- Today we have few articles but really good ones.
- Read an in-depth analysis of Lending Club’s business, triggered by the poor article in Bloomberg.
- A great article showing clearly how the 2016-student-loans-market compares unfavorably to the mortgage-market pre-2008.
- Did you know that Kabbage is using FICO’s rules engine ? Neither did I.
- And in New Zealand a very interesting relationship between Harmoney and the NZ regulator.
- Camino Financial, lender targeting Hispanic SME business owners, raises $2mil series seed. I continue to believe the p2p lending space is on an upswing.
- A very long but a quick read article on Lending Club’s and how Bloomberg’s article was poorly documented, poorly researched, and poorly written. Great charts and an article that is worth a read. Note: the data clearly shows that borrowers with verified income perform less well. I can’t think of any reason why that is except if the verification process itself annoys borrowers who have other loan options and pushes them to take the loan elsewhere because of “less paperwork”. A very interesting conundrum.
- Do you remember that in 2006 one said: “people will not default on their homes, they need homes, it’s ok to give them a mortgage”? In 2016 people are saying “student loans are nearly impossible to be discharged in bankruptcy” so it’s safe to lend against them. Guess what the government did in both cases: it deformed the credit market through intervention. What happen in 2008 ? What is going to happen with the student loans ? We will write about it when it happens. In the meantime, here is a great point of view based on quotes and research from an MIT professor at Sloan.
- For entrepreneurs: How and why Kabbage hired FICO for their rule engine. Also, a must-read, and a great lesson about reinventing the wheel.
- A great article showing how the regulator looked at Harmoney’s business model. Approved it. And now the regulator is suing Harmoney because of their business model. It’s fun to do business in New Zealand. Do we need regulators to regulate regulators ? Not that my opinion matters much, but charging a fee that is a % of the loan principal paid at disbursement , in my eyes, does not make Harmoney a creditor as the fee is paid regardless if the loan is paid back or not.
- News Comments
- United States
- Camino Financial Raises $2 Million in Equity Financing, (PR Newswire), Rated: AAA
- How Bloomberg’s ‘Expose’ On LendingClub Is Wrong, (Seeking Alpha), Rated: AAA
- The student loan conundrum, (Washington Times), Rated: AAA
- How Kabbage knows if it can lend you $100,000 in just seven minutes, (IT Business), Rated: A
- European Union
- EstateGuru.co - a p2p lending platform is launching its first investment opportunity in Latvia, ( Company Press Release), Rated: B
- New Zealand
- Commerce Commission attempts to establish how the Consumer Finance Act 2003 applies to peer-to-peer lender’s ‘platform fee’; Harmoney ‘disappointed’ with action, (Interest), Rated: AAA
Camino Financial Raises Million in Equity Financing, (PR Newswire), Rated: AAA
Camino Financial, an online credit marketplace focused on extending small business loans to Hispanic-owned businesses, the largest and fastest growing underbanked business segment in the U.S. The Company recently raised an additional $2 million in equity financing.
Last November, Camino Financial raised $750,000 in seed capital backed by institutional and angel investors.
The company raised an additional $2 million series seed financing led by Collaborative Fund and Hunt Holdings LP, in collaboration with Comcast Ventures Catalyst Fund and Impact America Fund, among other influential angel investors..
Twin brothers, Sean and Kenny Salas, incubated Camino Financial while completing their MBAs at Harvard Business School. As sons of an entrepreneur who lost her Mexican restaurant chain after 25 years in business, the Salas twins founded Camino Financial to empower small business owners with knowledge and capital.
How Bloomberg’s ‘Expose’ On LendingClub Is Wrong, (Seeking Alpha), Rated: AAA
A recent post on LendingRobot analyzed the performance of LendingClub’s repeat borrowers (Disclosure: I work as an IAR at the company that posted this analysis) and concluded that:
- LendingClub did not hide that it was issuing loans to repeat borrowers.
- Repeat borrowers had delinquency rates 4% lower than the platform average and had a higher ROI.
The Bloomberg article states:
“That Sims was able to use an algorithm and a home-built computer to pinpoint problematic loans…”
Saying that the loans are problematic supposes that these loans perform worse than average. As seen in the previously mentioned study, repeat borrowers perform better than the average loan.
