November 21, 2019.
It’s expensive to be poor.
There are two banking systems in the world today: one for people with money (or good credit), and another for people without. But neither of these systems work well. People with money have gotten used to clunky user experiences and low expectations. For everyone else—that’s most Americans and billions worldwide—banks don’t serve them well, or sometimes, at all. Not only that, but the current system actually extracts from the very people who need financial services the most. The poorer you are, the fewer options you have and the more you will pay.
Written by Angela Strange & Originally published at this link.
The system is clearly broken. No one has been able to fix it, despite many attempts to do so. Big financial institutions already spend billions of dollars a year simply to maintain existing systems and comply with regulations, leaving little room for new product development. Meanwhile, compliance requirements and infrastructure complexity make it challenging for new companies to afford to even enter the market. But what if software could address these hard, structural problems, enabling even non-fintech companies to provide financial services for their customers? What if, as a result, people didn’t just put up with, but actually love, their banks? Right now, despite how often people use financial services, few would list their bank as a beloved brand among their daily products.
But the next bank does not have to start out as a bank. I believe the next era of financial services will come from seemingly unexpected places. Just as Amazon Web Services dramatically lowered the cost and complexity of launching a software business, unleashing thousands of new companies, the “AWS for fintech” stage has arrived in banking. Before AWS, it could cost $150,000 a month for a business to run compute and storage; it now costs roughly $1,500 a month. In much the same way, we are nearing the point at which any company can start or enable financial services. Consumer apps across multiple categories—home-screen fixtures highly valued by users—are becoming banks. It’s not as crazy as it might sound: Many drivers already consider Lyft and Uber their de facto bank. These ridesharing companies didn’t even exist a decade ago.
Nothing happens for years and then, when things change, they change suddenly.
How we got here
Why couldn’t all this happen through incremental improvements to the current system? And why haven’t more startups been able to scale? To fully understand why we’re starting to see such dramatic change, it’s worth unpacking how we got here in the first place.
The popular narrative for why “it’s expensive to be poor” is that large banks are to blame for high fees and a lack of product innovation. To some extent, that’s true. But classic financial services institutions also face structural problems due to legacy tech and the burden of a physical footprint.
Banks have been under pressure for some time, especially as consumers have moved online. Most of these institutions have existed for decades—centuries, in some cases—by acquiring customers in physical branches. The bank then owned that person’s full financial lifecycle: from first checking account to first credit card, first brokerage account to first mortgage, and so on. In the age of ecommerce, however, banks no longer have that same relationship with their customers. Instead, consumers can choose their financial services from several different providers online.
Then the financial crisis hit. Well-intended regulations such as the Durbin amendment (part of Dodd-Frank) limited transaction fees on debit swipes, the theory being that if merchants paid less in interchange they would pass on savings to consumers in the form of lower prices. These regulations were meant to help merchants and spur economic activity, but they significantly reduced revenue for large banks. Some estimated a revenue drop of over $6 billion a year—no small hit to absorb! To make up for those losses, many banks banished free checking, increased minimum balances, and raised overdraft fees. Ironically, the Durbin amendment had an adverse effect on the consumers it was intended to help. Since most Americans live paycheck to paycheck, these seemingly incremental banking fees significantly impact their lives.
For banks, raising fees was much easier and faster than lowering the hard, fixed costs of physical branches and associated staff. On top of that, existing software infrastructure can seem like a bottomless money pit: Banks keep incrementally adding and patching fixes for new compliance rules on top of older, hard-to-change systems, leading to a tangled, spaghetti-like mess of software. At some of the larger banks, 75 percent of the IT budget goes toward maintenance. Beyond software, there’s a large manual labor force. Ten to 15 percent of the workforce at larger banks is devoted solely to compliance. Citigroup alone had 30,000 of its 204,000 employees working in compliance last year, largely devoted to tasks like manually reviewing alerts triggered by anti-money laundering (AML) systems and filing suspicious activity reports.
Many of these maintenance and compliance costs are passed along to customers in the form of higher fees. Those costs leave little room in the budget for innovation. In other industries, startups could typically come in with fresh approaches and better technology. But in financial services, getting a fintech company going under the current conditions is hard, requiring multiple partnerships, entrenched financial-industry insider knowledge, and established connections and capital.
Here’s what it would take to launch a new “bank” that offers just two basic financial services products, a checking account and a debit card:
- The new bank obviously needs to comply with regulation. In the US, this is most often achieved by finding a sponsoring bank partner. (This tactic is much faster and has a higher likelihood of success than applying for a license.) A regulated bank agrees to “lend” the new bank its license in exchange for a financial cut of whatever the new bank is offering. Typically, that means the sponsoring bank gets more deposits without having to pay to acquire those customers.
- For a startup, finding the right sponsor bank partner is difficult: There’s no directory of these banks to identify potential partners and contacts.
- While the sponsoring bank does get the benefit of more business (e.g., more deposits), it also carries additional risk. It must ensure that the startup properly complies with KYC (Know Your Customer), AML (Anti Money Laundering), and so on. Given this risk, when approached by potential fintech startups, how does a bank thoughtfully and efficiently evaluate a startup partner?
- After a startup decides to partner, there’s still much more time and effort involved to build out the rest of the product. There needs to be a card processor (more negotiation, more costs); a card issuer or relationship with a card network (like a Visa or MasterCard); a card printer (more back-and-forth); some way to hook into other bank accounts for data; a way for consumers to make payments; vendors to help with compliance; and so on and on and on.
Under this system, it’s hard for existing banks to get better, it’s hard for new banks to get started, and it’s hard for them to partner with each other, even when incentives are aligned.
