|03/04/2021||Apple Bank partners with Upstart|
|12/16/2020||Upstart soars after IPO|
|05/20/2020||Upstart announces new credit API|
|02/14/2020||US Senators want answers from Upstart|
|02/08/2020||Upstart refutes discrimination accusations|
Upstart Says Covid Had No Material Impact on Loan Performance, Believes All Loan Underwriting Will be Powered by AI in the FutureMarch 17, 2021
Yet another online consumer lender has reported that the Covid-era was good for business. Upstart, which went public in December, recorded $1M in profit in Q4 and $6M in profit for the year. Prosper Marketplace, an Upstart competitor, reported an $18.5M profit for 2020 just days earlier.
“Despite the COVID-19 pandemic, we delivered strong growth and profits in Q4 and for the full year 2020,” Upstart CEO Dave Girouard said in the company earnings announcement. “This combination is rare among FinTechs and demonstrates the growing advantages of AI-based lending.”
Upstart actually grew its revenue in 2020 by 42% over the previous year while keeping loan performance steady.
“We’re happy to report that the COVID-19 pandemic had no material impact on the returns that our bank partners and loan investors experienced this past year.”
The company is going full speed ahead on AI-based lending. “We believe virtually all lending will be powered by AI in the future, and we’re in the earliest stages of helping our bank partners successfully navigate that transformation.”
Keenly aware that AI is an overly used buzzword, the company reminded investors about what its AI can actually do.
Our AI models, like all AI systems, are fueled by incredible amounts of data and sophisticated software to interpret that data, while most lenders consider only a handful of variables as part of a lending decision, Upstart’s model considers more than 1,000 variables about each applicant. You can think of these as the columns in a spreadsheet. And as of December 31, 2020, our model was trained on more than 10.5 million unique repayment events.
These are like the rows in the spreadsheet. And we continually upgrade the machine learning software that interprets this data, enabling us to price the next loan on our platform just a bit more accurately. Upstart goes far beyond a singular AI model predicting default risk. We have discrete AI model that improve the entire lending process, including identity fraud, income misrepresentation, loan stacking, prepayment risk, fee optimization, and more.
But of course, our model that targets default risk is the centerpiece of our system. It predicts not just the likelihood that a loan will default, but when that default can be expected to happen.
Upstart also intends to bring that technology to auto lending. The company simultaneously announced that it had acquired Prodigy Software, Inc, a tech that’s been used to assist with selling more than $6B worth of cars.
“…2021, from our perspective with auto, is really a building year,” said Girouard, “And the acquisition of Prodigy, we certainly view as an accelerator toward the point of sale, the majority of the market that happens at the dealership.”
Upstart, the online personal lender that uses non-traditional data like a college education, job history, and residency to evaluate borrowers, is moving forward with an IPO.
The company revealed its financial statements in an S-1 filed on Thursday. In 2019, Upstart generated $164.2M in revenue and had a net loss of $5M. For 2020 through Sept 30th, revenue was at $146.7M with a net income of $4.5M.
The company said that in 2020, 98% of its revenue was generated from platform, referral and servicing fees that it receives from its bank partners. Their bank partners “include Cross River Bank, Customers Bank, FinWise Bank, First Federal Bank of Kansas City, First National Bank of Omaha, KEMBA Financial Credit Union, TCF Bank, Apple Bank for Savings and Ridgewood Savings Bank.”
Upstart borrowers tend to have limited or no credit history, which is where its AI-driven models with 1,600 variables come into play.
“Our bank partners have generally increasingly retained loans for their own customer base and balance sheet,” the company wrote in its S-1. “In the third quarter of 2020, approximately 22% of Upstart-powered loans were retained by the originating bank, while about 76% of Upstart-powered loans were purchased by institutional investors through our loan funding programs.”
Upstart was valued at $750M during its 2019 Series D.
In 2017, deBanked referred to Upstart as the Tesla of alternative lending.
