21 reasons we’re nearing the end of a secular bull market

Editor's note: the following is a perspective on US markets.

All valuation measures (price/cash flows, price/book, etc.) on the S&P 500 are at all-time highs except for P/Es but we are at 2.5 standard deviations on P/E. Even median valuations are near all-time highs at a P/E of 19. The equal weight indices are lower than the median stock valuations because of more individual equity outliers on the low end. But even the equal-weighted S&P 500 index is near its all-time valuation highs at a 17 P/E. The equity risk premium is negative, which has only occurred in the 1998-2000 period and right before the 1987 crash. On normalized earnings the 12-month forward P/E is about 25 vs. 30 in March of 2000 as we now have above-trend earnings.

The highest growth in the economy is coming from the Mag 7 which makes up 37% of the S&P 500. That means the rest of the economy must be growing slower than in the past as there is only a finite amount of growth in the economy (earnings growth is correlated with GDP growth over the long term and GDP growth is lower now and will be even more so in the future than the last century due to demographics and less immigration). You can’t have it both ways. The biggest cap stocks have much higher growth than in the past. So, the argument that the other 493 are relatively cheap is flawed. How are the other 493 cheap historically if they are growing a lot slower than they have in the past.

Many investors believe the Mag 7 are going to slow down their growth and the other 493 are going to pick up their growth. Even if that happens the S&P 500 overall is going to decline because the Mag 7 dominate the index. Also the overall index would still be way overvalued as the median stock is near all-time highs at a 19 P/E. Nobody has discussed these issues, and I have not heard a good response from any of my peers. But it is so basic!

Some investors argue that P/Es were higher in the late 1990s when both short and long interest rates were higher, giving the current stock market more room to run. However, the core PCE inflation rate averaged 1.5% from 1997-2000, compared to the much higher core PCE inflation rates of the last few years. Also, expecting another 3 standard deviation event is not a good investment strategy as it happens only 0.3% of the time.

The P/E of the S&P 500 advanced from 9 in 2009 to 22 in 2022 (coincides with calendar years!) So one multiple point per year which of course wasn’t linear. But that is how secular bull markets work. In 2012 we were at an 11 or 12 P/E and profit margins and P/EBIT were at all-time highs. The sovereign debt crisis was solved, US banks were securely capitalized, corporate earnings had been growing rapidly since 2009, and interest rates were low. Job growth was 1.8 million (recovering 42% of jobs lost during the Great Recession) and GDP growth was 2.3%. There was a long runway ahead to get back to full employment. No way anyone could justify an 11 or 12 multiple then vs. 23 today. S&P 500 EBIT margins are only slightly higher today than in 2012. That is how secular bull markets work. P/E ratios expand slowly over a 15-to-20-year time period. Can you imagine if someone in 2012 said that the P/E should be 23 or if someone said today’s P/E should be 11 or 12. They would be the laughingstock of the investment community. But they would be rational arguments. Status quo is so powerful.

If the S&P 500 today had the same P/E as in March of 2009, the S&P 500 would be 2772 instead of 6800. All of the increase above 2772 is just subjective valuation by investors. But subjective valuation has a very wide historical spread and we are at the very high end of that spread.

Valuations don’t correlate with short-term returns but correlate very highly with 10-year returns. A valuation model I use that has been made better by Cypress Capital shows the highest correlation with future 10-year returns. Since 1970, the average error for 10-year annual returns has been around 2 to 3%. Thus, if the model forecasts 5% annualized returns for next 10 years, the actual results with high probability would likely be between 3.5% and 6.5%. It has only been inaccurate during bubbles (the late 1990s and of course the last couple of years). It has predicted 0% 10-year returns only three times, and that was in 2000, January 2022 and again now.

