According to the average returns showed by LC, there is a delta between the interest rate for the borrower and the actual return. Wide Delta. Instead of receiving 17% on D's, we get 8%. Half the return is lost to defaults. What if you could marketplace consumer loans, but in this case, compel the borrower to provide proof ownership of a vehicle, home ownership, or partial ownership. If you look at the detailed borrower stats, they show if they own home (Own) or mortgage (partially own). This may work because at the outset of the loan, everything thinks it will work; the borrower is not thinking about defaulting so it may not seem significant to use his assets as collateral (because he doe snot think they will beneeded to repay the principal anyhow). But now, consider, that many of the marketplace loans are backed either fully or partially by these assets. A $3.5K loan request can be backed by a $7K used car. A $40K loan for a large purchase may be backed by a mortgage in which he's bough 50% of 400K house.
Could such a model improve yields? The overall interest rate would have to be lower (but perhaps not by much). But most importantly, the Surrender $ from Default will go down. I mean look at how ominous the numbers look like the MINUTE a loan payment is late. That won't be the same if we knew there's real collateral. One may ask- well if they have a car and wanted to consolidate debt, why wouldn't they just sell the car if it was that important. But that's not how debtos think. They are optimists assuming both.
I think it might be interesting for a closer look.
Are you new to lending? Not sure how much I need to explain vs. assume you know. So I'll bullet point
1) This would bring net returns lower, not higher
2) This would likely increase the servicing fee you pay for lower yielding notes
3) This wouldn't work well with a market place lender structure.
4) collateral you describe is far from an assurance of reduced losses. It could actually increase the credit costs.
Sent from my SAMSUNG-SM-G935A using Tapatalk
bdonovan's questions which started this thread are terribly interesting!
The implied question in his post is how to increase the probability that a borrower will pay back a loans.
Another way to put it is, how to reduce defaults.
The better you can predict default rate, the better returns.
I do not know how interest rates are determined for an individual. I suppose that FICO is a primary variable used in that formula. I assume that FICO is a good predcitor of financial reliability-risk.
I suppose if you did a deep interview, a full psychological assessment, and took one of the borrower's children as a hostage, you could increase probability of repayment. I suppose that more in depth data collection would cost more. There has got to be some set of data which are the best predictors of probability of default. I guess those variables could be determined using factor analysis.
Since any statistical prediction only has a particular margin of error, the more loans the less the variance.
This all and bdonovan's question, makes me want to learn more about the science and math of making loans.
I had taken stats in grad school, but I just don't remember enough to solve this problem, now.
I do wonder if LC gathers enough data and provides the data to us for us to be able to improve the prediction of default rates.
The idea has been considered especially on the small business side of things, but there are several hurdles:
1) Collecting and monetizing collateral increases servicing costs dramatically. You need a large ticket loan for the economics to work. Collecting is also a scummy business which the (formerly) reputable marketplace lenders didn't want to get into.
2) Assessing the value of collateral during the underwriting stage is also difficult and again increases costs.
3) Collection laws vary widely state to state. From a compliance standpoint, its difficult to administer.
Moreover, most collateralized loans such as auto and HELOCs are already low cost, so its not clear that investor demand would be there.
The other problem is that this lending works because the product is homogenous. It doesn't make sense to assemble pools of unlike consumer credit into a portfolio. The marketplace model relies on homogenous loans to make this stuff work. When you add variation, it makes it impractical to do it. Market place lenders are already at a cost disadvantage, since their cost of funding is so high. Layering on additional costs and segmenting loans by weird collaterals just doesn't make sense when trying to optimize a cost structure to compete against banks and other lenders. Automating all that stuff you've described is much more difficult than you'd think. Credit score cards are very complex generally, but not so much on just homogenous consumer credit. This is a reason most successful market place guys do some form of consumer lending.
