Latest / How Canadian Markets Work / Episode 23: Reading the Labour Market

Transcript

Episode 23: Reading the Labour Market · How Canadian Markets Work

0:00
So for today's deep dive, we are pulling from
0:03
a rather fascinating stack of sources. Yeah,
0:05
we've got a lot of interesting material this
0:07
time. We really do. We have these. incredibly
0:10
dense methodological manual straight from government
0:13
labor bureaus, some empirical research papers
0:17
that analyze survey data distortions, and some
0:21
really sharp economic analysis on the widening
0:24
gap between official statistics and the actual
0:27
realities of the labor market. Which is a gap
0:29
that I think a lot of people feel even if they
0:31
don't know the exact math behind it. Exactly.
0:34
And so our mission for you today is simple, but
0:36
we think it's crucial. We want to equip you with
0:39
the tools to actually read labor market data.
0:42
So that you never get fooled by a misleading,
0:45
overly cheerful headline again. Which happens
0:47
a lot more than you'd think. It really does.
0:50
And to kick things off, I want to pose a riddle
0:52
directly to you, the listener. It sounds like
0:54
a statistical impossibility. But it actually
0:57
happens in the real economy all the time. Oh,
0:58
this is a good one. Right. So here is the scenario.
1:02
How can the national unemployment rate fall from
1:05
6 % down to 5 .5 % without a single new person
1:08
actually finding a job? It sounds like a trick
1:11
question, honestly. It does. I mean, it goes
1:13
against everything we're taught to look for.
1:15
Yeah, because we are so conditioned by the evening
1:18
news to believe that a falling unemployment rate
1:21
just... Mathematically equals a booning hiring
1:24
market. Right. Less unemployment means more jobs.
1:28
Right. That's the assumption. But the answer
1:31
to your riddle is actually quite sobering. That
1:35
headline rate drops when a group of people simply
1:38
hit a wall, give up, and stop looking for work.
1:40
They just stop looking. Exactly. The very moment
1:43
they stop searching, the statistical machinery
1:45
essentially just erases them. They vanish from
1:47
the official unemployment count entirely. Wow.
1:50
And that disappearing act is exactly why we need
1:52
to dig into these sources, because the headline
1:55
number, the one that drives stock market rallies
1:58
and dictates interest rate policy, it's often
2:00
a mirage. It really is. To understand how a human
2:03
being can just, you know, vanish from the data,
2:06
we really need to look at the rigid architecture
2:08
of how statistical agencies categorize the working
2:11
age population, because they don't just look
2:13
at who is working and who isn't. Far from it.
2:16
The government sorts the entire working age population
2:20
into three very specific, highly inflexible buckets.
2:25
Okay, three buckets. Yeah. So the first bucket
2:27
is the employed bucket. Now, the threshold to
2:29
get into this bucket is remarkably low. How low
2:32
are we talking? Well, it basically asks, did
2:35
you do any paid work during the survey's reference
2:37
period? which is usually the week containing
2:40
the 12th of the month. Any paid work at all.
2:42
Any paid work. That includes part -time work,
2:45
gig work, or being self -employed. If you earned
2:48
a paycheck for even a few hours of work, say
2:50
driving an Uber for one evening, you were swept
2:53
right into the employed bucket. So the bar for
2:56
success, statistically speaking, is just any
2:58
paid labor at all, even an hour. Yes, exactly.
3:01
Now the second bucket is the unemployed bucket,
3:03
and this is where the rules become incredibly
3:04
strict. Right, because you don't just go in there
3:07
if you don't have a job. No, not at all. You
3:09
do not get put into this bucket just because
3:11
you lack employment. To be officially unemployed,
3:14
you must have no work. You must be currently
3:16
available to take a job. And most crucially,
3:20
you must be actively searching for one. I want
3:23
to pause on that phrase, actively searching,
3:25
because our sources highlight just how subjective
3:28
that can be. What actually counts as an active
3:31
search to the person conducting the survey? And
3:33
that is exactly the friction point. If a surveyor
3:36
contacts your household and you say you've been
3:39
browsing job boards or reading industry newsletters.
3:42
Which feels like searching. It feels like searching
3:44
to you, sure. But to the Bureau, that does not
3:46
count. To be actively searching, you must have
3:49
taken specific concrete actions in the last four
3:52
weeks. Like what kind of actions? submitting
3:55
a resume, interviewing for a position or registering
3:57
with an employment agency. If you haven't done
4:00
those specific things in the last four weeks,
4:02
you are pushed out of Bucket 2 and straight into
4:05
Bucket 3. The catch -all bucket. Right. the not
4:09
in the labor force bucket. This holds absolutely
4:12
everyone else. So full time students, retirees.
