WHY DO MARKETPLACE LENDING PLATFORMS NEED UNIFIED FINANCIAL ANALYTICS?
Marketplace lending platforms are still in the process of earning their keep against the formal credit institutions and other alternative sources of credit. The emerging players in the sector are all lagging in terms of the quality of their risk analytics, so it is easy to understand why their focus is other than on unifying analytics across the sector.
Online lenders are strong at using algorithms and Big Data to score borrowers and make credit decisions within minutes. Consistently, in interviews we conducted in the course of writing this book, entrepreneurs in marketplace lending pointed out the strength of their credit scoring algorithms over FICO scores and other “old-fashioned” metrics that banks use to assess borrowers. Whenever we asked about their risk management practices, platforms mentioned credit scoring as their only risk management practice. We explained that how, by doing this, marketplace lenders are measuring counterparty risk, and counterparty risk only. There are at least three grave issues with this single-minded approach to risk.First, predicting future behavior by looking at past actions is like driving by looking in the rearview mirror. If there are any bumps in the road—such as unforeseen market turmoil— predictions based on historical data go out the window. Second, using Big Data algorithms to score borrowers is hardly an exercise in risk management; it is risk measurement at best, and then again, it only focuses on counterparty risk. Third, marketplace lenders seem to forget about the other risks that might lead to losses for their investors. Platforms are hardly measuring credit risk at the same level of detail as banks. They apply no portfolio analysis and stress testing, and they fail to measure credit exposure on the level of portfolios, counterparty and contract levels. Because the loans are uncollateralized, lenders have net exposure only.
There is no analysis of market risk in credit exposure and losses, no expected loss analysis on the level of the counterparty and the portfolio, and no identification of seniorities. Meaningful and professional analytics for investors are missing in marketplace lending. Lenders are flying blind in terms of market risk, credit risk, and concentration risk. The latter is notoriously difficult to measure and manage, and it goes undetected even though platforms have the data that would allow them to identify concentration risk.As we pointed out, in Chapters 11 and 12 on banking risk management, there is much more work necessary than simply scoring borrowers for their ability to repay a loan. A borrower’s online activity and history with paying bills may be spotless, but his ability and willingness to pay has a lot to do with the strength of the market. What will happen to the loan book of a platform when market conditions change? What will happen when inflation speeds up, and when market interest rates rise? Predictive models for borrower behavior fail to forecast market behavior. We have investigated some stress scenarios in Chapter 13 with the analysis of a portfolio of marketplace loans, and we have seen that returns suffer even under light stress. Unfortunately, marketplace lenders are unprepared to deal with market risk.
It seems that by the term “risk management,” platforms often understand managing their own risk only, such as legal risk from lenders who have lost capital by investing in marketplace loans. Sure, platforms can always point to the fine print in their terms of service and claim they did all they could to score borrowers properly. Some marketplace lenders go farther and claim that borrowers are indifferent about defaults because by the time defaults happen, borrowers will have made most of their money back thanks to the high interest rates. The hands- off approach of marketplace lenders works when the economy is doing well, there are few defaults, and interest rates outpace inflation.
But when the economy tanks, when borrowers are underwater and defaulting en masse, lenders may demand that platforms do a little more work. Marketplace platforms should at least be on the same level as the formal credit sector when it comes to analyzing the risk of their loans. In fact, because the data of online lenders is cleaner and more centralized, they could easily surpass banks in the quality of their analytics. To cut a long story short, they need to up their game when it comes to analytics for their investors.Retail investors have no other option than to rely on the relatively simple analytics that marketplace platforms provide on their websites. Several third-party services exist that allow lenders to analyze the loan books of some platforms in a little more detail. Third-party services range from professional offerings for institutional investors to labors of love of peer- to-peer lending enthusiasts. Unfortunately, they hardly take into account all the risk factors we described in Part Two of this book, and they give investors just half the picture about risk in their portfolios. To be fair, many credit institutions are in the same position. However, bank loans are less risky because they require collateral as a security they can liquidate if defaults occur.
If a marketplace lender or third-party analytics platform has an engine that takes into account the factors we have pointed out in our excursion into credit risk management in Chapters 9-12, they could deliver an accurate assessment of the potential risk for marketplace loans. Platforms might use this information to optimize their loans, and they might want to make it available as a competitive advantage to their investors.
