The growth in U.S. margin borrowing raises a practical question for investors and lenders: can they withstand a correction without being forced to sell into it? Our central argument is that borrowing should be judged against the cash and collateral available under stress. A larger loan balance creates greater dollar exposure, but its danger depends on whether many investors must raise cash at the same time.
The borrowing has grown
FINRA reports $1.480 trillion in customer securities margin debt for September 2026, up $26.2 billion from August and 20.8% from December 2025. In the same reporting population, September 2016 stood at $540.615 billion: the ten-year increase is approximately 174%. The chart below extends the perspective to January 2000, while marking the earlier change in reporting coverage. [1]
That expansion warrants scrutiny. It does not mean financial vulnerability has risen by 174%. Stock market values, collateral, investor wealth and inflation have also changed. To judge the increase in risk, we need to compare debt with the assets supporting it and assess borrowers’ liquid resources. Comparisons stretching back before February 2010 also cross a change in reporting coverage. [1]
The wider financing picture includes hedge-fund borrowing and derivatives. Our workbook separates these categories because overlapping loans, different reporting populations and collateral deposits cannot be added into a reliable national borrowing total. An options margin deposit secures an obligation; it does not disclose how much money was borrowed to fund the position. [2]
Systemic exposure beyond FINRA
FINRA is one window into the financing system. It aggregates customer securities-account debt across its member brokers, including holdings traded on NYSE and Nasdaq markets. The missing exposure is not simply a list of other stock indexes: public reporting does not allocate these loans by index or isolate the borrowing behind particular options strategies. [1]
Our accompanying workbook records $7.952 trillion in secured borrowing by the qualifying hedge funds covered by OFR’s second-quarter 2026 data, including prime brokerage, repo and other secured financing. This is a broader reporting population across asset classes and domiciles, and may overlap FINRA debt. It cannot be added to FINRA’s $1.480 trillion to produce a clean U.S. investor-loan total. OFR identifies collateral declines, margin calls and dependence on short-term creditors as linked vulnerabilities. [2, 13]
Options and futures create another route for cash demands. Initial margin deposits secure obligations; variation margin and changing requirements can create further liquidity needs. Clearinghouse collateral figures show resources posted against risk, not the amount borrowed to fund positions. Derivatives can hedge a portfolio as well as increase its exposure, so neither contract notional nor collateral deposits should be treated as extra loan balances. The systemic concern is simultaneous cash demand across these channels. [2]
How the pressure can spread
The first stage is pressure on individual positions. A decline in collateral value can leave a borrower below the lender’s required equity cushion. Depending on the account and contract, the investor must contribute cash or eligible assets, reduce the position, or face liquidation. Higher margin requirements can intensify that pressure.
Consider a simplified account holding $100 of securities against a $50 loan. After a 30% price decline, its assets are worth $70 and its equity is $20, or 28.6% of asset value. If the lender requires 35% equity, it would need $4.50 of additional cash applied to the loan, or roughly $12.86 of securities sold with the proceeds applied to the loan. These are illustrative assumptions, not an estimate of current accounts or a universal margin rule.
The second stage is selling across portfolios. Investors may sell readily traded assets to raise cash for a stressed position elsewhere. Major stocks and Treasuries can therefore come under pressure even when the original problem lies in another market. Investors who repay their loans through asset sales can still amplify a correction; default is not necessary.
The third stage is pressure on financing providers. If brokers and dealers restrict credit, demand larger collateral cushions, or reduce their willingness to absorb sales, other investors may have to cut positions. These stages can overlap. The outcome depends on concentration, hedges, financing terms and the depth of available buyers.
