Jim Paulsen says the U.S. is now living through the longest technology-led equity cycle of the past century, but by his reading of the data, it is also the worst-performing one.
His argument starts with a simple observation: tech has led for so long that many investors now treat it as the default buy-and-hold asset class. Drawdowns are often framed as chances to add exposure, not as potential signs that a cycle may be changing.
A market where one sector can lift the whole index
Paulsen points to last Thursday as a clean example of the current market split. Ten of the 11 sectors in the S&P 500 fell on the day, with losses ranging from -0.39% to -1.50%. The index still rose 0.64% because information technology, the only advancing sector, jumped 3.4%.
That kind of session helps explain why so many investors now question whether they need to own much outside tech at all.
How Paulsen defines a tech leadership cycle
The article examines relative total return data for U.S. technology stocks going back to 1926. For the period before 1989, Paulsen uses the computer, software and electrical equipment industry groups from the Kenneth R. French public database and compares them with the total U.S. market index. From 1990 onward, he compares the S&P 500 information technology sector index with the broader S&P 500.
He identifies six major long leadership cycles since 1926. His method is qualitative: a cycle runs from a low in relative total return to the next major high, provided there is no meaningful long stretch of underperformance in between. The aim is not to map every bull or bear market. It is to isolate the long periods in which sustained market leadership came from technology shares.
Cycle 1: March 1927 to August 1929
The first major cycle ran from March 1927 to August 1929 and roughly marked the end of the Roaring Twenties. It was driven by technologies and industrial innovations that reshaped U.S. life and the economy.
- Commercial radio and mass communication expanded under Radio Corporation of America.
- Large-scale automobile production spread through companies such as General Motors.
- DuPont and Union Carbide pushed new work in polymers, plastics and synthetic chemicals.
- Public enthusiasm for commercial aviation climbed after Charles Lindbergh’s transatlantic flight in 1927.
Cycle 2: May 1932 to December 1937
The second cycle, from May 1932 to December 1937, followed the stabilization of the U.S. banking system after the Great Depression. The U.S. moved away from the traditional gold standard and devalued the dollar, injecting large amounts of liquidity back into the financial system. That helped break the deflationary spiral of 1929 to 1932 and pushed cheap capital toward high-growth industrial assets.
During the period, the Dow Jones Industrial Average rose nearly 500%. Radio became a central medium for news, entertainment and political speeches, including Franklin Roosevelt’s fireside chats. Better airmail contracts and safer passenger aircraft helped aviation and logistics spread more widely. Federal spending under the New Deal, including infrastructure, rural electrification and modernization projects, also supported technology manufacturing.
Paulsen adds that the creation of the Securities and Exchange Commission, along with stronger disclosure standards and action against manipulation, helped restore trust in Wall Street.
Cycle 3: 1952 to 1960
The next leadership phase lasted from 1952 to 1960 and was powered by postwar business innovation, heavy Cold War defense spending and plain speculation.
- The transistor moved into commercial use, shifting electronics away from vacuum tubes.
- The Cold War and space race intensified, especially after Sputnik in 1957.
- Early mainframe computers started to automate complex operations.
- Programming languages such as FORTRAN and COBOL emerged.
- Investors chased companies whose names included endings such as “-tron” or “-tronics.”
Cycle 4: 1964 to 1967
The 1964 to 1967 surge is described as one of the peaks of Wall Street’s “go-go years.” It was driven by space-age innovation, aggressive mutual fund trading and widespread retail speculation.
Investors rotated away from old-economy blue chips and into glamour stocks. The main ingredients were:
- High research spending being treated as proof of future earnings power.
- An “electronics” craze, where company names containing “-tronics” or “computer” could lift valuations.
- The rise of aggressive individual portfolio managers, the so-called gun-slingers.
- A small group of elite, innovative tech companies becoming must-own market favorites.
Cycle 5: 1992 to 2000
The 1992 to 2000 run ended in the internet bubble. Paulsen ties it to the commercialization of the web, unusually large venture capital inflows and broad retail speculation.
- The World Wide Web emerged.
- Investors prioritized market share and user growth over traditional financial measures.
- Y2K concerns lifted corporate IT spending.
- Online brokers and affordable personal computers opened active tech trading to retail investors.
- Interest rates stayed relatively low for much of the period.
