Pay Today, Earn Tomorrow: Why Valuations Matter in the Age of AI
The relationship between current asset prices and future returns has captivated investors and researchers for generations. At its core, this relationship reflects a fundamental economic principle: the price you pay today is the primary driver of the return you will earn tomorrow.
When assets trade at elevated valuations relative to their fundamentals (like earnings or sales), future returns tend to disappoint. Conversely, periods of depressed valuations have historically preceded periods of superior long-term performance.
Today, this discussion is not merely academic. As of late 2025, the U.S. stock market, as measured by the S&P 500, is trading at valuations that rival the peak of the 1999-2000 “dot-com” bubble. This surge has been famously led by the “Magnificent 7” stocks, a handful of mega-capitalization technology companies whose performance has become synonymous with the market itself. The excitement surrounding Artificial Intelligence (AI) has provided a powerful narrative, drawing parallels to the “new economy” enthusiasm that defined the late 1990s.
This leaves investors standing at a difficult crossroads, asking two critical questions:
- Are we in a speculative bubble driven by AI hype that is destined to burst?
- If the price for U.S. stocks is so high, what returns can I realistically expect from my portfolio over the next decade?
In our latest white paper, we address these questions head-on. We first establish the theoretical and empirical link between valuations and returns. Second, we analyze the “bubble” question by comparing the current environment to historical precedents, using data to separate hype from reality. Utilizing empirical work from leading researchers and asset managers like Research Affiliates, Vanguard, and J.P. Morgan, we evaluate the predictive power of various valuation metrics – including the cyclically adjusted price-to-earnings (CAPE) ratio. The evidence consistently demonstrates that while valuations are poor predictors of next year’s returns, they exhibit robust predictive power for returns over 10-, 15-, and 20-year horizons. This makes these tools particularly valuable for the exact people who need them most: long-term investors such as pension funds, endowments, and, most importantly, individual retirement savers.
Finally, we present 10-year expected returns for key global equity asset classes, synthesized from leading capital market forecasters, to provide investors with a practical, forward-looking guide for setting expectations and building resilient portfolios. We find that U.S. large-cap equity valuations are at historic percentiles, implying sobering future returns. Conversely, these same metrics suggest that international developed markets, emerging markets, and U.S. value and small-cap stocks offer significantly more attractive long-term prospects. The article concludes with practical implications for asset allocation and setting realistic financial goals.
Download the full white paper by clicking the button below or read on for a summary.
To understand why high prices lead to low returns, we don’t need a crystal ball. The logic is rooted in simple financial mathematics and the observable, long-term behavior of markets.
Theoretically, valuations are linked to expected returns through dividend discount models, particularly the Gorden Growth Model, which states that the expected return on a stock is the sum of its dividend yield plus its expected future growth rate in dividends (or earnings). We address this in depth in the white paper, but you can think of it this way:
A stock’s price is the “present value” of all its future earnings. If the price is bid up to very high levels, it simply means that more of the future good news is already “priced in,” leaving less room for future returns for the person who buys it at that high price.
A critical feature of valuation ratios is their tendency towards mean reversion over long horizons (10, 15, 20 years). While valuations can deviate substantially from their historical norms for extended periods—sometimes for many years—they eventually gravitate back toward their long-term averages. This mean-reverting behavior is the engine that underpins the predictive power of valuation metrics, particularly at 10-year or longer horizons where short-term market “noise” and sentiment diminish.
To gauge market valuations, researchers and practitioners use several key metrics. None are perfect, but some are far more reliable than others. They include:
- Traditional Price-to-Earnings (P/E) Ratio – the most common, but it suffers from severe limitations.
- Price to Book (P/B) and Price-to-Sales (P/S) Ratios – while both are moderately predictive, they are also less robust than earning-based measures that account for profitability over a full cycle.
- Cyclically Adjusted P/E/ (CAPE) Ratio, often called the “Shiller P/E” after its creator, Nobel laureate Robert Shiller – the most reliable and widely cited metric for long-term forecasting, extensive empirical evidence supports CAPE10’s predictive power. As we review in the white paper, original and subsequent research has shown that a high starting CAPE ratio predicts low 10-year returns, and a low starting CAPE ratio predicts high 10-year returns for developed and emerging markets.
- Q Ratio, first introduced by Nicholas Kaldor in 1966 and later popularized by Nobel Laureate James Tobin (hence it is commonly known as Tobin’s Q) – its predictive power is most pronounced over long investment horizons. However, there are some important caveats in using the Q ratio regarding replacement cost, types of assets marked for the market, and types of economic environment.
Each of these metrics is covered in depth in our white paper (which you can download here).
Is there an AI Bubble?
This brings us to our first critical question. With the CAPE ratio for U.S. large-cap stocks (S&P 500) standing at a staggering 40.42 as of November, 2025[1], we are at a level higher than nearly any point in history, except for the peak of the 1999-2000 dot-com bubble.
The above chart is from https://www.multpl.com/shiller-pe. Past performance is not a guarantee of future returns.
