Earnings Recessions vs. Economic Recessions: How They Differ

An earnings recession and an economic recession are two different events that happen to share a word. Treating them as one is among the most common mistakes I see in how investors read the cycle, and it usually pushes in the wrong direction: toward assuming the economy is about to break because index profits have rolled over for a couple of quarters.

The two do overlap, and often. A CaixaBank Research note published in September 2016 counted eight episodes of falling S&P 500 earnings since the early 1970s and found that six of them coincided with the start of a US recession. That's a strong relationship. But the exceptions are the interesting part. In 1998, index earnings fell by more than 7% while the economy showed no sign of weakness. The 2015–2016 profit slump was still under way when that note went to press, and it ended without a recession too. Then came 2022–2023. Three profits-only declines since 1998 is not a rounding error.

So this is a comparison, not a forecast. What follows is what each kind of recession actually measures, how they connect to stock prices through two separate channels, two historical analogues where profits fell and the economy kept growing, and a short checklist for telling which one you're looking at while it's happening.

What Each Kind of Recession Actually Measures

An economic recession in the US is whatever the National Bureau of Economic Research's Business Cycle Dating Committee says it is. The committee looks for a significant decline in activity spread across the economy and lasting more than a few months, and it weighs employment, real income, real spending and industrial production. It is not the two-negative-quarters-of-GDP rule of thumb, and it is never quick. The committee dated the business-cycle peak of February 2020 in June 2020, and it didn't name April 2020 as the trough until July 2021.

An earnings recession has no committee at all. The working convention is two or more consecutive quarters of year-over-year declines in S&P 500 earnings per share. That's it. And the population behind that number is very different from the population behind GDP.

• Coverage. The NBER looks at the whole economy: small businesses, government, households. Index earnings come from 500 large listed companies, weighted toward the biggest earners, with a large share of their profits earned outside the US.
• Units. Recession dating leans on real (inflation-adjusted) measures. Earnings per share are reported in nominal dollars.
• Speed. Index earnings are known within weeks of a quarter's end as companies report. NBER turning points arrive months, sometimes more than a year, after the fact.
• Drivers. Recessions are driven by demand, employment and credit. Index earnings can also swing on things that barely register in GDP: a commodity price, the dollar, a write-down at a handful of very large companies.

The nominal-versus-real point deserves a second look, because it changes how the signal behaves across regimes. When inflation is running hot, companies with any pricing power can grow nominal earnings while real earnings shrink, so an earnings recession is harder to trigger even as the real economy softens. In a disinflationary regime the reverse holds. A modest real slowdown can print as an outright nominal decline because there's no inflation tailwind to hide it.

Two Channels From Macro to Stock Prices

Stocks respond to a downturn through two channels, and the difference between the two kinds of recession is mostly about which channel is doing the work.

The first is the earnings channel. In a true economic recession, revenue falls across most industries at once, and operating leverage turns that revenue decline into a much larger profit decline. Fixed costs don't shrink when sales do. That's why index earnings typically fall far more than GDP does in a recession, and why the 2007–2009 decline in as-reported S&P 500 earnings was so severe once write-downs at banks and insurers were added on top of the slump in operations.

The second is the discount-rate channel. Equity prices are a stream of future cash flows discounted back at some rate, and when that rate falls, the multiple investors will pay for a dollar of earnings tends to rise. I've written before about when the link between interest rates and equity multiples holds and when it breaks. In a genuine recession the Federal Reserve usually cuts, which partly cushions the multiple even as earnings fall. The two channels pull in opposite directions.

In an earnings-only recession, that cushion is often missing. The Fed isn't cutting because the economy doesn't need it. In both analogues below, the Fed was raising rates while profits were falling. So the two channels pointed the same way, which can make a profits-only slump feel worse in the market than its economic footprint would justify.

Economic recession: earnings fall broadly, the Fed usually eases, and the discount-rate channel partly offsets the earnings channel. Earnings-only recession: earnings fall narrowly, the Fed often holds or tightens, and the two channels reinforce each other.

What Happened the Last Time Profits Fell Without a Recession: 2015–2016

The historical analogue most people reach for is 2015–2016, and for good reason. It's the cleanest case of a profits recession with an identifiable cause and no economic contraction behind it.

The cause was oil. Crude fell from above $100 a barrel in mid-2014 to below $30 in early 2016, and the energy sector's profits went with it. In late February 2016, FactSet's Earnings Insight showed blended fourth-quarter 2015 S&P 500 earnings down 3.6% from a year earlier. Excluding energy, the same figure was up 2.5%. By September 2016, analysts polled by FactSet were bracing for a sixth straight quarter of year-over-year declines, which would have been the longest streak since FactSet began tracking the series in 2008.

