Econ Papers Daily / 期刊论文
定向增值税激励的政策一致性：来自中国2023年增值税加计抵减方案的经验证据
Policy consistency in targeted VAT incentives: evidence from China’s 2023 additional input VAT deduction schemes
摘要
China’s additional input VAT deduction (AD) policy has shifted from temporary service-sector relief to targeted support for integrated circuits (IC), advanced manufacturing (AM) and industrial mother-machine (MM). Using a 2023 China Social Accounting Matrix (SAM) and a VAT-specific 47-sector CGE model, this paper evaluates the 2023 policy package and related counterfactuals. Under the Keynesian short-run closure, the full package reduces GDP, government revenue, VAT revenue, employment, and consumption relative to the no-AD benchmark and yields negative EV, CV, and EB values reported as absolute monetary changes. Under the neoclassical closure, however, the signs of the GDP and welfare effects reverse, indicating that the estimated macroeconomic effects are closure-dependent. Within the Keynesian closure, the qualitative direction of the results remains robust to Global Trade Analysis Project (GTAP) elasticity replacement, proportional elasticity changes and alternative statutory VAT-rate assumptions, although the magnitudes vary. These findings show that institutional policy objectives do not automatically produce aggregate gains and that policy assessment should jointly consider sectoral eligibility, deduction rates, VAT-rate design, fiscal consequences, and macroeconomic adjustment assumptions. KEYWORDS: Value-added tax additional deduction tax incentives policy effects CGE model JEL CLASSIFICATION: H25 H32 C68 E62
中文摘要
中国增值税加计抵减政策已从临时性的服务业纾困转向对集成电路、先进制造业和工业母机的定向支持。本文运用2023年中国社会核算矩阵和针对增值税的47部门可计算一般均衡模型，评估2023年政策组合及相关反事实情形。在凯恩斯短期闭合下，相较于无加计抵减基准，完整政策组合降低了GDP、政府收入、增值税收入、就业和消费，并产生以绝对货币变化表示的负EV、CV和EB值。然而，在新古典闭合下，GDP和福利效应的符号发生反转，表明所估计的宏观经济效应依赖于闭合设定。在凯恩斯闭合内，尽管幅度有所变化，结果的方向性在替换全球贸易分析项目弹性、按比例调整弹性以及替代法定增值税率假设下仍保持稳健。这些发现表明，制度性政策目标并不会自动带来总量收益，政策评估应综合考虑部门适用资格、抵减率、增值税率设计、财政后果以及宏观经济调整假设。关键词：增值税 加计抵减 税收激励 政策效应 可计算一般均衡模型 JEL分类：H25 H32 C68 E62