Popular in Social Sciences

Banks and Policy Organisations

Macroeconomic modelling of CBDC: a critical review

European Central Bank Ulrich Bindseil

Social Identities and Divisions

A new explanation of antisemitism: Jew hatred as a civilisational phenomenon

Boston University Liah Greenfeld

Banks and Policy Organisations

Financing Private Credit

Federal Reserve Bank of New York Nina Boyarchenko

Economics

Failing Banks

Massachusetts Institute of Technology Emil Verner

Culture and Society

The City as a Growth Machine: Toward a Political Economy of Place

New York University Harvey Moltoch

American Politics

Keeping the Republic

Boston College Marc Landy

British Politics

Entrepreneurial thinking: the politics and practice of policy impact

University of Sheffield Matthew Flinders

Interactions with Powerful Female Colleagues Promote Diversity in Hiring

View more insights in Social Sciences - Economics - Labor Economics

View more figures from Occidental College

View the page of Jorgen Harris

Read the Study

Next in Economics

Macro uncertainty, unemployment risk, and consumption dynamics

Joonseok Oh | University of Southampton

Geopolitical Oil Price Risk and Economic Fluctuations

Lutz Kilian | Federal Reserve Bank of Dallas

Jorgen Harris discusses the effect of hearing cases alongside female judicial colleagues on the probability that a federal judge hires a female law clerk.

Image courtesy of interviewee. December 7, 2023

Log-in or Sign-up to Faculti
Currently viewing this subject insight as a guest. You have insight(s) remaining for this month. Login to view 8000+ figures on the platform.

Generating Workspace...

Export as PDF

The study examines whether working with women influences federal appellate judges in hiring female clerks. The context is the federal appellate court system, where judges have significant power and their clerkships are critical for future legal careers. Currently, women represent about 25% of appellate judges but make up around 40% of clerks, despite over half of recent law graduates being female. This video video indicates that women are underrepresented in top legal positions compared to their male counterparts.

Previous research in social psychology suggests that working with diverse individuals can reduce prejudice and change attitudes. The study utilizes a unique setting in appellate courts, where judges are randomly assigned to cases and staff (clerks) to analyze how interactions with female colleagues affect hiring decisions. The findings reveal that increased interactions with women lead to a higher likelihood of hiring female clerks, specifically a 7% increase for each standard deviation rise in such interactions.

However, the effect does not persist over time, suggesting that while exposure to female colleagues positively influences hiring decisions, it does not lead to permanent changes in attitudes. The study implies that fostering workplace diversity at high levels (like judicial appointments) may create a feedback loop, promoting more gender diversity in lower ranks and ultimately benefiting the legal profession.

The broader implications of the research suggest that meaningful interactions among diverse groups can reduce prejudice and improve workplace dynamics, calling for policies that encourage such interactions to combat societal segregation and its negative effects.

Language:

Export as PDF

X-TIMESTAMP-MAP:LOCAL=00:00:00.000,MPEGTS=900000

00:00.375 --> 00:02.505
What we're doing in this study is we're looking at whether

00:02.505 --> 00:05.145
working with women increases the likelihood

00:05.145 --> 00:07.945
that federal appellate judges hire female clerks.

00:07.965 --> 00:10.025
And so the context of this study is the

00:10.025 --> 00:11.185
federal appellate court system.

00:11.855 --> 00:13.305
This is the level right

00:13.305 --> 00:14.985
below the Supreme Court in the United States.

00:15.445 --> 00:17.385
So for most cases, these are the judges

00:17.445 --> 00:21.185
who have the final word on the interpretation of US law.

00:21.255 --> 00:22.865
It's a essentially the elite

00:22.865 --> 00:24.545
of the judicial of the judicial field.

00:24.545 --> 00:27.105
It's a very powerful, very powerful group of people.

00:27.885 --> 00:30.225
The appellate court clerkships, the, the jobs

00:30.225 --> 00:31.785
that we're looking at for, you know,

00:31.785 --> 00:34.305
for entry level positions are also a really important

00:34.625 --> 00:37.665
stepping stone, both to positions in the judiciary itself,

00:38.165 --> 00:41.225
but also to other elite positions within law.

00:41.685 --> 00:44.305
People who become corporate law partners, for example,

00:44.365 --> 00:45.625
are very frequently people

00:45.625 --> 00:47.425
who have had appellate clerkships in the past.

