The ClearBridge Recession Risk Dashboard remains firmly in green expansionary territory; despite AI labor fears, supportive fiscal and monetary policy should power the U.S. economy and corporate earnings in 2026.
Anatomy of a Recession
The Long View: Party Like It's (Not) 1999
Key Takeaways
- Despite bubble fears, there are several key differences between the late 1990s and today that we believe bode well for U.S. equities in the year to come.
- A pickup in productivity from AI, combined with moderating wage gains, a softening labor market, weaker shelter prices and lower commodity costs could all push inflation lower as we move deeper into 2026.
- While valuations are elevated, we believe equities will “grow into the multiple” in 2026 with strong earnings fueled by ongoing AI capex strength as well as fiscal and monetary stimulus.
“Party like it’s 1999” is a phrase made famous by the musician Prince’s 1982 song, which experienced a renaissance amid Y2K fears and has since entered the lexicon meaning to celebrate intensely because the future is uncertain. The phrase seems apt to describe the current investment landscape, given the similarities between the late 1990s dot-com bubble and today. These include lofty valuations, strong market momentum and a focus on growth stocks. However, we believe there are several key differences between the present environment and 1999 that will keep the markets moving higher in the year to come, which may be a surprise to many.
1999 vs. 2025: Economic Differences
The first key difference is that the U.S. economy is poised to benefit from both fiscal and monetary stimulus in 2026, a potent combination typically only seen coming out of recessions. The net impulse from the One Big Beautiful Bill (OBBB) is expected to deliver ~1% of GDP this year with supercharged tax refunds providing support to low- and middle-income households.
Tax refunds are typically spent rather than saved, suggesting that much of this cash will make its way back into the economy relatively quickly. The COVID stimulus payments provide a good example of this dynamic, with research from the Peter G. Peterson Foundation showing that households earning below $75,000 spent around 80% of the initial stimulus payments they received. That figure dropped but stayed above 50% for households with over $150,000 in earnings. While the 2026 tax refund bonanza will likely fade in the second half of 2026, the OBBB’s fiscal impact should continue in 2027 and 2028 but at lower levels of support around 0.5% of GDP according to Congressional Budget Office and Wolfe Research estimates.
Exhibit 1: Tax Tailwind
Sources: Wolfe Research, CBO, Macrobond. Based on CBO’s Baseline Budget Projections from January 2025’s report The Budget and Economic Outlook: 2025 to 2035. There is no assurance that any estimate, forecast or projection will be realized.
The benefit from this stimulus should be significant because wage growth — the largest source of spending power for most Americans — has continued to moderate following the post-pandemic spike. Although this moderation has strained lower income cohorts and led to the “K-shaped” economy, it is somewhat encouraging from a macro perspective. Typically, maturing economic expansions see accelerating wages that often spook the Fed into tightening to prevent a wage-price inflationary spiral. This, in turn, can choke off economic growth and help set the stage for a recession. However, this dynamic is not in place today, which marks a second key difference between the present and 1999.
Exhibit 2: Wage Trend, Expansion’s Friend
Gray shading marks recessionary periods. As of Dec. 16, 2025, latest available as of Dec. 31, 2025. Sources: US Bureau of Labor Statistics (BLS), NBER, Macrobond.
Cooling wages have resulted in the Wage Growth indicator on the ClearBridge Recession Risk Dashboard remaining in solid green territory, which is also where the overall signal continues to reside. The dashboard did see two changes in December: ISM New Orders dropped back to red after the second consecutive monthly reading below 48, while the Yield Curve improved to green as it steepened above 50 basis points. Housing Permits and Retail Sales remain on hold due to lingering delays related to the U.S. government shutdown in October and November, but these issues should hopefully be resolved in the coming weeks.