Interesting comments all. Just a random thought I had. I get that there would be different costs in collecting on the collateral but bear in mind that LC already has steep costs for collection; it's 18% as it is, and 30% of litigation is involved. But some customers simply don't have cash, and if unsecured, perhaps think the law doesn't require them to 'liquidate' and get rid of the car to pay it back. It seems unfair- but less so if you explicitly agreed to it in the first place. Generally, partially secured loans seem to be more economical for the lender - I presume they wouldn't lower their interest rates if the overall cost (including cost to collect on the collateral would be higher); simple example:
"One way to reduce your interest rate (and maybe increase the amount you can borrow) is to turn a personal loan into a, "partially secured," loan by pledging collateral (like a paid off vehicle) that is worth more than 10% of the value of the loan. Ask your loan officer if you'd like to pursue this option."
https://cuonlineuhs.org/loans/loan-types/So I think the real question would be- is the (reduction in default loss - variance in cost of collection) > Lower Return from Lower Interest Rate. Regulatory hurdles- agreed; although many many other companies seem to be doing partially secured loans- I would think the model is understood. Assessing value is work- but I can't see how it's
less complicated than taking a person with a bunch of data points and assessing a risk profile. Cars have titles that can be verified and paired with the Blue Book, and public websites like Trulia have at least a ballpark on just about every home out there.
I do like the idea of taking the borrower's kids as hostage as well! My former manager might be one of those who defaults as he often joked he wanted to 'downsize' by letting go of the kids and dog.
Every bank on every street corner does the sort of lending you describe. But they have loan officers and back office staff devoted to the underwriting, collateral, filing uccs, doing collateral inspections, etc. But the difference is they are balance sheet lenders not trying to make every loan identical. But for an assembly line model, like mortgage or consumer credit, these sort of modifications make the efficiency ratio , price per loan produced, and defect rate all increase. At least that is what I'd suspect. Someone may figure it out
Sent from my SAMSUNG-SM-G935A using Tapatalk
Since I temporarily work at a loan company at a branch, I can say that the process is pretty easy for loans w/ collateral with only a few extra steps added. Basically pictures are taken of the car for proof of milage and condition and obtaining the title to transfer title of the vehicle. All underwriting is done at headquarters so cost is centralized. Filing this paperwork should not add any significant cost to the application process.
When you say dealer do you mean an auto dealer? Basically any loan w/ collateral are with cars that are paid off or have value left over to cover the loan, or in LCs case at least a good portion.It would be in LCs interest to turn to collateralization to recoup some of the investor fees lost from loans charged off and keep investors happy.
I love this discussion. While I'm not sure secured consumer loans are feasible and scalable on a platform such as LC, the subject of how to reduce borrower default is top-of-mind for me. Everyone here hopes their custom filter is some sort of "secret sauce" that will somehow juice returns by lowering default rates. Whether they hand-pick notes or rely on a subscriber service, I think we all are searching to reduce risk and increase return.
One of the only innovative ideas I've heard of in the fintech space is what SoFi is doing when one their borrowers loses her job. Since job loss is one of the primary reasons for a loan default, this intervention strikes me as attacking the problem head-on. Apparently SoFi has some sort of professional network that can match borrowers career services team to get them re-employed in job placement programs. I don't have any visibility into whether this is successful, but SoFi's incredibly low default rate is impressive.
This whole idea seems quite appealing to me, much more so than taking any one's children hostage or hiring the repo man to nab a car. What if LC has some similar service? I don't know how it would exactly work, or even if it would be feasible, but if we're truly interested in this being "peer-2-peer", I'd love to help those borrowers I've lent to find a way to pay their obligations. If the lending community of a specific note could collectively leverage their personal network for employment opportunities, that might prove to be a win-win for both the lenders and borrower. Sure, this is a pipe dream. But one can always wish, right?
Back in the old school days before this stuff was popular, there was a service some question askers required borrowers to use. It verified outstanding debts to us investors and showed how much. I think it even showed when it was paid off. I wrote LC once telling them they should buy the website and implement it. They never did but prosper bought something kinda similar a while back. I still think it could help turn LC from a commodity provider to a relationship business
Sent from my SAMSUNG-SM-G935A using Tapatalk