da-kuk
Upstart (NASDAQ:UPST) has been faced with a tough business environment since interest rates began their steep rise in early 2022. Still, hope is on the horizon. The company has secured committed funding, substantially lowered its cost basis, and launched new products. Interest rates will also trend downward soon. Upstart’s shares have already played this scenario and appreciated substantially. While its short-term potential may now be limited, the company still presents an interesting long-term opportunity. I believe Upstart to still be a buy for now, but price matters at these levels and it might be a good strategy to wait for shares to take a moderate dip. Moreover and when looking at competition, Upstart’s execution has clearly been falling behind in the past year. One should therefore keep a close eye on if the company manages to turn the tide in the coming months.
Turbulent Operational Development
Upstart runs an AI lending marketplace in the United States and aims to be a technology provider connecting borrowers with lenders via its network and its AI loan models. On the funding side, Upstart works with bank partners and credit unions, sells asset-backed securities (ABS) on the capital market, and has lately entered into agreements with “committed capital” partners. While the business with ABS transactions worked well in periods of low interest rates and made up the dominant portion of Upstart’s funding, it struggled when rates started to increase. Upstart then experienced “severe constraints on the funding side of the business” due to a reduced inclination of the capital markets to take on subprime loans in times of increasing default probability. Part of the truth is however also that Upstart’s loans under-delivered relative to their expected return in 2021 and 2022 cohorts and are only now back to target.
Quarterly loan performance vs. expectations (Upstart Q3/2023 earnings presentation)
As a result of this funding constraint, management decided to take loans on the company’s balance sheet, which resulted in a strongly negative stock market reaction and harsh criticism as many investors had based their thesis on investing in a technology company without meaningful credit exposure (Upstart has always held “R&D loans” on its own balance sheet). Additionally, management aimed at securing committed funding that would be stable through cycles and, after some back and forth with its balance sheet lending, finally succeeded with this endeavour in Q1/2023. At this point, many believed that Upstart would return to hyper growth mode and deleverage its balance sheet, yet its business only bounced back slightly and keeps stagnating since – now being demand constraint and only able to approve few borrowers. This came as a surprise and the company’s share price reacted with a whip effect by first increasing about sixfold and then crashing again by almost 70%.
Why The Business Still Struggles
I was just as surprised as the market by the fact that Upstart did not manage to get back on track after securing committed funding and have since re-evaluated my investment thesis. Upstart’s management claims that the constraints are to a large part due to the legal rate cap at 36% where borrowers can no longer be serviced. Yet, while conversions are low, Upstart’s contribution margins are at a record level. That made sense at times of constrained funding, as firms should be expected to set their prices high when supply (with loans) is limited to maximize profit and margin.
But once sufficient funding is secured, serving more customers at a slight loss of margin would be rational if the additional gain from quantity outweighs the loss from (contribution) margin contraction. Looking at Upstart’s numbers and comparing them to historic figures, I believe that serving more consumers at a slightly lower contribution margin would clearly improve profits. This leads me to conclude that Upstart is also constrained from its funding side with partners requesting rents and thus rates which are too high to clear the market in a sensible manner. The reason for this could be increased risk aversion and a lower confidence in Upstart’s models leading to further risk adders. In fact, Upstart’s management stated in the Q3 call that some committed capital partners require higher rates while others like Upstart to co-invest own capital as a sign of confidence in the models.
At the same time, and just looking at the headline numbers of Upstart’s AI technology, the benefits of its offering should be more obvious, particularly to traditional lending partners (banks, credit unions, car dealers, etc.). With more than 100 banks and credit unions as well as 69 car dealerships (at Q3/23) as partners, I have some trouble understanding why this side of the business is not more successful at gaining market share. Upstart’s models provide more inclusive lending, clearly better risk separation, and a high degree of automation, which should be a great incentive for these partners to run more and more of their own credit decisions on Upstart models. Even more since Upstart’s models have strongly improved and credit performance is back on track.
The potential to save time, money, and provide a better product in parallel should in my eyes exert more allure for partners. It might be the case that I’m missing something here as Upstart does not talk too much about this topic in earnings releases, but it gives me a lot of trouble trying to understand why there is no increased penetration from this side of the business being the traditional channel for lending (I’d be happy to discuss this in the comment section if you have any thoughts).
