WEBVTT
00:00:00.900 --> 00:00:01.900
♪ [music] ♪
00:00:03.800 --> 00:00:05.700
- [Narrator] Welcome to
Nobel conversations.
00:00:07.300 --> 00:00:10.240
In this episode, Josh Angrist
and Guido Imbens,
00:00:10.240 --> 00:00:12.000
sit down with Isaiah Andrews
00:00:12.000 --> 00:00:14.303
to discuss how the research
was initially received
00:00:14.900 --> 00:00:17.736
and how they responded
to criticism.
00:00:18.700 --> 00:00:19.300
At the time, did you feel like
you are on to something,
00:00:20.400 --> 00:00:24.000
you felt like this was the beginning
of a whole line of work
00:00:24.000 --> 00:00:27.400
that you felt like was going
to be important or...?
00:00:27.600 --> 00:00:30.100
Not so much that it was
a whole line of work,
00:00:30.100 --> 00:00:32.600
but certainly I felt like,
"Wow, this--"
00:00:32.600 --> 00:00:34.700
We proved something
that people didn't know before,
00:00:34.700 --> 00:00:39.000
that it was worth knowing.
00:00:39.000 --> 00:00:40.000
Yeah, going back compared to
my job market papers having--
00:00:41.600 --> 00:00:45.900
I felt this was actually
a very clear crisp result.
00:00:46.400 --> 00:00:48.400
But there were definitely
was mixed reception
00:00:48.900 --> 00:00:53.200
and I don't think anybody said that,
00:00:53.300 --> 00:00:56.900
"Oh, wow, this is already,
something."
00:00:57.100 --> 00:00:59.600
No, which is the nightmare scenario
for a researcher
00:01:00.300 --> 00:01:03.200
where you think you've
discovered something
00:01:03.300 --> 00:01:04.800
and then somebody else says,
"Oh, I knew that."
00:01:05.000 --> 00:01:08.000
But there were definitely was
a need to convince people
00:01:08.000 --> 00:01:09.000
that this was worth knowing,
that instrumental variables
00:01:09.000 --> 00:01:12.800
estimates a causal effect
for compliers.
00:01:13.200 --> 00:01:18.000
Yeah, but even though it
took a long time
00:01:18.600 --> 00:01:20.400
to convince a bigger audience,
00:01:20.700 --> 00:01:24.600
sometimes even fairly quickly,
the reception was pretty good
00:01:24.800 --> 00:01:27.000
among a small group of people.
00:01:27.200 --> 00:01:31.500
Gary, clearly liked it a lot
from the beginning
00:01:31.800 --> 00:01:34.600
and I remember, because at that point
Josh had left for Israel,
00:01:34.700 --> 00:01:37.400
but I remember explaining it
to Don Ruben
00:01:37.600 --> 00:01:43.700
and he was like,
"Yeah, this really is something here."
00:01:43.800 --> 00:01:47.200
Not right away though,
Don took some convincing.
00:01:47.500 --> 00:01:48.400
By the time you got to Don,
00:01:48.500 --> 00:01:51.500
there have been some back
and forth with him
00:01:51.800 --> 00:01:53.500
and in correspondence actually.
00:01:53.700 --> 00:01:56.700
But I remember at some point
getting a call or email from him
00:01:56.700 --> 00:02:02.300
saying that he was sitting
at the airport in Rome
00:02:02.500 --> 00:02:03.700
and looking at the paper
and thinking,
00:02:03.700 --> 00:02:07.000
"Yeah, no actually,
you guys are onto something."
00:02:07.000 --> 00:02:09.800
We were happy about
00:02:09.800 --> 00:02:10.800
but that took longer
than I think you remember.
00:02:10.800 --> 00:02:12.500
Yeah, it wasn't right away
00:02:12.600 --> 00:02:13.600
[laughter]
00:02:13.700 --> 00:02:16.500
because I know that I was back
in Israel by the time that happened.
00:02:16.500 --> 00:02:18.300
I'd left for Israel in the summer--
00:02:18.400 --> 00:02:22.300
I was only at Harvard for two years.
We had that one year.
00:02:22.600 --> 00:02:25.700
It is remarkable, I mean, that
one year was so fateful for us.
00:02:25.900 --> 00:02:27.200
- [Guido] Yes.
00:02:28.500 --> 00:02:30.200
I think we understood there was
something good happening,
00:02:30.200 --> 00:02:34.000
but maybe we didn't think it was
life-changing, only in retrospect.
00:02:34.400 --> 00:02:35.400
♪ [music] ♪
00:02:35.800 --> 00:02:38.300
- [Isaiah] As you said, it sounds like
a small group of people
00:02:38.300 --> 00:02:39.300
were initially quite receptive,
00:02:39.300 --> 00:02:41.000
perhaps took some time for
a broader group of people
00:02:44.400 --> 00:02:46.700
to come around to seeing
the LATE framework
00:02:46.700 --> 00:02:47.700
as a valuable way to look
at the world.
