← 返回 Avalaches

在1990年代初,约有500万(5 million)名年轻人进入劳动市场,遭遇严峻的求职景象。作者回忆纸本申请后仍收到拒绝信,即使在一场由1980年代末存款与贷款危机(savings and loan crisis)及1990–91海湾战争导致油价飙升部分引发的美国衰退后,联邦圣路易斯储备银行(Federal Reserve Bank of St. Louis)的数据把青年失业率定为13.4%,而这一阶段被称为「jobless recovery」,即经济增长未能为新进者创造工作。

今天,求职主要转为数位化,但新进者面临的压力在结构上看似相似。22–27岁大学毕业生失业率从2022年底的4.1%上升到去年底的5.6%。加州大学伯克利分校公共政策与经济学教授 Jesse Rothstein 指出,当劳动市场转软时,先受影响的是新进劳工。若入门职缺不足,青年将失去爬升职涯的踏脚石,也会限制长期经济需求;UCLA 经济学教授 Till von Wachter 对1980年代那一波世代的研究显示伤害效应(scarring):至少前10年收入较低,且离婚率更高、生育数较少、健康较差,并有更高肺、肝、心脏疾病死亡率。

AI 被视为主要不确定因素。斯坦福经济政策研究所(SIEPR)报告显示,从2022年底到2025年7月,像 software developers 等「AI-exposed」职位出现13%的相对下降,其中22–25岁下降6%,而年长劳工就业仍持续上升。经济学界仍分歧:有人预期AI会自动化底层入门任务;亦有如 MIT 的 David Autor 认为 AI 是放大劳工能力与判断的工具,会重新配置而非完全替代需求。欧洲多国和中国已采用学徒与补贴模式——例如丹麦、德国、瑞士与雇主共担、 美国透过公共资助医疗教育与住院医师计划、中国补助职业学校及制造业/高科技园区入门职位;美国若推行类似联邦方案,可能有助重建职业阶梯的第一级。

246120d0c173.png



In the early 1990s, around 5 million young people entering the labor market met a harsh welcome. The author recalls rejection letters from paper applications, even after a major U.S. recession triggered partly by the late-1980s savings-and-loan crisis and the 1990–91 Gulf War oil-price spike. Federal Reserve Bank of St. Louis data put youth unemployment at 13.4%, and the period became known as a “jobless recovery,” where growth failed to generate jobs for new entrants.

Today, entry into work is mostly digital, but the pressure on first-time applicants appears structurally similar. The unemployment rate for 22–27-year-old college graduates rose from 4.1% at the end of 2022 to 5.6% at the end of last year. UC Berkeley professor Jesse Rothstein says a softening labor market hurts new entrants first. If entry-level jobs are scarce, young workers miss stepping-stone positions, which also constrains long-run economic demand. UCLA economist Till von Wachter’s research on the 1980s cohort shows scarring: lower earnings for at least 10 years, plus higher divorce, fewer children, poorer health and higher mortality from lung, liver and heart disease.

The article treats AI as the main uncertainty. SIEPR reported a 13% relative decline in “AI-exposed” jobs such as software developers from late 2022 to July 2025, with a 6% decline for ages 22–25 while older workers’ employment kept rising. Economists are split: some expect AI to automate away entry tasks; others like MIT’s David Autor argue it augments workers and shifts, rather than destroys, demand. Governments already use apprenticeship and subsidy models—Denmark, Germany and Switzerland with employer co-funding; U.S. public support for medical training; China’s vocational and entry-level subsidies—and a federal U.S. program in this vein is proposed to rebuild the first rung of the ladder.
2026-04-15 (Wednesday) · 3974c32d4f3ab38966b5754b5bfc80674cf2f4aa