I’m sorry everyone, about Affirm

Investment Analysis Clubs Saul’s Investing Discussions
1

I’m sorry everyone, about Affirm

I’ve tried three times now to take a small position in Affirm, but each time I just can’t get myself to hold it, and I exit in a few days. There’s something about it so that I can’t get all excited about it. It just doesn’t seem like what I’m looking for. Perhaps Pavlos explained it better than I have (and I quote, somewhat shortened and a tiny bit paraphrased by me):

Upstart’s only job is to assess the risk. Upstart doesn’t need cash to underwrite (approve/disapprove) the loan. All it needs is the technology and the personnel to keep feeding their AI/ML models to learn and improve…

Affirm’s job is to take the risk. Affirm says yes or no to clients who wish to use the BNPL feature and charges the merchant a % of the sale right at the spot. And then makes back a portion of the interest earned through the loan given to the end consumer. But in reality, the client doesn’t take on the loan. Affirm does that. That’s why Affirm needs to pay the bank a fee for servicing the loan as well as a % when originating the loan in the first place. So, you could say that Affirm is like the customer of Upstart who needs access to credit. Affirm then goes on and pays the merchant the amount of the sale minus its % fee as explained….

So, it’s like opening a business and having someone (Affirm) standing outside telling people come on in buy this for free today and you can pay me (Affirm, not the actual business owner) later with a small interest. This person then goes around forming good relationships with other business owners telling them that I can bring in a lot more customers just by allowing me to take on the risk. You as the merchant will get your money no matter what. Even if the client doesn’t pay me back. I’ll cover that too!..

Thus, Affirm is first and foremost a company that makes money from loans, followed closely by the money they are able to extract from merchants from both transactions and interchange fees. Upstart is solely a technology company.

So why in the world would I want to invest in Affirm, really? Especially if I can invest in Upstart instead.

Here’s another way Pavlos looked at it

I like to go over the social media (especially Instagram) of the companies I follow to get a sense of what kind of customer they are looking for. Upstart’s message is that it wants to help people get their finances right by giving out financial self-improvement tips regularly. You can see lots of examples here https://www.instagram.com/upstart/

…Upstart speaks to the logical part of humans. How to be better financially, save on your interest rates, create a budget, assess your income and spending, make your home efficient, etc.

Affirm, on the other hand, https://www.instagram.com/affirm/ tells a different story. The theme here is a bit more celebrating, colorful, and shows lots of people taking trips, adventures, enjoying new clothes and buying stuff with the motto of buy now and pay over time. I guess the whole point of Affirm is to help people buy what they want now and split the cost along the way, which for a lot of people, may not be doing them a favor….

I hope that what Pavlos wrote helps explain my decision about Affirm.

Best,

Saul

143 Likes
2

From a month ago, and yet $AFRM has continued to rock n’ roll. Truly, Caveat Emptor here:

https://www.reuters.com/technology/buy-now-pay-later-surges-…

Sept 9 (Reuters) - A third of U.S. consumers who used “buy now, pay later” services have fallen behind on one or more payments, and 72% of those said their credit score declined, a new study published by personal finance company Credit Karma showed.

The study, conducted by software firm Qualtrics, surveyed 1,044 adult consumers in the United States last month to measure their interest in buy now pay later (BNPL) and found 44% had used these services before.

The usage figure was slightly up from a similar survey conducted by Credit Karma for Reuters in December, while missed payments was down from 38%.

15 Likes
3

PeregrineTrader, thanks for the article but there was one basic question that I can’t wrap my head around BNPL in general:

What’s stopping a merchant from accepting multiple BNPL services?

If BNPL is in any way similar to credit cards, then merchants would want to accept as many payment methods as possible. I have never been to a store that only accepts Visa and rejects MasterCard. What’s stopping Walmart and Target from adding Afterpay or Klarna besides on top of Affirm?

There are lots of e commerce sites already taking multiple BNPL - shein.com takes Afterpay, Klarna, and Zip, all three of them are BNPL with 4 installments.

If the price movements on Affirm recently was driven by news of onboarding large retailers (Amazon, Walmart, Target) then it seems like a short term price action to me and not a long term business advantage.

I am by no means a finance person and I really appreciate any input on this. Thank you.

