WEBVTT

00:00:00.030 --> 00:00:01.830
♪ [music] ♪

00:00:03.630 --> 00:00:05.630
- [Narrator] Welcome 
to Nobel Conversations.

00:00:06.930 --> 00:00:10.300
In this episode,
Josh Angrist and Guido Imbens,

00:00:10.300 --> 00:00:13.380
sit down with Isaiah Andrews 
to discuss the key ingredients

00:00:13.380 --> 00:00:16.700
in their Nobel-winning
collaboration.

00:00:16.700 --> 00:00:19.590
Josh and Guido, first, 
congratulations on the Nobel Prize!

00:00:19.590 --> 00:00:20.620
Thank you.

00:00:20.620 --> 00:00:23.190
- [Isaiah] The work you did 
together, particularly the work

00:00:23.190 --> 00:00:24.778
on the Local Average 
Treatment Effect,

00:00:24.778 --> 00:00:25.986
or LATE framework

00:00:25.986 --> 00:00:28.900
was cited as one of the big reasons
you won the prize.

00:00:29.200 --> 00:00:32.600
At the same time, you only
overlapped at Harvard for a year--

00:00:32.900 --> 00:00:34.300
if I'm remembering correctly--

00:00:35.210 --> 00:00:36.650
it would be great to hear 
a bit more

00:00:36.650 --> 00:00:38.500
about how you started 
this collaboration

00:00:38.500 --> 00:00:40.964
and what made your working
relationship productive.

00:00:40.964 --> 00:00:43.600
Are there ways in which you felt 
like you complimented each other,

00:00:43.600 --> 00:00:46.470
what got things started
on such a productive trajectory?

00:00:46.790 --> 00:00:50.624
Your job talk, as I recall, Guido,
it wasn't very interesting

00:00:50.624 --> 00:00:52.740
but I think it was 
a choice-based sampling--

00:00:52.740 --> 00:00:54.600
It was. It was.
[laughter]

00:00:54.600 --> 00:00:56.280
I was a very marginal hire there

00:00:56.280 --> 00:00:58.336
because they didn't 
actually interview me

00:00:58.336 --> 00:00:59.830
on the regular job market,

00:00:59.830 --> 00:01:03.000
but I think they were 
very desperate to get someone else

00:01:03.000 --> 00:01:04.900
to actually teach their courses.

00:01:05.920 --> 00:01:08.470
It was after they had 
a couple of seminars already

00:01:08.470 --> 00:01:11.123
and they're still looking 
in econometrics,

00:01:11.135 --> 00:01:13.940
- so Gary called me and kind of--
- [Josh] Gary Chamberlain?

00:01:13.940 --> 00:01:16.700
Gary Chamberlain called me and
interviewed me over the telephone.

00:01:17.400 --> 00:01:19.474
He said, "Okay, well, 
why don't you come out

00:01:19.474 --> 00:01:20.828
and give a talk?"

00:01:21.089 --> 00:01:23.411
- [Josh] I remember this talk 
a little bit.

00:01:23.411 --> 00:01:26.941
I remember the dinner that 
you and Gary and I had.

00:01:29.000 --> 00:01:32.900
I remember not being very excited
about your job market paper,

00:01:33.600 --> 00:01:38.220
but I saw that Gary was and
luckily, Gary's view prevailed...

00:01:38.580 --> 00:01:39.600
Yes.

00:01:39.600 --> 00:01:41.900
- [Josh] ...and Harvard 
made you an offer

00:01:42.400 --> 00:01:46.056
and I think we started talking to 
each other pretty pretty soon

00:01:46.056 --> 00:01:49.810
after you arrived
in the fall of 1990, right?

00:01:49.810 --> 00:01:52.990
As I said, I came and I didn't
have a very clear agenda.

00:01:52.990 --> 00:01:55.700
I was a little intimidated
getting there.

00:01:56.000 --> 00:01:59.776
But Gary kind of said, 
"No, you should talk to Josh."

00:01:59.776 --> 00:02:04.578
You should go to the labor seminar,
kind of see what these people do.

00:02:04.578 --> 00:02:06.990
They're doing very 
interesting things there."

00:02:07.470 --> 00:02:08.880
I listened to Gary.

00:02:09.880 --> 00:02:10.938
As we did.

00:02:10.938 --> 00:02:14.834
As we did in the those days 
and ever since.

00:02:14.834 --> 00:02:16.700
I think it helped that 
we were neighbors.

