YT YouTube2Text

YouTube2Text → Transcripts → Supplemental Lecture: Human capital in developing countries

Supplemental Lecture: Human capital in developing countries — Transcript

by CEPR & VideoVox Economics · 18,649 words · 3,152 segments · language en · Watch on YouTube

Open in transcriber Summarize with AI Download SRT/VTT

Full transcript

0:02when we get started here
0:05um thank you for coming um
0:08to this third lecture uh first lecture
0:11on thursday the first supplemental
0:13lecture
0:14of the class um my name is joe kobaski
0:18i'm uh one of the co-leaders with doug
0:20golin of the steg program
0:22this course is being offered
0:26by organized through the steg program
0:28which is supported by the united
0:30kingdom's foreign commonwealth and
0:31development office
0:33as part of the uk effort
0:37we have people in the front row as
0:39panelists
0:41you can interrupt use the microphone we
0:43encourage you to interrupt with
0:45questions
0:45just as in a normal classroom but
0:48otherwise if you're not interrupting
0:49please mute yourselves
0:52those of you who are viewing as a
0:54webinar you can address questions in the
0:56q
0:57a you can upvote them we'll be trying to
1:00um answer those as we can
1:04and if there's you know particular
1:05popular questions uh
1:07um todd shulman will um leave some time
1:12at the end
1:13to answer the questions uh the lecture
1:16notes
1:16are posted at the stag website uh there
1:19will be a youtube recording this is also
1:21being broadcast
1:22via youtube so you can go back and watch
1:24it on youtube i want to thank todd
1:26schulman from the minneapolis fed for
1:29teaching the class today
1:31again he's doing this out of the
1:32kindness of his heart i want to thank
1:34mandy and kirsty from cepr for
1:39everything that they've done
1:40for logistics and um i want to announce
1:43that
1:44tomorrow we have uh bertolt herrendorf
1:47who will be talking about structural
1:48transformation at the same time
1:514 p.m london time um okay todd go ahead
1:56all right thanks i want to say first
1:58thanks
2:00to both of you for organizing this whole
2:02enterprise and and
2:03these lectures in particular and for
2:06inviting me today so what i'm supposed
2:08to talk about is
2:09human capital in developing countries
2:12since i work for the fed i have to stop
2:13and make sure you all see here that
2:15these are my views about human capital
2:16in developing countries and not the
2:18federal reserve banks
2:22so i like to start by just making sure
2:24we're on the same page i think we often
2:26talk about human capital but here's
2:27going to be my formal definition of it
2:29i want to think about it as being the
2:31stock of productivity enhancing
2:33attributes embedded in people
2:35and for clarity i want to make sort of
2:37two points right up front
2:39the first is that in general we should
2:40think about this as being a vector as
2:42being a multi-dimensional stock it
2:44includes a wide variety of attributes
2:46it includes the knowledge of workers
2:48their skills the habits they develop
2:50over the course of their life their
2:52health
2:52which helps them be more productive on
2:54the job so i want to think about this as
2:56being a long vector of attributes that
2:58people have
2:59the second kind of key attribute
3:02component of human capital is that it's
3:03generally produced through investments
3:06okay so there's some active uh
3:09expenditure of time or resource that
3:10goes into enhancing human capital
3:13for example education is used to produce
3:16knowledge
3:16training or practice to produce skills
3:18and habits exercise to improve your
3:20health etc etc etc
3:23i think it's really common when we talk
3:24about human capital to jumble all of
3:26this stuff together
3:28primarily to be honest for measurement
3:30reasons we can't really
3:32measure human capital and often we can't
3:34really measure your knowledge or your
3:36skills exactly
3:37what we find it easiest to measure is
3:39how long you spent sitting in classrooms
3:41how many years you went to school
3:43maybe we can measure how many minutes a
3:45day you exercise even if we can't
3:46measure your health
3:47but at least at the outset i want you to
3:49kind of be clear these are distinct
3:50objects
3:52human capital has a long history in
3:54economics although it wasn't always
3:55referred to by that name if you go
3:57all the way back to adam smith you'll
3:58see he talks about the skills of workers
4:00and how
4:01workers with different skill levels have
4:02different productivities
4:04that's the same idea but it was really
4:06brought to the forefront of economics in
4:08the late 50s and early 1960s when
4:11becker minster and others started to
4:13work out a full theory of human capital
4:16what i mean by a full theory of it is
4:17that they have a models
4:19where workers or their families are
4:22making investments and they have
4:24rational cost benefit type
4:25considerations you know how long should
4:27i go to school
4:28i should go until the return on going to
4:30school is equal to the return on other
4:32activities i might undertake
4:33things like that okay from that time
4:36there's been a huge explosion in the
4:38literature
4:38and there's you know a great deal of
4:40work on human capital
4:42in general what i want to do today
4:46since there's been so much work because
4:47i'm going to focus on two very
4:49specific things that i think fit well
4:51with the theme of macro development and
4:53stag and
4:54that you're working through the first
4:57thing i want to do is measure
4:59as best we can or really estimate might
5:01be a better word
5:02cross-country differences in the stock
5:03of human capital
5:05when i do that i'm going to try to
5:06convince you of two things i think
5:08there's sort of
5:09two emerging uh there is an emerging
5:11consensus here which is that
5:13cross-country differences and human
5:14capital stocks are large
5:17that's a statement that has to be
5:18relative to something and i'm going to
5:20think of it as
5:20human capital stocks or large
5:22differences in human capital
5:24stocks are larger than what the early
5:27consensus held
5:28the early consensus would be like what
5:30julieta taught last week
5:31and i'll try to be more precise about
5:33that in a minute
5:35the second is i'm going to say that it's
5:36large relative to cross-country income
5:38differences
5:39that is to say that human capital plays
5:41an important role
5:42in thinking about why rich countries are
5:44much richer than poor ones
5:46there's relatively less evidence on
5:50the role that it plays for example in
5:52structural transformation
5:53i know that's a key theme of stag so
5:55i'll try to highlight a few references
5:56and
5:57some of the work that's being done in
5:58that area but i think that's an area
6:00that we need a lot more work on
6:03since i'm going to be focusing on these
6:04cross-country issues
6:06in other words comparing human capital
6:08stocks between rich and poor countries
6:10or middle income and poor countries
6:12i'm going to have relatively less to say
6:14on the vast and excellent literature on
6:16within country human capital differences
6:18that's just a function of how much time
6:19we have today i don't have time to get
6:21into it
6:22do you take as given that um there are
6:25large differences in human capital
6:26stocks
6:27across countries or across sectors
6:29within a country
6:30then as somebody who works on macro
6:32development or who's interested in
6:33structural transformation you might want
6:35to build human capital into your
6:36thinking and
6:37one way that we we of course make
6:39progress here is by building models
6:41and so i'm going to try to give you some
6:42ideas the second thing i want to do is
6:44illustrate the key model mechanisms that
6:46you can use to think about human capital
6:48differences across sectors and across
6:49countries
6:51so i'm going to build on some very
6:52classic results especially from becker
6:54and ben parath i'm going to show you
6:56three maybe four if i have enough time
6:57of the main mechanisms that people use
6:59to generate human capital differences
7:02as the results of workers and their
7:04families endogenously deciding how much
7:06human capital to acquire that's kind of
7:08the
7:08underlying idea
7:13i think it's always useful to at the
7:15outset say sort of what's the underlying
7:17challenge that we confront
7:19and with human capital i think it's
7:20really a big one and i like to pose it
7:23as a simple thought experiment which is
7:25imagine i ask you
7:26even pick within a country take two
7:29workers
7:30imagine i ask you to explain why some of
7:32those workers are more productive than
7:33others what drives productivity
7:34differences between workers
7:37even if you focus on a country like i'm
7:39an american i live in the united states
7:41you know why are some americans more
7:43productive than others
7:45i'm hard-pressed to tell you and i don't
7:47think the economic literature has
7:48settled on an answer how much of that is