00:02:47.700 --> 00:02:50.000
I guess, in over the course of that,
00:02:50.100 --> 00:02:52.200
were their periods
where you thought,
00:02:52.300 --> 00:02:53.200
maybe the people saying
this wasn't a useful way
00:02:53.300 --> 00:02:55.800
to look at the world were right?
00:02:55.800 --> 00:02:58.400
Did you get discouraged?
How did you think about?
00:02:58.400 --> 00:03:00.900
I don't think I was discouraged
but the people who were saying
00:03:00.900 --> 00:03:03.900
that we're smart people,
well informed metricians,
00:03:05.000 --> 00:03:08.000
sophisticated readers
00:03:08.900 --> 00:03:11.800
and I think the substance
of the comment was,
00:03:11.800 --> 00:03:15.600
this is not what econometrics
is about.
00:03:16.300 --> 00:03:20.700
Econometrics was being transmitted
at that time was about structure.
00:03:21.300 --> 00:03:24.700
There was this idea that
there's structure in the economy
00:03:25.100 --> 00:03:27.100
and it's our job to discover it
00:03:27.200 --> 00:03:31.200
and what makes it structure
is it's essentially invariant
00:03:31.900 --> 00:03:34.800
and so we're saying,
in the LATE theorem,
00:03:34.900 --> 00:03:39.100
that every instrument produces
its own causal effect,
00:03:39.300 --> 00:03:42.100
which is in contradiction to that
to some extent
00:03:42.400 --> 00:03:45.300
and so that was where the tension was.
00:03:45.300 --> 00:03:46.300
People didn't want
to give up that idea.
00:03:46.300 --> 00:03:47.700
Yeah, I remember once
people were started
00:03:51.200 --> 00:03:56.100
arguing more more vocally
against that,
00:03:56.900 --> 00:04:00.700
that never really
bothered me that much.
00:04:01.000 --> 00:04:03.700
It seems clear that
we had a result there
00:04:04.900 --> 00:04:08.100
and it was somewhat
controversial,
00:04:08.100 --> 00:04:09.100
but controversial in a good way.
00:04:09.100 --> 00:04:10.400
It was clear that people felt
00:04:10.700 --> 00:04:12.800
they had to come out against it because--
00:04:14.100 --> 00:04:17.500
Well, I think what
we think it's good now
00:04:17.500 --> 00:04:20.800
we might not have loved it
at the time.
00:04:20.800 --> 00:04:21.800
I remember being somewhat,
the more upset--
00:04:21.800 --> 00:04:26.400
there was some dinner
where someone said,
00:04:26.700 --> 00:04:28.300
"No, no, that paper with Josh,
00:04:28.800 --> 00:04:32.600
that was doing a disservice
to the profession."
00:04:32.600 --> 00:04:34.400
We definitely had
reactions like that.
00:04:34.800 --> 00:04:38.200
At some level, that may be
indicative of the culture
00:04:38.400 --> 00:04:40.000
in general in economics
at the time.
00:04:41.400 --> 00:04:44.300
I thought back later,
what if that'd happened now,
00:04:44.600 --> 00:04:48.200
if I was a senior person sitting
in that conversation,
00:04:48.300 --> 00:04:52.200
I would call that out because it
really was not appropriate--
00:04:53.000 --> 00:04:54.200
- [Josh] But it wasn't so bad.
00:04:54.600 --> 00:04:57.300
I think the criticism is--
00:04:57.700 --> 00:05:01.500
It wasn't completely misguided,
it was maybe wrong.
00:05:01.800 --> 00:05:04.700
No, no, but you can say
the paper is wrong
00:05:05.400 --> 00:05:07.300
but it's saying that
00:05:07.300 --> 00:05:08.300
it's a disservice
to the profession,
00:05:08.300 --> 00:05:09.300
that's not really--
00:05:09.300 --> 00:05:10.300
Personal.
00:05:10.300 --> 00:05:13.900
Yes, and doing that, not to me,
00:05:13.900 --> 00:05:14.900
but in front of
my senior colleagues.
00:05:14.900 --> 00:05:17.700
But nobody was saying
the result was wrong
00:05:17.700 --> 00:05:18.700
and I remember also,
00:05:18.700 --> 00:05:20.900
some of the comments
were thought-provoking
00:05:20.900 --> 00:05:24.000
so we had some negative reviews,
00:05:24.400 --> 00:05:26.300
I think on the average
causal response paper.
00:05:26.500 --> 00:05:30.600
Somebody said, "These compliers
you can't figure out who they are."
