0:00:00.000,0:00:02.550
♪ [music] ♪

0:00:03.800,0:00:05.800
- [Narrator] Welcome to [br]Nobel Conversations.

0:00:07.100,0:00:08.100
In this episode,

0:00:08.100,0:00:11.570
Josh Angrist and Guido Imbens[br]sit down with Isaiah Andrews

0:00:11.570,0:00:14.600
to discuss how the field [br]of econometrics is evolving.

0:00:16.100,0:00:18.750
- [Isaiah] So Guido and Josh, [br]you're both pioneers

0:00:18.750,0:00:21.500
in developing tools for[br]empirical research in economics.

0:00:21.500,0:00:22.930
And so I'd like to explore

0:00:22.930,0:00:25.300
sort of where you feel like [br]the field is heading,

0:00:25.300,0:00:28.079
sort of economics, econometrics,[br]the whole thing.

0:00:28.510,0:00:31.290
To start, I'd be interested to hear

0:00:32.200,0:00:35.200
about whether you feel like[br]sort of the way in which

0:00:35.200,0:00:38.510
the local average treatment [br]effects framework sort of took hold

0:00:38.800,0:00:42.100
has any lessons for how [br]new empirical methods in economics

0:00:42.100,0:00:44.300
develop and spread[br]or how they should.

0:00:44.560,0:00:45.960
- [Josh] That's a good question.

0:00:46.610,0:00:47.790
You go first.

0:00:47.790,0:00:49.460
(laughter)

0:00:49.700,0:00:53.180
Yeah, so I think [br]the important thing

0:00:53.180,0:00:58.550
is to come up [br]with good convincing cases

0:00:58.550,0:01:02.207
where the questions are clear

0:01:02.400,0:01:05.720
and where kind of the methods[br]apply in general.

0:01:05.720,0:01:07.560
So one thing I--

0:01:08.070,0:01:12.000
Kind of looking back[br]at the subsequent literature,

0:01:12.200,0:01:16.700
so I really like the regression[br]discontinuity literature

0:01:16.700,0:01:19.670
[where there were] clearly a bunch[br]of really convincing examples

0:01:19.670,0:01:21.319
and that allowed people to kind of

0:01:22.300,0:01:27.200
think more clearly, look harder at[br]the methodological questions.

0:01:27.400,0:01:28.800
Kind of do clear applications

0:01:28.800,0:01:30.600
that then allow you [br]to kind of think about,

0:01:30.600,0:01:33.600
"Wow, do this type of assumption[br]seem reasonable here?

0:01:33.600,0:01:38.000
What kind of things do we not like[br]in the early papers?

0:01:38.500,0:01:39.802
How can we improve things?"

0:01:39.802,0:01:42.300
So having clear applications[br]motivating,

0:01:43.400,0:01:46.400
these literatures, [br]I think it's very helpful.

0:01:46.800,0:01:49.300
I'm glad you mentioned[br]the regression discontinuity, Guido.

0:01:49.300,0:01:53.300
I think there's a lot of [br]complementarity between IV and RD,

0:01:54.700,0:01:57.060
Instrumental Variables and[br]Regression Discontinuity.

0:02:00.350,0:02:03.260
And a lot of [br]the econometric applications

0:02:03.260,0:02:04.520
of regression discontinuity

0:02:04.520,0:02:07.230
are what used to be called [br]"fuzzy" RD,

0:02:07.230,0:02:11.620
where, you know, it's not discrete [br]or deterministic at the cutoff,

0:02:11.620,0:02:14.900
but just the change[br]in rates or intensity.

0:02:14.900,0:02:18.740
And and the late framework helps us[br]understand those applications

0:02:18.740,0:02:21.140
and gives us a clear interpretation

0:02:21.140,0:02:25.000
for say, something like, [br]in my paper with Victor Lavy,

0:02:25.000,0:02:28.100
where we use Maimonides'[br]rule, the class size cutoffs.

0:02:28.600,0:02:30.200
What are you getting there?

0:02:30.400,0:02:31.970
Of course, you can[br]answer that question

0:02:31.970,0:02:33.900
with a linear [br]constant effects model,

0:02:34.200,0:02:36.310
but it turns out [br]we're not limited to that,

0:02:36.310,0:02:39.889
and an RD is still very powerful [br]and illuminating,

0:02:40.630,0:02:42.100
even when, you know,

0:02:42.100,0:02:46.000
the correlation between the cutoff[br]and the variable of interest,

0:02:46.000,0:02:48.980
in this case class size, [br]is partial,

0:02:48.980,0:02:51.000
maybe even not that strong.