Cycle 6: 2006 to 2026
The current cycle, running from 2006 to 2026 in Paulsen’s framing, rests on four structural changes in computing, communication and monetization.
- The mobile revolution from 2007 to 2015, when smartphones kept users connected at all times.
- Cloud computing from 2006 to the present, bringing infrastructure-as-a-service and subscription models.
- Digital advertising and platform dominance during the 2010s.
- Artificial intelligence and semiconductors from 2023 to 2026.
Why he says tech is not a simple buy-and-hold trade
Paulsen’s first broad conclusion is that diversified tech exposure has not historically behaved like a permanent buy-and-hold winner. Stories built around exceptional names such as IBM or Apple can create that impression, but the sector as a whole looks different.
Across all months since 1926, U.S. tech stocks beat the broader market only 50.5% of the time. In his view, that is close to a coin flip. During long leadership periods, the share of outperforming months rises to 58.1%. In the remaining months, it falls to 44.5%.
For Paulsen, the implication is clear: the big payoff comes not from owning tech forever, but from distinguishing long leadership phases from long periods of lagging performance.
He argues that, with the exception of the third cycle, each major leadership phase was followed by persistent underperformance until tech roughly gave back its excess and returned to something near market-level standing relative to its 1926 starting point. That is why he says long-term buy-and-hold in tech may produce little more than market-like returns with much higher volatility.
Leadership periods are long, but lagging periods are longer
The six long leadership cycles shown in his chart lasted 93 months on average, or 7.75 years. They often included painful pullbacks, but the relative trend usually remained upward for long stretches.
The longer-term lagging periods were even longer. Paulsen counts five complete long underperformance cycles since 1926, and they lasted 132 months on average, or 11 years.
That matters because the current outperformance run, which began in 2006 and continues today, has been so extended and so steady that investors may be forgetting what the other side of the history looks like.
Macro conditions do not offer a stable signal
Another point in the piece is that tech stocks have performed well and poorly under very different macroeconomic backdrops. Paulsen does not see a stable relationship between technology leadership and any single set of economic conditions.
Tech did well during the prosperity of the 1920s, during the recovery from the Great Depression in the 1930s and during the boom of the 1990s. It did badly in the 1940s even with solid real growth. It lagged from 1960 to 1965 even when real GDP growth was mostly firm. It also lagged in the 1980s despite generally decent growth.
The inflation record is mixed as well. Tech underperformed during World War II-era inflation, was roughly in line with the market during the severe inflation of the 1970s, and performed well during the inflation period linked to the pandemic and the Iran war in 2020.
His conclusion is that tech cycles depend more on innovation itself, on research and other investment spending, on how long and how far tech has already outperformed or underperformed, and on whether investor sentiment is too pessimistic or too optimistic.
Major turning points tend to be violent
Paulsen also says long tech leadership cycles usually begin and end with force. In his charts, the starts and ends of the six phases are mostly V-shaped rather than gradual.
Relative total return bottoms can reverse sharply higher while investors are still questioning whether the previous long tech bear market is truly over. Tops can break just as quickly, with little in the way of a rounded warning pattern. Many investors are still waiting for the next dip-buying chance when the decline starts to accelerate.
That is one reason he calls long-horizon timing in tech so difficult.
The indicators he watches
Paulsen does not present a mechanical model for deciding whether tech is still a good long-term buy or approaching a new long decline. He does, however, list a set of indicators that he thinks matter:
- How long tech has already been the market’s leadership group.
- How large its outperformance over the broad market has been in the current cycle.
- How excited and aggressive U.S. corporate behavior has become, including capital spending, cash-flow use and debt use.
- How dominant technology has become in media and market culture.
- How much investor enthusiasm, certainty and complacency is visible.
History, he notes, shows how difficult those calls can be. It would have been easy to say in 1935, 1955, 1997, 2020 or 2022 that the long tech bull market was over, and each time that call would have come too early. It also would have been easy in 1942, 1948, 1989 or 2002 to jump back into tech, and those decisions would have proved costly.
The current cycle stands out for length, not return
Against the prior five major tech bull markets since 1926, the current one is exceptional in one respect above all: duration.
Figure 2 shows that the previous five cycles lasted 63.4 months on average, or 5.28 years. Before the present run, the longest was the third cycle, which ended in 1960 after 99 months, or 8.25 years. The internet-bubble cycle lasted 87 months.