00:48.325 --> 00:50.185
The other thing to mention in terms of context,

00:50.185 --> 00:52.545
because we're looking at gender, this is a context

00:52.615 --> 00:55.585
with a real low level of gender diversity in a way

00:55.585 --> 00:57.985
that's pretty typical of other elite professions.

00:58.485 --> 01:00.985
So about a quarter of federal appellate judges are women.

01:01.315 --> 01:03.985
About 40% of the people who get these clerkships are women.

01:04.525 --> 01:07.585
In contrast, about 50%, a little more than 50% now

01:07.645 --> 01:09.385
of recent law graduates are female.

01:10.045 --> 01:12.825
So, so we see that women are a lot less likely

01:12.825 --> 01:14.825
to get these appellate clerkships than are men.

01:15.665 --> 01:16.985
Interestingly, when we look one level down,

01:16.985 --> 01:19.825
so we look at the district court clerkships, about 60%

01:19.825 --> 01:21.305
of the people getting those jobs are women.

01:21.365 --> 01:22.545
So it doesn't look like women are

01:22.545 --> 01:23.585
less interested in these jobs.

01:23.965 --> 01:25.745
It looks like they're substantially less likely

01:25.805 --> 01:28.065
to get the very top jobs than are their male

01:28.065 --> 01:29.185
colleagues from the same law schools.

01:29.485 --> 01:31.665
The question of whether working with people

01:31.725 --> 01:33.265
or interacting closely with people

01:33.265 --> 01:35.345
who are different from you affects your beliefs

01:35.345 --> 01:38.185
and affects your attitudes is a very old question.

01:38.335 --> 01:39.345
This goes back to work

01:39.345 --> 01:42.025
by a social psychologist in Gordon Allport in the 1950s

01:42.685 --> 01:44.185
who essentially said the best way

01:44.185 --> 01:45.745
to reduce someone's prejudice is

01:45.745 --> 01:48.185
to have them work closely in a cooperative manner

01:48.185 --> 01:49.505
with someone who's different from them.

01:50.375 --> 01:54.025
However, it's a really hard question to get evidence for

01:54.295 --> 01:58.025
because in most contexts, when we see a person working

01:58.025 --> 01:59.385
with someone who's different from them,

01:59.725 --> 02:01.145
that's gonna be that person's choice.

02:01.245 --> 02:02.665
And so who we see you working

02:02.665 --> 02:04.705
with is gonna tell us a lot about what your beliefs

02:04.705 --> 02:06.625
and prejudices were before you work with that person.

02:07.485 --> 02:09.745
As a result of that, the best evidence on this question

02:10.235 --> 02:12.345
tends to come from contexts where someone

02:13.125 --> 02:15.905
is closely interacting with someone different from them out

02:16.085 --> 02:17.545
for reasons other than their own choices.

02:18.405 --> 02:20.825
So, you know, we either do that via experiments

02:20.845 --> 02:22.985
or we do that via what we call natural experiments,

02:22.985 --> 02:26.385
which is context where some institution puts you in contact

02:26.385 --> 02:27.785
with someone else without your, your,

02:28.855 --> 02:30.145
your decisions playing a role.

02:30.765 --> 02:33.665
So I would say my sort of what I think are, are, are some

02:33.665 --> 02:35.545
of the best pieces of research on this question.

02:35.545 --> 02:39.065
There's a couple recent experimental papers looking at

02:39.075 --> 02:42.465
where essentially the researchers made amateur soccer

02:42.465 --> 02:44.945
leagues and assigned people to teams.

02:44.965 --> 02:49.505
Either teams that are, homogeneous in terms of religion

02:49.505 --> 02:51.385
or cast, or teams that are heterogeneous, right?

02:51.385 --> 02:53.705
Where you're work, you're playing with lower cast people,

02:53.705 --> 02:55.425
you're, you're a Christian playing with Muslims,

02:55.725 --> 02:57.385
and what they saw is that being assigned

02:57.385 --> 02:58.825
to teammates who are different from you

02:59.465 --> 03:03.605
Really increased measurable aspects of your,

03:04.045 --> 03:05.085
openness to those people, right?

03:05.085 --> 03:06.165
You're more likely to sit next

03:06.165 --> 03:08.245
to a Muslim at the award ceremony if you had a Muslim

03:08.345 --> 03:09.525
on your team, right?

03:09.525 --> 03:11.445
You're also more likely to express positive

03:11.445 --> 03:12.565
views of Muslims, et cetera.

