全球税收博弈下中国短期跨境资本流动的结构跃迁与政策响应机制------基于生成式人工智能语义指数的嵌入分析

The Structural Transition and Policy Response Mechanism of China􀆳s Short-term Cross-border Capital Flows under the Impact of Global Taxation: Embedding Analysis Based on Generative Artificial Intelligence Semantic Index

摘要:

2025年4月,美国启动新一轮全球税收重构计划,推动对多国贸易伙伴实施“对等征税” 机制,并迅速引发中国采取反制措施。全球税收规则的制度性重塑由此成为新型外部冲击源,在短期内强化国际资本的套利动机与趋利行为,诱发跨境资本流动的周期错配与结构跃迁。相较于传统基于利差与汇率波动的分析框架,此类制度性冲击更易通过政策预期变化与情绪放大效应触发非线性响应,传统方法难以刻画其动态演化过程。聚焦全球税收博弈背景下中国短期跨境资本流动的非对称调整机制,基于 2008—2024年季度数据构建多门槛回归模型,并引入生成式人工智能对政策文本进行语义建模,构造税收政策语义强度指数( TPSI) 。研究发现,全球税收冲击显著改变了资本对利差与汇率预期的响应结构,呈现区间跃迁与非对称特征;TPSI在识别资本情绪方向与流动趋势方面具有显著领先性;在高语义强度区间内,宏观审慎工具调节效应边际减弱,短期资本流动面临“ 高频脱锚” 风险。基于此,建议建立以语义预期识别为基础的跨境资本预警机制,前瞻性提升资本流动治理能力,并积极参与全球税收规则重构进程,增强中国在国际金融体系中的制度韧性与税收主权保障能力。通过将生成式人工智能语义建模引入跨境资本非线性分析框架,拓展了制度性冲击下资本流动结构跃迁的识别方法,并为提升宏观审慎政策的前瞻性与精准性提供了新的实证依据。

Abstract:

In April 2025, the United States launched a new round of global tax restructuring, promoting a “reciprocal taxation” mechanism toward multiple trading partners, which quickly triggered countermeasures from China. The institutional reshaping of global tax rules has thus emerged as a novel external shock, intensifying international capital’s arbitrage incentives and profit-seeking behavior in the short term and inducing cyclical mismatches and structural shifts in cross-border capital flows. Compared with traditional analytical frameworks based on interest rate differentials and exchange rate fluctuations, such institutional shocks are more likely to generate nonlinear responses through changes in policy expectations and amplified market sentiment, making their dynamic evolution difficult to capture using conventional methods. Against this background, the symmetric adjustment mechanisms of China’s short-term cross-border capital flows under global tax competition was examined. Using quarterly data from 2008 to 2024, a multiple-threshold regression model and introduce generative artificial intelligence was constructed to perform semantic modeling of policy texts, thereby developing a tax policy semantic intensity index (TPSI). The results show that global tax shocks significantly alter the response structure of capital flows to interest rate and exchange rate expectations, exhibiting clear regime shifts and asymmetric effects. The TPSI demonstrates strong leading properties in identifying the direction of capital sentiment and flow dynamics, effectively compensating for the lag of traditional variables in capturing policy expectation changes. Moreover, under high semantic intensity regimes, the marginal effectiveness of existing macroprudential tools weakens, increasing the risk of “high-frequency decoupling” in short-term capital flows. Accordingly, it proposes establishing a cross-border capital early-warning system based on semantic expectation identification, proactively enhancing capital flow governance, and actively participating in the reconstruction of global tax rules to strengthen China’s institutional resilience and safeguard its tax sovereignty within the international financial system.