The company filings show how concentrated the damage was. Halliburton's revenue fell from $32.87 billion in fiscal 2014 to $15.89 billion in fiscal 2016, a drop of 51.7% (15.89 ÷ 32.87 − 1), and its net income swung from a $3.50 billion profit to a $5.76 billion loss, according to Halliburton's 10-Ks for fiscal 2014 and fiscal 2016. Exxon Mobil's net income went from $32.52 billion in fiscal 2014 to $7.84 billion in fiscal 2016, down 75.9%, per Exxon Mobil's 10-Ks for those years. And the pain reached past the oil patch into the companies that sell equipment to it: Caterpillar's revenue fell from $55.18 billion in fiscal 2014 to $38.54 billion in fiscal 2016, down 30.2%, according to Caterpillar's 10-Ks for fiscal 2014 and fiscal 2016.

Now set that against the economy. The NBER dated no recession. The unemployment rate kept drifting down through both years. Consumers were, if anything, helped by cheaper gasoline. The Fed raised rates for the first time in almost a decade in December 2015, a move it would not have made into a contracting economy, and then waited a full year before raising again.

The market did notice. The S&P 500 went through two sharp corrections, in August 2015 and again in early 2016, but it never fell into a 20% bear market, and earnings growth resumed before the end of 2016. An investor who read the earnings recession as an economic one and sold on that basis was selling into a sector problem, not a cycle turn.

The Second Analogue: 2022–2023, When the Fed Was the Cause

The more recent case had a different mechanism, which is exactly why it's worth putting next to 2015–2016. FactSet's Earnings Insight of September 1, 2023 reported blended second-quarter 2023 S&P 500 earnings down 4.1% year over year, the third straight quarter of declines. Again the NBER dated no recession.

This time the cause wasn't one sector's commodity price. It was the fastest Fed tightening in four decades running into margins that had been inflated by 2021's demand surge, so the decline was broader than in 2016 and shallower in the sectors that drove it. And because inflation was high, the nominal print understated the damage. A 4.1% nominal decline in a period of 3%-plus inflation is a meaningfully larger decline in real terms.

The order of events is what I'd underline. The S&P 500 fell roughly a quarter from its January 2022 high to its October 2022 low, which means most of the price damage happened before the earnings decline was even finished. That was the discount-rate channel doing its work first. Prices moved on the rate path; profits caught up later. The earnings recession, when it came, was close to old news. It's the same sequencing I described in how Fed policy regimes shape sector leadership, and the credit side of that story, the slow fuse that runs from higher rates to defaults, is covered in how bankruptcy waves cluster at the end of tightening cycles.

But I'd be careful about turning two cases into a rule. The 2022–2023 episode might look different in a decade, once the full credit cycle has played out. I'm less sure about the clean "no recession" verdict on that period than I'd like to be, simply because the lagged effects of a tightening cycle that large are still being counted.

Telling Them Apart in Real Time

You won't get the NBER's verdict in time to use it. What you can do is ask a handful of questions about the earnings decline itself, each of which leans one way or the other.

1. How broad is it? Recompute the index earnings change without its largest losing sector. In February 2016 that single step turned a 3.6% decline into 2.5% growth. A decline that survives the exclusion is far more likely to be economic.
2. Is it revenue or margin? Falling revenue across most sectors points to weak demand. Steady revenue with compressed margins points to costs, rates or a specific input price.
3. What is the labor market doing? Earnings can fall without job losses. Recessions almost never happen without them. Employment data and the yield curve, whose record I went through in [the yield curve as a recession signal](/blog/yield-curve-recession-signal), are better recession evidence than profits alone.
4. Which way is the Fed moving? A Fed still raising rates into an earnings decline is telling you it doesn't see a contraction, and it also means the discount-rate channel offers no cushion.
5. Nominal or real? Adjust for inflation before deciding how bad a decline is, especially when inflation sits well above 2%.

Concentration works in both directions, too. FactSet's Earnings Insight of October 16, 2026 put blended third-quarter 2026 S&P 500 earnings growth at 29.0%, and about 20.0% excluding just two companies, Micron Technology and NVIDIA. An index-level number that leans that heavily on a few names can swing sharply on their results alone. The breadth test that explained 2016 applies on the way up as well: a headline growth rate tells you less about the economy when a small group of companies is producing a large share of it.

Key Takeaways

• An earnings recession (two or more quarters of falling S&P 500 earnings) and an economic recession (an NBER-dated contraction across the economy) often coincide, but profits have fallen without a recession at least three times since 1998.
• The key difference is the channel. In a real recession the Fed usually eases and lower rates partly offset falling earnings. In a profits-only recession the Fed often keeps tightening, so both channels hit stocks at once.
• Breadth is the fastest test. In 2015–2016, energy explained the whole index decline: Halliburton's revenue fell by more than half between fiscal 2014 and fiscal 2016 while the economy kept growing.
• Read the labor market and the Fed alongside profits, and adjust for inflation. A falling nominal earnings line on its own is weak evidence that the economy is turning.