00:02:16.700 --> 00:02:20.774
We both lived in Harvard's
junior faculty housing,

00:02:22.030 --> 00:02:25.040
partly because housing costs
were very high in Cambridge

00:02:25.040 --> 00:02:27.200
relative to our salary, 
which was very low.

00:02:27.800 --> 00:02:29.456
I think it also made a difference,

00:02:29.456 --> 00:02:31.212
neither of us came from Cambridge,

00:02:31.212 --> 00:02:33.360
so there were a lot of MIT people

00:02:33.360 --> 00:02:36.027
who already had
their whole networks,

00:02:36.441 --> 00:02:37.800
kind of our collaborators.

00:02:37.800 --> 00:02:39.460
♪ [music] ♪

00:02:39.460 --> 00:02:41.850
- [Josh] Well, I think
we figured out

00:02:41.850 --> 00:02:43.800
a mode of working together, also.

00:02:43.800 --> 00:02:46.410
We had kind of a regular date, 
so we were neighbors

00:02:46.860 --> 00:02:48.760
and we often did 
our laundry together.

00:02:49.510 --> 00:02:51.912
We didn't have laundry
machines at our apartments.

00:02:52.500 --> 00:02:55.450
But we used to do our laundry
and we were talking

00:02:55.450 --> 00:02:59.072
and you had a way 
of fairly systematically,

00:03:00.330 --> 00:03:03.300
addressing questions that
would come up in our discussions

00:03:03.300 --> 00:03:05.562
and the one thing that 
I was very impressed by,

00:03:06.400 --> 00:03:08.728
our early interaction,
is you would follow up.

00:03:09.993 --> 00:03:11.500
You would write some things down.

00:03:11.500 --> 00:03:13.250
Looking back at those days,
sort of clearly,

00:03:13.250 --> 00:03:16.460
I just had a lot more time 
to actually think.

00:03:16.460 --> 00:03:17.915
I look at my junior
colleagues now--

00:03:17.915 --> 00:03:19.580
[Guido] You don't have time 
to think now.

00:03:19.580 --> 00:03:22.753
[Josh] No, but for me that is 
kind of one thing,

00:03:22.753 --> 00:03:24.850
but I feel now a lot of 
my junior colleagues

00:03:24.850 --> 00:03:27.200
don't actually have a lot 
of time to think.

00:03:27.200 --> 00:03:29.710
People are just doing 
so many projects,

00:03:29.710 --> 00:03:31.800
and it's actually so hard

00:03:31.800 --> 00:03:34.560
and there's so much pressure 
on people to publish.

00:03:34.560 --> 00:03:38.081
I remember spending a lot of time 
sitting in my office

00:03:38.081 --> 00:03:41.505
and thinking, 
"Wow, what shall I do now?"

00:03:41.505 --> 00:03:42.835
[laughter]

00:03:43.410 --> 00:03:45.096
But it would give me a lot of time

00:03:45.096 --> 00:03:46.883
to actually think about 
these problems

00:03:46.883 --> 00:03:48.790
and trying to figure it them out

00:03:48.790 --> 00:03:50.610
and I could actually go to seminars

00:03:52.000 --> 00:03:56.651
and then the next day have coffee 
or lunch with Josh or Gary

00:03:56.651 --> 00:03:58.640
and actually talk about 
those things.

00:03:58.640 --> 00:04:00.169
- [Isaiah] You guys 
weren't actually

00:04:00.169 --> 00:04:01.558
at Harvard together all that long,

00:04:01.558 --> 00:04:03.300
you started working
together pretty quickly.

00:04:03.300 --> 00:04:06.600
Were you both in the mindset that 
you were looking for co-authors,

00:04:06.600 --> 00:04:09.260
or looking for a particular type
of types of co-authors at the time

00:04:09.260 --> 00:04:11.600
or was it more sort of 
fortuitous than that?

00:04:11.600 --> 00:04:13.020
- [Josh] I think we were lucky.

00:04:13.020 --> 00:04:15.444
I don't remember
that I was looking.

00:04:16.130 --> 00:04:17.900
I think, it was more fortuitous.

00:04:18.370 --> 00:04:21.411
I said I came in, 
I'd done my job market paper,

00:04:21.411 --> 00:04:23.110
and another paper for my thesis

00:04:23.110 --> 00:04:25.500
and I was just very happy 
to come to Harvard

00:04:25.500 --> 00:04:27.701
and suddenly there were all these
seminars to go to,

00:04:28.262 --> 00:04:30.712
and lots of interesting people 
to talk to,

00:04:31.200 --> 00:04:35.805
but it wasn't a very
conscious thing on my part.