7:50human capital
7:51versus how much of its other
7:52characteristics like luck or the job you
7:55type
7:56you happen to work in or how much of
7:57it's driven by things like
7:58discrimination
7:59you know those are questions that people
8:01are wrestling with
8:03now in other literatures and i don't
8:05think they're settled issues
8:07even if we agree that it's human capital
8:09i think we're also going to have a hard
8:10time putting our finger on exactly what
8:12it is
8:13and an easy way i can say this to you is
8:15that i have the good fortune to work
8:17with ed prescott
8:19ed is like me a phd educated economist
8:23so we have the same job we have the same
8:25employer
8:26we have the same education level and yet
8:29ed is vastly more productive than i am
8:31at producing economics papers why is
8:34that i think we can all agree it's
8:35something that ed has it's some human
8:37capital characteristics that he has that
8:38i don't have but if i ask you to put
8:41your finger on exactly what those are
8:42to measure them you would have a hard
8:44time doing that
8:46what we're going to do today is
8:47something that's even tougher than that
8:49we're going to try to investigate these
8:50differences across countries
8:52the set of relevant human capital
8:54characteristics that we might be
8:55interested in is going to be even larger
8:58and the set of confounding or
8:59environmental factors that we need to
9:00think about and control for is even
9:02larger
9:03and so in general i want you to just
9:05acknowledge right up front
9:06measuring cross-country differences in
9:08human capital stocks is really tough
9:11i'm going to give you my best guess you
9:12know the best approaches that i think
9:13we've discovered so far i'm going to
9:15give you the answers that they lead to
9:17but i want you to be throughout this
9:19vaguely dissatisfied
9:20because i think a vague and persistent
9:22sense of dissatisfaction
9:24is a great tool for research you know if
9:25you're never quite happy with what's
9:27been done
9:27and you're always a little uncomfortable
9:29with the assumptions that's a great way
9:31to get started in your own work
9:33all right so let's start with part one
9:35how do we measure human capital
9:36differences
9:41the techniques i'm going to show you two
9:43different ways to go about doing this
9:44and they're both going to be closely
9:46tied to development accounting
9:48so just like julieta's lecture last
9:50friday so she's done a great
9:51introduction to this
9:53i'm just going to remind you a little
9:54bit of the basics that i need and then
9:56we'll get going
9:58so for development accounting what you
10:00really need the main assumption you
10:01start with is about the aggregate
10:03production function
10:04so i'm going to think about capital y
10:06that's output
10:07that's ppp adjusted gdp of country c
10:11as being a cobb douglas production
10:13function of the physical capital
10:15measured in ppp adjusted dollars raised
10:19to the capital share
10:20and then tfp total factor productivity
10:24and the total human capital stock of
10:25country c
10:27okay and julieta walked you through last
10:29time we think that we have good evidence
10:31that this production function
10:32provides a useful fit to the data we
10:34think we roughly know what these
10:36factor shares the alphas are so this is
10:38sort of a good starting point for us
10:41just like she did i'm going to divide
10:43both sides through by the stock of
10:45workers in the country
10:46so then i can think about gdp per worker
10:48in country c
10:50and that's going to be determined by
10:51three components that she talked about
10:53the capital to output ratio raised to a
10:55power
10:57total factor productivity and the human
10:59capital stock per worker
11:01people will sometimes call this last
11:03piece the average human capital you do
11:05actually technically need extra
11:06assumptions to say that
11:07i may slip and call it that myself but
11:09anyway human capital per worker
11:13so generically in the development
11:15accounting literature what we're trying
11:16to do is
11:17understand differences or variation in
11:19this gdp per worker as a function of
11:22how much capital to output every country
11:23has the tfp in every country
11:26and the human capital per work in every
11:28country
11:29today i'm going to be really focused on
11:31what can we say about human capital
11:33so i'm going to lump the rest of it
11:35together and just call it z
11:37other stuff that i'm not interested i'm
11:38not going to try to measure i have
11:39nothing to add there
11:41and so we're just going to ask what can
11:43we say about human capital
11:47like julieta did i'm going to take logs
11:48of this equation so basically we're
11:50going to investigate
11:51variation or differences in log gdp ppp
11:54adjusted gdp per worker
11:56that's pen world tables 10.0 that's
11:59going to be in part a function of
12:00capital output and tfp differences
12:03and part human capital per worker
12:05differences
12:06okay so that's just to get you back to
12:08where julieta left us and what we're
12:09going to do now is consider what what is
12:11sort of the
12:12the best evidence about how we can
12:14measure this object h
12:16and think about its contribution to the
12:18ppp gdp per worker difference as we see
12:23i've already established that that's a
12:24challenge and and julieta did a nice job
12:26with this but
12:27i do want to take one extra minute and
12:28point out that regressions are probably
12:30not going to be the way that we
12:32identify or measure h or its importance
12:34for cross-country income differences
12:37she gave you one perspective on this
12:38which is that you should think about
12:39this as being a production function and
12:41the estimation of production functions
12:43is a very difficult task
12:46another way to think about this is that
12:48i like to borrow from
12:50mancus bookings papers
12:53he has a very nice discussion of the
12:55problems with cross-country regressions
12:57and his view is that this is really a
12:59very simple and practical problem
13:02at the end of the day if i want to
13:03investigate human capital using
13:05cross-country
13:06regressions how many countries do i have
13:08something like a hundred
13:09is typically how many countries you're
13:10going to be able to get data for
13:12and if i want to think about
13:13cross-country differences you know how
13:15many right hand side variables would i
13:17want to include how many things do i
13:18think might generate cross-country
13:20income differences including human
13:21capital
13:22skills capital institutions etc etc
13:26and his point that he makes very nicely
13:28is just that if you only have 100
13:29observations and you have hundreds of
13:32things you might think about as being
13:33determinants of
13:34gdp per worker it's kind of a hopeless
13:36endeavor
13:38and so we need to find a way to
13:39circumvent this we need to find a way so
13:41that we can have more than just 100
13:42observations to help us learn about
13:45human capital so how are we going to do
13:48that
13:52sorry 10 a.m here so i still need coffee
13:55um we're going to build on two
13:57assumptions that are going to allow us
13:59to make substantial progress and this
14:01was first worked out by bills and clean
14:02out a 2000 aer paper
14:05those two assumptions i'll just state
14:07them first and then we'll talk a little
14:08bit about their significance
14:10the first is that there's perfect
14:11substitution among different labor types
14:14which is sometimes called the efficiency
14:15units assumption
14:17what this assumption means is that for
14:19the purposes of producing
14:20aggregate output we all supply
14:23labor there's only one input called
14:25labor
14:27some of us just happen to provide more
14:28or less than others
14:30so you know the way to think about this
14:32is okay if you
14:33add up enough todd shulman's eventually
14:35you can have a bill gates and we just
14:37have to figure out how many todd
14:38shulman's it takes
14:39to provide the labor input of the bill
14:41gates or whoever your famous
14:42your favorite famous person is but there
14:45exists some number and we know what that
14:46is we can solve it
14:48okay the second assumption is that labor
14:51markets are competitive
14:53and what that really means is that
14:55workers and firms take prices as given
14:57or that prices are set by marginal
14:58products
15:00okay under these two assumptions
15:03we can use the firm's first order
15:05conditions to characterize the demand
15:07for workers who have hi units of human
15:10capital in country c
15:12so the competitive labor markets
15:13assumption we're going to use that first
15:15to say that the wage of worker i in
15:17country c is equal to their marginal
15:19product of their labor
15:22then we're going to take the derivative
15:24we're going to just find what that
15:25marginal
15:26product is and it's going to depend on
15:27the labor share the capital output ratio
15:30tfp and the human capital of the worker
