Panel 1: Approaching Labor Market Definition (3 of 7)


DR. DRENNAN:
Okay, everyone. Let’s get started. Okay. So if we are presented
with a particular merger or other market behavior
that raises a potential antitrust concern in the
labor market, how does one investigate that? How do you gather the
right information in a hopefully efficient way,
make a determination as to whether there is or
is not a violation of antitrust laws? Our current panel will
discuss how we might approach market definition
in the context of exploring these
labor issues. Market definition is only
one step in this process. But it can be an important
step, as the breadth of the market may go a long
way to explaining why the behavior we are
investigating should or should not be expected
to lead to harm. For this discussion, we
are extremely fortunate to have a very
distinguished panel assembled here. While I’m sure it’s only
going to touch on some of these topics, hopefully
this discussion will build upon the excellent
overview provided by Professors Marinescu
and Prager. Okay, so let me
introduce our panel. Seated directly next to
me is Patrick Greenlee. Dr. Greenlee is an
outstanding economist who’s been with the antitrust
division since 1995. During this time, he’s
worked on several important matters,
including matters that raise buy-side or
monopsony issues. In 2000-2001, he was
selected as the Victor H. Kramer Foundation fellow
at the University of Chicago Law School. Dr. Greenlee has taught
economics in the PhD and master’s programs at the
University of Maryland, of Virginia and
Johns Hopkins. He’s also an active
researcher who’s published articles in several
economic and antitrust journals. And let me take this
opportunity to cover Patrick, myself and the
other DOJ moderators that follow with just
a disclaimer. We DOJ speakers do not
purport that our views reflect those of the U.S. Department of Justice,
and thus nothing in this presentation may be
cited in any enforcement proceeding
against the U.S. Department of
Justice, right. The purpose of today is
to stimulate discussion. Sometimes that involves
playing devil’s advocate, et cetera. MR. GREENLEE:
Thank you, Ron. DR. DRENNAN: Yes. I saved you from
having to do it. Seated next to Patrick
is Dean Harvey. Dean Harvey is an attorney
with Lieff Cabraser Heimann & Bernstein. He represents individuals
and companies in antitrust, business
tort, employment and intellectual
property litigation. He has experience
representing employees asserting antitrust
claims, including high-tech workers,
doctors, rail equipment workers and fast
food workers. He has also represented
consumers in litigation against generic
drug makers. Dean was previously with
the antitrust division and was a law clerk for
the Honorable James V. Selna of the U.S. District Court
in California. Seated next to Dean
is Dr. Kevin Murphy. Dr. Murphy is a professor
of economics at the Booth School of Business in the
University of Chicago. He is also a research
associate at the National Bureau of Economic
Research and a member of the Hoover Institution. He has received the highly
prestigious John Bates Clark Medal, the MacArthur
Fellowship and the John von Neumann Award. Professor Murphy has
published scores of scholarly articles across
a wide range of topics, including economic growth,
competition, relative wages, unemployment, as
well as income inequality and how it is linked to
demand for skilled and unskilled labor. And finally, at the end
of the table is Dr. Orley Ashenfelter. Dr. Ashenfelter is a
professor of economics at Princeton University. Professor Ashenfelter has
also been the director of the industrial relations
section at Princeton, a director at the U.S. Department of Labor,
a Guggenheim fellow. He is a recipient of
the IZA prize in labor economics, the Mincer
Award for lifetime achievement in labor
economics and the Karel Englis Medal. His specialization
include labor economics,
econometrics and the intersection of
economics and law. His research has explored
wage rates and many other issues related to the
economics of labor markets. So that’s our panel. As I said, I’m very
happy, and thank them all for being here and
their preparation. What I’d like to start
with, before we get into it, is make use of having
our attorney on the panel and ask for just a little
antitrust labor 101, if you will, and what role
— and within that, what role dose market
definition play. Is it always relevant or
is it only in a subset of cases? So, Dean, would you? MR. HARVEY:
Thank you, Ron. First, let me say it’s
an honor to be here. I think it’s an important
topic and I’m delighted that it’s receiving all
the attention that it’s getting lately. Before I start, I’d love
to get a sense of the audience. If you consider yourself
not to be an antitrust specialist, could you
please raise your hand? Good. Who here considers
themselves primarily to be an employee or
a labor lawyer? Great, okay. So in terms of the basics
of antitrust in this area, I think the main
reference is the joint DOJ/FTC 2016 guidelines
for HR professionals. As was stated by I think
two of our prior speakers, the antitrust laws apply
in the employment setting in just the same way as
they apply in the product setting. Competing employers cannot
agree to set salaries, just as competing sellers
of products cannot agree to set prices. Competing employers also
cannot agree not to hire or even not to actively
solicit each other’s employees. And this is kind of a
subtle distinction where it is just as unlawful
under this guidance to agree not to affirmatively
seek out an employee but still allow someone to
apply and get hired. So that, if it’s naked,
is still a per se violation. The relevant harm here
is suppressed pay to employees, not increased
price to consumers. So there is I think a big
debate in the antitrust world about the consumer
welfare standard, should that be the end
all and be all. I think this should play
an interesting role in that debate because the
prices to consumers play almost no role
in these cases. It’s all about whether
the conduct at issue suppresses pay
to employees. When agreements are naked,
that is, when they are not reasonable necessary
to some pro-competitive collaboration between
the companies, then the misconduct, the agreement
is considered to be a per se violation of the
antitrust laws. That is, it is an
unlawful act without doing anything else to
investigate it. If the agreement is
arguably or it is in fact reasonably necessary to
some pro-competitive endeavor, then a more
deferential standard, the rule of reason, may
apply, in which case the first step of that may
be to define a relevant market, which is what
our panel is about. Violations can also give
rise to private causes of action for trial damages. And in those cases,
typically it’s a class of employees who were
affected by the misconduct who bring a private class
action to seek trial damages for lost pay. I should also mention
that per se violations are also potential criminal
violations, as was described earlier today. In terms of market
definition, I think there was a description of that
earlier this morning. So I won’t repeat it. But I think I would just
ask why do we bother with market definition. The reason why we do it
is to test whether the alleged monopolist or the
alleged monopsonist or a group of competing
employers have enough power to have an
impact on price. That is, do they have the
ability to suppress pay below the
competitive level. And I think with that,
I’ll stop on the 101 portion of this. DR. DRENNAN: Sure. Thanks. So, and before we — as
we get in closer to market definition itself, maybe
a time for our other panelists, starting with
Patrick, to think about — to talk about how do labor
market concerns compare to antitrust concerns
more broadly. MR. GREENLEE: Thanks, Ron. So my comments here are
going to echo a little bit of what we heard from our
economists on the initial panel and that’s just
let’s focus on thinking about labor markets
in a merger context. So this would be sort
of the example of the hospital study that was
at the end of our intro. Horizontal merger
guidelines are written focusing on competing
