Monday, March 29, 2021

Buy Now Pay Later Apps

If your financial life is going well, you’ve probably never used a Buy Now Pay Later (BNPL) app and may not have ever heard of them.  Here I look at what return they make on their money, and who pays them this return.

For a good explanation of BNPL apps, see Preet Banerjee’s video where he covers what they are and why you should avoid them.  In a typical case, if you are online buying a $100 item, you might encounter an offer to pay $25 now, and then $25 more in 2, 4, and 6 weeks.  From your point of view, this looks like an interest-free loan, but the BNPL company might only pay the retailer $94.  So, the BNPL company makes $6 over 6 weeks.

BNPL Returns

For this example, Preet calculates the BNPL company’s return as $6 on $94 invested over 6 weeks (42 days).  This works out to an annual uncompounded rate of (6/94)*(365/42) = 55%.

However, the BNPL company didn’t wait 6 weeks for the whole $100.  In fact, it got $25 of this money right away.  So, we could say that their investment was only $69 for 6 weeks.  This works out to an annual uncompounded rate of (6/69)*(365/42) = 76%.

But this still doesn’t account for the fact that the BNPL company got $25 in 2 weeks and another $25 in 4 weeks.  If we calculate the internal rate of return on investing $69 to get back $25 in 2, 4, and 6 weeks, it works out to 4.29% every 2 weeks.  Compounding this annually gives a whopping 199%!  This assumes that all returns get reinvested.

But what if we want to calculate the uncompounded return where the BNPL company makes a single investment and never reinvests returns?  The easiest way to think of this is to imagine the BNPL company entering into a series of $100 purchases where they invest $69 on each purchase to get back $25 in 2, 4, and 6 weeks. We will treat $2 from each payment as profit, and the remaining $23 as a return of the BNPL company’s working capital.

Suppose the BNPL company starts with $138 to invest.  They immediately invest $69 in one online purchase, leaving $69 of their capital uninvested.  Two weeks later, they get back $23 of their capital (and a $2 return), and they invest another $69 in a purchase, leaving them $23 of capital.  Two more weeks later, they get back $46 of their capital (and a $4 return), and they invest another $69 in another purchase (with no leftover capital).

From here on, every two weeks the BNPL company gets back $69 of their capital to invest in another purchase, and they get a $6 return.  They can continue this indefinitely having 3 purchases on the go at all times, all with just the initial $138 of working capital.  If we constrain the BNPL company to winding up all transactions by the end of the year, their total return on $138 is 26 payments of $2, $4, $6, $6, $6, …, $6, $6, $6, $4, $2.  This is a total return of $144, and an annual uncompounded return of 144/138 = 104%.

If we just focus on the steady-state condition where the BNPL company makes $6 every 2 weeks on a $138 investment, the uncompounded return is (6/138)*(365/14) = 113%.

So, which return value is correct?  I’d say the uncompounded return is 113%, and the compounded return is 199%.  Either way, this is well over the usury level of 60%.  However, it’s not at all clear who, if anyone, is being charged this interest.  The deal looks interest free to the consumer, and it looks like a 6% fee to the retailer.

Who Pays for BNPL Returns?

If we look at the retailer and customer as a single entity, they accept $94 from the BNPL company and return $100 spread over 6 weeks.  So the retailer and customer as a combined borrower are paying usurious interest rates in this example.  But it’s not easy to define how they are treated as individuals.

Ultimately, the high cost of the BNPL “service” has to be paid by some combination of consumers and retailers.  Who pays depends on how much the retailer is able to increase prices to cover this cost.  Sadly, it’s not just the consumers who use BNPL apps who pay this price.  When retailers raise prices, everyone pays more no matter how they pay.


  1. If I understand this correctly, the argument to not use BNPL is that while it appears to be an interest free loan, the net effect is higher prices for everyone, whether you use it or not.

    Should this same argument be applied to premium credit cards that offer higher rewards than basic cards? They charge retailers more when they are used. If everyone used them, it would lead to higher prices. If I'm the only one that doesn't use them, then my decision to use or not use the card doesn't substantially effect the price I pay, but if I don't play the game, I lose out on the benefits. Overall, I'd prefer the rewards and associated high fees weren't options. But since they are, I may as well maximize my personal value since I may end up paying the higher price if I don't, but without any of the value.

    I believe both of these cases are classic examples of tragedies of the commons. What appears like a win/win is really a lose/lose (with some middle man like credit card companies or a BNPL company winning at the expense of consumers and the retailers).

    Do you generally agree with this @Michael? Is there anything different that makes BNPL different than premium credit cards? From the sounds of it, the percentage gap between the benefit to me and the cost to the retailer sounds much higher. So the "tragedy" may be larger with BNPL. But if I don't use it, I may still pay for it anyways if others use it, right?

    Around Christmas time, I bought a somewhat pricey GPS/fitness watch as a gift from Amazon. It was sold by and fulfilled by Amazon. When I went to checkout, it offered a "free" BNPL scheme of 20% plus full taxes up front, then 20% of the non-tax amount per month for the next 4 months. While it would've made no impact on my decision to buy or not buy, I saw no reason not to take it, so I took the offer. I wonder if Amazon is using a BNPL company behind the scenes? If Amazon is effectively it's own BNPL provider (credit card statements show payments to Amazon on each payment, so perhaps?) then there probably isn't quite as much of a cost to Amazon, other than the amortized cost of rolling out the service, plus the cost of the deferred revenue. I bet Amazon's opportunity cost of the deferred revenue, based on their IROR, is definitely much better than my benefit of deferred payment. So perhaps the "value gap" is actually quite high, even with an internal BNPL scheme. The value to someone who doesn't pay off credit cards might be much higher however.

    1. Hi Returns Reaper,

      You'd have to look to Preet's video for why people shouldn't use BNPL offers. I think it had to do with nearly half of BNPL revenue coming from late fees. I don't know if this has to do with the financial troubles of BNPL customers or if it comes from some tricks BNPL uses to get people to miss payments.

      To the extent that I was making any statement, it is that either BNPL shouldn't exist for usury reasons. Failing that, those who use it should have to pay its costs and not everyone else. The same goes for premium credit cards. One could argue that handling cash has costs too, but they aren't as high as retailer costs from premium credit cards. Those who use these cards should pay a higher price at retailers.

      All this said, I don't blame individuals for serving their own interests. I don't blame those who use BNPL, and I don't blame those who use premium credit cards. However, retailers collectively need more power to fight back against credit card companies.

      For now BNPL seems to only exist in environments associated with shopping addiction, so BNPL is probably driving higher sales, and retailers are happy enough. However, if some BNPL company gets large enough to force their payment methods on retailers who don't want to accept it (the way Visa and Mastercard jam premium credit cards down retailers' throats), I'd like to see them stopped.

    2. I pretty much agree with everything you say Michael.

      I'd be happy if there was a reasonable way to prevent intermediaries from enticing consumers with some form of benefit only to pressure profits of retailers, resulting in higher prices for all consumers.

      If, in the case of BNPL, the benefit is truly desired by the consumer, the consumer should pay for that benefit. Some may choose to do that. But with premium credit cards I think very few people would pay for the rewards they offer.

      Credit cards themselves as a form of payment likely do provide value to the overall marketplace. But there is simply no reason for rewards to be tied to their use.

  2. Good catch, Michael. I'll add a note in the video description to link to this post. I should have factored in the upfront payment collected in my example.

    1. Hi Preet,

      I assumed your purpose was to demonstrate in the easiest way possible that the interest rate was very high. Your method achieved that goal. I wanted a more accurate figure to demonstrate that the interest rate was usurious.