The Student Loan Shadow Default

Today’s Lead-Lag Report post is sponsored by Tuttle Capital

Every phase of the AI buildout has faced challenges — compute in 2021, memory in 2023. Now it’s the connections between the chips. Lasers, optical transceivers, silicon photonics, specialty wafers, and foundry capacity are becoming the highway data has to travel on inside AI data centers — and that highway is increasingly moving from copper to light.

The Tuttle Capital Pure Play Photonics ETF (FOTO) is designed to capture that shift directly, with concentrated exposure to the companies making optical interconnects possible. Launched May 29, 2026 on Cboe. Learn more at fotoetf.com.

Credit reports are catching up to the payment pause, but the real macro hit arrives later

Key Highlights

• Student-loan balances edged down to $1.65 trillion in 2026Q2, but the share 90 or more days delinquent climbed to 10.6%, up from 10.3% the prior quarter.[1]
• The New York Fed counts roughly 1 million new defaults in 2025Q4 and another 2.6 million in 2026Q1.[2]
• Federal Reserve research finds the payment restart reduced spending by $12.20 per week per $10,000 of debt.[3]
• The current retail tape still looks firm: June sales rose 0.2% month over month and 6.7% year over year.[4]
• The risk is concentrated: even with new-default flows moderating in Q2, the accumulated credit-score damage and balance-sheet impairment will keep weighing on marginal demand without producing an economy-wide collapse.

The Delinquency Number Is Back

For three years, the student-loan system ran with its most important macro signal muted. Payments were paused, collection activity was constrained, and the credit file did not immediately show the full cost of missed obligations. That did not erase the liability. It moved the visibility of the liability into the future.

The future is now showing up in the household credit data. Outstanding student-loan balances stood at $1.66 trillion in the first quarter of 2026 and edged down to $1.65 trillion by the second quarter, one of only two debt categories to decline that quarter, while the share of balances 90 or more days delinquent kept climbing, from 9.6% in 2025Q4 to 10.3% in 2026Q1 and 10.6% in 2026Q2.[1] The stock of impaired debt is still rising even as new inflows into serious delinquency have started to moderate. That distinction between stock and flow is where the macro signal lives and where consensus is now getting the story wrong. The stock reflects damage already delivered to household balance sheets and credit files. It is the source of the drag on future spending. A moderation in the flow, welcome as it is, does not repair the credit files of the roughly 3.6 million borrowers who defaulted in the prior two quarters.

The flow data are more striking than the stock. The New York Fed estimated roughly 1 million federal borrowers entered default in the fourth quarter of 2025 and another 2.6 million in the first quarter of 2026, driven by the mechanical resumption of credit reporting after the payment pause.[2] The second-quarter update, released on August 11, showed the pace of new transitions into serious delinquency dropping to 7.83% from 12.88% a year earlier, a significant deceleration in the flow. However, the New York Fed cautioned that re-reporting of previously defaulted student debt is still distorting the underlying signal, and the credit-file impairment already delivered to the prior wave of defaulters is a stock effect that persists regardless of what happens to the flow going forward. The narrative shift from Q1's 'record default' headlines to Q2's stabilization is real at the margin, but neither headline captures what actually matters for the macro.

The headline is not that defaults returned. It is that the reporting lag first hid the cash-flow shock from the retail data, and now the moderation in new-flow defaults is being read as an all-clear when the stock of impaired credit files continues to widen.

The Shadow Default Is A Credit-Report Event

The repayment calendar explains why the signal feels abrupt. Interest resumed in September 2023, first payments were due in October 2023, and the twelve-month on-ramp ran through October 2024.[5] Reporting of first delinquencies began in the first quarter of 2025. More than 17% of borrowers then fell at least 90 days past due at least once, according to New York Fed researchers.[2] A borrower can therefore experience the budget shock first, the credit-score shock second, and the formal default much later. The sequence matters for anyone reading only monthly sales data.

The credit file is not a neutral ledger. The New York Fed estimates that a new 90-day delinquency can reduce scores by 171 points for a superprime borrower and 87 points for a subprime borrower.[6] In its later analysis, average scores for recent defaulters fell 91 points, from 567 to 476, between the third quarter of 2024 and the fourth quarter of 2025.[2] That can change the price and availability of auto credit, revolving credit, insurance, and housing finance even before a household stops spending altogether.

The delinquency is also correlated with other stress. Nearly 40% of recent defaulters with auto loans were past due, as were 56% of those with credit cards and 20% of those with mortgages.[2] The right interpretation is not that student loans have already triggered a universal credit event. It is that a payment resumption can expose a fragile household balance sheet that looked healthier while the file was protected.

