, requires a great deal of good data – which isn’t easy to obtain. For datasets to be statistically relevant, there must be a large enough number of transactions, each of which should contain enough data to work with. This includes information about the outcomes of those transactions.
P2P Lending Hits a Speed Bump
According to the Link Asset Services Marketplace Lending Index , if you had deployed money on the largest P2P lending platforms you’d have earned a composite net annual return of about 4.1% in the first quarter of 2019. This net figure, calculated after losses and fees, has been falling steadily for three years, from a recent high of 6.4% in the second quarter of 2016. These low returns and increased defaults have started to spook some investors. BondMason recently announced that his company would wind down its marketplace-lending activities owing to declining returns.
Yet a net return of just more than 4% from the asset class needs to be put in perspective. Assuming the average duration of the index’s basket of P2P loans is about three and a half years, it’s also possible to compare the P2P asset class’ returns with more conventional fixed-income investments. P2P returns are well above the risk-free rate on three-year UK government bonds, which stands at less than 0.5%. P2P loans also provide a higher return than baskets of investment-grade sterling corporate bonds.
The risk is that this premium might be entirely eaten away by defaults shooting up in a recession. However, that recession will also result in increased interest rates being charged by lending platforms, which in turn might result in a strong rebound for investor returns.The upshot? Despite its current travails, P2P’s glory days may be ahead of it.
Popularity of P2P Lending High in Asia
Established banks and technology specialists from start-ups to giants such as Amazon have spoken about their desire to use advanced data analysis to create new, more personalised types of financial products at lower costs. But a recent report from the Bank for International Settlements — the central bank for central banks — warned that firms could use the same information to exploit customers. For example, studying customer data could allow lenders to overcharge by working out the maximum rate a borrower would be willing to pay for a loan, rather than competing to give them the cheapest available deal.
Data-driven banking is in its early stages, but regulators are not alone in keeping a close eye on developments. The idea that the regulator could retroactively “move the goalposts” to allow customers to complain about old products was the biggest source of criticism from banks during the PPI scandal.
The FCA, however, is unmoved by such criticism. AI-influenced products and services are a new area, meaning that many boundaries on acceptable practice have yet to be agreed, and questions over how banks can use customer data are part of a broader debate on the ethics of new technology.
As Digital Lending Moves from Frill to Necessity – Where Are We at?
Five years ago, many banks viewed digital lending merely as an alternative sales channel or simply an additional option