Household formation is the least glamorous statistic in American macroeconomics and one of the most consequential. The Census Bureau counts a household every time one or more people occupy a housing unit of their own; formation happens when a young adult leaves a parent’s home, a couple marries and signs a first lease, a first child arrives and a nursery must be furnished. When formation slows, the absence shows up everywhere: apartment construction, appliance sales, school enrollment a few years on, state sales-tax receipts. This essay is about the largest deliberate drag on that process for adults under 35 — the roughly $1.6 trillion in student loan debt carried by more than 40 million Americans — and about the statutes, scoring conventions, and tax provisions that convert delayed weddings, postponed children, and foregone starter homes into someone else’s interest income.
I have spent enough evenings inside Congressional Budget Office baseline tables and Federal Register notices to distrust the word “technical” when it is attached to budget rules. Technical is where the transfers hide. The student loan system is the best-documented case in modern federal policy: a machine assembled across six decades, every part stamped with a roll-call vote, every part moving money from young households to creditors and to the Treasury’s own ledger.
What Household Formation Actually Measures
Start with the measurement rules, because they decide what gets counted and what gets missed. The Census Bureau defines a household as all the people occupying a single housing unit. Household formation occurs whenever a new unit acquires occupants — a graduate leaving the parental home, roommates splitting into two apartments, newlyweds renting a one-bedroom. The headship rate, the share of adults who head their own household, is the scoreboard demographers watch.
Two adjacent measures matter here. Doubled-up households — adults living with parents or non-partner roommates out of economic necessity — are formation that failed to happen. And the under-35 homeownership rate tracks the capital-accumulation half of formation: the purchase that historically turned a renter into a household with equity. At the pandemic low point, roughly half of American adults under 30 were living in a parent’s home, a share unmatched since the 1930s. That is not a lifestyle trend. It is a balance-sheet statement.
The Statute Trail: This Debt Was Designed, Not Discovered
Treat the $1.6 trillion as engineering, not weather. The Higher Education Act of 1965 created a guaranteed-loan program: banks lent, the federal government absorbed nearly all default risk, and banks kept the interest spread — socialized risk, privatized return, from the first page onward. The Education Amendments of 1972 built Sallie Mae, a government-sponsored enterprise modeled on Fannie Mae, to give bank-originated student loans a liquid secondary market. The Middle Income Student Assistance Act of 1978 extended subsidized borrowing up the income scale. The 1992 reauthorization raised loan ceilings and opened unsubsidized federal loans to everyone; university bursars read those ceilings as a price floor, and tuition rose to meet the available credit.
Then came the hardening. Between 1976 and 1998, Congress stripped bankruptcy protection from federal student loans in stages. The Bankruptcy Abuse Prevention and Consumer Protection Act of 2005 extended comparable protection to most private student loans. The College Cost Reduction and Access Act of 2007 created income-based repayment and Public Service Loan Forgiveness. The Health Care and Education Reconciliation Act of 2010 cut banks out of origination and replaced them with federal direct lending plus contracted servicers. The Bipartisan Student Loan Certainty Act of 2013 fixed the pricing formula. And the 2025 budget reconciliation law unwound the most recent repayment plan and rebuilt the system once more — the third or fourth such rebuild in two decades, each retaining the same chassis: accrual, capitalization, decades-long horizons, and, eventually, taxable forgiveness.
The Bankruptcy Carve-Out
Student loans can be discharged in bankruptcy only by proving undue hardship — in most courts, the Brunner test, which asks a debtor to show that a minimal standard of living is impossible, that it will stay impossible, and that repayment was attempted in good faith. Judges have summarized the standard as a “certainty of hopelessness.” No credit card, auto loan, or casino marker enjoys this protection. The stated rationale is taxpayer protection. The operative effect is that student lenders — the federal government chief among them — hold a lien on future earnings that no other creditor can obtain.
The Rate Formula and the Origination Fee
Since 2013, new federal loans have been priced at the high yield on the 10-year Treasury note at the May auction plus a statutory spread: 2.05 percentage points for undergraduates, 3.6 for graduate borrowers, 4.6 for PLUS loans, fixed for the life of the loan. On top of that, the Education Department collects an origination fee of just over one percent — about $106 per $10,000 borrowed — deducted at disbursement, before a single class is attended, and scored in the budget as deficit reduction. An excise on borrowing, paid up front, by the population with the least wealth. Both parties celebrated the 2013 compromise as “market-based.” A market rate, charged by a creditor with no competitor and powers no market lender holds: administrative wage garnishment without a court order, seizure of tax refunds, and — since the Debt Collection Improvement Act of 1996 — garnishment of Social Security checks.
