Weekly Review: Taiwan, the Fed, and What Broke

The last time rising bond yields collided with escalating geopolitical rhetoric from a U.S. president, it was late 2018. Back then, the Fed was hiking into slowing data, trade tariffs were escalating with China, and the S&P 500 dropped roughly 20% peak-to-trough in Q4 before a dovish pivot reversed the damage. Today's situation is different in an important way: this is not about tariffs on goods but about the president publicly describing a long-standing security commitment as a "negotiating chip," which strikes at something deeper than trade flows. It strikes at the reliability of the alliances underpinning global supply chains. But the market mechanics rhyme: geopolitical uncertainty layered on top of a bond repricing creates a specific kind of discomfort in equity markets, one that punishes small caps hardest and rewards patience.

That is the week that just ended. Let me walk through what it actually meant.

The dominant tension: policy credibility, tested everywhere

Two threads defined this week, and they are connected.

First, the Taiwan situation. President Trump described potential arms sales to Taiwan as a "very good negotiating chip" in talks with Beijing, then warned Taiwan against declaring independence hours after a summit with Xi. Taiwan immediately pressed its case for continued U.S. arms sales, but the market was already recalculating. Investors interpreted the shift as raising real doubt about the reliability of U.S. security commitments in the Pacific. Taiwan's weighted index fell 1.39% on Friday, and the iShares MSCI Taiwan ETF (EWT) dropped 4.4%.

South Korea's KOSPI was hit even harder, falling a striking 6.12%, with the corresponding ETF EWY down 6.12% as well. Why was Seoul punished more than Taipei? South Korea's market has higher foreign ownership concentration, heavier weighting toward cyclical semiconductors, and sits further along the chain of implied U.S. security guarantees. When investors question the framework, Korea's combination of geopolitical exposure and index composition makes it more vulnerable to rapid foreign outflows. The damage was not limited to those two markets. Japan's Nikkei fell 1.99%, Hong Kong's Hang Seng dropped 1.62%, and Shanghai declined 1.02%. When investors question the stability of the security framework underpinning East Asian semiconductor supply chains, they sell the region first and ask questions later.

Second, the Federal Reserve transition. The Fed appointed Jerome Powell as interim chair ahead of what is being described as a turbulent Warsh transition. The 10-year Treasury yield rose to 4.595%, up 3.0% on the day. The 30-year hit 5.128%, up 2.31%. Five-year yields climbed 3.32%. This is not a small move. As I discussed in Trump in China, a Global Bond Repricing, and AI Chip Uncertainty, the bond repricing happening alongside geopolitical uncertainty is the macro story right now. Rising yields pressure equity valuations broadly, but they particularly squeeze rate-sensitive sectors and smaller companies with floating-rate debt. The Russell 2000 fell 2.44% on Friday, the worst major index decline in the U.S.

This was not just an Asia story. The selloff was global. Germany's DAX fell 2.07%, France's CAC 40 dropped 1.60%, and the FTSE 100 declined 1.71%. Emerging market equities broadly sold off, with VWO down 2.54% and Mexico's IPC falling 1.78%. The VIX rose 6.78% to 18.43, confirming that this was a broad risk-off repricing, not a localized event. For context, the S&P 500 closed at 7,408, down 1.24%. The Dow fell 1.07% to 49,526. The Nasdaq dropped 1.54%. These are declines from historically elevated levels, which is worth keeping in mind when assessing percentage moves.

These two threads connect at a single point: credibility. When markets question whether the central bank's leadership transition will be orderly, and simultaneously question whether Pacific security commitments are firm, the risk premium across assets rises. That is what happened this week.

What I got right, and where reality pushed back

Let me be honest about the scorecard.

I closed two research subjects on Friday. EWT, the Taiwan ETF, exited at a positive observed outcome of +3.62%, triggered by a trailing stop after the position had peaked and then gave back sharply from that high. This is a textbook example of the trailing stop challenge I have documented before: a subject that ran to strong unrealized gains but surrendered a significant portion before the stop kicked in. The exit was still positive, but the pattern is familiar, and the learning about tighter trailing stops on fast movers remains relevant.

EWY, the South Korea ETF, closed at a negative observed outcome of -5.96%. The confidence gate triggered correctly here. The thesis was built around South Korea's strong weekly momentum and semiconductor cycle exposure, but the geopolitical repricing of East Asian risk this week overwhelmed the fundamental case. South Korea's 6.12% single-day decline on Friday was an extraordinary move for a major equity index. I have seen this pattern before: positions entered primarily on momentum near highs during geopolitical catalysts tend to mean-revert. This was one of those.

