TW-F-03: The New Southbound Policy and Taiwan's Southeast Asia Turn (1994–2026)

1. Key Takeaways

• Taiwan's southbound turn is a thirty-year idea that succeeded only on its third articulation, and then only partially. The Lee Teng-hui administration's 1994 Go South policy (南向政策, Nanxiang Zhengce) and its 1997 second wave, and the Chen Shui-bian administration's 2002 reaffirmation, all attempted to redirect Taiwanese outbound investment from the People's Republic of China toward Southeast Asia. All three precursors were overwhelmed by the structural economics of the China option — linguistic proximity, wage differentials, supply-chain agglomeration in the Pearl River and Yangtze deltas — and by the 1997–98 Asian Financial Crisis, which destroyed the near-term investment case for ASEAN just as the second wave launched. The 2016 New Southbound Policy (新南向政策, Xin Nanxiang Zhengce) is the third major articulation, and the first to coincide with — rather than fight against — the underlying market currents.

• The New Southbound Policy's distinctive design was people-centred because the diplomatic toolkit was unavailable. Taiwan has formal diplomatic relations with none of the eighteen target countries (the ten ASEAN members, six South Asian states, and Australia–New Zealand). It cannot sign conventional free-trade agreements with most of them, cannot exchange ambassadors, and cannot join the regional architecture (RCEP, most ASEAN-centred forums). The Tsai Ing-wen administration therefore built the policy around instruments that do not require recognition: visa liberalisation, scholarship and degree pipelines, medical and agricultural cooperation programmes, SME linkages, and tourism promotion. The "people-centred" (以人為本) framing was both a normative choice and a structural workaround — the only channels open were sub-state and societal.

• The policy's deepest driver was the post-2014 political settlement on cross-Strait economic exposure. The 2014 Sunflower Movement's blockage of the Cross-Strait Services Trade Agreement marked the end of the Ma-era integration trajectory and established a durable domestic-political ceiling on further institutionalised economic dependence on the PRC. The New Southbound Policy, announced in Tsai's 20 May 2016 inaugural, was the affirmative half of that settlement: if Taiwan would not deepen westward, it needed a credible eastward-and-southward alternative. The policy is therefore best read not as trade promotion but as the economic-geography expression of the post-Sunflower consensus on de-risking from China.

• The headline result is real but the attribution is contested. By 2023, Taiwanese approved outbound investment into the eighteen New Southbound countries exceeded approved investment into the PRC for the first time [TBD-VERIFY: the commonly cited 2023 figures are approximately US$5.2 billion to NSP countries against approximately US$3.0 billion to China, per Investment Commission/Investment Review data — confirm series and definitions], and the China share of approved outbound investment had fallen from a peak above 80 per cent (circa 2010) to roughly a tenth [TBD-VERIFY: peak-year share and 2023–2025 shares]. Critics note that the same crossover occurred for Japanese, Korean, and Western firms without any southbound policy — rising Chinese wages, the post-2018 US–China tariff war, COVID-era disruption, and the "China+1" diversification wave moved capital south regardless. The honest assessment is that the New Southbound Policy lowered frictions and built infrastructure for a relocation the market was driving anyway.

• The people flows are the policy's most transformative and least-celebrated dimension. Taiwan hosts approximately 750,000–800,000 Southeast Asian migrant workers [TBD-VERIFY: end-2024/2025 stock by nationality — Indonesia, Vietnam, the Philippines, and Thailand are the four sending states], who staff its factories, construction sites, fishing fleet, and — critically — its long-term-care system for a hyper-aged society. A separate marriage-migration cohort of Southeast Asian spouses, accumulated since the 1990s, has produced a "new second generation" (新二代, xin er dai) of bicultural Taiwanese now reaching adulthood, whom the policy explicitly framed as a talent bridge. Taiwan's demographic cliff makes these corridors structurally permanent and likely to deepen regardless of which party governs.

• The labour-rights record is the policy's standing contradiction. A people-centred southbound policy coexists with documented and persistent abuses in the very people flows it celebrates: forced-labour findings in Taiwan's distant-water fishing fleet [TBD-VERIFY: the US Department of Labor's listing of Taiwan-caught fish among goods produced by forced labour, and the Greenpeace/EJF investigation series], the exclusion of approximately 200,000+ household caregivers from the Labor Standards Act's protections, and broker-fee debt structures that Southeast Asian governments and civil-society groups have repeatedly criticised. The gap between the policy's soft-power narrative and the migrant-labour record is the sharpest critique levelled by the target countries themselves.

• The geopolitical value of the policy runs through Washington as much as through the region. From roughly 2018, US officials explicitly framed the New Southbound Policy as complementary to the Free and Open Indo-Pacific strategy, and the post-2020 supply-chain-resilience agenda gave the southbound logic a security rationale Washington could underwrite. Taiwanese semiconductor-ecosystem suppliers' moves into Vietnam, Malaysia, Singapore, Thailand, and India — and the courtship of an Indian fab ecosystem [TBD-VERIFY: the status of the Tata–PSMC Dholera fab partnership and any TSMC-adjacent Indian engagements] — fused the southbound turn with the US-aligned "friend-shoring" architecture. The policy thereby acquired a geopolitical patron its 1990s precursors never had.

• The target countries take the factories while keeping the one-China observance — and this hedging is the policy's operating environment, not its failure. Vietnam absorbed the largest share of Taiwanese relocation while policing its one-China commitments; the Philippines' post-2022 alignment shift created the most permissive environment (and the most direct stake, given the West Philippine Sea frontline and the 150,000-strong Filipino worker population in Taiwan); India built a labour-mobility agreement and a semiconductor courtship while avoiding any political signalling. Beijing's counter-pressure — diplomatic démarches, economic inducements, tourism weaponisation — has constrained the policy's visibility but not its substance, because the target states want Taiwanese capital, technology, and remittance corridors more than they fear symbolic ambiguity.

• The Lai Ching-te administration has continued the policy while folding it into a larger economic-security frame, raising the institutionalisation question. Under the post-2024 "integrated diplomacy" framing and the Vice Premier Cheng Li-chiun-led trade-negotiation architecture, the New Southbound brand persists [TBD-VERIFY: current official branding — "New Southbound Policy 2.0"/"NSP-plus" formulations have circulated without a definitive relaunch], but its content is increasingly indistinguishable from generic supply-chain de-risking. The open question for the 2030s is whether the southbound turn has become a structural feature of Taiwan's political economy that would survive a KMT presidency — the evidence of the migrant-labour dependence, the investment stock, and the second-generation demographic suggests yes — or whether it remains a DPP-branded programme vulnerable to the next alternation.

