Should I pay off debt or invest? Start by reading your rates

Every payday, Ruth sends $150 to her credit card and $150 to the Roth IRA she opened two winters ago. The split felt right the evening she set it up: debt is the responsible thing and investing is the hopeful thing, so half and half honors both. It has run, untouched, for two years. She could not tell you the card’s interest rate without opening the app, and she has never once multiplied anything. Her arrangement is close to universal, because an even split is what fairness looks like, and fairness is the instinct most of us reach for when two good causes want the same dollar.

Here is the problem with that instinct, and it is the entire subject of this article. Fair is how you treat people. Rates are how you treat money. The card and the Roth are not two children owed equal affection; they are two numbers, one of them certain and one of them hoped for, and the numbers are rarely close enough for a split to be the right answer. The internet argues this question as a matter of identity, debt-free people against market people, each side with slogans. The arithmetic does not have a side. It has a ranking, your own statements already contain it, and the gap between reading it and not reading it compounds quietly, every month, for years. Ten minutes here is worth real money, and I will show the dollars rather than assert them.

The urge to split is so common it has been measured

I want to be fair to Ruth, because the research says she has company. In 2019 the American Economic Review published a study of how people with multiple credit cards allocate their repayments, built on linked card records rather than surveys. The economically correct move is boring and mechanical: send every spare dollar to the highest rate first. That is not what people do. In the paper’s own words: “Repayments are not allocated to the higher interest rate card, which would minimize the cost of borrowing.” What people actually do is match payments to balance sizes, splitting proportionally across their debts, a pattern the authors call balance matching, and it held so strongly that it explained more than half of the predictable variation in repayments. A companion study later found the same behavior in American credit bureau data. People split by shape and by feel, at national scale, and the cost of doing so never appears on any statement as a line item, which is exactly why the habit survives.

The regulators who watch this space describe two repayment strategies, and the difference between them is this whole question in miniature. The Consumer Financial Protection Bureau’s guidance on reducing debt describes the first one plainly: “This approach focuses on your debts like credit card and student loan debts with the highest rate of interest.” That is the rate-reading method. The second, often called the snowball, clears the smallest balance first for the lift of a finished account. The CFPB presents both without moralizing, and so will I: momentum is worth something real to the person who needs it. But only one of the two methods is free, and it is the one that reads the rate.

What debt actually charges

So read the rates. Here is the landscape this year, each number from the organization that measures it, and the shape of it surprises most people.

Two things stand out. The first is the gap at the top: the credit card row is not merely the biggest number, it is three times the size of everything under it. A carried card balance is a different kind of object from every other common debt, and any answer to the debt-or-invest question that does not treat it separately is not being straight with you.

The second is stranger, and it is specific to right now: the middle of the table has collapsed onto one spot. A new undergraduate loan at 6.52%, a new mortgage at 6.58, a new car loan at 6.97: as of this writing, the big ordinary debts all sit within half a point of each other, and they sit almost exactly on the boundary where this decision genuinely gets hard. The order every dollar should follow walks that boundary in full, but the short version is that somewhere around 6 to 7%, paying a debt down starts rivaling what a patient investor can honestly expect from broad markets over decades, with the difference that the debt’s rate is a certainty and the market’s is not. Below the line, cheap debt and investing can coexist for years. Above it, the sure thing wins. The 2026 landscape puts an unusual number of people directly on the line, which is why the honest answer this year is so often “read your own paperwork,” not a slogan in either direction.

That is also why the method outlasts the numbers in the table. It is a comparison, not a threshold handed down from anywhere, so it survives the rates moving in either direction. If mortgages fall back toward 3%, that debt drops well below the line, and carrying it while investing becomes the easy call rather than the arguable one. If they climb toward 9%, the same mortgage crosses to the other side and paying it down starts outbidding the market. Nothing about the reasoning changes in either world. Only which side of the line your own paperwork lands on does, which is why the rate on your statement outranks any figure I can publish.

One practical note on where to look: the number you want is the APR line of the statement, the debt’s cost written as a yearly percentage. For federal student loans there is a detail worth knowing that most pages skip, stated on Federal Student Aid’s own site: “A fixed rate will not change for the life of the loan.” Loans from different years carry permanently different prices, a 2021 loan and a 2026 loan are different animals, and a household’s student debt is often a bundle of vintages that deserve to be ranked individually rather than treated as one blob.

