But in a revolving product like card, we’re only able to reserve for the loss content related to the balances that are on the books at the end of the quarter as opposed to the projected loss content for the account. So getting to your question then, when we have elevated seasonal balances in the fourth quarter, we expect a portion of those balances to pay down very quickly. And therefore, those specific balances are likely to have very, very low loss content given the life of the balance is far shorter than the life of the account. So all else equal, the coverage ratio in the fourth quarter has a bit of natural downward pressure from that elevated denominator as you suggest. But looking ahead, there’s a bunch of factors that can impact where the allowance goes from here beyond that single effect.
In periods where future losses may increase we would replace the low loss content of the current quarter with the projected higher loss content in a future period. And for what it’s worth, those assumptions also then carry into that reversion period. As we have growth with seasonally adjusted balances, Rich mentioned this before, CECL significantly pulls forward that allowance cost of growth. And then the third factor is coming out of a period where we have unusually low losses like we’ve experienced, over the last couple of quarters, you have lower recoveries to offset the forecasted loss content. So, all of those things can put upward pressure on allowance but we can also have revisions to our economic assumptions, to delinquency flow rates, to just our overall loss content.
And so, there’s pressures in the other direction. And so, I appreciate your bearing with me for a long-winded complex answer, but I think we all saw the complexity and pro-cyclicality of CECL play out during the pandemic when we had to make a bunch of assumptions as the pandemic played out, we built a sizable allowance only to release virtually all of it over the subsequent quarters. And so, it’s just a very difficult thing to predict given all of the assumptions at play, which is why we are trying to focus you on Mako and having delinquencies as a leading edge indicator of Mako because that is ultimately where the real economic cost is felt.
Richard Shane: Got it. No, it’s a great answer, and I’m glad to bear with you. I’ll probably read it in the transcript about 7 more times.
Operator: Our next question comes from the line of Moshe Orenbuch with Credit Suisse.
Moshe Orenbuch: And I was hoping to talk a little bit about marketing. I mean, Rich, you did mention that you were primarily in card and primarily in the upscale customer. But could you just talk a little bit about, number one, what you might be doing kind of in nonprime and how we should think about whether that total marketing spend given what you see is likely to be higher in 23 or not?
Richard Fairbank: Yes. So Moshe, I’m glad you asked the question because I would not want the net impression to be — I think what you were saying is that, is our marketing primarily just in the upscale customer segment? Differentially relative to years ago, we’ve certainly had a shift to the higher end in terms of our marketing. The marketing is also so expensive at that end. And then, also the marketing at the higher end tends to, in some sense, lift the boats across the franchise. So, the marketing at the higher end is carrying a lot on its shoulders, Moshe. But, we do a lot of marketing in the mass market of the card business, including in the higher end of the subprime segment of the market. This is — our strategy here is — well, it changes all — we tweak it around the edges all the time.