Urban Institute data point in the same direction. In August 2025, 21% of borrowers had been at least 60 days past due during the prior 24 months, compared with 3% during the pause and protection period and 19% in 2019. Urban projects that its August 2026 measure could exceed any point since at least 2015.[7] The shadow default is the interval between a household losing flexibility and the economy recognizing why.

Why The Macro Hit Lags

The spending channel is measurable, but it is not instantaneous. Federal Reserve researchers studied 55 million people, 89 million cards, roughly $800 billion of sales, and 18,178 ZIP codes, retaining nearly 98% of eligible student balances.[3] After the payment announcement, spending fell $6.20 per week for every $10,000 of debt. After payments resumed, the reduction reached $12.20 per week per $10,000.[3] The difference is the important part: the system reacts more strongly when a scheduled debit becomes a real household constraint.

The Fed translates that effect into an annual reduction of about $630 per $10,000 of balance, $1,590 for the median borrower, and $2,980 for the mean borrower. In aggregate, it estimates an $80 billion annualized spending reduction, roughly 0.3% of GDP and 0.4% of PCE.[3] That is large enough to matter at the margin and too small, by itself, to look like a recession in the headline data. A concentrated drag can be economically real while remaining statistically easy to miss.

The current retail numbers show why the lag matters. Census reported June retail and food-services sales of $768.6 billion, up 0.2% month over month and 6.7% year over year, with the April-to-June quarter up 6.4% from a year earlier.[4] June unemployment was 4.2%, payrolls rose by 57,000, and average hourly earnings were up 3.5% year over year.[8] Those are not readings of a consumer in free fall. They are the surface conditions under which an uneven payment shock can remain buried in category and regional dispersion.

This is why the retail tape should not be treated as a clean falsification of the student-loan thesis. Aggregate sales can hold up as higher-income households, online channels, and essential categories carry the index. The question is whether high-exposure borrowers reduce discretionary purchases, delay financed durables, or lose access to credit in the months after reporting and collections normalize.

Collections Create A K-Shaped Consumer

The next phase is not simply a larger delinquency print. It is the interaction between credit reporting and collections. The Department of Education temporarily delayed administrative wage garnishment and Treasury Offset activity while it implemented repayment reforms, while continuing to report defaults to credit agencies.[9] A borrower can therefore lose credit access before the government removes cash from a paycheck. That ordering creates a slow, asymmetric transmission mechanism.

The Federal Reserve household survey shows how uneven the burden already is. Forty-five percent of adults with student loans were not currently required to make payments, while 73% of those required to pay said they paid the full amount and 41% reported recent difficulty.[10] Among required payers, full-payment rates ranged from 42% for households earning under $25,000 to 92% for households earning at least $100,000. Seventy-six percent of those reporting difficulty cited affordability-related reasons.[10] The average borrower is therefore the wrong unit of analysis. The marginal borrower is the macro variable.

The New York Fed also provides the limiting case. Delinquent or newly defaulted borrowers represented only 2% of the credit population; their balances represented 2.7% of auto loans, 2% of credit-card balances, and 1% of mortgage balances.[2] That makes broad contagion unlikely on the evidence currently available. It does not make the spending channel irrelevant. It means the effect should first appear in the places where a small loss of credit or cash flow changes a purchase decision.

The additional risk is the delayed SAVE cohort. About 7 million borrowers on SAVE remained on a delayed repayment timeline, creating the possibility of a second default wave around the nine-month mark.[2] If that cohort re-enters the reporting system while wage-garnishment activity stays delayed, the macro story will look contradictory: credit stress worsens, but immediate cash extraction remains incomplete. That contradiction is the shadow default in its purest form.

What Would Prove This Wrong

The contrarian thesis must have an observable failure condition. It is not enough to point to a rising delinquency rate and declare victory. The mechanism requires a sequence: reporting normalizes, affected households lose flexibility, and spending or credit access deteriorates in the areas and categories with the highest exposure.

This thesis is wrong if the next two quarters show (a) student-loan 90+ day delinquency on balances rolling over decisively, not just stabilizing, and new-default flows continuing to moderate as they did in Q2, (b) no meaningful deterioration in credit access or other credit-file downstream effects for the roughly 3.6 million borrowers who defaulted in the prior two quarters, and (c) retail sales holding trend growth in real terms with no relative underperformance in categories concentrated among borrowers between ages 25 and 44. The Q2 print moved partially in the direction of (a). Two more quarters of the same, combined with (b) and (c) both holding, would falsify the thesis.