It is difficult to understand why marketplace lending platforms feel they have no need for better risk analytics. Perhaps they see themselves as a service provider only, and less as a credit institution. Still, improving their approach to risk management on behalf of their investors is in the best interest of marketplace lending platforms themselves.
Most platforms believe they have barely scratched the surface of what is possible in terms of market share. They have aggressive plans to take a bigger bite of the multi-trillion-dollar credit market in the U.S. with plans to branch out into car loans, student loans, and mortgages.1 Nevertheless, originating collateralized loans is quite a different ballgame from unsecured consumer credit. Having physical or financial collateral in the mix requires more complex models to assess risk. Stricter regulation will inevitably follow when online lenders become bigger and more systemic, and marketplace loans will have to become more secure. Regulators and the market will need to understand better the impact of alternative lending on the financial sector in general.So far, investors in marketplace loans are left to their own devices as far as risk analysis is concerned. Some platforms make their loan book available, but investors will still need to program their own risk model to extract some knowledge from the available data. Stressing a loan portfolio is a job that quickly exceeds the average retail investor’s Excel skills, no matter how VBA-savvy he may be. Sure, if someone is knowledgeable and patient enough to program a financial model in MATLAB, C, or Java, they can simulate the expected performance of a marketplace loan portfolio with the publicly available data, in similar ways that banks do. We proved this in Chapter 13. It helped in the analysis that one of us (Akkizidis) holds a PhD in applied mathematics and programs financial risk models for a living. Still, to get the entire picture, we would need additional data not only about borrowers, but also about individual lenders, as well as data from other marketplace lenders. Platforms are sitting on these data, and it would make a lot of sense for them to pool data to understand better the risk in their asset class. This starts with standardizing the formats in which they collect and store data. If they aim to reinvent the future of credit, then why should they have an interest in keeping their investors in the dark?
Another argument for unified analytics is the aspect of pooling individual services that are expensive for platforms to develop and maintain on their own.
Running a marketplace lending platform is expensive, which forces the platform operator to underwrite as many loans as possible. However, what if a community, such as a town with 5,000 residents, would like to run its own lending platform? Residents already know each other and might already extend credit to one another. Bringing this informal exchange of credit into the digital age with a marketplace lending platform makes a lot of sense. At the same time, running a small platform would be too expensive at the current time. However, if the individual marketplace lending services were available for rent, such as in software as a service (SaaS), scale would matter less. Once it has become affordable, a community, such as a company, club, church, or other special interest group may set up their own lending platform on which members could extend credit to each other. For example, a town might set up its own lending platform that a community bank may operate. Existing networks could also have proprietary lending platforms: a phone company, for instance, already has lots of data about its users, and they are already processing payments. It would be a small step for a phone company to allow its subscribers to extend credit to each other, especially when it can build a marketplace lending platform in a modular way. An add-on analytics module might be useful for such community platforms, should they emerge.We will discuss in this chapter how exactly we could develop and implement a system for unified analysis. Because marketplace loans are still relatively simple uncollateralized annuities, we recommend lending platforms already think ahead to a time when their loans might have more moving parts, such as physical and financial collateral or several currencies when lending abroad. The benefits of such a system go beyond advantages to lenders— borrowers, platforms, and third parties who might partake in marketplace loans will gain as well. On top of that, unified analytics in marketplace lending would be a boon for banks also.
Did we manage to spark your curiosity? Let’s focus in more detail on these advantages.17.1.1 Advantagesforlenders
Marketplace lenders are the most obvious group to profit from more transparency of the risk of the loans they invest in. When they gain a more accurate picture of their exposure—regardless of whether they like to pile on more risk or try to avoid it—they can make a more educated decision in their asset allocation and improve the performance of their investment. Professional money managers already do this, and retail investors should have access to the same tools that let them perform credit risk analysis by themselves. Being able to provide lenders with stronger analytics would certainly be nice to have for platforms, the only obstacle being the high cost of developing industrial-strength tools in-house. How could marketplace lending platforms offer strong analytics without breaking the bank?
A framework for better analytics should enable investors with the following insights:
■ Lenders should gain a clear idea about all financial risks including risk factors they have exposure to—such as interest rates, currencies, inflation, and asset prices. Expected and unexpected losses that may arise from exposure to market risk should also be clear.
■ Unified analytics would allow comparison of marketplace loans across different platforms in the entire sector when it comes to risk exposure and expected returns. Investors then have a clear view of the entire market.
■ Robust analytics should measure all risks based on real-world stressed market conditions, stressed default probabilities, in addition to behavior expectations. Identifying these risks next to counterparty credit risk would be another powerful enabler for lenders.