00:04:36.300 --> 00:04:39.200
Looking back, I think there
was a moment for me,

00:04:39.700 --> 00:04:42.933
where I was discussing 
instrumental variables,

00:04:42.933 --> 00:04:46.020
potential outcomes, 
treatment effects with Guido

00:04:47.000 --> 00:04:49.859
and we had
a pretty good discussion,

00:04:49.859 --> 00:04:54.539
but then he also sent me some notes

00:04:55.770 --> 00:05:00.695
and the notes were very methodical
write-up of our discussion

00:05:00.695 --> 00:05:02.966
and what you thought,

00:05:03.600 --> 00:05:07.457
we had been concluding 
in a fairly formal way

00:05:07.457 --> 00:05:09.887
and I thought, 
"Well, that's great."

00:05:10.340 --> 00:05:13.020
Talk is cheap, right, 
but with somebody...

00:05:13.020 --> 00:05:15.514
- [Guido] Yeah, but--
- ...really writes out their story.

00:05:15.900 --> 00:05:18.500
- [Guido] For me, it really helps
writing things down

00:05:18.500 --> 00:05:22.771
and I do remember working with Josh

00:05:22.771 --> 00:05:26.163
and sitting in my office 
and writing things out

00:05:26.633 --> 00:05:28.880
and you guys have all
had the discussions with Gary

00:05:28.880 --> 00:05:32.212
where afterwards we need 
to then sit down

00:05:32.212 --> 00:05:33.920
and actually write things up

00:05:33.920 --> 00:05:36.641
to figure out exactly 
what was going on.

00:05:37.810 --> 00:05:39.360
I think the other thing 
we had, Guido,

00:05:39.360 --> 00:05:41.655
is we had some 
very concrete questions

00:05:41.655 --> 00:05:43.610
that came from applications.

00:05:43.610 --> 00:05:45.000
- [Guido] Yeah.

00:05:45.600 --> 00:05:47.465
A lot of econometrics, in my view,

00:05:47.870 --> 00:05:51.421
that we were schooled in 
was about models,

00:05:51.421 --> 00:05:55.411
here's a model and what can 
you say about this model?

00:05:57.110 --> 00:06:00.200
I think we were thinking about, 
here's a particular scenario,

00:06:00.500 --> 00:06:03.800
draft eligibility is an instrument
for whether you serve in the Army.

00:06:04.830 --> 00:06:06.300
What do we learn from that?

00:06:06.300 --> 00:06:07.300
- [Guido] That's right.

00:06:07.300 --> 00:06:11.930
That's right, and that's sort of 
where your influence

00:06:11.930 --> 00:06:14.890
on the way I do research now 
is still very clear--

00:06:14.890 --> 00:06:16.970
♪ [music] ♪

00:06:16.970 --> 00:06:19.240
- [Isaiah] Zooming out 
a little bit, just thinking about

00:06:19.240 --> 00:06:20.940
when you guys started 
working on this,

00:06:20.940 --> 00:06:22.353
when you started working together,

00:06:23.100 --> 00:06:24.410
any thoughts for folks

00:06:24.410 --> 00:06:26.890
who are just interested in
finding productive

00:06:26.890 --> 00:06:28.170
co-authors being productive?

00:06:28.170 --> 00:06:30.900
I mean, Guido already mentioned
the importance of having time,

00:06:30.900 --> 00:06:32.100
right, which it is.

00:06:32.100 --> 00:06:34.860
It is very easily not to have 
a lot of time to think--

00:06:34.860 --> 00:06:36.250
You definitely have to make time.

00:06:36.250 --> 00:06:38.569
That's a great question 
though, Isaiah,

00:06:40.400 --> 00:06:42.738
and I tell my students that

00:06:42.738 --> 00:06:46.500
you should pick your co-authors 
as carefully,

00:06:46.500 --> 00:06:48.805
maybe more carefully 
than you pick your spouse.

00:06:49.390 --> 00:06:51.774
You want to find co-authors who,

00:06:53.800 --> 00:06:57.724
you have some complementarity

00:06:58.120 --> 00:07:00.985
and that's what makes 
a strong relationship.

00:07:03.200 --> 00:07:05.050
You don't want to work 
with somebody

00:07:05.050 --> 00:07:06.900
who sees the world exactly like you

00:07:09.370 --> 00:07:11.754
and as much as Guido 
and I agree about things,

00:07:12.071 --> 00:07:14.820
we often disagree 
about things to this day

00:07:15.565 --> 00:07:18.688
and it's fruitful to have 
those discussions

00:07:19.250 --> 00:07:21.400
and we had complimentary skills.