• As a genre, the New Southbound Policy is the leading case study in diplomacy without diplomatic relations. Taiwan demonstrated that a middle-sized economy under recognition blockade can build dense, functional, semi-institutionalised relationships through education, labour, health, agriculture, and investment channels — a toolkit closer to development cooperation than to classical statecraft. The closest comparator, South Korea's New Southern Policy (2017) [TBD-VERIFY: Moon Jae-in administration launch date and the Yoon-era successor framing], pursued similar diversification with full diplomatic relations and still achieved arguably less relative repositioning — which suggests the binding constraint on southbound strategies is firm-level economics, not diplomatic form.

2. The Southbound Prehistory (1994–2016)

2.1 The Lee-Era First Articulation: the 1994 Go South Policy

The southbound idea entered Taiwanese statecraft in the early Lee Teng-hui years, and it entered as a counter-China policy from the first. Following the 1987 lifting of martial law and the 1990–91 legalisation of indirect mainland investment, Taiwanese manufacturers — the Taishang (台商) — began moving labour-intensive production across the Strait at extraordinary speed: footwear, garments, toys, and then electronics assembly migrated to Guangdong and Fujian, drawn by shared language, kinship networks, and wages a fraction of Taiwan's. By 1993 the Lee administration had concluded that this westward tide was creating a strategic dependence that Beijing could weaponise, and that the state should actively redirect a portion of the outflow toward Southeast Asia.

The first Go South policy was promulgated in 1993–94 as a set of Ministry of Economic Affairs investment-encouragement guidelines targeting seven ASEAN economies — Thailand, Malaysia, Indonesia, the Philippines, Singapore, Vietnam, and Brunei [TBD-VERIFY: the precise country list and the formal title of the 1994 guidelines, conventionally dated to a January 1994 Executive Yuan adoption]. Its instruments were investment-guarantee agreements, state-linked investment vehicles, official credit support, and high-level political signalling. The signalling was the most visible element: in February 1994 Lee Teng-hui conducted his celebrated "vacation diplomacy" (度假外交) tour — golf-and-holiday visits to the Philippines, Indonesia, and Thailand during which he met President Fidel Ramos, President Suharto, and Prime Minister Chuan Leekpai in deliberately informal settings that allowed the host governments to deny any breach of their one-China commitments [TBD-VERIFY: the precise meeting formats; the Suharto meeting at Bali is the best documented]. The tour established the template that every subsequent southbound push has used: substantive engagement wrapped in deniable form.

The first wave produced a measurable surge. Taiwanese approved investment into ASEAN rose sharply in 1994–96, and Taiwan became one of the largest foreign investors in Vietnam (where it ranked among the top sources of committed FDI through the 1990s and 2000s [TBD-VERIFY: Vietnam FDI rankings by source country for the 1990s]), a major investor in Malaysia's Penang electronics cluster, and a significant presence in Thai and Indonesian manufacturing. But the redirection was relative, not absolute: mainland-bound investment continued to grow faster, and much of the ASEAN-bound flow would likely have occurred without the policy, since the same labour-cost arbitrage logic applied.

2.2 The 1996–1997 Second Wave and the Asian Financial Crisis

The policy's second articulation came in 1996–97, in the shadow of the Third Taiwan Strait Crisis. The 1995–96 missile crisis hardened Lee's conviction that economic dependence on the mainland was a security exposure, and in September 1996 he announced the jieji yongren (戒急用忍, "no haste, be patient") doctrine — restrictions on large-scale, high-technology, and infrastructure investment in the PRC, including a US$50 million ceiling on individual mainland investment projects [TBD-VERIFY: the ceiling figure and the categories restricted]. The Go South policy was simultaneously extended and re-emphasised in 1997 as the affirmative complement to the westward restrictions: capital deterred from China should flow south.

The timing was catastrophic. The Asian Financial Crisis broke in July 1997 with the Thai baht collapse and tore through precisely the economies the policy targeted — Thailand, Indonesia, Malaysia, the Philippines. Currency collapses, banking crises, and (in Indonesia) the fall of Suharto and the May 1998 anti-Chinese riots — in which ethnic-Chinese businesses, including Taiwanese-linked ones, were attacked — destroyed the near-term investment case. Taiwanese ASEAN-bound investment fell sharply after 1997, while the PRC, whose currency held and whose growth continued, looked safer than ever. The structural lesson the Taishang drew from 1997–98 was the inverse of the one the government intended: China was the stable destination, Southeast Asia the risky one.

2.3 The Chen-Era Third Push and the Westward Tide

The Chen Shui-bian administration (2000–2008) inherited both the southbound idea and the westward problem, and its record illustrates the gap between declared policy and economic gravity. Chen reaffirmed the Go South policy in 2002 [TBD-VERIFY: the 2002 reaffirmation's form — conventionally dated to remarks and an Executive Yuan re-endorsement in August–September 2002], adding new emphasis on Vietnam, which was emerging as the most receptive destination. But Chen's overall cross-Strait economic policy moved the other way: the 2001 Economic Development Advisory Conference replaced jieji yongren with "active opening, effective management" (積極開放,有效管理), substantially liberalising mainland investment, before the 2006 reversal to "active management, effective opening" (積極管理,有效開放) re-tightened it. The liberalising phase coincided with China's 2001 WTO accession and the great mainland boom; the result was the largest westward investment surge in Taiwanese history.

The numbers tell the story of the precursors' defeat. By the late 2000s, the PRC accounted for the overwhelming majority of Taiwan's approved outbound investment — the share peaked at over 80 per cent around 2010, the year the Ma administration signed ECFA [TBD-VERIFY: the conventionally cited peak is 83.8 per cent of approved outbound investment to China in 2010 — confirm against Investment Commission series]. An estimated one million or more Taiwanese were living and working on the mainland at the peak [TBD-VERIFY: estimates of the Taiwanese expatriate population in the PRC range from 400,000 to over 1 million across the 2000s–2010s]; the Taiwanese electronics industry's assembly tier — Foxconn/Hon Hai above all — had built its global position on mainland labour; and China (with Hong Kong) was approaching 40 per cent of Taiwan's total exports. Two southbound waves had not bent this curve at all.