sellers selling products to consumers. But they say that they
apply equally as well to consuming buyers
purchasing inputs. We can think of labor
being as one particular category of input and
then ask how well do the horizontal merger
guidelines approach apply. So whether it’s selling
goods to consumers or purchasing inputs, either
labor services or other widgets that get added to
create a final product, a merger’s going to have
adverse competitive effects when the merging
parties compete against each other vigorously. In some sense, it’s
kind of obvious. It’s also going to need
to be the case that there are not a large number of
third-party alternative for that input supplier
to sell their services to switching, you know, a
particular employer, thinking about switching
— I’m sorry, a particular laborer
thinking about switching from one job to another,
it’s got to be costly or — I’m sorry, an employer
has market power or differentiation if it’s
the case that it’s difficult for an employee
to switch to another job or an input supplier to
sell their products to another creator,
another producer. Then there’s the
other category. So we have in the merger
guidelines that sort of need to be satisfied if
we’re going to end up with an adverse
competitive effect. These include that entry
or expansion by other competitors would not
defeat an attempted exercise of market power. So in the input case or a
labor case, this would be a situation in which, as
a result of a merger, two firms attempt to reduce
wages and, as a result, some other employer
perhaps already established in the market
or in an adjacent market swoops in and says, hey
look, I can purchase labor very cheaply now because
these guys are trying to exercise market power
following their merger. So entry expansion could
potentially defeat an attempt at merger-induced
market power. Another potential effect
would be whether or not a merger makes
collusion easier. So in the antitrust
lingo, this would be whether or not the
merger’s causing — you know, taking out a
maverick who otherwise is behaving very
competitively, very aggressively with respect
to purchasing the input. And then finally, in
merger review, we have to consider the possibility
that there might be some efficiencies that result
from a merger and assess whether those efficiencies
are actually specific to the merger, but they
couldn’t be otherwise achieved and sort of what
the magnitude of those are. Is it enough to
counterbalance some opportunity to
exercise market power. So, you know, that’s a
reminder of what the basic components of the
merger guidelines are. At the division, we do
look at output markets and we look at input markets. In many cases, a
particular input market might be much broader than
the output markets we are focused on. So, for example, consider
a hypothetical merger of two producers who make
manufactured goods, maybe kitchen appliances
of some type. And they use railroad
services to deliver their finished goods. So they are purchasers
of the input, you know, railroad services. It’s probably not an input
market that we would spend a lot of time
focusing on if we think that these two particular
producers of manufactured goods don’t represent a
very large portion of purchasers of rail
services, you know, in a particular geographic
area or nationally. So that would be an
example where we might have two firms competing
very aggressively head to head in an output market
with very few competitors in the output market. But they’re not able to
exercise market power in purchasing the input. So that’s a case where the
input market may be much broader, more competitive
than the output market. We do have examples
that are the reverse. So in many agricultural
markets, it’s not economical for an
individual grower to ship their finished good,
whether it’s a crop that they’ve harvested or
livestock that they’ve fed, it’s not economical
for them to ship them far, long distances to a
processor to, you know, turn that steer
into steak. And as a result, it may be
the case that the market for the finished goods,
the, you know, grocery store items might be
national or perhaps even broader. But for a particular
grower, you know, on their farm, they may have
very few alternatives for where they can
sell their input. It may be that, hey,
two processors that are the closest to that
particular farm are the ones that are merging. If th alternative
for that particular grower is quite
a bit — quite a distance away, you can imagine that
the merged firm would be able to strategically
reduce the price that they pay that particular
grower for their good. So this would be an
exercise of buyer market power. So we look at input
markets and we don’t — it’s not the case that
we assume that an input market is the same size
or breadth as an output market. I guess before I conclude
this part, you know, input markets, labor
is an input. So we think that the
guidelines should apply or the approach should apply. But I do recognize
that there are several characteristics of
labor markets that — characteristics not
necessarily unique to labor markets but maybe
play an especially prominent role in those,
in labor markets. The first one is that, you
know, if you’re thinking about a specific
interaction between an employer and an employee,
it’s not just the wage that matters presumably. It’s other work
conditions, sort of, you know, flexibility
in hours. Some of those things, like
where the establishment is located,
maybe is fixed. Some of those other
attributes might be things that the employer can
adjust, like maybe I’m going to hire fewer
workers and people are going to have to work
more hours if they want to keep the job, that
sort of thing. And that’s not to say
that in output markets we don’t face a
similar issue. You know, we may be
able to observe price. We might not know all
the attributes of a particular product. But this may be a larger
issue in labor markets. Another feature of labor
markets that might be different than many output
markets or even other input markets would
be that employment is typically, in many cases,
a long-term relationship. So we don’t have a spot
market where — in most, in many labor situations,
we don’t have a spot market where each day or
each week we’re deciding what the prevailing price
for labor is going to be that week and we
go week to week. Instead, long-term
relationships, in that setting, you can think of
it being the case that, you know, the
employer’s making some relationship-specific
investments, you know, training the employee on
how to do the job and the employee is
also making some relationship-specific
investments. They’re learning how to
do this particular job or the peculiarities
of given employer. Finally, the last two
attributes that are perhaps distinctive for
labor markets that we don’t see as much in
other markets would be the labor markets are
characterized as being search markets and
being matching markets. So search market being
that it takes effort for a jobseeker to
actually find a job. It also takes effort by
the employer to try to identify productive,
useful workers to work at that job. So there’s search plays a
perhaps prominent role. And matching, here I
mean that the benefit of having an
employer/employee relationship, the quality
of that relationship matters to both the
employee, sort of how satisfied they are with
their job and all that it provides, but the employer
also cares about the productivity or how
useful, how reliable the particular employee is. And you think of this
as being quite a bit different than, you know,
a grocery store that’s selling bottles
of shampoo. They probably don’t care
just how I’m using the shampoo or what my
relationship is with it. And so, in this way, labor
markets might be quite a bit different than a lot
of the markets we study that are not
labor markets. DR. DRENNAN:
Thanks, Patrick. In that regard, let’s
maybe now turn to our labor economist to
discuss for us antitrust practitioners what are the
important aspects about labor markets
to recognize. Kevin, would you like — DR. MURPHY: Yeah, thank you. Patrick covered
a lot of it. A couple of things I would
bring in, and it’s not — you know, there’s not
