近年來，前往海外工作已逐漸成為不少台灣年輕人職涯規劃中的選項之一。隨著國際分工日漸緊密，勞動力的跨國流動不再只是少數人的選擇，現在已是愈來愈普遍 的現象。本文以台灣教育長期追蹤資料庫(Taiwan Education Panel Survey， TEPS)及其後續調查(TEPS-B)為基礎，合併兩份資料，衡量青少年時期早期國際經驗對成年後海外就業意願、創業意願、就業傾向與結婚意願的長期影響。本文採用普通最小平方法(Ordinary Least Squares， OLS)進行基準估計，並進一步採用工具變數法 (Instrumental Variables， IV)處理國際經驗取得之內生性問題，以家庭國際化相關背景 因素作為工具變數，進行因果識別。本文對國際經驗的定義不同於現有文獻多聚焦於 正式留學計畫，本研究擴展至出國旅遊、與外國人接觸及國外居住等日常形式，並同 時探討其對多個人生選擇的影響。 在 OLS 估計結果中，不同類型的國際經驗對各項人生選擇的影響方向與強度有所 差異。在海外就業意願方面，各類國際經驗均有顯著的正向影響，本文發現國際接觸 越深，出國工作意願越強。在創業意願方面，並非所有類型的國際經驗均有同等效果， 「與外國人接觸」是其中影響最明顯的，直接的人際跨文化接觸可能才是引發創業動 機的核心因素，而出國旅遊與國外居住的效果則相對薄弱。在就業傾向方面，國際經 驗對進入公部門就業呈現一致的負向關聯，其中以國外居住經驗的負向效果最為明顯， 顯示國際接觸可能擴大個體對私部門與跨國職涯的期待，進而降低公部門的相對吸引 力。在結婚意願方面，OLS 估計顯示出國旅遊經驗有溫和但顯著的正向影響，然而此 一結果在工具變數估計後消失，顯示該正向關聯較可能反映選擇偏誤，而非真實的因 果效果;跨文化互動與國外居住的效果則未達統計顯著。 在工具變數估計結果中，海外就業意願是所有依變數中因果效果最為穩健的一項。 以父母國際化態度加總指標與家中訂閱外文報紙雜誌為工具變數時，國際經驗對出國 工作意願的正向影響在多個衡量方式下均達顯著，且係數明顯大於 OLS 估計，顯示 OLS 低估了國際經驗的真實因果效果。就業傾向方面，部分工具變數組合的估計亦顯 示負向效果，方向與 OLS 一致，但各工具變數組合之間一致性有限，因果推論仍需審 慎。創業意願的 OLS 正向關聯在控制內生性後消失，這與選擇偏誤假說的預測方向一 致:家庭資源較豐富、個人能力較強者，本身就更傾向創業，同時也更有機會累積國際經驗，顯示 OLS 所觀察到的正向關聯主要來自此一偏誤，廣義國際接觸對創業意願 無可辨識的正向因果效果。結婚意願的 OLS 正向效果在控制內生性後同樣縮小並轉為 負向且不顯著，家庭環境穩定、社經地位較高的家庭，既能提供子女更多出國機會， 也更傾向傳遞對婚姻正向的價值觀，形成同方向的偏誤，廣義國際接觸對結婚意願也 無可辨識的正向因果效果。 綜合 OLS 與工具變數估計結果，本研究發現:第一，國際經驗對出國工作意願有 明確的正向因果效果，且效果隨國際接觸深度增加而提升;第二，廣義的國際接觸對 創業意願與結婚意願無可辨識的正向因果效果，現有政策若以促進國際化作為提升創 業率或婚育意願的手段，缺乏充分的因果依據;第三，國際經驗使青年較不傾向選擇 公部門就業，這一負向關聯在因果識別後仍有一定支持，但推論需要保留審慎空間。 此外，值得注意的是，工具變數法識別的是局部平均處理效果(Local Average Treatment Effect， LATE)，本文估計結果僅代表因家庭國際化背景而改變國際經驗累 積行為的特定子群體，不能直接推論至全體樣本。在政策意涵上，出國機會往往集中 在家庭資源較充裕的族群，擴大補助以提升弱勢家庭子女的國際接觸機會，不僅是公 平性的考量，亦可能是提升整體人才國際移動力最直接有效的途徑。;Working abroad has become an increasingly common option in the career planning of young Taiwanese people. As global labor markets become more interconnected， cross-border mobility is no longer a choice made by only a few， but an increasingly widespread phenomenon. This study integrates data from the Taiwan Education Panel Survey (TEPS) and its follow-up survey (TEPS-B) to examine the long-term effects of early international experience during adolescence on adults′ willingness to work abroad， entrepreneurial intentions， employment sector preferences， and marriage intentions. The study employs Ordinary Least Squares (OLS) as a baseline estimation strategy and further applies an Instrumental Variables (IV) approach to address the endogeneity of international experience acquisition， using family internationalization background factors as instruments for causal identification. Unlike existing studies that predominantly focus on formal study abroad programs， this study adopts a broader definition of international experience that encompasses overseas travel， contact with foreigners， and experience living abroad， while simultaneously examining its effects on multiple life outcomes. The OLS estimates reveal that the direction and magnitude of effects vary across different types of international experience and outcomes. Regarding willingness to work abroad， all types of international experience show significant positive effects， indicating that deeper international exposure is associated with stronger motivation to work abroad. Regarding entrepreneurial intentions， not all types of international experience have equally significant effects. Contact with foreigners exhibits the most pronounced impact， suggesting that direct interpersonal cross-cultural interaction， rather than geographic mobility alone， may be the core driver of entrepreneurial motivation， while the effects of overseas travel and overseas residence are comparatively weak. Regarding employment sector preference， international experience is consistently negatively associated with entering the public sector， with overseas residence experience showing the most pronounced negative effect， suggesting that international exposure may expand individuals′ confidence