I am not sure why their investor presentation is very vague, obscuring the nature of what they actually do to make money.
Here is what I have found about Pagaya.
Pagaya uses machine learning and big data analytics to manage institutional money, with a focus on fixed income and alternative credit.
The company’s technology platform, Pagaya Pulse, runs on a suite of artificial intelligence technologies and state-of-the-art algorithms.
This company was founded in 2016 in Israel, currently headquarted in NY and Tel Aviv.
Cofounder/Gal Krubiner, CEO
Cofounder/Avital Pardo, CTO
Cofounder/Yahav Yulzari, Chief Revenue Officer
The company states the Pagaya’s network is fully-automated and has processed over 17 million applications in the last 12 months as of the end of June 2021, with a new application analyzed every second.
They are currently purchasing/packaging loans for sale in unsecured consumer, auto, credit card, point-of-sale; and Pagaya has plans to offer their solutions for mortgages and insurance related products.
When they say “its proprietary API seamlessly integrates into its next-gen infrastructure network of partners to deliver a premium customer user experience and greater access to credit.” They are referring to marketplace lenders and institutions, not the actual consumer (borrower).
According to KBRA reports (Pagaya had its first public ABS in October 2020:
To date, the Company has completed 16 securitizations for $4.9 billion since 2018 with 12 consisting of unsecured consumer loans.
Here’s how their consumer loan ABS works.
Pagaya purchases unsecured consumer loans from the following marketplace lenders (MPL) LendingClub, Marlette, Prosper, Avant, Upgrade. Pagaya has purchased loans from Prosper and LendingClub since 2016, Upgrade since 2018, and Avant and Marlette since 2020.
Each MPL continues to service the loans.
Pagaya does not rely on the MPL’s credit criteria when selecting loans for purchase and instead selects loans using a proprietary and independent credit analysis model utilizing machine learning technologies.
The model analyzes each loan prior to selection for purchase to assess its quality, risk-adjusted expected return, and probability of default.
Pagaya’s model incorporates a variety of factors, including: borrower and loan characteristics, performance of loans previously sold by the MPL, asset management factors, concentration limits, and legal restrictions.
Pagaya evaluates the risk of each borrower and loan and assigns a grade to each loan.
Pagaya then ranks the risk-adjusted expected return and characteristics of each loan offered by the MPL.
Pagaya purchases certain ‘custom’ loans that the MPLs originate specifically for Pagaya and also purchases loans that the Platform Sellers typically originate for other investors.
These ‘custom’ loans may not meet all of a Platform Seller’s traditional underwriting criteria (such as minimum FICO and DTI), but Pagaya believes it has the ability to properly grade such loans and that these loans offer a better investment opportunity.
According to Pagaya: ‘Custom’ Loans do not perform differently than non-Custom Loans with the same “Pagaya Loan Grade”
Then, these selected loans by Grade are packaged together into the ABS for sale to institutional investors.
Presumably, Pagaya gets a fee from the institutions that buy these loans/securities
“Through the first three months of 2021, Pagaya registered revenue reflecting $300 million annually, and a profit of $100 million for the year.”
So that means they generated $75 million in revenue in Q1 2021 and a profit of $25 million in Q1 2021.
It looks like they generated $100 million in revenue in Q2 2021.
The company’s revenue in 2020 stood at $94 million.
Looks like they took 9.72% of ‘network volume’ (I’m assuming the loans purchased volume?) as revenue (" Network volume is defined as the gross dollar amount of assets that are originated by lenders enabled by Pagaya A.I. technology and are acquired by institutional investors")
“Pagaya has tripled its workforce over the past year and currently employs 350 people in total, 250 of them in Israel and the rest in the U.S.”
On Linkedin, 426 employees listed as of today.
There are 20 glassdoor reviews, 4.6 rating, 100% approve of CEO
They are definitely growing fast.
I haven’t looked closely at default data yet but will when I get more time.
But it does seem from Barclays Hedge fund ranking for 2020, they were the #3 ranked Fixed Income - Long-Only Credit fund, so they are doing something right:
#1 Renminbi Bond Fund, 19.50% CAGR, 48.72AUM
#2 DWS Concept Institutional Fixed Income 16D, 18.87% CAGR
#3 Pagaya, 13.89% CAGR, 314.65 AUM
#4 Eastspring US Corporate Bond Fund C, 10.59% CAGR
Similarities between Pagaya and Upstart:
-Pagaya tries to increase financial inclusion by using AI/ML underwriting to increase loan application conversion rates for MPLs. They help MPLs lower their FICO requirements or DTI requirements. They attempt to increase loan volume while keeping defaults the same.
Differences between Pagaya and Upstart:
-Pagaya does not appear to have any consumer facing business or consumer branding. They do not control the customer’s experience. They do not service loans. They do not appear to verify consumers or detect fraud (seems like the MPLs do that?)
-Upstart does have a consumer facing business, can control the customer experience, services the loans, and performs fraud/verification checks.
-Pagaya, from what I have found so far, does not seem to have mitigated regulatory risk in regards to fair/non-discriminatory AI “underwriting”. I guess they skirt this by dumping the regulatory risks solely on the MPLs, even though the MPLs originates ‘custom’ loans specifically for Pagaya.
-Upstart defends against potential regulatory adverse action with its CFPB No Action Letters and partnering with NAACP, SBPC and constant regulatory contact on the hill.
-Pagaya says they have “16 million+ trailing data points since inception” but they have evaluated “17 million applications in the last 12 months”. How does that work? Shouldn’t there be billions of data points?
-Pagaya is a competitor to Upstart, by assisting other MPLs to originate more volume that can take market share away.
-Pagaya is a direct competitor to Theorem (https://www.theoremlp.com/) who also uses data science/AI/ML to buy marketplace loans.
-I’ll follow this company closely. They’re growing fast. I don’t care for SPAC forecasts, those are not reliable. Once they’re public and have their first earnings report I’ll make a decision whether they would be a good stock to include in my portfolio.