a complete difference between labor markets
and product markets. There’s some overlap. I think many of the
ingredients are similar. But I think the mix is
often quite different. So, for example, in
many product markets, particularly the ones
we tend to study in antitrust, we often
see relatively few alternatives on the buyer
side, that you look at if I stop buying from firm
A, the vast majority of customers go to either B,
C or D, for example, in a market where there’s four
primary sellers of the product. We’re sort of used to
dealing with kind of small numbers,
oligopoly-type models in those markets. Many labor markets, not
all labor markets, don’t look anything like that. If you look at where
people go when they leave a firm or where people
come from when they go to a firm, often
very diffuse. People go many, many
different places. You know, if you look at
employer data and you ask, you know, where do
people go when they leave, often you’ll find no more
than 5 percent of them go to any one firm, that
they go all over the place. And some go in
the same industry. Some go in other
industries. Some change occupations. Some don’t. You look at plant
closings, where people go. Again, not so often a big
concentration of where they go to. If you look at data on
where people are hired from, you see much
the same patterns. That’s kind of a much
more diffuse nature. Now, that’s going to
depend on the geography. In a small town, that
probably isn’t as true as it is in a big city. But the kind of boundaries
of where workers’ alternatives are and the
commonalities of those alternatives across
workers are much less it tends to be than for
customers in a product market. The second part that’s a
little bit different as well is that virtually
all labor contracts are negotiated. And you might say, well,
wait a minute, what about the case where
people place a wage. But there are very few
contracts where you just post a wage and hire
anybody who wants to show up at that wage. Almost any contract,
whether it’s a contract where we negotiate
individually the terms of the deal or one where I
post a wage rate are ones where the employer makes
the decision whether you get hired or not. And because of that, the
effects on me of various labor market practices
might be quite different for different workers. That is, it’s not
sufficient to ask the question what happened
to the wage rate. You also ask did I get the
job because if the wage rate’s higher but you
don’t get the job, you didn’t benefit very
much from that. Now, it might
affect other wages. You’ve got to think
about that too. But there are two
aspects, even in a very commoditized labor market:
what’s the pay, whether I got the job, whether I
got a promotion, more opportunities for some
workers often translate into less opportunities
for others, so things you have to worry
about there. So I think those are
two I think significant differences. A third one that makes
really all input markets more complicated from an
economic standpoint is usually we fall
back on output. What happens to output
in an output market to judge the difference
between efficiency effects and competitive effects? More efficient markets,
we expect output to rise. Less efficient markets,
we expect output to fall. And that’s a good test for
thinking about which way it goes. Unfortunately, in input
markets, that test doesn’t work. That is, efficiency
enhancement can reduce usage of inputs, right? The goal is not to
use more inputs. The goal is to
produce more output. And the effect on
inputs of efficiency is ambiguous. On the one hand, more
output tends to lead to more inputs. On the other hand, more
output per unit input tends to lead
to less inputs. So separating efficiency
explanations from other explanations, not as easy
in input markets because if inputs go down and
maybe even prices of inputs are driven down
because demand for input falls, that doesn’t rule
out efficiency as the driver. So you can’t use the same
test you’d use in output markets. One thing I will say,
and this is true in some output markets, is
frictions are a big part of the labor market. They have to be. It’s individually
transacted, right? It’s not generally
done on a big scale. Each person is supplying
your own labor. They don’t search
jobs for other people generally. There are intermediaries
that help you search for jobs. So frictions are
an important part. But in a world where
there are frictions and antitrust aren’t the same
thing, you always have to remember individual
firm-level elasticities are something
interesting to measure. But they don’t tell
about wage suppression. So for example, if we were
to look at the grocery business, we’d say in a
grocery business, the typical grocery store
faces an elasticity maybe of about five. We wouldn’t conclude from
that that groceries are priced 25 percent above
the competitive level because grocery stores
face an elasticity of about five. So be careful what
you make out of the measurements that we get. The basic issue is you
can move back up, in this case, the labor demand
curve and therefore the gap between marginal
product and wage doesn’t measure the effect
on the wage rate. I’ll stop there. DR. DRENNAN:
Thank you, Kevin. Orley, what are some
thoughts you have on what we should know about
labor markets as we think through these problems? DR. ASHENFELTER: Oh,
you do have to push the button. Now you can maybe
hear me now. Let me say a few
things about that. It’s a pleasure to be
here, and especially because I think, at least
for many decades, I’ve always thought there were
missed opportunities in the use of and
implementation of antitrust remedies in the
context of labor markets. Historically the most
obvious cases have been in the healthcare
business. But there have been
others as well. And almost everyone has
been very reluctant to make this connection. So it’s really a
great pleasure. I consider it a victory
lap to see that someone is actually using
antitrust cases. I’ve testified in a few. The first point I’d like
to make is I think, with respect to the market
definition issue, it makes a huge difference whether
what you’re doing. So Dean Harvey has made
the case that for many antitrust applications,
what we’re looking at are direct effects, the
no-poaching agreements in Silicon Valley were
a result ultimately, originally of the origin
of that is because Steve Jobs and George Lucas
made a deal where Jobs bought the computer
division of Lucas’ film industry company at a time
when Lucas had bought the New York Institute of
Technology computer science department. And notice the
way I said that. He bought it. That’s the way he
thought about it. Moved it to Marin County
and the idea was they would make
animated pictures. But they couldn’t actually
produce any animated pictures at that time
because Moore’s law, the famous example of how
computer power increases at an exponential rate,
just hadn’t quite caught up to the software
they were using. And it was during a
period when Jobs wasn’t at Apple. He had been kicked out. The deal they made was he
would buy this animation production facility from
Lucas and they wouldn’t hire from each other
and that was the deal. That’s what led to the
other — I think Jobs realized this
worked out so well. It made a lot of sense
to make this arrangement with Google and Adobe
and everybody else. Well, and by the way, in
case what you don’t know, what he bought turned
out to be Pixar. I’m sure everyone, if you
have either children or grandchildren, you’ve
seen Toy Story. That was their first
movie, and it’s gone on forever since. Now of course it’s owned
by Disney, which raises a whole separate question
because there was actually a lawsuit
brought over the issue of animators and Lucasfilm
and Pixar and others were defendants. But of course Disney owns
all those companies now. So there’s no — there
isn’t going to be any more lawsuit about that. They can do whatever
they want with their animators. They’ve all merged up. So I think the point I
want to make is that whether you think
carefully about market definition depends. The importance of that
depends on whether you’re looking at direct