“You hear so much about how Tesla cars will drive themselves, how Google or Amazon home assistants talk to you to as if you’re human,” said Dave Girouard, Upstart co-founder, in an interview back then. “In lending we are the first company to apply these types of technologies to lending.”
Girouard’s co-founder Paul Gu, who serves as SVP of Product and Data Science, was only 21 when Upstart launched in 2012. He’s now 29.
Anna M. Counselman, the third co-founder, is SVP of People and Operations.
Upstart is planning to raise $100M from its IPO.
Upstart, an AI lending platform, welcomed longtime industry advocate Nat Hoopes to the team this week, to lead as Head of Government Policy and Regulatory affairs. Hoopes previously served as the Marketplace Lending Association executive director (MLA), where he grew the trade group and advocated on behalf of its members.
“My hope is to bring the energy that I did in growing the organization [MLA] and also just in tackling a lot of different workstreams to Upstart,” Hoopes said. “But also, deepen their ties with the DC policy community.”
Hoopes is excited to join the Upstart team and advocate for the company to state and federal legislators. Hoopes intends to address the development of two main issues as he enters his new office: facilitating better credit reporting with the help of AI, and using better credit to bring financing options to disenfranchised minority communities.
Upstart uses non-traditional data like a college education, job history, and residency to evaluate borrowers for personal loans. The company recently introduced an AI-powered Credit Decision API to deliver instant credit decisions. Upstart added auto loans to the platform in June, so the new API works with personal, student, and auto loans.
Hoopes said he and Upstart shared a similar motivation: to provide credit to people and improve financial futures, especially to people unfairly blocked from receiving credit.
“I think because of the structural inequality that we have in our society, a lot of minority groups get really left behind and stuck in a low credit score environment,” Hoopes said. “By using more data, and using it in new ways with artificial intelligence we can really level the playing field.”
Hoopes said that he has already seen Federal regulators in the FDIC and the OCC, and the CFPB working on using AI learning in credit underwriting. He said the Fed is planning out how to help banks adopt more of these models to approve more people.
“I think that’s a key initiative,” Hoopes said. “A key area where I’ll be working for Upstart: Engaging with regulators on how to help banks get more comfortable in serving more customers,”
While advocating for banks to use the credit capabilities of partners like Upstart, Hoopes said he would be devoted to ensuring decisions are made with equality and inclusion in mind. Hoopes will stay on as a member of the MLA board, and working in concert with his responsibilities advocating at Upstart.
“At MLA, I helped develop the diversity and inclusion strategies for our part of the fintech industry,” Hoopes said. “I’ll remain active on those issues at Upstart both collectively with other members of the industry as a member of the MLA.”
Hoopes referred to the Diversity and Inclusion strategy released by MLA last month. Board members signed off on the paper, written with the help of the National Urban Leauge. League president and CEO Marc Morial and Representative Gregory Meeks (D-NY) to create a vision of an inclusive fintech industry.
Hoopes addressed what he said was the failure of the American credit scoring system. For instance, according to Upstart’s study in 2019, 80% of Americans have never defaulted, yet only half have a prime credit score. It’s a problem he says disproportionately affects minority borrowers.
According to a Federal Reserve study, more than three times as many Black consumers (53%) and nearly two times as many Hispanic consumers (30%) as White consumers (16%) are in the lowest percentiles of credit scores.
Hoopes said Upstart does not collect racial data from applicants but cites a CFPB test that found Upstart’s platform increased access to credit across race and ethnicity by 23-29% while decreasing annual interest rates by 15%-17%.
The rise of fintech has already rocked the banking and traditional lending industry and now it’s disrupting FICO, the traditional credit scoring method that’s been in place since the mid-1900s.
