What Is Going on with NYC Jobs?

Introduction

In some respects, 2026 has gotten off to a reasonably good start for the U.S. economy: the stock market, by most measures, remains near an all-time high; business sentiment has improved somewhat; inflation remains subdued; and, as of February 20, tracking GDP estimates by the Atlanta and New York Fed range between 2.4% and 3.1% for the first quarter.

In contrast, job market indicators have generally been signaling weakness, both locally and nationwide. For the U.S. as a whole, job creation picked up somewhat in January, though it was still sluggish; in New York City, January data are not yet available, but there was no net job creation outside the Health & Social Assistance sector during 2025. In recent weeks, jobless claims have remained subdued nationwide; in New York City, the recently-ended nurses strike did drive claims up, but outside the health sector, they have remained low. In contrast, consumer confidence—heavily influenced by perceptions of the job market—fell to a 9½-year low nationally in January and was just shy of a 5-year low in New York State.

In this Spotlight, we dig beneath the surface of the employment data to get a better sense of what is going on, what are the underlying causes, and what may lie ahead. Given strong evidence that young workers are having a particularly hard time finding meaningful employment, along with signs that this group is most vulnerable to the job displacement effects of Artificial Intelligence (AI), this Spotlight also focuses on the recent and potential role of AI usage on the contours of the job market.

Recent Employment Trends

Job Creation Anemic & Narrowly Based

Both in NYC and the rest of the U.S., job creation over the past year has been exceptionally weak. Based on recently revised data, U.S. private-sector employment increased by just 0.3% over the course of 2025, averaging only 31,000 jobs added per month. New York City fared only somewhat better with a gain of 0.8% (2,700 jobs/month), based on current data. However, we estimate that, after upcoming revisions (to be released in early March), this gain will be revised down. The box below provides a brief explanation of the annual benchmark revisions.

But even the revised data likely overstate the underlying strength in the job market. The only sector with any significant job gains over the past year—again, both locally and nationally—has been Healthcare & Social Assistance. Because these tend to be relatively low-paying jobs that most often serve the local community and are largely driven by government funding, they are not drivers of economic growth, nor are they reflective of underlying economic conditions. In fact, outside of this sector, employment actually declined over the course of 2025, both locally and nationally. And for New York City, this trend is expected to be weaker still once revised data are released in early March, as illustrated in Chart 1 below.

Chart 1

Sources: U.S. Bureau of Labor Statistics; NYC Office of Management & Budget; Moody’s economy.com; Office of the NYC Comptroller

Despite its diverse industry mix, New York City’s economy is driven primarily by certain key industries, namely Finance, Information, and Professional and Business Services. These sectors are the main contributors to office space occupancy and pay higher wages in comparison to others, thus generating multiplier effects with positive spillovers to the rest of the economy. While preliminary data show an increase in employment in Professional and Business Services and Information between Q4 2023 and Q4 2025, we project that this job creation will be revised away with upcoming benchmark revisions as shown in Chart 2. We project employment growth in Financial Activities (which comprises the Finance, Insurance, and Real Estate subsectors) will be revised upward over the two-year period to a gain of about 9,000 jobs, which is still fairly modest. In recent months, Professional & Business Services have seen some pickup in job creation, but Information and Financial Services have seen modest job losses.

Chart 2

Sources: U.S. Bureau of Labor Statistics; NYC Office of Management & Budget; Moody’s economy.com; Office of the NYC Comptroller

While New York City and U.S. employment numbers look weak, new claims for unemployment insurance have held steady, signaling a low hire, low fire economy. Apart from a recent spike in initial claims in Health Services & Social Assistance—likely a consequence of the Nurse’s strike—claims have only been modestly above last year’s levels.

Table 1

Source: NY State Department of Labor (NYC)

The Worker Adjustment & Retraining Notification (WARN) gives advance notice about large scale layoffs or site closures. For most of 2025, WARN data for New York City was subdued, especially in comparison to 2024, except for a large but temporary spike in October.

Job Openings and Labor Turnover Survey (JOLTS) data for December 2025 show a national hiring rate of 3.3%, below pre-pandemic levels, whereas New York State’s hiring rate was stronger, rising to 3.7% in December—up from a year earlier and above pre-pandemic levels—as shown in Chart 3. Layoffs and discharges in the state also fell to a five-month low of 1.3% in December. The rate of job openings across New York State also fell noticeably in December, though it is around the middle of the range that had prevailed before the pandemic.  JOLTS data are not available at the sub-state level—i.e. for NYC.

Chart 3

Sources: U.S. Bureau of Labor Statistics; Moody’s economy.com

Despite sluggish job creation, the proportion of working-age adults with jobs ended 2025 at a record high in New York City, despite a nationwide decline. This reflects the fact that employment has rebounded beyond pre-pandemic levels, but the adult population has not. Though New York City’s ratio is still a bit lower than the nation’s, the difference is historically narrow at less than 1 percentage point, as shown in Chart 4. Throughout the 1980s and ‘90s, the gap was well over 8 percentage points. This, along with the incipient pickup in hiring, suggests some resilience in labor demand.