4:15
Yeah, full time students, retirees, people staying
4:17
home to care for family. But importantly, it
4:20
also holds what economists call discouraged workers.
4:23
Right. The people who just hit a wall. Exactly.
4:26
These are people who desperately want a job,
4:28
but maybe the local factory closed, they've sent
4:30
out a hundred resumes with no response, and they're
4:33
so frustrated they just stopped applying in the
4:35
last four weeks. Okay, so let's tie this back
4:37
to the riddle from the beginning, because the
4:39
math is where the illusion really takes shape.
4:42
It is. The denominator for the headline unemployment
4:45
rate, what we call the total labor force, is
4:48
only bucket one plus bucket two. It's just The
4:51
employed plus the officially unemployed. Right.
4:53
It entirely ignores bucket three. It does. And
4:56
to see how easily this data can be manipulated
4:59
by despair rather than success. Let's run the
5:02
math on a tiny isolated economy. OK. I'm picturing
5:05
a very small town. Yeah. Imagine a town with
5:08
exactly 100 working age people. 60 of them are
5:11
employed and six of them are actively sending
5:14
out resumes. So they're in bucket two. Exactly.
5:16
And the other 34 are retirees or students in
5:20
bucket three. In this scenario, our labor force
5:23
is the 60 employed plus the six looking, so 66
5:27
people. Got it. The labor force is 66. Right.
5:30
So to get the unemployment rate, we divide the
5:32
six active searchers by that labor force of 66.
5:36
That gives us an unemployment rate of roughly
5:38
9%. Which is pretty high. It is. Now, let's let
5:42
a year pass. The economy stagnates. Not a single
5:47
new job is created in this town. The exact same
5:50
60 people are employed. But of those six original
5:53
job seekers, three of them get completely exhausted
5:56
by the endless rejections. They just give up.
5:58
They give up. They stop setting resumes. So they
6:00
transition from bucket two into bucket three.
6:02
So our labor force shrinks. Shrinks from 66 down
6:05
to 63. Now let's run the headline calculation
6:07
again. We divide those three remaining active
6:10
searchers by the newly shrunken labor force of
6:12
63, which gives us an unemployment rate of roughly
6:16
4 .7%. Wait, hold on. A near halving of the headline
6:19
rate. Yes. It dropped from 9 % to under 5 % just
6:23
because three people got tired of sending out
6:25
resumes. That is exactly how the math works.
6:27
But the economy... didn't create a single job
6:30
in that scenario. If you just saw that 4 .7 %
6:34
flash on a news ticker on a Friday morning, you
6:37
would assume a massive economic boom. You would
6:39
completely assume a victory. When the underlying
6:42
reality was actually localized despair, the aggregate
6:46
statistic improves precisely because the individual
6:49
reality got worse. Wow. It reminds me of sort
6:52
of the modern dating market. Or maybe the housing
6:55
market. Oh, how so? Well, think about it. If
6:57
you are single and you are constantly going on
7:00
terrible first dates, you know, sending messages
7:02
on apps, you are an active searcher. Right. You're
7:04
in bucket two. Exactly. But it's exhausting.
7:08
And if you get so burned out that you just delete
7:10
the apps and stop going out, you haven't magically
7:13
found a partner. You are still single. But you
7:15
just evaporated from the market entirely. Yes.
7:18
The statistics for those remaining on the apps
7:20
might suddenly look less crowded and more favorable,
7:23
but it's only because a huge cohort just walked
7:26
away in frustration. That captures the dynamic
7:29
perfectly. You removed yourself from the denominator.
7:32
Right. Which naturally raises a critical question
7:35
for anyone trying to analyze the economy. If
7:37
the headline unemployment rate can actively mislead
7:40
you by shrinking when conditions deteriorate,
7:43
how do you find the truth? We definitely need
7:45
the antidotes to the headline. And looking at
7:48
our sources, they point heavily toward a metric
7:50
called the participation rate. Yes, the participation
7:52
rate is vital context. It measures the share
7:55
of the entire working -age population that is
7:57
actually in the labor force, meaning they're
8:00
either working or actively looking. So it looks
8:02
at the whole picture. Exactly. When you read
8:04
this alongside the unemployment rate, the true
8:07
narrative emerges. If unemployment falls, but
8:11
participation also falls. That is the illusion
8:15
we just walked through. People are deleting the
8:16
app. Right. People are leaving, not finding work.