Having said that, I also expected auto lending to pick up much more strongly and outpace any declines due to the difficult environment in personal lending. After all, auto lending is a much larger and more secured market and initial market share gains in this nascent stage should have already contributed more meaningfully to the top line. In the latest earnings presentation, Upstart however reported declines of 75% QoQ and 78% YoY from an already low level in lending volume, making it obvious that the situation here is even more difficult compared to personal loans. According to the CEO, Upstart currently faces similar difficulties in auto as the personal loan space did a few quarters ago. The aim is now to recalibrate and improve the models in order to adapt to the new level of risk and changed economic environment. Yet, a decline like this at a time when competitors (see below) are gaining market share is worrisome.
Business Model Vs. Competitors
As described above, there is a lot to be excited about when considering the benefits of AI-based lending. In my opinion, Upstart is right in saying that traditional personal lending is broken in the face of all kinds of discriminative biases, limited risk separation (due to limited analysis of customer data), higher defaults and risk adders, and often subjective decision taking, particularly in regional bank branches. Additionally, offline lending processes take of course much longer to close while Upstart is now at 88% automation rate with instant approval and expects this figure to stay high even when conversion rates pick back up. In an economic sense, AI-based lending delivers welfare improvement by strongly reducing transaction costs and asymmetry of information, benefitting all parties involved. As such, the case for AI lending solutions is strong and Upstart is a leading player in this field.
Risk separation of Upstart’s AI models (Upstart Q3/2023 earnings presentation)
So why is Upstart not executing more effectively? It seems like its partners are only adopting its AI models very slowly and exercise a lot of caution. Upstart therefore used to be very successful when investors on the capital markets were in desperate need for high-yielding assets and default rates were artificially depreciated by governmental aid. As soon as this changed, they quickly jumped ship and Upstart’s business was in trouble. Moreover, elevated rates and heightened risk lead to more unattractive offers to customers of which many cannot be served below the legal rate cap of 36%.
I’ve been uneasy with the company’s business model for a while now and felt that management should give a lot more consideration to if the current pricing scheme is still the right one, respectively how its competitors are faring with different approaches. For instance, Upstart’s pricing model is predominantly based on fees for loan origination, which makes its top and bottom line very vulnerable to the credit cycle. I understand that lending partners prefer such an arrangement in the introduction phase of a new technology and Upstart had to build some trust. Yet, Upstart now has a large and experienced partner network and higher shares of SaaS revenue could introduce more continuity into Upstart’s revenue stream. Such a revenue model also increases incentives for customers to further the usage of Upstart’s AI models and is a more fitting pricing scheme for a technology company.
Moreover, competitors like Pagaya (PGY) and SoFi (SOFI) follow different strategies and both have fared the turbulent times better than Upstart. SoFi for instance is a bank itself who is not depended on funding partners and has the freedom to conduct its lending business more autonomously. It is of course not a great comparable, but due to Upstart’s funding issues and corresponding discussions among investors on if becoming a bank would be suited to better the situation, I wanted to mention this. Pagaya’s business on the other hand is very similar to Upstart as it also uses AI models to address personal and auto lending, but with a focus on larger partners compared to Upstart. While Upstart’s management believed that large banks would develop own solutions in-house and focused instead on regional banks and credit unions, Pagaya is now in discussions with 80% of the top 25 US banks and has already signed a few including a top 4 bank in Q3/2023.
In the auto business, Pagaya is far more successful than Upstart by regularly delivering large partnerships and growing consistently. It also delivers very high growth (x6 Q1 on Q3 and x2 Q2 on Q3) with its point-of-sale solution and is partnered with the world’s largest buy-now-pay-later provider Klarna. Furthermore, Pagaya structures loans and serves as the largest ABS issuer in the US. According to Upstart’s CEO in the Q3 earnings call, this business is currently in high demand with banks and credit unions deleveraging balance sheets and seems to be countercyclical to lending. It thus might be a very nice way to stabilize revenues and expand the application of credit models which would have been a great thing for Upstart. At the moment, Pagaya’s revenues are growing YoY and it just raised its full year outlook.
