本周由斯蒂芬妮·弗兰德斯主持的“特朗普经济学”播客，深入探讨了美国对超级富豪征税这一紧迫问题，探讨了当前系统为何举步维艰，以及拟议的解决方案是否可行。讨论邀请了彭博社记者凯特琳·莱利（她主要负责国会和税收政策报道），以及哈佛大学经济政策教授、奥巴马总统经济顾问委员会前主席贾森·弗曼。 弗兰德斯开篇指出，国际货币基金组织（IMF）一次兵棋推演中发现的一个令人惊讶的“末日场景”：并非杀人机器人，而是人工智能（AI）对“所得税基础的文明终结式攻击”。如果AI摧毁了高薪工作，政府的主要收入来源（所得税）可能会崩溃。尽管公司和大型投资者会从中获利，但他们的大部分收益将来自征税很少或未征税的资本利得，这将加剧超级富豪现有的避税行为。这一背景促使欧洲和美国进步派政治家呼吁征收财富税，并以加州提议对亿万富翁征收一次性5%的税收为例。 凯特琳·莱利证实，对富人征税“越来越成为民主党人关注的焦点”，尤其是在他们展望2028年大选之际。鉴于目前不在执政地位，这些想法仍停留在理论层面，但却是进步派和温和派民主党人之间辩论的核心，特别是在如何不进一步增加国债的情况下为雄心勃勃的社会项目提供资金。她指出，美国税收体系严重依赖工资收入，却未能充分捕捉富人（通常通过持有资产）积累和扩大财富的方式。随着时间的推移，传统上针对财富的税收，例如遗产税和企业税率，已被削弱。埃马纽埃尔·赛斯和加布里埃尔·祖克曼的研究表明，最富有的400个家庭现在支付的有效税率低于最底层50%的家庭。 贾森·弗曼虽然个人认为税法不够累进，但他澄清说，直到最顶层的千分之一人群，税收系统 *确实* 是累进的。然而，他承认累进性有所下降，这主要是由于企业税率从35%降至21%，这使得顶层的资本收入者不成比例地受益。他区分了在痛苦最小化的情况下增加收入的目标，以及利用税收来改变行为的目标。 弗曼对诸如伯尼·桑德斯提出的对亿万富翁征收每年5%的财富税，或加州的一次性税收等提议表示怀疑。他认为，在州甚至国家层面（如法国），富人可以轻易搬迁以避免此类税收，从而削弱税收收入，只是在地理上转移了亿万富翁，而非减少了他们。美国对全国性财富税而言，一个更重大的障碍是宪法。弗曼解释说，美国宪法禁止未经人口比例分配的直接税，而第16修正案只为所得税提供了例外。他断言，大多数宪法学者认为，当前的最高法院将裁定全国性财富税违宪。凯特琳·莱利表示同意，称之为“一个非常大的障碍”。 讨论随后转向现有宪法框架内更可行的替代方案，重点是解决允许富人避免资本利得税的结构性问题。凯特琳强调了“遗产税基础递增”（step-up-in-basis）规则，即资产的成本基础在所有者去世时被重置，实际上免除了所有者生前累积的资本利得。富人还经常以其资产为抵押进行借贷，在不触发应税事件的情况下获取流动性。解决这些问题的提议包括调整遗产税基础递增规则、对贷款征税，或对“未实现收益”征税（如怀登参议员所提议）。 贾森·弗曼支持解决这些结构性问题。他指出，简单地提高现有税率（所得税、资本利得税、企业税）是一种直接且符合宪法的途径。对于未实现收益，他承认对市场波动征税和评估私人资产存在后勤上的复杂性。尽管怀登参议员的提案试图解决这些问题，但事实证明它“极不受欢迎”。弗曼建议，对死亡时获得的收益征税，尽管不如在收益累积时每年征税那样全面，却能带来“85%的收益”，而只产生“20%的政治痛苦”，使其成为一个更审慎的选择。 两位嘉宾都讨论了公众与财富税之间复杂的关系。尽管美国人普遍支持对富人多征税，但遗产税或财富税等具体政策却面临极大的不受欢迎，即使在那些不太可能受影响的人群中也是如此。弗曼分享了一些轶事，说明了公众对遗产税情感上、常常是非理性的抵制。 最后，弗兰德斯询问了人工智能使这些讨论变得更加紧迫的可能性。弗曼同意，人工智能很可能使对财富征税成为一个“更具现实意义的问题”，预测它将随着时间推移降低劳动收入份额，并提高资本收入份额。这一转变凸显了弥补资本税收“漏洞”的必要性。然而，他提醒不要忘记“老式的收入方面”，在可预见的未来，这仍将构成大部分收入。

This week's "Trumponomics" podcast, hosted by Stephanie Flanders, delves into the pressing issue of taxing the mega-rich in the United States, exploring why the current system is struggling and whether proposed solutions could work. The discussion features Caitlin Riley, a Bloomberg reporter covering Congress and tax policy, and Jason Furman, Professor of Economic Policy at Harvard and former chair of President Obama's Council of Economic Advisors. Flanders opens by highlighting a surprising "doomsday scenario" identified by an IMF wargaming exercise: not killer robots, but a "civilisation-ending attack on the income tax base" due to AI. If AI destroys well-paying jobs, the government's primary revenue source (income tax) could collapse. While companies and big investors would profit, much of their gains would be in lightly or untaxed capital gains, exacerbating existing tax avoidance by the super-rich. This context fuels the growing calls from progressive politicians for wealth taxes in both Europe and the US, citing California's proposed one-time 5% tax on billionaires as an example. Caitlin Riley confirms that taxing the wealthy is "increasingly front of mind for Democrats," especially as they eye the 2028 elections. Currently out of power, these ideas remain theoretical but are central to the debate between progressive and moderate Democrats, particularly regarding how to fund ambitious social programs without further ballooning the national debt. She notes that the US tax system heavily relies on wages, failing to adequately capture how the wealthy accumulate and expand their wealth, often through asset holdings. Over time, taxes that traditionally targeted wealth, such as the estate tax and corporate tax rate, have eroded. Research