23 Likes
4

Investi Analyst elaborated this in his following seeking Alpha article which I believe makes a lot sense:
https://seekingalpha.com/article/4452738-affirm-holdings-moa…

Convenience and Flexibility are key: Most consumers would prefer to have only 1 or perhaps 2 max for their BNPL provider when they go shopping. Nobody wants multiple BNPL solutions at checkout, so this could have been one of the reasons why Amazon chose to do a partnership since Affirm already has a large number of consumer and merchant reach.

Consumers’ trust is a key to success in BNPL space.

Zoro

8 Likes
5

In addition to what Saul and Pavlos posted (thank you both for sharing your thoughts and insights on this because I’ve been wrestling with whether to purchase Affirm in a similar way), the following information posted by PeregrineTrader is really why I have so far shied away.

[PeregrineTrader’s post]
Sept 9 (Reuters) - A third of U.S. consumers who used “buy now, pay later” services have fallen behind on one or more payments, and 72% of those said their credit score declined, a new study published by personal finance company Credit Karma showed.

[Me here]
If 1/3 of BNPL users have fallen behind on payments, what a risk! Also, even more troubling, is 72% of BNPL users say their credit scores declined! What does that say about the feasibility for long-term adoption of this service if it actually hampers a person’s chances to get credit? I’m not sure, but it does not feel positive to me.

Although as an Upstart investor, to look at it from a different perspective, I guess the decline in credit scores that appears to be correlated with BNPL use could be a potential increase to Upstart’s TAM in future quarters :slight_smile:

6 Likes
6

You have raised some very good points regarding responsible consumption. Affirm is not a buy yet as its profitability is unproven.

However, the point which their CEO is labouring, again and again, is that they are competing against credit cards. The first source of irresponsible consumption is a credit card because it clubs all expenses from a cup of coffee (regular expense) to a treadmill (occasional purchase) all in one account and payable after few days.

Credit cards are good for regular expenses but not so good for occasional high ticket items. How great it will be if I get an option to plan the payment for my occasional big-ticket purchase and not exhaust my credit limit.

Afrm/ or some other BNPL can be a card issuer like Mastercard/ Visa. Also, BNPL players can get extra discounts/exclusive deals from sellers for their cardholders.

Compulsive shoppers are already exhausting their credit cards and then going for personal loans to streamline their credit card debts.

Just a few thoughts. No brickbats, please.

7 Likes
7

Affirm reminds me of Wimpy, from the cartoon Popeye,
“I’d Gladly Pay You Tuesday For A Hamburger Today”

2 Likes
8

Let me make clear, that for anyone invested in Affirm, I hope that you make a million on your investment. It just wasn’t for me. I don’t wish the company any harm at all. There are just some companies that don’t click for each of us, and that was one for me.
Saul

31 Likes
9

The way I see it is BNPL companies, like AFRM, play on the big interest rate differential charged by credit cards by offering a lower cost solution. Their market is retail customers who leave a balance on their credit card. UPST aims to refinance larger ticket items at better terms. Different markets and different risk profiles. Just because BNPL firms take more risk it does not mean their model is not viable. Tell this to Master card and Visa.

5 Likes
10

Upstart’s only job is to assess the risk. Upstart doesn’t need cash to underwrite (approve/disapprove) the loan. All it needs is the technology and the personnel to keep feeding their AI/ML models to learn and improve…

Affirm’s job is to take the risk.

This is the difference between Visa/Mastercard and Discover/Amex models. Visa/MC is asset light and better. However DFS and Amex work satisfactorily.

You do get paid interest for assuming the risk. It is like insurance. Your success depends on how well you price your premiums. Affirm may use Upwork if they don’t have the right AI models.
I am thinking they will build and tune their own models.

I invested in DFS 15 years ago and it has worked out well.

2 Likes
11

Sept 9 (Reuters) - A third of U.S. consumers who used “buy now, pay later” services have fallen behind on one or more payments, and 72% of those said their credit score declined, a new study published by personal finance company Credit Karma showed.