15:33and here's where the perfect
15:35substitution assumption shows up
15:37it says that the marginal product of
15:39labor depends on how many units of human
15:41capital they provide
15:43in a world where there's not perfect
15:45substitution but there's different types
15:46of labor there's an
15:47extra term that shows up here which is
15:49the price of my type of labor
15:51so for example if you're used to
15:53thinking about models where there's
15:54skilled and unskilled labor and they're
15:56different inputs maybe with a ces
15:57production function
15:59then you would have a term here that
16:00says what's the price of skilled or
16:02unskilled labor
16:04the fact that we don't have that comes
16:05from the perfect substitution assumption
16:08so that's what those two assumptions buy
16:10us
16:11if you stare closely at this equation
16:13you'll see that there's the capital
16:14output
16:15in t of p terms just like before that's
16:16what i labeled z
16:18on the last slide so we've got that same
16:19z that's what this really tells us is
16:22that the wage of worker i in country c
16:24depends on that country stuff
16:26and the workers human capital
16:30i find it most useful to explain
16:34the power of these assumptions by
16:35turning this whole thing around
16:39what this says is that if a worker has
16:42twice as much human capital they're
16:43going to make twice as
16:44twice as high of a wage but in practice
16:47what we really want to do is
16:48invert that and use it the other way
16:50around we want to say if i see a worker
16:53making twice as high wages they must
16:55have twice the human capital
16:58and the reason that's really useful is
17:00on the last slide i was saying boy if we
17:02want to measure human capital using
17:03cross-country evidence we only have 100
17:05or so countries maybe 200 at the most
17:08here i'm saying well if i see a worker
17:10with twice as high of wages i know they
17:12have twice the human capital
17:14now i'm using evidence on workers wages
17:16how many observations do i have on the
17:18wages of workers
17:20billions at this point it really is
17:22billions of observations
17:23and so this is going to give me a lot
17:25more evidence it's going to allow me to
17:26draw on a lot more data
17:28to discipline the measurement of human
17:30capital
17:32and now you can really think about okay
17:33so what i need what i want to be able to
17:35do i said
17:37is i want to be able to say if i see a
17:38worker who has twice as high of wages
17:41he must have twice the human capital now
17:44you can really see what those
17:45assumptions are doing for me
17:47if there was imperfect substitution
17:48across labor types that inference no
17:50longer holds because i have to worry
17:52well maybe he's just a different type of
17:53worker and that type of worker has a
17:55different price
17:57if labor markets are not competitive
17:58then the fact that a worker makes twice
18:00the wage could just be due to
18:02discrimination or market power
18:05or luck or all sorts of other factors
18:08okay
18:09so that's what those assumptions bias
18:10they're both really important
18:11assumptions i'll come back and talk a
18:13little bit about them
18:14later on i don't have time to dig into
18:16the whole literature on these
18:17today but i think i want you to
18:19understand that's what they get us
18:21they allow us to take this from being a
18:23cross-country problem
18:24to a cross-worker problem and that's
18:26really what's unlocked the measurements
18:28of human capital
18:30sorry todd can i ask a question on this
18:32yes please
18:33so just irrespective of the two
18:35assumptions we are already
18:37we also make an assumption on kind of
18:39ruling out human capital externalities
18:41here right
18:42we are not as we are already assuming
18:44that h
18:46aggregate h doesn't have a direct effect
18:48on a right on on z
18:54yes there is a small literature that
18:57tries to measure those human capital
18:59externalities and i'm going to skip it
19:00today
19:01and and there's really it's a practical
19:03thing um
19:05it is extremely difficult to measure
19:07externalities
19:08if you think about what i'm doing here
19:10the underlying idea is that i'm saying
19:12uh i can learn about the productivity of
19:14a worker from looking at their wages and
19:15their marginal products
19:17and sort of the defining feature of
19:18externalities is of course that they're
19:20not priced in the market which makes
19:22them hard to measure through this
19:23approach
19:24so i can point you to the couple of
19:26papers that i think have made the best
19:27progress on that that is a
19:28much trickier issue and one that i'm not
19:30going to get into today
19:34thank you all right
19:38so suppose i take this equation very
19:40simple equation right and i just take it
19:42in logs
19:44then i have a very nice log linear wage
19:46equation which says that the log wage of
19:48worker i and country c
19:50is additive in two terms there's a part
19:54that depends on the country you live in
19:55that captures tfp and capital output
19:57ratios
19:59and there's a part that depends on your
20:00human capital
20:03what's really nice about this is if you
20:05lay it up next to
20:07okay let me do this differently there's
20:09a very nice connection here
20:11to the enormous labor literature in
20:14particular the micro labor literature
20:16that starts with jacob mintzer
20:17and his 1974 book in that book
20:21minser argues that log wages in the
20:24united states
20:26are approximately linear in schooling
20:28and quadratic in experience
20:30so he investigates a whole sets of
20:32different ways to model wages in the
20:34united states
20:35and he says the best way to model wages
20:37is to think about them as being
20:39the log wages linear and schooling
20:40quadratic in experience
20:42this was enormously influential so tons
20:44of people have estimated
20:46equations of this form which is now just
20:48called the mincer wage equation
20:50not just in the united states but in
20:52countries all over the world
20:54and so what's really another neat
20:56feature about these two assumptions is
20:58that they
20:59allow us to have a wage equation that
21:01looks very
21:02much like what people have been doing in
21:04micro labor for
21:0540 50 years so if you kind of stare at
21:08those there's a bunch of connections i
21:10want you to see
21:12if you estimate a wage equation of this
21:14type within a country
21:16you'll see that the intercept captures
21:19all of these country specific effects
21:24okay so that's the first piece of
21:25connection the second thing is that of
21:28course this huge labor literature that's
21:29been
21:30estimating this minster wage equation
21:31they're including things like schooling
21:33and experience on the job those
21:36sound like attributes of human capital
21:38those sound like things we want to
21:39include in our measure of human capital
21:41so
21:41they have been estimating exactly these
21:44types of
21:44ideas and what they're doing when they
21:48estimate them these coefficients beta 1
21:50beta 2 beta 3 they speak to the value
21:54of those things now the way somebody in
21:57the minster literature thinks about this
21:58is they tell you the wage value how much
22:00does it increase wages
22:02you know how much does a year of
22:03schooling increase my wages that's what
22:05beta 1 picks up
22:07but under the assumptions i've made
22:08those also tell me how much more
22:09productive a year of schooling makes me
22:13and so those connections says that now i
22:15can bring to bear all of the evidence
22:17from this minster wage
22:18literature not just in the united states
22:20but really around the world at this
22:22point
22:23to tell me what sorts of characteristics
22:26of human capital i might want to include
22:29and what are the value
22:30of those characteristics and in this
22:32specification
22:34is experience just measured in years of
22:36experience
22:37yeah i'll be more precise in a second
22:39what almost everyone ends up doing is
22:41something called potential experience
22:44which is the years since you graduated
22:46school
22:48and then you're going to do a quadratic
22:50people i mean the literature is built
22:51from that over the years but that's
22:53really going to be the the
22:54starting point for everything okay thank
22:57you
22:57um excuse me so um
23:00by looking at this equation i can
23:01imagine that this is going to have
23:04huge errors terms so
23:07i'm i'm wondering how well this um
23:11like human capital factors are actually
23:12going to explain the differences in
23:14wages
23:15i'm just saying like if these factors
23:17explain only 10 percent of the
23:18dispersion in wages maybe it's not
23:20i mean they might be contributing
23:22factors but they're not going to give us
23:23a good measure
23:24that's a great point it's not so much
23:27about the standard errors it's about the
23:28r squared in a typical application the r
23:32squared of this regression will be 20 to
23:3430