History shows both amplification and resilience
Recent history offers a useful comparison. The New York Fed describes the unwinding of leveraged Treasury cash–futures trades as a likely contributor to market disruption in March 2020. In April 2025, relatively stable repo funding and continued dealer intermediation helped prevent a comparable broad unwind of that trade. [3]
The April 2025 correction was substantial. The Federal Reserve reported that major stock indexes had fallen almost 20% from their February peaks by the second week of April. It identifies excessive leverage and liquidity mismatches as conditions that can produce self-reinforcing margin spirals. The episode nevertheless showed that a sharp decline need not become a systemic funding crisis. [4]
What the public data cannot answer
The sources assembled so far do not establish the percentage of U.S. securities purchases funded with borrowing, or the proportion of investors unable to meet calls under a particular shock. Outstanding debt is a balance at a point in time; purchases are transactions over a period. Dividing one by the other would not measure the share of purchases financed on margin. [1, 2]
A meaningful stress exercise would test market declines of different sizes alongside tighter collateral requirements, reduced financing and imperfect hedges. It would distinguish losses from cash demands, and cash demands from forced sales. Without account-level liquidity and position data, any systemwide result must be presented as a scenario with explicit assumptions.
Put options add another dimension
The Kobeissi Letter has drawn attention to an exceptional session in S&P 500 put trading. In a post indexed October 19, it reports that Friday’s dollar-notional turnover reached a record $3.3 trillion, 330% above the 2024 daily average. It also reports total Cboe put volume of 9.6 million contracts, the fourth-highest since 2015. These are the source’s reported figures; the broader Cboe contract count should not be described as SPX puts alone. [14]
The historical comparison gives the scale perspective. Kobeissi says that before the 2020 pandemic, daily S&P 500 put notional volumes ranged from $100 billion to $500 billion. The reported $3.3 trillion session is therefore 6.6 times the top of that range and 33 times the bottom. Those multiples are our arithmetic using Kobeissi’s figures. The 330% comparison means 4.3 times the 2024 average—not 3.3 times. [14]
The comparison is a peak session against an earlier reported daily range. It is not an inflation-adjusted measure of investor fear, and it does not control for the growth of the index or changes in options-market participation and contract turnover. It shows how large the reported activity was; it does not establish why every trade occurred.
Notional turnover describes the scale represented by traded contracts. It is different from premiums paid, outstanding positions and borrowed money. We are not adding $3.3 trillion to FINRA margin debt. We are considering unusually heavy options activity alongside an already substantial borrowing balance.
What the independent data show
Options-market expansion extends beyond the reported SPX session. Cboe reports that total index options volume rose 25% and ETF options volume 27% year to date through the second quarter of 2026. Its March study also reports a 36% year-over-year rise in combined Russell 2000 index and ETF options average daily notional volume. These are mixed call-and-put measures over specified periods. They establish broader growth in options activity, but do not verify that put-only turnover or bearish positioning is reaching records across every major index. [22, 23]
Cboe’s own October 19 report provides a separate, directly sourced observation. Across SPX and SPXW, it records 2,590,723 put contracts and 2,177,950 call contracts traded, giving a put/call ratio of 1.19. It also records 13,407,659 outstanding put contracts. These are Monday’s contract counts, not an independent reproduction of Friday’s dollar-notional peak. [16]
The distinction matters to the argument. Trading volume measures transactions during a session; open interest measures contracts still outstanding. A put can protect a long portfolio, express a bearish view or form part of a wider strategy. Large turnover alone does not reveal the balance of net bearish exposure.
Cboe’s research explains why gross options volume cannot by itself establish the resulting hedging pressure. Depending on dealers’ net positions, their adjustments can either amplify a market move or cushion it. That is a reason to examine positioning and liquidity together, rather than assume that every put transaction creates additional selling pressure. The research explains a mechanism; it does not establish current dealer positioning. [15]
Why this assessment is needed now
We are assessing leverage now because inflation, renewed monetary tightening, elevated bond yields and geopolitical uncertainty can affect both valuations and the cost of carrying debt. The concern is that these pressures could coincide with a market decline and expose borrowers’ dependence on continued financing.