There has got to be a better way
Today, technology allows innovative new companies to be created and enables existing banks to better serve the needs of customers. Specifically, that’s happening through (1) new financial-services infrastructure companies providing APIs (application programming interfaces); (2) new distribution channels that enable better, differentiated products to spread more easily and at lower customer acquisition cost; and (3) better data that allows companies to assess and assign risk more precisely.
First, the infrastructure: We are at the beginning of a growing ecosystem of banking infrastructure companies—an “API economy”—that both startups and incumbents can draw on. These Lego-like companies specialize in building and providing specific building blocks for banking (for instance, KYC/AML compliance). By providing APIs to their services, the companies democratize their expertise. This means that any one company doesn’t have to know every single thing there is to know about complex regulations; another company that specializes in that area has created an API for others to use. It also means that it’s easier to create new financial services companies of all sizes and of all kinds. Rather than having to build or maintain regulatory systems themselves, they can just “plug in” to that expertise.
Not only are new entrants using this software infrastructure to get started faster and more cheaply, incumbents are beginning to augment, or even replace, some of their legacy systems. Instead of going the way of other brick-and-mortar retailers—many of which either went out of business or became glorified showrooms for ecommerce—the beauty of the API economy for banking is that it lets everyone participate, play to their strengths, and concentrate on their core offering. The demand for better and more inclusive financial services is big enough that there’s room for many players in the market to succeed as large, stand-alone companies. All of this results in better products at less cost, serving a wider range of consumers.
This also means that almost any company can offer banking services. Whereas existing consumer services used to have only two options to monetize—either charge for the product or sell advertising—now, companies can layer on financial products. What if your ridesharing app became your bank and you could pay for goods as frictionlessly as you hop in a car? What if your favorite gaming company or streaming service or consumer product becomes a beloved financial services company, thanks to tech? Or hell, what if your toothbrush company could offer you…dental insurance?
Seem far-fetched? It’s not. The thing that excites me most about this future is that it can unlock new services for banking the unbanked, built by entrepreneurs who come from the very communities they are trying to serve, whether geographically or through personal experience. The people who understand the problems in their communities will likely build products that better serve them: Who better to build banking services for those on food stamps than entrepreneurs who grew up on food stamps?
The key is that today, better products can spread more easily and cheaply thanks to new distribution channels like messaging and social media, as well as through the non-fintech brands you already use, resulting in lower customer acquisition costs (CAC). A product that is exponentially better than the status quo could spread by referral, creating an organic-growth cost advantage for the company. If the company doesn’t have a high fixed cost structure, then they don’t have to recoup their costs through high fees, thus expanding the range of customers that it can serve. Luckily, the bug that limits better service in the legacy system—in which banks are incentivized to recoup their customer acquisition investment through higher fees—becomes a feature for newer companies, which have low cost structures and efficient distribution strategies.
Great technology shifts aren’t just about fixing existing problems, though, They’re also about opening access to help more people. This is where data is the last piece of the puzzle: Through sophisticated data science and machine learning, we can now unlock more data sources to better assess risk for people who lack sufficient data or are “credit invisible” in the current system. Today, 79 million Americans have credit scores below 680 (the point at which rates can dramatically increase) and 53 million don’t have enough data to even generate a FICO score. Most banks will assume those people are high-risk and charge them higher interest rates (or not serve them at all) by default. Yet we’re swimming in much more, far superior data (not just 5 factors!) than when these credit-assessment systems were invented.
Data science and machine learning can also help us understand the best signals for determining one’s willingness and ability to pay. New experiments in assessing credit-worthiness—such as monitoring rent and cell-phone payments, as well as cash-flow underwriting—have been promising. Globally, companies are using even more creative data types to effectively predict loan repayment, such as how up-to-date your phone operating system is, the number of friends you message with regularly, and even whether you charge your phone fully at night.
People who were previously hard to gauge now become new customers. When more people have access to fair credit, income inequality declines, spurring opportunity and economic growth.
Such data shifts affect not only how money flows around debt, but around income, too. What if people could get paid sooner, instead of having to wait two weeks? Reliable workers (whether salaried or hourly) who are falling short on cash before payday should be able to get access to their earnings for work they have already completed. With data, we know when you’ve worked, what you’ve already earned, and can offer this service. The current model disproportionately punishes the poor. A shortfall before payday means they have to engage with the more sometimes pernicious payday players. Some states tried to solve this problem by using lending caps to discourage payday lending at usurious rates, but the unintended outcome is that the people who need it most now have access to…nothing. It’s another example of good intentions, misapplied. But software can flow around the hard limits of legacy models, better aligning intentions and desired results.
The first wave of fintech companies 10 years ago proved that physical locations aren’t a requirement for banking. Now, the next wave is unlocking the rest of the infrastructure required for building better financial services—banking licenses, payment processors, regulatory compliance, and so on. Previously companies would have had to painstakingly acquire or partner (slow and expensive!), build from the ground up (also slow and expensive!), and figure out a patchwork of compliance and IT (still slow and very expensive!). New financial services companies can leverage best-of-breed infrastructure, create differentiated products that help lower customer acquisition costs, and draw on better data sources to serve many, many more people.
Software not only lets us bypass hard structural problems, it lets us build entirely new kinds of companies and services. All of this is to say: fintech is eating the world.
The banking system today favors the privileged, while large swaths of the population have no options at all. Technology lets financial services players of all kinds innovate beyond the constraints of the existing systems to better reflect the world we live in. Instead of chipping away at deeply ingrained inefficiencies, we can use its very structural limits to build new kinds of companies from the ground up. We can do better than a two-tier system. It shouldn’t be expensive to be poor. And as more people enter the economy by getting plugged into better financial services, everyone wins.