Much has been made about how we as investors have incomplete information regarding Upstart and the performance of there loans when judging the company.
We can look at indirect measures like loan origination volume, market penetration, bank partners, originator growth etc to get a read on these variables we don’t have great visibility for. At scale, these parties likely have a much better indicator of the true performance and value of Upstart loans vs. outside investors.
The biggest downside of this is that they are back-wards looking indicators, and might not reflect current dynamics. ( Loss rates increasing… etc. )
For example:
If Upstart’s loan isn’t much better than the average FICO based loan, then there’s no reason to take it. So if consumers are taking it, you could infer that there loans MUST be better than their other offers. Loan volume and marketshare are an indicator of this.
UPST has substantially increased loan originations and market share over the last 9 quarters:
Annualized market share = ( Quarterly Origination Volume * 4 )/ personal TAM (96 Billion)
Quarter Origination volume Annualized Market Share
Q4'19 - 1.044 Billion 4.350%
Q1'20 - 1.123 Billion 4.68%
Q2'20 - 164 Million 0.68%
Q3'20 - 909 Million 3.79%
Q4'20 - 1.249 Billion 5.20%
Q1'21 - 1.729 Billion 7.20%
Q2'21 - 2.795 Billion 11.6%
Q3'21 - 3.130 Billion 13.0%
Q4'21 - 4.098 Billion 17.07% !!
(Imagine UPST’s market cap if they originate 17% of all loans… )
There loans MUST be resonating with consumers, or else this growth trend wouldn’t exist. Of course, if there losses are also increasing, then there bank partners won’t like this. This rapid growth suggest upstart is able to offer substantially lower interest rates than their competition. They have stated something to the effect of " our rates are already so competitive there is no point in lowering them more." There rates are so good there is no point in lowering them more… This statement matches the data from this paper on them. (https://www.nber.org/papers/w29840 )
The NBER paper evaluates what would happen if a traditional models added alternative data ( they would get a lot better at approving people), but they are projected to still offer significantly higher interest rates than Upstart. So there is some secret sauce more than just incorporating alternative data. This is on Page 39 Panel B.
Upstart and Banks - https://www.nber.org/papers/w29840
One interesting item in the NBER report is that the key metric tracked by Upstart for borrowers with different credit score is Internal Rate of Return ( page 6 ) IRR is a " measure of whether you’ve made a worthy investment". What’s interesting is that IRR actually increases as the FICO score is lowered; so if Upstart can evaluate the risk of this category better, their bank partners can generate better returns. ( Page 43)
( https://www.indeed.com/career-advice/career-development/how-…)
A Question: Would the difference between initial over-collateralization and target over/collateralization be a good indicator that Upstart loans are performing as expected or at least not worse? Or maybe the Credit enhancement?
Similar stats apply with banks.( Look at UPST auto-retail growth trends. Remarkable growth!!)
https://info.upstart.com/hubfs/Hosted%20PDF%20Content/Upstar…
This report (though likely biased) also gives a good indicator of Upstart’s value proposition for banks.
Another interesting article on UPST:
https://protectborrowers.org/upstart-naacpldf-sbpc-fair-lend…
I think Upstart being friends with everyone, and not trying to disrupt the banks, will work wonderfully for them.
Still Long UPST. Maybe a bad idea. Time will tell 