03:12.905 --> 03:14.445
So those sorts of, you know, those sorts

03:14.445 --> 03:15.925
of experiments can be very powerful.

03:16.505 --> 03:18.885
The big challenge with those is that there's only so much

03:18.885 --> 03:21.125
that you can experimentally vary in someone's life

03:21.125 --> 03:23.565
because doing an experiment like this is really expensive.

03:23.905 --> 03:26.405
And of course, right, you, we, we can't interfere

03:26.405 --> 03:28.565
with people's lives too much for the sake of research.

03:28.985 --> 03:31.165
So the other thing that we do, right, as much

03:31.165 --> 03:33.885
as we would love to as researchers, the other thing

03:33.885 --> 03:36.125
that we tend to do is we look at contexts

03:36.175 --> 03:37.965
where somebody else is essentially

03:37.965 --> 03:39.045
doing an experiment for us.

03:39.585 --> 03:41.365
So the classic example of this is

03:41.365 --> 03:43.965
to look at freshman year roommate assignments in college.

03:44.425 --> 03:46.485
We say, okay, you know, you fill out your housing form,

03:46.485 --> 03:48.685
you send it in, and then they basically pull your name out

03:48.685 --> 03:50.005
of a hat and stick you with a roommate.

03:50.425 --> 03:53.125
And what we can see in that, in, in that body of research is

03:53.125 --> 03:54.285
that being randomly assigned

03:54.285 --> 03:56.165
to a freshman year roommate who's different from you,

03:56.185 --> 03:57.405
for example, if you're a white student,

03:57.405 --> 03:58.685
you're assigned to a black roommate.

03:59.155 --> 04:00.485
That seems to have, again,

04:00.525 --> 04:02.565
a measurable effect on things like who your friends are

04:02.565 --> 04:05.205
and what your stated racial attitudes are, at least

04:05.205 --> 04:07.165
for a little while, at least for that first year of college.

04:07.985 --> 04:11.605
The thing that's been very hard for us to look at is a,

04:11.605 --> 04:12.925
looking at effects of these sorts

04:13.085 --> 04:14.565
of interactions on powerful people,

04:14.675 --> 04:17.645
because powerful people don't tend to get assigned to stuff

04:17.665 --> 04:19.205
by drawing their names out of hats.

04:19.745 --> 04:22.565
Um, it's also hard for us to look at workplace interactions,

04:22.915 --> 04:25.085
both because it's hard for workplace interactions

04:25.085 --> 04:27.285
to generate useful data and also, again,

04:27.285 --> 04:29.165
because these are contexts where,

04:29.165 --> 04:31.805
usually someone's being assigned to a workplace interaction

04:32.385 --> 04:34.365
for specific intentional reasons, right?

04:34.505 --> 04:36.605
Not, not a, as a result of random chance.

04:36.985 --> 04:39.525
So what we realized is the appellate courts have

04:40.185 --> 04:42.125
two properties that make them really perfect

04:42.265 --> 04:43.365
for studying this question.

04:44.105 --> 04:48.045
The first thing is that in order to make the decisions,

04:48.225 --> 04:49.525
as in with the intention

04:49.525 --> 04:51.405
of making the decisions in these cases,

04:51.595 --> 04:53.445
fair judges are assigned

04:53.445 --> 04:56.205
to cases at random conditional on which judges are working

04:56.205 --> 04:57.325
within a particular court.

04:57.785 --> 04:59.045
So I'm in California, which

04:59.045 --> 05:00.405
is represented by the ninth Circuit.

05:01.105 --> 05:02.565
If a case goes to the ninth circuit,

05:02.655 --> 05:05.125
there will be three judges chosen at random out of the 60

05:05.145 --> 05:07.525
or so judges in the ninth circuit to hear that case.

05:08.465 --> 05:09.805
As a consequence of the fact

05:09.805 --> 05:11.605
that the judges work together in groups

05:11.745 --> 05:14.205
and are assigned to cases at random, it means

05:14.205 --> 05:16.085
that judges are assigned to each other at random

05:16.185 --> 05:17.685
as well, if that makes sense.

05:18.145 --> 05:21.445
And so, you know, if I end up spending a lot of time working

05:21.445 --> 05:22.805
with you or not very much time working

05:22.805 --> 05:24.045
with you over the course of a year,

05:24.345 --> 05:27.245
that's gonna come down primarily to luck of the draw.

