Fintech – or financial technology – is a much-hyped buzzword on Wall Street that brings to mind nimble startups slaying stodgy bank ‘dinosaurs.’ The reality is more complicated. Fintech is growing up: Financial institutions are increasingly viewing these disruptors as partners while startups are learning that they need the scale and regulatory expertise of the incumbents. Both sides have a lot to learn, and benefit, from each other, according to speakers at the recent “Fintech: The Impact on Consumers, Banking, and Regulatory Policy” conference at the Federal Reserve Bank of Philadelphia.
“We are actively seeking startups for our members to partner with,” said Robert Nichols, president of the nearly 6,000-member American Banking Association (ABA). He said the ABA has “kicked the tires” of some startups to try to match them with banks, and it has invited five disruptors to present at its annual convention, to be held this month in Chicago. The ABA also created a ‘fintech playbook’ designed for smaller banks. Collaboration is a good thing. “The banks have trusted relationships, these [startups] have ideas to enhance the customer experience,” said Nichols in a follow-up interview with Knowledge@Wharton.
Large and small banks are innovating on their own, too, Nichols said. Capital One has integrated its services with Amazon’s Alexa digital assistant and its video-enabled device, Echo Show. Consumers can ask Alexa for their account balance, request that it track their spending or even make a payment. Bank of America is set to debut its chatbot Erica on the bank’s mobile app to help customers with personal finance decisions. Also, more than 30 banks are using Zelle, a service that lets people send money to each other in minutes. It started in 2011 as a collaboration among Bank of America, Wells Fargo and JPMorgan Chase.
Nichols also cited the experience of Live Oak Bank in Wilmington, N.C. It launched a startup called ‘nCino’ that developed a cloud-based technology meant to make the commercial lending process more efficient. NCino is now being used by 130 banks, including eight of the top 20. This month, Live Oak announced the creation of Apiture, a digital banking joint venture with First Data Corp., an ATM payments technology company. Meanwhile, Boston-based Eastern Bank spun off Numerated, a startup that lets customers apply for a small business loan in two minutes and get funding within two days. Eastern hired fintech entrepreneurs to work with its bankers to create an innovation lab that led to the startup.
Alternative Data, Fintech Charter
The central bank is getting in on the act as well. The Federal Reserve Bank of San Francisco launched a fintech portal in May to help companies navigate the regulatory system and show them where to go for further assistance, said Tracy Basinger, its director of financial institution supervision and credit. More than 30 firms have already reached out for help. “We have no doubt that fintech innovation can introduce real benefits for consumers” and small businesses, especially those that are the underserved.
“The banks have trusted relationships, these [startups] have ideas to enhance the customer experience.”–Robert Nichols
The Office of the Comptroller of the Currency (OCC), which regulates more than 1,600 banks including the country’s largest institutions, has proposed a plan to charter fintech companies as special purpose national banks. However, several state bank regulators are suing the OCC to stop the plan, saying such a move could threaten consumers, according to CNBC. Instead, state regulators back Vision 2020, which lays out the framework for a 50-state fintech licensing and supervisory system.
Basinger said regulators do not want to get in the way of fintech innovation. She cited the Consumer Financial Protection Bureau’s (CFPB) “no-action” letter in September as one example of both sides learning from each other. In the letter, the CFPB declined to initiate supervision or enforce any action on Upstart Network, a company that uses alternative data in making credit and pricing decisions.
Upstart wanted to use non-traditional data such as a customer’s education and work history as part of its lending criteria. The CFPB is allowing it, but Upstart has to provide lending and compliance information to educate the regulator on the impact of using alternative data in lending. The letter is “a form of acknowledgement that with this additional scrutiny and analysis, the data can possibly be used to responsibly expand access to credit or to provide less expensive credit,” Basinger said. They want to avoid the risk of discriminating against borrowers with scant education, further disadvantaging the underserved.
While the use of non-traditional, or alternative, data sources is gaining ground, information culled from social media remains a challenge especially for use in making lending decisions. “We honestly have not spoken to many lenders at all that are able to show us how they’re using that data,” Basinger said. Still, “there are a lot of firms giving some serious thought to this and potentially testing it internally.” One problem with using social media data — such as what people post on Facebook — is that these habits change over time. Lenders need to be able to make credit decisions based on factors with more longevity, she said.
For that reason, the FICO credit rating score remains “extremely powerful,” said Spencer Robinson, head of strategy at Kabbage, a startup that uses algorithms based on large data sets to make lending decisions to small businesses. However, he said, FICO is an intentionally broad score that applies to wide swaths of population. Kabbage supplements traditional credit information with other data to get a more granular view of the borrower. As a result, young folks and immigrants without a long credit history might be approved for loans more readily and at lower rates.