Yet beneath the failed investment redirection, an unplanned societal southbound turn was already under way — and it would prove the more durable legacy of the period. Taiwan formally opened to Southeast Asian contract labour in 1989–1992 (the 1992 Employment Service Act created the legal framework [TBD-VERIFY: the 1989 construction-sector opening under the Fourteen Major Construction Projects and the 1992 Act's sequence]), and the worker population grew from tens of thousands in the mid-1990s to over 300,000 by the mid-2000s, drawn first from Thailand and the Philippines and then increasingly from Indonesia and Vietnam. In parallel, commercial marriage brokerage built the spouse corridor: tens of thousands of Vietnamese and Indonesian women married Taiwanese men across the late 1990s and 2000s, concentrated in rural and working-class households, before regulatory tightening (the 2007–08 prohibition of for-profit brokerage [TBD-VERIFY: the amendment chronology]) slowed the flow. By the time the New Southbound Policy was announced in 2016, Taiwan already had half a million Southeast Asian workers and spouses inside its borders — a societal interpenetration the investment-led Go South policies had never contemplated and never claimed. The 2016 policy's people-centred framing was thus partly a recognition after the fact: it elevated to strategy what migration had already made demographic reality, and it inherited both the asset (the human network) and the liability (the brokerage abuses, the caregiver exclusion, the assimilationist policy legacy) of the unplanned corridor.

2.4 The Lesson Inventory

The failed precursors left a usable inventory of lessons that the 2016 designers explicitly drew on. First, exhortation cannot beat arbitrage: a policy that asks firms to accept higher costs for strategic reasons will fail; a policy that works must ride a cost structure already turning in its favour. By 2016, mainland wages had quintupled from their early-2000s levels, the demographic dividend had ended, and the arbitrage had genuinely reversed — the third wave launched with the market, not against it. Second, investment-only policies are shallow: the 1990s waves built factories but no societal fabric, so each crisis erased them; the 2016 design therefore led with people — students, workers, spouses, patients, tourists — on the theory that societal interlinkage is more durable than capital flows. Third, the diplomatic ceiling is real but survivable: the vacation-diplomacy template proved that substantive relationships could be built beneath the recognition threshold, provided Taiwan supplied things the partners wanted and asked for no political price. Fourth, a southbound policy needs an external patron: the 1990s waves operated in a permissive but indifferent international environment; the 2016 wave would eventually acquire active US alignment, which changed the partners' calculus of risk.

3. The New Southbound Policy (2016–2024): Design and Instruments

3.1 The Tsai-Era Design

Tsai Ing-wen named the reorientation in her 20 May 2016 inaugural address, and the Executive Yuan released the New Southbound Policy Promotion Plan on 5 September 2016 under Premier Lin Chuan, with a presidential-office New Southbound Policy Office under James Huang (黃志芳) providing initial coordination [TBD-VERIFY: the office's disbandment, conventionally dated to 2018 when coordination passed to the Executive Yuan's Office of Trade Negotiations]. The institutional and operational detail is covered in TW-E-02; what matters for this document's arc is the design logic, which differed from the precursors in four ways.

The target set widened from ASEAN to eighteen countries: the ten ASEAN members, six South Asian states (India, Pakistan, Bangladesh, Sri Lanka, Nepal, Bhutan), and Australia and New Zealand. The South Asian extension — above all India — reflected both the scale logic (India as the next labour-cost frontier and the largest unaligned market) and the geopolitical logic (India as the Indo-Pacific's other China-wary democracy). The ANZ inclusion signalled that the policy was about the democratic Indo-Pacific, not merely cheap labour.

The framing was explicitly de-risking, post-Sunflower: the policy was announced by a president elected on the back of the 2014 Sunflower Movement's repudiation of the CSSTA and the Ma-era integration trajectory. Tsai's formulation — that Taiwan should "bid farewell to the past over-reliance on a single market" — made the China referent unmistakable without naming it. The New Southbound Policy was the constructive half of a two-part settlement whose restrictive half was the indefinite shelving of further cross-Strait economic agreements.

The four pillars institutionalised the people-centred workaround: economic and trade collaboration; talent exchange; resource sharing (medical, technological, agricultural); and regional links. Because Taiwan could not sign FTAs with states observing the one-China principle, could not join RCEP, and could not open embassies, the pillars were deliberately weighted toward instruments below the recognition threshold. The result was a foreign policy executed largely by the ministries of education, health, agriculture, and labour rather than by MOFA — a structural novelty.

The metric of success shifted from investment shares to relationship density: the official discourse measured students enrolled, patients treated, visa-free arrivals, sister-hospital agreements, agricultural demonstration sites, and SME matchmaking events alongside trade and investment figures — partly genuine philosophy, partly hedge against the precursors' fate of being judged solely on investment redirection they could not command.

3.2 The Instruments

Visa liberalisation was the cheapest and fastest lever: visa-free entry for Thailand and Brunei (August 2016), the Philippines (November 2017), and progressively eased e-visa and conditional-entry schemes for Indonesia, Vietnam, India, and others [TBD-VERIFY: the full liberalisation chronology and current status of each scheme]. Tourist arrivals from the eighteen countries rose steeply through 2019, partially offsetting Beijing's post-2016 throttling of mainland tour groups — the first demonstration that the southbound turn could blunt a specific PRC coercion instrument.

The education pipeline became the policy's signature. Southeast and South Asian enrolment in Taiwanese tertiary institutions roughly doubled across the first term — from approximately 28,000–32,000 students from NSP countries in 2016 to approximately 55,000–60,000 by 2019–20 [TBD-VERIFY: Ministry of Education enrolment series; Vietnam, Indonesia, and Malaysia are the largest senders] — driven by Taiwan Scholarship expansion, industry-academia "2+i" programmes that combined study with internships, and Taiwanese universities' own demographic desperation: with domestic cohorts collapsing, Southeast Asian recruitment became a survival strategy for mid-tier institutions, aligning institutional self-interest with state policy. The pipeline created a future-elite bet — alumni networks across ASEAN bureaucracies and firms — whose returns mature on a twenty-year horizon.