The Latest Developments In Q3
Upstart’s numbers have lately lagged competitors’ metrics. Revenue from fees came in at $147 million, slightly below guidance of $150 million and down 14% YoY, yet marginally up from last quarter’s $144 million. Operating expenses were down YoY (but slightly up sequentially due to increased marketing and sales activities), enabling Upstart to achieve positive adjusted EBITDA for the second quarter in a row and contribution margins to remain at record highs. However, the R&D loan portfolio continued to create negative charge-off’s. Loans held on the balance sheet were down just over 7% to $776 million due to an ABS transaction completed early in the quarter. Of these loans, $447 million are held for R&D purposes (of which, 85% are auto). Upstart now holds an unrestricted cash position of $517 million and $425 million in net loan equity at fair value. For Q4/2023, the company expects all metrics to be approximately flat and an adjusted EBITDA of about zero.
Overview Q3 financials (Upstart Q3/2023 earnings presentation)
In the earnings call, management stated that Upstart’s ability to approve borrowers is constrained due to a macroeconomic environment of low consumer savings and high credit default rates. Its UMI index correspondingly remained at an elevated level, indicating that default risks are elevated and economic conditions are unfavourable for credit quality. Management subsequently attributed the slight revenue miss to more conservative underwriting standards, which also stem from increased risk from prime borrowers. This customer segment is experiencing adverse effects with some delay compared to riskier borrower categories and the newly developed models are able to track these distinctions better so that lending volume unexpectedly decreased vis-à-vis guidance. Still, the CEO believes Upstart’s personal lending models are now well calibrated with the largest model improvement to date and newly issued credit expected to perform in line now.
Having said that, Upstart experiences credit demand at a very high level. Due to the challenging environment and the low ability to approve loan inquiries, resulting conversion rates are low as well. Management attributed a large part of this low conversion to the 36% rate cut-off level. However, as rates go down, conversion is supposed to increase easily due to the fact that Upstart can offer loans to consumers that are currently over the cut-off line. Moreover, management does not expect the high automation level to inadvertently go down again when conversion increases, so that heightened contribution margins are sustainable – particularly due to the cost savings measures implemented in the past quarters (headcount reductions, software savings, etc.). The CFO, therefore, believes Upstart to be in a much better position compared to before the crisis in many tangible ways.
Management also admitted that funding markets are distracted by opportunities from the banking sector as banks aim to manage their balance sheets by unloading assets to the capital market. Upstart is therefore pursuing more committed capital partnerships to stabilize its funding side. They noted that there are generally two different categories of partners: some want to invest to earn a premium on their asset returns while others care about predictability. These funding partners care a lot about if Upstart’s predictions are reliable, i.e., if predicted loss estimates are accurate. They thus like Upstart to co-invest as a show of confidence, which Upstart has been doing since the beginning of the year. This increasingly exposes Upstart to credit risk on its balance sheet. Then again this can enable additional profits if loan performance continues to be above expectations.
Summary of Upstart’s committed capital co-investment (Upstart Q3/2023 earnings presentation)
In the earnings call, Upstart’s management furthermore conveyed a lot of excitement about the new HELOC product. The product is now available in the first markets and preliminary feedback seems to be good, particularly regarding the speed (5 days instead of a month) and convenience it introduces. It additionally diversifies Upstart’s business in terms of (1) it is a prime product with low default rates and (2) it is countercyclical to personal loans because HELOCs are more effective in a high-rate environment. Another big advantage according to management is customer acquisition since more than 80.000 homeowners apply for loans on Upstart per month and many of them would be better served with the lower rates of a HELOC. By integrating them, Upstart aims to become more customer centric instead of product centric.
Interest Rates Are About To Become Tailwinds
The most important catalyst for Upstart in the coming quarters will be the interest rate development which is why I would like to quickly lay out the current macroeconomic situation as well as the FED’s commentary: after acting too late (see below) on already hot economic indicators, the FED executed the steepest rate hike cycle of all times to tame inflation while energy prices around the world surged in the aftermath of Russia’s invasion of Ukraine. Even though the hike cycle has now presumably ended, the impacts of monetary policy are lagged and the true extent of this cycle’s effects will only hit the economy in a few months.
With an effective rate of beyond 5% and inflation at about 3%, the FED is in strongly restrictive territory. At the same time, the US job market has been cooling slightly and wage growth remains well in line. FED chair Powell therefore changed his wording substantially after the last FOMC press conference and believes thinking about cuts is now more substantiated than looking into further rate hikes, with the feeling of being on a good way while the battle on inflation is not yet won. New housing starts have lately recovered on already decreasing mortgage rates, however, predictions for 2024 still remain moderate with an expectedly weak job market.
