I have spent four decades in the recruiting business. I am Joe Pelayo, an executive recruiter and the managing partner of Joseph Michaels International Recruiters. I started in this work in 1986, opened the firm in 1990, and I have spent the years since then on both sides of the table: companies trying to hire, and people trying to land the next role. That is why I do not treat the next recession as an abstract chart. I treat it as a hiring freeze that starts quietly, then becomes a wall.
The next recession will not wait for a committee to finish its briefing book. Rate policy that pretends otherwise is how you turn a slowdown into a wreck.
Monetary policy works with a lag. That is not a slogan. It is the least controversial fact in central banking, and it is the one officials keep treating as optional. By the time unemployment is obviously rising, hiring has already frozen, credit has already tightened, and firms have already decided that the cheapest way to protect margins is to stop replacing people. Cutting after that point is not “support.” It is cleanup. Cleanup is more expensive, more political, and less effective than prevention.
I have watched that sequence in every cycle since the late 1980s. The official recession date is a rearview mirror. The real date is when a hiring manager calls and says the search is on hold, or when a strong candidate who would have had three offers last year cannot get a second interview. If you wait for the government to name the downturn, the damage in the labor market is already underway.
That is the ordinary case for moving early. This cycle has an extra feature that makes waiting even dumber: AI is compressing the time between “this job is exposed” and “this job is gone,” and we do not have a training system that can keep up.
The lag is the policy
A recession is not a light that flips on. It is a sequence. Orders slow. Inventories look heavy. Managers freeze headcount. Then they cut contractors. Then they stop backfilling. Then the official data finally admits what workers already knew.
If you wait for the National Bureau of Economic Research to stamp the file, you have already wasted the window where cheaper credit could have kept investment, housing, and new firm formation from collapsing at the same time households lose hours. The 2008 lesson was not “never tighten.” It was that late recognition plus late easing turns a financial shock into a multi year labor scar. The 2020 lesson was that speed matters more than elegance when the demand side falls out.
The current posture is the opposite of speed. The Fed just raised the funds rate to 3.75 to 4 percent, with inflation still above target and officials talking about another hike. Growth is not collapsing today. Unemployment is still around 4.1 percent. That is exactly when institutions get overconfident. They look at the lagging indicators, declare the labor market “resilient,” and treat preemptive easing as a moral failure.
Resilience is not the same thing as spare capacity for a shock. Household interest burdens are already high. Mortgage rates have made housing a gated community. Non mortgage interest costs have been flashing the same warning they flashed before previous downturns. You can hike into that if you are convinced inflation is the only risk that counts. You should not pretend the employment mandate will wait politely until the PCE print cooperates.
This downturn will not be a clean demand story
The next recession, if it arrives on schedule with history, will land on top of a labor market that is already being rearranged by software that improves faster than people change careers.
That is the part the “just retrain them” crowd refuses to price. I hear the slogan in conference rooms. I also hear what companies actually ask for when they retain a search: more output, fewer seats, and a candidate who can already do the job with the new tools on day one. They are not budgeting two years for a national upskilling campaign. They are trying to fill a role this quarter.
AI is not a single invention that arrives, sits still, and waits for community colleges to rewrite their catalogs. Capability is compounding. Tools that were novelty two years ago are now default workflow. Firms do not need a sci fi takeover to change hiring. They need a cheaper way to do the same output with fewer junior seats, fewer back office hours, and a higher bar for the remaining humans. That is already showing up where you would expect it: slower wage growth in exposed occupations, weaker entry level hiring, and concentrated pain among people who were supposed to be the pipeline.
The official unemployment rate can look fine while the on ramp collapses. Recent graduates in exposed fields feel it first. Experienced workers feel it next, when the firm decides the work can be supervised rather than staffed. By the time the aggregate number moves, the matching problem is already ugly: vacancies in the new roles, surplus people from the old ones, and a gap in skills, geography, and credentials that a 25 basis point speech cannot close.
After four decades of placing managers, directors, and executives, I can tell you what that gap looks like in real time. A company will pay a premium for the person who already lives in the new stack. The same company will pass on a loyal, competent professional whose tasks just got absorbed, not because that person is worthless, but because the firm does not have the runway to retrain while customers, lenders, and boards are watching the next quarter. Goldman style estimates of displacement over a decade sound manageable in a spreadsheet. They are not manageable if the displacement arrives in bunches during a downturn. Historically, people knocked out by automation take longer to get back and take a bigger real pay cut when they do. Recessions make that worse. Combine a cyclical hiring freeze with a structural substitution shock and you do not get “creative destruction.” You get a cohort that never quite recovers.
You cannot train a country at the speed of the model
The popular answer is always the same: upskill. Teach everyone AI. Make prompt engineering a civic duty. Convert every displaced clerk into an “AI enabled knowledge worker.”