00:07:21.400 --> 00:07:24.821
I was very empirical. 
I'm not really an abstract thinker.

00:07:25.500 --> 00:07:29.800
Guido was great at figuring out 
what the principles were.

00:07:30.100 --> 00:07:34.500
Yeah, that's right and I totally
agree, kind of [inaudible].

00:07:34.700 --> 00:07:37.829
These are incredibly
important relationships

00:07:37.829 --> 00:07:42.400
and you see a lot of
people working together

00:07:42.400 --> 00:07:45.679
and not necessarily 
working very well

00:07:47.000 --> 00:07:51.990
and then it's very hard often 
to get out of this relationship.

00:07:52.900 --> 00:07:56.000
A good partnering is a
beautiful thing, like a marriage.

00:07:56.500 --> 00:07:58.500
It produces wonderful children,

00:07:59.500 --> 00:08:03.408
the fruits of the scholarship are
potentially wonderful

00:08:03.408 --> 00:08:07.666
and they exceed the capacity of
the partners to do it on their own

00:08:07.900 --> 00:08:10.842
but a bad co-authorship 
can be very destructive

00:08:10.842 --> 00:08:14.336
and time consuming and painful,
just like a bad marriage.

00:08:16.190 --> 00:08:18.922
Arguments may start about 
who did what when

00:08:18.922 --> 00:08:22.640
and intellectual property
type issues,

00:08:22.640 --> 00:08:24.700
especially when it when 
it goes a little sour

00:08:24.700 --> 00:08:27.602
and somebody thinks the other party
is not pulling their weight.

00:08:30.100 --> 00:08:32.070
There's more co-authorship
now in economics,

00:08:32.070 --> 00:08:34.250
I think that's been documented, 
much more.

00:08:34.250 --> 00:08:35.270
- [Guido] Yes.

00:08:35.270 --> 00:08:37.460
There's more teams
and there's larger teams

00:08:38.100 --> 00:08:40.651
and I think that's great, 
I love working on teams.

00:08:40.651 --> 00:08:43.070
We do work on schools 
with big teams.

00:08:43.070 --> 00:08:49.230
I work often with PI teammates 
like Parag Pathak and David Autor

00:08:49.230 --> 00:08:51.100
and then a team of 
graduate students,

00:08:51.100 --> 00:08:54.260
but I see that the students 
are not always,

00:08:54.260 --> 00:08:56.480
in some ways they're a little 
too promiscuous,

00:08:56.480 --> 00:08:58.700
in my view, in their partnering.

00:08:58.700 --> 00:09:00.420
They don't think it through.

00:09:01.180 --> 00:09:03.600
It's difficult to think it through.

00:09:03.930 --> 00:09:08.500
I think, for me, working
with people always has involved

00:09:08.500 --> 00:09:11.061
spending a lot of one-on-one
time with people,

00:09:11.431 --> 00:09:12.981
you need to figure out 
how they think

00:09:14.712 --> 00:09:17.465
and what kind of problems 
they're interested in

00:09:17.465 --> 00:09:19.546
and how they think about 
these problems,

00:09:19.546 --> 00:09:23.127
how they like to write, 
to make that--

00:09:23.600 --> 00:09:26.060
And it takes some maturity on
everybody's part.

00:09:26.060 --> 00:09:28.339
- Yes. Yes.
- [Isaiah] In what sense?

00:09:28.339 --> 00:09:30.700
Just in the sense of knowing 
what's going to work for them,

00:09:30.700 --> 00:09:32.640
knowing when things are
versus aren't working?

00:09:33.200 --> 00:09:36.131
- [Josh] Maturity in the
sense of having some judgment

00:09:36.600 --> 00:09:40.151
to be able to face it honestly,
if it's not going well,

00:09:40.151 --> 00:09:42.603
sometimes you have to have 
some difficult discussions.

00:09:43.583 --> 00:09:45.075
Is it worth continuing?

00:09:46.400 --> 00:09:48.795
"I was hoping you would do this, 
and you didn't,"

00:09:49.100 --> 00:09:51.560
maybe it turns out 
there's some feeling

00:09:51.560 --> 00:09:53.445
in the other direction, 
the same way.