On September 16, the Fed raised its policy-rate target by 0.25 percentage point to 3.75–4.00%, citing elevated inflation. The 10-year Treasury yield is a separate, market-determined longer-term rate, although expectations about future policy influence it. Higher longer-term yields can raise refinancing costs, lower existing fixed-rate bond prices and increase the return investors require from other assets. The benchmark is a financing backdrop, rather than a precise clock for a correction. [10, 21, 26]
Why the 10-year Treasury yield matters worldwide
The 10-year U.S. Treasury yield is a widely watched reference for medium- and long-term dollar financing. Investors use Treasury yields as a starting point when comparing returns and credit spreads across sovereign and corporate debt. The scale, liquidity and role of the Treasury market in global dollar portfolios give changes in this benchmark consequences beyond the United States. It is a global reference point for financing and valuation, rather than a single rate governing every international trade or currency. [24]
The chart below follows only the 10-year maturity. The horizontal axis shows calendar years and the vertical axis shows yield in percent. Every observation represents a constant 10-year maturity, even as the underlying Treasury securities change. The latest available FRED observation is 5.09% on October 16, 2026. This is down from 5.24% on October 9, so the latest weekly move was a decline. [21]
The history begins in 1972, the first full calendar year after the United States suspended official dollar-to-gold convertibility in August 1971. This starting point supplies monetary-system context; it does not establish that ending convertibility caused each subsequent movement in yields. [25]
The benchmark is market-determined and incorporates expectations for Federal Reserve policy, inflation and economic activity, together with the term premium and supply-and-demand conditions. A rise can increase financing costs and reduce existing fixed-rate bond prices, but it cannot by itself isolate weaker demand for U.S. debt. Its relevance to this article is how changing financing costs and collateral values could interact with investor leverage. [24, 26]
Tariffs and conflict add supply-side pressure. In his September 29 assessment, New York Fed President John Williams identified tariffs, Middle East supply-chain and energy disruptions, and AI-related demand as inflation drivers. He also said tariffs were no longer adding to goods inflation and that broader, persistent spillovers had not yet been demonstrated. Iran-related escalation remains a risk to energy supply and transport; the stress scenario is a prolonged disruption that keeps costs elevated while financing becomes less accommodating. We should distinguish this risk from a verified attribution of every price increase to the conflict. [12]
Federal debt adds a separate constraint. CBO’s February 2026 current-law baseline projects publicly held debt rising from 101% of GDP in 2026 to 120% in 2036, with rising net interest costs driving much of the projected increase in deficits. That is evidence of sustained fiscal pressure, rather than proof of an immediate debt spiral or Treasury default. Treasury liquidity also matters because leveraged investors use government securities in financing and trading strategies. [6]
Inflation and household cash flow
The latest inflation and employment releases sharpen the financing question. September CPI rose 3.7% over the year, up from 3.4% in August, and 0.6% over the month. Gasoline accounted for more than half of the monthly rise. Core CPI, excluding food and energy, increased 2.6% over the year and 0.3% over the month. Inflation pressure therefore deserves attention, but the energy contribution should be distinguished from the underlying trend. [17]
The labor chart uses two clearly labelled axes because job counts and percentages measure different things. The payroll line is the total number of jobs; it is not the monthly change. Unemployment is a percentage level, while real earnings is an annual percentage change. Average earnings can also shift when the mix of employed workers changes, particularly during the pandemic, so an aggregate spike need not represent a pay rise for a typical individual.
August core PCE inflation provides another measure of underlying price pressure, but its reference month differs from September CPI. The two measures should not be treated as identical baskets or synchronous observations. [18]
Employment data illuminate a different vulnerability. BLS describes September payroll employment and unemployment as little changed. Nominal average hourly earnings rose 3.0% over the year, while real average hourly earnings fell 0.7%. These releases do not demonstrate a wage-price spiral. They indicate limited job growth and a loss of hourly purchasing power on average, with outcomes differing across workers. [19, 20]
For this assessment, the implication is conditional: if living costs absorb more income while employment opportunities expand slowly, some households may have less flexibility to save, service debt or meet unexpected cash needs. These aggregate releases do not measure the cash reserves of leveraged investors, and they do not prove a recession. They help explain why the combination of persistent price pressure and fragile income growth belongs in a discussion of financial resilience.