■ Calculating unexpected losses from all potential risks is also a must. This would trump the current median expected rate of return that most platforms content themselves with.
■ When we know the links and potential correlations between lenders and other peers (i.e., systemic and concentration risk) and between peers and guarantors that might provide credit enhancements (i.e., wrong way risk), we will have the full picture of the risk inherent in the entire marketplace lending sector. In the long run, this can set the foundations for establishing an integrated platform that is linked to other online lending platforms, similar to clearing houses.
Even though marketplace lending has undergone a name change from its earlier moniker peer-to-peer lending, individual persons lending to others is the endearing idea that keeps the sector in the limelight. We have already discussed the fact that retail investors gaining wider access to credit is a vital empowering feature of marketplace lending. Clarity about the actual risk and return of the loans they invest in, the corresponding income, and the future cash flows discounted to the present time will raise the confidence of the investing public. If investors have certainty about a positive income and potential expected losses during the lifetime of the contract, their financial literacy improves. What could be a better starting point for marketplace lenders to grow market share and launch additional products?
With unified analytics, investors gain more than a better understanding of their own exposure to loans from an individual platform. When they can compare their investments in credit across different originators and platforms, both in the alternative space and traditional financial institutions, they can structure their investment portfolios to make them more robust, with better performance and fewer losses. Lenders will also benefit from unbiased and meaningful reports of the performance of their credit portfolios.
17.1.2 Advantagesforborrowers
The current consensus among marketplace lending platforms is that they will have to increase interest rates if they are to invest more in their analytics capability. This may ultimately lead to higher rates for borrowers and lower profits for investors. However, this reasoning is similar to complaining about the cost of a tune up for a car that will ensure the engine's integrity in the long term. With unified analytics across the sector, marketplace lending has a better chance of achieving escape velocity, to penetrate its current market deeper and branch out into other credit segments. If this is the case, borrowers are likely to benefit. They will gain the confidence that deals are fair for both sides, especially when collateral is involved. When individual borrowers and SMEs know that the terms of a loan are fair—considering riskfree probabilities and real world future assumptions—they are more likely to pay, which is ultimately in their own best interest. Of course, borrowers also have a stronger tool to shop for the best loans across several platforms.
Better analytics can also help foster more responsible borrowing behavior when borrowers have better information about their own exposure—or over-exposure—to credit risk. A lower default rate in the sector will ensure that even longer-term loans have the lowest possible interest in the future.
A unified analytics framework that is useful for borrowers should offer the following insights:
■ It should clearly identify market risks that come with loans, such as fixed and variable interest rates, currency risk, prices of assets that borrowers could pledge as collateral.
■ Borrowers should be aware of the consequences of defaulting on their obligations. These include possible fees and other penalties; the exercise of collateral and other credit enhancements; losses to other peers and the portfolio of the platform that result from the default of an individual borrower, via concentration and systemic analysis; and the impact of a default on the credit rating and other credibility scores.
■ Borrowers should also have full transparency about possible options, such as prepayment. They should know whether prepayments are a good strategy, for instance, to roll over loans. Of course, risk-free probabilities and real-world assumptions about the future should inform fair value calculations of loans.
17.1.3 Advantages for marketplace lending platforms
By their own account, most marketplace lending platforms believe their risk management is superb, despite being extremely limited in reality. Platforms offload financial risk management of their loans to lenders; yet, to expect retail investors to run their own stress tests borders on cynicism. It is unclear what will happen when borrowers are under stress, and platforms should be aware of the negative consequences of widespread losses if they occur. Marketplace lenders gain much from offering more robust analytics to lenders and borrowers. When they are transparent about their product, all involved parties benefit.
In the run-up to the financial crisis of 2007/8, banks sold fuzzy products to investors who only cared about yield on paper. At the end of the day, it was the investors who shouldered the losses, and banks suffered reputational damage that still reverberates. Conversely, in postcrisis conditions, increased transparency will stimulate competition. If marketplace lending platforms have a serious interest in reinventing the way credit works, they should set an industry standard that centers on transparency. In the end, this stands to increase the reach of individual lending platforms and the size of the market in general.
Until regulators require platforms to calculate analytics with the same granularity as banks is just a matter of time. If platforms invest in this capability now, they are taking proactive steps to shape the dialogue about regulations and to be ready when compliance requirements ramp up. They should be in a position where they can prove they are providing stable and safe financial products. This is important for regulators and other financial partners, such as banks. As an added bonus, platforms will also understand better the financial health of their platform in regards to risks and profitability.