This is a slogan dressed up as a plan.
Training at national scale is slow, uneven, and bottlenecked by the same institutions that were already mediocre at ordinary career transitions. People have rent, kids, shifts, and cognitive load. Many of the jobs being thinned out were the jobs that used to teach the next job. You cannot pull the bottom rungs off the ladder and then lecture the people standing on the ground about lifelong learning.
I have watched career changes for forty years. A serious transition is not a weekend course. It is months of lost income, a damaged story on the resume, a relocation the family cannot afford, and a hiring manager who wants proof, not potential. Even strong people need time and a market willing to take a chance. A tight, frightened market does not take chances. It hires the safest resume and leaves everyone else outside.
Even the optimistic forecasts assume massive churn: tens of millions of roles changing, not a tidy swap of old titles for new ones. A 22 percent global workforce churn number is not a TED Talk. It is a statement that a huge share of people will have to change what they do while the target keeps moving. The skills premium for people who already have the new stack is real. That is exactly why the people without it fall behind faster.
There is also a timing problem nobody in the “reskilling will save us” camp wants to say out loud. The half life of a course is now shorter than the enrollment cycle. By the time a state workforce board procures a vendor, designs a certificate, and graduates a cohort, the tooling has changed. Firms will not pause adoption so the syllabus can catch up. They will adopt, measure output, and stop hiring for the tasks the model absorbed.
So the claim that we can train everyone in parallel with AI growth is not ambitious. It is false. Some people will adapt quickly. A larger group will adapt slowly. A large group will not adapt on any timeline that prevents lost years of income. If policy assumes the first group is the whole population, it is planning to leave the rest behind and then act surprised when politics gets ugly.
What a preemptive cut is actually for
A rate cut in advance of a recession is not a magic shield against automation. Interest rates do not teach anyone linear algebra. Anyone selling that is lying.
What cheaper money can do is keep the transition from happening in a credit crunch.
It can keep housing from locking people in the wrong cities. It can keep small firms from dying before they become the employers of record for the new work. It can keep capital expenditure from becoming an exclusive club of the ten companies already building the data centers. It can keep household balance sheets from snapping when hours fall. It can make it less suicidal for a worker to take time out of the market to switch occupations.
That last point is the one I see most clearly as a recruiter. People do not reinvent themselves while they are drowning. They reinvent themselves when they can still make the rent, still get a call returned, and still believe a good company will take a meeting. A preemptive cut does not create that by itself. It buys the slack in which planning can work.
That is the adult version of “plan this out.” Planning is not a press release about STEM camps. Planning is buying time and slack so the people who will be displaced are not also being crushed by high debt service, frozen hiring, and a central bank that waited for confirmation.
The inflation objection is real. Cutting into sticky prices and an energy shock is not free. If you ease sloppily, you can re anchor expectations the wrong way and spend the next decade paying for it. That is not an argument for doing nothing until the unemployment rate has already done the ugly work. It is an argument for looking through one off supply noise, watching real labor demand rather than last month’s headline, and moving before the financial system starts doing the Fed’s job for it by choking credit.
There is a difference between “ignore inflation” and “do not hike into a labor market transition you cannot retrain your way out of in real time.” The first is reckless. The second is the dual mandate taking the second half of the sentence seriously.
Leave people behind and the politics will not stay technical
If AI growth continues and policy treats displaced workers as a rounding error until the recession is official, the backlash will not be a thoughtful debate about productivity. It will be a demand to stop the technology, freeze hiring of models, or turn the state into an employer of last resort after the damage is done.
That is the worst of both worlds: slower growth and a more frightened labor market.
The better sequence is obvious and unfashionable. Ease before the contraction is undeniable. Accept that mass instant retraining is a fantasy. Build actual transition capacity while people still have jobs: portable benefits, faster credentialing that employers recognize, relocation that housing supply makes possible, and a labor market loose enough that a 40 year old whose tasks just vanished can find the next role without a two year exile.
I have spent my working life trying to make those matches. The matches get harder when money is tight, companies are scared, and the skill target is moving every quarter. You will not get a decent outcome by waiting for the recession to introduce itself. Recessions are late, rude, and uninterested in your training montage.
Cut early enough that the landing has a runway. Plan as if millions of people will not make the jump on the first try, because they won’t. I have seen enough careers, enough hiring freezes, and enough “we will retrain later” promises to know how that story ends. The alternative is to discover, after the fact, that “everyone will just learn AI” was a lullaby for people who were never going to be the ones left standing outside the new economy.
Joe Pelayo, C.P.C.
Managing Partner
Joseph Michaels International
Email: jpelayo@josephmichaels.com
Website: https://josephmichaels.com/
“One of the Top 75 Recruiters in the United States.” – Recruiter Life Magazine. Bringing Together Great Talent and Growing Companies in the USA and around the World for over three decades.