0:02:52.000,0:02:54.999
So there was definitely a kind of,[br]a parallel development.

0:02:54.999,0:02:56.400
It's also interesting,

0:02:56.600,0:02:59.780
you know, nobody talked about[br]regression discontinuity designs

0:02:59.780,0:03:01.220
when we were in graduate school,

0:03:01.220,0:03:05.300
it was something that other social[br]scientists were interested in,

0:03:05.800,0:03:09.507
and that kind of grew up [br]alongside the late framework

0:03:09.507,0:03:14.787
and we've both done work on [br]both applications and methods there

0:03:14.787,0:03:18.377
and it's been very exciting [br]to see that kind of develop

0:03:18.377,0:03:19.800
and become so important.

0:03:20.000,0:03:23.700
It's part of a general evolution,[br]I think, towards, you know,

0:03:24.000,0:03:27.700
credible identification strategies[br]causal effects...

0:03:28.630,0:03:30.200
less, you know, making econometrics

0:03:30.400,0:03:33.300
more about causal questions[br]than about models.

0:03:33.640,0:03:34.650
In terms of the future,

0:03:34.650,0:03:37.860
I think one thing that LATE [br]has helped facilitate

0:03:37.860,0:03:42.700
is a move towards more creative,[br]randomized trials, where,

0:03:42.700,0:03:44.400
you know, there's[br]something of interest,

0:03:45.500,0:03:50.700
it's not possible or straightforward[br]to simply turn it off or on,

0:03:51.000,0:03:54.584
but you can encourage it [br]or discourage it.

0:03:54.584,0:03:58.200
So you subsidize schooling [br]with financial aid, for example.

0:03:59.000,0:04:02.080
So now we have a whole [br]framework for interpreting that,

0:04:03.600,0:04:06.900
and it kind of opens[br]the doors to randomized trials

0:04:06.900,0:04:10.300
of things that that maybe would,[br]you know,

0:04:10.300,0:04:12.471
not have seem possible before.

0:04:14.500,0:04:17.700
We've used that a lot in the work[br]we do on schools in our--

0:04:17.700,0:04:21.160
in the Blueprint Lab at MIT,

0:04:22.360,0:04:26.600
we're exploiting random assignment[br]and in very creative ways, I think.

0:04:28.100,0:04:32.300
- [Isaiah] Related to that, do you see sort[br]of particular factors that make for

0:04:32.400,0:04:34.300
useful research and econometrics.

0:04:34.400,0:04:37.400
You've alluded to it?

0:04:37.400,0:04:40.300
Having a clear connection to[br]problems that are actually coming up.

0:04:40.300,0:04:45.000
And empirical practice is often a good[br]idea. I'll send it. Always a good idea.

0:04:45.700,0:04:50.100
I often find myself sitting in an[br]economy metrics Theory, seminar.

0:04:50.700,0:04:52.500
Say the Harvard MIT seminar

0:04:53.400,0:04:57.000
and I'm thinking what problem is[br]this guy solving who has this?

0:04:57.100,0:04:59.800
This problem and you know,

0:05:01.600,0:05:04.700
sometimes there's an[br]embarrassing silence if I ask

0:05:04.900,0:05:08.300
or there might be a[br]fairly contrived scenario.

0:05:08.800,0:05:11.600
I want to see where the tool is useful.

0:05:12.500,0:05:14.900
There are some purely foundational tools.

0:05:14.900,0:05:17.600
I do take the point, you[br]know, there are people who are

0:05:18.200,0:05:22.500
working on conceptual[br]foundations of you know,

0:05:22.600,0:05:25.300
it's more becomes more like[br]mathematical statistics.

0:05:25.800,0:05:27.000
I mean, I remember an early example

0:05:27.100,0:05:28.100
I believe that that I,

0:05:28.200,0:05:32.500
you know, I struggled to understand was[br]the idea of stochastic Equity continuity,

0:05:32.500,0:05:36.900
which my one of my thesis advisors Whitney[br]knew he was using to great effect and

0:05:37.500,0:05:39.900
I was trying to understand[br]that and there isn't really.

0:05:40.600,0:05:45.200
It's really foundational. It's not[br]but an application that's driving that

0:05:46.100,0:05:47.300
at least not immediately

0:05:48.600,0:05:53.200
but but most things are not like that[br]and so there should be a problem.