This is not a matter of opinion. According to data from Research Affiliates, this above-40 reading is in the 98.7th percentile of all historical values since 1880.[2] Prices are, by this reliable measure, exceptionally high.
The narrative driving these prices is the transformative potential of Artificial Intelligence, leading to obvious and anxious comparisons to the dot-com bubble. Is history repeating itself?
The parallels are undeniable: Like the internet of the 1990s, AI is truly revolutionary technology with the potential to reshape the entire economy. Media and public interest are intense, creating a “fear of missing out” (FOMO) that drives speculative investment. And just like the 1990s, new valuation paradigms (today using AI models or deprioritizing traditional cash-flow analysis, for example) and a narrow market concentration (today led by the “Magnificent 7”) exist.
But there are crucial differences. Most importantly, today’s leaders – Microsoft, Google, Nvidia, Apple – are among the most profitable corporations in world history. Unlike the many dot-com companies that lacked a clear path to profitability (remember Pets.com?), these are established, dominant business models that are using AI to enhance their existing moats.[3] These businesses are also benefitting from a mature technological ecosystem of global high-speed internet, cloud computing, and massive existing datasets – which were unavailable in 2000.[4] Furthermore, artificial intelligence is already generating tangible economic value in vital sectors like healthcare (medical diagnosis), industry (predictive maintenance), and law (contract analysis).[5]
So, are we in a bubble? The answer is nuanced.
We are not in a dot-com-style bubble of unprofitable, speculative companies destined for bankruptcy. We are, however, in a valuation bubble, where the prices for high-quality, profitable companies have been bid up to levels that are mathematically unsustainable from a future-return perspective.
As finance professor Luciano Floridi argues, the hype is real, and the speculation is outpacing the reality of deployment, which history shows can take decades, not months. The market is pricing in decades of perfect execution and transformative growth today.[6]
For a much deeper dive into the valuations of individual stocks, please refer to Exhibit A of the white paper. Click here to download.
What are the realistic, long-term expected returns for major equity asset classes?
This brings us to our second critical question. If the CAPE ratio for U.S. large caps is at the 98.7th percentile, what returns should we expect?
We have synthesized 10-year nominal (before inflation) return forecasts from three of the world’s leading asset managers who specialize in this type of modeling: Research Affiliates, Vanguard, and J.P. Morgan.
It is crucial to note that these are not 1-year guesses. They are 10-year, annualized projections based on rigorous models that start with current valuations (like CAPE), add expected earnings/dividend growth, and factor in mean reversion.
The results are stark and unanimous.
10-Year Equity Return Forecasts (2025-2035)
The table below presents the 10-year annualized nominal return forecasts for the specific asset classes as of September 30, 2025. Since then – through November 28, 2025, most stock assets classes have increased in value modestly – from about 1% to 4% (driven in part by a rebound in technology stocks during the week of Thanksgiving).
Analysis of the Forecasts
The message from the data is overwhelming:
- U.S. Large-Cap Returns Are Projected to Be Anemic. The consensus forecast for the S&P 500 (U.S. Large Cap) is in the 3%-7% range, with Research Affiliates and Vanguard at the low end. The picture for U.S. Large-Cap Growth—the home of the “Magnificent 7″—is even more dire. Research Affiliates projects a paltry 1.50% per year, and Vanguard projects 1.3% – 3.3%. After accounting for inflation (projected at 1.6%-2.6% by Vanguard), the real return could be zero or negative for a decade.
- Valuations are the Reason. This isn’t a bet against AI or American innovation. It is a simple accounting of the starting price. As stated previously, the CAPE ratio for U.S. Large stocks is above 40 (98.7th percentile). In contrast, the CAPE for “Dev ex US Large” (EAFE) is a historically average 19.9 (51.6th percentile).[7] Investors are paying twice as much for every dollar of cyclically-adjusted U.S. earnings as they are for international earnings. The forecasts simply reflect this gap.
- Value and Small-Cap Stocks Appear Far More Attractive. Within the U.S., the brightest spots are in value and small-cap stocks. Morningstar’s Q4 2025 outlook notes that while U.S. large-cap growth stocks are at a 12% premium to their fair value, small-cap stocks are trading at a 16% discount. This valuation gap is reflected in the forecasts: Research Affiliates projects 8.68% for U.S. Small-Cap Value, nearly 6x its forecast for U.S. Large-Cap Growth.
- International and Emerging Markets Show Robust Potential. The forecasts for Foreign Developed (EAFE) and Emerging Markets (EM) are consistently in the 7%-8% range across all providers. This is more than double the forecast for the S&P 500 from Research Affiliates and Vanguard. After a decade of U.S. market dominance, valuations and, therefore, future return prospects, have flipped in favor of international diversification.
What does it mean for investors?
This research has profound, practical implications for investors.
- Strategic Asset Allocation
The most important application of this data is in strategic asset allocation – setting the long-term, target percentages for your portfolio. Institutional investors like pension funds use these exact forecasts to determine their optimal mix of assets. When U.S. equity valuations are at historic highs, prudent risk management calls for rebalancing away from the most expensive assets (U.S. Large-Cap Growth) and allocating toward those with better long-term prospects (Value, Small-Cap, and International/EM stocks).
