On Friday, November 20, the S&P 500 and the Dow both closed in correction and the Nasdaq composite turned negative for 2026. The S&P 500 fund (SPY) fell 3.64%, the Nasdaq-100 fund (QQQ) fell 5.67%, and Nvidia dropped 14.01% to $125.53 on its seventh straight down day. If you run a small business, your first reaction is probably "that's Wall Street's problem." For most of you, it is, for now. But there's one number from Friday that matters more to you than any stock price: high-yield credit spreads sat at 352 basis points. That's calm.
Calm credit is the window. Your bank is still lending on the terms it was lending on last week. Banks tighten lines and cut limits after spreads blow out, not before. So the useful question isn't "will stocks keep falling?" Nobody knows. It's "what would I want my credit to look like if lending seized up in six weeks, and can I fix that this week while the phone still gets answered?"
This post borrows a lesson from 1907, when J.P. Morgan organized a private rescue of the New York banking system. The lesson isn't about Morgan being a hero. It's that the rescue arrived only after credit had already frozen, and the terms were brutal. You don't want to be the borrower waiting on a rescuer. You want your documents signed before anyone needs rescuing.
What happened
Friday's session was a violent equity de-risking, driven by AI-heavy names. Besides the index and Nvidia moves above, reports said the SEC is examining potential accounting issues at Nvidia, and Anthropic shares slid on their second day of trading after its IPO. NPR covered the Nasdaq going negative for 2026, and CNBC's report as carried by The Loadstar describes the S&P 500 and Dow in correction. MarketScreener has the reported SEC examination of Nvidia, and Modern Ghana has the Anthropic slide.
Bitcoin fell to a six-month low near $58,260 after a $1.8 billion liquidation wave, then bounced back above $60,000. CoinDesk ties the next test to Monday's stock open.
Now the part that matters for borrowers. The VIX closed at 28.44 (the Nov 22 brief notes that reading is three days stale). But the high-yield option-adjusted spread held at 352 bps, and BBB investment-grade spreads were 117 bps. Apprised.news's markets desk names roughly 400 bps as the stress threshold. The economy underneath looks fine on paper: Q3 GDP of +3.4% annualized, unemployment of 4.1%, jobless claims of 203,000. The less comfortable part is rates. Fed funds are at 3.88% against headline CPI of 4.08%, a negative real policy rate, and Bitcoin Magazine and RealClearPolitics both flagged Fed hike risk. Add tanker incidents in the Strait of Hormuz, such as two tankers reported near Oman, and you get energy-cost uncertainty on top.
The historical framing in this post comes from Apprised.news lens analysis, specifically the markets desk's November 22 brief, the markets desk's November 21 brief, and the tech desk's November 21 brief. The markets desk's read: Friday was "flow-driven, AI-concentrated de-risking, not yet a credit-confirmed regime break," and it expects choppy equity weakness over the next few weeks rather than an immediate floor or a systemic event. It also names what would change that. High-yield spreads breaking toward 400 bps while the Fed signals hikes would turn this into a funding-cost story. A VIX back under about 24 with QQQ reclaiming $608.45 would favor a bounce.
One more data point from the insurance desk: six December catastrophe bonds were still arriving on schedule through the Nasdaq's 5.28% drop. That suggests risk capital was still open for business as of this week. And the health desk notes hospitals bracing for a surge in uninsured patients, with nearly half losing money per Kaufman Hall. That's a squeezed counterparty, the kind small suppliers to hospitals should notice.
The historical parallel
In October 1907, a run on New York's trust companies turned into a full panic. There was no central bank to lend against collateral. J.P. Morgan, then the country's most powerful private banker, summoned bankers to his library and organized pools of money to prop up failing trusts, and he made the call on whom to save. The relevant facts for an owner are simple. The rescue happened after credit had seized. The rescuer chose who got help and set the terms. And the episode was a big part of why Congress later created the Federal Reserve.
The Apprised.news lens takes on this are worth reading plainly. The markets desk says that with high-yield spreads at 352 bps there's "no funding crisis to rescue, only an equity drawdown, and no one is convening a pool." The parallel is the reminder that private rescues appear when credit seizes. The tech desk adds that Morgan lent on character, and that the AI trade has the same structure of huge capital extended on belief in a counterparty's books, now questioned for Nvidia. No equivalent organizer is visible in the current reporting.
Translate that to your world. In a squeeze, nobody convenes a pool for a 12-person contractor. The "rescuer" for a small business is its own banker, and your banker decides based on the paperwork already in the file and the covenants already signed.