00:05:31.500 --> 00:05:31.900
Right.
00:05:32.000 --> 00:05:33.000
It's one thing to say
00:05:33.000 --> 00:05:35.200
you're estimating
the effect of treatment
00:05:35.200 --> 00:05:36.200
on the treated
or something like that.
00:05:36.200 --> 00:05:38.400
You can tell me who's treated,
00:05:38.600 --> 00:05:42.600
people in the CPS,
you can't tell me who's a complier.
00:05:42.800 --> 00:05:46.800
So that was a legitimate challenge.
00:05:46.800 --> 00:05:47.800
That's certainly fair
and I can see why
00:05:47.800 --> 00:05:53.700
that part made people
a little uneasy and uncomfortable.
00:05:53.800 --> 00:05:54.200
Yeah.
00:05:54.300 --> 00:05:56.400
But it's a at the same time
00:05:56.900 --> 00:06:00.800
because it showed that you couldn't
really go beyond that,
00:06:01.500 --> 00:06:05.500
it was very useful thing
to realize.
00:06:05.500 --> 00:06:06.500
I remember on the day,
00:06:06.500 --> 00:06:11.100
we got to the key result
that I was thinking,
00:06:11.100 --> 00:06:12.100
"Wow, this is as good as it gets.
00:06:12.100 --> 00:06:16.500
Here we actually have an insight
but clearly--"
00:06:17.500 --> 00:06:21.000
And we had to sell it.
00:06:21.000 --> 00:06:22.000
For quite a few years,
we had to sell
00:06:22.000 --> 00:06:24.800
and it's proven to be quite useful.
00:06:25.500 --> 00:06:29.300
I don't think we understood that it
would be so useful at the time.
00:06:30.100 --> 00:06:34.600
No, I did feel early on this was
a substantial insight.
00:06:34.600 --> 00:06:35.600
- [Josh] Yeah we [learned] something.
00:06:35.600 --> 00:06:40.400
But I did not think
goals were there.
00:06:40.500 --> 00:06:42.400
I don't think we were aiming
for the Nobel.
00:06:42.650 --> 00:06:43.650
[laughter]
00:06:43.900 --> 00:06:46.300
We were very happy to get
that note in Econometrica.
00:06:47.600 --> 00:06:48.600
♪ [music] ♪
00:06:49.900 --> 00:06:52.800
- [Isaiah] Are there factors
or are ways of approaching problems
00:06:53.200 --> 00:06:55.600
that lead people to be better
at recognizing the good stuff
00:06:55.600 --> 00:06:56.600
and taking the time to do it
as opposed to dismissing it?
00:06:56.600 --> 00:06:57.700
- [Josh] Sometimes
I think it's helpful.
00:06:57.700 --> 00:07:00.900
If you're trying to
convince somebody
00:07:00.900 --> 00:07:01.900
that you have something
useful to say
00:07:01.900 --> 00:07:04.500
and maybe they don't
speak your language,
00:07:04.700 --> 00:07:09.900
you might need
to learn their language.
00:07:10.300 --> 00:07:10.850
Yes, yes, exactly.
00:07:10.850 --> 00:07:11.400
That's what we did with Don,
we figured out how to--
00:07:12.200 --> 00:07:16.800
I remember we had a very hard time
00:07:16.800 --> 00:07:17.800
explaining the exclusion
restriction to Don,
00:07:17.800 --> 00:07:19.700
maybe rightfully so,
00:07:20.000 --> 00:07:24.400
I think Guido and I
eventually figured out
00:07:24.400 --> 00:07:25.400
that it wasn't formulated
very clearly,
00:07:25.400 --> 00:07:28.700
and we came up
with a way to do that
00:07:28.700 --> 00:07:29.700
in the potential outcomes framework
00:07:29.700 --> 00:07:32.700
that I think worked
for the three of us.
00:07:33.400 --> 00:07:35.800
Yeah, well, it worked for
the bigger literature
00:07:35.800 --> 00:07:38.100
but I think what you're saying
there is exactly right,
00:07:38.100 --> 00:07:39.100
you need to figure out
how not just say,
00:07:41.400 --> 00:07:43.900
"Okay well, I've got this language
and this this works great
00:07:43.900 --> 00:07:45.900
and I've got to convince someone
else to use the language.
00:07:45.900 --> 00:07:47.600
You could first figure out
what language they're using
00:07:49.600 --> 00:07:51.200
and then only then,
can you try to say,
00:07:51.200 --> 00:07:55.700
"Wow, but here you thinking of it
this way,"
00:07:55.700 --> 00:07:56.700
but that's actually
a pretty hard thing to do,
00:07:56.700 --> 00:08:00.000
get someone from
a different discipline,
00:08:00.200 --> 00:08:02.300
convincing them, two junior faculty
in a different department
00:08:02.300 --> 00:08:03.300
actually have something
to say to you,
00:08:03.300 --> 00:08:06.600
that's that takes
a fair amount of effort.