FICO, which is the credit scoring system created by Fair Isaac Corp, is getting a makeover. The UltraFICO Score, which is scheduled to launch in early 2019, will pull from a consumer’s checking, savings and money-market accounts and add the data to their credit profile. It creates a broader credit picture, one that is designed to lead to more lending approvals than the static formula provides, as long as a consumer manages their cash well. Reports suggest the FICO score could jump by 20 points or more for millions of borrowers.
Meanwhile, fintech startup Upstart has been in the consumer lending business for the past five years. Upstart takes a two-pronged approaching, using more variables and more machine-learning algorithms than the traditional credit-scoring method.
“Using a variety of machine learning algorithms lets you pick up new insights from data,” said Upstart Co-Founder Paul Gu.
The company’s approach has influenced banks that frequently approach Upstart, a couple of which have become partners that are using a fully branded version of Upstart.com.
It’s not surprising considering Upstart is experiencing a lower loss rate versus its banking competitors. Upstart’s Gu explained the average lender issuing a personal loan to someone with a FICO score in the 660 range will typically experience a loss rate of 14%. Upstart’s loss rate is half that.
“That same 660-type borrower in our portfolio has an annual loss rate of 7%. That’s a pretty staggering difference and translates into benefits for our borrowers,” Gu told deBanked, adding that if the company can cut the loss rate, they can, in turn, lower the interest rate. Certainly, non-fintech lenders are paying attention.
“I don’t want to claim credit for anything FICO is thinking about. But I do think we are showing the industry at large that there is a huge amount of opportunity out there, and that you can go after it with technology that is available today. The potential benefits for consumers and your business are enormous,” Gu said.
Upstart’s early focus was on younger consumers with no real credit history but with an education history, which Gu said has yielded great success for the company. Since then they’ve expanded to pursue other groups of people who have similarly been “lost in the cracks” of the traditional credit scoring system, including certain occupations.
Gu explained that while lenders typically examine a potential borrower’s income level and their debt-to-income ratio, there’s more to it than that. Upstart most recently has created a way to include data based a potential borrower’s occupation, which he points out is tricky to quantify.
“Occupations are combinations of words that are hard to group in a useful way for the purposes of data analysis,” said Gu. Nonetheless, Upstart and its team of nearly a dozen data scientists have poured research into employers and occupations to create a classification system and determine how to turn words into numbers to use in their machine learning model.
“It’s not shocking that some professions are more highly correlated with repayment than others. Nurses, for example, are very reliable in paying back their loans,” Gu explained.
Upstart, which has issued consumer loans to fewer than 300,000 borrowers, has made it their mission to constantly improve upon their models to find cracks. “Our best estimate suggests we’ve solved only 8% of the total opportunity so far,” he said.
It’s early days for FICO’s new credit scoring system, but according to reports lenders have already begun to show an interest. Experian has reportedly partnered with fintech startup Finicity to publish the broader credit profile to banks. But the increased competition doesn’t seem to bother Upstart.
“I think there are starting to be efforts made by other players in the space to do some of the things we’re doing. Some of the lower hanging fruit we were uncompeted for earlier might have a little bit of competition. That said, the thing people don’t realize is how much room for improvement is still left,” Gu said.
As for FICO, their new feature is a side-product to the traditional credit-scoring system, not a replacement, which could impact the pace of adoption and innovation. “This kind of technology investment should take 95% of their attention, not 5%,” said Gu.
Online lender Upstart considers more than 10,000 variables such as an applicant’s education, academic performance, and employment background, according to their website, a proprietary system they say is used to detect “future prime” borrowers. But according to a recent Kroll Rating Agency report, their borrower base looks prime even by traditional standards in that their average borrower is 28 years old, earns $95,000 a year and has a FICO score of 690. Upstart lends money (through Cross River Bank) to individuals for a variety of purposes including student loan refinancing and debt consolidation.
In the Kroll report, Upstart asserts its belief that its use of additional data points will outperform traditional credit models, but concedes that their system has not been tested through economic cycles.