These pressures justify examining how a correction could unfold. The latest releases show inflation pressure alongside limited job growth and weaker real hourly earnings, rather than a uniform signal of overheating. Our study tests an adverse combination of risks against the cash reserves, hedges and financing capacity that can absorb shocks. The ten-year increase measures outstanding margin debt, not the share or growth of stock purchases financed with loans. [10]
Policy support has limits
Treasury buybacks serve cash-management and market-liquidity objectives. They should not be presented as proof that Treasury is preventing default or guaranteeing stock valuations. Treasury’s stated program objectives also distinguish buybacks from Federal Reserve monetary policy. [5]
The indicators worth monitoring are debt relative to collateral and market value, financing costs, changes in collateral requirements, concentrated positions and declining market depth. Rising debt alone cannot tell us when a crisis will occur. It does make the system’s ability to raise cash without disruptive selling a more consequential question.
Conclusion
The reported $3.3 trillion put-option session adds a further dimension to this assessment. Relative to Kobeissi’s stated pre-pandemic $100–$500 billion daily range, the reported peak was 6.6–33 times as large. That comparison adds historical scale to the discussion of borrowing and financing pressure; it is not an extra loan balance or proof of an imminent crash. [14]
The reason for this assessment is the combination of greater reported borrowing and pressures that could make borrowing harder to sustain. Inflation and geopolitical supply disruptions can keep financing expensive. Elevated yields can challenge valuations and weaken bond collateral. If earnings or incomes then disappoint, margin calls may arrive when investors have less spare cash.
That combination could turn an ordinary correction into forced selling across portfolios. Its severity depends on leverage relative to collateral, liquid reserves, hedges, lender requirements and the willingness of dealers to continue financing positions. Neither the debt increase nor the current macroeconomic indicators proves that a crisis is imminent.
Investors and lenders should therefore test falling asset prices, sustained financing costs and tighter collateral requirements together. Identify the calls that can be paid from available cash, the positions that must be sold and the exposures shared across borrowers. Our conclusion is a case for preparation now: the larger borrowing balance raises the dollar stakes, and the overlapping risks make the ability to withstand a cash squeeze worth measuring before markets test it.
Sources and methodology
[1] FINRA, Margin Statistics and historical download, retrieved October 20, 2026. September 2026 balance, January 2000–September 2026 monthly history and reporting definitions. https://www.finra.org/rules-guidance/key-topics/margin-accounts/margin-statistics
[2] SP500_Profitability_Review.xlsx, version 3, October 13 2026. Borrowing and collateral tab. Historical figures and calculations are taken from the reviewed workbook; earlier decades require attention to its coverage notes.
[3] Federal Reserve Bank of New York, Recent Developments in Treasury Market Liquidity and Funding Conditions, May 9 2025. https://www.newyorkfed.org/newsevents/speeches/2025/per250509
[4] Federal Reserve, Governor Lisa D. Cook, Financial Stability, May 23 2025. https://www.federalreserve.gov/newsevents/speech/cook20250523a.htm
[5] TreasuryDirect, FAQs about Treasury Securities Buybacks, accessed October 13 2026. https://treasurydirect.gov/help-center/faqs/buyback-faqs/
[6] Congressional Budget Office, The Budget and Economic Outlook 2026 to 2036, February 11 2026. Current-law projections, not realized outcomes. https://www.cbo.gov/publication/61882
[7] BLS, Consumer Price Index August 2026, September 11 2026. https://www.bls.gov/news.release/cpi.nr0.htm
[8] BLS, Employment Situation September 2026, October 2 2026. https://www.bls.gov/news.release/archives/empsit_10022026.htm
[9] BLS, Productivity and Costs Second Quarter 2026 Revised, September 3 2026. https://www.bls.gov/news.release/prod2.nr0.htm
[10] Federal Reserve, FOMC statement, September 16 2026. https://www.federalreserve.gov/newsevents/pressreleases/monetary20260916a.htm
[12] Federal Reserve Bank of New York, John C. Williams, Unwavering Dedication, September 29 2026. Speaker’s assessment and outlook. https://www.newyorkfed.org/newsevents/speeches/2026/wil260929
[13] Office of Financial Research, Hedge Fund Monitor, Borrowing by type and SEC Form PF methodology. https://www.financialresearch.gov/hedge-fund-monitor/categories/leverage/chart-23/
[14] The Kobeissi Letter, put-option volume post indexed October 19, 2026. Reported figures and original chart credited to Goldman Sachs. Source.
