Let’s start with the “scary” chart — the earnings yield vs 10-Year Treasury yield (see Chart 1). Earnings yield (the inverse of PE) has been falling with the equity market rally while bond yields are rising on the back of higher inflation. Since late-2023, the gap between the two has almost vanished — and this, according to the bears suggest that investors are not being paid the required premium to hold risky stocks instead safer government bonds. Historically, expensive stocks relative to bonds will eventually correct, that is, stock prices must fall.
But this argument only tells half the story. The other half of the equation, and arguably the most important part, is the expected earnings growth.
Valuation expansion driven by strong growth expectations
See also: The AI grid rewrites the value chain
Yes, S&P 500 stocks are currently trading at a forward price-earnings (PE) of 21.4 times, which is high from a historical perspective. Forward PE has averaged around 18.3 times since 1990 (see Chart 2), but elevated PE alone does not necessarily mean an expensive or risky market.
A simplified valuation model for stocks is (read the sidebar for a more detailed explanation):
See also: The great data delusion — where to invest for AI winners
Price or Value = E/(r-g) or PE = 1/(r-g)
Where:
P = stock price
E = Earnings
g = long-term earnings growth expectation
r = discount rate (risk-free rate + equity risk premium)
Mathematically, PE rises when (r-g) is compressed — as is currently the case. Although the benchmark 10-Year Treasury yield (risk-free rate) has risen, growth expectations (g) have risen even faster.
For more stories about where money flows, click here for Capital Section
Wall Street is coming off one of its strongest quarterly earnings results in years. According to data service provider FactSet, 1Q2026 earnings are now estimated to grow 27.7% year on year (with actual results from 91% of S&P 500 companies), more than double the expected growth just before the reporting season started at end-March. Revenue is up 11.4% y-o-y and crucially, net margin has risen to 14.7%, the highest on record going back to 2009.
The profit margin expansion underscores the dominance and pricing power of US businesses globally. Unsurprisingly, the biggest contributors to the earnings growth were the Magnificent 7 companies, whose blended earnings growth was a strong 61% y-o-y. That said, the rest of the 493 (of S&P 500 companies) did not perform too shabbily either, delivering remarkably robust profit growth of 16.4% y-o-y on average during the quarter.
The growth estimate for 2026 has been revised upwards, to 21.5%. In short, rising valuations are accompanied by higher earnings growth, and there is no compression in the risk premium, which remains within its historical range. This is contrary to what Chart 1 implies — that risk premium has narrowed too much by just looking at PE and 10-Year yields. This can be seen clearly in Chart 2 and Chart 3.
In other words, investors are not accepting unusually low compensation for risk (due to excessive speculation), such as that seen during the late-1990s (dotcom bubble) and just before the global financial crisis (GFC) in 2007.
Of course, this does not mean there is no valuation risk; it’s just that the risk is to growth expectations. Therefore, the right question to ask is — can Corporate America continue to deliver the expected earnings growth, which will be driven primarily by AI?
Strong productivity gains from early AI adoption
The numbers so far say yes. Earnings for S&P 500 companies have grown at a double-digit clip since 2024, due, in no small part, to strong productivity growth (see Chart 4). More importantly, margin-earnings growth was driven by a sharp increase in total factor productivity (TFP) — compared to previous periods going back to 1990 — as opposed to simply adding more labour and/or capital (see Chart 5).
Call this the efficiency-innovation component of growth (doing/producing more with less).
There is a confluence of reasons behind the productivity gains, such as the post-pandemic digitalisation and continued transformation in the US economy towards higher-value software-platform businesses — that are highly scalable with strong network effects, intellectual property (IP) and brand names. US companies are globally dominant and excel at monetising innovations and protecting that created value (we will elaborate on this next week). Note the increased contributions from IP and research and development to US productivity gains in Chart 5.
We believe TFP gains in the last few years are due, in large part, to low-hanging returns from early and rapid adoption of AI. For instance:
- Faster AI-assisted coding and software development;
- Better ad targeting;
- Cheaper content creation;
- Automation tools like AI chatbots; and
- Improved white-collar productivity and workflow
Whether AI can continue to deliver sustained productivity gains and the huge returns (from new products, scientific breakthroughs and new industries) currently expected by markets — and built into stock valuations — is the multi-trillion-dollar question.