The monitoring list is therefore narrow. Track the New York Fed delinquency and default flows, the share of SAVE borrowers reaching repayment milestones, credit-score and new-account outcomes, and retail categories that depend on financed or discretionary demand. Compare high-exposure ZIP codes with lower-exposure areas rather than relying on the national sales headline. The thesis should be downgraded if the credit file worsens without any corresponding change in access or spending, because then the signal is administrative rather than macroeconomic.

For now, the evidence supports a delayed, concentrated drag rather than an economy-wide collapse. The payment pause did not remove the liability. It separated the liability from the moment when the credit system and the consumer budget could see it. The point is not that the economy has collapsed; it is that the same liability can be invisible in the aggregate until the credit file forces a budget response.

Few understand this.

Notes

[1] Federal Reserve Bank of New York, Household Debt and Credit, 2026Q2 (released August 11, 2026). https://www.newyorkfed.org/microeconomics/hhdc and 2026Q1 report at https://www.newyorkfed.org/medialibrary/interactives/householdcredit/data/pdf/HHDC_2026Q1

[2] Federal Reserve Bank of New York, “Federal Student Loan Defaults Return After Pandemic Pause,” May 12, 2026. https://libertystreeteconomics.newyorkfed.org/2026/05/federal-student-loan-defaults-return-after-pandemic-pause/

[3] Board of Governors of the Federal Reserve System, FEDS Notes, “Debt Payments and Spending: Evidence from the 2023 Student Loan Payment Restart,” September 5, 2025. https://www.federalreserve.gov/econres/notes/feds-notes/debt-payments-and-spending-evidence-from-the-2023-student-loan-payment-20250905.html

[4] U.S. Census Bureau, Advance Monthly Sales for Retail and Food Services, June 2026. https://www.census.gov/retail/sales.html

[5] Federal Reserve Bank of New York, “Federal Student Loan Defaults Return After Pandemic Pause,” May 12, 2026. https://libertystreeteconomics.newyorkfed.org/2026/05/federal-student-loan-defaults-return-after-pandemic-pause/

[6] Federal Reserve Bank of New York, “Credit Score Impacts from Past-Due Student Loan Payments,” March 26, 2025. https://libertystreeteconomics.newyorkfed.org/2025/03/credit-score-impacts-from-past-due-student-loan-payments/

[7] Urban Institute, “Student Loan Repayment Since the Payment Restart,” January 2026. https://www.urban.org/sites/default/files/2026-01/Final_Student_Loan_Repyament_Since_the_Payment_Restart.pdf

[8] U.S. Bureau of Labor Statistics, The Employment Situation, June 2026. https://www.bls.gov/news.release/empsit.nr0.htm

[9] U.S. Department of Education, “U.S. Department of Education Delays Involuntary Collections Amid Ongoing Student Loan Repayment Improvements,” January 16, 2026. http://www.ed.gov/about/news/press-release/us-department-of-education-delays-involuntary-collections-amid-ongoing-student-loan-repayment-improvements

[10] Board of Governors of the Federal Reserve System, Economic Well-Being of U.S. Households in 2025: Credit, May 26, 2026. https://www.federalreserve.gov/publications/2026-economic-well-being-of-us-households-in-2025-credit.htm

[11] Federal Reserve Bank of New York, Household Debt and Credit, 2026Q1. https://www.newyorkfed.org/medialibrary/interactives/householdcredit/data/pdf/HHDC_2026Q1

[12] Federal Reserve Bank of New York, “Federal Student Loan Defaults Return After Pandemic Pause,” May 12, 2026. https://libertystreeteconomics.newyorkfed.org/2026/05/federal-student-loan-defaults-return-after-pandemic-pause/

[13] Federal Reserve Bank of New York, “Credit Score Impacts from Past-Due Student Loan Payments,” March 26, 2025. https://libertystreeteconomics.newyorkfed.org/2025/03/credit-score-impacts-from-past-due-student-loan-payments/

[14] Urban Institute, “A First Look at Student Loan Borrowers Exiting Default via the Fresh Start Program,” May 2026. https://www.urban.org/sites/default/files/2026-05/Final_A_First_Look_at_Student_Loan_Borrowers_Exiting_Default_via_the_Fresh_Start_Program.pdf

[15] U.S. Department of Education, “Default,” Federal Student Aid. https://studentaid.gov/articles/default/

[16] MOHELA, Understanding Credit Reporting. https://mohela.studentaid.gov/DL/resourceCenter/understandingCredit.aspx

[17] TransUnion, “May 2025 Student Loan Update,” May 2025. https://newsroom.transunion.com/may-2025-student-loan-update/

[18] Federal Reserve Bank of New York, Student Debt research page. https://www.newyorkfed.org/microeconomics/topics/student-debt