violation. There was no need for a
market definition in the case of the no-poaching
because they were basically negative
agreements. And even though it’s
no doubt the case that animators or software
engineers don’t have a million different places
they can work, it’s not really relevant for the
fact that there was a set of negative agreements. So I think a lot of
what Dean’s point about there’s no need for,
or irrelevant really, relevancy of markets,
it’s very similar to the example in the case of
price-fixing in the product market. You don’t need to know. As long as two firms are
engaged — you may have to know something about
it for collecting — figuring out how to
collect the money you’re going to collect
off the damages. But you don’t need to know
about what that market was to do that. Now, there’s a second
issue however. And this is a much more
interesting in some ways and les developed. And I don’t take any
credit for this. And that’s the issue of
mergers and what effect they have on
labor markets. Actually, Dan Hosken,
who’s trained as a labor economist, is in this
room, a friend of mine. We wrote several papers
doing ex-post evaluations. This is now a
business, I guess. EU has the whole — an
entire office that just does ex-post
evaluations of mergers. But we never thought to
actually ask whether any of these mergers
affected wages. (Laughter.) And we’re
labor economists. So that goes to show you
that it’s easy to forget. I regret that and I don’t
know if it would have shown anything. But it would have been
interesting to try. And at the time, it just
didn’t occur, to don’t think, to either of us
to think carefully about that. We were more interested
in what happened to prices. So, and there, I think, on
the merger side, this is very similar, measuring
demand elasticities and crossflow elasticity is
the critical issue for a merger analysis or any
kind of merger simulation that’s going to involve
something like that. And that’s similar, I
think, in the case of the labor market. So market definition now
becomes a pretty critical issue. And as Kevin said, a lot
of it has — the one way — the easiest way to
think about it is where do these workers — where
do they move to, to and from. You could think of there’s
easy cases like, you know, maybe a commuting
zone that has two or three major hospitals and
they’re thinking about merging. It’s almost like the
elasticity, the commuting zone elasticity of supply
of registered nurses, say, is not going to be — is
not going to be very high. So that could easily lead
to some suppression of wages. But you’d need to figure
out, it would matter how much competition there was
in that area and how much people moved across from
one part of the firm to the other. And that is not so
different, I don’t think, from exactly the
same problem. We get into the same
problem in the product market measuring cross
price elasticities. We should never — I used
to have a red light in our — we had all of our
computers were in one room at one time. And we had a red light. And if you were measuring
elasticities, you had to turn the red light on
because it was dangerous. It was a very
dangerous thing to do. So that’s a — now, I
would say one further thing about that. So that’s very
similar, I think. And the burden’s I don’t
think are any greater. There’s one little extra
piece to this though that applies, I think. It hasn’t been mentioned
yet, and it’s not very well researched. And I’m not sure
anybody can study it, which applies to both the
merger issue as well as the naked no-poaching or
product wage-fixing cases. And that’s the issue of
third parties who collect information and
disseminate information on wage rates across firms. The BLS has a long history
of doing industry wage surveys. They always have been
historically considered pro-competitive, so you
could figure out what’s going on. But almost everybody,
including — I mean, I have seen some very
privileged information at some universities where
it’s pretty obvious that there is a — the
administrators are taking a look at the salaries of
economists, observing how much higher they are than
many other people in their universities, but
observing also that that’s true in other places. It’s quite very detailed
information, a lot of times being exchanged
in a little unclear way. You know, some third
party collects information from a bunch of firms and
then gives you back some information. Supposedly you can’t tell
exactly what the other places look like. I don’t know. It’s very easy to see how
that could be misused. And I don’t know that
it’s possible to — I don’t know how
you find that out. I don’t know how often —
it’s my impression that those third-party data
collection enterprises exist at very low levels. Wage rates are collected
and shared at for workers that are not at a
very high level. They certainly are
compared at universities where people are
at very high level. So that’s an issue. It’s come up in the
guidelines as to what can or can’t be shared. But it’s one of those
areas where you can see something related
to it in litigation. But otherwise, you’re
never going to see much because no one’s ever
going to tell you what exactly they’re sharing. So the guidelines for the
HR people, they may just be guidelines. You don’t really know
whether any of that’s actually being
implemented. That’s always been a
disturbing feature for me. Now, there’s one last
thing I just want to say and it’s related to
Kevin’s point about the differences between
product and labor markets. When you think about
fixing prices in the product market, you can
think about the concern you have is that you
don’t want to lose demand. So you don’t want
to lose sales. So the concern a firm has
is that, you know, you may get engaged in some
kind of price-fixing. And then, there’s an
issue about how much demand you’re
going to lose. And that’s where this
cross-substitute building matters so much. But the question is
what’s the fixation on. Is the fixation on
determining the quantities, because you
can do it that way. You can agree on
quantities sold, and that’ll just
determine the price. Or is it on prices, and
then trying to fiddle around with a void,
quantity declines. Now, the same thing is
true in labor markets. Much of what people try
to — I think much of the noncompetitive behavior
comes with issues associated with turnover. So the goal typically is,
at a given wage rate, you’re going to have
a certain turnover. What you’d like to do is
keep that wage rate the same perhaps and
reduce turnover. Noncompetitive
behavior can do that. But so, in a way, it’s the
relationship — you need to really know a little
bit about what the turnover is before you can
say how much the wage is or isn’t suppressed. And I think it might be
the case in labor markets that the fixation is
probably greater on reducing turnover than it
is on the suppression of wages. So right now, for
example, in the U.S. economy, or certainly in
the town I live in, you can walk down the street. In fact, I just passed
through an airport where there’s actually a posting
of the wage rate at the McDonald’s. That’s how hard it was for
them to — this was in California, hard for
them to get people. They’d actually
posted the wage rate. I’d never seen
that before. Pretty high too, more than
the minimum wage there. And so clearly there I
don’t think the wage is probably fixed. But then, the issue is how
can they reduce turnover. So when you think about an
investigation, you need to think a little bit
more carefully about the quantity side, not just
the price side because things that restrict the
quantity, I mean, this is true in the product market
side too, but may be more important in the
labor market side. DR. DRENNAN:
Thank you, Orley. Patrick, you talked
before about sort of the guidelines approach and
how it applies generally to an investigation. How about specifically now
focusing on the market definition aspect of it? Can you just touch
briefly on thoughts about what questions arise? MR. GREENLEE: Yeah, sure. So this would echo a
little bit about what we heard in the economics
presentation initially and it’s perhaps going to
preview some disagreement or different views here
