A fully remote offer can look like the best outcome on a spreadsheet. For a new graduate, the New York Fed’s Liberty Street note from 1 June 2026 says the remote shift is also the main reason younger graduates have had a harder time finding work. The planning question is whether the role you are about to accept still has someone in the room who will tell you what good looks like.
The note looks at graduates under 29 and at more experienced graduates. Unemployment for the younger group rose from 3.1 percent to 3.7 percent. For experienced graduates it moved from 1.9 percent to 1.8 percent. Remote work, the authors write, explains about 64 percent of the rise in young-graduate unemployment. In-person work brought more feedback and more mentorship. Those are the figures on the page. They are not a rule that every remote job is a bad job, and they are not a promise that an office will hire you.
Read the age cut before you borrow the number. “Under 29” is not the same group as “22 to 27” on the New York Fed’s college labor market page, and it is not the unemployment rate for all bachelor’s holders over 25. This post uses the Liberty Street comparison only. If a chart does not say who is in the denominator, it is not ready to plan a year around.
What the 64 percent is doing
The mechanism in the note is straightforward. When experienced people work from home, the informal teaching that used to happen beside a desk happens less often. Employers who can already staff the work with people who need less oversight have fewer reasons to hire someone who still needs that oversight. The younger unemployment rate edges up. The experienced rate does not. Most of that gap, about 64 percent in the authors’ accounting, lines up with the spread of remote work.
That is an explanation of a national change. It does not tell you to refuse a remote offer that pays the rent and teaches a skill you can show. It does tell you to ask, before you accept, who reviews your work and how often you will hear from them. A remote role with a weekly review and a named mentor is a different object from a remote role where you are handed a login and a ticket queue. The note’s point about feedback is something you can ask in an interview without pretending to have read the regression.
A public post from the New York Fed pointed at the same finding. Separate from that, Ernie Tedeschi wrote on 17 October 2026 that things other than AI are going on in the entry-level market, while allowing that AI may still have an effect in some fields. His post is an economist’s reading. It does not add a new table. Use it as a brake on a one-cause story. The Liberty Street note is the table this post relies on. A thread that blames a single technology for every quiet inbox is ahead of what this source measured.
How to read an offer against that note
When you compare two offers, write three lines that have nothing to do with the logo.
Where the work happens on a normal week, in the words of the offer or the posting. Who you would sit near, or meet, in the first month. What artifact you will have finished by day 60 that a stranger could look at. If the remote offer loses on the second and third lines, the Fed’s note is a reason to weigh the in-person offer more carefully than the group chat will. If the in-person offer is a commute to a room where no one reviews junior work, the note does not rescue it. The research finding is about feedback and mentorship. A badge swipe is a weaker stand-in for that.
Fields with a built-in place to stand, such as a classroom, a clinic, or a site, have always had this structure. You do not need this paper to know that student teaching puts you next to a teacher. You do need it when a software, operations, or analyst role is sold as “flexible” and the flexibility means you will not meet the people who know how the work is judged.
Write the location into your own plan for the year. “Remote” and “two days on site” are facts. “They seemed chill” is not. When a posting does not say, ask once, and record the answer next to the role. A plan that says “any job” will accept the first thing that arrives, including the one that leaves you without the feedback the Fed associates with in-person work.
A semester plan that names the room
In the term before you recruit, pick two target roles and write where each kind of work usually happens, based on postings you have actually read. If every posting you respect is hybrid, your plan should say so, and your questions for the fair or the screen should include the hybrid pattern. If you need a paid job this term and the only offer is a night shift on campus, take the facts of that job seriously. Campus work can be the room where someone learns your name. It counts when you can describe the task. It does not become a mentorship story just because it was in a building.
Keep a short list of people who have already seen you work: a supervisor, a lab partner, a client from a class project. The Fed’s point about feedback is also a point about who can describe you later. A remote internship that never puts you on a call with the person who would write that description is a thinner asset than the posting photo suggests.
Do not rearrange a major around one paper. The 3.1 to 3.7 move is real and it is small next to the share of young graduates who are employed. The planning use is local. When two roles are close, prefer the one where someone will watch the work and tell you the truth about it.
How Flightpath helps
The portal plan by year can hold the constraint in a sentence you write: on-site or hybrid, a named reviewer, a day-60 artifact. When you pin a job, keep the location the posting states with the row. The weekly check-in is the moment you look at that sentence again, before you add another role that ignores it.
Flightpath does not rank a remote role above an on-site role, or the reverse. It ranks from facts you confirmed and from eligibility. A role you cannot take stays labeled and is never placed above one you can take. People you could ask about the team come from public contacts, or you get a search link. A note that asks where the team sits still waits for your approval before it is sent. The Fed’s percentage stays in the article. Your plan stays specific to the offers in front of you.