00:00:00
The US labor market has an
interesting mix going on right
00:00:03
now. You have hiring leveling
off and not adding to what we've
00:00:07
seen at the same levels in
recent years, but we aren't
00:00:11
seeing a lot of people being
laid off either. It does seem
00:00:14
like we are also seeing a bit of
rightsizing, to a degree, for a
00:00:19
variety of reasons. So what does
this tell us about the state of
00:00:22
the labor market right now?
Peter Cappelli is a Professor of
00:00:25
Management and Director of the
Center for Human Resources at
00:00:28
the Wharton School. Peter, great
to catch up with you again.
00:00:30
Thanks for your time.
- Thank you, Dan.
00:00:33
How are you reading the
00:00:34
labor market right now and all
that we are seeing go on?
00:00:39
You know, it really depends who
you're talking to. I've been
00:00:43
talking to our Ph— to our MBA
students, anyway. And you know,
00:00:47
from their perspective, it
doesn't look great. And I think
00:00:52
the reason why, if you look
at the job openings from the BLS
00:00:55
data, they're really down for
some segments of the economy,
00:01:00
and mainly it's down for white
collar jobs, especially finance-
00:01:05
related jobs. It's fallen by
about half since last year. If
00:01:09
you're in construction, openings
are actually up. So, you know,
00:01:13
it does seem to depend where you
are, quite a bit. I think, you
00:01:19
know, there is just a ton of
uncertainty about how things are
00:01:22
going. And businesses hate
uncertainty. So you know, they
00:01:27
think they're kind of sitting on
the sidelines.
00:01:29
Which I guess then ties into
the fact that we aren't seeing
00:01:33
more layoffs or more job cuts,
is that they're not certain
00:01:37
we're going to be bad as an
economy, but they're not certain
00:01:40
we're going to be great either,
and you still want to have some
00:01:43
level of status quo.
00:01:45
Yeah, no, I think that's right.
There's also something quirky
00:01:48
about the labor market that is
important to understand. It's
00:01:52
not new. But now, the current
decades or so, when a vacancy
00:01:58
occurs, it's much more likely
that it's filled by somebody who
00:02:02
already has a job someplace
else. So that creates what
00:02:07
researchers call vacancy chain
problems. So you know, if Dan
00:02:11
quits and Peter then takes Dan's
job, it creates another vacancy
00:02:17
over here, and then somebody
takes that person's job, and
00:02:21
then, bang, bang, bang, bang,
bang, you could have a cascade
00:02:24
of vacancies and a cascade of
job openings, right? - Right.
00:02:28
So when employers start hiring and
creating a new position, it
00:02:34
creates, in the data, several
positions over time, right?
00:02:38
Because people are moving from
one to the other, it's kind of
00:02:41
what we used to call frictional
unemployment in those days.
00:02:44
And then when they stop hiring, then
not only is there not their
00:02:48
initial job, but there's not
that whole chain of vacancies
00:02:52
that happen afterwards. So as
soon as hiring and creating new
00:02:56
positions stops, then it looks
like things have really fallen
00:03:01
off a cliff compared to the way
they were before, right?
00:03:04
And does that also mean that
we're also not seeing the level
00:03:07
of expansion within companies
that maybe we're used to seeing?
00:03:10
Companies adding— instead of
having 10 people in a— in a
00:03:14
segment and maybe raising that
to 20 because business is good,
00:03:18
they're not even doing that
right now.
00:03:20
Right. And that is the big
puzzle, right?
00:03:22
Because the economy seems
00:03:23
to be doing pretty well in the
fourth quarter. Way better than
00:03:27
expected, 4% plus growth rate.
And you would think you'd be
00:03:32
hiring in that context. And I
think this is the quirky thing.
00:03:35
It doesn't fit usual modeling
very well. And that is that
00:03:39
business leaders, particularly
in corporations, have in the
00:03:42
back of their heads a kind of
idea about what they should be
00:03:47
doing with respect to hiring.
And that is driven heavily by
00:03:50
investors who really don't like
hiring. They really don't like
00:03:55
employment costs. They see them
as fixed costs, even though you
00:03:59
can lay people off and they're
actually not fixed. So I think
00:04:02
there has been a kind of
resetting of their expectations,
00:04:05
driven by investors who, for
some reason— maybe it started in
00:04:09
tech. Think that companies are
overstaffed all over the place.
00:04:13
- Right. - And when they start
cutting jobs in tech, then
00:04:16
everybody else in tech looks
around and says— the boards say,
00:04:19
why aren't we cutting? And then
they cut, and then everybody
00:04:23
else gets the same pressure. The
cuts in white collar jobs are
00:04:27
easier to make because you don't
see the consequences
00:04:31
immediately. If you cut a
production worker and a
00:04:35
frontline worker, you know, you
don't have people to staff the
00:04:38
restaurant. You don't have
people to move stuff in the
00:04:40
warehouse. You cut a white
collar job, people, at least for
00:04:44
a little while, try to cover the
work, right? - Yeah.
00:04:47
So you can imagine that it's not
00:04:49
hurting you as much
when you cut white
00:04:50
collar jobs. But eventually it
turns around and bites you too.