The medical, agricultural, and SME programmes supplied the resource-sharing pillar: the "One Country, One Centre" (一國一中心) medical-cooperation architecture pairing Taiwanese medical centres with seven partner countries; agricultural demonstration farms and rice/aquaculture technical missions continuing TaiwanICDF's older development-cooperation tradition; and SME credit and matchmaking facilities. Individually modest, these programmes built the ministerial-level working relationships — health ministry to health ministry, agriculture agency to agriculture agency — that substitute for diplomatic infrastructure.

The Taiwan-desk infrastructure rounded out the toolkit: dedicated Taiwan desks in target-country investment agencies, expanded TAITRA offices, the Taiwan-ASEAN studies centres, and bilateral investment-agreement updates where attainable (the updated Taiwan–Philippines bilateral investment agreement of 2017 and the Taiwan–India bilateral investment agreement of 2018 were the headline cases [TBD-VERIFY: signature dates and entry into force]). The 2024 Taiwan–India migrant-worker memorandum of understanding [TBD-VERIFY: signed February 2024, operationalisation status] extended the desk infrastructure into labour mobility, opening a corridor whose long-run significance may exceed the investment agreements.

4. The Results Ledger (2016–2026)

4.1 Trade and Investment: the Crossover

The trade record shows genuine but bounded reorientation. Taiwan's exports to the eighteen New Southbound countries grew substantially across the decade — ASEAN's share of Taiwanese exports rose from roughly the high teens toward the low twenties in percentage terms [TBD-VERIFY: ASEAN-6/ASEAN-10 export shares 2016 vs 2024–25], while the China-plus-Hong-Kong share, which peaked at approximately 43–44 per cent in 2020–21 on pandemic-era electronics demand, declined to the low-to-mid thirties by 2024–25 [TBD-VERIFY: the peak figure and the current share — the decline is among the most-cited data points in the diversification debate]. The export composition matters as much as the shares: a large fraction of the southbound export growth is semiconductors and intermediate goods feeding Taiwanese-owned assembly operations that relocated from China — which means the southbound trade rise and the China trade decline are substantially the same supply chains moving, not new demand discovered.

The investment record contains the policy era's emblematic data point: in 2023, Taiwanese approved outbound investment into the eighteen New Southbound countries exceeded approved investment into the PRC for the first time since records of the comparison began [TBD-VERIFY: approximately US$5.2 billion against approximately US$3.0 billion is the commonly cited pairing; confirm against Investment Commission/Investment Review Office series and check 2024–25 continuation]. Set against the 2010 peak — when China took over 80 per cent of approved outbound investment — the reversal is dramatic. The destination concentrations are Vietnam (the largest single recipient of the relocation wave, especially in electronics assembly and footwear/textiles), Thailand (printed circuit boards above all — the Thai PCB cluster built substantially by Taiwanese firms post-2019 [TBD-VERIFY: the scale of Taiwanese PCB investment commitments in Thailand, frequently cited in the multi-billion-dollar range]), Singapore (the financial and regional-headquarters node), Malaysia and the Philippines (semiconductor assembly-test and electronics), and India (the fastest-growing frontier from a low base).

4.2 The Semiconductor-Era Extensions

The post-2020 chip era fused the southbound turn with the semiconductor diaspora documented in TW-G-03. As TSMC's leading-edge expansion went to Arizona, Kumamoto, and Dresden, the broader ecosystem's cost-sensitive tiers went south: assembly, testing, packaging, substrates, PCBs, and passive components moved into Vietnam, Thailand, Malaysia, and Penang's revitalised cluster, with Taiwanese OSAT and component firms among the leading investors. Foxconn's India expansion — iPhone assembly in Tamil Nadu and Karnataka growing from marginal to a meaningful share of global output [TBD-VERIFY: the India share of iPhone assembly, rising across 2023–26] — made a Taiwanese firm the central corporate instrument of India's electronics-manufacturing ambitions.

The India fab courtship is the frontier case. The approved Indian fab projects centre on the Tata Electronics–Powerchip (PSMC) partnership at Dholera, Gujarat — a mature-node fab with Taiwanese technology transfer [TBD-VERIFY: project status, node range, and production timeline] — alongside assembly-test projects with Taiwanese participation. TSMC itself has not committed an Indian fab, and the Taiwanese government has been notably cautious about leading-edge transfer southward; the southbound semiconductor story is one of ecosystem-tier migration, not crown-jewel relocation. That distinction preserves the silicon-shield logic (the strategic framing treated in TW-M-01 and TW-G-03) while still giving India and ASEAN a growing stake in Taiwanese industrial health.

4.3 The People Flows

The migrant-worker corridor is the largest human dimension. Approximately 750,000–800,000 Southeast Asian migrant workers were employed in Taiwan by 2024–25 [TBD-VERIFY: Ministry of Labor stock figures — the conventional breakdown places Indonesia and Vietnam at roughly 250,000–280,000 each, the Philippines at roughly 150,000, Thailand at roughly 60,000–70,000], concentrated in manufacturing, construction, the distant-water fishing fleet, and household long-term care. The caregiver segment — overwhelmingly Indonesian women — is the de facto backbone of elder care in a society aging faster than nearly any other, and the dependence deepens every year as Taiwan's working-age population shrinks (the demographic complementarity treated in TW-O-01 §5). The 2024 India labour MOU extends the corridor to a fifth major sending state.

The marriage-migration generation supplies the policy's most distinctive societal layer. Taiwan's "new residents" (新住民, xin zhumin) — foreign spouses accumulated since the 1990s — number approximately 580,000–600,000, of whom Southeast Asian spouses (led by Vietnamese) constitute roughly 170,000–200,000 [TBD-VERIFY: National Immigration Agency series and the PRC/SEA split]. Their children — the "new second generation," several hundred thousand strong and now reaching university age and the workforce — were explicitly framed by the policy as a bicultural talent reservoir: heritage-language programmes entered the national curriculum (Southeast Asian languages became electives in the 2019 curriculum reform), and scholarship and employment schemes targeted the cohort. The framing has been criticised as instrumentalising — assigning the children of working-class migration the burden of national strategy — but it marked a real discursive shift from the assimilationist 1990s–2000s treatment of Southeast Asian spouses.