While the US economy has cooled but does not appear to be suffering too much from high rates yet, other parts of the world seem to be coming down harder (China, Germany, etc.) which will impact US growth in 2024 too and very probably further drag down inflation. Financial markets are expecting the FED to start reducing rates with the first cut in March and more cuts to follow. The currently inverted yield curve (yield of long-term bonds below the yield of short-term bonds) is another sign that financial market expect interest to tend lower in the long-run (many investors see this as a sign of a coming recession, yet this may often only coincide as the FED reacts to a recession by lowering interest rates).
Valuation
With the first rate cuts in sight, Upstart’s share price has already started to appreciate meaningfully from the low 20s to high 40s and with it, the company’s valuation. The market is putting a lot of advanced hope into Upstart and its valuation has climbed to a level unsustainable if it does not soon regain its operating strength. Conversely, Upstart’s closest competitor, Pagaya, is not experiencing a similar recovery yet. The chart below shows how Pagaya’s price-to-sales multiple remains practically flat while Upstart almost doubled in the past few weeks and is now at more than 5 times Pagaya’s level. Yet, to be fair, Pagaya is not a well-known company and might just fly a bit under the radar.
In a historic perspective, Upstart has had far higher multiples, but I doubt that these levels are a realistic comparison. I therefore use discounted earnings to explore if the share price is sustainable in a scenario with declining interest rates and a recovery in Upstart’s business. In this scenario, I assume that the drop in top and bottom line starting 2022 is only transitory and a swift recovery is possible thanks to the initiatives the company has been setting in motion in the past quarters. Management at least seemed to believe so in the Q3 earnings call.
For a simple model of 5 years and at a discount rate of 12%, I can imagine Upstart to grow into its current valuation if it can return to about half of its pre-crisis 2021 earnings level within two years, then grow at 50%, 40% and 30% for the next three years. This seems not particularly aggressive considering Upstart’s growth rates in 2020/21 and with the organizational and technological progress it has made. For comparison, I further added a pessimistic scenario in which Upstart only recovers slowly and growth rates remain moderate. In this case, the latest share price appreciation would in fact not be sustainable.
| Pessimistic | Neutral | Positive | Optimistic | |
| Share of 2021-EPS in yr. 2 | 1/4 | 1/2 | 2/3 | 3/4 |
| Subsequent growth in % | 30-30-30 | 50-40-30 | 50-50-50 | 80-70-60 |
| Years to pre-crisis EPS | >5 | 4.2 | 3.0 | 2.3 |
| Target price | $19 | $46 | $92 | $150 |
In my opinion, once Upstart manages to get its business back on track, it can even outgrow its current valuation. I have therefore additionally included more favourable scenarios in my simple model which show that if Upstart would to some extent retrace its pre-crisis growth rates, it could return to triple digit share prices. For these two scenarios, I have also included a higher exit P/E multiple of 30 instead of 25. In my view, this is in the cards considering that management aims to enter and expand into further product sectors.
Conclusion
Upstart’s future prospects are difficult to ascertain at the moment. Its lending models have underperformed in the past quarters and while personal loan performance is back on track, its auto lending business is still heavily under water. The management team has already achieved a reduced cost base, committed funding, and technological progress and is pushing for new committed capital partnerships and the introduction of the new HELOC product. I see a lot of potential in this technology and believe that Upstart can turn things around if these initiatives bear fruit. With the upcoming interest rate cuts, Upstart’s operative position will improve for sure and the market seems to think so too by already lifting the stock higher. I therefore rate Upstart a buy, yet with a cautious eye on its increased valuation. If the stock experiences setbacks and Upstart manages to get back into strong growth mode, I recommend seizing the opportunity.
Having said that, Pagaya comparatively seems to execute better while being much cheaper and I will keep a close eye on how this trend continues in the next earnings releases. In the meantime, I hedged my bets and have sold part of my Upstart stake after the latest run up, since the position was getting too large in my portfolio and to extend my position in Pagaya. For the moment, Upstart’s shares are clearly performing better but I will keep an eye on if this trend inverts. If Upstart keeps being outpaced on execution, I will give deeper thought on downgrading the stock and to redistributing even more of my position to Pagaya.