by Emmanuel Saez and Gabriel Zucman indicates that the top 400 wealthiest households now pay a lower effective tax rate than the bottom 50% of households. Jason Furman, while personally believing the tax code isn't progressive enough, clarifies that up to the top one-tenth of one percent, it *is* progressive. However, he acknowledges it has become less so, largely due to the reduction in the corporate tax rate from 35% to 21%, which disproportionately benefits capital income earners at the top. He distinguishes between the goal of raising revenue with minimal pain and using taxes to change behavior. Furman expresses skepticism about proposals like Bernie Sanders's annual 5% wealth tax on billionaires or California's one-off tax. He argues that at state or even country levels (like France), the wealthy can easily move to avoid such taxes, undermining revenue collection and simply shifting billionaires geographically rather than reducing them. A more significant hurdle for national wealth taxes in the US is the Constitution. Furman explains that the US Constitution prohibits direct taxes unless apportioned by population, with the 16th Amendment providing an exception only for income taxes. Most constitutional scholars, he asserts, believe the current Supreme Court would rule national wealth taxes unconstitutional. Caitlin Riley agrees, calling this a "very hard stop." The conversation then shifts to more viable alternatives within the existing constitutional framework, focusing on addressing structural issues that allow the wealthy to avoid capital gains taxes. Caitlin highlights the "step-up-in-basis" rule, where the cost basis of assets is reset at death, effectively forgiving capital gains accumulated during the owner's lifetime. The wealthy also often borrow against their assets, accessing liquidity without triggering taxable events. Proposals to address this include adjusting step-up-in-basis, taxing loans, or taxing "unrealized gains" (as proposed by Senator Wyden). Jason Furman supports tackling these structural issues. He notes that simply raising existing tax rates (income, capital gains, corporate) is a straightforward, constitutionally sound approach. For unrealized gains, he acknowledges the logistical complexities of taxing market fluctuations and valuing private assets. While Senator Wyden's proposal attempted to address these, it proved "exceedingly unpopular." Furman suggests that taxing gains at death, while not as comprehensive as taxing them annually as they accrue, offers "85% of the benefits" with only "20% of the political pain," making it a more prudent option. Both guests discuss the public's complex relationship with wealth taxes. While Americans generally support taxing the wealthy more, specific policies like inheritance or wealth taxes face significant unpopularity, even among those unlikely to be affected. Furman shares anecdotes illustrating the emotional, often irrational, resistance to estate taxes. Finally, Flanders asks about AI's potential to make these discussions more urgent. Furman agrees that AI will likely make taxing wealth a "much more live issue," predicting it will lower the labor share and raise the capital share of income over time. This shift underscores the need to address the "holes" in capital taxation. However, he cautions against forgetting the "old-fashioned income side," which will still constitute the majority of income for the foreseeable future.