"History doesn't repeat… but it rhymes." — Mark Twain
◉ THE PRESENT
The September payrolls report hits the tape this morning with Wall Street expecting 84,000 new jobs — half the 162,000 that shocked the market five weeks ago. The number matters, but the setting matters more. The 10-year Treasury yield touched 5.34% yesterday before settling at 5.24%, the highest level since 2007, while the S&P 500 hovers at 7,666 and consumer confidence sits at a twelve-year low of 81.9. The Fed just raised rates to 3.75–4.00% and is weighing whether to go again in October.
NFP consensus: 84K | Prior: 162K | 10Y: 5.24% | Fed funds: 3.75–4.00% | Brent: $102 | S&P 500: 7,666 | UE: 4.1%
The last time a key payrolls number showed this kind of month-over-month drop while yields sat above 5% and the Fed was still hiking, it was Friday, June 2nd, 2006. Almost nobody remembers what happened that day. What happened afterward is harder to forget.
◉ THE ECHO — AUGUST 25, 1987
The Morning Nobody Believed
The traders on the big desks in lower Manhattan had it figured out. The economy was running at 3% growth. Housing prices were still climbing in Phoenix and Miami and Las Vegas. Ben Bernanke — four months into the job as Fed chairman, still trying to prove he was no softer than Greenspan — had just lifted rates to 5.00% on May 10th. Everyone expected another solid payrolls number, somewhere around 170,000. The screens at 8:30 that Friday morning showed 75,000.
Less than half what the Street had penciled in. S&P futures dropped in the first sixty seconds. The 10-year yield, which had been sitting around 5.11%, fell eight basis points by lunch. Oil was above $71 a barrel and climbing. For about three hours that morning, it looked like the American economy had a problem nobody had noticed.
Then the rationalizing started. April had come in at 126,000. The three-month average was still respectable. GDP was strong. Corporate earnings were beating estimates left and right. And the wage data in the report was cool — average hourly earnings rose just 0.1%, which meant the Fed didn't need to worry about a wage-price spiral. By noon the selling had stopped. By the close, the S&P had steadied at 1,288. The market had decided the weak number was noise.
The Fed agreed. Four weeks later, on June 29th, Bernanke raised rates to 5.25% in a unanimous vote. Ten to zero. The FOMC statement said economic growth was "moderating" but core inflation was "elevated." The weak payroll number didn't rate a mention. It was the last rate hike of the entire cycle, and nobody at the table knew it.
What nobody was watching — what nobody wanted to watch — was the housing market. The S&P/Case-Shiller National Home Price Index would later show that nationwide prices peaked that exact summer, July 2006. Housing starts had already rolled over from their January high of 2.27 million units. The personal savings rate had gone negative the year before for the first time since the Great Depression. Americans were spending more than they earned, and the collateral backing that spending was a house whose value had just stopped going up. The weak payrolls number on June 2nd was the thermometer. The disease was leverage, stacked eighteen layers deep in securities that almost nobody outside a handful of trading desks fully understood.
◉ THE RHYME — WHAT'S IDENTICAL
Both moments sit at the exact same crossroads: the labor market is flashing yellow, yields are at multi-year highs, and the stock market is refusing to blink.
◉ THE DIVERGENCE — WHAT'S DIFFERENT THIS TIME
The bubble is in a different room. In 2006, the overleveraged asset was housing — something that touches every household, every bank balance sheet, every county assessor's office in the country. When housing cracked, it pulled the entire consumer economy down with it. The potential 2026 excess is in AI infrastructure — data centers, chips, power contracts. If AI capex disappoints, it hits corporate balance sheets and equity valuations first. The transmission to Main Street is narrower, which could mean the damage is more contained. Or it could mean it hides longer before anyone feels it.
The Fed is lower in real terms. Bernanke's 5.00% rate in June 2006 sat above core PCE of 2.1%, giving him a real rate close to 3%. Today's 3.75–4.00% against a 3.0% core PCE leaves the real rate near 1%. The Fed has less weight pressing down on the economy right now, but it also has less room to cut if something breaks. Whether that's a cushion or a trap depends on what breaks first.
Oil is rising for different reasons. The 2006 oil climb was driven by Chinese and Indian demand — a structural force that took years to play out. Today's oil spike comes from the Iran-Hormuz supply threat, which is geopolitical. Supply shocks can reverse in a weekend if a diplomat picks up the phone. But they can also spiral in ways that demand cycles don't, because nobody can forecast a missile strike.
The plumbing risk has moved. In 2006, the dangerous leverage was hiding in mortgage-backed securities and CDOs that even their creators struggled to value. In 2026, the risks that keep people up at night — the trillion-dollar Treasury basis trade, the concentration of AI capex debt — are more visible. Risk you can see is easier to manage. Unless seeing it makes you think someone else is managing it.
◉ THE RECKONING — WHAT HAPPENS NEXT
After the weak June 2006 payrolls print, the S&P 500 sold off to 1,220 by June 14th — an 8% correction from its May 5th high of 1,326. A nasty two weeks. Then something happened that always happens after the first crack: the market turned around and ran. From that June low, the S&P climbed for sixteen months, all the way to 1,565 on October 9th, 2007. Twenty-eight percent. Anyone who panicked on the weak jobs number and sold missed one of the best rallies of the decade.
But here's the rest of the story — the part that matters. Housing prices peaked in July 2006 and started falling. Quietly at first, then not quietly at all. Subprime mortgage delinquencies doubled between mid-2005 and mid-2007. In June 2007, Bear Stearns had to rescue one of its two hedge funds that had loaded up on subprime CDOs. In August, BNP Paribas froze three money market funds, locking investors out of $2.2 billion because the bank couldn't figure out what the assets were worth. By September, Northern Rock was experiencing the first bank run in Britain in over a century. Bernanke finally cut rates on September 18th, 2007 — fifteen months after that last hike — but by then the fire was already behind the walls.
The recession officially started in December 2007. The S&P peaked at 1,565 and fell to 676 by March 2009. A 57% decline. The round trip from the last hike to the market bottom took thirty-three months.
The pattern is clean: weak payrolls, the Fed ignores the signal, the market rallies, and the real problem shows up twelve to eighteen months later, when it's too late to get in front of it.
The smart money in mid-2006 didn't sell stocks on the weak jobs report. They kept riding the rally. But they started buying protection — put options on homebuilder stocks, credit default swaps on subprime mortgage bonds. Michael Burry had begun putting on those trades more than a year earlier, in mid-2005, when the cost of insurance was still cheap because everyone believed the weakness was temporary. Those positions eventually paid off roughly seven to one.
The pattern says the market can rally from here — maybe for months. But the pattern also says the thing to watch isn't the payrolls number itself. It's the leverage underneath: $600 billion in AI infrastructure spending financed at yields above 5%, and a trillion-dollar Treasury basis trade that assumes volatility stays low. Protection is cheapest when nobody thinks they need it. Right now, nobody thinks they need it.
◉ TOMORROW’S WATCH
Monday's ISM Services PMI for September will test whether the services economy — where most Americans actually work — is absorbing the oil shock or starting to crack under it. In June 2006, the ISM Services index held above 57 even after the weak payrolls print, buying the market sixteen months of false confidence before the real damage arrived.
*Disclaimer: This is a paid advertisement for BluSky AI Regulation A offering. Please read the offering circular at invest.bluskyaidatacenters.com