Chart 4

Source: NY State Department of Labor (NYC); U.S. Bureau of Labor Statistics (USA)

A Weak Labor Market for New Grads

The nationwide unemployment rate has steadily trended upward in the past several years, after reaching a post-pandemic low of just 3.4% in April 2023—about the same level as in February 2020. But while unemployment has grown across demographic groups, young college graduates in particular have seen their standing in the labor market decline to a greater degree than the overall workforce.

Chart 5 plots the nationwide unemployment rate disaggregated by age and education level, revealing an important trend among young college graduates. Young workers in general have always had a higher unemployment rate than the overall workforce, while college-educated workers have had a lower unemployment rate. Historically, the positive employment effect of a college education surpassed the negative impact of youth, meaning that young college graduates had a lower unemployment rate than older workers without a B.A. But the relative size of this gap steadily shrunk through the 2010s, and since late 2022 the trend has altogether reversed, with unemployment among college graduates age 22-27 now exceeding that of older workers without a degree.

Chart 5

Source: U.S. Census Bureau Current Population Survey (IPUMS), based on analysis from the Federal Reserve Bank of New York

The causes of this reversal are debated. Some commentators point to the rise in artificial intelligence as a critical factor, noting that the technology is particularly capable of automating the work of entry-level employees in traditionally high-paying, white-collar jobs that require a college degree. Others argue that economic uncertainty caused by erratic U.S. government policy, in particular since the beginning of the second Trump administration, has made employers especially reluctant to take chances on hiring new college graduates. This idea is consistent with the low-hire, low-fire economy described above. However, neither narrative accounts for the fact that the magnitude of a college education’s positive impact on employment relative to the negative employment impact of youth has been shrinking since the early 2010s, indicating that the trend defies any simplistic or individual explanation.

Chart 6 explores how unemployment among these age/education groups has unfolded in New York City. As in the country overall, unemployment among recent college graduates now exceeds that of older workers without a college degree. However, the reversal emerged later in New York City than nationwide (around late 2024 compared to late 2022, respectively). Part of this delay could result from college education carrying an even higher employment premium in New York City than elsewhere in the country, given the predominance of high-skill jobs in Finance, Information, and Professional and Business Services—thus holding unemployment down for young college graduates. The fact that the unemployment reversal does not begin until November 2024 could indicate that, at least in New York City, the reversal has more to do with recent economic uncertainty caused by federal policymaking and AI technology, as hypothesized above. However, it is worth keeping in mind that the Current Population Survey’s (CPS) smaller sample size in New York City relative to the country as a whole makes these trends more volatile overall.

Chart 6

Source: U.S. Census Bureau Current Population Survey (IPUMS)

Chart 6 reveals another notable trend in New York City unemployment: its steep decline among young, non-college-educated New Yorkers over the course of 2025, a pattern which is not apparent nationally. While the trend appears stark, when we compare the two equivalent groups in Charts 5 and 6 (“Age 22-27, non-college”) we see that young, non-college-educated New Yorkers’ diminishing unemployment in 2025 appears more like a reversion to the unemployment level of their nationwide counterparts.

In other words, young, non-college graduate unemployment reached a much higher pandemic-era peak in New York City than in the U.S. (26.9% versus 13.8%, respectively, in early 2021) and remained elevated by an average of 7.1 percentage points over the national average in 2022 through 2024. This sustained, elevated unemployment rate was in large part a result of sluggish recovery in both the Retail and Leisure and Hospitality sectors in New York City, both of which disproportionately employ non-college graduates.

By December 2025, the unemployment rate among young, non-college graduates in New York had fallen much closer to that of their nationwide counterparts—to 7.6% versus 7.2%, respectively. Part of the 2025 decline in New York City may have come from a reduction in the supply of available workers in this demographic, a result of both restrictive federal immigration policies and a shrinking asylum-seeker population—both of which would disproportionately impact New York City. While the extent to which asylum seekers are adequately represented in the Census’ Current Population Survey is unclear, their exit from New York City’s labor market would regardless constrict labor supply relative to demand, thus driving down unemployment in the short-term.

Non-college graduates’ declining unemployment in New York City is also highly likely to be driven by continued job growth in the Health Care and Social Assistance sector—which more frequently employs those without a college degree—while other sectors remain stagnant.

AI and Jobs

Our August 2025 spotlight explored the usage of generative AI at work, and how it might affect or already be affecting the job market. The spotlight relied on national data from Anthropic’s generative AI platform, Claude, to examine the share of AI usage in each occupation group. We found that hiring has slowed in occupations with heavier usage, and usage is positively correlated with income. In addition, Anthropic’s data make a distinction between automation (where AI totally replaces human effort for a given task) and augmentation (where AI assists and collaborates with a worker to accomplish a task). A majority of Claude conversations were classified as augmentation.

Relative to the total number of employees in each occupation, usage was most concentrated in Computer and Mathematical; Arts, Design, Entertainment, Sports, and Media; and Life, Physical, and Social Science occupations. Computer and Mathematical occupations represented by far the highest share of Claude conversations, even before accounting for the relatively small share of U.S. employees. On the other hand, occupations that typically involve more physical action and less academic education, such as Healthcare Support; Food Preparation and Serving; and Transportation and Material Moving occupations saw a very low share of Claude conversations, especially given the number of employees in those occupations.