0:05:53.800,0:05:56.900
And the I think it's on the it's on

0:05:57.100,0:05:59.600
On the, the seller of that sort of thing,

0:06:00.100,0:06:04.400
you know, because there's opportunity[br]cost the time and attention and effort

0:06:04.400,0:06:07.200
to understand things to, you[br]know, it's on the seller to say.

0:06:07.400,0:06:08.900
Hey, I'm solving this problem

0:06:09.400,0:06:12.900
and and here's a set of results[br]that show that it's useful.

0:06:12.900,0:06:15.200
And here's some insight that I get.

0:06:16.200,0:06:17.200
As you said, Josh, great,

0:06:17.200,0:06:20.700
sort of there's been a move in the[br]direction of thinking more about causality

0:06:20.700,0:06:22.800
in economics and empirical[br]work in economics,

0:06:22.900,0:06:24.800
any consequences of sort of the Wilds,

0:06:24.800,0:06:27.000
the spread of that view. That[br]surprised you or anything.

0:06:27.100,0:06:31.400
UB was downsides of sort of the way that[br]he could empirical, economics has gone

0:06:31.500,0:06:32.100
sometimes.

0:06:32.100,0:06:38.500
I see somebody does Ivy and they get a[br]result which seems implausibly large.

0:06:38.800,0:06:40.200
That's the usual case.

0:06:42.500,0:06:43.900
So it might be, you know,

0:06:43.900,0:06:48.600
an extraordinarily large causal effect[br]of some relatively minor Intervention,

0:06:49.100,0:06:52.800
which was randomized or for[br]which you could make a case that

0:06:52.900,0:06:56.900
that there's a good design.[br]And then when I see that,

0:06:57.100,0:06:57.600
That and,

0:06:57.600,0:06:58.500
you know, I think,

0:06:58.500,0:06:58.900
you know,

0:06:58.900,0:07:02.600
it's very hard for me to believe that this[br]relatively minor intervention has such

0:07:02.600,0:07:04.000
a large defect,

0:07:04.100,0:07:07.400
the author. Well, sometimes[br]resort to the local average,

0:07:07.400,0:07:09.100
treatment effects theorem and say,

0:07:09.400,0:07:12.700
wow, these compliers, you know,[br]they're special in some way.

0:07:13.300,0:07:17.600
And, you know, they just benefit[br]extraordinarily from this intervention

0:07:18.100,0:07:22.100
and I'm reluctant to take that[br]at face value. I think, you know,

0:07:22.100,0:07:24.100
often when effects are too big,

0:07:24.300,0:07:26.900
it's because the exclusion[br]restriction is failing. So

0:07:27.100,0:07:31.700
Don't really have the right endogenous[br]variable to scale that result.

0:07:32.000,0:07:35.700
And so I'm not too happy to see

0:07:35.800,0:07:38.800
you know, just sort of[br]a generic heterogeneity

0:07:38.900,0:07:43.800
argument being used to excuse something[br]that I think might be a deeper problem.

0:07:45.300,0:07:47.400
I think it played somewhat[br]of an unfortunate roll pin.

0:07:47.400,0:07:52.300
The discussions kind of between reduced[br]form and structural approaches where

0:07:52.600,0:07:54.200
I feel that wasn't quite

0:07:55.000,0:07:59.300
right. The instrumental[br]variables assumptions are

0:08:00.400,0:08:05.200
at the core structural assumptions about[br]Behavior. They were coming from economic

0:08:07.100,0:08:09.900
thinking about the economic[br]behavior of agents,

0:08:10.300,0:08:15.000
and it's somehow it got[br]pushed in a Direction.

0:08:15.100,0:08:17.600
That I think wasn't[br]really very helpful. If

0:08:18.800,0:08:21.700
the way I think, initially the

0:08:22.800,0:08:27.300
we wrote things up. It was it was describing[br]what was happening, there was set of

0:08:27.500,0:08:32.200
methods. People were using be[br]clarified what those methods were doing

0:08:32.900,0:08:38.500
and in a way that I think[br]contain a fair amount of insight,

0:08:39.100,0:08:45.000
but it somehow it got pushed into a corner[br]that I think was not necessarily very

0:08:45.100,0:08:48.600
or even just the language of[br]reduced form versus structural.

0:08:48.600,0:08:51.100
I find kind of funny in[br]the sense that the right

0:08:51.100,0:08:53.100
the local average treatment[br]effect model, right?