What the order is worth, in dollars

The ranking sounds abstract until you watch two orders of operations run side by side, so here is the arithmetic, start to finish. Take someone with a $5,200 card balance at the G.19 average of 22.15%, and $446 a month of genuinely spare money after the minimums and the bills. One version of this person kills the card first: every spare dollar goes to the balance until it dies, which takes 14 months, and then the full $446 flows into investing at an assumed 7% a year. The other version invests from day one: $350 a month into the market, with the remaining $96 holding the card’s interest at bay so the balance never grows and never shrinks. Same paychecks, same five years, nothing different but the order.

Three honest readings of that chart. First, the endpoint: the card-first path finishes about $4,000 ahead on identical paychecks, which is what the gap between a certain 22 and a hoped-for 7 costs when it runs for five years. Second, the beginning: for the first year the two lines are nearly on top of each other, which is precisely why nobody feels the error while making it. The cost of the wrong order is invisible at the moment of choosing and large only in accumulation, the least helpful possible schedule for learning a lesson. Third, the illusion, and I think this is the most useful thing here: at every point of those five years, the invest-first version has the bigger brokerage balance, $25,100 against $23,900 at the end. The account you check for encouragement looks better on the losing path. Only the net, investments minus what you owe, tells the truth, and the interest paid along the way seals it: the card-first path paid about $710 of card interest in total, while the invest-first path paid about $5,760 to keep its balance merely standing still.

That is the whole case, shown rather than asserted. And it is worth saying what the chart does not claim: at 22% nothing about it is close, but rerun the same five years against a 4% debt and the ranking flips, with investing ahead and the gap pointing the other way. The chart is not an argument for paying off debt. It is an argument for reading the rate before deciding, which is a different and better rule.

On the line, the tiebreakers are personal

The table above put most 2026 debt near the 6 to 7% boundary, where arithmetic alone stops settling the question. That band is where honest judgment lives, and a few things genuinely belong in it.

An employer match outranks everything, including the card. If your job matches retirement contributions, that is a 50 to 100% certain return on the matched dollars, a bigger number than any rate in the table, and it comes first no matter what else is true. What to do with your money first prices the whole hierarchy if you want it derived.

Certainty is worth something, and how much is a fact about you. Paying down a 6.5% debt earns exactly 6.5, guaranteed, every year, in every market. Investing the same dollars is expected to earn more across decades and will spend some years down sharply along the way. A person who would sleep through those years can lean toward investing at the line; a person who would lie awake, or sell in a bad stretch, collects more from the guarantee than the spreadsheet shows. Neither temperament is a flaw. Pretending you have the other one is.

Flexibility runs the opposite direction, and it is the tiebreaker people forget. A dollar of extra mortgage principal is committed: it lowers no monthly payment and comes back only when you sell or refinance. A dollar invested in an ordinary account stays reachable. Killing a card is different again, because it frees its whole minimum payment immediately, which is the fastest cash-flow improvement available to most households. Two footnotes complete the picture. Some interest, notably on student loans and mortgages, is deductible for the people who qualify, which quietly lowers that debt’s effective rate and weakens the case for rushing it; the qualifying rules are narrower than most people assume, though, so it is a footnote to check against your own return rather than a discount to assume you have. And none of this ranking starts until a small cash cushion exists, because spare dollars sent anywhere while a single surprise bill would land on the card have been aimed too early.

What this looks like on payday

Run it once and the system is set. Minimum payments on everything, always, since missing them costs more than any rate in the table above. Any employer match, collected in full. Then every spare dollar to whatever is left on your list above roughly 7%, highest rate first, until that list is empty. From there the honest answer really is both: cheap debt on its schedule, investing on its automatic one, in whatever proportion your temperament and the tiebreakers above suggest. The order never changes; only where you currently stand in it does.

Ruth, from the opening, ran her numbers on a Tuesday. The card was at 23%, a point above the national average, and the Roth’s hoped-for return was not within shouting distance of that. So the split ended: everything spare went to the card, which died seven months later, and then everything spare went to the Roth, which is a larger monthly contribution than she has ever made in her life. Total time spent deciding, after two years of not deciding: about ten minutes. The fairness instinct got no vote, and strangely, both of her good causes ended up better served for it.