We also used New York City employment data to estimate Claude’s usage within the city, with the assumption that New York City and the U.S. exhibit similar usage patterns across occupations. We found that AI usage was concentrated in higher-paying occupations, while the city’s largest occupations, such as healthcare and retail, saw little usage, but also tended to pay less. Therefore, mirroring U.S. trends, the effects of AI will likely hit NYC’s highest paying occupations first, while the city’s currently growing occupations are likely to remain more insulated, for the time being.

In January 2026, Anthropic released new data that cover a more recent period (up to November 2025) and include several new statistics. Most notably, the usage data for the U.S. are now disaggregated into individual states.  This enables us to directly assess New York’s usage instead of inferring state-level trends from national data.

Relative to its population, New York uses Claude more than any other U.S. state, trailing only Washington, D.C. Consistent with the August spotlight, usage is heavily concentrated in Computer and Mathematical occupations, but the difference is much starker in New York than in other locations: 43% of New Yorkers’ work-related Claude conversations are in Computer and Mathematical occupations, followed by 8.5% in Education, and 8% in Arts, Design, Entertainment, Sports, and Media. In the U.S. as a whole, only 27% of Claude conversations are for Computer and Mathematical occupations, 14% in Education, and 10% in Arts and Media. Thus, the analysis from the August spotlight understated the extent to which AI usage is concentrated in New York’s Computer and Mathematical occupations.

Normalizing for the number of employees in each occupation highlights significant differences in AI usage across geography, as shown in Table 2 below. In each column, the relative representation factor is equal to the percent of Claude conversations belonging to an occupation divided by the percent of the total employed population who work in that occupation (e.g., Computer and Mathematical occupations in New York represent 43.14% of Claude conversations, but Computer and Mathematical employees make up only 3.14% of the state’s workforce, so the relative representation factor is 13.72). Occupations with relative representation factors above 1 see disproportionate Claude usage for their size, while occupations below 1 are underrepresented. The general trends are similar between the U.S. and New York, with the exception of Community and Social Service Occupations, which is overrepresented in the U.S. but underrepresented in New York. As the raw data suggests, New York’s Claude usage is very heavily concentrated in Computer and Mathematical occupations, although the U.S., while still concentrated, has stronger AI representation in Life, Physical, and Social Science occupations; Arts, Design, Entertainment, Sports, and Media occupations; and Educational Instruction and Library occupations.

Table 2

Source: Anthropic Economic Index, OEWS. Note: in our August 2025 spotlight, we referred to the relative representation factor as the “overrepresentation factor,” but we use the two terms interchangeably.

Compared to the older data from the August spotlight, the U.S.’s Claude usage has become slightly less concentrated in Computer and Mathematical; Arts, Design, Entertainment, Sports, and Media; and Life, Physical, and Social Science occupations. This does not mean that overall activity has declined (in fact, it has increased across the board), but instead indicates that other occupations are “catching up” in their AI usage, so the idea that some job categories are “safe” from AI adoption is unlikely to hold over time.

Outlook

Periods of economic growth are typically characterized by diminished layoffs and robust hiring. Yet, despite GDP growth of roughly 2.5%, recent job creation has been minimal, particularly in areas outside of Healthcare & Social Assistance. One notable feature of the current environment is the difficulty recent college graduates, even those with more advanced degrees, have had finding the types of career-starting jobs for which they trained. This pattern is not uniform across education groups: in New York City, unemployment among non-college workers has declined, reflecting the uneven growth in the labor market and demographic changes. A number of potential forces appear to be at play.

In recent years, both the national and local economies have experienced a series of largely unprecedented developments: the pandemic-driven surge in remote work; pro- and anti-immigration policies; the evolution and spread of AI; and increasingly restrictive trade policies. Given these cross-currents, it is difficult to definitively attribute today’s labor market challenges to any single cause. Still, at least some of the reluctance to hire new talent likely stems from uncertainty on the part of employers about the future evolution of AI. And, it is clear that AI usage has broadened to more and more occupations in just the past six months. While the long-term effect of AI proliferation may include productivity gains and higher-paying roles, the near-term adjustment appears uneven, posing particular challenges for young college graduates and those with more advanced degrees seeking to enter professional occupations.

---

Acknowledgements

This report was prepared by Jason Bram, Director of Economic Research; Amber Born, Economic Development Analyst; Andre Vasilyev, Assistant Director for Economic Development; Aliyah Sahqani, Economic Research Analyst; Jonathan Siegel, Chief Economist; Krista Olson, Deputy Comptroller for Budget; and Francesco Brindisi, Executive Deputy Comptroller for Budget and Finance. Archer Hutchinson, Creative Director and Danbin Weng, Multimedia Designer led the design, with assistance from Angela Chen, Senior Website Developer, Martina Carrington, Web Developer, and Cindy Zhao, Web Developer.