05:27.665 --> 05:29.285
So that gives us this very useful context

05:29.295 --> 05:31.845
where we see these powerful people being assigned

05:31.845 --> 05:33.845
to cases at random in a similar way

05:33.845 --> 05:35.845
that roommates would be assigned to, to roommates at random.

05:36.425 --> 05:37.925
That's the first really useful thing.

05:38.305 --> 05:41.325
The second really useful thing is that judges have staffs

05:41.385 --> 05:44.765
of law clerks to help 'em research their legal decisions

05:44.825 --> 05:46.165
and write their legal decisions.

05:46.165 --> 05:47.765
These are essentially apprentice judges.

05:48.235 --> 05:49.805
They tend to hire these people out

05:49.805 --> 05:51.725
of the top law schools in the pro in the country.

05:52.405 --> 05:54.285
Normally the top graduates of those top law schools

05:54.905 --> 05:56.565
and judges make a hiring decision

05:57.065 --> 05:59.485
As individuals and they do so almost every year.

06:00.065 --> 06:02.205
So if you get a law clerk, you know, a clerkship,

06:02.205 --> 06:03.765
you're not gonna work for the ninth circuit,

06:03.785 --> 06:05.365
you're gonna work for a specific

06:05.365 --> 06:06.685
judge who decided to hire you.

06:07.105 --> 06:08.285
And as a consequence of that,

06:08.345 --> 06:10.245
we can see the judges make decisions

06:10.515 --> 06:12.645
that reflect something about how they believe women

06:12.645 --> 06:17.325
and men perform in the judiciary on a regular basis, right?

06:17.325 --> 06:18.805
Pretty close to every single year.

06:19.745 --> 06:22.805
The last thing that's really nice here is that the decisions

06:22.805 --> 06:23.885
that judges make are public,

06:24.265 --> 06:25.685
at least on cases that they publish.

06:25.825 --> 06:27.805
And so as a result of this, it's not just

06:27.805 --> 06:29.565
that judges are assigned at random to working

06:29.565 --> 06:31.925
with each other, it's also that we can see a paper trail

06:32.435 --> 06:33.805
generated by those interactions

06:33.805 --> 06:35.365
that tell us who was working with whom.

06:35.745 --> 06:37.965
So what we did with this, what we did

06:37.965 --> 06:40.165
with this is essentially we said we're gonna construct a

06:40.165 --> 06:42.165
measure where we ask what percentage

06:42.165 --> 06:44.605
of your interactions every year were with female colleagues.

06:45.185 --> 06:47.325
And then what we do is we say, we expect that

06:47.325 --> 06:49.605
to be random conditional and which court you're in

06:49.605 --> 06:51.205
and which year, which year we see you in.

06:51.865 --> 06:53.325
So, you know, essentially

06:53.325 --> 06:55.245
what we do is we regress our dependent variable,

06:55.245 --> 06:56.845
which is did you hire at least one woman

06:57.395 --> 06:58.965
onto the percentage unit

06:59.025 --> 07:00.885
of interactions you had with female colleagues?

07:01.185 --> 07:02.485
And what we saw is that working

07:02.485 --> 07:05.085
with female colleagues really substantially increases your

07:05.085 --> 07:06.245
likelihood of hiring a woman.

07:06.705 --> 07:09.205
Um, your likelihood of hiring at least one woman increases

07:09.205 --> 07:11.805
by about seven percentage points if you have a one standard

07:11.805 --> 07:12.925
deviation increase in your

07:12.925 --> 07:14.325
interactions with female colleagues.

07:14.865 --> 07:15.885
So one thing, you know,

07:15.885 --> 07:18.525
so we see this large substantial increase in your likelihood

07:18.525 --> 07:19.925
of hiring a woman in the year

07:19.925 --> 07:22.125
after you interact with female colleagues at a higher rate.

07:22.745 --> 07:24.965
We don't see any evidence of persistence, right?

07:24.965 --> 07:27.285
So the next year you're gonna have a new set of interactions

07:27.305 --> 07:28.685
and you're not gonna have some sort

07:28.685 --> 07:30.165
of permanent shift in your beliefs.

07:30.585 --> 07:32.725
That's not something that that surprises us too much

07:32.725 --> 07:34.565
because these are very routine interactions.

07:34.565 --> 07:36.725
These are interactions that judges have every single year.