The Scoring Trick: How Borrowers Became a Profit Center
Here is the accounting convention doing the quiet work. Under the Federal Credit Reform Act of 1990 (FCRA), when the government lends, the budget records the expected lifetime cost as a single number: projected payments minus projected receipts, discounted to present value using Treasury interest rates. Because Treasury rates treat the government’s own portfolio as essentially riskless, long-horizon, income-contingent student loans routinely score as negative subsidies — profit, in plain English. The Congressional Budget Office’s alternative, fair-value accounting, discounts at market-equivalent rates to reflect the real chance that borrowers default, drop out, or die with balances outstanding. The gap between the two methods has run to tens of billions of dollars a year. Every claim that “the government profits from student loans” rests on the convention that assumes lending to the most financially fragile age cohort in the country is risk-free.
The measurement rules flatter the system elsewhere, too. The official cohort default rate — the statistic that decides whether a college keeps access to federal aid — counts defaults only within the first three years of repayment, a window Congress widened from two years in 2008. Longer-horizon default simply falls out of the frame. The New York Fed’s household debt data tell the fuller story: delinquency concentrates among borrowers with small balances, often under $10,000, who attended for-profit or two-year colleges and frequently left without a degree — a pattern documented by economists Adam Looney and Constantine Yannelis. The people the system was least designed for are the people it extracts from most reliably.
And then there is capitalization: when a payment is smaller than the month’s accrued interest — the ordinary condition of income-contingent plans for low earners — the unpaid interest is added to principal, and the borrower begins paying interest on interest. The plans marketed as mercy were engineered to grow balances first and forgive them, perhaps, two decades later. That is not a glitch. It is the design.
The Evidence on Family Formation
Homeownership and the Underwriting Wall
Does student debt actually delay household formation? The data answer yes, through three channels. The first is underwriting. Debt-to-income ratio (DTI) — total monthly debt payments divided by gross monthly income — long sat at a 43 percent ceiling for qualified mortgages, and a $400 monthly student loan payment consumes roughly $60,000 of borrowing capacity at recent rates. Federal Reserve Board economists Alvaro Mezza, Daniel Ringo, Shane Sherlund, and Kamila Sommer, tracking borrowers who left school in the late 2000s, estimated that each additional $1,000 of student debt lowered the probability of owning a home within a few years by roughly one to two percentage points. Multiply that across a cohort carrying $30,000 and the arithmetic writes itself.
The second channel is the down payment: every dollar of monthly debt service is a dollar not compounding toward a 20 percent stake. The third is geography — economists have linked higher balances to lower rates of moving for a better job, because a borrower in repayment cannot bankroll a job search in a distant, expensive city.
Marriage, Children, and the Spouse-Counting Rule
Some of the marriage effect is mechanical: panel studies, notably economist Slavyana Gicheva’s, associate higher student debt with lower probabilities of first marriage by the mid-twenties. But the 2007 repayment statute added its own penalty. For a married couple filing jointly, both incomes count in the payment formula, so the monthly bill can jump the month after the wedding; filing separately shields a spouse’s income but forfeits other tax benefits. A rule choice, made in statute, that literally prices marriage for indebted couples. Survey evidence extends the pattern to childbearing — higher balances, later first births, and child-care costs arriving precisely when repayment peaks.
The Spending Effect: A Generation Financing Its Financiers
The macroeconomic channel is simple: debt service is consumption that never happens. Median payments run in the low hundreds of dollars per month, and the Federal Reserve’s Survey of Household Economics and Decisionmaking finds, year after year, that roughly a third of American adults could not cover a $400 emergency with cash or its equivalent — with student borrowers reporting more skipped medical care and more purchases postponed indefinitely. The distributional kicker is substitution: a household whose loan payment crowds out groceries does not stop spending. It borrows elsewhere — on credit cards, five-year auto loans, installment plans. A federal program charging 6.5 percent functions, at the margin, as a feeder for plastic charging three times that.
The system has a second act, and it involves the borrowers’ parents. Parent PLUS loans — federal credit extended to parents with minimal underwriting — now total roughly $100 billion. Under the 1996 offset law, defaulted federal student debt is garnished from Social Security checks without a court order; a Government Accountability Office review found more than 100,000 beneficiaries subject to such offsets, most of them in their fifties or older, many pushed toward or below the poverty line. A program built to finance the next generation’s human capital is quietly consuming the last generation’s retirement income.
The Tax Provisions: Relief Calibrated to Be Small
The tax code contributes a pair of rules that look generous and are priced, precisely, not to be. First, the student loan interest deduction (26 U.S.C. § 221): up to $2,500 of interest, deductible without itemizing, but phased out for single filers with incomes in the high five figures and for couples at roughly double that. Run the arithmetic for a borrower with $60,000 at 6.5 percent — about $3,900 of first-year interest, of which the code will acknowledge $2,500. Created by the Taxpayer Relief Act of 1997, the deduction has survived every reform since, not because it is powerful but because it is cheap. A gesture calibrated to survive a budget table.
