The most direct signal that US stocks have peaked: South Korean chip exports?

Whether U.S. stocks can continue to rise may not depend on the bond market, but rather on monthly data within Asia's semiconductor supply chain.

Recently, Bloomberg macro strategist Simon White pointed out that investors are currently overly fixated on the question of "how high U.S. Treasury yields need to go to crush the equity market," which is essentially a false premise. The real indicator to track is the chip export data from South Korea and Taiwan—if a significant decline emerges, it will provide a 3- to 6-month advance warning that the capex cycle has peaked, prompting a shift in earnings expectations and putting sudden pressure on U.S. equities in the current high-yield environment. So far, these export figures do not show signs that the investment cycle is about to reverse.

High yields have yet to bring down the equity market, mainly because earnings expectations are rising at an unprecedented pace.

The aggregate expected earnings for the S&P 500 over the next three years are growing at an annual rate of 40%, far outpacing the negative impact of higher discount rates on present cash flow values. At the same time, the implied equity risk premium calculated by Aswath Damodaran, finance professor at the NYU Stern School of Business, currently stands at 3.67%. When combined with Treasury yields, the S&P 500’s expected return has recently climbed close to 9%, explaining the market’s resilience in a high-yield environment.

However, this situation is highly dependent on earnings expectations remaining at current levels. Once expectations start to shift, yields above 5% will quickly turn from manageable background noise to a major risk for valuations.

A prevalent misconception in the market is that equities will collapse automatically once yields breach a certain threshold. Simon White believes this framework overlooks the underlying drivers of higher yields.

The current upward move in nominal yields in the U.S. is almost entirely driven by real rates, not inflation expectations (breakeven rates). This distinction is crucial: rising real rates reflect the market’s pricing of stronger-than-expected economic growth; if inflation expectations were the main driver, it would signal rising fiscal risk and directly impact earnings forecasts. The former poses much less threat to equities than the latter.

From a relative valuation perspective, U.S. equities are currently only modestly overvalued versus Treasuries, with little pressure for a major rebalancing. Yields have exerted a negative impact on stock valuation multiples (Forward P/E)—investors doubt the sustainability of rapid earnings growth, and higher discount rates compress the present value of long-term cash flows—but strong revenue and profit margin growth are bridging this gap, keeping the S&P 500 robust overall.

The main pillar supporting the stock market is exceptionally strong earnings expectations. The aggregate projected earnings growth for the S&P 500 over the next three years reaches 40%, a rate not seen historically.

This growth rate is so substantial that it more than offsets the discounted cash flow losses caused by higher rates. It is this force that has kept the Damodaran framework’s implied equity risk premium in a positive zone of 3.67%—a sharp contrast with the negative ERP derived from simply using rolling historical earnings.

The issue, however, is that these earnings expectations are highly concentrated in the narrative of AI-driven capital spending. Should the tech capex cycle cool down, a rapid downgrade in earnings forecasts would make equities suddenly vulnerable in a 5%+ yield environment. At that point, high valuations would lose earnings support and face discount rate compression—a double whammy.

The mainstream market view favors using analyst forecasts for capex or individual announcements from large tech companies to judge the capex cycle. Simon White argues that both approaches have obvious shortcomings: analyst forecasts lag the cycle, and single-company announcements are nearly impossible to predict in advance.

In comparison, export data from South Korea and Taiwan function as natural leading indicators. Any systemic capex slowdown will show up in these countries’ export figures first, with a lead time of about 3 to 6 months.

The accessibility of the data is another advantage: export statistics are published monthly and promptly at the start of each month; additionally, South Korea offers a preliminary estimate for the first 10 days of exports, usually released around the 21st of each month, giving investors a higher-frequency tracking window.

Simon White emphasizes that when chip exports begin to fall, the impact will be twofold: it will directly dampen earnings expectations, making equities suddenly fragile under high-yield conditions; on the other hand, since chip exports lead nominal GDP, a downward revision in growth expectations will also push yields lower. The exact sequence remains uncertain—an investment shock could simultaneously drag down both stocks and yields, or it might first hammer earnings expectations and then fully expose the constraint of high yields.