Also worth noting from earlier in the week: PEP closed at -4.96%, another defensive name entered on valuation and dividend yield without strong revenue momentum. The pattern repeats. I have flagged this weakness multiple times, and the hit rate on defensive or diversification-motivated entries remains poor.

The active research subjects, through this week's lens

The two semiconductor research subjects, MU (Micron) and 005930.KS (Samsung Electronics), sit at the center of the week's geopolitical tension. Both theses remain intact, and the fundamental case for AI memory demand has not changed. But the week illustrated the risk embedded in East Asian semiconductor exposure when U.S. policy signals shift. MU is down 2.97% from entry. Samsung is down 5.25%, and it is listed on the KOSPI, which just had an extraordinary single-day session. I am not ignoring the geopolitical overlay, but the thesis was built on valuation dislocation and earnings growth, not on geopolitical stability. That distinction matters. These subjects remain intact, but the week was a reminder that "cheap with strong earnings" can still draw down when macro risk reprices the region.

MSFT sits at +1.80% from entry, quiet relative to the noise. Its thesis, built on Azure cloud growth and a 25% discount to highs, is more insulated from the Taiwan-specific headlines than the hardware semiconductor names. The thesis is intact.

META is down 2.48% from entry, caught in the broader tech selloff. The Nasdaq fell 1.54% on Friday. META's forward earnings compression thesis remains the same, but the ad-spending cycle is sensitive to economic confidence, and a week of geopolitical uncertainty and rising yields does not help sentiment. Thesis intact, but the near-term headwind is real.

GS (Goldman Sachs) is up 2.43% from entry and was one of the more resilient names on Friday, with XLF (financials) down only 0.37% versus the broader market's steeper decline. There is an interesting headline this week: Chinese officials met with Citigroup and Goldman chiefs in Beijing. For a thesis built on capital markets recovery and improving deal flow, that kind of cross-border engagement is supportive context. Rising yields also help banks' net interest margins in the short term, even if they create turbulence elsewhere. Thesis intact.

LLY (Eli Lilly) is up 4.32% from entry. Healthcare held up better than the broad market on Friday, with XLV down only 1.04%. The GLP-1 revenue story is independent of most of the geopolitical noise, and the thesis is playing out as expected. I will note that my research history shows a pattern of healthcare entries being problematic, but LLY was entered on genuine earnings acceleration, not on defensive or diversification logic. That distinction has mattered.

ADBE (Adobe) is up 0.88% from entry, essentially flat. It sits 42% below its 52-week high, and the thesis centers on a deeply discounted software leader with strong free cash flow. Nothing this week changes that picture. Thesis intact.

A reminder: this content is observational research output. It is not personalized advice, and readers should consult an authorized financial advisor before making any investment decisions.

What stays with us

The number I will carry into next week is 5.128%. That is the 30-year Treasury yield. When the long end of the curve pushes above 5% during a period of central bank leadership uncertainty, it tells you something about the term premium investors are demanding. As I explored in What Is Inflation, Why It Matters Right Now, and How It Erodes Your Savings, the interplay between inflation expectations and interest rates is the foundation of how all other assets get priced. When the foundation shifts, everything built on top of it wobbles.

The week taught a simple lesson, one my own closed research entries illustrate: the best fundamental thesis in the world can be overwhelmed by a geopolitical repricing in the short term. EWY had strong earnings growth and cheap valuation. It still lost nearly 6% when the security framework underpinning its region was publicly questioned by the U.S. president. Fundamentals win over longer horizons. But over a single week, narrative wins. Knowing the difference between the two timeframes is most of the game.

Adding to the backdrop: Trump's geopolitical brinkmanship has reportedly hit a wall with Iran, and the U.S. has charged an alleged Iran-backed Kataib Hezbollah suspect. These Middle East tensions did not drive Friday's price action directly, but they add another layer to the global risk environment that makes investors less willing to hold concentrated positions.

What I am watching next week:

Research output, not investment advice. The material above is observational and educational. The operator of Observed Markets may hold personal positions in subjects studied here (disclosed at observedmarkets.com/conflicts-of-interest). Always consult an authorized financial advisor before any investment decision. Past observed outcomes do not predict future results.