The labour-rights frictions cut against the people-centred narrative. Taiwan's distant-water fishing fleet — the world's second largest — has faced repeated forced-labour findings: the US Department of Labor added Taiwan-caught fish to its List of Goods Produced by Child Labor or Forced Labor in 2020 [TBD-VERIFY: listing year and current status], and Greenpeace, the Environmental Justice Foundation, and Taiwanese NGOs have documented wage withholding, excessive hours, and deaths at sea among the largely Indonesian and Filipino crews. Household caregivers remain outside the Labor Standards Act, without statutory minimum wage, rest-day, or hours protections [TBD-VERIFY: the status of the long-pending household-services legislation]. Broker-fee debt bondage persists despite direct-hiring reforms. Sending-state governments — Indonesia most pointedly — have periodically suspended or threatened to suspend deployments over protection disputes [TBD-VERIFY: the Indonesian placement-cost dispute of 2020–21 and its resolution]. The critique is not peripheral: it is the gap between the policy's self-presentation and its largest actual people flow.

4.4 The Soft-Power Record

The tourism file deserves separate note because it was the arena in which the southbound turn first directly duelled a PRC coercion instrument. Beijing cut mainland group tourism to Taiwan after the 2016 transition — arrivals from the PRC fell from a 2015 peak of approximately 4.2 million to under 2.7 million by 2019 and to effectively zero in the post-COVID, post-2022 period [TBD-VERIFY: the arrival series and the current status of the group-tour ban]. The New Southbound markets filled a substantial share of the gap: arrivals from the eighteen countries rose from roughly 1.5 million (2015) to approximately 2.5–2.7 million (2019) on the back of the visa liberalisations [TBD-VERIFY: Tourism Administration series], and in the post-pandemic recovery the southbound markets — alongside Japan and Korea — became the structural core of Taiwanese inbound tourism. The substitution was economically imperfect (per-capita spending differed; hotel and tour operators geared to mainland groups suffered real losses that fed KMT-aligned criticism) but strategically instructive: it was the first demonstrable case of the southbound architecture absorbing a deliberate PRC economic punishment, and it permanently reduced the tourism lever's coercive value.

The soft-power ledger is genuinely positive if hard to quantify. The COVID-era mask diplomacy of 2020 — donations to Southeast and South Asian partners under the "Taiwan can help" banner — gave the policy a visible humanitarian register at the precise moment Beijing's pandemic diplomacy was souring regional opinion. The Audrey Tang-era digital-governance outreach (open-government tooling, digital-democracy workshops, the civic-tech network across Asian democracies) supplied a distinctive niche no competitor occupied. Taiwanese popular culture, food, tourism, and the university experience of tens of thousands of Southeast Asian graduates built a societal familiarity that the 1990s waves never attempted. Against this stands the hard limit: soft power has not moved any target state one millimetre on recognition, international-organisation votes (the exclusion architecture is treated in TW-N-01), or public alignment — nor was it realistically designed to.

5. The Geopolitical Dimension

5.1 The US Alignment

The New Southbound Policy launched as a Taiwanese hedge and matured into a node of the US Indo-Pacific architecture. From 2018–19, AIT and State Department officials explicitly paired the policy with the Free and Open Indo-Pacific strategy — joint Taiwan–US programmes under the Global Cooperation and Training Framework (GCTF) ran workshops for Southeast and South Asian officials on public health, cybersecurity, energy, and women's empowerment, effectively making Taiwan a junior co-provider of US-aligned regional public goods [TBD-VERIFY: the GCTF's expansion chronology and the addition of Japan and Australia as full partners]. The post-2020 supply-chain-resilience agenda completed the convergence: when Washington began pressing semiconductor and electronics supply chains to exit China, the Taiwanese firms executing the exit moved precisely along the southbound corridors the policy had spent four years lubricating. The "China+1" wave was a multinational phenomenon, but for Taiwanese firms it had a ready-made policy scaffold — and for the policy, the wave supplied the market force its precursors had fatally lacked.

5.2 The PRC Counter-Pressure

Beijing's response operated on three tracks. Symbolic-diplomatic: pressure on target states to refuse ministerial visits, downgrade Taiwan-desk visibility, and reaffirm one-China formulations — pressure that intensified around each Taiwanese electoral cycle and each US–Taiwan upgrade. Economic competition: the Belt and Road Initiative's infrastructure offer dwarfed anything Taiwan could finance, and Chinese FDI competed directly in several target economies; Beijing could also punish Taiwan directly, as with the post-2016 tour-group throttling and the selective agricultural import bans (pineapple, 2021) that the southbound markets partially absorbed. Recognition raids: the post-2016 capture of Taiwan's remaining diplomatic partners (treated in TW-F-02 and TW-N-01) proceeded in parallel, reminding target states of the costs of visibility.

The counter-pressure shaped form more than substance. No target state abandoned its one-China observance; none stopped taking Taiwanese factories, students, or technical cooperation either. The equilibrium that emerged — hedging by every partner — is best understood as the policy's permanent operating environment: the target states harvest Taiwanese capital and remittance corridors while denying Beijing any pretext, and Taiwan accepts invisibility as the price of access.

5.3 The Regional Reception's Texture

The reception varied with each partner's China exposure and labour interests. Vietnam became the largest economic beneficiary and the most disciplined political hedger — absorbing the biggest share of Taiwanese relocation while policing symbolic boundaries, a posture consistent with its general great-power balancing. The Philippines moved furthest: the post-2022 Marcos-era alignment shift, the West Philippine Sea confrontation (PH-F-06), and the explicit Philippine military acknowledgment that a Taiwan contingency would involve the Philippines created the most permissive environment among the eighteen — eased visa regimes, expanded economic-zone courtship of Taiwanese chip-adjacent investment, and an unusual willingness to discuss Taiwan in security terms, sharpened by the 150,000-plus Filipino workers who would need evacuation in a crisis. India ran a twin-track courtship — semiconductor and electronics investment pursued energetically, labour mobility opened, think-tank and parliamentary contacts thickened — while keeping official formulations rigorously unchanged; the India track is the policy's largest long-run upside and its slowest-moving file. Indonesia and Thailand engaged primarily as labour-sending and investment-hosting states, with the protection of their nationals in Taiwan a recurring bilateral irritant that gave them practical leverage. Singapore and Malaysia maintained their longstanding quiet functionality. The labour-sending states' interests deserve emphasis: for Jakarta, Manila, and Hanoi, the Taiwan relationship is substantially about their citizens' wages and safety — which makes Taiwan's migrant-labour record not merely a rights issue but the operative bilateral agenda.

