Dan Feshbach, founder and CEO of MeasureOne, thinks that disclosure of loan-level data can do for the student loan market what it did for mortgages.
Feshbach should know; he founded LoanPerformance (now a subsidiary of First American Corp.), which became a conduit for mortgage lenders that were reluctant to “crack the tape,” and make this data available to investors. Providing it helped to create a market for their private label mortgage securities.
MeasureOne has enlisted lenders and issuers of private student loans to contribute to a deal-level database; it is now enlisting contributors for a loan-level database.
The $63 billion of private student loans outstanding represent around 6% of the $600 billion-plus overall student loan market, which is dominated by federally guaranteed loans. Still, a number of whole loan portfolios have traded in the past year as investors cut their teeth on analyzing these deals. One transaction that MeasureOne worked on attracted interest from 40 different investors and six investment banks.
ASR sat down with Feshbach in early February to discuss the similarities between mortgages and student loans, the appetite for analyzing and transacting student loan portfolios on Wall Street, and the potential for securitizing loans that the U.S. Department of Education makes directly to students.
ASR: Do you see the same potential for the student loan market that you saw for the mortgage market when you founded LoanPerformance?
Feshbach: I do, though the potential is different. And I’m not alone. Before we started MeasureOne, we were approached by several clients who asked us: “Can you do something, like LoanPerformance, in the student loan space?” Although there are significant differences between mortgages and student loans, to some degree, they are more similar than other types of consumer credit. For example, these loans have longer “lives” than credit card and auto loans. They often involve whole families, not just individuals, and there are both public and private lenders involved.
Getting into this marketplace has been challenging. We’ve spent three years getting to know investors, lenders, and issuers here.
To enlist them in our student loan database, and get them to provide the data we need to build out collaborative databases, we’ve had to show them what’s in it for them. But lenders and issuers, at least in the private sector, are coming to see the advantages. One investment banker recently told me that our deal-level data is the first standardized data on the student loan industry he has ever seen. Having standardized data, he believes it could significantly increase the speed of analysis and therefore the number of deals that traders and bankers can review and bid on. So this would substantially increase their productivity and ability to transact.
The deal-level data is our first standardized database, and it’s available today. The next step will be loan-level securities data. We are not there yet but we do have two participants contributing: Goal Financial and South Carolina Student Loans.
To get to the next level, issuers have get comfortable they can use this data to expand participants in the market, increase their rating levels, and improve pricing and liquidity.
But the declining volume of student loan securitization has taken its toll on the investor and dealer communities. And this has slowed the learning process. Today, there are fewer staffers on Wall Street devoted to this asset class, even though it is now the third-largest class. Similarly, investors have never had deep benches of analysts or traders [for student loans].
Today, if you want to analyze individual deals, you have to go to each issuer’s website, download remittance reports or use limited data available through Bloomberg, Intex and Yieldbook, etc. Or you can get easy access to standardized deal-level (and eventually loan level) data through us.
If nothing else, the growing direct-lending federal portfolio will make this more critical. After all, how long can the government continue to originate $100 billion dollars in new loans a year without any securitization?
Did you have the same challenge in the mortgage market?
Feshbach: When we started the LoanPerformance database 25 years ago, a lot of people wondered why even build a database on mortgages, when delinquencies were so low and foreclosures so rare? But over time, they realized that our data was giving them early warnings on performance issues in niche parts of the market, like IOs [interest only] and pay-option-ARMs. Also, it could be used to create new products, to modify loans and to benchmark your performance against other lenders.
In the student loan market, we see these same issues and opportunities. Interestingly, the dynamics of who is making the riskier loans is reversed in the student loan market vs. the mortgage market. In the mortgage market, the private sector was the driving force behind subprime loans, and the government lending — by Fannie and Freddie — was pretty tight, at least initially. It eventually loosened to respond to what was happening in the non-agency market and government pressure to meet affordable housing goals
The situation is just the reverse in student lending. Our recent Private Student Loan Performance Report shows that private student loans are performing significantly better than those originated through government programs. That’s because private lenders have stricter underwriting and are assessing ability to repay.
Obviously, government programs have a different mission, and it’s not our place to comment on that; except to note the performance is much different. Right now seriously delinquent rates (borrowers three payments behind) are running at 3% in the private market according the recent MeasureOne Performance Report vs. nearly 21% in the total market, according to a New York Federal Reserve study. The private sector is only 6% of the total market so it’s not the a factor in deteriorating student loan performance .
How big is the student loan market? Isn’t it much smaller than the mortgage market?
Feshbach: The student loan market is big: It’s recently passed the $1.2 trillion mark, putting it ahead of credit cards and making it the third largest debt market, after treasuries and mortgages. The student loan market is really two markets: a very small private market of approximately $90 billion and a federal market of $1.1 trillion. When subprime mortgages hit $1 trillion, demand for our service [LoanPerformance] grew. As an information company we attracted more customers as the market grew because we were providing information to help make investment decisions whether short or long specific securities or sectors.
In this one, [we have a] slightly different role. First, we need to educate the market on the value of better data and what all the counter parties can do with it. Prior to our first report in December, and the release of our securities database in January, there was only minimal data available to the private student loan market. The government provides even less [information about federally guaranteed loans].
In the mortgage market, a couple of companies — Prudential Home Mortgage, now part of Wells Fargo, and Residential Funding Corp. — were pioneers in making data available to investors and it helped them create a market for their private label securities. Other issuers wanted to follow their lead but not all firms wanted “to crack the tape” and build the analytical and reporting capabilities. So LoanPerformance filled that role: we became a conduit for industry loan-level and securities data.
I can see MeasureOne playing a similar role in this market.
It’s understandable that investors would want loan-level data about private student loans, but why would they need it for federal student loans if the principal and interest payments are guaranteed, or at least 97% guaranteed?
Feshbach: True, your risks are capped, but you still have real issues with the timing of cash flows. Depending on what bond you hold it could be significant. You’re not going to lose money, but have opportunity cost. I’d be surprised if anyone has good projections on cash flows on FFELP [Federal Family Education Loan program] paper. Estimating the percentage of borrowers going into forbearance or deferment is difficult. That’s why loan-level data is so critical to making much better projections. In an ideal world, just like mortgages, you could look at the student loans associated with any securitization. This would be accessible to investors in those deals.
How does Reg AB affect student loan backed securities?
Feshbach: Reg AB II doesn’t apply to past securitizations, and FFELP-based securities have limited disclosures. Some municipal issuers such as New Jersey do not publish remittance reports as surprising as that might seem in this post Dodd Frank world. There’s a regulatory debate as to whether there should be loan level or rep line’ level, in other words, summary data where you take a number of loans with similar characteristics and group them to create a sort of synthetic loan.
On one side, there are issuers who are concerned that investors don’t have the capabilities to analyze loan level data. On the other side, there are investors who’ve worked in the mortgage market and expect to work with very large quantities of data. They have plenty of computing power to analyze student loan portfolios or securitizations.
A lot of whole loan student loan portfolios have traded in the last year. Investors have cut their teeth on these deals and proven they have the analytical skills to evaluate portfolios and build modeling tools.
We’ve also made our tools available to investors bidding on those assets. It’s not a very big market, whole loans trades, but enough are done today that there’s a large base of investors able to work with loan level data.
One private student loan transaction we worked on had 40 different investors and six investment banks were looking at the pool, though not all of them bid. This was in early fall of last year.