Recession fears are escalating alongside a tremendous level of government policy uncertainty, specifically regarding tariffs and DOGE spending cuts. The stock market has taken notice, with a correction underway among a couple of key indexes. There are a lot of threads to pull here, but let's start with a broader look at the history of cycles—including across markets and the economy.

A few years ago, in the midst of 2022's bear market in stocks, we put together the graphic below that shows every recession (orange bars) and equity bear market (blue bars) in the post-WWII era. The individual boxes represent monthly increments, and the date range to the left represents the entire cycle—from either the beginning of the bear market or the beginning of the recession, to the final completing of the cycle(s).

As shown, recessions and bear markets don't always overlap. There were four recessions—1945, 1953-54, 1960-61, and 1980-81—that did not have an overlapping bear market (although there was one in short order in 1981). There were also four bear markets that did not overlap with recessions: 1946-47, late-1961-62, 1966, and 1987.

The visual above also highlights the lagged nature of the declaration of recessions' start (red boxes) and end (green boxes) dates. Since 1978, the official arbiter of recessions has been the National Bureau of Economic Research (NBER), which provides start/end dates by month, always in retrospect. In contrast to conventional wisdom, the NBER's definition is not two consecutive quarters of negative gross domestic product (GDP) readings.

The actual definition per the NBER is "a significant decline in economic activity that is spread across the economy and that lasts more than a few months." The view of NBER's Business Cycle Dating Committee (BCDC) is that "while each of the three criteria—depth, diffusion, and duration—needs to be met individually to some degree, extreme conditions revealed by one criterion may partially offset weaker indications from another."

To determine the months of peaks and troughs, the NBER's BCDC specifically monitors "real personal income less transfers, nonfarm payroll employment, real personal consumption expenditures, wholesale-retail sales adjusted for price changes, employment as measured by the [Bureau of Labor Statistics'] household survey, and industrial production." The NBER has "no fixed rule about what measures contribute information to the process of how they are weighted."

The indicators the NBER monitors are a combination of coincident and lagging economic indicators. When the NBER declares recessions, they also simultaneously announce the start month (red boxes in the visual above)—with an average historical lag of seven months. Once the NBER declares recessions as complete, they also simultaneously announce the end month (green boxes in the visual above)—with an average lag of a whopping 15 months!

The moral of the story is that by the time the NBER announces recessions' starts, they have either already been well underway, or at times already over; while by the time they announce recessions' ends, recoveries were already well underway. In fact, in both 1991 and 2020, when the NBER announced the start of those recessions, it turned out they were already over.

Recessions have some patterns, but each one is different, especially in sequencing. The causes (and effects) can vary as well. Early warnings signs include slowing economic growth due to factors like rising interest rates, inflation, or external shocks. Business and consumer confidence typically declines, and the stock market generally becomes more volatile/weaker. Typically, it's a key crisis event that pushes the economy into a recession.

For visual representation, a number of years ago we created a dominoes chart showing how recessions typically unfold. The version below references the Global Financial Crisis cycle. We opted not to show how things unfolded during the last recession since it was pandemic-driven, and therefore not a "traditional" array of triggering events.

In terms of weakness so far in this cycle, we have seen the following dominoes take a hit:

• Stocks fall (though only in correction territory so far)
• Consumer confidence falls
• Price-to-earnings ratios (P/Es) fall (tied to the correction)
• Volatility rises

When thinking of the dominoes that may fall in the current cycle, we clearly need to add in a domino for inflation turning back higher. Also key to watch include purchasing managers indexes (PMIs), financial conditions, and the earnings outlook. Today, we are in the midst of a crisis of uncertainty—specifically regarding government policy. The trade policy uncertainty index, shown below, is up on a parabolic spike.

Uncertainty has dented "animal spirits" and "soft" (survey-based) economic data, but it's starting to hit some of the "hard" data as well. Citi's widely watched Economic Surprise Index, having improved markedly between last summer and immediately post-election, has moved back down and now sits in net-negative territory.

The widely watched GDPNow "nowcast" from the Atlanta Federal Reserve is deep in negative territory in terms of how first quarter gross domestic product (GDP) is tracking, as shown below. This is in part due to weaker consumer spending trends so far this quarter, but also a significant spike in imports relative to exports—clearly reflecting front-running of tariffs before they were to take place. Caveat: Some of the spike in imports was of non-monetary gold (gold not held as a reserve asset), which is not driven by tariff concerns, so reality is likely milder than what's shown here.

The number of "stagflation" mentions in news stories is soaring as well, as shown below.

The Challenger Gray & Christmas job-cut announcements tally jumped to more than 100% year-over-year, as shown below.

Over the past two months, the percentage of small businesses saying now is a good time to expand has fallen by eight percentage points…among the largest declines in the National Federation of Independent Business (NFIB) survey's history. The drop has quickly reversed part of the post-election surge, which is in contrast to optimism remaining high in the aftermath of the 2016 election (until the pandemic erupted).

Alongside growth worries, the 10-year Treasury bond yield moved down to a recent low of 4.16% from a mid-January high of 4.8%. Up until recently, bond yield moves were tied more to inflation risk's ups and downs; but this latest downtrend is more about growth concerns. Lower yields reflecting lower inflation would typically mean better stock market performance; but that has not been the case over the past month. As shown below, the correlation between bond yields and stock prices recently shifted from inverse to positive.

Shortly after the peak in bond yields, equities began their descent. Eerily, at the onset of the pandemic, stocks peaked on February 19, 2020, and exactly five years later, stocks peaked on February 19, 2025. Shown below are the maximum drawdowns for three key indexes, both at the index level, but importantly also at the average member level. At the index level, the S&P 500, Nasdaq, and Russell 2000 are all in correction territory. In terms of the average member within each index, we are in correction and/or bear market territory per those maximum drawdowns.

In our 2025 outlook, we mentioned how we often chuckle at the notion of "markets hating uncertainty," as if there are ever truly certain times. Not even three months into the year, and it's quite clear that markets, businesses, and consumers are not embracing the uncertain backdrop in which we find ourselves. Much of that has already been reflected in confidence metrics which, admittedly, went through a similar soft patch a couple years ago, giving a "false" recessionary signal. The catch at the time, though, was that underlying growth and the hard data were resilient; it truly was a "vibecession." Fast forward to today, and the "vibespansion" that was evident in so many sentiment metrics just a few months ago looks to be receding quickly. If the soft data stay soft long enough, we see more risk of the hard data catching down. The only thing we can be certain of is that things will remain uncertain for the time being.