Where the comparison breaks
Credit isn't frozen. In 1907, the trouble was banks and trusts that couldn't meet withdrawals. Right now, as of Friday's data, the pain sits in equity valuations and crypto, not in lending markets. A 352 bps spread is below the roughly 400 bps stress line the desk names. Comparing Friday to October 1907 would overstate things badly. The honest reading is "equities broke, credit hasn't joined yet."
There is a public lender now. Morgan acted because no one else could. Today the Federal Reserve exists precisely to lend in a panic. That cuts both ways. It lowers the odds of a 1907-style seizure. But the Fed is also the source of the current worry, since fed funds sit below headline inflation and hike chatter is rising. A rate rise can raise your floating-rate costs without any panic at all.
The shock is different. 1907 started with a failed attempt to corner a copper stock and spread through hidden debt in trust companies. This one starts with AI valuation fears, an SEC examination reported through unnamed sources, and a flow-driven selloff. The markets desk itself says Friday's forced-selling mechanics look like one regime-break read seen from two angles, not independent confirmations. Reports of an SEC probe are not findings. Treat the Nvidia story as unresolved.
Your exposure is probably indirect. Most small businesses don't hold much Nvidia. Your risk runs through customers whose wealth or revenue is tied to tech and equities, through your lender's appetite, and through floating-rate debt. That's a narrower and more manageable problem than a banking panic, but it's real.
What it means for a small business
The decision is simple to state and a little annoying to do: use the calm to renew, extend and document your credit, then build a plan with trigger numbers so you aren't deciding under stress. Don't count on a Morgan. Here are two hypothetical examples with the arithmetic shown.
A 12-person HVAC company in Ohio
Say the company has $1.2 million in revenue and a $150,000 bank line that renews in March. The line funds about $60,000 of parts inventory ahead of summer. If spreads widen toward 400 bps, banks commonly tighten terms or trim limits on renewal. Suppose the owner calls in November and asks two things: can the renewal be done early at current terms, and can the drawn balance be termed out over 24 months?
Say the early renewal and term-out cost about $1,500 in fees. Compare that with the downside: if the line were cut from $150,000 to $90,000 at a March renewal, the company would lose $60,000 of capacity, exactly the amount needed for summer inventory. Paying $1,500 to protect $60,000 of funding is a 2.5% cost. The owner should also ask what covenants apply and whether a lower limit could be triggered by a covenant breach. Get the answers in writing. And be aware that the bank may say no, or may reprice. That's information worth having in November rather than March.
A boutique accounting firm with tech-heavy clients
Say 40% of this firm's $900,000 in billings come from tech employees and startups. If clients' equity wealth is down 20% to 30%, a planning assumption (not a forecast) is that payments run 10% to 15% slower and the firm loses about 5% of discretionary work. Here's the arithmetic. Slow payment: 10% to 15% of $360,000 (the tech-exposed 40%) is $36,000 to $54,000 of cash delayed. Lost discretionary work: 5% of $900,000 is $45,000. The swing runs from about $45,000 to roughly $100,000 depending on how the assumptions combine, and a harsher combination reaches the $135,000 end of the range. The firm has a $50,000 line.
So the owner builds a 13-week cash forecast with receivables stretched by 15 days and checks the week when cash first drops below one payroll cycle. If the $50,000 line doesn't cover the gap, that's the argument for asking the bank about a larger limit now, while the firm's numbers still look good.
Both owners should write down three trigger numbers and the action each triggers. Examples: high-yield spreads near 400 bps, VIX staying above the high 20s, or your own days-sales-outstanding rising 10 days. Actions might be freezing hiring, drawing on the line, or calling the lender. If you have floating-rate loans above roughly $100,000, price a fixed-rate or swapped option too. On, say, $400,000 of floating-rate equipment debt, a one-point rate rise adds $4,000 a year. If fixing 60% of the balance costs a one-time $2,500, you can compare those directly, and the fix pays back in well under a year if rates rise a full point. If rates fall instead, you've paid for insurance you didn't need. Say that out loud in your notes so you aren't surprised either way.
Also rank your customers by exposure to the AI and equity selloff, and your suppliers by financing fragility, and call the top three of each. A web-design shop that gets 30% of $600,000 in revenue from AI-adjacent startups should stress-test losing its top three clients (say $90,000) and decide in advance what cost cuts and credit draw would cover three months.
One caution on drawing your line. Drawing down only makes sense if you have a real use or a reserve plan. Keep the money separate from operating cash and document why you drew it. Some lenders read a sudden large draw as a distress signal.
For a related look at how a different cost shock lands on small firms, see Oil Near $100 on Hormuz Fear: What Morgan's 1907 Panic Teaches Owners.
Copy-paste prompts
Use these with an AI assistant, and paste in your own documents and numbers. Each one tells the AI to flag what it can't determine. Check the output against your loan agreement and with your banker before acting.