Exhibit 3: U.S. Recession Dashboard
‘Not Available’ reflects data that has not been updated due to the government shutdown. Data as of Dec. 31, 2025. Sources: BLS, Federal Reserve, Census Bureau, ISM, BEA, American Chemistry Council, American Trucking Association, Conference Board, Bloomberg, CME, FactSet and Macrobond. The ClearBridge Recession Risk Dashboard was created in January 2016. References to the signals it would have sent in the years prior to January 2016 are based on how the underlying data was reflected in the component indicators at the time.
The trend of moderating wage gains over the past few years has stood in stark contrast to an economy that has continued to deliver solid growth. Gross domestic product (GDP) has grown by an average of 2.8% on a real basis since the end of the first quarter of 2023. This momentum has shown little sign of slowing recently, with the third quarter of 2025 coming in at 4.3%. At the same time, the unemployment rate has risen over a full percentage point to 4.6% from the 3.4% lows in April 2023. As a result, some observers are blaming the rise of artificial intelligence (AI) for recent labor market weakness.
AI does appear to be contributing toward softer hiring in the technology industry and for entry-level roles in particular. However, a broader review shows that occupations where AI adoption is the lowest have seen the greatest rise in unemployment, suggesting that other factors have been driving labor weakness. Job growth for industries more rapidly adopting AI is actually positive, driven by AI augmentation as opposed to substitution — a dynamic we explored in last month’s blog.
Exhibit 4: Lower AI Adoption, Weaker Labor
Data as of Nov. 30, 2025; latest available as of Dec. 31, 2025. Sources: Wolfe Research, Haver Analytics.
AI adoption has been advancing at an extremely fast pace compared to past innovations. Positive technological breakthroughs typically lead to a pickup in productivity and a drop in inflation (or even deflation). This dynamic usually takes time to bear out as was the case during the 1990s. Given the rapid adoption of AI over the past few years, however, we believe there’s a strong possibility this lag could be compressed. If this proves true, the risk to inflation could end up being to the downside, not the upside, in 2026.
The “January effect” is likely to be larger than normal once again this year. However, a pickup in productivity combined with moderating wage gains, a softening labor market, weaker shelter prices and lower commodity costs could all push inflation lower as we move deeper into 2026. Further disinflation — the aforementioned scenario largely represents a continuation of 2025 trends — would likely be bullish for financial markets as it would allow for further Fed easing should employment growth remain lackluster (50k-75k jobs per month).
Another 2025 trend likely to continue in the new year is the rapid growth of AI capital expenditures (capex) as the AI infrastructure buildout continues. Despite fears that AI capex has reached bubble territory, current levels of spending are well below the peak seen during prior innovative technological cycles in the U.S., as a percentage of GDP. For example, AI investment accounts for about 1% of the U.S. economy today compared to 3% during the late 1990s tech/telecom bubble. Should history repeat, AI capex could surprise to the upside in the years to come, providing continued support for both the economy and markets.
Exhibit 5: More AI Capex Ahead?
Data as of Oct. 20, 2025; latest available as of Dec. 31, 2025. Source: Bureau of Economic Analysis, Goldman Sachs Global Investment Research.
1999 vs. 2025: Market Differences
The funding source of this capex brings us to the third key difference between the late 1990s and today. Today’s spending is largely financed from corporate free cash flow (FCF), whereas the tech/telecom capex buildout was primarily underwritten in the capital markets through debt and/or equity issuance. Recently, smaller players have begun to increasingly tap debt markets and even the hyperscalers have begun to dip a toe. However, the tech sector’s aggregate capex spending equates to under 40% of FCF, well below the mid-1990s peak of 67%. With today’s leaders flush with cash, so far there has been less of a need to raise capital to fund the AI buildout. Put differently, debt financing will likely become more prevalent in 2026 (and beyond), but we are not currently near concerning levels.
Exhibit 6: Similar Capex Boom, Different Funding
Data as of Dec. 31, 2025. Sources: Datastream, Goldman Sachs Global Investment Research.
