Board of Governors of the Federal Reserve System

Figure 1. New Job Openings and New Job Postings

This is a line chart titled “New Job Openings and New Job Postings.” The x axis ranges from 2011 to 2026 in months. The Y axis, labeled “Index, 2022 = 100,” ranges from 0 to 150. The data are monthly. There are two variables charted on the plot. The first, labeled “Job Openings (JOLTS),” is designated by a solid black line; the second, labeled “Job Postings (Lightcast),” is designated by a solid red line. Both series are normalized so that their 2022 average equals 100. Over 2011 to mid-2010s the two series diverge, with job openings running well above job postings—openings hover around 65 to 90 while postings climb from about 30 to 50—reflecting the gradual expansion of online posting coverage. The gap narrows steadily through the late 2010s as postings rise toward the openings series. Both series show a sharp swing around 2020, dropping abruptly at the onset of the pandemic and then spiking, with openings briefly reaching about 140. They peak together in 2021–22 near 100 to 115 and then decline in tandem through 2023–24, converging closely and ending near 80 to 90 by 2026. Overall, the two measures track one another increasingly closely over time, particularly from 2022 onward.

Return to text

Figure 2. Manufacturing Share of New Job Openings and New Job Postings

This is a line chart titled “Manufacturing Share of New Job Openings and New Job Postings.” The x axis ranges from 2011 to 2026 in months. The Y axis, labeled “Percent,” ranges from 5 to about 9, in percent. The data are monthly. There are two variables charted on the plot. The first, labeled “Openings (JOLTS),” is designated by a solid black line; the second, labeled “Postings (Lightcast),” is designated by a solid red line. Each series measures the manufacturing sector’s share of, respectively, total job openings and total new job postings. Both series are volatile and fluctuate broadly between about 5.5 and 8 percent for most of the sample, moving together without a strong trend. The manufacturing share of openings rises notably in 2021, peaking around 9 percent, while the postings share stays closer to 7 percent during that period. In the most recent years both series drift lower and then tick back up, ending near 6.5 to 7 percent by 2026. Overall the manufacturing share of both openings and postings has remained within a fairly stable band over the full period, with the openings series showing a more pronounced spike during the post-pandemic recovery.

Return to text

Figure 3. Share of Postings Requiring AI and Computer Skills

This figure is titled “Share of Postings Requiring AI and Computer Skills” and contains two line-chart panels. In both panels the x axis ranges from 2011 to 2026 in months, the data are monthly, and two variables are charted against separate vertical axes: “AI Skills” is measured on the left axis, labeled “Percent,” and is designated by a thick dark-blue line, while “Computer Skills” is measured on the right axis and is designated by a thinner green line. The left panel, titled “Manufacturing,” has a left axis ranging from 0 to 12 percent and a right axis ranging from 20 to 40 percent. There, the AI-skill share rises gradually from about 2 percent in 2011 to roughly 4 to 6 percent through the late 2010s, with a brief spike near 8 percent around 2018, and then climbs steeply at the end of the sample to about 11 percent by 2026. The computer-skill share is much higher and comparatively flat, fluctuating around 34 to 38 percent for most of the period and easing slightly to about 34 percent at the end. The right panel, titled “All Postings,” uses the same axis scales. There, the AI-skill share is roughly flat near 2 to 3.5 percent for most of the sample before rising sharply at the very end to about 8 percent, while the computer-skill share moves down more recently from around 24–26 percent to a low near 20–22 percent. In both panels the AI-skill share accelerates markedly in the final part of the sample.

Return to text

Figure 4. Types of AI Skills Required in Manufacturing Postings

This figure is titled “Types of AI Skills Required in Manufacturing Postings” and contains two line-chart panels. In both panels the x axis ranges from 2011 to 2026 in months, the Y axis is labeled “Percent,” and the data are monthly. Each panel charts three variables: “Broad AI,” designated by a dark-blue line; “Machine Learning,” designated by a medium-blue line; and “Generative AI,” designated by a light-blue line. Note the difference in axis scale between the two panels. The left panel, titled “Manufacturing,” has a Y axis ranging from 0 to 15 percent. There, Broad AI rises from about 2 percent in 2011 to around 4 to 5 percent through the 2010s (with a spike near 8 percent around 2018) and then accelerates to about 11 percent by 2026. Machine Learning follows a similar but lower path, rising from near 0.5 percent to about 8 percent at the end of the sample, while Generative AI stays near zero until roughly 2023 and then climbs to about 1 percent. The right panel, titled “Manufacturing Production Occupations,” has a much smaller Y axis ranging from 0 to 2.5 percent. There, Broad AI fluctuates around 0.7 to 1 percent for most of the sample before rising sharply at the end to about 2.3 percent; Machine Learning stays near 0.2 to 0.3 percent before ticking up to about 1.5 percent; and Generative AI remains essentially at zero until a small increase at the very end. In both panels AI-skill requirements rise steeply in the most recent years, but the levels are far lower for production occupations than for manufacturing overall.

Return to text

Figure 5. Average Posted Wages by Skill Type

This figure is titled “Average Posted Wages by Skill Type” and contains two line-chart panels. In both panels the x axis ranges from 2012 to 2026 in months, the Y axis is labeled “Dollars,” and the data are monthly. Each panel charts three variables: posted wages for overall manufacturing postings, designated by a thick black line; wages for postings requiring AI skills (“Postings with any AI skill”), designated by a dark blue-gray line; and wages for postings requiring computer skills (“Postings with any computer skill”), designated by a green line. The left panel, titled “All Manufacturing,” has a Y axis ranging from about 50,000 to 150,000 dollars (the overall series labeled “Mfg Postings”). Postings requiring any AI skill consistently command the highest wages, starting near 80,000 to 100,000 dollars, dipping around 2020, and then rising steeply to about 150,000 dollars by 2026. Wages for postings with any computer skill and all manufacturing postings track more closely and lower—around 50,000 to 65,000 dollars for much of the sample—before climbing after 2021 to roughly 105,000 to 120,000 dollars by the end. The right panel, titled “Manufacturing Production Occupations,” has a Y axis ranging from 20,000 to 100,000 dollars (the overall series labeled “Mfg Prod Postings”). There, the AI-skill series is highly volatile—with sharp spikes in the earlier years—but trends upward strongly after about 2022, reaching roughly 70,000 to 100,000 dollars. Postings with any computer skill and all production postings again move together and lower, rising from around 35,000 to 40,000 dollars early in the sample to about 55,000 to 62,000 dollars by 2026. In both panels, postings requiring AI skills carry a persistent and widening wage premium relative to computer-skill postings and to manufacturing postings overall.

Return to text