8:19
But if unemployment falls while participation
8:22
rises, that means new people are entering the
8:25
market and getting hired. That is... genuine
8:27
organic growth. OK, that makes sense. But hold
8:29
on. If we rely on the participation rate, doesn't
8:32
that have a pretty massive demographic blind
8:35
spot? It can. Like, say we have a massive wave
8:38
of baby boomers retiring. They leave the labor
8:41
force naturally. Right. They go into bucket three.
8:43
So the participation rate plummets. But that's
8:46
not a sick economy. That's just the passage of
8:48
time in an aging population. How do we separate
8:51
economic distress from just you know, natural
8:53
aging. That is an excellent pushback. And honestly,
8:57
it's a trap a lot of amateur analysts fall into
8:59
when they just look at raw participation. To
9:01
filter out the noise of an aging population,
9:03
economists use a much sharper tool. Which is?
9:06
The prime age employment to population ratio.
9:08
The prime age EPOP. Right, EPOP. It isolates
9:12
adults between the ages of 25 and 54. OK. That's
9:16
the demographic that should theoretically be
9:19
working if the economy is healthy. It simply
9:21
asks. out of all the prime -age adults in the
9:25
country, what percentage actually have a job?
9:27
So it completely sidesteps the murky question
9:30
of who is actively searching. Exactly. It strips
9:33
out the distortions of college students and retiring
9:36
boomers. If the prime -age EPOP is rising, you
9:39
know the labor market is genuinely pulling people
9:41
into productive work. So for you listening, this
9:43
is a highly practical takeaway. The next time
9:46
a jobs report drops, look right past the headline,
9:49
unemployment rate. look at the participation
9:51
rate, and specifically that prime age employment
9:54
ratio. Yeah, that's where the real story is.
9:56
It's like the headline is the marketing department,
9:58
but the prime age ratio is the audit of the company's
10:01
books. That's a really great way to frame it.
10:02
But, you know, even when we look at the prime
10:04
age ratio, we still have to contend with another
10:07
distortion our sources highlight. Which is? These
10:10
surveys treat all jobs as completely identical.
10:13
They treat the economy as just a uniform block.
10:16
Oh, right, because earlier you mentioned that
10:18
working just a few hours a week puts you in the
10:20
employed bucket. A job is a job is a job, according
10:22
to the data. Exactly. And that leads to the involuntary
10:25
part -time problem. What does that look like
10:27
in practice? Well, if a person is working 10
10:30
hours a week at a checkout counter, they go into
10:33
the exact same statistical bucket as a corporate
10:36
architect. working 60 hours a week. Which is
10:38
wild. It is. Now, if that 10 hour worker is,
10:41
say, a semi -retired person who only wants 10
10:44
hours, that's a perfect match. Yeah. But if that
10:46
person is a parent who desperately needs 40 hours
10:49
of work to feed their family but can only find
10:52
a single shift, they are still counted as a full
10:56
success story in the headline rate. Wow. It's
10:59
like looking at a company's gross revenue without
11:01
ever checking its profit margins. The top -line
11:04
revenue of the employment rate looks massive
11:06
and healthy. But underneath, the quality of those
11:09
jobs might be deteriorating. People are employed,
11:12
but they are severely underemployed. Which is
11:15
why the statistical agencies actually publish
11:17
a whole spectrum of unemployment rates. They
11:19
range from U1, which is the narrowest, all the
11:22
way up to U6. And the headline rate we always
11:24
hear about on the news, that's U3, right? Yes,
11:26
U3. But U6 is the broadest measure. It includes
11:30
those discouraged workers who gave up, and crucially,
11:33
it includes those involuntary part -time workers.
11:36
And how much of a gap is there usually between
11:38
the headline rate, U3, and this broader U6 rate?
11:41
It's substantial. Even in a healthy economy,
11:43
economy, the U6 rate is often three to four full
11:47
percentage points higher than the headline rate.
11:49
That's a huge difference. And during a recession,
11:52
that gap can absolutely explode. So if you only
11:55
look at the headline, you are completely blind
11:58
to the millions of people who are just clinging
12:00
to the margins of the labor market. OK. So if
12:02
the definitions are this broad and the aggregates
12:05
mask this much underemployment, Let's talk about
12:08
the Friday morning data drops. Oh, the famous
12:10
Friday payrolls? Exactly. Every month, these
12:13
numbers are released and financial markets go
12:15
into an absolute frenzy. Billions of dollars
12:18
move in milliseconds. But based on our sources,
12:21
should we even trust a single month's data point?
