Creative Landscape

FinTech companies embrace creative applications of new technology Jason Ash

Every year on my birthday, my grandmother mails me a card with a check inside. When I was young, she encouraged me to deposit this birthday money in a savings account, saying the bank would pay me even more money to keep it there and it would grow over time. (Her advice likely played a role in my eventual career choice as an actuary.) This year, I had the good fortune to spend my birthday with my grandmother in person. As expected, though with slightly less flourish after 30 years, she wrote the annual check and passed it to me across the dining room table. I assured her I would deposit it, and I took the opportunity to prove it.

I asked her if she wanted to watch me deposit the check. I unlocked my smartphone with my fingerprint, logged into my bank account, took a picture of the check, and showed her my new and improved balance. This led to a long conversation about other technology in finance, like mobile payments with Venmo, free stock trading with Robinhood and budgeting tools like Mint that link directly to a person’s bank account. (I decided to save Bitcoin and the Blockchain for another day.) These innovations make remote deposit, which was first implemented in 2003,1 seem like a relic of the past. But while the novelty may be gone, consumers today consider simplicity and convenience to be requirements, rather than perks, of financial services.

What is FinTech?

FinTech, a neologism that blends the words “finance” and “technology,” is the name for these and other endeavors that use technology to improve financial services. But not all FinTech companies are tech-driven startups looking to disrupt the status quo. Some companies have found a lucrative niche by collaborating with existing institutions, and many of those same existing institutions have established internal venture groups to fund new innovations directly. Others, like Blackrock, have entered the space through acquisitions,2 and it is not hard to guess which large companies are backing new services like Google Wallet and Apple Pay. In other words, it would be a mistake to think that FinTech is separate from or outside of the broader financial services industry. Instead, it encapsulates most of what is innovative in finance today—not because FinTech companies in Silicon Valley are the only innovators, but because FinTech companies have fully embraced creative applications of new technology.

However, innovation often causes disruption, and FinTech’s rapid growth has been far from smooth. Among the challenges it still faces are integrating into the financial services ecosystem, sculpting a fair and responsible regulatory landscape, and overcoming setbacks that threaten to distract from progress. For this industry to be guided responsibly, all financial professionals, including actuaries, should play a role in its continued evolution.

Actuaries possess skills that are particularly well-suited to many challenges facing the FinTech industry today. Yet actuaries are woefully underrepresented within FinTech companies. My hope is that our profession will find a way to partner with FinTech. Thousands of new job listings request actuaries in every way except by name. For example, a FinTech lender seeks a director of Data Analytics who is “highly analytic, paired with strong communication skills and a proven track record of translating data insights into actionable outcomes that result in business growth.” Another company seeks a director of Risk Modeling & Forecasting who is to be “responsible for risk analysis, developing underwriting/pricing/response/forecasting models and developing analytic tools.” Among an actuary’s most transferable skills is an aptitude for understanding the world through data and translating this understanding into strategic action. And, if there is one thing all FinTech companies share, it is an unceasing hunger for data and the value that can be derived from it.

Nearly all FinTech companies benefit from a virtuous cycle of data that helps them improve and expand their services over time. Charles Moldow, a partner at the venture firm Foundation Capital and an early FinTech investor, wrote about the benefits of this virtuous cycle for online marketplace lenders. He says: “Data from loan performance feeds back into the marketplace lender’s model, creating an even more accurate model. As the accuracy of the data and model increases, the marketplace lender can offer borrowers lower rates. As rates decrease, more borrowers flock to the platform, driving more data into the model.”3

Financially Creative Lending Practices

Consider SoFi, an online lender founded in 2011 whose name stands for Social Finance. Its founders hypothesized that graduates of elite universities would be highly likely to repay their student loan debt, and they began a pilot program to refinance $2 million of loans from Stanford students at lower interest rates. Eventually, as SoFi offered more loans and gathered more data, it improved credit underwriting, refinanced a higher number of loans and lowered its rates even more. In 2016, after having gathered data from more than $19 billion of loans, SoFi publicly stated that it would no longer use FICO credit scores in its underwriting models, believing instead that data—such as employment history and monthly free cash flow4—cultivated from its member base of more than 300,000 individuals provided better predictive power. By some estimates, FICO is used by 90 percent of all lenders in the United States,5 so this move reveals how much confidence SoFi has in its proprietary machine learning models.