00:09:54.155 --> 00:09:56.800
And Josh is very good
[chuckles]

00:09:56.800 --> 00:09:59.600
in the being honest,
part from the beginning,

00:10:00.000 --> 00:10:01.310
- [Josh] For better or worse.

00:10:01.310 --> 00:10:03.896
- [Guido] I would write this stuff 
and then I remember

00:10:04.600 --> 00:10:08.192
the first version of the paper 
with Rubin,

00:10:08.963 --> 00:10:10.590
Josh was in Israel at the time,

00:10:12.700 --> 00:10:14.165
Don and I were in Cambridge

00:10:14.165 --> 00:10:16.010
and so I would talk 
with Don regularly,

00:10:16.010 --> 00:10:18.600
but Don wasn't really doing
much writing in those days,

00:10:18.600 --> 00:10:21.801
I would write things 
and then I would fax them to Josh

00:10:21.801 --> 00:10:22.946
and they would come back,

00:10:22.946 --> 00:10:26.733
the first page
just one big cross, "No,"

00:10:26.733 --> 00:10:29.555
second page, one big line, "No"

00:10:30.570 --> 00:10:33.470
and that would go for awhile
but he still does that.

00:10:33.470 --> 00:10:36.800
I sent him the first draft 
of my Nobel lecture,

00:10:36.800 --> 00:10:39.309
and Josh goes, 
"No, no!"

00:10:40.330 --> 00:10:44.649
I've gotten some PDF comments 
like that from Josh, very helpful.

00:10:45.150 --> 00:10:46.720
Omit needless words.

00:10:47.390 --> 00:10:51.717
I have few co-authors
who are willing to do that.

00:10:54.280 --> 00:10:58.400
Especially as you get older, 
it's harder to put up with that.

00:10:59.300 --> 00:11:02.905
I would find it harder now to start
working with people who did that

00:11:03.810 --> 00:11:05.830
early on in a co-author 
relationship.

00:11:05.830 --> 00:11:08.900
It's also very hard because 
you need to have enough trust.

00:11:12.520 --> 00:11:16.559
Josh, for being willing 
to be very critical,

00:11:16.559 --> 00:11:20.470
he was also willing 
to admit being wrong.

00:11:20.470 --> 00:11:22.500
♪ [music] ♪

00:11:22.500 --> 00:11:25.370
- [Josh] But you have to be on
the lookout for good partners,

00:11:25.800 --> 00:11:28.010
somebody who can help you 
answer questions

00:11:28.010 --> 00:11:29.710
that you can't answer yourself.

00:11:30.200 --> 00:11:33.000
I think there's a natural tendency 
for people to gravitate

00:11:33.000 --> 00:11:37.365
to people who are similar 
in outlook and skills

00:11:37.779 --> 00:11:39.763
and that's not as useful.

00:11:40.220 --> 00:11:43.131
- [Guido] Josh is right, nowadays 
it's very tempting

00:11:43.131 --> 00:11:46.134
to find people who think 
about the same problems

00:11:46.134 --> 00:11:49.176
you're already thinking about,
who think along the same lines

00:11:49.730 --> 00:11:56.283
and that may not lead 
to very novel stuff.

00:11:58.500 --> 00:12:00.838
But at the same time finding people

00:12:00.838 --> 00:12:02.800
who actually have 
very different ideas,

00:12:02.800 --> 00:12:04.380
it's going to take a lot of time.

00:12:04.680 --> 00:12:07.570
Guido, you mentioned in passing 
how working with Josh

00:12:07.570 --> 00:12:10.270
has influenced how you do research,

00:12:10.270 --> 00:12:11.880
could you say a little more 
about that?

00:12:11.880 --> 00:12:14.430
I'd also be interested 
to hear from Josh,

00:12:14.430 --> 00:12:16.910
did working with Guido influence 
the way that you do research?

00:12:16.910 --> 00:12:19.460
- [Guido] Nowadays, I'm much 
more conscious of the fact that,

00:12:20.620 --> 00:12:21.990
for me, good economic research

00:12:21.990 --> 00:12:25.270
comes out of talking to people 
doing empirical work,

00:12:25.600 --> 00:12:29.300
and it's really not reading 
econometrica

00:12:30.020 --> 00:12:31.420
or the reading the stats journals,

00:12:31.420 --> 00:12:34.550
but it's actually talking to people
doing empirical work,

00:12:34.550 --> 00:12:37.034
going to the empirical seminars.