23:34meaning you explain about 20 to 30
23:36percent of the variation in log wages in
23:38the sample
23:41and you're exactly right now you have to
23:42think okay but then what
23:45one hypothesis is that there's non-human
23:48capital stuff going on that drives wage
23:50variation
23:51right block the employer you happen to
23:54meet with discrimination etc
23:57another is that maybe i'm doing a lousy
23:58job of measuring human capital
24:02and i don't know that we've ever figured
24:03out exactly what it is but those 20 to
24:0530 percent r squared's you know they are
24:07a bit worrying in that sense and i
24:08agree completely yeah
24:12okay so these really tight connections
24:14are going to lead us to the to the first
24:15approach to measuring human capital
24:17stocks which i'm going to call the
24:18constructive approach
24:20and i don't mean that as opposed to
24:21destructive but i mean it to point out
24:23that you were literally going to
24:24construct human capital piece by piece
24:26with this approach so what you're going
24:27to do
24:28it's just a simple four-step procedure
24:31you're going to select
24:32some dimension of human capital that
24:34you're going to try to measure so think
24:35about it is i'm going to select
24:36schooling
24:38i'm going to measure each nation's stock
24:41how much schooling does every country
24:43have
24:45then i have all of this evidence from
24:46the menstrual waves equations about the
24:48value of schooling
24:50so if i know how much schooling every
24:51country has and i know how valuable it
24:53is how much it raises productivity
24:55then i can use that to aggregate and
24:57find a measure of human capital
24:59from schooling and i can in theory do
25:01this for a whole variety of attributes
25:03and so that's what i want to build on
25:04first and show you first
25:06i like to show this one first because
25:07this is the way the literature actually
25:09developed
25:11and so it's a good way to kind of get a
25:12sense of the intellectual history of the
25:14measurements of human capital
25:17so the way this literature started is
25:19from
25:21measuring years of schooling and i think
25:24the reason is kind of hopefully obvious
25:26which is the years of school years of
25:28schooling is
25:29it's the most obvious it's the most
25:31visible it was the first thing that
25:33mincer put in the wage equations and
25:34talked about right it was right there
25:35from the beginning
25:37it's also one of the earliest things
25:38that we got data on
25:40and richard hit on this in his in his
25:42first lecture data availability and the
25:44growth
25:45of data availability has been a
25:46fundamental driving force for macro
25:48development research
25:50so human caps right years of schooling
25:52is one of the very first things we have
25:53because it turns out that many countries
25:56collect data on how many students or
25:58what share of young people are enrolled
26:00in school
26:02and knowing how many people enrolled in
26:03school if you just kind of add that up
26:05through a through a cumulative process
26:06you can figure out sort of the average
26:08years of schooling in the population
26:11barrow and lee are kind of the most
26:13well-known people for providing these
26:14estimates and so that's what i've
26:16plotted here on the left
26:17this is simply average years of
26:19schooling in the population age
26:2115 or above from their data plotted
26:23against gdp per worker
26:25measured in 2017 us dollars from pen
26:28world tables 10.
26:29both easy to get data sets online so
26:32what you can see is that there's
26:33enormous variation in years of schooling
26:35and this makes this a very likely and
26:37obvious culprit for including in our
26:39measures of human capital
26:41the richest countries in the world
26:43especially those that started getting
26:44educated early
26:45they have something like 13 to 14 years
26:47of schooling on average in their
26:49populations
26:50the poorest countries in the world that
26:52might be as low as two years of
26:53schooling
26:55okay people don't talk about it a lot
26:58but i think it is kind of interesting to
26:59look at the long run dynamics of this
27:01process and so
27:03uh lee and lee and one of these leads is
27:05the same as this lee
27:06built on the barrel and lead database
27:08but went back in time and measured
27:11or took the best possible estimates of
27:12enrollment as far back as you could go
27:16and they argued that it was possible to
27:17measure educational attainment from 1870
27:20onward for many countries so what i've
27:21shown you here is just sort of the
27:22long-run development of education which
27:24i like to do to give you a sense of the
27:26historical perspective
27:28so you can see for example that part of
27:30the reason why the united states is very
27:32educated today is that it's been
27:33educated for a long time
27:35way back in 1870 it was already the most
27:37educated country in the world
27:39you can see here japan has caught up
27:41fairly quickly it used to be relatively
27:43less educated and has grown quickly
27:46china is also becoming educated quickly
27:48the nice thing about china and one
27:50reason i want to show you this chart is
27:51that i want you to understand human
27:52capital is very slow moving
27:55because it's human capital of the
27:56workforce and so if you start investing
27:59a lot
28:00building you know secondary schools
28:02building colleges
28:03paying people to go to college you know
28:05making tuition free all of those
28:07policies
28:07you do that today
28:10you're going to affect the workforce but
28:12very slowly because the workforce
28:14includes some people who are 60 65 70
28:16years old
28:18and so your 60 65 70 year olds will not
28:20be highly educated for another 30 40
28:22years and that's sort of what's slowing
28:24china down they're investing a huge
28:25amount in education
28:26but it takes a long time for this to
28:28play out
28:30and then i've included down here a
28:31couple of sub-saharan african countries
28:33just to make the point that
28:34even within regions there's huge
28:36diversity of experiences so you can see
28:37here
28:38ghana has had phenomenal growth they're
28:42investing a ton in education and in fact
28:44they're up around china
28:45and brazil at this point whereas niger
28:48is lagging badly and is way behind
28:51okay so this is all to say that
28:54education i think is sort of the most
28:55obvious most visible it's something
28:57we've been investing in a lot for a long
28:59time this is the first thing we're going
29:00to want to include in our measure of
29:01human capital
29:04the second thing we need is an estimate
29:07of what's the value
29:09of education now right from the very
29:12beginning in the united states there was
29:14the literature argued that wages were
29:15about log linear in schooling
29:18if that's true then the value of
29:20schooling is just
29:21the slope it's just sort of what's beta
29:241 for my regression before what's the
29:25return to schooling as people in the
29:27micro literature often call it
29:30mincer argued from the very beginning
29:32that looked like something in the range
29:33of eight to ten percent
29:35was right for the united states that
29:38value seems to be fairly persistent over
29:40time we get
29:40fairly similar numbers now
29:44what's nice about this is that as i said
29:46people have run regressions of this type
29:47in many countries around the world and
29:49so people have been collecting what is
29:50the return to schooling for countries
29:52around the world
29:54and that's what's plotted here on the
29:56right this is the data from the most
29:58recent database
30:00i've tried to always include the source
30:01of the data and you can actually click
30:03all of these links if you have the
30:04slides open and it'll take you to the
30:05reference if you want to get any of this
30:08so this is a database that has collected
30:11these returns to schooling and i've just
30:13plotted what is the return against gdp
30:15per worker
30:17you can see that there's substantial
30:19variation right you see some countries
30:21where it seems like the estimated return
30:23is two three four percent you can see
30:24some countries where it's like twenty
30:26thirty percent
30:27what's maybe surprising but shows up
30:30persistently in studies that do this
30:33is that the return to schooling on
30:34average does not vary much between poor
30:36and rich countries
30:39sometimes you'll find it's a little
30:40higher in poor countries sometimes you
30:42find it's the same
30:43but not that different so i think what
30:46many people have done is take this as
30:48being evidence of saying that the return
30:49to schooling around the world
30:50a good estimate would be eight to ten
30:52percent okay
30:59so this leads us to the classic estimate
31:01of human capital i should have put some
31:02references here i'm sorry but this
31:04is what a classic estimate would be like
31:05holland jones 99 or caselli's handbook
31:08chapter from 2005.