12:25
The short answer is absolutely not. Really? Not
12:27
at all? No. Treating a single month's employment
12:30
data as gospel is one of the most dangerous things
12:33
an investor or analyst can do. And it comes down
12:36
to how this data is actually gathered. OK, how
12:39
is it gathered? In the US, for example, the Friday
12:41
jobs report is actually two separate surveys
12:44
mashed together. You have the household survey
12:46
and the establishment survey, which is also known
12:48
as the payrolls. And they often violently contradict
12:51
each other, right? Why is that? Because they
12:53
measure different things using completely different
12:55
methodologies. The household survey calls about
12:58
60 ,000 households and asks people about their
13:01
status. So it counts people. It counts people.
13:03
Got it. But the establishment survey asks about
13:06
119 ,000 businesses how many people are on their
13:09
payroll. So it counts jobs. OK, so if I take
13:12
on a second job to make ends meet. That is exactly
13:15
the distinction we need to make. If you take
13:18
a second job. The Household Survey still just
13:22
sees one employed person. Right. I'm still just
13:24
me. But the Establishment Survey sees two distinct
13:28
jobs. Oh, wow. So, payrolls can surge, making
13:31
it look like a booming economy, when in reality
13:34
it's just the same people taking on multiple
13:36
part -time gigs just to survive inflation. That
13:39
is a massive discrepancy. And it gets messier.
13:42
The payroll survey relies heavily on something
13:44
called the birth -death model. Birth and death
13:46
of businesses, I assume? Right. The Labor Bureau
13:49
has to mathematically guess how many new businesses
13:51
opened and closed that month, because they obviously
13:54
can't survey a company that doesn't exist yet,
13:56
or one that just went bankrupt and isn't answering
13:59
the phone. Makes sense. But this modeling often
14:01
overestimates job creation at the start of a
14:04
recession, and it underestimates it during a
14:07
recovery. And then you add in, say, a random
14:09
weather anomaly, like a massive blizzard during
14:12
the reference week that keeps hourly workers
14:14
off the payroll, and the numbers get completely
14:17
scrammed. Precisely. Which is why the rule of
14:20
thumb among serious economists is to ignore the
14:23
single month noise. Just ignore it. Yeah. And
14:25
only look at the three month or six month rolling
14:27
average. The trend is the actual signal. The
14:30
single month is a volatile guess that will likely
14:34
be revised twice before it's even finalized.
14:37
That is so important to keep in mind. Well, let's
14:40
pivot slightly to what happens after someone
14:42
actually lands securely in that employed bucket,
14:44
because having a job is only half the equation.
14:47
Very true. Our sources make a point of biving
14:50
into wage growth, which really feels like the
14:52
missing link here. The participation rate tells
14:55
us people are working But it doesn't tell us
14:57
if the labor market is actually tight enough
14:59
to force employers to pay them more. And this
15:02
brings us to the crucial difference between nominal
15:04
wages and real wages. Right. When the data says
15:07
wages grew by 4 percent, that is nominal growth.
15:10
That's the raw numbers printed on paychecks.
15:12
But nominal growth in a vacuum is meaningless
15:15
without weighing it against the cost of living.
15:17
Because if your rent and groceries go up faster
15:19
than your paycheck, you are objectively poorer,
15:22
even if you got a raise. Exactly. Why are wages
15:25
so tricky in the data? Why do they sometimes
15:28
look strong, even when the economy is clearly
15:31
failing? Because nominal wages are notoriously
15:34
sticky. Sticky? Yeah. Think about the psychology
15:38
of a business owner during a downturn. If revenue
15:41
drops, they very rarely call their employees
15:44
into a room and say, hey, everyone, you're all
15:46
getting a 5 % pay cut. Yeah, that would destroy
15:48
morale. People would quit on the spot. Right.
15:51
So what do they do instead? They freeze wages.
15:53
Exactly. They give 0 % raises. And they let inflation
15:57
do the dirty work of cutting their real labor
15:59
costs over time. Ah, that's clever but brutal.