Housing bubble

The housing bubble is significantly understated. From December 2019 to July 2022 average housing prices increased 43% or 29% after inflation using a housing price model that combines median and average like for like property sales from Zillow, Redfin, HUD, FHFA and Case-Shiller and National Association of Realtors. Housing prices compound annual rate of appreciation for 2.5 years was 15.1% or an 11% real annual compound rate. Nothing in the early 2000s housing bubble came close to these annual compound rates for the same time period. From July 2022 to August 2025 my model shows housing prices rose 7% while the CPI increased 9%, thus increasing at a rate below inflation. HUD data shows housing prices actually dropped during this time frame by 7%, and in real terms declined over 15%. But HUD data tracks lower income housing buyers. Overall, from December 2019 through August 2025 housing prices increased 51% with real annual compound appreciation of 4.6%. To put it in perspective from December 1999 to December 2005 housing prices increased 43% with a real compound annual returns of 5.9%. Last century housing prices rose on average about 1% per year above the inflation rate, thus the 4.6% real annual returns of the last six years are well above historical appreciation. Still significantly overpriced.

The big difference between the current housing bubble and the bubble in the early 2000s is the lack of leverage today. Bubbles can burst in different ways. There can be a sudden burst as in the 2007-08 period or just a slow deflating of the bubble over many years. When housing prices became overheated in the late 1980s they proceeded to increase at a 2.2% annual compound rate from December 1989 to December 1997 vs. a 3% CPI annual compound rate. Eight years of negative real annual returns in housing prices allowed housing prices to revert to their mean in real terms. That is what has been happening for the last three years.

Secular bull and bear markets

This secular bull market which began in March 2009, is now 16 years and 9 months long. The last two secular bull markets lasted 17 years and 7 months (August 1982-March 2000) and 16 years and 7 months (June 1949-January 1966). There has been a consistent pattern of alternating secular bull and bear markets since 1815 with most lasting between 15 and 20 years. The secular bull markets usually have average real compound annual returns of 12 to 16% vs. the long-term average of 6.5%. The current secular bull market which began in March of 2009 has so far produced real compound annual returns of 13.7% vs 15.1% from 1949-66 and 15.5% from 1982-2000.

Secular bear markets have also mostly lasted between 15 and 20 years and have produced flat or slightly negative real compound annual returns. The lowest real annual compound returns were negative 7.6%, from March of 2000 to March of 2009. Although these were the worst bear market real returns, the bear market only lasted 9 years.

My thesis has always been that the current secular bull market would not be quite as strong as in the past as we only experienced a 9-year bear market (despite the more negative annual returns) but it has been almost as strong. Interesting that the short 8-year secular bull market from 1921-29 produced almost double the real annual compound returns of other secular bull markets at 28%, almost making up for the shorter time period.

There also have been very consistent major 25-to-33-year cycles where the stock market makes a major decades low in valuation and negative returns. These occurred in 1861, 1896, 1921, 1949, 1982 and 2009. That would mean the next major secular low is scheduled to occur between 2034 and 2041. That is relevant today because it takes many years for the low to develop with the top usually occurring at least a decade before the secular bottom. Expect the S&P 500 P/E to fall below 10 sometime during that time period wiping out at least a whole decade of earnings.

All of these secular market patterns point to a likely end to the current secular bull market within the next two to three years and more likely sooner, with this year being my best estimate.

Timing of AI bubble bursting

The timing and circumstances are almost too perfect for the AI bubble to burst within the next couple of years to end the current secular bull market. It seems too easy to call. Almost every CEO of the major hyperscalers have publicly stated that we are in an AI bubble but that they have to continue to invest or else risk being left behind in the race. There are myriad events that could go wrong in the next year or two to burst this bubble.

First, we are in a race with China for AI dominance and China has at least 20 times more effective training runs per dollar than the US. China’s $42 billion private investment in 2025 delivers the equivalent real world training output of $1-1.4 trillion in US terms. At some point in the future are American hyperscalers going to realize similar efficiencies or will more world capital flow to Chines AI to get higher returns for each dollar spent.