0:08:53.100,0:08:55.300
The potential outcomes[br]model is a nonparametric.

0:08:55.300,0:08:56.200
Structural model,

0:08:56.200,0:08:58.600
if you want to think about it, as[br]you sort of suggested, he does.

0:08:58.600,0:08:59.400
So, there's something,

0:09:00.100,0:09:03.700
there's something a little funny about[br]putting these two things in a position when

0:09:03.800,0:09:06.600
yes, well, that language, of[br]course, comes from the area, the

0:09:06.900,0:09:09.800
70s equations framework that we inherited.

0:09:10.400,0:09:12.400
It has the advantage that people seem

0:09:12.400,0:09:15.000
to know what you mean[br]when you use it, but might

0:09:15.100,0:09:18.200
That people are hearing different. Different[br]people are hearing different things.

0:09:18.300,0:09:20.900
Yeah. I think I think veggies[br]Farmers had become use

0:09:20.900,0:09:23.200
a little bit of the[br]pejoratives. Okay? Yeah.

0:09:23.300,0:09:28.300
The word, which is not really quite[br]what it was originally intended for.

0:09:30.100,0:09:34.100
I guess something else that strikes[br]me in thinking about the effects of

0:09:34.100,0:09:38.200
the local average treatment effect[br]framework is that often folks will appeal to

0:09:38.200,0:09:41.800
a local average, treatment effects[br]intuition for settings. Well, beyond

0:09:42.000,0:09:44.900
ones, where any sort of formal[br]results has actually been

0:09:45.000,0:09:49.700
Shhhhht. And I'm curious given[br]all the work that you guys did to,

0:09:49.900,0:09:53.200
you know, establish late results in[br]different in different settings. I'm curious

0:09:53.300,0:09:57.800
any thoughts on that. I think there's[br]going to be a lot of cases where

0:09:57.900,0:10:02.200
the intuition does get.[br]You get you some distance,

0:10:02.800,0:10:07.600
but it's going to be somewhat limited[br]and establishing formal results. There

0:10:08.400,0:10:12.700
may be a little tricky and there may[br]be only work in special circumstances,

0:10:13.100,0:10:13.500
you need.

0:10:14.600,0:10:19.500
And you end up with a lot of formality[br]that may not quite capture the intuition

0:10:19.900,0:10:23.200
sometimes I'm somewhat uneasy with them[br]and they are not necessarily the papers.

0:10:23.200,0:10:25.000
I would want to ride that the

0:10:25.100,0:10:30.000
but I do think something do intuition[br]orphaned US capture part of the

0:10:30.200,0:10:31.100
of the problem.

0:10:33.100,0:10:36.300
I think, in some sense we were[br]kind of very fortunate there

0:10:36.900,0:10:40.500
in the way. The late paper go handle. It.[br]Don't know if that, actually the editor,

0:10:40.600,0:10:41.700
made it much shorter

0:10:42.100,0:10:46.300
and that then allowed us to kind of[br]focus on very clear, crisp results

0:10:47.100,0:10:49.800
where if, you know, this,

0:10:50.000,0:10:54.200
this is somewhat unfortunate tendency in[br]the commercialization of having the papers.

0:10:54.600,0:10:58.800
Well, you should be able to fix that, man.[br]I'm trying to take some time to fix that.

0:10:59.400,0:11:02.700
I think this is an example where it's sort[br]of very clear that having it. Be sure.

0:11:02.900,0:11:08.000
It's actually impose that no paper can[br]be longer than the late paper that wow.

0:11:08.800,0:11:14.300
Great. At least no Theory. No Theory Pig.[br]Yeah, and I think, I think they're well,

0:11:14.500,0:11:16.800
I'm trying very hard to get[br]the papers to be shorter.

0:11:16.800,0:11:18.700
And I think there's a lot of value

0:11:19.200,0:11:22.600
today because it's often the second[br]part of the paper that doesn't actually

0:11:23.700,0:11:26.500
Get you much further[br]and understanding things

0:11:27.000,0:11:31.700
but and it does make things much[br]harder to read and, you know,

0:11:32.400,0:11:33.700
it sort of goes back to

0:11:34.200,0:11:38.500
how I think he kind of a trick should[br]be done to you should focus on the see.

0:11:38.700,0:11:41.300
It should be reasonably[br]close to empirical problems.

0:11:41.500,0:11:43.900
They should be very clear problems.

0:11:44.800,0:11:48.900
But then often the the theory[br]doesn't need to be quite so long.