Use this one before you call your banker about a renewal.
I run a [type of business] with $[revenue] annual revenue, a $[amount] bank line of credit renewing in [month], and these debts: [list]. I will paste my loan agreement and my latest financial statements below. Credit spreads are calm but stocks just fell into a correction. Draft a checklist and a script for a call with my banker to renew or extend my line now, including questions about covenants, term-out options and fees. Work only from the documents I paste. If a term or covenant is unclear or missing from my documents, say so instead of guessing.
Use this one to build the stress forecast.
Using these numbers, build a 13-week cash forecast template and a stress case where receivables slow by 15 days and revenue drops 10%: weekly sales [..], payroll [..], rent [..], debt payments [..], average days to collect [..]. Show the week my cash first goes below one payroll cycle. List every assumption you made, and tell me which inputs you are least sure about.
Use this one to rank which customers could hurt you.
Here is my customer list with revenue share per customer: [list]. Rank them by likely exposure to a tech and AI equity sell-off, estimate revenue at risk in a mild, moderate and severe case, and suggest three actions for each tier. Use only what I tell you about each customer. Where you would be guessing about a customer's industry or finances, say you don't know and tell me what to look up.
Use this one to pick your trigger numbers.
Explain in plain English what high-yield credit spreads, the VIX and negative real interest rates are, why a small business owner might watch them, and help me choose three trigger levels for my business with the action to take at each. Base the triggers on my own numbers below (revenue, debt, cash on hand, days to collect): [paste]. Say clearly that these are planning thresholds, not predictions, and note where you are unsure.
Use this one when you have floating-rate debt.
I have $[amount] in floating-rate loans; my loan documents are pasted below. Compare keeping floating, fixing now, or paying down with cash on hand under a scenario where rates rise one point and a scenario where they fall one point, and list the questions I should ask my lender. Show the arithmetic. If my documents don't say what prepayment or swap terms apply, tell me that instead of assuming.
Download the kit
The kit is a printable checklist and set of worksheets: a banker-call checklist and script, a 13-week cash-flow worksheet with a stress case, a customer and supplier exposure table, a trigger-and-action plan with example thresholds, a floating-versus-fixed comparison worksheet, and the pasteable prompts above. Get it here: Credit Calm Window Kit.
Fact-check notes and sources
- Apprised.news markets desk, Nov 22: SPY, QQQ and NVDA moves, the Morgan lens, and the conditions that would reverse the read. link
- Apprised.news markets desk, Nov 21: the 352 bps high-yield spread, 117 bps BBB spread, VIX of 28.44, fed funds of 3.88% versus CPI of 4.08%, and the roughly 400 bps stress threshold. link
- Apprised.news tech desk, Nov 21: the Morgan lens on AI-trade financing and the Nvidia drop after the reported SEC examination. link
- Apprised.news insurance desk, Nov 22: six December cat bonds still arriving on schedule through the Nasdaq's 5.28% drop. link
- Apprised.news health desk, Nov 21: hospital margins and the uninsured surge. link
- NPR: Nasdaq negative for 2026 and the S&P 500 selloff. link
- The Loadstar (CNBC report): S&P 500 and Dow in correction. link
- MarketScreener: reported SEC examination of potential accounting issues at Nvidia, based on unnamed sources. link
- Modern Ghana: Anthropic shares slide on second day of trading. link
- CoinDesk: Bitcoin bounce above $60,000 and the Monday stock-open test. link
- Bitcoin Magazine: Fed hike risk. link
- RealClearPolitics: Fed policy-error commentary. link
- gCaptain: tanker incidents near Oman and the Strait of Hormuz. link
- Worked examples: the HVAC, accounting and other business examples are hypothetical, built on planning assumptions, not data about any real firm.
Related reading
- Oil Near $100 on Hormuz Fear: What Morgan's 1907 Panic Teaches Owners: the same lens applied to an energy-cost shock.
- Your AI Vendor Can Lock You Out Tomorrow. Here Is a One-Page Continuity Plan.: the same "plan before you need it" habit, applied to a key supplier.
- Should You Build or Buy a House in 2026? The 25-Year Math: a worked example of how rate assumptions change a long-term decision.
- RAG and Agents Over Your Own Documents, Without the Hype (Part 4 of 5): how to point an AI at your own loan and cash documents, as the prompts above assume.
- The FBI Logged $20.9 Billion In Internet Crime Losses Last Year. Here Are Six Reflexes Small Business Owners Should Have Already Built.: worth reading before you change banking arrangements in a nervous market.
This post is informational, not legal, financial or security-consulting advice. It draws on AI-written Apprised.news analysis of public reporting; check the linked sources before acting. Mentions of third parties are nominative fair use. No affiliation is implied.