16:02
It is. And this means that aggregate nominal
16:05
wage data rarely drops sharply. To see the true
16:08
health of the worker, statistical agencies have
16:11
to calculate real wages, adjusting that nominal
16:14
stickiness against inflation. So if real wages
16:16
are negative? If real wages are negative, the
16:18
consumer is bleeding out, no matter how low the
16:21
headline unemployment rate is. That brings us
16:23
to one of the most profound insights in our source
16:25
material. We have this noisy, subjective data.
16:30
We have sticky wages. We have surveys contradicting
16:33
each other. It's a lot to untangle. It is. And
16:36
all of this culminates in a reality that feels
16:38
completely counterintuitive, and that is that
16:41
employment data is the ultimate lagging indicator.
16:44
It really is like looking through the rearview
16:46
mirror. Employment data tells you where the economy
16:49
was months ago. not where it is today, and certainly
16:52
not where it is going. But why is the lag so
16:54
severe? I mean, why doesn't it reflect right
16:56
now? Because hiring and firing are incredibly
16:58
expensive, high -friction decisions. If you run
17:01
a company and sales dip for a quarter, you don't
17:05
instantly fire your best engineers. Right, because
17:07
severance is expensive. Exactly. And the loss
17:09
of institutional knowledge is devastating. Plus,
17:12
if the dip is temporary, rehiring and training
17:14
replacements next year will cost you a fortune.
17:16
So businesses just hold on. Yeah, they practice
17:18
what's called labor hoarding. They hold onto
17:20
their staff, hoping the storm passes. They only
17:23
fire people when they're absolutely backed into
17:25
a corner. It makes me think of steering a massive,
17:29
heavy cargo ship across the ocean. But the captain
17:32
is only allowed to navigate by looking out the
17:34
back window at the wake trailing behind the ship.
17:36
That's a great visual. When the captain turns
17:39
the wheel, the immense momentum of the ship means
17:41
it takes miles to actually change course. And
17:44
you only see the result of that turn in the churning
17:46
water minutes later. Exactly. So a strong jobs
17:49
report isn't a high definition photograph of
17:52
the economy today. It is a photograph of cautious
17:54
decisions made by executives three, four, or
17:57
maybe six months ago. The cargo ship analogy
18:00
is spot on. The momentum is just huge. And this
18:04
is exactly why the unemployment rate is almost
18:07
always the very last metric to recover after
18:09
a recession ends. Right. Because they're still
18:11
nervous. The stock market bottoms out. GDP starts
18:14
growing. But unemployment stays painfully high
18:17
for another year because employers are still
18:20
staring at the wake, waiting for absolute certainty
18:22
before they take on the long term liability of
18:24
a new salary. So when we step back and look at
18:27
this entire machine, the picture is incredibly
18:30
nuanced. The headline, unemployment rate, is
18:33
really just a thin slice of the story. A very
18:36
thin slice. And it's heavily dependent on the
18:38
rigid definition of who is actively searching.
18:41
If discouraged, workers give up, the headline
18:44
artificially improves. Right. We have to look
18:46
at the prime age employment ratio to see through
18:49
the demographic noise. We have to watch the U6
18:52
rate to catch the invisible struggle of underemployment.
18:55
And we have to average out the monthly survey
18:57
anomalies and weigh any wage growth against the
19:00
silent theft of inflation. And above all, we
19:02
must remember that we are analyzing a delayed
19:05
reaction. The labor market is the final domino
19:08
to fall into recession, and it's the last one
19:10
to be stood back up in a recovery. Which leaves
19:12
us with a rather provocative thought to carry
19:14
with you. Oh, definitely. We've established that
19:16
employment is a lagging indicator. It's the wake
19:19
of the cargo ship. But we also know that financial
19:22
markets are inherently forward looking. They
19:24
try to price in the future before it happens.
19:26
So if you are watching a news ticker and you
19:29
decide to adjust your retirement portfolio based
19:31
on a sudden official jump in the unemployment
19:34
rate. What are you actually doing? That is the
19:37
million dollar question. If the data is lagging
19:40
and the market has already moved on, by the time
19:42
a recession is finally officially printed in
19:44
the headline unemployment rate, could you be
19:47
making major financial decisions based on the
19:49
ghost of an economy from six months ago? It is
19:52
a question every investor needs to ask themselves.
19:55
Are you trading on tomorrow's reality? or are
19:58
you just reacting to last season's weather? Something
20:00
to carefully consider the next time the breaking
20:03
news banner flashes across your screen. Until
20:05
next time, keep looking past the headlines and
20:08
thank you for joining us on this deep dive.