[15] Cboe, Mandy Xu, Volatility Insights: Much Ado About 0DTEs, September 8, 2023. Used for the distinction between gross volume and net dealer exposure, not as a current positioning estimate. Source.
[16] Cboe Daily Market Statistics, October 19, 2026 selected date. SPX + SPXW contract counts, open interest and ratio. Source.
[17] BLS, Consumer Price Index—September 2026, released October 14, 2026. Source.
[18] BEA, Personal Income and Outlays—August 2026, released September 30, 2026. Source.
[19] BLS, Employment Situation—September 2026, released October 2, 2026. Payroll estimate preliminary. Source.
[20] BLS, Real Earnings—September 2026, released October 14, 2026. Source.
[21] Board of Governors of the Federal Reserve System, Market Yield on U.S. Treasury Securities at 10-Year Constant Maturity, Quoted on an Investment Basis [DGS10], H.15, retrieved through FRED October 20, 2026. Daily series, January 1972–October 16, 2026. Source.
[22] Cboe, State of the Options Industry: Options Market Continued to Break Records in Q2 2026, July 21, 2026. Index and ETF volume includes puts and calls. Source.
[23] Cboe, Mandy Xu, Small Caps, Big Momentum: The Rise of 0DTE Trading in Russell 2000 Index & ETF Options, March 13, 2026. Combined Russell index and ETF options, not put-only volume. Source.
[24] Federal Reserve Bank of New York, Michael J. Fleming, The Benchmark U.S. Treasury Market: Recent Performance and Possible Alternatives, April 2000. General explanation of Treasury securities as pricing and risk-reference benchmarks. Source.
[25] Federal Reserve History, Nixon Ends Convertibility of U.S. Dollars to Gold and Announces Wage/Price Controls. Suspension in August 1971. Source.
[26] Federal Reserve Bank of New York, Treasury Term Premia. Distinguishes expected future short-term rates from the term premium in Treasury yields. Source.
[27] Monthly BLS and BEA series via FRED, retrieved October 20, 2026: CPIAUCSL, CPILFESL, PCEPILFE, PAYEMS, UNRATE, and CES0500000003. All underlying series are seasonally adjusted. Price growth = 100 × (index / index 12 months earlier − 1). Real hourly earnings growth = 100 × [(earnings / CPI) / (earnings / CPI) 12 months earlier − 1]. Payroll jobs are PAYEMS × 1,000; the graph displays millions. Unemployment is the published rate. Data are current-vintage observations and may contain revisions.
Options methodology. The $3.3 trillion and pre-pandemic daily range are attributed to Kobeissi; 6.6–33 times and 4.3 times are arithmetic comparisons based on its figures. The original dollar-notional daily dataset was not obtained and the source chart has not been independently replotted. Its right boundary reads “10 Dec 2027”; a displayed axis can extend beyond the last observation, but that explanation and the exact circled session date remain unverified. The source’s “Friday” wording is retained without inferring a date from the image. October 19 Cboe contract counts are a separate observation.
Revision scope. This edition preserves the original article’s sections, supporting illustrations and source list. The main monthly margin-debt chart and its leading text use FINRA’s October 20 download through September 2026. The supplemental workbook graphic retains the earlier dates identified in its caption; it is not presented as a new observation. Treasury coverage now uses only the 10-year FRED history beginning in 1972. Inflation and employment context uses the releases identified above.
Detailed borrowing and dollar history
Philippe Martin
Founder & Co-CEO
Iversen | Martin