Upstart has raised $88.35 million in equity to-date. The Kroll Report was prepared in anticipation of a $163 million securitization transaction that is expected to close this month. They expect to be making $100 million of loans per month by the end of the year.
It’s been three years since we launched the Upstart lending platform, and today we’re pleased to announce we’ve raised $32.5M to take our business to the next level. The funding round was lead by Rakuten, a global leader in internet services and global innovation headquartered in Japan, and by a large US based asset manager. Existing investors Third Point Ventures, Khosla Ventures, and First Round Capital also participated. We’re particularly excited to have Oskar Mielczarek de la Miel, Oskar Miel, Managing Partner of the Rakuten FinTech Fund join Upstart’s Board of Directors.
With more than 50,000 loans originated, Upstart has the highest consumer ratings in the industry, has Net Promoter Scores (NPS) in excess of 80, and has delivered industry-leading returns to loan investors.
Leaders in Technology and Data Science for Lending
Upstart was the first platform to leverage modern data science and technology to power credit decisions, automate verification, and deliver a superior borrower experience. In 2014, we were first to launch next-day funding. In the last year, we virtually eliminated loan stacking on the Upstart platform, a central cause of credit issues in online lending. Today, more than 20% of our loans are fully automated, helping us attract the best quality borrowers with a superior experience.
As a result of our efforts, we’ve seen unparalleled credit performance, with 2016 cohorts our strongest yet. Upstart loans are funded in four distinct ways: 1) whole loan sales to institutions, 2) retention by Upstart’s originating bank partner, 3) sales to Upstart itself, and 4) via individuals in our fractional market. Furthermore, we expect our first loan securitization transaction within a few months.
2017 and Beyond!
We’ve focused considerable effort on our credit quality and loan economics, and the results speak for themselves. We aim to originate more than $1B in loans in 2017, and expect to reach cash flow profitability this year.
But that’s not all. We’re also thrilled to announce that Sanjay Datta has joined Upstart as CFO. Sanjay was formerly VP of Global Advertising Finance at Google, having spent a decade to help build and internationally expand Google’s $80B core economic engine.
Those that know my history at Google will understand why I’m excited to tell you about “Powered by Upstart”, a Software-as-a Service offering derived from Upstart’s top-rated consumer lending platform. From rate requests through servicing and collections, this new service brings modern technology and data science to the entire lending lifecycle.
Anna, Paul, and I founded Upstart to bring the best of Google to consumer lending. Upstart was the first platform to leverage modern data science and technology to power credit decisions, automate verification, and deliver a superior borrower experience. In 2014, we were first to launch next-day funding. As of today, more than 20% of our loans are fully automated and we expect this percentage to increase significantly through 2017. With more than 50,000 Upstart loans originated, we have the highest consumer ratings in the industry and have delivered industry-leading returns to loan investors. With Net Promoter Scores (NPS) in excess of 80, we’re excited about the impact we’re having.
FinTech is disrupting all areas of financial services. As a leading tech platform in marketplace lending, Upstart aims to partner with financial institutions rather than compete with them. Given the pace of change in lending, technology partnerships will be critical in the years to come, and Upstart aims to be a partner the industry can rely on.
But “Powered by Upstart” is not just software – it’s a turnkey solution that provides all necessary document review, verification phone calls, fraud analysis, and (optionally) customer service, loan servicing and collections.
Software-as-a-Service in lending
SaaS has grown exponentially in the last decade because of its obvious virtues: rather than buying, installing, configuring, hosting, and supporting software yourself, the software is delivered over the cloud. It’s more reliable and always up to date. Delivering cloud software can be challenging in any industry. Usability, reliability, and performance are the minimum to play, and effective change management is critical to success. As the team that delivered Google’s SaaS platform before it was called “cloud”, we understand these challenges.
Of course, the regulatory environment in lending raises the bar even higher. We’ve long demonstrated our commitment to trustful and compliant lending, and we’re likewise committed to delivering robust and compliant lending software.
