The Tip Is the Interest Rate

Zero Percent, and Then a Screen Asking for a Tip

I want to be honest that I thought these were fine. An app fronts you $100 four days before payday, takes it back automatically when your check lands, and charges no interest. Compared to a $35 overdraft fee that is not a close call. That was my entire analysis, and I held it for a long time without examining it.

The thing I had not done was multiply.

The Consumer Financial Protection Bureau published a data spotlight on this market and calculated a 109.5 percent illustrative APR on a typical employer-partnered advance. Not 10 percent. Not 19 percent. A hundred and nine. The product that advertises no interest carries an effective rate roughly four times a bad credit card.

Why a Small Fee Becomes an Enormous Rate

An APR is a fee expressed as an annual rate, and annualizing anything over a very short window makes it look absurd, because it is absurd.

Take $100 borrowed for eight days. A $3.18 expedite fee, which is the average the CFPB found across the market, is 3.18 percent of the amount. Eight days is roughly one forty-fifth of a year. Multiply and you are at about 145 percent annualized. The fee stayed small. The clock stayed short. The rate is a function of the ratio between them.

This is the same arithmetic that makes payday lending indefensible, and it is the reason payday lenders spent two decades arguing that APR is a misleading way to describe a two-week product. They lost that argument, mostly. The apps are running it again with better design.

I keep coming back to how well the interface hides this. Nothing on the screen is a lie. There genuinely is no interest. It is just that interest was never where the money was.

The Tip Is Not Optional in Any Way That Counts

Here is the mechanism I find most impressive, in the way a very good magic trick is impressive.

The tip is presented as voluntary and framed as supporting the service. It is preselected. Declining it takes an extra tap, sometimes buried behind a smaller link, occasionally attached to a note about how tips help the company serve people like you. California's regulator looked at this and found that 73 percent of transactions that solicited a tip received one, and described multiple strategies lenders use to make tips almost as certain as required fees.

Almost as certain as required fees. That is a regulator's phrasing and it is doing a lot of work.

The District of Columbia sued EarnIn over exactly this, alleging the real cost averaged an APR over 300 percent. On the specific example in the filing, a $100 advance disclosed at zero percent carried an $11 tip and a $4 expedite fee, which works out to a true APR of 498 percent. Every element of that was voluntary. All of it was disclosed. The disclosure is what makes it legal and the design is what makes it near-universal, and those two facts sit together comfortably.

The FTC also sued Dave in November 2024, alleging the company misrepresented advance amounts, charged undisclosed fees, and collected tips without consent. The Justice Department took the case over later.

The Number That Changed My Mind Was 27

A single advance with a $3 fee is a rounding error. I would have defended it, and honestly I still would in isolation.

The CFPB found the average user took 27 advances in a single year. Roughly 82 percent of transactions carried a fee. Average total fees came to $106 a year.

Twenty-seven advances is not an emergency. Twenty-seven is a subscription. It means the paycheck is arriving four days late every single cycle, permanently, because the advance is repaid out of the check that was supposed to cover the period the advance was covering. You borrow $100 on Tuesday, payday comes Friday, $100 leaves immediately, and now Friday's check is $100 short. So you borrow again.

That is not a criticism of anyone in it. The math of a tight budget makes the first advance completely rational. It is just that the product does not solve the gap, it relocates it, and then charges $106 a year to keep relocating it. Nothing in the app ever displays that annual figure, which strikes me as the single most revealing design decision in the entire category.

What the Rules Actually Say Right Now

This is the part where I expected to find clear regulation and did not.

The CFPB proposed an interpretive rule taking the position that these products are credit and that fees and tips must be disclosed as finance charges. Consumer advocates supported it on the straightforward grounds that money advanced and repaid with a fee is a loan regardless of what the marketing calls it. Then a Regulation Z rule published on December 23, 2025 addressed the non-application of Truth in Lending to earned wage access products, which moved the ground in the other direction.

I am not going to pretend the regulatory picture is settled, because it is not. What I will say is that the fight is entirely about whether these count as loans, and the fact that anyone has to litigate that question tells you which way the economics point.

Some state regulators have not waited. Several have moved to treat the products as lending and impose rate caps, which is why availability and terms differ noticeably by state.

What I'd Tell Myself at Nineteen

Not never use one. That advice is useless to someone whose rent clears on the 30th and whose paycheck lands on the 1st, and pretending otherwise is how personal finance writing loses people.

What I would say is: skip the expedite fee. The standard transfer is free and takes one to three business days. Most of the cost in this product is people paying $3.18 to receive money hours earlier, and if you can plan even two days ahead, the entire fee disappears. That one change does more than any other.

Decline the tip. It is genuinely optional, the company is not a person, and 73 percent of people not declining it is a design outcome rather than a moral one.

And count. Open the app and count how many advances you took in the last twelve months. If the number is over five, the product has stopped being an emergency tool and become a structural part of your budget at somewhere around 100 percent annualized, and the actual problem is a payday timing gap that a $500 buffer would close permanently. Building that buffer from a standing start is its own difficult project, and the honest version of how to do it is worth more than anything an app can front you.

I wish someone had made me multiply. The multiplication is the whole story.