Companies like SoFi estimate a borrower’s probability of default using information from a short online loan application that is later linked to data from other sources. These lenders have found creative ways of extracting every piece of information they can while making the application process as painless as possible. In fact, lenders may prefer to keep the application simple to reduce adverse selection,6 reasoning that any unnecessary complication may drive away higher-quality applicants who aren’t desperately seeking funds. To do this, lenders use intelligent filters to decide whether to verify certain application information, such as whether the applicant’s stated income is within a threshold of the average income for people with the same job in the same geographic area. Some lenders use services like Plaid, another FinTech company, to connect directly to a person’s bank account and calculate monthly free cash flow. Lastly, they purchase thousands of data points for each applicant from credit bureaus like TransUnion. While SoFi may not use FICO in its models, like most lenders, it still relies on a “track record of meeting financial obligations” established through other credit information like outstanding debt, payment history and number of open credit lines.7

In just a few seconds, a credit model uses this data to approve or decline the applicant. Behind the scenes, the credit model is calibrated with millions of historical records made up of thousands of attributes each, and the latest machine learning algorithms are pitted against one another to identify those with the most predictive power. Logistic regression, a staple in probabilistic forecasting, is commonly used. But other techniques, notably ensemble methods like random forests or gradient boosting trees, are becoming increasingly important.

These models construct decision trees from loan attributes in a historical training data set to separate good loans from bad ones. One decision tree might have thousands of decision points, such as “applicants with no late payments on credit cards.” A record that meets this criterion is sent in one direction to more decision points deeper in the tree, while a record that fails is sent in another direction. Each record in the training data set traverses many layers through the tree before the model predicts its probability of default. This prediction then is compared against the actual outcome to measure the model’s accuracy. As a final step, hundreds of decision trees are generated and combined—hence the name ensemble models—to reduce overfitting and improve the model’s predictive power for data it hasn’t yet seen.

Marketplace lenders understandably are private about the exact models and techniques they use to underwrite potential borrowers. After all, underwriting, like it is for insurers, is a critical component of a lender’s competitive advantage. But if SoFi’s wager on its own models pays off—and, to be clear, this is an ongoing risk that the company faces—it would allow SoFi to compete even more effectively against banks as well as other industry incumbents, using its data in creative ways beyond student loan refinancing. In fact, building on its early success, SoFi has expanded to offer wealth management services, mortgage refinancing and term life insurance, which together suggest the company’s desire to engage continuously with its members from college graduation through retirement. But if a successful FinTech company is one that starts by focusing on a single service, gathers data and uses this data to expand its services to meet more diverse needs, then what ultimately separates it from the financial institutions it sought to disrupt? Increasingly, the lines are being blurred as FinTech companies play a larger role in providing financial services to consumers.

Keeping Up With Regulations

One consequence of this rapid technological development is that the regulatory environment has struggled to keep pace. In an effort to create a more consistent regulatory standard, early this year, the U.S. Office for the Comptroller of the Currency (OCC), the bureau responsible for chartering and supervising banks, announced it would allow FinTech companies to apply for special purpose bank charters.8 (Incidentally, Abraham Lincoln signed the law establishing the OCC and the national banking system, which was considered innovative at the time. If only it were possible to hear Abraham Lincoln’s thoughts on FinTech!)