31:11their estimates of human capital that
31:12you get these very traditional estimates
31:14are built on years of schooling and in
31:16particular they're built on the
31:18observation that there are large
31:19differences in years of schooling across
31:21countries
31:22and it seems like a year of schooling
31:23raises wages by eight to ten percent
31:26under those assumptions i described that
31:28would mean that a year of schooling
31:29makes a worker eight to ten percent more
31:31productive
31:33okay um here there's a bunch of
31:36references there's you know people have
31:37been building these databases on the
31:39returns to schooling so i've given you a
31:40bunch of different references that i
31:42think all end up pointing in the same
31:43direction
31:44poor and rich countries have similar
31:46returns maybe poor countries it's a
31:47little higher
31:50and that'll get us to a measure of human
31:51capital like what juliette gave on last
31:53friday
31:54which says that log wages and log human
31:56capital they're about
31:58linear in years of schooling so if i
32:00want to measure a country's
32:01human capital stock i'm going to say
32:04that the log of human capital is just 10
32:06times their average years of schooling
32:09and this is what people are in practice
32:11frequently doing
32:12okay if you do that just to give you
32:15some statistics for what i think is the
32:17classical measure of human capital
32:19you'll find that human capital in the
32:2190th percentile country divided by the
32:2310th percentile country that differs by
32:25about a factor of two
32:27whereas the total income difference is
32:28about a factor of 19 between those two
32:30countries
32:31okay so human capital that this is
32:33important you know i think we would all
32:35like to have
32:35twice as much human capital twice as
32:37much output that would be great
32:39but relative to the size of the thing
32:41we're trying to explain it's small
32:42and sort of that's you know here's the
32:44metric for that
32:45the 90 to 10 ratio of human capital
32:48divided by gdp worker that's like 11
32:50so we haven't explained that much and
32:51then people said well it seems like
32:52maybe human capital is not so important
32:56i prefer a little bit different
32:58development accounting metric
32:59everybody's got their favorite this is
33:00like one of these never-ending debates
33:03i kind of like to look at the covariance
33:04of log human capital log output divided
33:07by the variance of log output but
33:08it really doesn't matter what metric you
33:10use this gets you 19
33:12at the end of the day if this is your
33:13measure of human capital you're going to
33:14find it's not that important
33:19questions before i start showing you
33:20more components
33:25all right i have a i have a question may
33:27i ask yep
33:28please so i have a question about the
33:31years of schooling so is there any way
33:33that
33:33researchers include their the quality of
33:36schooling also
33:38give me two slides
33:43not for the basic thing but eventually
33:44people started to think about that
33:45that's the right answer
33:49all right so the second thing i want to
33:50include and juliette did put this into
33:52her measure of human capital on friday
33:54is
33:54experience think about this as right
33:57from the very beginning
33:59minser argued that experience was an
34:01important component of human capital
34:03more experienced workers have higher
34:05wages
34:06under the assumptions we've made that
34:07implies that more experienced workers
34:09are more productive
34:11up here i answered this question earlier
34:14i'm sorry i can't
34:15remember the names from flashing them in
34:16the screen but what is experience really
34:18almost always that means potential
34:20experience
34:22we would love to measure real experience
34:24like how many years or hours or months
34:26have you been working
34:28to do that really you need panel data i
34:31don't think anyone
34:32trusts the you know imagine asking you
34:34how many hours have you worked in your
34:35life for months or years
34:36you know the answer would be nonsense
34:38let's be honest i have no idea how many
34:40months or
34:40years i've worked in my life we wouldn't
34:43trust that answer
34:44and we don't have panel data for poor
34:46countries so instead of using actual
34:48experience what almost everybody does is
34:50they measure what we call potential
34:51experience that's how many years has it
34:52been since you graduated
34:54the idea is that's when you really
34:56started working you could have really
34:58started working after you finished your
34:59last degree
35:00all the time since then that's how much
35:02experience you could potentially have
35:04that's where the word potential comes
35:05from
35:07and then in practice that ends up being
35:09measured as something like age minus
35:10schooling minus six
35:13okay if you measure experience that way
35:15you're going to find that the level of
35:17experience is really similar in poor
35:19rich countries
35:22and the reason is actually very simple
35:23very intuitive
35:25poor countries you know you are
35:26developing countries have are younger
35:29they have very young populations they
35:31have young workforces and so they have
35:32lower ages on average among their
35:34workers
35:35but they go to school less so they have
35:37lower schooling
35:39if you construct experiences age minus
35:41schooling minus six
35:42what you're going to find is they have
35:43lower ages they have less schooling
35:46but the experience they have is almost
35:47exactly the same as what rich countries
35:48have on average
35:51okay so that's not what's going to be
35:53different
35:54it's not how much experience they have
35:56it's the value of that experience that
35:58we see in the labor market
36:00so what i'm plotting down here is the
36:02experience profiles for a select set of
36:04rich countries
36:06these are the five richest countries in
36:07the database from this paper
36:09and then profiles for select poor
36:11countries which is the five
36:13poorest countries that were in the
36:14database for the same paper
36:17so the way you read these profiles let
36:18me just pick one to start with so this
36:20dash gray here the highest one is
36:22germany
36:23you always estimate a profile so that it
36:26starts at zero
36:27and then what you're measuring is how
36:29much do wages go up as you get more
36:31experience and get older
36:33and so that's what's being plotted here
36:35when you get to five years of experience
36:36in germany the wages have gone up by 50
36:38percent as you get towards 15 years of
36:41experience their wages have gone up by a
36:42hundred percent
36:44and eventually in germany the highly
36:45experienced workers make about 120
36:47percent more than brand new workers
36:50okay you can see here united states
36:52canada france australia you're getting
36:54between
36:5570 to 120 higher wages as you become
36:58experienced
36:59on average among rich countries you
37:01about double your earnings over your
37:02life
37:04um just when they're when they're doing
37:06these wages what are they using for wage
37:08like is it like pre-tax post tax like
37:10how are they defining your
37:12wage that's going to make a huge
37:16difference for something like us versus
37:17canada even like
37:19yeah so
37:23they as we i'm in here somewhere in the
37:25at all now but now it's embarrassing
37:26because i have to remember
37:27i think we're using pre-tax for most
37:30countries and the reason is that pre-tax
37:31is what is measured in most countries
37:33yeah that would seem to that would make
37:34sense to me
37:36okay thanks yeah
37:39um on the right hand side we're doing
37:41the same thing we're measuring these
37:42same wage profiles but in a group of
37:44poor countries and you can see