The first merchant cash advance enthusiast ended up the richest man in the history of the world. Jakob Fugger was the cash king of Europe 500 years ago, and his climb to wealth indirectly caused the Protestant Reformation. One of the pivotal events in western history, the Reformation led to the eventual “fad” of democratic representational government— all because some guy bought the future receivables of a silver mine.
In Jakob Fugger the Rich, historian Jakob Strieder writes the Fugger enterprise began as one of the upstart merchant families of the Renaissance. The Fuggers were traders and cloth merchants from Augsburg, Germany. They created a network of aristocratic clients, furnishing weddings and parties through trading warehouses in modern-day Venice, Florence, and Austria. Jakob Fugger I lent some money around, but when Jakob Fugger II joined the family shipping warehouse in Venice, he looked for a better return on capital.
According to International Business History: A Contextual and Case Approach, Fugger entered an agreement to supply some cash- 23,627 Florins to a silver mine owned by Archduke Siegmund in 1487.
Siegmund had plenty of silver laying around for collateral; he just needed cash for the day-to-day. It was a collateral-backed loan, common today: if he couldn’t pay it back, the Fuggers would get paid in silver. The transaction worked so well that a year later, Siegmund reapplied, this time in a revolutionary way. Siegmund would get 150,000 florins, and the Fuggers would get paid the future receivables of the silver mine: unrefined and cheap future silver for cash now.
The problem, written by historian Greg Steinmetz in The Richest Man Who Ever Lived, was the Church. Any interest-based transaction was specifically outlawed, though there were hundreds of lenders during this era. The line from Luke 6:35, “Lend and expect nothing in return,” was taken by the Church to mean an outright ban on usury, defined as the demand for any interest at all.
Even savings accounts were considered sinful, but Venetians ignored these rules as they preferred making money to pleasing God, entombed in the motto “First Venetians, then Christians.” Fugger began accepting deposits like a bank to his clients, with a 5% return to investors.
But convicted usurers could be excommunicated and denied a Christian burial, a nightmare for a capitalist who relied on a Christian network. Fugger did not worry about punishment or the apparent sin of money lending, but as he became a fixture in European society, his reputation became increasingly vulnerable.
Fugger needed the laws to be changed, or at least relaxed, or his lending business was in trouble. In 1515, he wrote a letter to Pope Leo X and funded a debate in the St. Petronius Basilica in Bologna. The debate ran for five hours, a back and forth of philosophy, scripture, and rampant crowd heckling. In the end, it was declared a tie, but Pope Leo X that year signed a papal “bull” reforming the concept of usury.
Originally, the Church pointed to the philosopher Aristotle’s model for determining what was okay to charge for and what wasn’t. Aristotle had said that charging someone for a cow because it produced milk was fine, but money was a dead thing and unfair to profit from.
A silver mine produced silver and as such paying cash for the future proceeds of the mine had allowed Fugger to more or less carry on his business. It wasn’t called merchant cash advance back then but he applied that model wherever he could. Not everyone in need of money had a business, however, and it was critical that he be allowed to charge interest when circumstances called for it.
More than a millennium after Aristotle, Pope Leo X found that risk and labor involved with safeguarding capital made money lending a living thing. As long as a loan involved labor, cost, or risk, it was in the clear. This opened a flood of church-legal lending: Fugger’s lobbying paid off with a fortune.
Jakob Fugger was off to the races and he greatly expanded his financial services business. Historian Dennis McCarthy found that the Fugger family grew their war chest nine times over in the next seventeen years, a gain of 927%. Their funding efforts bought a trading empire, and they entered into agreements with nobles that placed entire countries as collateral.
McCarthy wrote: That was one of the problems with the Fugger model- “how does one take possession of Austria or France or Spain when its rulers default or lag behind debt repayment schedules?”