that we have on the panel. But in terms of market
definition, when we’re analyzing, say, a merger,
the idea is trying to identify which collection
of competitors, whether it’s people selling
products in an output market or employers or,
you know, purchasing inputs, which collection
of economic agents compete against one another. And so, the hypothetical
monopolist test or the hypothetical monopsonist
test is basically just saying, hey, we want to
make sure that our focus is broad enough that it
includes all competitors that can have a
disciplining effect on potential exercises
of market power. So, you know, on the sell
side, the question is if I have a candidate
product market, would a hypothetical monopolist
that controlled all of those, so eliminated the
competition between those products, would they be
able to significantly increase the price that
they sold the product to. If it turns out that, no,
they wouldn’t be able to because there would be a
sufficient diversion to a product that was not in
your candidate market, then the analyst would
have to expand to include other products that make
the candidate product market larger. So the goal here is to
make sure that our focus is not too narrow, that
we’re including all sort of significant competitors
that can have an effect on competition so that it’s
not the case that we’re looking at merger and we
think, oh look, this is a two-to-one, when in fact
there are several other competitors out there
that, even after the merger takes place, would
be able to help keep prices in check. So I described it in
terms of a hypothetical monopolist on the
output market. The similar issue on
hypothetical monopsonist would be just the
mirror image of that. Would a collection of
buyers, you know, if there was a hypothetical
merger between them, they acted as a monopsonist,
would they be able to suppress the price they
paid for the input by a small but significant
non-transitory amount, so SSNRP, instead of SSNIP,
for reductions in the price paid versus
increases in the price demanded. So that’s the sort
of market definition exercise that we typically
perform, especially in merger contexts. But understand that Dean
perhaps has a different view with respect, at
least in the non- merger context, of whether
this applies. MR. HARVEY: So let me
ask another question. For the employment lawyers
or labor lawyers in the room, who here has ever
in their careers, in litigating those cases,
have ever defined a relevant market? Oh, one person, okay. That’s interesting. I’d love to hear more
about that, by the way. But, so no one else raised
their hand and people generally laughed because
I think the question strikes us as absurd. And I think there’s a
profound truth in that. Why does it strike
us as absurd? I think there’s a very
good reason for it. So let me give
you an example. Suppose, in a
discrimination case, we learn that a firm, as a
matter of policy, pays its female employees 10
percent less, just an internal rule. And we learn about this. Okay, an antitrust lawyer
would say, oh, let’s define a relevant market. Let’s figure out if that
employer controls a relevant share of the
geographic market skills at issue that
the women have. And if that employer
doesn’t have, say, a 50 percent share, then we
say, all right, well, I guess that discriminatory
policy, while it exists, couldn’t possibly
have an effect, right? That’s insane. And so, and why
is that insane? I think what I would
submit is that every employer has a certain
amount of market power over its employees. The only question
is how much. And so, for — when these
cases have been litigated in the antitrust context,
they’re perhaps all — I don’t know if I can think
of a single example to the contrary — are examples
where there’s a certain act that has occurred
that the plaintiffs are challenging, a no- poach
agreement, a wage-fixing agreement. There, as Orley indicated,
I think the appropriate way is, one, not to
concoct some market definition step that only
exists because we call ourselves antitrust
lawyers and it has nothing to do with what
we’re looking at. You look at the standard
array of econometric tools in the toolbox. And if you have a before
and after, you try to control for as much as you
can and you try to figure out the impact of the
misconduct on the pay. And that’s the claim,
if you are a private plaintiff, or if you’re a
government looking at it, that’s the way you try
to assess the harm. I do agree that, in the
merger context, you do have to do some version
of market analysis. And I think to do the
kinds of research that we heard about this morning,
you have to figure out what kind of markets
are you talking about. You have to have
some to do that. And that research
is very valuable. But I think when we’re
assessing certain acts that we deem to be
anticompetitive, I don’t think they
have any place. And I think actually
that they’re very counterproductive. We could conclude
erroneously, such as my example of gender
discrimination, that there is nothing to see here,
when in fact there very likely is. Some other differences
that I think are important between the
kinds of product cases that antitrust lawyers
are comfortable with and labor cases, that I would
just say in addition to what’s been described,
one is that, as Patrick and others said, it’s a
matching market, which is really interesting where
both sides are trying to figure out all kinds of
characteristics about the other to decide whether
they want to engage in this long-term
relationship where someone is going to be in the
office down from you and you’re going to work with
that person for a long time. I know there’s a
sort of cliché. When lawyers are applying
to law firms, the question is do you want to work
with this person at 2 a.m. on some brief. You don’t ask these kinds
of questions when you’re deciding what brand
of soap to buy in the grocery store, right? They’re totally
different situations. And that bond is akin to
a family relationship. It’s an emotional bond. And there have been
studies about this where, for example, people have
been asked to rank the most important events in
their lives, the most disruptive. Right up there with a
death in their family or the death of a family
member and a divorce is getting fired
from your job. It’s a major event in
your life in a way that, you know, not being able
to buy a product just doesn’t compare. One other aspect that I
think is very interesting is that the value of
an employee in the marketplace can go down
the more the employee switches. So if an employee say only
works for an employer for a month and then switches
again and switches again, and you’re an employer and
you’re looking at this resume where you see the
person can’t hold a job for longer than a month,
very unlikely you’re going to hire that person. There’s no equivalent
like that in the product space. You’re not penalized as
a customer if you switch products. And it’s one of the
sources of friction that Kevin described. With respect to, you know,
what do we compare it to, to try to figure out if
there’s some misuse of market power, you’re
comparing current pay rates against some
competitive pay rate. And the perfectly
competitive pay rate would be where the wage equals
the marginal benefit of that employee. And the employer isn’t
pocketing some surplus of that and the employee
isn’t either. That I don’t think is
a fair description of probably any
employer/employee relationship. Perhaps at the start, but
then think about this. This is a long-term
relationship generally. So when you first decide
where to work or you first hire an employee, that
pay rate is generally set in a fairly
competitive way. You’re telling that person
what the pay rate is and they may make comparisons
and they decide to start working. Once that bond has been
created though, over time, the employer is
investing training and other activities
in the employee. The employee becomes
more knowledgeable, more valuable, more productive
over time and does the pay for that employee
match lockstep with that increase in value. I would say probably
almost never. The employer is absorbing
some of that over time. And, for example, I think