37:46in poorer countries earnings wages go up
37:49much less over the course of your life
37:51so you know older and experienced
37:54workers make
37:55maybe 30 more than the youngest workers
37:57in those countries
38:00so we can show this for as many
38:02countries as we can get our hands on so
38:03what i'm going to show you now is what
38:04is the height of this profile what's the
38:06value to having between 20 and 24 years
38:09of experience
38:10that's what's plotted on the y-axis so
38:13basically it's the height of that
38:14profile at 20-24 years
38:16plotted against gdp per worker
38:20and so you can see that that what i just
38:21showed you is a fairly persistent
38:23finding
38:24are richer countries wages go up a lot
38:26more
38:27over your life as you get more
38:28experience and in the poorer countries
38:30they go up a lot less and you can see
38:32here you know bangladesh vietnam india
38:34guatemala earnings are going up by wages
38:37go up by about 20
38:39as you gain experience over your life
38:40not very much
38:43now in practice you have to be a little
38:46careful about saying that that's
38:47definitely human capital
38:49right there's many things that could be
38:51going on in in in the
38:53paper we we talk a lot about other
38:55hypotheses of what could be happening
38:57here
38:58i think a lot of ancillary evidence has
39:00come together to lead people to conclude
39:02that this
39:03probably is human capital that workers
39:05are developing more human capital over
39:07their lives in rich countries
39:09than in poor countries sort of a recent
39:12piece of evidence that i like a lot that
39:13points in this direction
39:14comes from a paper by ma it's all it's
39:17brand new it might even be 2021 at this
39:19point
39:20which looks at how much training workers
39:22receive
39:23over their lives and so what they do is
39:25they collect data from a wide range of
39:27countries on
39:28the share of workers who get training
39:29from their employer
39:31and they're plotting what share of
39:33workers get training against gdp per
39:34worker you can see there's a humongous
39:36gradient here
39:38in poor countries basically no one gets
39:40training from their employers and in
39:41which countries maybe half of the
39:42workforce typically gets training from
39:44an employer
39:46that's a function of two things one is
39:48that in poor countries many people are
39:49self-employed
39:50you can't get training from your
39:51employer if you're self-employed
39:54i think that's still fair though in the
39:56sense at least to me it seems like
39:57i think it's probably legitimate to
39:59think that a lot of self-employed
40:00workers aren't learning much as they do
40:02their jobs
40:02i mean who would they learn it from who
40:04would teach them anything
40:07the second thing of course though is
40:08that even for people who work for
40:10employers who work for wages are more
40:12likely to get training from their
40:13employer in rich countries
40:15so i think about this as being one
40:16likely culprit for why we see wages
40:18rising more over their life maybe
40:20workers are getting more training
40:21they're doing more investments they're
40:22learning more over their lives
40:24they actually provide a quantitative
40:26theory that suggests this is probably a
40:27big component of it
40:30if we take oh sorry if we take this
40:33as reflecting workers learning more over
40:35their lives and rich than poor countries
40:37we want to add that into our measure of
40:38human capital
40:41and i'll do that in a second do these
40:44results
40:44change dramatically at all if you
40:46separate it out between like men and
40:48women for example
40:49have you done that
40:53a little bit but not much surprisingly
40:56women have
40:57lower returns to experience in almost
40:59every country
41:02but that is mostly because the potential
41:05experience is a less good measure of
41:07experience for women than for men
41:10men work almost continuously since they
41:11graduate women have many more
41:13interruptions over the course of their
41:15lives
41:15right yeah true yeah exactly all right
41:19yeah can i ask a question about the uh
41:22about this graph actually
41:25um i guess is it
41:29like you you say that it's not exactly
41:34well no sorry uh let me let me reword it
41:36i guess are we
41:38could it also possibly be that just the
41:40there's kind of like a
41:42cap to the capacity of a country to use
41:44experience of a worker
41:46where like in the more rich countries
41:51having more experience actually lets you
41:53move
41:54up like a wrong or a ladder or something
41:56but in a poorer country
41:58having more experience doesn't really do
42:01anything for you
42:02if there's less capacity for the
42:04country's like
42:07uh like tfp to kind of use that
42:10experience
42:11so i guess i'm like years of experience
42:14maybe isn't
42:15completely translatable across countries
42:19i think that the the paper that we had
42:22to read
42:22kind of talks about this a little bit
42:24the one that you did with hendrix
42:28sorry if that was very vague no no i
42:30understand exactly what you're driving
42:31yet
42:32um it's a good question
42:35and it's a deep one so if i go back a
42:38little bit
42:41really this goes back to the production
42:43function
42:44if you look at this production function
42:46it says look
42:47there's constant but decreasing returns
42:49to capital
42:50constant and decreasing returns to human
42:52capital they're completely separable and
42:53there's really no interesting
42:55interactions between these factors here
42:57if i double your human capital in the
42:59united states you get 2 raised to the 1
43:01minus alpha output
43:02if i double your human capital vietnam
43:04you get 2 raised to the 1 minus alpha
43:06extra output it's the same everywhere
43:09and then i think i could give you a list
43:12of papers and especially it's a common
43:14thing to see in the conclusion of a
43:15paper to say
43:16yeah you know it's really interesting
43:17what if there's an important interaction
43:19between the technology we have
43:21and our ability to use schooling or
43:23experience that's an interesting idea
43:24for future research
43:27and that's kind of where it always
43:28lingers okay but that's
43:30that's really where this point right
43:32here is where you can really see why
43:34why i'm not talking about ideas like
43:35that it's embedded in the production
43:37function right from the very beginning
43:40thank you
43:44all right i don't want to bore you to
43:46death by showing you a million of these
43:47things so there's a couple more things
43:49that people are going to keep track of
43:50time too there's a couple more things
43:51that people have looked at
43:52including education quality and health
43:55would be the two other things you might
43:56think about adding into this measure
43:59so on the left hand side here i have
44:00plotted a measure of education quality
44:03which is
44:03math scores on the pisa exam which is a
44:06very common and very popular exam
44:09if your country takes it you'll know
44:10about it because they'll report the
44:12report they'll report the scores in the
44:13newspaper and complain about how you're
44:15not doing well enough on it unless
44:16you're in singapore or finland where
44:18they do great
44:19so what's plotted here is the score on
44:21math in 2018 and you can see yes people
44:23in poor countries score
44:24much worse on the math exam
44:28health is probably the trickiest thing
44:30to measure
44:31and again it's like even in the united
44:33states you know how do i measure the
44:34health of two people
44:36that's a very difficult thing so here's
44:38one proxy for health
44:40adult survival rate that's the
44:42probability of the 15 year old will live
44:44to 60.