The Bloomberg article states:
“In Portland, over coffee, Sims said he’d found thousands of instances from 2009 through 2011 in which LendingClub seemed to allow borrowers to split their loans in two if their first attempt to get a loan didn’t find any takers.”
It shouldn’t be a surprise that a borrower can have more than one loan since the following is stated on LendingClub’s website:
The Bloomberg article states:
“It also chose not to verify the incomes of most of its borrowers, arguing that unverified loans performed as well as verified ones. The practice is defensible-credit card issuers rarely verify incomes-but banks generally do so for the sort of fixed-rate loans LendingClub offers. In any case, the practice cut against Laplanche’s story about his standards being more rigorous than those of conventional banks.”
The article neglects to show the data on LendingClub’s income verification. In fact, LC has verified the income of over 50 percent of borrowers since 2010.
What is more interesting is looking at the charge-off rate of these loans. Write-offs for non-verified borrowers are consistently several percentage points lower than verified borrowers.
The article states:
“Even so, in loan applications, borrowers have made reference to bankruptcies, divorces, and medical procedures-information that could be linked to them publicly if someone were determined to do so. ‘What’s scary is that everybody who has taken out a loan is in there,’ Sims said. ‘You don’t expect that if you apply for a loan, everybody is going to be able to find your income and everything else. It’s this black hole of bad stuff.'”
This is partially true. Data for issued loans is publicly available, though the company redacts personally identifiable information. Data for applicants who did not receive a loan is heavily truncated. Here is a sample of these entries:
The article states:
“British asset management firm Baillie Gifford, at one point LendingClub’s second-largest shareholder, sold its entire stake.”
This statement leads the reader to believe that all investors are jumping ship. Most surveyed analysts don’t recommend selling the stock.
In addition, the article leaves out that, as one major investor was selling, another was buying. After all, there are two sides to every stock transaction. For instance, Shanda Group bought an 11.7% stake in LendingClub in May, and purchased an additional 3.4% of the company in June to become LC’s largest shareholder.
Regarding the stock price, the article notes that after Renaud left the company:
“LendingClub’s shares fell again, losing half their value by the end of the week.”
The article neglects to share that the stock has rebounded 60% from its lows in mid-May.
The article states:
“The company arranged about $2 billion in loans last quarter-and has more than $500 million in cash, according to regulatory filings.”
Actual cash and cash equivalents listed in the SEC filing is $572.9 million. The article neglects to mention that LendingClub also has an additional $259.5 million in sellable securities, which brings the company’s total available reserves to $832.4 million. The company could lose $80 million per quarter for more than two-and-a-half years before this reserve is exhausted, something other peer lending companies probably envy.
The article states:
“… Sims tuned in to an earnings call, learning that LendingClub had lost more than $80 million. Its CFO was also resigning.”
These two sentences put together imply that the CFO was forced out due to the company losing $80 million. The article leaves out that the CFO decided to leave earlier in the year, and was asked to stay until the company was able to “navigate current events.”
Reading articles like the recent Bloomberg piece makes it sound like there are still “shady things” to be uncovered, but this is unlikely. With the stock currently trading at 2x book value, the bad news of lower revenue and originations for the next two quarters has already priced into the stock. I expect that if the company’s recent key hires succeed in quelling regulators and attracting institutional money, the stock will rebound quickly.
The student loan conundrum, (Washington Times), Rated: AAA
At every political event I’ve been to this election cycle — Democrat or Republican, down-ballot or top of the ticket — a millennial always asks the same question to whomever is running.
“What are you going to do to make college more affordable and lower student loans?” they ask sincerely. And there’s no question as to why — average annual tuition at private universities have skyrocketed to $26,740 in the 2014-2015 school year from $9,882 in the early 1980s.
What these millennials really need to understand, is that high-college tuition is partially the result of well-intended but ill-conceived liberal policies that were kicked into high gear under President Barack Obama.
For Mr. Obama took the student loan industry public when he passed the Affordable Care Act in 2010, making the Department of Education the originator of about 90 percent of U.S. student loans. Today, student loan financing roughly equals nearly a quarter of the government’s annual federal borrowing.
