Underwrite or Lose (Money)!

June 3, 2022

By Max Levchin

What follows is an expanded version of my thoughts on Affirm’s expected credit performance in a downturn from the most recent earnings report. I hope it sheds some additional light on our approach and how it is different from those of some of our competitors.

With inflation running at multi-decade highs, the Federal Reserve is hard at work trying to cool down the overheated economy while trying to avoid causing a prolonged recession. Whether the proverbial landing is hard or soft, it is reasonable to expect at least something of an economic downturn in the relatively near future as interest rates inexorably rise.

Consumer credit performance is always a concern for investors when times are tougher. How will Affirm fare in the months and quarters to come?

We’ve heard Affirm pattern-matched to credit card issuers, but that’s intellectually lazy, as there really are more differences than similarities. I’ll get a little detailed here, in the hopes of helping investors understand Affirm better.

Let’s start with credit cards, specifically credit cards as used by the revolvers – people who use them to finance things (vs transactors who generally do not carry a balance.)

Powering the plastic is a revolving line of credit (hence the name). Its issuer (let’s call them the bank – not always the case, but close enough) underwrites you pretty much once, when you apply. You borrow from this credit line with each transaction, up to your limit (which will generally go up as you spend more and pay your bills), and the only real rule is you have to make the minimum payment to the bank every month.

This works really well during economic booms. You get the stuff you want now and keep most of your cash to do whatever with, and the bank gets interest income as you build up a big balance by making only the minimum payments, which are easy so long as you have a steady job. It’s a win-win!

Of course, during an economic bust, you could lose your job or just decide to conserve cash. The bank, meanwhile, might really want you to start paying down your balance and might even cut your credit limit to reduce their exposure. (This can be a self-defeating move – you still owe your balance plus interest, and the card is less useful to you.)

Without a job, your ability to make minimum payments is hampered, and your balance could go unpaid for a really long time, which is a problem – for you because you’ll get hosed by the compounding interest – but also very much for the bank.

Which brings me to Affirm, which was conceived as something of an anti-credit card.

One of the foundational ideas behind our company is that revolving lines of credit – while unbelievably convenient – aren’t really good for most people. With cards, once you carry a balance, you pay interest on every single thing you buy – a couch or a cup of coffee. Calculating what purchases really cost you is near-impossible. In my opinion, this is the key reason behind many Americans’ bad relationship with debt. Credit cards make it easy to spend, hard to pay off your ever-compounding balance, while hiding their profit margin in the fine print and complicated math.

When we started Affirm, we wanted to build something truly transparent, where each purchase plan had a clear end-date, there were no surprises or gotchas, and most importantly, you knew exactly what each purchase would cost you. We believe that by making every transaction an explicit borrowing event, we don’t just protect Affirm from excess risk, we protect our borrowers from overextending themselves.

Every time you want to use Affirm to buy something, you have to ask (or apply, in credit lingo) to be approved for that specific transaction. Sure, we make it easy and convenient to ask, but we will still look at your credit situation at that very moment and decide — and if we believe you won’t be able to pay off your loan, we will, in fact, decline your application – with compassion and transparency – without fail.

We underwrite every single loan application using proprietary technology that we’ve been developing for a decade, using proprietary data we’ve been building up for nearly as long. Given the sheer scale of transactions we process - north of 10 million last quarter - the dials we turn to control risk are very fine-grained. Our process involves looking at credit report data, but could also involve some Affirm-specific stuff, like what we know about the merchant and the thing they are about to sell you.

It may be worth stressing that we do in fact underwrite every transaction. I highlight this because a few other BNPL providers (who mostly specialize in the 6-week loan variety) have been famously on the record about not underwriting their transactions at all. Their stated view has been that these low-value, ultra-short-term loans don’t really need underwriting. They approve more or less everyone, and maybe some people won’t pay them back, but they can write those losses off and treat them as a cost of borrower acquisition.

The trouble is, in a less healthy economy, this “cost of borrower acquisition” could go up quite a bit. Without at least a point of view on which of your would-be borrowers are more likely to become delinquent, your only real risk management tools are to either decline those already delinquent (which doesn’t help much) or to decline lots of first-time applicants to kill growth and brace for impact (which is what banks tend to do in a recession.)

OK, you might say, this is all well and good going forward. But what about all the loans that were made during the more benign credit environment? Our weighted average loan term is also really short, as in 5 months. That means that in the event of a downturn, much of our outstanding back book (the loans we held or sold) will have been paid off, and it’s that newly-tuned “front book” that will have the greater impact on our financial results.

So as you might expect – and I am definitely biased here – we are confident in our ability to deliver strong growth while driving positive credit outcomes consistent with maintaining attractive unit economics.

I should mention that during the very brief recession of 2020, we saw applications nearly quadruple at many of our merchants almost overnight. Looking ahead, we believe that the secular trend for paying over time without fees and gotchas will continue to strengthen across cycles.

It is our mission to improve people’s lives, and we fully intend to rise to the occasion and meet this demand — and we absolutely plan to maintain strong unit economics by only extending credit that we believe can and will be repaid. Hopefully, this gives you a pretty good sense of what one might expect from Affirm in a downturn.