How student loans work in the United States

A student loan is money borrowed to pay for education, charging interest, repaid in instalments once a defined period has passed. The mechanism is ordinary lending. What makes it behave unlike other consumer debt is the gap at the front: the money is borrowed years before the first payment is due, so interest has time to build before anyone asks the borrower for anything.

Everything below describes the United States, where a student loan is an individual debt with a stated principal, a stated rate and a repayment plan the borrower selects. That scope is not a formality. England, Wales and Australia run income-contingent systems collected through the payroll or tax system, where what comes out each year is a percentage of income above a threshold rather than an amount set by the balance. They are not interchangeable with each other either: the England and Wales schemes charge interest and write off whatever is outstanding after a set number of years, while an Australian HELP debt carries no interest, is indexed to a price or wage measure instead, and has no write-off clock attached to it. These are different instruments, and reading the mechanics of one onto another produces wrong answers.

Three things separate a United States student loan from a mortgage:

• What you owe can exceed what reached the school. Federal loans deduct an origination fee from each disbursement, and principal is measured before that deduction.
• Each year of borrowing is normally its own loan, with its own rate and disbursement dates, so a graduate holds several loans rather than one.
• The rate on a federal loan is not priced on your credit. It is set by a statutory formula in the year the loan is made and then fixed for the life of that loan. Private student loans are priced on credit and can be variable.

Subsidised is spelled subsidized in American English. Same thing.

The two words settle one question: who pays the interest during the periods when the borrower is not paying it.

On an unsubsidised loan, interest accrues from the day the money is disbursed and nobody covers it. While the student is enrolled, through any grace period afterwards, and through most postponements, that interest accumulates, and the borrower owns every day of it.

On a subsidised loan, the interest accruing during qualifying periods is paid by the federal government as it accrues. Nothing accumulates across those periods, so where every period before repayment qualifies, the balance entering repayment equals the principal borrowed. Time spent in a period that does not qualify accrues like any unsubsidised loan.

Read that carefully, because the subsidy is not a discount on the rate. Both loans can carry the same stated rate and produce the same schedule once repayment begins. The whole value of the subsidy is delivered before the first payment, and it shows up in the opening balance rather than anywhere in the payment formula. Two students who borrow the same amount at the same rate can start repayment owing different amounts, and the schedule then treats those different amounts identically.

One consequence is worth carrying around, and so is the limit on it. On an unsubsidised loan, paying the interest as it accrues reproduces the subsidised opening balance exactly: there is nothing left to capitalise, so repayment starts on the original principal and every figure from the first payment onwards is identical.

It does not reproduce the subsidised cost, and the difference is the whole point of a subsidy. A subsidy is a transfer: the interest is paid, but somebody else pays it and the borrower never does. A borrower who pays it themselves has paid it. The two loans match from the first repayment onwards and differ by every dollar handed over before then.

In the United States, eligibility for subsidised borrowing has been restricted to undergraduate students demonstrating financial need under the federal aid formula, with the amounts and the qualifying periods set by statute and revised from time to time. Treat the two categories as mechanisms and check the current rules for who qualifies and for how long.

Accrued interest and principal are two separate balances, and knowing that is most of understanding a student loan. On a United States federal student loan, interest is computed each day on the principal balance only:

$$\text{interest for one day} = \text{principal} \times \frac{r}{\text{days in the year}}$$

Unpaid interest collects in an accrued-interest balance that does not itself accrue. That is the difference between a student loan and a credit card, where unpaid interest joins the balance each month and compounds from there. The contrast is about how the balance grows and about nothing else. A year of non-payment on a federal student loan still runs through delinquency to default, which brings collection costs, offset of tax refunds and wage garnishment without a court order, so a gentler accrual rule is not a gentler position to be in.

Capitalisation is the event that ends the separation. The accrued interest $A$ is added to the principal $P$:

$$P_{\text{new}} = P + A$$

and from that moment the former interest is principal, so it accrues like principal. Nothing is charged twice. What changes is the base the daily calculation runs on.

Worked in full below: $20,000 unsubsidised at 6 percent, four years of enrolment. $4,800 accrues, and capitalising it starts a ten year standard plan at $24,800.

The first row is the one that gets dropped, and dropping it overstates the case for paying early by about four times. The left column repays $26,644.92 and also hands over $4,800 before repayment begins, so what it costs is both rows and not just the bottom one. Measured that way, capitalisation costs $1,594.78, which is the interest charged on the interest and nothing else.

The $6,394.78 gap between the two bottom figures is a different quantity, and it is the one most often quoted as the cost of capitalisation. It is what the right-hand column pays relative to somebody who handed over nothing during enrolment and owed nothing for it, which describes a subsidised borrower rather than a borrower who paid as they went. So $6,394.78 measures the subsidy and $1,594.78 measures capitalisation.

