ECFA Endgame

Taiwan's share of exports going to China (including Hong Kong) has been declining steadily from its 2020 peak of 43.8% ^[1]. By 2025, the figure had fallen to 25.3% — still Taiwan's largest single export destination, but on a scale no longer comparable to five years earlier.

What drives this trend is not political will, but three market forces:

First, U.S.-China tech decoupling reshaped the procurement chain. The main customers for Taiwan's semiconductors have shifted from China's security, communications, and consumer-electronics sectors toward American AI chips and servers; those orders move through U.S., Japanese, and European channels, bypassing China ^[5].

Second, China's manufacturing upgrade replaced Taiwanese supply. As China's own chemical, machinery, and textile suppliers advanced technologically, their reliance on Taiwanese intermediate goods declined; Beijing's policy preference for "de-Taiwanization" accelerated this substitution ^[6].

Third, Taiwanese firms relocating out of China took the ancillary trade with them. Taiwanese businesses' purchases of Taiwan-made components for their China production have shifted along with their factories to Southeast Asia and India, producing a "trade follows the Taiwanese firm" effect ^[7].

Set against this backdrop, the end of ECFA's early-harvest list looks more like confirmation of a process already under way than the cause of a new rupture.

One question that has been argued over for years is: how much did the ECFA early-harvest list actually contribute to Taiwan's exports?

According to estimates from TAITRA and the DGBAS, the sectors that benefited most (petrochemical feedstocks, machinery components, cotton textiles, and agricultural products) saved tariffs each year through ECFA equivalent to roughly NT$149 billion in export value, about 3.8% of Taiwan's total exports ^[2].

That figure is not zero, but it is not a lifeline either.

The hardest hit were small and medium-sized traditional manufacturers in central and southern Taiwan — petrochemicals, machinery, textiles — that had long relied on ECFA's tariff preferences to stay competitive in the China market. For these firms, the end of ECFA means a shrinking market, rising costs, and a sharp drop in competitiveness ^[8].

Taiwan's export transition has found its biggest support in the U.S. market.

In 2024, Taiwan's exports to the United States grew 26.7% year on year, and the U.S. replaced China as Taiwan's largest export market ^[3]. AI servers, advanced chips, advanced packaging — these are exactly where American tech-sector procurement is concentrated, and Taiwan's position within it is hard to dislodge.

Southeast Asia's absorption capacity is also growing. The relocation of Taiwanese firms has driven Taiwan's exports of intermediate goods to Vietnam, Thailand, and India; ASEAN's overall share of Taiwan's exports rose from 14% in 2020 to 19% in 2025 ^[1].

But this transition carries a structural imbalance: the beneficiaries are mainly electronics and semiconductor exporters, not traditional industry. For SMEs in petrochemicals, machinery, and textiles, Southeast Asia is itself their competitor, not a substitute market ^[9].

Taiwan's share of exports to China has fallen, but that does not mean cross-Strait economic dependence has disappeared. It has simply changed form:

Capital ties: the historical cumulative total of Taiwanese investment in China is estimated to still exceed US$153 billion ^[4]. These assets remaining in China are both a commercial interest for the firms involved and a political hostage. Even where business has stopped, the difficulty of liquidation and divestment has led most Taiwanese firms to stay put and watch ^[10].

Residual technological dependence: in certain fields, Taiwan still depends on the China market — from sourcing components for semiconductor equipment to supplies of specific chemical raw materials ^[11]. Full technological decoupling is harder and slower than trade decoupling.

Personnel exchange: with cross-Strait exchanges restricted, legal disputes, asset protection, and personnel management for Taiwanese firms in mainland China have grown more complicated, and the uncertainty around legal and personal-safety risk has risen ^[12].

The national level: geopolitically, the end of ECFA removes one of Beijing's important economic levers over Taiwan. With the share of exports to China falling from 43% to 25%, Beijing's ability to use trade sanctions as coercion has already weakened considerably ^[13]. But the remaining Taiwanese-owned assets and personnel in China are still a chip Beijing can squeeze, especially under conditions of strategic ambiguity.

The industrial-intermediary level: for Taiwanese SME suppliers in traditional manufacturing, the short-term pressure is real — ECFA's tariff preferences have vanished, Chinese substitutes have grown more competitive, and Southeast Asian markets are also competing for their orders. The transition window is 2–3 years; some firms will make it, some will not ^[14]. Firms in the electronics and semiconductor supply chain do not feel this pressure, but if their customers (non-AI chipmakers other than TSMC) still do business in China, they too could face indirect impact.

The SME (S2) level: the most directly affected are SMEs in sectors that benefited from the ECFA early-harvest list (petrochemicals, machinery, cotton textiles). Next are upstream and downstream suppliers dependent on Taiwanese firms' supply chains in China. Most SMEs outside these two categories feel the effect indirectly — mainly through the overall business cycle and the exchange rate (reduced exports to China → a shift in Taiwan's export momentum → effects on the exchange rate and inflation) ^[15].

The end of the ECFA early-harvest list is a milestone in Taiwan's declining economic dependence on China, not a sudden rupture. Because of that, the impact on Taiwan overall has been smaller than the outside world expected.

But this framing also makes it easy to overlook several things. First, the fruits of decoupling have gone mainly to the electronics and semiconductor industries, while SMEs in traditional manufacturing bear a pure burden, with no accompanying compensation ^[16]. Second, the residual dependence in capital and personnel still means the "informal linkage" across the Strait is far tighter than official statistics suggest. Third, Taiwan's large-scale pivot to the U.S. market has raised the concentration of Taiwan's exports to the U.S. — trading one form of external dependence for another.

What matters is how Taiwan, after decoupling, builds more resilient and more diversified external economic ties. ECFA's story is not over yet.