Resilient consumer spending
Aside from massive AI infrastructure spending and productivity gains, broader consumption growth is also underpinned by a stable labour market as well as the wealth effect from record high stock and house prices. Unemployment remains low at 4.3% (although new job additions have slowed) and wages are growing at a decent clip, albeit slower than during the early days of the pandemic recovery.
The positive wealth effect, we believe, has a strong impact on consumer spending. It is estimated that at least six of every 10 Americans own stocks, either directly or via mutual funds and pension-retirement schemes. US households are also sitting on substantial wealth gains in their homes — home equity has risen significantly from the GFC lows of around 46% to over 71% currently.
In a nutshell, the average US household is in good financial shape — household debt to GDP has been trending lower, currently around 68%, down from the GFC high of 100%.
Even though the consumer confidence gauge (what people feel) has fallen to a record low, it is contradictory to actual spending, which has not changed drastically. Nevertheless, this does suggest rising risks, should the job market weakens or stock prices collapse.
We think inflation — and interest rates — will stay higher structurally as compared with the post-GFC years, and this will create increasing cost of living pressure on lower-income households. That said, we do not expect runaway inflation like that of the 1970s and early-1980s, in part because the wage-driven inflation mechanism is much weaker today (weak union power), and inflation expectations remain anchored and are offset by tech-AI productivity gains.
Conclusion: Invest if you believe in the AI-led productivity growth — or not
Historically, Corporate America has excelled at monetising innovations and protecting the value created for shareholders through durable pricing power. The S&P 500 has rewarded investors with annualised 11.3% compounded returns since 1990. The market clearly believes that US companies, especially Big Tech, will continue to be dominant globally in the winner-takes-most AI era.
Chart 6 compares the current PE versus expected earnings growth for the S&P 500 and by sector. Notably, expected growth for the IT sector (Nvidia, Broadcom, Apple, Microsoft) and, to a lesser extent, communication services (Alphabet and Meta), energy and materials (related to AI infrastructure) is robust. Therefore, if you believe in this AI narrative, and the implied long-term growth expectations, then current S&P 500 valuation may prove reasonable. Should AI disappoint, and (r-g) expands again, valuations will compress violently.
If, however, you don’t believe in the AI narrative — that the huge capex spending will generate the required returns for the hyperscalers-AI companies and productivity gains for businesses — don’t hedge by investing in non-tech related stocks. For instance, in the real estate, consumer discretionary and consumer staples sectors, where valuations have been dragged up by the equity market rally (PE higher than S&P 500) but expected growth rates are relatively low and current risk premiums are below their long-term averages (see Chart 7). If you believe that growth expectations will fall far short and the AI bubble will burst, then the smart move would be to reduce your equity exposure and hold a higher percentage of bonds and cash.
Portfolio commentary
The Malaysian portfolio fell 0.6% for the week ended May 20, holding up better than the broader market, given our relatively high cash holdings. The only winner was United Plantations (+0.4%) while the biggest losers were Hong Leong Industries (-4.0%), Maybank (-1.1%) and LPI Capital (-0.3%). Total portfolio returns now stand at 215.5% since inception. This portfolio is outperforming the benchmark FBM KLCI, which is down 6.1% over the same period, by a long, long way.
The Absolute Returns Portfolio too ended lower, down 2.4% with all stocks finishing the week in the red. The losses pared total portfolio returns to 33.6% since inception. The top losers were ChinaAMC Hang Seng Biotech ETF (-6.9%), Ping An Insurance - A (-6.9%) and Ping An Insurance - H (-4.6%).
The AI portfolio, on the other hand, gained 1.2% as tech and AI stocks continued to outperform. Last week’s gains boosted total portfolio returns to 19.2% since inception. The top gainers were Naura Technology (+12.7%), Hewlett Packard Enterprise (+5.4%) and Marvell Technology (+5.0%) while the top losers were Minth (-9.5%), Unusual Machines (-4.0%) and Amazon.com Inc (-1.9%).
Disclaimer: This is a personal portfolio for information purposes only and does not constitute a recommendation or solicitation or expression of views to influence readers to buy/sell stocks, including the particular stocks mentioned herein. It does not take into account an individual investor’s particular financial situation, investment objectives, investment horizon, risk profile and/or risk preference. Our shareholders, directors and employees may have positions in or may be materially interested in any of the stocks. We may also have or have had dealings with or may provide or have provided content services to the companies mentioned in the reports.