- Setting Realistic Return Expectations
Perhaps the most valuable application of all is in setting realistic financial goals. For the past decade, many investors have become accustomed to double-digit annual returns from simple S&P 500 index funds. The valuation data from nearly all major forecasters warns that this period is over.
An investor whose financial plan assumes a 10% annual return from U.S. stocks is setting themselves up for failure. A plan that assumes a more modest return would be far more realistic and encourages the right behavior: saving more, delaying retirement, or reducing spending goals. Communicating these sober expectations helps prevent disappointment and encourages appropriate risk-taking.
- The Fallacy of Market Timing
Those investors who are more risk-averse may also desire to lower the allocations to equities – although there is no guarantee that stock valuations will decline in the short term (and valuations may even go much higher).
Be aware that a critical limitation of this data is that it tells you nothing about when returns will occur. Valuations have virtually no predictive power for 1-year returns, which are driven by momentum, sentiment, and short-term news.
The late 1990s bubble illustrates this perfectly: the CAPE ratio signaled extreme overvaluation as early as 1996, yet the market continued to soar for four more years before the bubble burst. An investor who sold everything in 1996 would have missed substantial gains, even though their long-term thesis was correct.[8]
The value of these metrics lies in setting long-term strategic allocations, not in making short-term tactical trades. Again, however, those investors who are more risk-averse might choose to lower their allocations to equities, realizing that they might be sacrificing returns from equities should equity returns continue their recent trajectory.
In Summary
The past 25 years of rigorous research have firmly established that current asset valuations contain meaningful information about future long-term returns. The evidence consistently demonstrates an inverse relationship: high prices today lead to low returns tomorrow, particularly over horizons of 10 years or longer.
Today, this evidence points to a clear and present challenge. U.S. large-cap stocks, driven by a powerful AI narrative and concentrated in a few technology giants, are trading at valuations in the 98th percentile of history. This is not a repeat of the 2000 bubble, as today’s market leaders are highly profitable and mature companies. However, for investors, the high price may have well “pulled forward” future returns, leaving little on the table for the next decade.
The forecasts from Research Affiliates, Vanguard, and J.P. Morgan all confirm this, projecting low-single-digit returns for U.S. large-cap growth stocks.
But this is not a counsel of despair. It is a counsel of diversification. The very same metrics that paint a bleak picture for the S&P 500 tell a far more optimistic story for U.S. small-cap stocks, U.S. value stocks, and international stocks in both developed and emerging markets. In these unloved corners of the market, valuations are reasonable, and 10-year expected returns are two to three times higher.
For long-term investors, the core message remains clear: valuations matter. While they cannot predict the timing of market movements, valuation metrics provide our best window into long-term expected returns. In an uncertain world, this information is invaluable for asset allocation, risk management, and setting the realistic expectations that are the foundation of all successful financial plans.
To download the full white paper, including Exhibit A: A Deeper Dive into AI-Related Stock Valuations, CLICK HERE.
About the Authors
Ron A. Rhoades, JD, CFP®
Ron Rhoades is an Associate Professor of Finance at the Gordon Ford College of Business, Western Kentucky University. He also serves as a financial advisor at Scholar Financial, a practice within XY Investment Solutions LLC. With a background as both an attorney and a CERTIFIED FINANCIAL PLANNER™ professional, Ron is a nationally recognized authority on the fiduciary duties of financial advisors.
Chris Brown, Ph.D., CFP®
Chris Brown is a faculty member in the Department of Finance at the Gordon Ford College of Business, Western Kentucky University, and a financial advisor at Scholar Financial, a practice within XY Investment Solutions, LLC. He holds the CERTIFIED FINANCIAL PLANNER™ designation and a Ph.D. in Personal Financial Planning. His research and teaching focus is on behavioral finance, retirement planning, and evidence-based investment strategies.
Sources
[1] Source: https://www.multpl.com/shiller-pe
[2] See Arnott, R. D., Beck, N., & Kalesnik, V. (2016). Shiller’s CAPE: Market timing and risk. Research Affiliates Publications.
[3] Id.
[4] See, e.g., Taherdoost, supra n.6, and Al-Maani, supra n.7. See also, e.g., Investing.com. (2025, October 6). Why the AI boom may defy history: 4 reasons this time could be different.
[5] Floridi, L. (forthcoming). Why the AI Hype is another Tech Bubble. Digital Ethics Center, Yale University.
[6] Id.
[7] Research Affiliates, LLC. (2025). Asset allocation interactive data (As of September 30, 2025) [Data file].
[8] This value of 25.2 was already considered quite high compared to the historical average (which was around 16-17 at that time). The ratio continued to climb rapidly throughout 1996 and 1997, reaching approximately 28 by January 1997. The CAPE ratio finally peaked at around 44 during the Dot-com bubble in 2000.
Tiffany Isselhardt/Canva.