00:12:38.770 --> 00:12:40.400
When I was at Berkeley,

00:12:41.250 --> 00:12:44.167
David Carr and Raj Chetty, 
as colleagues there

00:12:44.780 --> 00:12:46.700
and I would talk to them 
and listen to them,

00:12:46.900 --> 00:12:48.200
trying to figure out

00:12:48.601 --> 00:12:53.534
how are they solving their problems
and other things there

00:12:53.534 --> 00:12:57.111
where I'm not really quite happy 
with the way they're doing things

00:12:57.111 --> 00:13:01.097
and trying to look for 
methodological problems,

00:13:02.710 --> 00:13:07.532
where there's some more 
general solutions possible.

00:13:07.900 --> 00:13:11.700
I tried to tell it to my students
that I encourage them to work

00:13:11.700 --> 00:13:13.199
as research assistants also,

00:13:13.199 --> 00:13:18.704
for the people doing empirical work
at Stanford.

00:13:19.530 --> 00:13:22.190
There was no subbing what I learned
while I was in graduate school,

00:13:22.190 --> 00:13:25.600
but it really came out of 
working with Josh,

00:13:25.600 --> 00:13:26.990
as well as talking to Gary,

00:13:26.990 --> 00:13:30.367
Gary was always encouraging 
of doing that

00:13:30.367 --> 00:13:31.860
and because he done that himself,

00:13:31.860 --> 00:13:36.210
he'd worked with on empirical 
problems with Zvi Griliches

00:13:37.810 --> 00:13:39.500
early in his career.

00:13:39.500 --> 00:13:40.500
Yeah.

00:13:40.500 --> 00:13:44.600
Well, I became more more interested
in the econometric theory

00:13:45.410 --> 00:13:47.100
through our interaction,

00:13:47.100 --> 00:13:51.960
and I think empiricists are often 
impatient with econometric theory,

00:13:51.960 --> 00:13:55.040
partly because empirical work
is very time-consuming,

00:13:55.040 --> 00:13:59.100
and you may have a sense 
that something is

00:13:59.300 --> 00:14:00.740
convincing and sensible

00:14:00.740 --> 00:14:04.100
and you haven't really fully 
made the case for that,

00:14:04.100 --> 00:14:05.100
but you're convinced

00:14:05.100 --> 00:14:07.005
and that motivates you 
to pursue it,

00:14:07.860 --> 00:14:09.729
like the draft lottery story.

00:14:10.604 --> 00:14:12.441
I was pretty sure that was
worth doing

00:14:14.420 --> 00:14:19.613
and I came away from 
working with Guido

00:14:19.613 --> 00:14:22.348
seeing that there was 
the potential to say something

00:14:22.733 --> 00:14:25.220
more than just about 
that particular problem,

00:14:26.560 --> 00:14:29.307
and I think over the those early
years in the 90s,

00:14:31.600 --> 00:14:33.360
our thinking evolved together

00:14:33.360 --> 00:14:35.500
that there's actually 
a framework here,

00:14:35.500 --> 00:14:37.270
a way to solve a lot of problems

00:14:37.270 --> 00:14:40.700
and I think that that is the power
of the LATE framework,

00:14:40.700 --> 00:14:42.630
is it answers a lot of questions.

00:14:42.630 --> 00:14:44.500
♪ [music] ♪

00:14:44.500 --> 00:14:46.300
- [Isaiah] In some sense, 
did you find that,

00:14:46.300 --> 00:14:48.560
email versus facts 
versus in-person,

00:14:48.560 --> 00:14:51.462
the medium mattered
to how collaboration went

00:14:51.462 --> 00:14:53.190
or they're ways that you felt like

00:14:53.190 --> 00:14:54.760
it was the most useful 
to collaborate?

00:14:55.300 --> 00:14:57.361
To me, I think 
what matters most is,

00:14:58.251 --> 00:15:00.588
initially you have a period of--

00:15:00.588 --> 00:15:02.150
We needed that initial period,

00:15:02.150 --> 00:15:06.477
that was very intense with 
almost daily interaction

00:15:06.477 --> 00:15:08.358
and we also became friends.

00:15:08.900 --> 00:15:13.512
You don't develop the kind of 
friendship, electronically usually

00:15:13.512 --> 00:15:15.000
[laughter]

00:15:15.000 --> 00:15:18.731
but once you have that foundation 
you can be pen pals

00:15:19.300 --> 00:15:24.314
and we did use email, 
though it wasn't as useful then

00:15:25.500 --> 00:15:28.400
but it worked, 
we definitely had a lot of faxes.