0:11:49.000,0:11:49.300
Yeah,

0:11:51.100,0:11:53.400
I think they had things have

0:11:53.600,0:11:54.700
On a little off track.

0:11:56.400,0:11:58.400
The relatively recent change has been a

0:11:58.500,0:12:02.200
seeming big increase in demand for[br]people with sort of econometrics.

0:12:02.200,0:12:04.800
Causal effect, estimation[br]skills in the tech sector.

0:12:05.000,0:12:09.000
I'm interested either of you have[br]thoughts on sort of how that's gonna

0:12:09.200,0:12:11.600
how that's going to interact with[br]the development of empirical methods,

0:12:11.600,0:12:14.300
or Empirical research, and[br]economics. Going forward, sort of

0:12:14.600,0:12:21.000
whether sort of a meta point, which[br]is there's this new kind of employer

0:12:21.800,0:12:26.000
the Amazons and the Uber and, you know,

0:12:26.200,0:12:27.600
Riser world

0:12:28.000,0:12:29.300
and I think that's great.

0:12:29.300,0:12:33.200
And I'd like to tell my students about[br]that, you know, especially at MIT.

0:12:33.200,0:12:37.000
We have a lot of computer science[br]Majors. That's our biggest major

0:12:37.400,0:12:42.800
and I try to seduce some of those folks[br]into economics by saying, you know,

0:12:43.200,0:12:45.700
you can go work for these,

0:12:45.800,0:12:48.400
you know companies that[br]people are very keen to

0:12:48.700,0:12:50.800
work for because the work seems exciting,

0:12:52.000,0:12:56.000
you know that the skills that you get in[br]econometrics are are as good or better.

0:12:56.100,0:13:01.100
Better than than any competing discipline[br]has to offer. So you should at least

0:13:01.400,0:13:04.200
take some econ, take some[br]econometrics. And some econ.

0:13:04.800,0:13:07.000
I did a fun project with a uber

0:13:07.600,0:13:12.900
on labor supply of Uber drivers and was[br]very, very exciting to be part of that.

0:13:13.100,0:13:15.400
Plus. I got to drive for Uber for a while

0:13:15.900,0:13:20.700
and I thought that was fun tonight. I did[br]not make enough that I was attempted to

0:13:21.100,0:13:25.100
give up by a mighty job, but[br]I enjoyed the experience.

0:13:25.300,0:13:26.000
I see a

0:13:26.200,0:13:30.900
Cho challenge to our model[br]of graduate education here,

0:13:31.700,0:13:37.400
which is if we're trading people[br]to go work at Amazon, you know,

0:13:37.900,0:13:42.900
it's not clear. Why? You know, we should[br]be paying graduate stipends for that.

0:13:43.200,0:13:45.400
Why should the taxpayer effectively

0:13:46.100,0:13:51.400
be subsidizing? That our graduate education[br]in the u.s. Is generously subsidized?

0:13:51.400,0:13:56.000
Even in private universities. It's[br]ultimately there's a lot of public money.

0:13:56.100,0:13:59.300
Me there. And I think the[br]traditional rationale for that is,

0:13:59.500,0:14:03.900
you know, we were training, Educators and[br]Scholars, and there's a great externality

0:14:04.300,0:14:05.700
from the work that we do.

0:14:05.700,0:14:09.600
It's either the research externality,[br]or a teaching externality.

0:14:10.100,0:14:14.600
But, you know, if many of our students[br]are going to work in the private sector,

0:14:16.300,0:14:21.700
that's fine, but that maybe their[br]employers should pay for that.

0:14:22.300,0:14:25.100
He says, so different from[br]people working for a Consulting.

0:14:26.300,0:14:26.900
Trust me.

0:14:27.200,0:14:33.000
It's not clear to me that the number[br]of jobs in academics has changed.

0:14:33.100,0:14:37.600
It's just, I feel like this is a[br]growing sector whereas Consulting,

0:14:37.700,0:14:42.100
your right to raise that, it might[br]be the same for for Consulting.

0:14:43.300,0:14:44.400
But this,

0:14:44.500,0:14:47.500
you know, I'm placing more and[br]more students in these businesses.

0:14:47.500,0:14:50.400
So, it's on my mind in[br]a way that I've sort of,

0:14:50.800,0:14:55.500
you know, not been attentive to consulting[br]jobs, you know, Consulting was always,

0:14:55.600,0:15:00.400
It's important and I think they'll so[br]there's some movement from Consulting back

0:15:00.400,0:15:02.600
into research. It's a little more fluid.

