XCEM: Skipping China Makes Sense, But Some Peers Do It Better

Columbia EM Core ex-China ETF (XCEM) provides emerging-market exposure excluding China and Hong Kong, but has concentrated risk: technology represents 48% of assets, while Taiwan and Korea account for 35% and 25%, respectively. TSMC and Samsung alone comprise 15.4% and 9.4% of the portfolio. Although XCEM has outperformed the broad EM benchmark since inception and over the past year, it trails ex-China peer FRDM on total and risk-adjusted returns.

Analysis

XCEM is not a diversified China-avoidance allocation; it is effectively a concentrated semiconductor and Korean-memory trade wrapped in an EM label. Its return path will be driven more by AI capex, foundry utilization, DRAM/NAND pricing, and KRW/TWD moves than by the growth or policy cycle across emerging markets. That concentration can be advantageous while AI infrastructure spending remains strong, but it creates correlated downside if hyperscaler capex guidance softens or the semiconductor inventory cycle turns.

The second-order issue is valuation and overlap: investors using XCEM alongside SMH, SOXX, QQQ, or direct TSM exposure may be materially more exposed to the same AI supply-chain factor than portfolio-level geographic labels imply. A weaker dollar is a near-term tailwind for local-currency EM returns, but TWD/KRW appreciation can also dilute USD-reported export competitiveness over a 6-18 month horizon. Geopolitical risk is asymmetric: Taiwan-related stress would impair both the fund's largest holding and its core diversification premise simultaneously.

No standalone trade is warranted on the ETF comparison alone. The more actionable implication is portfolio hygiene: treat XCEM as a high-beta Asian semiconductor allocation, not as a broad EM core holding. A durable rotation toward EM domestic-demand exposure would favor vehicles with lower Northeast Asia technology concentration, while continued AI earnings upgrades would keep direct foundry exposure more efficient than owning the broader basket.

Contrarian view: investors may overpay for the perceived safety of excluding China while retaining substantial regional and hardware-cycle concentration. If Chinese stimulus broadens into a regional electronics demand recovery, China-excluding funds could underperform broad EM despite avoiding headline regulatory risk; conversely, a semiconductor correction would expose the limited protection offered by the geographic screen.

Key Decisions for Investors

• Do not classify XCEM as a core diversified EM position; cap it as a satellite Asia-semiconductor exposure and aggregate its factor exposure with TSM, SMH, SOXX, QQQ and Korean memory holdings immediately.
• For a 1-3 month AI-capex continuation view, prefer direct long TSM over XCEM: direct ownership offers cleaner participation in foundry utilization and advanced-node pricing, while XCEM dilutes upside with non-TSM holdings. Reassess if TSM cuts full-year revenue growth or capex guidance.
• For EM exposure without adding semiconductor concentration, pair any XCEM allocation with a domestic-demand-oriented EM vehicle or reduce XCEM in favor of broader EM exposure; monitor relative performance versus EEM and FRDM over the next two earnings cycles.
• Use a downside-risk trigger rather than preemptive hedging: reduce XCEM/TSM exposure if major hyperscalers collectively signal AI-capex deceleration, or if TSM's utilization and gross-margin outlook falls below management's prior range. These would challenge the fund's dominant earnings driver over the following 6-12 months.