00:08:07.500 --> 00:08:10.200
Yeah, I wrote on a number of times,
in fairly long letters.
00:08:10.700 --> 00:08:14.500
I remember thinking
this is worth doing,
00:08:14.600 --> 00:08:17.600
that if I could convince Don
00:08:18.450 --> 00:08:19.450
that would validate the framework
to some extent.
00:08:20.300 --> 00:08:24.000
I think both you and Don were
00:08:24.000 --> 00:08:25.000
a little bit more confident
that you were right.
00:08:25.000 --> 00:08:27.500
Well, we used to argue a lot
00:08:27.500 --> 00:08:28.500
and you would sometimes
referee those.
00:08:28.500 --> 00:08:29.500
[laughter]
00:08:29.800 --> 00:08:30.700
That was fun.
00:08:30.800 --> 00:08:34.100
It wasn't hurtful.
00:08:35.200 --> 00:08:39.600
I remember getting
a little testy once,
00:08:39.600 --> 00:08:40.600
we had lunch in The Faculty Club
00:08:40.600 --> 00:08:44.400
and we're talking about
the draft lottery paper.
00:08:45.200 --> 00:08:47.600
We were talking about never takes
00:08:47.700 --> 00:08:51.000
as people wounded serve
in the military irrespective of
00:08:51.200 --> 00:08:53.700
whether they were getting drafted
00:08:54.500 --> 00:08:58.700
and you and Don said something
about shooting yourself in the foot,
00:08:59.800 --> 00:09:02.400
as a way of getting out of the military
00:09:02.400 --> 00:09:03.400
and that may be
the exclusion restriction
00:09:03.400 --> 00:09:05.300
for never takes wasn't working
00:09:06.300 --> 00:09:08.900
and then the other one would say,
00:09:08.900 --> 00:09:11.250
"Well, yes you could do that
00:09:11.250 --> 00:09:12.250
but why would you want
to shoot yourself in the foot?"
00:09:12.250 --> 00:09:12.825
[laughter]
00:09:12.825 --> 00:09:13.400
It got a little [out of hand] there.
00:09:13.400 --> 00:09:17.600
I usually go for moving to Canada,
for my example,
00:09:18.150 --> 00:09:19.150
when I'm teaching that.
00:09:19.700 --> 00:09:24.000
But he thinks it's tricky,
00:09:24.100 --> 00:09:27.900
I get students coming
from computer science
00:09:28.100 --> 00:09:29.900
and they want to do things
on causal inference
00:09:30.400 --> 00:09:33.700
and it takes a huge amount
of effort to figure out
00:09:33.700 --> 00:09:36.000
how they actually thinking
about problem
00:09:36.000 --> 00:09:37.000
and whether there's something there
00:09:37.000 --> 00:09:38.500
and so, now over the years,
00:09:38.500 --> 00:09:40.600
I've got a little more appreciation
for the fact
00:09:40.800 --> 00:09:42.100
that Don was actually willing to--
00:09:42.200 --> 00:09:46.000
It took him a while,
but he did engage first with Josh
00:09:46.400 --> 00:09:47.500
and then with both of us
00:09:47.700 --> 00:09:51.000
and rather than dismissing
and say,
00:09:51.500 --> 00:09:54.800
"Well, okay I can't figure out
what these guys are doing
00:09:54.800 --> 00:09:55.800
and it's probably just
not really interesting."
00:09:57.200 --> 00:10:00.300
Everybody always wants
to figure out quickly,
00:10:00.300 --> 00:10:04.000
you want to save time
00:10:04.000 --> 00:10:05.000
and you want to save your brain cells
for other things.
00:10:05.000 --> 00:10:07.000
The fastest route to
that is to figure out
00:10:07.000 --> 00:10:08.000
why you should dismiss something.
00:10:08.000 --> 00:10:09.000
Yes.
00:10:09.000 --> 00:10:11.300
I don't need to spend time on this.
00:10:11.500 --> 00:10:12.500
♪ [music] ♪
00:10:12.700 --> 00:10:15.300
- [Narrator] If you'd like
to watch more
00:10:15.300 --> 00:10:16.300
Nobel conversations, click here,
00:10:16.300 --> 00:10:18.600
or if you'd like to learn
more about econometrics,
00:10:18.700 --> 00:10:21.300
check out Josh's "Mastering
Econometrics" series.
00:10:21.800 --> 00:10:24.700
If you'd like to learn more
about Guido, Josh, and Isaiah
00:10:24.900 --> 00:10:26.400
check out the links
in the description.
00:10:26.702 --> 00:10:28.017
♪ [music] ♪