Orley said that, you know, if a certain kind
of misconduct decreases turnover, the employee
has fewer places to go. As that employee’s value
goes up, there’s less and less incentive to have
the employer increase the pay rate to match it. Another aspect of it
is in the price-setting mechanism. In the product space,
prices are changing all the time. They can often be
very discriminatory. You know, for example, for
those of us who flew here, you’re sitting on a plane
and you paid a certain price for your
plane ticket. The person sitting next to
you probably paid a very different price. And it’s a function of
this very complicated way that airlines try
to extract the most willingness to pay out
of every passenger. In the employee space,
you set a wage rate or a salary and that is very
sticky in a couple of different ways. One, it’s typically
a one-way ratchet. It’s very unusual for your
pay to actually go down in an absolute sense. It may not keep
up with inflation. But very rarely do you
have a 5 percent pay cut. So it’s a one-way ratchet. Secondly, your pay is not
set by in large through an individual negotiation,
in the vast majority of cases. Say, in the classic
example, the civil service pay scale in the
government, there’s no negotiation there. You apply for a certain
job and maybe there is in terms of what job
you slide into. But once you have that
title, there’s just a set pay rate and it’s fixed. And in terms of how those
pay rates are changed, typically it’s changed
once a year as part of a firm-wide, very complex
budgetary procedure where say there’s a merit
increase of say 1.5 percent. All kinds of factors
go into that. The CEO is involved, HR
professionals and they decide for employees in
good standing, everyone will get a raise of
at least 1.5 percent. Mayne there’s some
variance for performance, but — and that also
is very unusual. One big category that has
not come up today, and I think if there’s a blind
spot in the way people are looking at these
issues, I think it’s this. It’s that people
examining this space tend to ignore the dynamics
within the firm. They look only at the
dynamics outside the firm and ask, well, if
competition goes up, what effect does that have on
the pay within the firm. What we’ve seen in all the
cases we’ve litigated is that a large company does
not look only to external competition in
setting pay. There’s a pay structure
within the firm that is — one of the main goals
is to maintain morale, decrease turnover and
keep people happy within the firm because people
make comparisons primarily with other
people within the firm. And again, one of the
great things about this is that we all have our
own personal experience. It’s not — you know, as
fun as it is studying rocket markets, we’re not
all rocket scientists. But we’re all employees. So we all can really turn
on personal experience. Just ask yourself what
comparison means more to you, if you know of
someone who does your job at a different employer
who gets a raise and you don’t versus someone at
your own employer who’s just like you who gets
a raise and you don’t. Which one
bothers more? It’s the within-firm
comparisons. And firms know this. There’s a great deal of
research about this. And HR professionals get
graduate degrees in how to set up pay structures
within the firm. And so, that is important,
particularly in private cases, and also I would
say one of the slides that I saw this morning
made me think of this, in the analysis of hospital
mergers where there’s an analysis breaking up
different kinds of employees by skills. And I would love to see an
analysis of, within the same hospital, for
hospitals that have the greatest impact on the
highly skilled because of the within-firm dynamics,
is that also suppressing pay of the
lowest skilled. And those are I think very
interesting questions that I would love
people to look at. DR. DRENNAN: Thank you. Go ahead, Kevin. DR. MURPHY: I think your
study kind of did that. I mean, you don’t find
any effect on the — DR. DRENNAN: Microphone? DR. MURPHY: You didn’t —
I think the charts didn’t show any effect on the
least skilled groups in those pictures. You had effects on the
most skilled, but very little effects. A couple of things. One is, again, related
what I said before about own elasticities, I mean,
for more competitive effects, what
we care about is cross-elasticities,
whether it’s a merger or indeed even if it’s an
agreement of some type. Whether it’s legal or not,
maybe you don’t need to measure anything for
legality purposes because it’s per se or whatever. But for impact, you really
probably do care about cross-elasticity. What fraction of the
outside alternatives are being affected and how
important are those relative to other
alternatives? And that’s one of the
things that market definition helps with
because while it usually doesn’t answer things at
the end of the day, it gives you some
information about what those cross-elasticities
are likely to be and therefore the impacts are
likely to be which is I think exactly why when
they studied the nurses’ case, they had a prior
that it was likely to have much less of an
effect on one group of employees than
on another. So that’s one — that’s
one aspect that I’d like to pick up on. In terms of, you
know, well, if we see discrimination, then it
can’t be that firms are competitive, to me that
just does not tell you one way or another. Lack of discrimination
wouldn’t tell us the firms are not competitive. Go back to Gary
Becker’s 1950s work. We know that you could
have discrimination affect wages. And the key is, you know,
you see wage differentials by race or by gender
within a firm and you say, well geez, competition
would eliminate that. Yeah, that’d be great in
a world in which outside the firm blacks and whites
were paid equally and outside the firm men and
women were paid equally. But that’s not the
world we live in. In fact, the differentials
in pay between race and gender are bigger in the
economy as a whole than they are within
your typical firm. That is, the typical
within-firm differentials are smaller than the
aggregate differentials that you see. And therefore, I don’t
think it tells you one way or the other whether
there’s competition or not, that you see
differentials in pay between men and women or
blacks and whites within a firm. That’s the fact they’re
represented in the market as a whole. In terms of wage
structures, and I agree there are wage structures
out there in the marketplace and they’re
certainly a part of human resource practice. Some of the methodologies
people have used to try to prove that there are
wage structures, for example, regressing
the wage change for individuals on wage
changes for the average simply are not
informative for that question. Individual wages are going
to move with average wages, I don’t care, in
any system that makes a lot of sense. And people who purport
that a regression of individual outcomes on
average outcomes tells us that there’s this
transmission mechanism and all this other stuff,
they’re just playing games with statistics, to
tell you the truth. There’s nothing in valid
statistical methodology would tell that that
inference is valid. So I would be very careful
about those things. In terms of Orley’s point,
I think turnover is part of what firms think
about all the time. And turnover’s costly,
for the reasons I said before. Turnover’s costly
to workers. Turnover’s
costly to firms. And in fact, you were
talking about people being very competitive
when they start. We know that, for example,
for most senior workers, it’s not like firms —
both firms and workers are invested in
the relationship. We know, for example, in
a plant closing, older workers lose a lot. They’re doing much better
than their next best outside alternative. One o established
things in labor economics is that
if you have an unexpected closing of a plant or a
firm disruption, guys with tenure in those firms
are going to lose a substantial amount. They’re earning much more
than their next best outside alternative. That’s a reality
of labor markets. I’ll stop there. DR. DRENNAN: Orley,
your thought? DR. ASHENFELTER: I don’t