44:46and you can see that in rich countries
44:47most 15 year olds will live to 60. but
44:50if you go to poor countries you know
44:51on average maybe only 70 80 percent do
44:55we think that's probably correlated with
44:57bad health with the prevalence of
44:59diseases with their inability to get
45:01enough nutrition
45:02and to afford proper medicines you know
45:05it's not ideal measure but that's what
45:06we have to work with
45:10you can see though that as we go along
45:11in this constructive method
45:13you know what's the value of 50 points
45:15of ps of math score
45:17for productivity now we're starting to
45:20get this is getting a little tricky this
45:22gets a little harder and i'm going to
45:23show you the papers that have done this
45:24in a second
45:25but this is more challenging what's the
45:28value
45:28in terms of productivity of living in a
45:30country where you have 0.1 higher adult
45:32survival rate
45:35okay people continue to build on those
45:38minster wage equations and they have you
45:39know approaches and tricks in back of
45:41the envelope calculations to make that
45:42happen
45:43so let me show you what those results
45:44are
45:46what i'm going to do here is accumulate
45:48one step at a time
45:50a measure of human capital so i'm going
45:51to construct it literally as the
45:53literature has done over time
45:55and i'm going to ask how important is it
45:56for cross-country income differences
45:58using this covariance metric
46:01so i told you before if you measure
46:03human capital as years of schooling
46:04you're going to
46:05you're going to explain or account for
46:06about 19 percent of income differences
46:10if you add quality of schooling sort of
46:13a typical
46:14estimate is that that's actually about
46:15as important as the years of schooling
46:18for cross-country income differences
46:19that'll get you to about 38
46:22and i don't know if this is coincidence
46:24or not but when you add experience you
46:25almost always get about another 18 to 19
46:28percent here it's 18
46:31so this is 56 health
46:34the leading estimates of health are that
46:36it's much less important which
46:38i don't know i just always feel like
46:39it's important but anyway i have no i
46:41have no evidence i've never done any
46:42research i feel like it should be more
46:44important than that but i i've never
46:45proved it
46:46and someone said that schooling
46:48experience health we're up to maybe like
46:4959
46:50of cross country more than half of
46:52cross-country income differences are due
46:53to health if we add it up this way
46:55here's your list of references if you
46:57click on them they'll take you to the
46:58right papers to see
46:59see where these estimates come from
47:02being part of health is accounted for
47:04indirectly through like schooling
47:05experience like if you don't have good
47:07health then you're not going to school
47:11yep give me one second that's the next
47:12slide it's exactly i'm going to
47:14summarize
47:14and that's exactly what i'm concerned
47:15about
47:17so this first approach which i'm going
47:19to be done with now
47:21uh measures cross-country human capital
47:23differences constructively like
47:24literally constructing building it piece
47:26by piece
47:28you account for differences in stocks or
47:30rates of return
47:32and those resulting estimates are large
47:33maybe 60 percent of income differences
47:35but
47:35you know i think you should have some
47:36concerns at this point you should not be
47:38satisfied or happy with what we've done
47:41one thing is you have to be worried
47:42about those assumptions so here's a
47:44reference to the first paper that starts
47:45to rethink those assumptions we use for
47:47development accounting there's a whole
47:48literature that builds from him but he
47:50was the first
47:52the other two concerns just kind of
47:54conceptually it's hard to be exhausted
47:58what i mean by exhaustive is do i think
48:00we've accounted for or measured all of
48:02the different pieces of human capital
48:04probably not right i mean we added up
48:06years of schooling quality of schooling
48:08experience health if there's anything
48:09else that's not in there
48:11so we might imagine this number could be
48:13even bigger
48:14but at the same time and this is what is
48:16it jolene is that who that was
48:18yeah okay good um we may be double
48:20counting here
48:22right because health might very well
48:26affect the quality of schooling that a
48:27country has if the students have better
48:29health they may do better on the tests
48:30and that may show up as better quality
48:31of schooling
48:33likewise you know being in school a long
48:35time may help you learn more on your job
48:37so when i add up the value of schooling
48:39and the value of experience i may have
48:40counted something twice
48:42because some of the value you having
48:43experienced may have come from your
48:44schooling
48:45and so there's a concern here that by
48:47adding things up this way i may be
48:49misattributing miscounting or double
48:51counting things
48:54that's going to lead me to my second
48:56approach which i'm going to call the
48:57deductive approach
48:58so are there any questions before i meet
49:02um excuse me
49:07yeah hello prof yep somebody go
49:12hey hello i hear you yeah look
49:16yeah hello i have a little bit of
49:18question respect to a recent port
49:20contrary especially technological uses
49:22and that of that with the taste
49:24so in terms of measuring uh the human
49:27capital specifically in
49:29most of the developing countries like
49:31you gave an example like in ghana and
49:33other different countries and etcetera
49:34don't you think that the taxonomy
49:36might be perhaps determined more easily
49:39based on the technological uses
49:41that the people have within the country
49:43they say because i think uh with the use
49:45of technology
49:46the technical know-how especially in
49:48this uh
49:49environment of uh intensive uh economic
49:53that we are in so with all
49:54i mean uh the technical uses uh how do
49:57you think that poor control regardless
49:58of violence
49:59especially in the cultural project in
50:00terms of converging bigger convergence
50:03uh it is possible whenever poor
50:06countries are trying to educate the
50:07populations
50:08because i think there is always a peak
50:11in most of the western nations
50:12especially developed country
50:13and as a result agenda is always an
50:15opportunity for the poor country to
50:17catch up because they're convergent
50:18but how would they catch up when they in
50:21terms of the transfer of technology
50:23rnd i mean research and development is
50:26not advanced
50:27and the taxonomy issues also has a
50:30little bit of gap
50:31so with respect to the quality of
50:33education so what
50:35are some of the research you think that
50:36may be appropriate to ensure that
50:38an advancement or most of the least
50:39developing country can able to catch up
50:41more faster
50:44okay that was a lot to chew on um i
50:46think
50:48i heard two distinct questions and you
50:50can let me know if i missed something
50:53one thing i heard was um a general
50:55question about the interaction between
50:57human capital and technology
50:59yeah and i think my answer is going to
51:02be the same as the answer i gave a
51:03minute ago which is that if you look
51:05right here
51:06technology is a part of tfp here's human
51:09capital
51:11if you look at this you'll see that
51:13there are no really interesting
51:14interactions between these two things
51:16having more technologies good having
51:18more human capital is good but there's
51:20not really
51:21there's no idea that you have to have
51:23technology to make use of skills
51:25there's no idea that having skills makes
51:26technologies more valuable
51:29in everything i've done so far that's
51:30been missing if you send me an email
51:33i'll send you i mean there is
51:34very recent work that's trying to
51:36investigate these things like how much
51:37does technology
51:38aid the use of human capital and vice
51:40versa
51:42that's really frontier work at this
51:44point so i would say it's not very
51:45subtle that's why i didn't want to teach
51:47it but
51:47people are thinking about it the second
51:49piece was you know if
51:51education quality is important how do i
51:53get more of it i think that's the way i
51:54would interpret that question
52:00the uncomfortable and sad truth about
52:02this literature is that no one knows
52:04i can point you to some books and some
52:06papers and reports
52:08that's not really something that
52:11what's a polite way to say i'm somebody
52:15who's worked on this and i'll say it
52:16about myself
52:17that's not something my research has
52:18ever been very good at
52:21i can tell you that it's important but
52:22if you ask me like at a nitty gritty
52:24level what should i do in our schools to
52:26make sure that they actually get this
52:27quality that seems so important
52:28my papers are completely useless on that
52:30part that's the truth
52:33um i can point you to some manuals and
52:35like if you read the
52:36pisa reports they'll give you some ideas
52:38or best practices of countries i don't
52:40know how reliable those are but that's
52:41the best that's out there truthfully
52:44because what made me to have asked the
52:46question is simply because i'm also
52:48writing my uh this irritation solution