Afrm has not charged a single late fee as of yet. They have a lot of data on their clients and their CEO Max Levchin was a co-founder of PayPal and was an investor in Yelp. He is part of the PayPal mafia. Watch both of the video’s.

https://www.cnbc.com/2021/01/05/affirm-ceo-max-levchin-to-be…

Andy

4 Likes
12

A quick note on credit card companies, from my recent experience with Amex. I am a high net worth individual, and we pay our credit card balances every month without fail. Amex has, however, just introduced a new product (in Canada, at least) that allows for monthly instalments. You simply click on a button once your statement has been issued, and you can choose between 3, 6 and 12 months. The first option comes at ZERO cost. They just break up your payment. Too easy. The second and third come with low cost monthly fees. I’m not interested in those, but I’m guessing lots of people will be.

I find it hard to imagine the other cc’s not jumping on board.

Saul raised a good question that I haven’t found an answer to. Where will they get all the capital needed for the expected growth? Am I correct that they now have over $2b in outstanding loans, up from $100m a year ago?

Anyhow, I’m out for now. Too much uncertainty for me, and if I’m honest, I hadn’t done enough dd to be anywhere near this company. Following along from the sidelines.

6 Likes
13

The one question I have was why would I use affirm when I can get cash back or points for flights when I use my credit card? My guess is the buyers have max out their credit card limit. Not the only reason but a possible reason. I personally used affirm because my credit card was stolen and I needed to make a purchase and affirm was available and quick.

I also like to look at company mission statement.

Upstart - “Our mission is to enable effortless credit based on true risk.” Short, concise, and truthful.

Affirm - “Our mission has been to deliver honest financial products that improve lives” It goes on to say No fees, No Gotchas, No Surprises" Really no fees? It reminds me of a saying when it sounds too good to be true. It probably isn’t. There’s no free lunch. It caused somebody something.

Thank you Saul for your wisdom.

Lily

5 Likes
14

As someone who used affirm to purchase a pelaton for the wife the reason why I did it was bc it had no interest fees. So I could space the payment out two years and in theory use that money to either generate some returns that compound (savings, stocks, etc). I could have put in on a card just as easily but then I have to pay it back in a month, which is fine but I lose out on the time value of money. Also the points generated are always pennies on the dollar so not a huge motivation. I mean yes if your gonna buy something anyways might as well earn a few pennies back but it’s nothing close to the returns we’d expect to see here after two years.

7 Likes
15

Sry forgot to mention I have no positions long or short in affirm

16

I am long Afrm and I think the next two quarters they are going to kill it. They have been taken in by Shopify, Walmart, Amazon, and Target. That is really amazing. Yes there are other BNPL options involved too but Afrm is right there with them. They have Ai/ML driving their proprietary risk model and they approve 20 percent more loans than any of their competitors. All of this with less risk.

Here is the Delinquency rate of all loans from all commercial banks given by the St. Louis Fed.

 Graph and download economic data for Delinquency Rate on All Loans, All Commercial Banks (DRALACBN) from Q1 1985 to Q2 2021 about delinquencies, commercial, ...
Q4 2020: 1.64 Q3 2020: 1.58 Q2 2020: 1.50 Q2 2021: 1.32

https://fred.stlouisfed.org/series/DRALACBN

Here is the delinquency rate of all commercial loans by commercial banks according to the St. Louis Fed.

 Graph and download economic data for Delinquency Rate on Consumer Loans, All Commercial Banks (DRCLACBS) from Q1 1987 to Q2 2021 about delinquencies, ...
Q1 2021: 1.70 Q2 2020: 2.03 Q2 2021: 1.56 Q3 2020: 1.83

Here is the delinquency Rate on all Credit Card loans.

 
Delinquency Rate on Credit Card Loans, All Commercial Banks
[https://fred.stlouisfed.org](https://fred.stlouisfed.org) › series › DRCCLACBS

Graph and download economic data for Delinquency Rate on Credit Card Loans, All Commercial Banks (DRCCLACBS) from Q1 1991 to Q2 2021 about credit cards, ...
Q3 2020: 2.02 Q4 2020: 2.12 Q2 2020: 2.44 Q2 2021: 1.58

Here is Afrm’s Delinquency rate taken from their S1.

Our continuously-learning risk model benefits from increasing scale. As data from new transactions are incorporated into our risk algorithms, we are able to more effectively assess a given credit profile. Our model is robust enough to allow us to assess credit risk at a pre-defined risk level with a high degree of confidence, resulting in a weighted-average quarterly delinquency rate of approximately 1.1% for the thirty-six months ended September 30, 2020.