Which events trigger capitalisation is set by regulation, and that list has been both narrowed and widened, so check what applies to your loan. The mechanism is identical whenever one fires. On a private student loan, the triggers are whatever the contract says they are.

A standard repayment plan amortises the balance over a fixed term. Three inputs, one output. The payment is the single amount that, repeated for the whole term, lands the balance exactly on zero at the final payment:

$$M = P\,\frac{i}{1 - (1+i)^{-n}}$$

with $i$ the annual rate divided by 12 and $n$ the number of monthly payments. On $24,800 at 6 percent over ten years that is $275.33 a month. Of the first payment, $124.00 is the interest for that month and the rest reduces the balance, and the split shifts towards principal every month after that. The shape is the same on any level-payment debt, which is how amortisation works in detail.

Two rules about where the money lands matter more on student loans than on most debts:

• On a federal loan, a payment is applied to fees first, then to accrued interest, then to principal. That order is set by regulation rather than by the servicer. If an accrued-interest balance is sitting alongside the principal, early payments clear that before they touch the debt itself. A private lender applies payments in whatever order its own contract states.
• Paying extra and shortening the loan are not the same act. Federal loans can be prepaid without penalty, and money above the amount due runs through the same order, so it reaches principal once fees and accrued interest are clear. What it does not do by itself is change the schedule: a servicer can also push the due date forward, which leaves the borrower paid ahead and skipping months rather than finishing sooner. A standing instruction covering both halves, apply the extra to principal and leave the due date alone, is what settles which of the two happens.

The standard plan is the default in the literal sense: a borrower who chooses nothing lands on it. It is also the yardstick. Any plan producing a smaller payment on the same balance at the same rate is stretching the term, cancelling a balance at the end, or both. The loan payment calculator above runs the standard direction, from a balance and a rate and a term to a payment. The student loan payoff calculator runs the inverse, turning a payment someone would actually make back into a number of months.

An income-driven plan inverts the calculation. It derives the payment from income rather than from the balance: a percentage of income above a threshold that scales with household size, recertified each year. The balance is not an input, so the term and the total repaid come out of the payment rather than the other way round.

Two consequences follow from the arithmetic alone.

First, the payment can be smaller than the interest accruing over the same period. One month's interest on $24,800 at 6 percent is $124.00, so a payment of $150 leaves under a fifth of itself for the balance:

Same debt, same rate, and the interest more than triples. A payment below $124.00 makes the balance grow instead of shrink, which is negative amortisation.

Second, and it changes how that table reads: income-driven plans in the United States cancel whatever is left at the end of a set term, and some carry an interest benefit covering part of the accrual a low payment misses. So the table shows what a small fixed payment does to a balance, not what an income-driven plan costs. A borrower on one may never reach the end of the amortisation.

So a plan cannot be judged from its payment. A lower payment costs more if the loan runs to a zero balance, and can cost far less if a balance is cancelled. Whether a cancelled balance counts as taxable income is a matter of federal tax law, which has changed more than once, and a state can treat the same cancelled balance differently from the federal rule in the same year. The percentages, thresholds, terms and qualifying-employment rules are set by regulation and revised, so take the mechanism from here and the parameters from the rules in force.

Two different operations get called the same thing, and the difference is the whole point.

Federal consolidation combines existing federal loans into one new federal loan. The rate is a weighted average of the loans being combined, rounded up to the nearest eighth of a percentage point, so it is not a route to a cheaper rate. It keeps federal status and reduces several bills to one, and it changes how payments already made are counted towards cancellation, which is worth checking before doing it rather than after.

Refinancing means a private lender pays off the existing loans and issues a new private loan on its own terms. If your credit and income support a lower rate, the saving is real:

What the lower rate costs is the note. Federal repayment, postponement and cancellation provisions attach to the loan, not to the borrower, so refinancing ends access to income-driven payment formulas, to statutory deferment and forbearance, to cancellation for qualifying employment, and to statutory discharge on death or total and permanent disability. A private lender may write comparable terms into its own contract, but a contract term is one the lender sets and can revise, which is a different kind of promise from one written in statute. There is no route back either: a refinanced loan cannot be made federal again. That is a swap of a certain, quantified interest saving for a set of provisions whose value depends on circumstances nobody can price in advance.

Read the third row of that table before reading the second. $189.72 a month is below $275.33, and $9,349.30 of interest is above $8,239.70. The payment fell and the undiscounted total rose. Rate and term are separate decisions, and a quote bundles them into one number.

That last comparison needs one qualification, because every total on this page is a plain sum of payments with nothing discounted, and the third row spreads its payments over five more years than the first. Treating a dollar due in year fourteen as equal to a dollar due in year one is what makes the fifteen year row look dearer. Discount both streams at anything above roughly 1.4 percent a year and the ranking flips, because the later payments shrink faster than the extra interest adds up. Which figure to act on depends on what the money not paid in the early years would otherwise do, and that is a fact about the borrower rather than about the loan. The row is a comparison of cash totals, not a verdict.