have much to add to that really. I mean, much of the labor
market literature, as Kevin is alluding to here,
is about the fact that much labor doesn’t operate
in a simple spot market, which means that it isn’t
as if the value of a worker’s marginal product
is always equal to their wage. It could be higher at some
points and lower at other points in some kind of a
lifecycle context where there’s a long-term
relationship between the firm and the employee. What Kevin’s referring to
is the kind of classic analysis of pay in a firm
typically relates — I’ve seen these regressions in
virtually every country in the world. It’s one of the most
well-documented things that we have. There’ll be an effective
schooling or some kind of credentials, an effective
how much total labor market experience you have
and then, in most jobs, but not all, time spent
with the firm that you’re currently at will
also have an effect. That’s often associated
with specific human capital, something that
the employee gets some part of in order to
reduce turnover. So there is — I’d like
to comment on something related to this though
which is that — an interesting question to
me is I really — I don’t know who wrote
Delrahim’s speech, but congratulations, whoever
that was, especially to pick up on the Adam Smith
reference, which is one of my absolute favorites. He just read a
short part of it. It goes on very several
pages in Smith. In fact, there is no long
discussion of collusion amongst firms in the
product market, in the entirety of Smith. His main discussion of
collusion is amongst employers. And this tacit notion of a
tacit understanding about not raising wage rates. It sounds very
familiar, right? We are observing that
there’s excess demand everywhere in the economy
and very little upward pressure on wage rates. But I say that because it
does raise a question. There is, at least in my
opinion, there seems to be much greater interest,
at least at the public level, as, for example, in
a government agency, not so much on the
academic side. So you’ve heard a couple
of people speak about academic studies of a
failure of competition in labor markets. There’s not very
much of that. It hasn’t been
very successful. How do we know that? Well, what do the MIT and
Harvard graduate students write for their
dissertations? That’s what all the
economists are going to be in the next — or Chicago
or Princeton, whatever. We actually have a
student working on this. But it’s very rare. So the public interest
is much greater. And I think there’s a
sense maybe that there’s something more going on
here that’s disturbing to people. A couple of things
that I’ve noticed, for example, one of the
ancient, ancient, ancient things from my days as
an economist was that aggregate productivity
growth in the economy typically was equal
to real wage growth. That relationship
has been broken. It’s not true anymore. That happens to be
identical to the statement that labor’s share of
output is declining because they’re more or
less the same, not exactly the same because it turns
out there’s more to it than that. But it’s a little bit
surprising I think to a lot of people. It’s not clear, by the
way, whether it’s — I say labor share
is declining. It doesn’t mean capital
share is increasing. It could just be
margins are going up. Margins going up is one
of the natural things associated with mergers. And remember there’s
no way to distinguish between that. Labor’s share going down,
there are two ways that can happen. Wages can go down compared
to marginal products or prices can go up compared
to marginal values. So that’s not identified. The failure of
competition can be on either side of
that market. But it seems to me that
there’s more — for some reasons that I can’t
quite understand, there’s considerably more interest
in this in the general public. I don’t know. Maybe there are some
examples or maybe people are aware of failures of
competition that most of us don’t get to see very
much or maybe the mergers are now at such a scale
that people, it’s suddenly dawned on them that maybe
there is going to be an effect in the
labor markets too. But it was a comment I
couldn’t resist because it’s relevant to what
Kevin was saying about many aspects of the
characteristics of labor markets that are unusual. DR. DRENNAN: Thank you. Go ahead, Kevin, please. DR. MURPHY: I just want to
follow up on what Orley said. I mean, labor’s share
can change for lots of reasons, some of which
would be related to competition in either the
input or output markets. But there are many other
things that would change labor’s share. So it’s not limited. You don’t want to think
change in labor share tells us immediately
there’s competition effects on either side. Obviously changes in
relative factor ratios and things like that like
would also potentially affect labor share
depending on the degree of substitutability
and the like. So you have to think
about that as well. DR. DRENNAN: Thank you. Dean, anything, early
indications to take away from the early litigated
cases about the implications for how
this is being viewed? MR. HARVEY: I think in
terms of the litigated cases, there are a
couple of lessons. One is what kind of
classes are courts certifying. So, for example, in the
no-poach space, the employers agree with
their competitors not to recruit each other’s
employees, period. It’s not as to a specific
category of employees. It’s all employees. And so, we’ve tried to —
when we’ve represented plaintiffs in these
cases, tried to figure out who’s injured. And, for example, in the
high-tech employees case, there you had certified
class of over 64,000 workers across
the country. It was a national class
that included everyone from a storyboard artist
at Pixar to a hardware engineer at Intel, all
part of the same class. Now, the court there was
not thinking of these employees as substitutes
for each other, you know, which is what you would
typically do in a product case. The class would consist of
customers of a particular product and you would
do the usual market definition product
substitute stuff. So that is just
happening in the cases, and I think for
very good reasons. The other takeaway, which
is kind of a corollary to that, is that courts
do take a good look at what’s going on within the
firm to try to figure out if pay is suppressed for
one group within the firm, to what extent is
that pay suppression translated into harm
for others at the firm. And I think, you know,
those are the two takeaways. DR. DRENNAN:
Okay, thank you. I know as an antitrust
practitioner, we often look to the academic
literature to give us insights about where to
look and what information to get and in instances
when it’s hard, difficult to measure a key
parameter, elasticity, et cetera, in the context of
an investigation, what benchmark values might
be out there that have already been produced. Kevin, do you have any
— sort of two parts. Do you have any thoughts
that are important for us to keep in mind as we
look to the growing academic literature
in this space and any recommendations for this
would be a great next thing to be addressed
in the literature going forward relevant
to these issues? DR. MURPHY: No, I think
one thing is you want to kind of look at the
academic literature and ask is it asking the
question that I need to answer here and is it
helping me with guidance because, as I said
before, there’s a wide range of labor markets. Labor markets, some
involve, you know, very specialized workers with a
small number of employers and many other labor
markets have very broad employee bases. So you don’t want to
generalize too far from what you saw in one
case to the other. Also I think the academic
literature needs to be more careful in terms
of, like I said before, differences between
effects on wages versus establishing differences
between, say, you know, the elasticity of supply
or something like that. Those two are not
synonymous, even though they are in
simple models. Another recommendation for
the academic literature is I think it’s kind of
odd to think of a static labor supply curve in most
of these marketplaces that, you know, you don’t
get the same number of applications if you’re a