As more students were able to borrow more money, colleges, in turn, increased their tuition costs, according to an analysis of the Federal Reserve Bank of New York.
Remember the government-induced housing bubble? Same thing. The more loans you can get — regardless of credit qualifications — the higher the prices become. Everyone is happy — until repayment time comes.
The employment market, although better than it was in 2009, is still in recovery and, as millennials will tell you, jobs aren’t plentiful and competition is more intense — they’re now competing against more college-educated youth.
More than 40 percent of Americans who borrowed from the government’s main student-loan program aren’t making payments or are behind, according to recent data.
“If student debt continues to skyrocket, the federal government may have to deal with as much as a $500 billion write-down when future defaults and loan-forgiveness programs are factored in,” Mr. Pianko wrote.
What’s the answer?
S.P. Kothari, a professor at the MIT Sloan School of Management, had an interesting idea: restructure the Federal Direct Loan Program to base loan amounts on field of study and earning potential.
“When the government is in the business of offering credit, as it is now with student loans, it should think hard about credit risk,” he wrote in The Wall Street Journal with Korok Ray, a professor at the Texas A&M University. “One of the chief lessons of the 2008 financial crisis was that mispricing credit risk can have catastrophic consequences. Yet the government’s Direct Loan Program mostly ignores the credit risk of students, treating them largely as identical in their long-term employment outcomes.
“If the housing market crash taught us anything, it’s that all bubbles will burst. It’s time for the government to start backing off,” she wrote.
How Kabbage knows if it can lend you 0,000 in just seven minutes, (IT Business), Rated: A
A partnership with Scotiabank will give access to the loans platform founded in 2009 to the bank’s customers in Canada and Mexico. The loans in Canada are for a six-month term and work as a line of credit, and can be for as little as $1,000. Even though Scotiabank describes Kabbage as a financial technology startup that’s disruptive to the loans business, it’s supported the firm since last year, when it invested in Kabbage through the ScotiaBank Digital Factory.
How Kabbage has been able to evaluate a mass number of loan applications and determine fair rates for those approved has changed since it first started operating to serve the eBay sellers market in 2009. At first, it was licencing a rules model to process the data from another third-party provider. But when it began struggling to make instant decisions and lacked a variety of financial models, Petralia knew she had to fix the problem.
irst, she tried to build an in-house rules engine.
“I made the mistake of assuming we could do it,” she says. “We couldn’t execute on that and we needed more support.”
Looking back on the experience, Petralia compares trying to build a new rules and decision-making software to trying to build a custom CRM for her company. She uses Salesforce.com for her CRM rather than trying to focus time and energy on something that’s not a core competency, and for the rules engine, she turned to FICO.
After hiring external consultants to evaluate options and going through an RFP process, Kabbage settled on the San Jose, Calif.-based firm that’s best known for creating the industry-standard score for evaluating consumer credit risk in the U.S. Namely, it adopted FICO’s Origination Manager Decision Module for the task.
That’s no surprise to Tim VanTassel, vice-president and general manager of FICO’s credit risk lifecycle line of business. He says that supporting alternative lending services like Kabbage is now part of of FICO’s business strategy.
FICO’s module crunches the data that Kabbage pulls in and applies the rules to it so loans can be automated, he explains. Since in the loans business, you expect that some segment of the loans will go bad, you charge enough interest to cover that loss. But the real trick is in determining the appropriate amount to both cover that loss while keeping your products priced to sell.
A typical Kabbage loan transaction could call on 30 different data sources and up to 50 different models to score that data at any different time. So at least one aspect of FICO’s software that helps ensure the best model is being used is a Champion-Challenger process.
This predictive analytics algorithm will compare different models against a historical data set to determine which model is most effective. By doing this, VanTassel says that Kabbage can hone its understanding of specific markets over time.
Petralia says the software also allows her team to test the deployment of models. That makes it easy for an analyst to try a model on a smaller portion of loan applications, and if it proves a good change, bump it up to cover a wider swath.
If she had to function without FICO, she could turn to competitors such as IBM for a rules engine, she says. But she’s content with her vendor.