Eventually, when lenders can compare marketplace loans with each other, a unified analytics engine can become best practice for all financial events, for all lending platforms and credit institutions, online and offline. At that point, marketplace lending has truly arrived on a level playing field with the established credit sector. Online lending platforms can then become ringleaders in the Hybrid Financial Sector. In the long term, platforms will standardize their analytics. Banks, of course, are also mostly running proprietary analytics, but the trend points toward a more unified system. In such a system, investors, markets and regulators have a common framework to assess risk. If platforms spearhead the approach and surpass banks on transparency, comparability, and regulatory compliance, they will have a leg up in the future.
Another reason points towards the need for unified analytics for marketplace lenders: when they re-bundle their offerings with a plug-and-play architecture, they can actually create a new kind of financial services company that banks will have a very hard time competing with. Instead of continuing on the path towards greater and greater fragmentation, where platforms serve one small niche demographic, they could go the other way. This would tip the scales in favor of the new entrants, and they might find themselves in a position of strength when bargaining with banks and other lending institutions for partnerships.
What are the first steps marketplace lenders can take towards unified analytics? Consistency of data with unified definitions and terminology of events, data formats, and time frames are a good starting point. Data lie at the core of understanding risks and rewards. Some marketplace lending platforms already have APIs in place for investors to execute transactions and keep track of their portfolios. However, they should also make data about their loan books available in real time.
17.1.4 Advantages for guarantors and protection sellers
For the time being, guarantors and protection sellers have little to do with marketplace lending. However, when online lenders offer collateralized loans, guarantors and protection sellers will enter the fold. Why do guarantors and protection sellers need unified analytics? Obviously, they would want to know the exact contractual payment obligations in case of a risk event and the associated pricing premiums. Guarantors need to know the real income and the potential claims so they can calculate the nominal value of the insured exposure. Market risk, counterparty risk, and concentration risk should build on real-world probabilities and stress conditions, instead of borrower analysis alone. With this information, guarantors and protection sellers can structure derivatives that insure them against financial risk. With derivatives, guarantors can hedge their exposure. They could also offer standardized insurance products, similar to credit default swaps, which platforms could then offer as a subscription to lenders.
17.1.5 Advantagesforbanks
Several FinTech services rebundling on a single platform would be a double whammy for banks: first, customers would learn that they can easily replace their banking services with add-ons that work in a plug-and-play architecture. Second, because those individual FinTech services operate on a single platform, regulators would have the opportunity to examine all of them via a single API. Regulators would have the full scope of operations on the platform. A FinTech platform standard for unified data would set an uneasy precedent for banks who benefit from complicating their operations with tangled data that take weeks and months to decrypt. If regulators have gotten used to clean digital records with immediate access from FinTech companies, there is just a small step to asking banks to provide the same convenience. Regulators will then be able to perform their own analysis and stress tests to identify financial risks and check for compliance in real time.
What is initially uncomfortable for banks would be a boon for consumers. Unified analytics would narrow the gap between the established financial sector and new entrants significantly. In the long run, it would also save banks and FinTech companies enormous time and effort needed for providing the results of analysis to comply with regulatory demands. Compliance has been one of the most expensive, unpleasant and risky operations in established financial institutions in recent years.2
The current leadership position of banks puts them in an ideal place to spearhead a unified analytics platform themselves. Banks have been slow to make their know-how available for a fee for the emerging FinTech sector, and marketplace lending in particular. If they jumped at the chance to actively lead the development of unified credit analytics and data standards, they might assure themselves pole position in the sector in the future. A possible point of entry into providing unified analytics for banks is to seclude a financial laboratory far from the mother ship, comprised of data scientists, financial mathematicians and statisticians (“geeks”). They could get to work on data collection standards, reporting standards, open APIs and robust interactive analytics that online lenders could use. A bank could launch such a venture with a relatively modest investment to the tune of what it costs to sponsor a yacht competing in the America's Cup. Banks have every reason to do this. When a new financial sector has formed without banks as members, it will be too late forever for them to regain their footing. If banks miss the boat, they will join the club of other companies that learned their services were no longer in demand after the digital revolution had disrupted their business model. Certainly, hindsight is 20/20, but banks will have no excuse to say that they didn't see the writing on the wall.
17.2