Second, the “tax bomb.” Under the Internal Revenue Code, cancellation of debt counts as gross income in the year it is forgiven. Income-driven plans forgive the remaining balance after 20 to 25 years — typically in a borrower’s forties or fifties, at peak earnings, often with a balance that capitalized upward for years. A $50,000 forgiveness becomes an $11,000 federal tax bill due the following April. The American Rescue Plan Act of 2021 excluded forgiven student loan balances from federal taxable income through tax year 2025. The expiration date is in the statute; someone wrote it. Several states tax forgiven balances even now. Forgiveness, as engineered, is a tax liability with a two-decade fuse.
Who Pays, Who Profits: The Ledger
Who pays: borrowers, who cover the Treasury spread, the origination fee, capitalized interest, and — at the end of the road — income tax on whatever is forgiven; non-completers, who owe the debt without the degree’s earnings premium and default at the highest rates; parent borrowers, whose Social Security checks are garnishable; and the macroeconomy, in the form of apartments never rented and cribs never bought.
Who profits: the Treasury’s scorekeeping, which books the program as revenue under FCRA conventions; servicers paid per account — Navient agreed in 2022 to a $1.85 billion settlement with nearly forty state attorneys general over allegations it steered borrowers into forbearance, which accrues interest, because forbearance was cheaper to process than income-driven plans; guarantee agencies and private collectors, paid a percentage of what they extract; the pre-2010 bank complex, which held guarantees covering nearly all principal and interest against default while keeping the interest spread; and, during the Title IV boom, the shareholders of for-profit college chains that collapsed under their own recruitment costs, leaving the debt behind.
One entry for the file: when Public Service Loan Forgiveness opened in October 2017, the Education Department approved fewer than 100 of the first 28,000 applications — paperwork failures, certification lapses, eligibility technicalities. A program’s generosity is measured by its denial rate, and that one set a record worth keeping.
Frequently Asked Questions
How does student debt delay household formation?
Through three mechanisms. Underwriting: loan payments raise debt-to-income ratios and shrink the mortgage a lender will approve — roughly $60,000 of borrowing capacity per $400 of monthly payment at recent rates. Savings: monthly debt service crowds out down-payment accumulation. Legislated rules: income-driven repayment formulas count a spouse’s income for joint filers, raising payments after marriage. Federal Reserve Board researchers estimated each additional $1,000 of debt lowered early-adulthood homeownership probability by roughly one to two percentage points.
Does student debt reduce consumer spending?
Yes, measurably. Monthly payments divert cash from consumption for decades, and Federal Reserve household surveys repeatedly find about a third of adults unable to cover a $400 emergency with cash, student borrowers especially. Where spending continues, it migrates to high-interest substitutes — credit cards above 20 percent APR — so the federal program effectively routes marginal household borrowing to the most expensive credit in the system.
Who profits from the current student loan rules?
Named beneficiaries, not conjecture: the Treasury, whose books record the program as profitable under FCRA scoring conventions; servicers paid per account, one of which, Navient, paid $1.85 billion in 2022 to settle allegations of forbearance steering; guarantee agencies and private collectors paid on recovery; and, historically, banks that held near-total federal guarantees on defaulted loans while keeping the interest spread.
What is the student loan “tax bomb”?
When an income-driven plan forgives a remaining balance after 20 to 25 years, the forgiven amount is treated as taxable income, producing a one-time tax bill — potentially five figures — in the borrower’s peak-earning years. A temporary exclusion enacted in 2021 covers federal forgiveness only through tax year 2025; unless extended, the liability returns, and several states tax forgiven balances regardless.
The Verdict: A Thesis, Not a Eulogy
A fair accounting notes the tradeoffs. Federal lending did open campuses to students whose parents could not write tuition checks; income-driven repayment does relieve genuine cash-flow distress; the average borrower with a completed degree is not the system’s casualty. None of that is the point. The point is that every relief mechanism has been built on a chassis — accrual, capitalization, non-dischargeability, taxable forgiveness, FCRA scoring that books borrower risk as public profit — designed by identifiable people for identifiable beneficiaries, and the chassis is where the wealth moves.
So here is the thesis, stated plainly: student debt is not a policy failure; it is a policy instrument — a five-decade transfer mechanism that uses young households’ family formation as collateral, books their risk as public profit, and calls the arrangement neutral accounting. The missing households are not a cultural trend. They are the measured externality of a discount rate, of a $2,500 cap, of capitalization schedules, of the spouse-counting formula, and of a bankruptcy exception no other creditor enjoys. Any reform that leaves that chassis intact — whatever its branding — is renovation, not removal. The rules were written by people, voted on by people, and can be unwritten by people. That is the whole argument.
This essay opens a running column on scoring conventions and measurement rules as distributional instruments. Next in the series: Parent PLUS and the Social Security offset — how the same 1996 statute that garnishes retiree benefits became the student loan system’s quiet second balance sheet. A glossary entry on FCRA versus fair-value accounting is in preparation, and reader questions for the FAQ above are welcome: send the measurement rule you find most suspicious, and it will get the forensic treatment.