This alternative data is not always something esoteric, like what time someone logs into Facebook. “There is a misconception of what this alternative data is,” Robinson said. “When you strip all of the sexiness off it, quite frankly, we’re doing the exact same thing that’s being done in lending for years. … We’re trying to answer the exact same questions that any lender is always trying to answer,” which is the customer’s ability repay the loan. Kabbage uses additional data that varies from person to person to help it look deeper into the borrower beyond the FICO score.
While the use of non-traditional, or alternative, data sources is gaining ground, information culled from social media remains a challenge especially for use in making lending decisions.
One example of alternative data used by online lender the LendingClub is the internet footprint of a customer. It doesn’t use social media information due to privacy concerns. Rather, the company uses things like a geocode IP address for fraud detection. If a borrower applies for a loan online and puts his address down as Boston but his computer’s IP address signals it’s in Russia, a red flag goes up. “That’s not in credit bureaus. That only comes from learning, capturing and analyzing the data,” said Siddhartha Jajodia, LendingClub’s chief investment officer.
Like the use of alternative data, fintech’s foray into artificial intelligence (AI) and machine learning is an evolution, not revolution. “Machine learning and AI and all of these buzzwords that are out there are quite frankly just that — they’re buzzwords,” Robinson said. “It’s just a new algorithm. Instead of a linear regression, they’re using a random forest. It’s not an inherently different thing.” Most people also think machine learning means computers learn on their own without being supervised. But he said fintech mainly uses supervised machine learning. In supervised learning, the computer is given a set of sample data to learn the concepts behind it and make inferences, which can be applied to new examples.
Bitcoins and Financial Stability
One of the more disruptive developments in fintech is the creation of cryptocurrencies like bitcoin. Unlike fiat, or hard, currency, a cryptocurrency is unregulated digital money controlled by its developers and used by the members of its virtual community, according to Asani Sarkar, an assistant vice president at the Federal Reserve Bank of New York. It is a decentralized product. No one provides a set of protocols, but there are a mutually agreed set of codes. Transactions are recorded on a virtual public ledger called the blockchain.
Like the use of alternative data, fintech’s foray into artificial intelligence (AI) and machine learning is an evolution, not revolution.
Proponents say using the bitcoin can reduce fees, time and risk in transferring value. For example, moving assets between two digital wallets could take 10 minutes instead of one to two days, Sarkar said. A range of businesses and nonprofits accept the bitcoin and many startups are proposing new applications and business models for it. However, Sarkar points to some sobering drawbacks of bitcoin as a currency. It is not widely accepted in and of itself — people or companies that use the bitcoin still have to convert it to hard currency for it to be useful. That adds fees and delays, which increase the risk for bitcoin users.
Large banks are getting out of handling bitcoins because of the risks involved, so bitcoin users have to keep looking for other banks willing to swap it for hard currency. Such additional complexity makes the value of bitcoin hard to pin down, Sarkar said. “The risk in this transaction between bitcoin and fiat currency has increased markedly over the past year or two. The risk is [that] the exchange rate of bitcoin and fiat currency is very volatile.” Moreover, the value of a bitcoin can vary by as much as 20% among exchanges — if the exchange even survives. As of 2013, 13 of 40 bitcoin exchanges have failed, he noted.
Another factor driving price volatility of the bitcoin is speculation. Buyers are investing in it to make a profit if the price goes up, as opposed to using it as money. The value of the bitcoin has risen by 400% this year, sparking concerns of a bubble, said William Nelson, executive managing director of The Clearing House. He noted that similar comments recently were made by JPMorgan CEO Jamie Dimon and Nobel Laureate economist Robert Shiller. However, Nelson believes that cryptocurrencies pose a limited threat to the U.S. financial system because it is not widely used. He thinks it will have more of an impact in economies where people do not have confidence in their own currency or they are avoiding controls on their money.
Some countries are actively exploring issuing their own digital currency, including Singapore and England, according to Gurvinder Ahluwalia, founder of Digital Twin Labs and former CTO for technical solutioning and blockchain at IBM. While the problems with cryptocurrencies are well-known, governments are studying it and seeking solutions. “Everybody is trying to figure out the level of readiness” in terms of the right business models, operations and level of risk, he said. That’s just part of the way technology works. “Innovation is never a clean process.”
Article by Knowledge@Wharton