A fourth key distinction between the late 1990s and 2025 is what has been powering the equity market higher. During the final surge of the dot-com bubble, P/E expansion was the primary driver of upside, a stark difference from 2025’s improving EPS expectations-powered rally. Within the technology sector specifically, this year’s 23.3% price move was more than fully driven by improving EPS expectations, while multiples contracted and actually detracted from returns. Similarly, over 80% of the S&P 500 Index’s price return in 2025 was driven by improving fundamentals (aka earnings). This illustrates that market participants today are engaging in less speculative behavior as compared with the heyday of the dot-com bubble.
Exhibit 7: Party Like It’s (Not) 1999
Note: Dot Com Peak was March 23, 2000; Contribution to price return based on change in sell-side consensus NTM EPS expectations and NTM P/E. Data as of Dec. 31, 2025. Sources: S&P, FactSet.
An additional sign that investors are behaving less exuberantly than in past bubbles comes from the lower valuation multiples assigned to today’s market darlings. The Magnificent Seven currently trade at 53.3x as a group, in the ballpark of what was seen during the peak of the Nifty Fifty and dot-com bubbles. However, a large portion of the current valuation is driven by Tesla; a “Magnificent Six” or Magnificent Seven ex-Tesla trades at a less lofty 27.4x, which bears far less resemblance to former speculative manias.
Exhibit 8: Are Valuations At Bubble Levels?
*Actual P/E ratios; forward P/Es unavailable for this period. Data as of Dec. 31, 2025. Sources: FactSet, Valuing Growth Stocks: Revisiting the Nifty Fifty; AAII, October 1998, Goldman Sachs. Company references are used for illustrative purposes and should not be construed as an endorsement of sponsorship of Franklin Templeton companies. This information is not intended as an investment recommendation, nor does it constitute investment advice.
Part of this excitement may be due to the Magnificent Seven’s superb earnings over the past three years. However, we continue to anticipate a rotation in leadership as earnings delivery broadens in 2026. Third-quarter earnings season was encouraging on this front, with U.S. companies largely reporting robust earnings and equities beginning to price this dynamic during the fourth quarter of this year. Given the hype around AI, it may surprise some to learn that, among the Magnificent Seven, only Google and Nvidia outperformed the S&P 500 in 2025. We also believe there is further room to run for this trade, which should be a tailwind for active stock pickers who can navigate the concentration risks associated with these recent winners.
Exhibit 9: Mag 7 Losing Altitude
Data as of Dec. 31, 2025. Sources: FactSet, S&P. Magnificent 7 data refers to the following set of stocks: Microsoft (MSFT), Amazon (AMZN), Meta (META), Apple (AAPL), Google parent Alphabet (GOOGL), Nvidia (NVDA), and Tesla (TSLA).
The Market’s Great Paradox
A comprehensive comparison of today’s backdrop and the final hurrah of the dot-com era shows more differences than similarities at this point. While valuations are elevated, we believe equities will “grow into the multiple” in 2026 with strong earnings fueled by ongoing AI capex strength, as well as fiscal and monetary stimulus. Additional upside could come from deregulation, AI-related productivity gains and further labor cost moderation, the latter of which could also open the door to additional Fed easing. Although AI will disrupt labor markets to some degree, we believe the impact will be more like the jobless recovery that followed the dot-com bubble with weaker job gains (50k-75k per month) as opposed to a recession. Additionally, AI should contribute to a disinflationary backdrop that should be bullish for financial assets.
Although it may feel at times like the markets are partying like it’s 1999, we believe the eventual hangover that will follow (such as in 2000-2003) remains further on the horizon. With this in mind, we are reminded of a quote from famous investor William O’Neil: “It is one of the great paradoxes of the stock market that what seems too high usually goes higher and what seems too low usually goes lower.” While the dot-com bubble does have some parallels to the present, it is important to also consider the risks from sitting on the sidelines during periods of large technological change. As such, we remain buyers of dips should any arise in the coming months.
Jeffrey Schulze, CFA
Related Perspectives
Anatomy of a Recession