Second, there is still no proof that AI companies are going to deliver a reasonable return on capital for all the money invested so far, at least in the next couple of years, which may lower demand for AI.

Third, even if demand for data centers continues to grow, will there be enough energy supplied to the data centers? Orders for nuclear and natural gas turbines are back ordered for the next 5 to 10 years and Trump is discouraging investment in renewable energy that could come online faster. Also, they need local community permits for these energy facilities and there is increasing NIMBY resistance, especially as the enormous demand for water to cool the data enters could deplete the local water supplies. And the latest research reveals higher cancer rates in communities surrounding the data centers.

Fourth, nefarious uses of AI could cause stricter regulation. Fifth, the bonus depreciation could actually backfire as it encourages even more AI investment that wouldn’t generate a high enough return on investment without the tax incentive, adding to the bubble.

Nividia (NYSE: NVDA) only trades at a 25 P/E on 12 month forward earnings estimates, with many investors believing it is very cheap vs. its growth rate and its historical P/E. But for all of its strengths, NVDA is still a cyclical company, and the stock market is saying that NVDA’s earnings are going to decline significantly sometime in the not-too-distant future. Cyclical companies trade at their lowest P/Es near the top of earnings cycles and their highest P/Es near the bottom of cycles. NVDA also has major concentration risk with the hyperscalers and potentially increasing competition. The stock market also does not like the circular agreements it has made with its customers, a real red flag and similar to the circular agreements back in the late 1990s. Memory chip prices are advancing this year at the highest rate since 1999, another eerie resemblance to 1999. The stock market could be wrong, but I wouldn’t bet against it.

Boxes checked for near end of secular bull market

Just about all of the boxes have been checked that an investor would look for to determine that the end of a secular bull market is near.

1. Near record bullish investor positioning.
2. Near record low junk bond spreads.
3. Market concentration at extreme highs.
4. Extreme valuations by all metrics.
5. Bubbles in multiple assets (gold, housing, high yield spreads, equities, crypto).
6. Unemployment near historic lows for a sustained period.
7. High leverage.
8. Speculative fever.
9. Long period of time without a down credit cycle
10. Outperformance of large cap growth stocks, especially technology.
11. Poor market breadth.
12. A new revolutionary technology that captures investors’ enthusiasm for at least a few years.
13. Overinvestment in the near term of the new technology.
14. Low to mid-teens annualized real total returns for the S&P 500 over a 15-to-20-year period, preferably 16 to 18 years.
15. Leading economic indicators declining.
16. Negative output gap.
17. Divergence of free cash flow from earnings (hyperscalers).
18. Circular agreements by the leading tech companies.
19. High labor participation rates especially prime age (18 to 55).
20. Investor margin debt increasing year over year, more than 500 basis points more than year over year gains in the S&P 500.
21. Historical high retail trading volume.

Now it is just a timing issue. The leverage is in government debt and in the median debt to EBIDTA of public companies (although interest coverage is still high) as well as the coming build-up of debt and negative cash flow of the hyperscalers. Market breadth as measured by the cumulative advance/decline line for all exchanges has not been good over the last couple of months despite the appearance of the broadening of the stock market. Overall, it has not been great since the 2022 low, but we need much worse breadth to start a secular bear market. The IPO market, which is one box that has not been checked, has not been especially exuberant. However, as explained earlier, we already had our once-in-a-generation speculative IPO market in 2020-21 and will likely not see another one of that magnitude in the near future. Another important box that has not been checked yet is rising interest rates/FED restrictive policy, although the Fed has been somewhat restrictive over the last couple of years.

David R. Snyder, CFA, is Managing Principal/Chief Investment Officer at Journey 1 Advisors, a registered investment advisor based in Pennsylvania. Disclosure: The opinions merely represent the opinion of the author as CIO of Journey 1 Advisors, LLC and intended to inform the readers about our investment philosophy and strategy. It is not intended to offer investment advice.