07:37.345 --> 07:40.365
So overall, this is I think, a really, promising and,

07:40.365 --> 07:41.925
and surprisingly positive result

07:41.925 --> 07:44.525
because what this suggests is that the type of work

07:44.525 --> 07:45.965
that most of us do, right?

07:45.965 --> 07:48.565
Most of most of us are not assigned to ju

07:48.565 --> 07:49.965
to colleagues at random in this way,

07:50.425 --> 07:52.365
but most of us do work closely with some

07:52.365 --> 07:54.365
of our colleagues over the course of our careers

07:54.365 --> 07:55.485
and on a year to year basis.

07:55.985 --> 07:57.245
And this suggests that those sorts

07:57.245 --> 07:59.165
of ordinary workplace interactions can have really

07:59.165 --> 08:00.845
significant effects on your decisions.

08:01.445 --> 08:02.845
I think there's a couple things that we're able

08:02.845 --> 08:04.085
to contribute to this literature.

08:04.745 --> 08:08.165
The first thing is that we're pretty unusual in being able

08:08.165 --> 08:10.485
to look at the decisions of powerful people.

08:11.185 --> 08:14.125
If we think about in working

08:14.125 --> 08:17.245
with other people influencing your beliefs, my

08:18.315 --> 08:20.135
my belief coming into this study would be

08:20.135 --> 08:23.415
that a college first year is gonna be more malleable than a

08:23.415 --> 08:26.135
70-year-old fellow federal judge, right?

08:26.275 --> 08:27.815
Um, when you're early in your life,

08:28.165 --> 08:31.095
your each experience is gonna have a bigger effect on you

08:31.095 --> 08:32.135
because everything is new to you.

08:32.735 --> 08:34.615
I wouldn't have been surprised if what we saw is

08:34.615 --> 08:36.615
that once someone reaches the end of their career,

08:37.265 --> 08:38.935
their beliefs are relatively fixed.

08:39.395 --> 08:42.135
So our work is suggesting that that might not be the case.

08:42.645 --> 08:46.255
That even these very powerful people are influenceable

08:46.255 --> 08:48.935
through professional interactions through exposure to people

08:48.935 --> 08:50.015
who are different from themselves.

08:50.515 --> 08:53.015
And that's important because if we think about

08:53.875 --> 08:57.885
gender racial diversity in the US labor market, the people

08:57.885 --> 08:59.805
who are making consequential decisions

08:59.805 --> 09:03.165
that influence diversity are generally gonna be older people

09:03.265 --> 09:05.285
who are established in their careers and who are powerful.

09:05.285 --> 09:06.845
This is a group of people we really care about.

09:07.785 --> 09:09.525
The other thing I think is important about this

09:09.525 --> 09:11.285
that we're able to contribute is

09:11.285 --> 09:13.245
that we're looking at hiring decisions.

09:13.255 --> 09:15.765
Those are really important high stakes decisions.

09:16.405 --> 09:18.245
A lot of the research in this, in this field,

09:18.305 --> 09:20.205
and this is again still very valuable,

09:20.205 --> 09:22.325
but a lot of the research in this field is looking at stuff

09:22.325 --> 09:25.685
like surveys or implicit association tests that try

09:25.685 --> 09:27.005
to measure someone's beliefs

09:27.505 --> 09:29.845
or they're looking at relatively small decisions

09:29.845 --> 09:31.165
that people are making, right?

09:31.185 --> 09:33.565
As I mentioned before, decisions like, do you sit next

09:33.565 --> 09:35.045
to a Muslim at the award ceremony?

09:35.635 --> 09:40.025
It's been difficult for us to see how this sort

09:40.025 --> 09:42.225
of intergroup contact affects the

09:42.465 --> 09:43.505
decisions we care the most about.

09:43.565 --> 09:45.625
And hiring decisions is a great example of this.

09:46.325 --> 09:48.945
The last thing I'd say that I think is really important here

09:49.685 --> 09:50.865
is the fact

09:51.025 --> 09:53.585
that these are ordinary professional interactions.

09:54.115 --> 09:55.425
These are things, these are the types

09:55.425 --> 09:57.185
of interactions we expect most professionals

09:57.245 --> 09:58.865
to have over the course of their career.

09:59.445 --> 10:01.585
And so even though the context we're looking at is very

10:01.585 --> 10:02.625
special and very unique,

10:03.005 --> 10:04.825
and you know, we're only able to do this work

10:04.825 --> 10:07.145
because of this special characteristics of this context,

10:07.845 --> 10:10.505
we think that it sheds light on a much broader array

10:11.125 --> 10:13.425
of important professional context

10:13.485 --> 10:14.705
in the United States and around the world.