0:15:02.900,0:15:03.500
The,

0:15:03.900,0:15:05.400
a lot of the work in the

0:15:06.400,0:15:09.800
in both domains. I have to say,[br]it's not really different but

0:15:10.100,0:15:13.700
you know, people who are working[br]in the tech sector are doing things

0:15:13.700,0:15:16.800
that are potentially of scientific[br]interest, but mostly it's hidden.

0:15:17.100,0:15:20.900
Then you really I have to say, you know,[br]why, why is the government paying for this?

0:15:21.800,0:15:25.500
Yeah, although yeah, I mean taquitos point,[br]I guess it. There's a, there's a data.

0:15:25.600,0:15:30.000
Question here of it has the sort[br]of total nanak. It sort of say

0:15:30.500,0:15:30.900
private

0:15:31.300,0:15:34.100
for-profit sector employment of econ Ph.D.

0:15:34.100,0:15:37.700
Program graduates increased or has[br]it just been a substitution from

0:15:37.900,0:15:40.200
finance and Consulting towards tack.

0:15:40.300,0:15:44.300
I may be a reaction to something[br]that's not really happening

0:15:44.400,0:15:48.200
so bad. I've actually done some work[br]with some of these tech companies.

0:15:49.100,0:15:52.300
So I don't disagree with Justice[br]point that we need to think

0:15:52.300,0:15:55.100
a little bit about the funding model[br]whose it was in the end paying for the

0:15:55.600,0:15:59.400
It education. But from a[br]scientific perspective.

0:16:00.100,0:16:03.500
The only do these places have[br]have great data and nowadays.

0:16:03.500,0:16:07.100
They tend to be very careful[br]with that for privacy reasons,

0:16:07.500,0:16:08.800
but also have great questions.

0:16:10.200,0:16:11.200
I find it very

0:16:11.600,0:16:13.300
inspiring kind of to listen to

0:16:13.300,0:16:15.800
the people there and kind of see[br]what kind of questions they have

0:16:15.900,0:16:17.300
and often their questions.

0:16:18.200,0:16:20.600
That also come up outside of these.

0:16:20.700,0:16:25.400
These companies have a couple of[br]papers with the rights in the chat.

0:16:25.800,0:16:26.300
And then

0:16:26.500,0:16:31.600
as soon as an atheist kind of where we[br]look at ways of combining experimental data

0:16:31.700,0:16:34.100
and observational data, and can it there.

0:16:35.500,0:16:38.600
Rights Chetty was interested[br]in what is the effect

0:16:38.700,0:16:44.600
of Early Childhood programs on outcomes[br]later in life? Not just kind of test scores,

0:16:44.600,0:16:48.300
but on earnings and stuff, and[br]we cannot be developed methods

0:16:48.600,0:16:51.500
that would help you shed[br]light on that, on the some,

0:16:52.700,0:16:55.000
in some settings and the same problems.

0:16:56.300,0:17:00.500
Came up kind of in this[br]tech company settings.

0:17:00.800,0:17:03.700
And so for my perspective, it's

0:17:04.400,0:17:07.500
the same kind of a stocking two[br]people doing a protocol work.

0:17:07.600,0:17:11.800
I tried to kind of look at these[br]specific problems and then try to come up

0:17:11.900,0:17:18.300
with more General problems that we[br]formulating the problems at a higher level.

0:17:18.500,0:17:22.900
So that I can think about solutions[br]that work in a range of settings.

0:17:23.400,0:17:25.500
And so from that perspective, the

0:17:25.700,0:17:30.300
His with the the tech companies I[br]just very valuable and very useful.

0:17:30.900,0:17:31.400
It's know.

0:17:31.700,0:17:33.700
We do have students. Now spent

0:17:33.800,0:17:37.200
doing internships there and[br]then coming back and writing

0:17:37.400,0:17:43.400
more interesting thesis, as a[br]result of their experiences there.

0:17:44.600,0:17:47.800
If you'd like to watch more[br]Nobel conversations, click here,

0:17:48.200,0:17:50.500
or if you'd like to learn[br]more about econometrics,

0:17:50.600,0:17:53.200
check out Josh's mastering[br]econometrics series.

0:17:53.700,0:17:55.500
If you'd like to learn more about he do.

0:17:55.600,0:17:58.300
Josh and Isaiah check out[br]the links in the description.