Speculative valuations may matter beyond the stock market: a new study finds that positive bubble shocks are followed by stronger economic activity and higher policy rates. Its model estimates that a bubble component reached nearly 25% of the S&P 500 price-dividend ratio in December 2024, a result the authors say may reflect enthusiasm for artificial intelligence. That estimate is model-based; it doesn’t establish that AI stocks were mispriced or that a market decline must follow.
Andreasen and Bro’s paper, published in the Review of Finance in 2026, addresses a persistent problem in market analysis: rising prices can reflect better expected fundamentals, lower required returns or speculative demand. A high valuation ratio alone can’t distinguish among them. The authors combine traditional market data with survey responses about expected dividend growth and stock prices to estimate the parts of valuations associated with expected dividends, expected returns and a rational bubble.
What the model measures
The approach starts with a present-value model, which relates share prices to expected future dividends and the returns investors require. The authors add survey-based expectations to help identify whether those fundamental drivers account for observed valuations. The remaining component is classified as a rational bubble under the model’s assumptions about expectations.
That distinction matters for interpreting the results. A price-dividend ratio can rise because investors expect stronger dividend growth or are willing to accept lower future returns. The study treats the bubble estimate as the portion its model cannot attribute to those explanations. It’s an analytical decomposition, not a direct measure of investor irrationality or a standalone signal that prices will reverse.
Using monthly U.S. data from January 1971 through December 2024, the authors estimate that the bubble accounts for 61% of variation in the S&P 500 price-dividend ratio. Expected returns account for another 42%, while expected dividend growth contributes little to explaining variation across the full sample. These reported variance contributions aren’t shares that add to 100%; they should be read as separate estimates of how each component relates to valuation movements.
Removing the estimated bubble substantially reduces excess volatility in stock valuations and eliminates evidence of explosive dynamics in the unadjusted price-dividend ratio. That finding suggests the speculative component helps explain why market valuations have moved more sharply than fundamentals alone would imply within the authors’ framework.
Historical episodes and the latest estimate
The model’s estimates rise and fall across several familiar market episodes. The bubble component reaches about 10% of the price-dividend ratio during the early-1970s Nifty Fifty boom, then falls to roughly 2% in the 1980s. It climbs during the dot-com period, reaching 27% in January 2001, and rises again around the global financial crisis.
A further increase accompanies the COVID-era market boom, with the estimate peaking at about 19% in November 2021. By December 2024, it reaches nearly 25% of the price-dividend ratio. The authors suggest this latest episode may be linked to investor enthusiasm around AI. The figure’s uncertainty bands are a reminder that these are estimates, not precise readings of an observable bubble.
For market professionals, the most useful distinction is between a valuation concern and a timing call. A model may indicate that prices rely more heavily on expectations than on measured fundamentals, yet it can’t say when that reliance will diminish. The paper allows bubbles to build or decline gradually, so a high estimate does not, by itself, identify an imminent turning point.
Market valuations and the wider economy
The authors also examine whether changes in the estimated bubble are followed by changes in economic conditions. They use innovations in the bubble estimate as an instrument in a structural vector autoregression. Following a positive shock, unemployment falls, industrial production rises and working hours increase.
Two possible channels could help explain the pattern. Higher equity values may raise household wealth and support spending. They may also improve companies’ market values and ease financing constraints, allowing investment to increase. These are plausible mechanisms discussed in the study; the results don’t establish that either one alone causes the observed responses.
Inflation rises only modestly after a positive bubble shock, and the response is generally not statistically significant. The contrast suggests that an expansion associated with rising equity valuations may not look like a conventional demand-driven boom across every economic measure.
Interest rates also respond. The estimated monetary policy rate reaches a peak increase of about 48 basis points eight months after a positive one-standard-deviation bubble shock. Over the same analysis, the spread between long-term Treasury yields and short-term rates declines, as do corporate bond spreads.
Those results connect valuation dynamics to broader financial conditions. They suggest speculative stock market movements can coincide with changes in employment, production and policy rates even when dividend growth and productivity don’t respond significantly. The analysis doesn’t show that policymakers react directly to a bubble estimate; they may instead be responding to the economic activity that accompanies it.
Reading the findings for portfolio risk
The study supports using more than a single valuation ratio when assessing market conditions. Survey expectations offer one way to test whether a high price-dividend ratio is consistent with expected fundamentals or whether the model assigns a larger share to speculation. That added perspective can inform scenario analysis, especially for portfolios concentrated in market segments where investor expectations play a large role.
Still, the bubble estimate is not a ready-made trading instruction. Even if speculative forces are present, they can persist, and the paper doesn’t supply a reliable date for their reversal. For portfolio construction, the more practical use is to consider how holdings might respond if expectations shift and valuations compress.
The economic findings also argue for linking market stress scenarios with macroeconomic assumptions. A decline in valuations could remove some of the support that rising prices provide to household spending or corporate financing. But the paper’s reported results concern positive shocks; they don’t show that negative shocks produce effects of equal size in the opposite direction.
What the chart shows
Figure 3 tracks the estimated speculative component of U.S. stock valuations from January 1971 to December 2024. One panel plots the estimated bubble over time; the other shows it as a fraction of the annual price-dividend ratio. Both include uncertainty around the estimates.

The results are hypothetical results and are NOT an indicator of future results and do NOT represent returns that any investor actually attained. Indexes are unmanaged and do not reflect management or trading fees, and one cannot invest directly in an index.
Research details
The paper, “Identifying a Stock Price Bubble and its Macroeconomic Implications,” is by Andreasen and Bro and appears in the Review of Finance, 2026. A version is available here.
The evidence points to a two-part takeaway: speculative components may explain a substantial share of valuation variation, and their movements may coincide with real economic effects. The estimates can sharpen discussion of market risk, but uncertainty around the model and the absence of a timing signal make them most useful as context for scenarios rather than as a forecast of a correction.
Stock Price Bubble Dissection Requires a look at Fundamentals and Speculation was originally published at Alpha Architect. Please read the Alpha Architect disclosures at your convenience.