and expectations regarding the private sector and cross-border careers， thereby reducing the relative attractiveness of public sector employment. Regarding marriage intentions， OLS estimates show that overseas travel experience has a modest but significant positive effect; however， this association disappears after IV estimation， suggesting that the positive relationship is more likely attributable to selection bias than to a genuine causal effect. The effects of cross-cultural interaction and overseas residence do not reach statistical significance. The IV estimates reveal that willingness to work abroad is the most robust finding in terms of causal identification. Using a combined instrument comprising the parental internationalization attitude index and household subscriptions to foreign-language publications， the positive effect of international experience on willingness to work abroad is significant across multiple specifications， with coefficients substantially larger than the OLS estimates， indicating that OLS underestimates the true causal effect of international experience. For employment sector preference， some IV specifications also yield negative effects consistent in direction with OLS， though consistency across instrument combinations is limited， warranting caution in causal interpretation. The positive OLS association for entrepreneurial intentions disappears after controlling for endogeneity， consistent with the prediction of a selection bias hypothesis: individuals from resource-rich families with stronger personal ability are both more inclined toward entrepreneurship and more likely to accumulate international experience， such that the positive OLS association primarily reflects this confounding rather than a causal effect of international exposure. Similarly， the positive OLS effect on marriage intentions shrinks and turns negative after controlling for endogeneity， though it remains statistically insignificant. Families with stable environments and higher socioeconomic status are more likely both to provide children with overseas opportunities and to transmit positive attitudes toward marriage， generating bias in the same direction. Broadly defined international exposure thus has no identifiable positive causal effect on either entrepreneurial intentions or marriage intentions. Synthesizing the OLS and IV results， this study draws the following key conclusions. First， international experience has a clear positive causal effect on willingness to work abroad， and this effect strengthens with the depth of international exposure. Second， broadly defined international exposure has no identifiable positive causal effect on entrepreneurial intentions or marriage intentions; policies that promote internationalization as a means to boost entrepreneurship rates or marriage intentions lack sufficient causal justification. Third， international experience is associated with a reduced tendency to choose public sector employment， a negative relationship that retains some support after causal identification， though the inference should still be interpreted with caution. It should further be noted that the IV approach identifies a Local Average Treatment Effect (LATE)， meaning the estimates reflect the experience of a specific subgroup， namely those whose international experience was influenced by their family internationalization background， and cannot be directly generalized to the full sample. In terms of policy implications， opportunities for international exposure tend to be concentrated among individuals from more resource-advantaged families; expanding subsidies to increase access for children from lower-income households is not only a matter of equity， but may also represent the most direct and effective means of improving the overall international mobility of Taiwan′s talent pool.