Companies granted a special purpose FinTech charter would be regulated like banks and supervised by the OCC. Today, many of these companies inhabit gray areas between existing regulatory frameworks. For example, Lending Club, which is not a bank, partners with banks like Utah-based WebBank to originate consumer loans. While Lending Club is not obligated to follow banking regulations, it nevertheless voluntarily chooses to comply with laws that apply to banks, like the Truth in Lending Act. In 2008, Lending Club also worked closely with the Securities and Exchange Commission (SEC) to create a new registered financial product called the Note, which is a small slice of a larger loan (as small as $25) that can be bought and sold much like a stock.9 Nevertheless, the OCC hopes that a special FinTech charter could move away from these bespoke regulatory arrangements, promote consistency and “make the federal banking system stronger.”10

This regulation also comes at a time of recent turbulence in this nascent industry. Despite an impressive track record of venture financing—$36 billion globally in 201611—the FinTech industry has had its share of setbacks, which serve as reminders that its long-term viability is not a foregone conclusion. In the online credit space alone, Lending Club was rocked by a scandal in 2016 in which its CEO resigned and its stock price fell more than 50 percent, calling into question the effectiveness of its internal quality controls.12 Also in 2016, Avant, a Chicago-based online lender, laid off 60 employees and requested voluntary severance from others.13 Prosper Marketplace recently announced it has significantly overstated investor returns on its website for several quarters,14 a failure that has the potential to draw scrutiny from the Consumer Financial Protection Bureau. Yet these events reveal opportunities to improve and shape a promising industry, not flaws in the FinTech value proposition.

Opportunities for Actuaries

The FinTech industry clearly has momentum, but it is still in its early stages. After achieving initial success brought about by rapid technological innovation, it now must establish itself as a dependable service that provides long-term value for consumers. The pace of innovation is unlikely to slow, and FinTech companies need professionals to guide the next phase of growth. These professionals should be technically proficient and trained extensively to understand data and its importance in guiding a business’s strategic plans. They should be able to navigate and contribute to an evolving regulatory landscape. And they should be comfortable managing the ongoing trade-offs of risk and reward. Actuaries are well-known for competency in these areas, and they also are backed by professional societies with impeccable reputations and experienced members that can help the FinTech industry mature and flourish.

Actuaries are among the most technically skilled of all professionals, and this analytical expertise is incredibly valuable for many FinTech companies. These companies employ teams of people—like those you might see within an insurance company—who build financial models, evaluate risk, monitor performance and modify products in response to market demand and emerging experience. The core analytical problem-solving process that comes so naturally to most actuaries can be a valuable resource for FinTech companies looking to iterate through new ideas.

Beyond their technical skills, actuaries can help FinTech companies manage the wider implications of rapid technological innovation, among them the complementary changes in regulation. There is already significant overlap in existing banking and insurance regulation, especially for insurance products that draw on elements from both industries, like variable annuities. Actuaries who helped craft regulation in the rapidly developing annuity markets of the 2000s, and those who continue to formalize principle-based reserving methodologies, could play a meaningful role in shaping the FinTech regulatory landscape.

Finally, a key strength of many actuaries is their ability to evaluate the trade-offs between risk and reward. A relative of mine who has spent his career in Silicon Valley once told me that the entire culture of Silicon Valley is based on taking risk. The young but promising industry of FinTech is no different. There is a risk of further setbacks like those that occurred in 2016, but its fundamental value proposition—that technology can be used to provide financial services more cost effectively than existing options, with a high degree of quality and personalization, and to a vastly wider population—is sure to have a lasting impact. Actuaries who think they understand the risks of moving to a career in FinTech might be concerned that individual companies will fail. But actuaries who understand the rewards of a career in FinTech can be confident they have a unique and valuable skill set that can be applied to some of the most exciting challenges in finance. From my perspective, the real risk is not taking a chance on FinTech.

Jason Ash, FSA, CERA, MAAA, is a former consultant at Milliman and analytics manager at Lending Club. He is the founder of Ash Analytics LLC in Seattle.