00:15:28.400 --> 00:15:31.300
I still have these faxes, 
long faxes

00:15:33.330 --> 00:15:36.885
and then in the summer, 
I would come to Cambridge,

00:15:36.885 --> 00:15:40.300
usually to the NBR meetings 
and hang around for a few weeks

00:15:40.300 --> 00:15:41.949
and you visited me in Israel.

00:15:41.949 --> 00:15:44.000
I visited in Israel.

00:15:44.000 --> 00:15:48.400
But yeah, there was good foundation
from that that year

00:15:49.000 --> 00:15:51.100
and in some sense that was enough.

00:15:51.800 --> 00:15:53.300
Nowadays,

00:15:53.300 --> 00:15:56.600
I have the co-authors
in lots of different places,

00:15:56.600 --> 00:15:59.100
but it's always been important

00:15:59.100 --> 00:16:01.380
to spend some time with people 
in the same place,

00:16:01.380 --> 00:16:04.900
so you understand how they work, 
how they think,

00:16:05.800 --> 00:16:07.250
even to the point that,

00:16:07.250 --> 00:16:08.690
you know when 
they actually respond,

00:16:08.690 --> 00:16:09.813
whether they respond quickly

00:16:09.813 --> 00:16:10.831
or whether that means

00:16:10.831 --> 00:16:12.240
they're not actually
doing anything,

00:16:12.240 --> 00:16:15.100
or that mean they're thinking hard 
about a problem

00:16:15.100 --> 00:16:17.300
and they just take longer,

00:16:18.020 --> 00:16:21.707
but you do need to
develop some understanding there.

00:16:21.707 --> 00:16:23.814
♪ [music] ♪

00:16:23.814 --> 00:16:26.900
- [Isaiah] We've talked about 
how your collaboration started,

00:16:26.900 --> 00:16:31.000
maybe just to step back slightly,
were they're sort of features about

00:16:31.000 --> 00:16:34.000
the environment at Harvard 
or in Cambridge, at the time,

00:16:34.000 --> 00:16:35.980
which you felt contributed to it?

00:16:36.290 --> 00:16:37.920
Coming from Brown,

00:16:37.920 --> 00:16:40.047
I felt it was very 
intimidating place

00:16:40.047 --> 00:16:45.200
because it clearly was a very,
very impressive set of people.

00:16:46.590 --> 00:16:48.620
Zvi Griliches was there, 
Dale Jorgensen--

00:16:49.350 --> 00:16:53.610
Gary, Jerry Hausman, Whitney Newey,
sometimes Jamie Robins.

00:16:53.890 --> 00:16:55.900
I mean, my view of that 
in retrospect,

00:16:55.900 --> 00:16:58.000
I can't say I loved every
minute of every talk

00:16:58.000 --> 00:16:59.270
I ever gave in that Workshop,

00:16:59.270 --> 00:17:00.990
but that was the highest powered,

00:17:00.990 --> 00:17:02.670
that was the group 
you wanted to reach

00:17:03.600 --> 00:17:07.664
and you would get extraordinarily 
insightful feedback,

00:17:07.664 --> 00:17:10.600
even if it wasn't always 
easy to swallow.

00:17:11.300 --> 00:17:12.500
Yeah, and I have for a while,

00:17:12.500 --> 00:17:15.940
I would basically give 
a talk every semester

00:17:15.940 --> 00:17:19.000
because we didn't have 
any money to invite people.

00:17:19.710 --> 00:17:22.000
Gary would say, 
"Well, why don't you give a talk?"

00:17:22.000 --> 00:17:23.260
[laughter]

00:17:26.800 --> 00:17:30.758
That was the arena for young people
with our interest.

00:17:30.758 --> 00:17:33.098
- [Guido] Yeah, it was really 
very impressive,

00:17:33.098 --> 00:17:35.448
but it was also quite tough--

00:17:35.448 --> 00:17:37.128
It was intimidating.

00:17:37.800 --> 00:17:41.000
People there had very strong
views on what they thought was

00:17:42.810 --> 00:17:44.310
the way you should do econometrics,

00:17:44.310 --> 00:17:46.310
the way the direction 
things should go,

00:17:49.340 --> 00:17:51.710
now, I would think things were
getting a little stale,

00:17:51.710 --> 00:17:55.550
that in fact, we were bringing in 
a lot of the new ideas...

00:17:55.550 --> 00:17:57.000
- [Josh] Yeah.

00:17:57.000 --> 00:18:01.530
...and that wasn't necessary
immediately appreciated.