00:04:53
How do you view the
question of the impact of AI
00:04:58
moving forward? And I'll ask you
that from a couple perspectives,
00:05:01
because recently we saw a
company like Dow Chemical cut
00:05:06
4500 jobs. And they said it—
directly said it was tied to AI
00:05:10
adjustments. Then you have a
company like Amazon that, you
00:05:14
know, has been cutting over the
last couple of years, just
00:05:17
announced 16,000. Didn't use the
same phraseology, but you got
00:05:22
the sense that some of it was
probably tied to that as well.
00:05:25
Right.
00:05:26
So I think— I didn't make up this
phrase, but it is attributed to
00:05:31
me now, and it's AI washing. And
that means that they're claiming
00:05:37
it's AI even when it's not AI.
And the reason is they're under
00:05:40
enormous pressure, from investors
in particular, to show that
00:05:45
they're using AI, because the
hope is AI will allow you to cut
00:05:49
jobs. Cut headcount. Immediately
jumps your valuations. You look
00:05:54
like a much more valuable
company when that— when you can
00:05:57
cut headcount. So they felt the
pressure to claim they were
00:06:00
doing it even when they weren't.
There's actually some survey
00:06:04
data from this from executives,
where about a third of them
00:06:07
claimed that they were reported
anonymously, that they were
00:06:10
doing this. That their AI stuff
was really performative. It
00:06:14
wasn't really doing very much.
So there's a lot of claiming
00:06:17
it's AI, even when it's probably
not. You know, companies that
00:06:21
are a little more honest like
this, which includes Amazon,
00:06:25
say— if you read what they
actually say, they say they
00:06:28
expect that the work of the
people being cut will be taken
00:06:34
over by AI. It hasn't yet,
right? - Right. - You know, my
00:06:38
looking at individual companies
and when they introduce AI is
00:06:43
that you just don't see
headcount dropping. What you see
00:06:48
is eventually able to take on
more work with the same number
00:06:53
of people. Now you might think
that's the same thing, but it's
00:06:55
really not. So if you think
about computer programmers, for
00:06:58
example, 30% of their time— only
30%— is writing code. So let's
00:07:04
say AI takes over that whole
task. They've still got 70% of
00:07:08
their time doing other things,
which is talking to clients and
00:07:11
people in the organization,
figuring out what our needs are,
00:07:14
managing their expectations,
getting budgets, all that stuff.
00:07:17
So if you took that 30% away,
you still have 70% of those— of
00:07:22
that work that has to be done.
So you're not, at least in the
00:07:25
short term, cutting any of those
people, nor can you just ram
00:07:29
them together and cut one
programmer, right?
00:07:32
Well, and I guess it brings up
an interesting point that you
00:07:34
kind of led into before, is that
there are many businesses out
00:07:38
there that maybe are still
trying to figure out what their
00:07:42
future hiring levels longer
term are going to look like.
00:07:46
- Yeah. - Because they don't know
exactly how AI is going to
00:07:50
impact their productivity, and
they don't want to make these
00:07:52
moves just yet.
00:07:54
Yeah, I think that's— you know,
that's a charitable view.
00:07:57
- Okay. All right.
- Whatever— it might be,
00:08:00
that they're just
cutting because they feel they
00:08:01
have to cut. And for sure,
though, I think they're
00:08:05
reluctant to staff up now. Part
of it might be because they're
00:08:09
not sure how AI will play out,
but I would say more of it is
00:08:13
just because of the uncertainty
generated largely by the
00:08:16
administration, trade policies,
federal government contracts,
00:08:20
all that kind of stuff. Where is
it going? It changes day by day.
00:08:24
And if you're a business person,
you want to place big bets. Even
00:08:28
medium bets, until you have a sense
of where that's going. And it—
00:08:32
you know, it hasn't settled down
right? So not surprisingly,
00:08:35
they're sitting on their
hands a bit.
00:08:37
So would it be safe to say then,
at least maybe in the short term,
00:08:41
that we're probably looking at
more of the same, like this
00:08:44
pattern that we've kind of been
in for the last several months,
00:08:47
that that's going to continue
while all of this kind of
00:08:51
works itself out?
00:08:53
Yeah, I mean, I think the
question is, will it
00:08:55
settle down? And it's not clear
it's going to settle down. But I
00:08:58
think at some point employers
have to decide, you know, we
00:09:03
just can't do nothing. We
just can't keep waiting, because
00:09:06
it's not going to settle down.
So maybe we have to start
00:09:09
placing bets. You know, at
the Wharton School and at other
00:09:12
institutions with students
graduating, we noticed this a
00:09:16
lot, because college graduates
take the first hit. The easiest
00:09:21
thing to do if you want to
shrink your workforce is to use
00:09:24
attrition. That means just don't
fill vacancies when they pop up.
00:09:29
And if you're going to fill
positions, the most challenging
00:09:32
way to do it is to hire new
graduates in and train them.
00:09:37
It's a longer term, slightly
longer term bet than just
00:09:41
pulling somebody from someplace
else and popping them in. So the
00:09:44
hiring that does take place has
shrunk down the most, I think,
00:09:48
for new college graduates,
right?
00:09:50
Peter, always great to get your
thoughts and your— and spend some
00:09:54
time with you. I appreciate it.
We will talk to you again very
00:09:57
soon, sir. Thank you.
00:09:59
Good. Thank you, Dan.
00:10:00
Peter Cappelli,
00:10:01
Professor of Management here at
the Wharton School.