That is pretty amazing to have a delinquency rate of 1.1 percent and approving 20 percent more clients. I think this proves out their AI model but is something that will have to be watched closely.

Andy

41 Likes
17

It is far, far, FAR too soon to be talking with any degree of certainty about affirm and its delinquency rates. It is assuming the credit risk of what is likely to be a portfolio of loans that have a much higher concentration of high risk borrowers than that of a typical bank. But we really don’t know, and won’t know for a year, at least. It seems to me that if the portfolio does turn out to be problematic, it will be a fairly sudden revelation, and affirm will almost certainly not have the capital cushion that a bank is required to have. So, for punters probably a couple of quarters of flying high, and after that…?

7 Likes
18

It is far, far, FAR too soon to be talking with any degree of certainty about affirm and its delinquency rates.

I am not talking certainty, its a data point. I don’t think you can have any degree of certainty about delinquency rates with any bank or credit cards as 2009 showed us. But what we can do is follow it and see how it does compared to the banks and credit cards. As of this time, I think you would agree, they are killing it against the banks and credit card companies. What the future holds is unknowable but then I am not betting on the future but on each quarter as I watch this one very closely.

Andy

9 Likes
19

‘‘It is assuming the credit risk of what is likely to be a portfolio of loans that have a much higher concentration of high risk borrowers than that of a typical bank. But we really don’t know, and won’t know for a year, at least.’’

I think that one of the features of using an AI model, is that it is continuously learning/training, and the model should be constantly adapting to new input. Every “loan” they make should provide new data to train the model further. That is (if I understand correctly), the model should get smarter every day.

So if a problem with delinquencies does develop, it should be detected very early and the model should adjust on the spot. At least this is my understanding.

5 Likes
20

It is far, far, FAR too soon to be talking with any degree of certainty about affirm and its delinquency rates.

I’m not sure I understand this statement. Perhaps this means we don’t know how delinquency rate will trend in the long run, which is true because AFRM has a limited history and of course we cannot predict the future loan performance for any company.

But for now, AFRM has an active loan portfolio and they report delinquency and chargeoff amounts, so we can measure these as of their last reporting period. Per the Fed’s method of measurement, a loan is delinquent after 30 days of nonpayment. AFRM charges off delinquent loans after 120 days. So, delinquency rates start to be known after 30 days of nonpayment, which is not a long time.

As of June 30, 2021:


**AFRM		Credit Cards - All Banks NSA (via Fed)**
Delinquency Rate	1.78%		1.48%
Chargeoff Rate		2.53%		2.54%

Chargeoff rate can be impacted by the age makeup and term structure of the loans on the balance sheet.

New loans have low chargeoff rates by definition because there has not been enough time for the new loans to be delinquent for 120 days and then be charged off. So, a growing company like AFRM may have young loans at its early stages of loan growth (when loan growth accelerates) and thus lower chargeoff rates because of this age effect. With steady or decelerating loan growth, this age effect reverses as the average loan age increases and these older loans, by definition, have higher charge-off rates.

These aging dynamics are also impacted by the term (duration) of the loans that are originated. Chargeoff rates for a loan cohort can be measured faster for shorter duration loans (we don’t need to wait 10 years to measure the final outcome of a 6-month or a 1-year loan, for example).

Most of AFRM’s loans are originated in the last 12 months, so these aging impacts should not be too large, but certainly should be kept in mind as the loan portfolio grows and/or the term structure of the loans they originate lengthens or shortens.

For AFRM, loan portfolio size as of June 30, 2021 (thousands)


**Fiscal Year Originated	Amount		% of Total**
2021			$1,878,866	93%
2020			$131,075	6.5%
2019 and older		$9,943		0.5%
Total			$2,019,884	100%

(fiscal year is Jul 1 to Jun 30)

AFRM
(https://www.sec.gov/edgar/browse/?CIK=1820953&owner=excl…)
Delinquency Rate (30d+), p. 132 ( 1.78% = (17,267 + 12,044 + 6,759)/2,019,884 )
Net Chargeoff Rate, total history ( 2.53% = 51,215/2,019,884 ); loan portfolio size p. 131

Federal Reserve
delinquency (https://www.federalreserve.gov/releases/chargeoff/delallnsa…)
chargeoff (https://www.federalreserve.gov/releases/chargeoff/chgallnsa…)

22 Likes