After gaining the good faith to lend in the Church’s eyes, the papacy itself became a Fugger customer. Positions in the Church were inseparable from social and political power, and the only way to get a place on the totem pole was by paying for a title. Just as the richest silver mine owners didn’t have the cash to pay for lunch- so did wealthy aristocrats need capital to afford positions in the cloth.
By the time Martin Luther “nailed” his 95 theses to the door of a church in 1517, he was rallying against the Fugger funding family and its stranglehold on the Roman Catholic Church.
It all came down to an in-house promotion. Albert Brandenburg brought a whole new meaning to the concept of “moneychangers in the temple.” A German Archbishop of Magdeburg, Brandenburg was promoted to Elector of Mainz: the second in command of the Holy Roman Empire. Unfortunately, he had to pony up 21,000 ducats to pay the Roman Curia (the Church’s admin)- for the title. Naturally, he didn’t have the cash, and the Fuggers stepped in.
Brandenburg got a loan on interest. To pay it back, he also paid Pope Leo X for the right to sell indulgences. Indulgences were contracts the church sold to forgive sins, allowing believers to purchase their way out of purgatory and into heaven. A fresh round of indulgences was printed to fund the construction of St. Peter’s Basilica, and Brandenburg was entrusted to sell them in 1517. (Their sale was later banned by the Church in 1567).
The sale of indulgences interlinked the Church with Fugger, and solidified Luther’s desire to maintain the Faith through an alternate system. Luther’s complaints spawned the Reformation, and his followers and independent revolutionaries like John Calvin would bring the rise of Protestantism, the Church of England, and ultimately what historian Alec Ryrie wrote as the foundation of modern mercantilism.
“I’m saying that there are some specific parts of modern life that derive directly from the Protestant Reformation. We couldn’t have these features if it hadn’t happened.” Ryrie said. “That combination of free inquiry, democracy, and limited government is pretty much what makes up liberal, market democracies. It runs the modern world.”
To this day, no one is sure of the extent of the Fugger fortune. Historian Mark Häberlein found that Fugger struck a deal with Augsburg Tax authorities in 1516: he agreed to pay an annual lump sum on the condition that his family’s true wealth would never be revealed. He died in 1525.
To get an idea of the extent of his wealth, we can base calculations on the cost of butchering a pig in 1522 (yes, that’s a real metric.) It cost one Gulden, a new coin minted in 1500 to butcher a hog. The German coin contained about the same amount of gold as a Florin.
Based on those ham prices, Jim Ulvog from Ancient Finances estimated that in 2017 a single florin would be worth ~$900, and other writers have put the florin in the same range. Though the true wealth of the Fuggers may never be known, when Charles V aimed to take control of the Holy Roman Empire in 1519, the Fuggers were lending Charles 543,000 guldens to buy votes: approximately $448 million. That’s just in a single deal.
It’s been said that merchant cash advances or sales-based financing is relatively new, but it could be argued that such transactions are so old that life as we know it in the modern world only exists because a guy 500 years ago was engaged in non-loan transactions to fund businesses in a manner that was Church-compliant and wanted to expand.
Fintech IPOs are back. Affirm, a fintech company whose platform offers “a point-of-sale payment solution for consumers, merchant commerce solutions, and a consumer-focused app,” is the latest company to file for an IPO.
Affirm’s S-1 was filed earlier today, revealing that they intend to raise $100 million. The company generated $509M in net revenue during its fiscal year ending June 30 and a net loss of $112 million.
|Fintech||Date Filed||Date Public||Amount Raised|
Unrelated to fintech, but still “tech” are pending IPOs for DoorDash and Airbnb.
upstart(founded by ex-googlers) further down the automated fintech ladder., , ahhh....,i use basic quant strategies, i use to be better at it ...now i am looking for nsfs lol sigh and google and amazon have an obscene amount of money,...
upstart(founded by ex-googlers) further down the automated fintech ladder....