00:18:01.530 --> 00:18:02.530
[laughter]

00:18:02.530 --> 00:18:04.300
- [Josh] But that's okay.
- And that's fine.

00:18:04.300 --> 00:18:07.730
We were pushed 
and a lot of great discussions

00:18:07.730 --> 00:18:12.887
in that workshop about 
what should we make of LATE

00:18:12.887 --> 00:18:15.687
but there were other questions
that were just as interesting,

00:18:15.687 --> 00:18:18.267
like the role of 
the propensity score,

00:18:18.267 --> 00:18:20.678
that was a big deal in the 90s

00:18:20.689 --> 00:18:24.300
and econometrics was 
moving towards that

00:18:25.000 --> 00:18:27.550
and there were a lot 
of great questions.

00:18:27.550 --> 00:18:28.790
Yeah,

00:18:29.430 --> 00:18:32.810
I learned a huge amount
there from the time I spent--

00:18:32.810 --> 00:18:34.800
- [Josh] I think the other thing 
that Guido and I

00:18:34.800 --> 00:18:36.800
both benefited from is we both,

00:18:37.400 --> 00:18:39.810
not at the same time, 
but in early in our careers,

00:18:39.810 --> 00:18:42.400
taught econometrics 
with Gary Chamberlain,

00:18:43.200 --> 00:18:46.500
and that was like 
an apprenticeship for us, I think.

00:18:46.800 --> 00:18:50.350
I taught a mixed graduate,
undergrad 1126,

00:18:50.350 --> 00:18:52.210
I don't know if they still have 
that number...

00:18:52.210 --> 00:18:53.500
- [Isaiah] Mmhmm, they do.

00:18:53.500 --> 00:18:55.480
...very interesting course 
that it had

00:18:55.480 --> 00:18:57.885
both graduate 
and undergraduate enrollment

00:18:58.800 --> 00:19:02.869
and it was relatively applied 
for an econometrics class,

00:19:03.270 --> 00:19:06.542
and I learned a lot by teaching 
that with Gary.

00:19:07.500 --> 00:19:11.618
But in that sense, 
Harvard was a great place,

00:19:11.990 --> 00:19:13.210
very flexible there.

00:19:13.600 --> 00:19:15.879
The other thing I remember 
about Harvard is,

00:19:16.710 --> 00:19:18.612
well I had very good students,

00:19:20.300 --> 00:19:22.672
I taught a lot of 
wonderful students

00:19:22.672 --> 00:19:24.995
who went on to have 
wonderful careers.

00:19:26.300 --> 00:19:28.346
Also, Harvard as an institution,

00:19:28.346 --> 00:19:31.400
you're probably are aware of this, 
Isaiah,

00:19:31.400 --> 00:19:35.350
as a junior faculty member, 
they didn't then ask much of us,

00:19:35.350 --> 00:19:37.140
other than teaching our classes.

00:19:37.640 --> 00:19:40.730
We didn't have administrative
concerns, to speak of.

00:19:41.300 --> 00:19:43.757
I think I went to two
faculty meetings

00:19:43.757 --> 00:19:45.356
in my two years at Harvard

00:19:46.600 --> 00:19:48.940
and so we're left--

00:19:50.920 --> 00:19:53.400
You were given a lot of freedom 
and flexibility.

00:19:54.740 --> 00:19:56.011
I went to the chair said,

00:19:56.011 --> 00:19:58.022
"Can I teach this course
with Rubin?"

00:19:59.620 --> 00:20:04.133
I think it was Friedman
at the time. It was like, "Fine."

00:20:05.200 --> 00:20:09.193
It wasn't really any concern about 
what what it was about

00:20:09.193 --> 00:20:11.790
and again, that was a very 
intimidating experience,

00:20:11.790 --> 00:20:13.350
but it was a great experience.

00:20:13.350 --> 00:20:14.375
♪ [music] ♪

00:20:14.375 --> 00:20:15.990
- [Narrator] If you'd like to 
watch more

00:20:15.990 --> 00:20:18.198
Nobel Conversations, 
click here,

00:20:18.198 --> 00:20:20.516
or if you'd like to learn more 
about econometrics,

00:20:20.516 --> 00:20:23.290
check out Josh's 
"Mastering Econometrics" series.

00:20:23.845 --> 00:20:26.713
If you'd like to learn more about 
Guido, Josh, and Isaiah,

00:20:26.713 --> 00:20:28.514
check out the links 
in the description.

00:20:28.514 --> 00:20:30.347
♪ [music] ♪