EstateGuru.co - a p2p lending platform is launching its first investment opportunity in Latvia, ( Company Press Release), Rated: B
EstateGuru is the leading Baltic peertopeer lending platform for secured short and mid term property loans.
Within two years the platform has provided funding for over 65 projects by investors from 27 European countries. In August 2016, the platform reached a remarkable milestone when loans provided through the platform surpassed €10 million. Now, EstateGuru is launching its next phase of success by entering the Latvian market.
EstateGuru’s 4500 investors are always ready to fund promising business plans provided security is in place.” Summarises EstateGuru CoFounder and CEO Marek Pärtel.
EstateGuru’s first Latvian project is a bridge loan secured by residential project at Zentenes 19a in Riga with the loan amount of up to €430,000 allowing investors to earn 12.5% p.a interest rate from their investment.
EstateGuru platform has provided successful funding for several highend development projects, the biggest being Mardi street residential development project in Tallinn city centre (€2,2 million). Even though the minimum investment is €50, biggest portfolios in the platform exceed €500,000.
The funding speed on the EstateGuru platform is increasing by the project with the current record being €31,000 in 31 minutes.
Average loan sizes through the platform are €162,227, with the LTV (Loan to value) ratio being 58.62% and historical average returns of 11.77%. The average loan term is 14 months. To date, €3.4 million of loans provided has already been repaid and investors have received a total of €409,771 in interest. None of the loans have defaulted and all present repayments are on schedule.
Commerce Commission attempts to establish how the Consumer Finance Act 2003 applies to peer-to-peer lender’s ‘platform fee’; Harmoney ‘disappointed’ with action, (Interest), Rated: AAA
Since its incorporation in May 2014, Harmoney has charged borrowers a ‘platform fee’ that is added to all loans funded through its platform.
Before December 2015 Harmoney set the fee at a percentage of the amount borrowed.
The Commission’s view is that the platform fee is a credit fee under the Act, and that Harmoney is a creditor. That was changed, but the commission says it understands that the lending transaction remains fundamentally the same
Harmoney says it is not a creditor, and that the fee is the revenue it earns for running its loans marketplace.
The commission says that while its case relates primarily to the original platform fee regime, the Court’s answers will apply to any fee structures that are similarly constituted.
There are now four P2P companies in operation, with different business models. Both Squirrel Money and Lending Crowd launched with flat fees, assuming they were party to the CCCFA. But LendMe has argued its fees are not credit fees under the CCCFA.
The Commission has made its application under section 100A of the Commerce Act which enables it to seek the opinion of the High Court on issues of law. This is the first time that the Commission has made an application under section 100A in a consumer credit case.
Harmoney’s joint chief executive and founder Neil Roberts said that prior to launching the Harmoney peer to peer marketplace the company documented the business model in detail following extensive legal advice and working with all stakeholders during the licensing process prior to being granted its peer to peer licence by the Financial Markets Authority.
The Commerce Commission disputes this comment, saying it was not given all the information about Harmoney’s set up and fees modelling prior to launch.
Through its PR company Harmoney has, later in the day, disputed the commission’s assertion and produced two documents, here and here that it says shows the commission was consulted on fees and other aspects of Harmoney’s business prior to launch.
For the year to 31 March 2016, Harmoney recorded a loss of $14.2 million before tax on revenues of $8.6 million. Roberts says, “We have invested heavily in the platform to open up a new asset class for retail investors and a frictionless experience for borrowers, a genuine alternative that creates competition in the financial markets. Like many tech start-ups we are not yet turning a profit, and continue to invest as an innovator in the P2P lending market.”
LendMe chief executive Marcus Morrison said his company was “having quite a bit of dialogue with Comm Comm (and their lawyer) late last year” but have not had any since.
“We made our stance and process very clear to them at that time. We believe that the our platform fees are not deemed credit fees under the CCCFA in that the borrowers’ ability to drawdown their loan is in no way contingent upon their payment of the platform fee (paid to LendMe) and that the fee is not in any way passed on to the lender. “In any case, virtually all of our borrowers have either been commercial entities or trusts and so are not subject to the requirements of the CCCFA,” Morrison said.