OT: Upstart (UPST) - Revised Expectations

DISCLOSURE: I own a tiny bit of Upstart. However, I no longer actively invest, so call
it a hobby. I’m using money I can afford to lose. A 100% permanent loss here wouldn’t
affect my retirement or other financial goals one bit. In other words, heads I win,
tails I scrape my elbow, but don’t get killed.

Note: the approach below is one suggested by Michael Mauboussin and Alfred Rappaport.
Any shortcomings are mine, not theirs.

Early this January I took a first swing at identifying what the market was expecting for
Upstart’s future performance. I created a model using the share price at the time and a
few performance drivers as inputs. Running several scenarios through the model gave me a
sense for what the market was expecting from Upstart, as implied by the share price.

Here’s the link: https://discussion.fool.com/ot-upstart-upst-part-1-expectations-…

A whole lot has happened with Upstart in the seven weeks since then. They released Q4 and
2021 year-end results, held their Q4 2021 Earnings Call, announced a share buyback, and
posted an intriguing Performance Award for Paul Gu. Several large shareholders reported
their stakes. And the price went up 16%.

With all this new information at hand, it’s time to update the model. As a reminder, the
model always is trying to answer these three questions:

  1. How much cash can Upstart’s business generate for its owners?
  2. When will they get that cash?
  3. How certain are we that they’ll actually get that cash?

Here’s the data I used to update the model:

Sales FY2021                                 $848.6M  <= From 10-K
Estimated Sales FY2022                     $1,400.0M  <= From guidance

Operating Margin FY2021                        16.6%  <= From 10-K
Estimated Operating Margin FY2022              12.2%  <= My estimate, based on guidance
Estimated Operating Margin after FY2022        15.0%  <= My estimate, based on Conf Call

Working capital investment rate last 5 years     10%  <= Calculated from 10-K's
Fixed capital investment rate last 5 years        6%  <= Ditto

Tax rate                                       21.0%  <= Rate in five years after NOLs exhausted
Inflation rate       7% 2022, 5% 2023, 2% thereafter  <= My estimate

Share price 2/25/2022                        $135.88  <= From Yahoo
Diluted shares 12/31/2021                      94.8M  <= From 10-K
Market Value of Equity                        $12.9B  <= Price * shares
Cash and Securities                            $937M  <= From 10-K
Debt                                           $803M  <= Ditto

Two key drivers are missing from this list: sales growth rate and cost of capital. Of
course, these are highly speculative. Rather than try to guess them, I look at a range
of values for them. Each set of values will tell me something different about investor
expectations. If I run enough scenarios, I’ll get a feel for what the market is expecting.

Here are a few examples:

Example A.
I think Upstart’s cash flows are unusually risky so I peg cost of capital at 20%. The
implied forecast period – when I’ll recoup my $138 share price – is 9 years with
a 50% annualized sales growth rate and 14 years with a 25% annualized sales growth rate.

Example B.
I think Upstart’s cash flows are moderately risky so I peg cost of capital at 15%. The
implied forecast period – when I’ll recoup my $138 share price – is 8 years with
a 50% annualized sales growth rate and 13 years with a 25% annualized sales growth rate.

Example C.
I think Upstart’s cash flows are run-of-the-mill risky so I peg cost of capital at 10%.
The implied forecast period – when I’ll recoup my $138 share price – is 6 years with
a 50% annualized sales growth rate and 11 years with a 25% annualized sales growth rate.

In all examples, I look to see if the implied growth rates look reasonable, given the
rate I picked for cost of capital. Instead of changing the cost of capital, I can keep
that fixed, and look at different sales growth rates. Or I can change the estimated profit
margins and see what effect that has on forecast period.

Please keep in mind that when I do this I’m not picking numbers out of thin air! The
numbers I’m using as inputs to the different scenarios are based on my assumptions about
the different drivers of the business. The inputs to the model are rooted in what I’ve
known previously and learned lately about Upstart’s market, strategy, operations, and
competitive landscape.

Morningstar has upgraded their fair value estimate for Upstart to $264. They maintain
their “extreme uncertainty” designation for this estimate and also add that Upstart has
a narrow moat. To me this means high cost of capital – in the 25% range. Given that
Morningstar’s target is $264, that translates to an implied forecast period of 9 years
with an annualized growth rate of 50% and 15 years with an annualized growth rate of 25%.
Quite the hurdle.

I mentioned earlier that Upstart recently posted a Performance Award agreement for
Paul Gu. It’s a remarkable document. In order to get the award, Mr. Gu has to be employed
at Upstart through January 1, 2029. After that date, he receives his award. But only if
specified stock price targets are met for 60 days or more. The targets range from a low
of $166 to a high of $616. The higher the price, the more units he gets and the more
he gets vested.

Let’s run a few of these prices through a couple of tries in our model. Let’s fix the
cost of capital at 10% for the first few tries. At a growth rate of 50%, Upstart’s value
gets to $166 by 2027. It gets to $616 by 2032. If we increase the growth rate to 75%,
the value gets to $166 by 2026 and to $616 by 2028. How reasonable an expectation is it
to assume Upstart will grow 75% per year for the next 6 years?

To repeat what I said in an earlier post, we are not building the James Webb Telescope
here. Our goal is NOT to precisely predict growth rates or profit margins or rate of
investment or cost of capital or the value of a share. All were doing is running some
scenarios based on what know and learn about the business to see which ones of them look
more achievable than others.

One other point. Just like Upstart sharpens its AI model the more data and experience
it gets, we do the same with this model. Every quarter that passes gives us more data
about Upstart’s drivers that we can feed to the model, and the more experience we get
with the model, the better we can interpret results.

Best to you,

P.S., Here’s another of the many things that fascinate me about Upstart. I mentioned
earlier that several shareholders reported their stakes in Upstart recently. One of
these was hedge fund manager Dan Loeb’s Third Point. At the end of September 2021, his
fund owned 12.4 million shares worth about $3.9 billion. In the next three months he
sold 8.4 million of these shares. That was roughly 12% of the float. It’s no wonder the
price cratered as it did. At the end of December, his fund owned 4 million shares worth
$605 million. His latest filing two weeks ago shows he unloaded these remaining 4 million
shares in January and early February.