6. The Lai Continuation (2024–2026)

6.1 Institutionalisation and Rebranding

The Lai Ching-te administration inherited the New Southbound Policy as an eight-year-old going concern and faced the second-generation policy problem: continue, relaunch, or absorb. The observable choice has been absorption-with-continuity. Lai's foreign-policy framing of "integrated diplomacy" (整合外交) folds the southbound architecture into a broader values-democracy-economics triad rather than treating it as a standalone flagship; the trade-negotiation machinery consolidated under Vice Premier Cheng Li-chiun's Office of Trade Negotiations carries the operational weight (the post-2024 re-architecture covered in TW-E-02 §11); and "New Southbound Policy 2.0" / "NSP-plus" formulations have circulated in official and semi-official discourse without a definitive relaunch document [TBD-VERIFY: the current official branding as of mid-2026 — whether a formal second-generation plan has been promulgated or the 2016 framework simply rolls forward]. The hung Legislative Yuan (TW-D-02) has constrained any budget-intensive relaunch, and the KMT-TPP bloc's budget actions in 2025–26 trimmed several programme lines [TBD-VERIFY: specific NSP-related budget cuts in the 2025–2026 budget battles]. The legislative politics are notable for what they are not: the KMT-TPP bloc has contested the policy's cost-effectiveness and individual programme lines, but no major party proposes withdrawing from the southbound architecture itself — the contestation is managerial, not directional, which is itself evidence of the policy's migration from partisan flagship toward settled state practice.

The substantive Lai-era overlay is economic security. The post-April 2025 Trump-2 reciprocal-tariff architecture (TW-E-02 §12, TW-D-03) scrambled the calculus of every relocation decision: Vietnamese and Thai assembly operations built to serve the US market now carried their own tariff exposure, and the administration's response — supply-chain-resilience incentives, US-market production commitments, screening of China-linked transshipment through southbound operations — effectively merged the southbound logic into a generalised friend-shoring management problem. Where the 2016 policy asked "how does Taiwan reduce China dependence?", the 2025–26 version asks "how does Taiwan position its offshore production map inside a US-centred trade regime?" — a question in which the eighteen southbound countries are venues rather than partners.

6.2 The Structural Assessment: What the Policy Changed vs What the Market Did

A decade in, the attribution debate admits a reasonably disciplined answer. What the market did anyway: the investment crossover. Rising Chinese wages, the 2018 tariff war, COVID, and the China+1 wave would have moved Taiwanese cost-sensitive manufacturing south under any government — Japanese and Korean flows followed the identical trajectory without any southbound policy. The KMT critique (the contestation framing in TW-E-02 §13) is on its strongest ground here. What the policy plausibly changed: the friction coefficients and the societal layer. Visa regimes, investment agreements, education pipelines, medical networks, labour corridors, and the Taiwan-desk infrastructure lowered the cost of every individual relocation and built the people-flow architecture that pure market forces would not have produced — no market mechanism doubles scholarship pipelines or creates One Country, One Centre hospitals. What neither achieved: dependence elimination. China-plus-Hong-Kong remains Taiwan's largest export market by a wide margin [TBD-VERIFY: 2025 share — low-to-mid 30s per cent is the conventional figure]; the critical-dependence map still runs through the Strait in specific categories (the residual ECFA early-harvest exposure treated in TW-E-01; mainland-sourced intermediate inputs; the Taishang operations that did not move because their market is China itself). Diversification has been real, partial, and concentrated in the US-facing export segment — which means Taiwan has traded a portion of its China concentration for a deepened US concentration, a substitution the 2025 tariff shock made suddenly visible.

The critical-dependence map, read carefully, shows where the diversification stops. Taiwan's import dependence on the mainland for specific intermediate inputs, rare-earth-processed materials, and certain electronic components persists; the Taishang firms serving China's domestic market — retail, food, services, and the mainland-market-facing manufacturing tier — have no southbound substitute and remain hostage to cross-Strait weather; and the financial exposure of Taiwanese banks' mainland books, though reduced from its 2014 peak [TBD-VERIFY: Financial Supervisory Commission mainland-exposure series], remains material. Conversely, the dependencies running the other way have grown: the PRC's residual reliance on Taiwanese semiconductors gives the relationship its remaining mutual-hostage structure, while the eighteen southbound countries now hold Taiwanese-owned production whose disruption in a Strait crisis would propagate through their own employment and exports — a quiet regionalisation of Taiwan risk that no communiqué acknowledges.

The honest ledger therefore reads: the southbound turn succeeded as relationship infrastructure and succeeded as accompaniment to a market-driven relocation, while the strategic-autonomy promise — a Taiwan no single external actor can economically coerce — remains unredeemed and may be unredeemable for an economy of Taiwan's size and trade structure.

7. The Policy in Perspective

7.1 The Middle-Power Economic-Statecraft Genre

The New Southbound Policy has become the standard case study in diplomacy without diplomatic relations. Its toolkit — scholarships, hospitals, farms, visas, labour corridors, SME finance, sub-ministerial working groups — demonstrates that a recognition-blockaded state can construct functional bilateralism of considerable density, provided it supplies concrete value and demands no political acknowledgment. The Taiwanese case suggests three genre rules: lead with what partners' domestic constituencies want (jobs, remittances, degrees, medical care); accept invisibility as a structural cost rather than fighting it; and let firms and universities, not the foreign ministry, be the principal agents.