The current cycle, by contrast, has reached month 241 and is now slightly past 20 years. Paulsen says that helps explain why so many investors now struggle to imagine holding much outside technology and why public market discussion has become so dominated by the sector.
But when he compares excess average annualized return relative to the broader market, the picture changes sharply. Figure 3 shows that since 2006, the current cycle has generated only 6% in excess average annualized market return. In his dataset, that is lower than even the weakest earlier tech leadership phase.
So while tech has still been a successful investment over recent decades, its relative performance efficiency versus the market has been the poorest of the six major leadership cycles.
Lower volatility, but also the weakest return per unit of risk
Paulsen says one reason the current cycle may have lasted longer is that it has looked more like a steady grind than an explosive boom. Figure 4 shows that because its return stream has been weaker than earlier long tech runs, the relative total return volatility experienced by investors has also been fairly low.
Today’s relative return volatility is close to what was seen in the big tech runs of the 1950s and 1960s. It is far below the volatility associated with the Great Depression cycle and especially the internet-bubble cycle. He treats the dot-com episode as an outlier, with relative return volatility well above every other cycle since 1926.
Even so, Figure 5 shows that the present cycle also delivers the lowest relative total return per unit of risk among all the long tech leadership phases in the sample. The 1960s IBM mainframe cycle produced the best risk-adjusted relative result, and the 1930s recovery cycle also ranked strongly because both generated much higher average annualized excess returns.
His conclusion: not a full exit, but a lower weight
Paulsen says the piece is meant more as commentary than as a forceful market call. Still, he writes that a growing list of warning signs leaves him increasingly uneasy.
Figure 1 shows the relative total return index for the tech sector has just risen to a 100-year high. The current 20-year leadership phase is twice as long as any prior cycle in the record. He says the recent AI boom has also materially lifted investor excitement and made corporate behavior more aggressive.
He describes companies as burning cash flow, pushing capital spending plans faster and adding debt. Financial media, in his view, is now heavily captured by technology and innovation narratives. Wall Street analysts are even tracking the frequency of the word “AI.” Videos of robots mowing lawns, performing surgery and breaking the 100-meter sprint record are spreading widely. After 20 years of success, he sees investor complacency around tech as clearly warmer, if not outright extreme.
Still, he stops short of calling an immediate top. He says it is hard to make that call when tech earnings appear to be surging and the sector has been winning for so long. Even if such a call proves right in the end, it could arrive five years too early.
That leads to a measured recommendation: do not sell everything, but move technology and other “new era” sectors to an underweight position.
An added warning from business investment behavior
The final chart in the article overlays the relative total return index for the U.S. technology sector since 1955 with detrended real investment per job. Bold green numbers mark the tops of prior long tech cycles and the current cycle’s place in that sequence.
Paulsen says there has been a fairly close relationship since at least World War II between tech’s relative performance and aggressive corporate spending per employee. When detrended real investment per job rises above zero, meaning investment per job is above average, it has often signaled that a tech-led market cycle is moving toward its end.
In the third, fourth and fifth cycles, once this measure moved well above zero, the cycle ended relatively quickly. In the last few quarters, the ratio has surged above zero again. He says the current level is comparable to the one that preceded the end of the third cycle.
His bottom line is cautious rather than dramatic. Tech leadership may continue for some time. But the warning signs are building, and the risks of concentrating in tech are rising. If investors cannot bring themselves to be bold against a 20-year trend, he suggests they can at least be a bit more careful.
Disclosure and risk language included in the piece
Paulsen says stocks are risky by nature and that any stock can lose a large part of its value at any time, including those mentioned in the article. He says investors should never rely on a single source for investment decisions. He describes himself as a retired investment strategist offering views and observations on markets, the economy and companies, and says readers should do their own research and consult qualified financial planners and investment advisers before making decisions.
He also states that the newsletter is for reference only and is meant for illustration and discussion, not financial, legal, tax or investment advice. The material reflects the author’s expectations as of the stated date only. Some information comes from third parties or is based on third-party sources. While such information is considered reliable, no representation is made as to its accuracy, completeness or timeliness.
The disclosure adds that the financial instruments discussed are speculative in nature and may involve risk to principal and interest; any prices or levels shown are historical or indicative only; and the material does not take into account the investment objectives, financial situation or needs of any specific investor. Recipients are responsible for making their own independent decisions regarding any securities, investment products or other financial instruments discussed.