big employer as you do if you’re a small employer. That means long-run supply
elasticities are likely quite different than
short-run supply elasticities. And when that’s the case,
even the short-run can’t be analyzed using the
short-run elasticity because a hiring decision
today affects supplies in the future as well. Another thing I would
say in the academic literature, be careful
about using things like vacancies or applications
as measures of quantities. They’re not. They’re measures of
propensities to transact at the market equilibrium. You know, it’s a measure
of how much slippage there is in some sense in
getting to that market equilibrium. They don’t really measure
supply and demand. They don’t really
measure the fundamental determinants of prices. Just like you wouldn’t say
if I have the only house for sale in my
neighborhood, I’ve got a monopoly on selling
houses in my neighborhood, you know? It’s like, no, that’s not
how prices for houses in my neighborhood are
going to be determined. When we go from three
houses for sale to one house for sale, it’s
not going to be a big difference in the way
that marketplace operates. Again, I think, you know,
so as the marketplace changes, you could see
vacancies move one way or the other. That reflects things other
than what we would think of as the principles
of supply and demand. That’s really what I have. So as much of a
cautionary tale for the literature as it is
for the practitioners. DR. DRENNAN: Before we go
on to Orley’s thoughts — DR. ASHENFELTER: Yeah, let
me make a comment about that. DR. DRENNAN: Go ahead. DR. ASHENFELTER: That’s
a very interesting question, a practical
question really of where do you look for potential
anticompetitive behavior that’s actionable. I don’t really have a
good answer to that. Many of the examples, many
of the applications we know about seem to be kind
of hit or miss. Let’s take the high-tech case. The way that really got
started is because I don’t know who did this, and
it’s never come out. Someone complained to
Department of Justice or explained there were no-
poaching agreements. We know that because the
same day that the case was filed, there was a
settlement filed, which I thought was, at the time,
extremely annoying because we didn’t learn anything,
like a two-page piece of information. Somebody was really
screwing around with no- poaching agreements. We don’t know how they
learned about it. We don’t know
where it came from. And now we say we’re not
going to do it anymore, end of story. Now luckily Lieff
Cabraser, here’s the man with the big bucks. Lieff Cabraser showed up
and a lot of discovery occurred after that. There’s a very, very nice
set of articles that I’ve collected in the Bloomberg
daily labor report that actually has kind of a
nice evolution of the thing as it came
about showing you the interactions amongst the
different firms and who interacted one
to the other. But I don’t know, I don’t
know where that — I don’t even know how that
case got started, right? It was like some anonymous
tip, I guess, that came here to this building,
which no one’s ever going to tell us about. DR. DRENNAN: Hopefully. (La DR. ASHENFELTER:
And even — (Laughter.)
Hopefully. I guess once you’re
retired, you wouldn’t think about it. And the same is true
sometimes in the product market cases, that
you don’t know. There’s something going
on and the first reaction you have is it’s shocking
that that’s actually going on, such
explicit collusion. It’s very difficult to —
I don’t know how — the hotline I guess is the
only thing I can think of for these extreme
kinds of cases. I don’t know that there’s
any very easy — I do think that one of the few
things, if you could find a way to follow
it, would be the information-sharing. I think it’s very
tempting, once you’ve engaged in some kind of
collusive agreement, to try to keep track on how,
what other people are doing. And really in order to do
that, you need some way to find out what the
information is on the other places. My favorite example is
it’s very well known. I don’t know how you’d
find out the answer to this. In the Ivy League schools,
I don’t know if you’re in this crowd. Chicago is kind
of a maverick. They’re like a tough,
don’t play ball with the people in the
tweed coats. But it’s kind of known
that there’s a piling up at one salary in most
of the Ivy League universities. There’s like a lot of
people make the same amount at a certain level. It’s not a secret. It’s not public. But it’s not a secret. DR. DRENNAN: Not anymore. DR. ASHENFELTER: Well, I
mean, it’s not — it was a — you know, it’s a
secret like if you have a trade secret and you drop
it off on the floor here andi pick it
up, it’s mine. It’s nothing to say about
— that’s how I learned about it. Now, the interesting
question is — and I’d love to know the
answer to this. There no way of finding
out — is so there’s a piling up I think in each
of these universities. There’s a pile-up
at the same spot. I don’t know how you
would ever find that out. I’m sure there’s a spot at
Penn, for example, where there’s a piling up level. And in fact it’s called
a most favored nation clause, right? You get a contract as an
employee that your salary will be as high as
anybody else’s in the institution. Kevin probably
got one of these. I don’t have
one of these. Chicago might not
do this, by the way. That’s a very
competitive place. It may not do that
kind of stuff. There’s up to no
limit I guess, right? There is no maximum. I only mention that as an
example because that’s one example people can
know about, right? But I know about it
as an accident, right? I can tell you about it. But it’s not going
to do any good. And the interesting
question is whether there are other situations
like that. We just don’t have any
way of finding out. So the cases, I notice
when they come up, they’re a lot of just
accidents associated with where you find them. And it’s not clear to me
there’s any easy way to say, you know, what would
be where you’d see the most concern. Lots of times I think
maybe something comes up as a result of another
investigation. In other words, there’s
something going on in some other area and then
like ancillary to that you hear about some agreement
that was something else. But all of that is really
hit or miss. And that’s why I think the effort of
enforcement, we need to make sure the enforcement
margin — I mean, what you do, you don’t have
to have that many cases, right? The HR people are going to
learn from the fact that there are other cases. They don’t want to go to
jail or they don’t want to have their companies pay
a ton of money, when it’s their fault anyway. So I think as long as
the enforcement level is sufficient, presumably it
creates some competitive efforts to — you know,
people drop out of these agreements saying, oh, I
don’t like the sound of that. And hopefully if there’s
enough enforcement, you’d get most people
to say that. DR. DRENNAN:
Okay, thank you. If none of the panelists
has any final thoughts — okay, Kevin? DR. MURPHY: Yeah. I mean, information-
sharing is always an issue. It’s an issue in product
markets, not just in labor markets. In fact, it’s not
that different there. You know, that is
information on what other firms are doing can be
useful for enforcing a cartel. The sad part is it’s also
useful even when you’re not in a cartel. A lot of farmers wake up
in the morning and watch the farm report and that
could be true even when they’re not colluding. They’re interested in
learning what’s going on in the labor market or
in the markets they’re involved in. So that’s what makes
those cases tough. There’s usually a reason
for information-sharing, at least ex-post
and the like. It’s valuable for
planning and other purposes. That doesn’t mean
it can’t be used for anticompetitive
purposes too. But that’s why these guys
get paid the big bucks for the tough job, you know? (Laughter.) DR. DRENNAN:
Okay, I feel a lot better about my pay now. Okay. Well, I want to thank this
great panel for a very interesting discussion and
for all the preparation that went into it. Thank you.

Leave a Reply

Your email address will not be published. Required fields are marked *