10:14.945 --> 10:18.585
Diversifying the tops of the profe of professions is likely

10:18.605 --> 10:20.945
to have real benefits for people at the entry level.

10:21.565 --> 10:22.625
One of the things, so for example,

10:22.685 --> 10:25.545
the Biden administration has made a real intentional effort

10:25.805 --> 10:26.905
to promote a diverse

10:26.905 --> 10:31.265
or a more diverse set of candidates to the bench, right?

10:31.265 --> 10:32.385
To the federal judiciary.

10:32.885 --> 10:34.665
And so sort of the most direct thing

10:34.665 --> 10:37.145
that we can say here is our work suggests

10:37.145 --> 10:38.745
that appointing more diverse judges

10:38.885 --> 10:40.545
to the judiciary is likely

10:40.565 --> 10:42.025
to result in more diverse people

10:42.055 --> 10:43.545
receiving appellate clerkships.

10:44.365 --> 10:46.505
And what we might expect is that means

10:46.745 --> 10:47.705
that there's gonna be this sort of

10:47.905 --> 10:48.985
positive feedback loop, right?

10:48.995 --> 10:50.225
We're diversifying the top

10:50.225 --> 10:52.385
of a profession now is gonna mean the people

10:52.405 --> 10:54.505
who are gonna enter the top of the profession later

10:55.015 --> 10:56.465
will be a more diverse group of people.

10:57.005 --> 10:58.785
So I think, you know, when we think about sort

10:58.785 --> 11:02.585
of is it important for us to have CEOs not all be men,

11:02.605 --> 11:04.905
is it important to have lot, you know,

11:05.155 --> 11:06.745
legal partners, not all be men.

11:07.095 --> 11:08.545
What our work is suggesting is that

11:08.545 --> 11:10.745
that's important not just for the people at the elite

11:10.745 --> 11:11.745
who are getting these top jobs,

11:11.805 --> 11:13.225
but also for the rest of us, right?

11:13.225 --> 11:15.945
For people whose, whose lives are affected

11:15.965 --> 11:17.985
by the decisions made by people in that elite.

11:18.405 --> 11:20.465
So I think that's sort of the most direct and, and

11:20.565 --> 11:24.055
and the easiest thing to say in terms of

11:24.055 --> 11:25.895
what the policy implications of our work are,

11:26.615 --> 11:28.415
I think the broader policy implication,

11:28.415 --> 11:29.895
and this gets a little bit fuzzier,

11:30.235 --> 11:33.015
but the broader policy implication is that

11:33.575 --> 11:36.815
segregation can have negative consequences in terms of

11:37.455 --> 11:39.295
maintaining negative beliefs about outgroups.

11:39.835 --> 11:44.175
And so in a wide variety of context, we might expect that

11:44.715 --> 11:48.975
if we can find ways to encourage people to interact with

11:48.975 --> 11:50.815
and work with people who are different from themselves,

11:51.275 --> 11:52.695
that's likely to reduce levels

11:52.755 --> 11:56.205
Of prejudice, you know, over, over our society.

11:56.625 --> 11:59.165
And that's tricky because as economists we think

11:59.165 --> 12:01.405
that people tend to not wanna work with those

12:01.465 --> 12:03.685
who have prejudice beliefs when those prejudices are

12:03.925 --> 12:05.005
directed at them, right?

12:05.005 --> 12:06.965
And people with prejudice beliefs tend not to wanna work

12:06.965 --> 12:08.165
with people they're prejudiced against.

12:08.585 --> 12:10.085
So what we tend to see is that, you know,

12:10.085 --> 12:12.845
when we let people do their own thing, we should expect

12:12.845 --> 12:14.125
to see a reasonable amount

12:14.125 --> 12:15.965
of segregation in the market, right?

12:16.025 --> 12:17.725
As people who experience prejudice sort

12:17.725 --> 12:19.325
of flee from the prejudice, people

12:19.325 --> 12:21.125
who are prejudice flee from those they don't like.

12:21.545 --> 12:23.285
So what this suggests is that that sort

12:23.285 --> 12:26.365
of natural process in the market might lead prejudices

12:26.365 --> 12:28.325
to maintain themselves longer than they would otherwise.

Faculti AI can make mistakes

Research

Policy

Analysis