The comparative set sharpens the assessment. South Korea's New Southern Policy, launched by the Moon Jae-in administration in November 2017 [TBD-VERIFY: launch date and the formal "New Southern Policy Plus" 2020 upgrade; the Yoon administration's 2022 absorption of it into the Korean Indo-Pacific Strategy], pursued near-identical diversification (from China dependence, toward ASEAN and India, with a "3Ps" people-prosperity-peace framing that visibly echoed Taipei's four pillars) with the full diplomatic toolkit Taiwan lacked — embassies, FTAs, summit diplomacy, RCEP membership. Its results were similar in kind: substantial investment growth into Vietnam and India, modest reduction of China trade share. The comparison is analytically valuable because it isolates the variable: if the diplomatically unconstrained version achieved no categorically better diversification, then the binding constraint on southbound strategies is firm-level economics and market structure, not diplomatic form — which both vindicates Taiwan's workaround toolkit and caps expectations for it. Japan's ASEAN position — built over five decades of ODA, production networks, and the 1977 Fukuda Doctrine's non-hegemonic framing — represents the genre's mature endpoint: what Taiwan's education and investment pipelines might compound into by mid-century, if sustained. The small-state hedging literature (Kuik, Goh, and the ASEAN-school analyses) frames the receiving side: the eighteen target states treat Taiwan as one hedging instrument among several, valuable precisely because it is economically substantial and politically unthreatening.

The genre's internal critique also deserves recording, because the target-country scholarship has made it consistently. Southeast Asian analysts have noted that the New Southbound Policy was designed in Taipei about the region rather than with it — that the eighteen-country architecture aggregates radically different partners into a single Taiwanese strategic category; that the policy's benefits flow disproportionately to Taiwanese firms and universities (the relocated factories repatriate profits, the student recruitment serves Taiwanese institutional survival); and that the "people-centred" vocabulary sits uneasily atop a labour regime the sending states have repeatedly contested. The fairest reading is that the policy is mutually beneficial but asymmetrically designed — a feature it shares with every great- and middle-power regional strategy, and one Taipei has partially answered through the genuinely demand-driven programmes (medical referral, agricultural missions, the labour corridors themselves, which exist because sending states want the remittances). The critique matters less as indictment than as a guide to the policy's durability: the components the partners themselves demand will outlive the components Taipei merely promotes.

7.2 The 2030s Questions

Three questions carry the southbound arc into the 2030s frames set out in TW-O-01.

The contingency question (TW-O-01 §2): every cross-Strait scenario stresses the southbound architecture differently. In a blockade-coercion scenario, the southbound production map becomes Taiwan's economic continuity plan — offshore Taishang operations would be the surviving revenue base — while the 750,000-plus Southeast Asian workers in Taiwan become eighteen capitals' most direct stake in the island's safety, and the evacuation problem (Manila's and Jakarta's above all) becomes a regional crisis-planning fact that quietly internationalises the Taiwan question. The southbound relationships are thus dual-use: economic in peacetime, consular-humanitarian leverage in crisis.

The demographic question (TW-O-01 §5): Taiwan's fertility collapse makes the migration corridor's deepening close to arithmetically certain — the long-term-care system, the construction sector, the fishing fleet, and increasingly mid-skill manufacturing cannot staff themselves domestically. The policy choice the 2030s forces is whether Taiwan converts a rotational guest-worker regime into a settlement-and-citizenship regime: the intermediate-skilled-worker pathway opened in 2022 [TBD-VERIFY: the retention-pathway scheme's uptake] points toward conversion, and the new-second-generation cohort supplies the precedent of successful incorporation. A Taiwan that is 5–10 per cent Southeast Asian-origin by mid-century is a plausible trajectory — at which point the southbound turn ceases to be foreign policy and becomes demography.

The institutionalisation question: would a KMT presidency keep the architecture? The 2024 KMT platform criticised the policy's returns while proposing no dismantlement, and the structural facts — the investment stock, the labour dependence, the university enrolment dependence, the second-generation electorate — have created constituencies in every camp. The likeliest KMT-era outcome is rebranding plus cross-Strait re-engagement layered on top, rather than reversal: the southbound turn appears to have crossed the threshold from programme to structure.

8. Conclusion

The thirty-year arc from the 1994 Go South policy to the 2026 New Southbound architecture is a study in the limits and the patience of economic statecraft. The idea was constant across three articulations — that Taiwan's safety requires that its economic centre of gravity not sit on the other side of the Strait — but the first two articulations failed because they asked firms to defy arbitrage, and the third succeeded, partially, because the arbitrage finally reversed. The policy's real innovations were not economic at all: the people-centred toolkit that built functional bilateralism beneath the recognition threshold; the education and labour corridors that converted foreign policy into demography; the fusion with the US-led supply-chain realignment that gave the southbound turn the external patron its precursors lacked.

The ledger as of 2026 is honest rather than triumphant. The investment crossover of 2023 is real and the China investment share's collapse from its 2010 peak is historic, but the market deserves much of the credit; the export-market dependence on China has declined without disappearing, and a new dependence on US market access has taken shape beside it; the soft-power gains are genuine and the migrant-labour record undercuts them; the regional partners have taken everything Taiwan offered while conceding nothing Beijing watches. What the southbound turn has unambiguously achieved is irreversibility: the factories, the students, the spouses, the caregivers, the second generation, and the eighteen governments' quiet stakes in Taiwan's continuity have made the Southeast Asia relationship a structural feature of Taiwanese national life rather than a programme of any administration. Whether that structure can bear strategic weight — in a coercion scenario, in a demographic emergency, in a post-alternation Taipei — is the question the 2030s will answer.

---

Primary Sources Consulted:

1. Executive Yuan, Republic of China, New Southbound Policy Promotion Plan (新南向政策推動計畫, 5 September 2016) and the subsequent Work Plan (November 2016); the mid-term review documents 2018–2024; the Lai-era trade-negotiation re-architecture documentation under Vice Premier Cheng Li-chiun.
2. Ministry of Economic Affairs (MOEA), Investment Commission / Investment Review Office, approved outbound-investment statistics by destination 1991–2026 — the series underlying the China-share peak (c. 2010) and the 2023 NSP-over-China crossover; Bureau of Foreign Trade / International Trade Administration trade-flow series.
3. Ministry of Labor, Republic of China, foreign-worker stock statistics by nationality and sector 2000–2026; the direct-hiring and intermediate-skilled-worker pathway documentation; the 2024 Taiwan–India labour-mobility memorandum of understanding materials.
4. Ministry of Education, Republic of China, international-student enrolment series with NSP-country disaggregation 2011–2026; the industry-academia collaboration programme documentation; the 2019 curriculum reform's Southeast Asian heritage-language provisions.
5. National Immigration Agency, Republic of China, "new resident" (foreign-spouse) statistics by origin 1990s–2026; the New Economic Immigration Bill drafts and related settlement-policy documentation.
6. Lee Teng-hui era primary materials: the 1993–94 Go South guidelines; contemporaneous coverage of the February 1994 "vacation diplomacy" tour (Philippines, Indonesia, Thailand); the September 1996 jieji yongren announcement texts.
7. Ku, Samuel C.Y., multiple articles on Taiwan's Southeast Asia policy under Lee and Chen (Contemporary Southeast Asia, Asian Survey, 1995–2012) — the principal academic record of the Go South precursors.
8. Glaser, Bonnie S., Scott Kennedy, Derek Mitchell, and Matthew P. Funaiole, The New Southbound Policy: Deepening Taiwan's Regional Integration (CSIS, January 2018) — the fullest external mid-term assessment of the policy's first phase.
9. Marston, Hunter, and Richard C. Bush, "Taiwan's engagement with Southeast Asia is making progress under the New Southbound Policy" (Brookings, July 2018); Bush, Richard C., Difficult Choices: Taiwan's Quest for Security and the Good Life (Brookings Institution Press, 2021).
10. Yang, Alan Hao (楊昊), National Chengchi University / Taiwan-Asia Exchange Foundation, multiple papers on the NSP's people-centred design and regional reception 2017–2025; the TAEF Southbound annual review series.
11. Rigger, Shelley, The Tiger Leading the Dragon: How Taiwan Propelled China's Economic Rise (Rowman & Littlefield, 2021) — the Taishang westward-investment record against which the southbound turn is measured.
12. Hsiao, Russell, ed., Global Taiwan Institute Global Taiwan Brief NSP assessment series 2016–2026; Templeman, Kharis, Hoover Institution Taiwan in the Indo-Pacific papers on economic diversification.
13. Congressional Research Service: Manyin, Mark E., and Susan V. Lawrence, "Taiwan: Political and Security Issues" (IF10275, multiple updates); CRS reporting on the Global Cooperation and Training Framework and US–Taiwan Indo-Pacific coordination.
14. Environmental Justice Foundation and Greenpeace East Asia investigation reports on forced labour in Taiwan's distant-water fishing fleet (2018–2024); US Department of Labor, List of Goods Produced by Child Labor or Forced Labor (2020 edition and updates, Taiwan-caught fish entry); Control Yuan and Fisheries Agency response documentation.
15. Lan, Pei-chia (藍佩嘉), Global Cinderellas: Migrant Domestics and Newly Rich Employers in Taiwan (Duke University Press, 2006) and subsequent work — the foundational scholarship on the caregiver corridor; Taiwan International Workers' Association (TIWA) advocacy documentation.
16. Cheng, Isabelle, and Lara Momesso, multiple works on marriage migration and the "new second generation" in Taiwan (2015–2024); National Development Council new-resident policy papers.
17. Korea, Republic of: Presidential Committee on New Southern Policy documentation (2017–2022) and the New Southern Policy Plus (2020) — the comparative case; analyses of the Yoon-era Indo-Pacific Strategy absorption.
18. Kuik, Cheng-Chwee, multiple works on small-state hedging in Southeast Asia (Contemporary Southeast Asia, International Affairs, 2008–2025); Goh, Evelyn, on ASEAN's omni-enmeshment — the receiving-side analytical frame.
19. ASEAN Secretariat and partner-country investment-agency statistics (Vietnam MPI, Thailand BOI, Indian government semiconductor-mission documentation) on Taiwanese inbound investment 1990–2026.
20. Taipei Times, Liberty Times, United Daily News, China Times — domestic coverage and contestation framings of the southbound policies 1994–2026; Nikkei Asia, Reuters, Bloomberg, Financial Times — the relocation-wave and India-courtship reporting 2018–2026.
21. American Institute in Taiwan (AIT) and US State Department statements pairing the NSP with the Free and Open Indo-Pacific strategy (2018–2025); Global Cooperation and Training Framework (GCTF) programme archive.
22. Tata Electronics–Powerchip (PSMC) Dholera fab announcement materials (2024) and India Semiconductor Mission documentation; Foxconn India expansion disclosures 2022–2026.

Related Documents:

• TW-E-02: New Southbound Policy 2016–2026 — the companion implementation ledger: the four-pillar operational detail, the CPTPP application trajectory, the Lai-era Office of Trade Negotiations re-architecture, and the Trump-2 tariff recalibration that this document summarises.
• TW-F-01: US–Taiwan Relations and Strategic Ambiguity 2016–2026 — the US-alignment channel through which the southbound turn acquired its external patron.
• TW-F-02: Cross-Strait Dynamics 2000–2026 — the dependence problem the southbound policies were designed to answer, and the PRC counter-pressure architecture.
• TW-F-04: Cross-Strait Military Balance and Asymmetric Defence 2017–2026 — the contingency scenarios against which the southbound architecture's dual-use value is assessed.
• TW-G-02: Taiwan's Semiconductor Ecosystem Beyond TSMC — the supplier and OSAT tier whose Southeast Asian relocation carries the semiconductor-era southbound extension.
• TW-G-03: Taiwan Semiconductor Diaspora and Chip Diplomacy 2020–2026 — the global fab architecture within which the India courtship and ecosystem-tier migration sit.
• TW-M-01: Taiwanese Identity and the Status-Quo Doctrine — the post-Sunflower political settlement whose economic-geography expression the New Southbound Policy is.
• TW-N-01: Taiwan in International Perceptions 1971–2026 — the recognition-blockade environment that forced the people-centred workaround.
• TW-D-02: Lai Ching-te Presidency 2024–2025 — the hung-legislature governing context constraining the Lai-era continuation.
• TW-O-01: Taiwan Megatrends — The 2030s Questions — §2 (contingency scenarios) and §5 (demographic cliff) supply the forward frames for §7.2 of this document.
• PH-F-06: West Philippine Sea Confrontation 2023–2026 — the frontline dimension of the Philippines' reception of the southbound turn.
• TW-C-09: back-reference added by symmetry sweep
• TW-D-01: back-reference added by symmetry sweep
• TW-H-PRES-04: back-reference added by symmetry sweep
• TW-K-01: back-reference added by symmetry sweep