Lending (Capital) in the Twenty-First century

Lending (Capital) in the Twenty-First century

Finance seems too big. Certainly, some banks are too big to fail. But what if finance isn’t big enough yet? Financial services are highly concentrated. Much like Thomas Piketty’s argument around income inequality — the wealthiest 1% of households control 20% of income here in the United States — the distribution of consumer financial services can be viewed in the same way — but on a global scale.

Right now only 50% of people worldwide have access to a savings account and just 20% have access to a loan product from a financial institution. There are roughly five billion people that banks don’t serve today. More than anything, finance is poorly distributed.

It’s staggering to think how banking, one of the oldest, most powerful, and most globalized industries avoids four fifths of humankind. Also telling is the fact that banks haven’t figured out how to profit from lending to those other 5 billion people (while still alienating plenty of their actual customers via poor service). Last year, in our white paper, we predicted that new marketplace lending platforms would originate $1 trillion in loans within the decade — but what if its potential is even larger? Lending holds the highest transformative potential for those excluded from banking. After all, it is through credit that money is created in the modern economy.

Data and new technologies are the key to unlocking credit globally

In the US, the ability to open a bank account, get a credit card, or get a loan to start building a life or business all hinges upon the approval of credit bureau data. Lending Club has built a multibillion-dollar business by accurately pricing the risk of borrowers that were overcharged by credit card operators. Emerging lending platforms are finding new ways to bypass and best FICO — the ‘gold standard’ of credit bureaus. In the developed world, banks have been sitting on (and not utilizing) a wealth of data available that would improve their credit decisions. Banks need to get smarter just to catch up with the underwriting capabilities of new marketplace lending platforms. The data problem also hurts non-bank lenders. At the lower end of the spectrum, most payday lenders refuse to report to bureaus the repayment history of their clients, as they fear graduating them into lower interest loans.

Further, when you have a privately owned repository of repayment data as you do with credit bureaus here in the US — it’s incredibly expensive to access it. A hard pull from a single US credit bureau costs a lender between $1-$2 at scale, depending on their size. While that works fine for an auto loan or mortgage where underwriting is done only a few times within a consumer’s lifetime, it’s entirely uneconomical to do so on a regular basis — like for an adjustable rate revolving loan or other small loan amounts (i.e., Can you lend me a couple bucks? This 6-month old printout shows I have a good FICO score!).

Across the world, the challenge is that most countries have no credit bureaus, and of those who have them, they only report negative payment behavior (as opposed to positive/negative behavior and a numerical score). Even then coverage is patchy, as most of the population is not listed in the credit bureau databases. Most utility and bill payments are not captured or reported. Moreover, they still carry a lot of the data collection by hand.

Underwriting based on behavioral data, using machine learning, neural networks, and other advanced statistical techniques, will provide the scalability required for financial institutions to offer more to those now without access. Mobile phones are making most of the world’s people accessible electronically for the first time, and provide a wealth of data that a stale stack of stamped bill receipts simply can’t compare with. Unless banks and traditional lenders think of themselves primarily as data and technology companies, they will become increasingly irrelevant.

There are great opportunities ahead, and a new notion of identity can emerge via new technologies. The bitcoin blockchain holds enormous potential here. I won’t get into the weeds of bitcoin, but given the distributed nature of the blockchain, the emergence of a decentralized credit bureau with global reach is now a real possibility. Using the blockchain ledger to create a global repository of the world’s credit transactions will establish a new way to assess underwriting and risk with data that was previously unavailable. The notion of a thin-file or no-file borrower can become irrelevant — people and companies will have full, instant access to their complete credit history, and make use of it to access loans in any country, regardless of how long these potential borrowers have been there.

Lowering costs, increasing transaction frequency

Every day most Americans swipe their credit cards for groceries, utility bills, or a night out on the town. Charges are applied — merchants pay roughly 30 cents plus 3% of the value of the transaction — and while expensive, merchants oblige because of the convenience credit cards offer to their customers. But imagine that you are a merchant who is also a part of the 3 billion people (50% of the world’s population) living on less than $2.50 a day. That 30 cents credit card fixed fee just does not add up. Who would pay a third of their income (and often half of their margin) just to swipe a card?

In this case, bitcoin — the Internet of money — holds the most potential to bring transaction costs down. Anyone can send 1000s of bitcoins across the world just as easily as sending one hundredth of a million bitcoin, without incurring transaction or remittance fees. Moreover, unlike cash, full information details on the transaction can be attached to it, and be programmed to be disbursed only after meeting whatever condition is agreed upon — all without any extra cost. Talk about reducing the costs of moving money around.

Most of the costs of holding and moving small balances will be stripped out of such a system (eventually including Deposit Insurance). Electronic delivery and instant disbursement of funds will take place via mobile phone. Daily amortization would be possible for microloans. Non-banks, all while still turning a profit and serving more clients, will perform many of the banks’ functions better. One really wonders what banks will use their branches for 20 years from now!

Change is coming, past a bank near you

The current banking system is not equipped to carry the task. Banks use mostly centralized legacy technology infrastructure, and rely heavily on expensive branches and manual processes. Moreover, given their role in the recent financial crisis, they will keep facing increasing capital and regulatory constraints. Banks will continue to pull from lending to ‘risky’ clients. Given their high cost base, banks can’t afford to underwrite and lend money in small amounts as would be convenient for clients.

Under this wicked cost structure, banks rely heavily on fees and tricky penalties to make money — no need to hold capital reserves to earn those. Take overdraft, which generates $30B+ in fees per year, on automatic overdraft loans of less than $30B annually. One recent FDIC study found that the average consumer would pay an implied APR of 3520% on a $24 ATM overdraft (at the median overdraft fee of $34), if the overdraft loan were paid back in two weeks. By the way, in 2007, total overdraft charges were $17.5B, so this revenue line has grown at a ‘healthy’ 9% CAGR since.

Marketplace lending platforms hold the key to making finance as big and distributed as it should be. Embracing new data sources and underwriting techniques, as well as using cost reducing and feature-enabling technologies like bitcoin, is where the next mass scale opportunities will be unlocked. Lending in the twenty-first century comes with the promise of transforming the world. The addressable market is immense — 5 billion people strong. There are plenty of loans yet to be made, and one trillion of them might just be the beginning.

This post was originally published March 27, 2015 on Medium. An earlier draft of this post was originally published at the LendIT blog, and at www.foundationcapital.com.