52:50on human capital development it will
52:51develop
52:52in developing countries so but uh i came
52:54across some uh
52:56i mean some issue with respect to the
52:59i mean the country productivity based on
53:02mainframe
53:02that stays to this stage that is to say
53:05from the
53:06before the turns after another
53:07advancement like in asia example like in
53:09korea
53:10like people select taiwan you know they
53:12started with uh radio activities like a
53:15radio receiver and all the different
53:18things you know before they try to
53:19advance
53:20but in the same vein and in certain
53:21countries in asia and so like malaysia
53:24where i'm stationed currently and
53:25basically china
53:27they were not advanced and typically
53:29compared to other countries like ghana
53:31was empowered with other korea
53:33by the end of the day ghana left behind
53:35malaysia has advanced
53:36korea has very intense technology i
53:39believe that technology i mean observe
53:42like a sort of leaf rock
53:43of the economy like in india okay
53:46because of
53:47the incremental human capital in india i
53:49think they tried to refrog
53:51we're in it of living over world today
53:53so if
53:54maybe a technology vis-a-vis human
53:56capital a main
53:58improvement or advancement per se can
54:00can be uh somewhat depressed let's say
54:02because
54:02we have different processes in terms of
54:04human capital such as f as is one of the
54:05deposited percent
54:07and income is one of them but can we
54:09measure the gdp per capita
54:11based on the individual percent based on
54:15the popular educated population within a
54:17country
54:18or based on the input of the country the
54:20total production
54:21part of a dpf so total factor of
54:24production tpf
54:26tft runner
54:30sorry i don't think i understood that
54:32question yeah what i try to say is that
54:35can we measure the country gdp
54:39i think we're going to have to go on i'm
54:41sorry okay
54:42okay
54:46if asked me again during the during the
54:48follow-up session i'm just not quite
54:49sure i'm grasping it um
54:52can i make a quick question yes please
54:55um so is there a sense here um we're
54:58making a causal statement
55:00somehow so we're saying human capital
55:01causes gdp to be larger
55:03but it can also be true that um richer
55:07countries have
55:08like more opportunities to increase
55:10human capital right because it's not
55:12only valuable
55:13for develop as a productive resource
55:16and thinking in cases such as latin
55:19america where
55:20like recently the like the fiscal
55:23government like the
55:24the country governments make a huge
55:26effort to increase education
55:28in all levels and it was only possible
55:30because the countries were
55:32a bit richer than before or they
55:34experience some kind of commodity bloom
55:35etc
55:37so right from the beginning actually
55:39this has been an obsession in this
55:41literature and
55:42if you go back and read bills in
55:43greenhouse paper the people who worked
55:45out these assumptions that we're using
55:47you'll see that that was exactly what
55:48they were wondering which is does
55:50getting more schooling make you rich
55:51or does getting richer lead you to get
55:52more schooling
55:54what i'm going to say today and the way
55:56i try to phrase these slides i sometimes
55:58mess up on this is we always try to say
55:59that
56:00human capital accounts for a lot of
56:03cross-country income differences
56:05and then that's a fancy way of saying we
56:07don't know where the schooling comes
56:08from
56:09it could be that tfp leads people to get
56:11more schooling and
56:12leads them to get richer i'm not going
56:15to be able to make a statement
56:16that's causal i'm trying not to make any
56:18causal statements
56:19it just seems like there are really
56:20large differences across countries in
56:22human capital
56:23and we want to think about that more
56:29all right so the second approach uh that
56:31i kind of want to investigate a little
56:32bit is going to be called
56:33the deductive approach and this is built
56:36on the idea that
56:37human capital you know by my definition
56:39at least is embedded in people
56:42and so what we're going to do is we're
56:43going to utilize the fact that migrants
56:46of which there are many carry their
56:48human capital to new countries when they
56:49move
56:50okay and so this is going to allow us
56:52the opportunity to deduce the importance
56:54of human capital
56:56what's different about this is that
56:57we're not going to try to specify what
56:59human capital is or to build it up piece
57:01by piece we're just going to say
57:03whatever
57:03human capital is the migrant has it when
57:06they move they take it to the new
57:07country let's see how important it is in
57:09that new country
57:11so if you remember you know the wages we
57:14specified under those maintained
57:15assumptions were that the log wage of a
57:17worker with human capital i in country c
57:20is additive and a country effect and a
57:22human capital effect
57:24and so the way i want you to think about
57:25this is well if i have a worker who
57:27migrates they take that human capital
57:29they used to have a country with this z
57:31they move to a country with a new z
57:32and that's going to give me some
57:33opportunity to see how they do
57:36even more intuitively the way i like to
57:38think about this is
57:41in general what would be really cool is
57:43if we could take the united states and
57:44give the united states the human capital
57:46investments of ghana
57:47and then say how does that affect the
57:49united states now of course we don't get
57:51you know we're macro economies we don't
57:53get to do randomized trials on entire
57:54countries
57:56but an immigrant gives us the best shot
57:58we have because we can see somebody with
57:59ghanaian human capital move to the
58:01united states and see how they do
58:04and that's kind of the underlying idea
58:06behind this exercise
58:08now of course there's a trade-off
58:09nothing comes for free in economics
58:11since we're going to be talking about
58:12migrants we have to deal with two
58:13additional concerns that are just going
58:15to be pervasive throughout this
58:16literature that thinks about migrants
58:18and human capital
58:19which is that the migrants of course are
58:21not randomly chosen you have to worry
58:22about the fact that they're selected
58:24they may be different than the random
58:25ghanaian
58:27and the second is that their human
58:28capital may not be fully portable
58:30right if you look at this equation what
58:32i'm thinking to myself is well if they
58:33take this human capital to another
58:35country what happens
58:37but of course the skills and the
58:38knowledge they have may not be as useful
58:40in the new country they may be the wrong
58:42type they may be discriminated against
58:43all of those types of concerns have to
58:45be dealt with
58:47okay so there's been a
58:50at the bottom here i talked i gave you
58:52some references for other papers there's
58:53a there's a
58:55sort of a small literature but a long
58:57one that long has been going on for
58:59about 20 years that investigates this
59:00idea
59:01the approach i'm going to show you today
59:03and that i recommended for the reading
59:05was to think about what are the wage
59:06gains at migration
59:08okay so the idea here is imagine that i
59:10could see
59:12the wage of a worker who moves from c
59:15to c prime both after he migrates and
59:18before he migrates
59:21if i go back and take this wage equation
59:23and just take the first difference of
59:25this just difference across two
59:26countries that worker has the same human
59:28capital in both countries
59:29so that's going to drop out and what i'm
59:32going to be left with is just
59:33the z the difference in z's between the
59:35two countries
59:37but that's exactly one of the two things
59:39i wanted to understand from my
59:41development accounting exercise which is
59:42how big is the
59:43z difference across countries
59:47so this would be a really easy way to
59:49quantify one piece of the development
59:51accounting puzzle
59:53the underlying intuition here i like to
59:55just say this like plainly
59:57suppose i have a worker who migrates
59:59from a poor to a rich country let's say
1:00:00the rich country is 10 times richer and
1:00:02make everything easy
1:00:04if that workers wages went up by a
1:00:06factor of 10
1:00:07i would say wow switching countries
1:00:09fixed everything it made this worker
1:00:10much richer
1:00:11it must be that country is what matters
1:00:14something about the institutions the
1:00:17capital the infrastructure i don't know
1:00:19but changing countries fixed everything
1:00:21i'm going to conclude that z must be
1:00:22what matters for cross-country income
1:00:24differences
1:00:26on the other hand suppose that that
1:00:27worker switches countries and their
1:00:29wages don't move at all they stay really
1:00:30low
1:00:32well then i would say boy the country
1:00:33didn't do anything the country must not
1:00:35really matter much for the wage this
1:00:36worker receives
1:00:38boy it seems like their low human
1:00:39capital must explain why they have such
1:00:41low wages
1:00:44now you'll notice that that really
1:00:45hinges crucially on
1:00:47thinking about who these workers are and
1:00:48whether their skills fully transfer and
1:00:50in the paper
1:00:51i'm going to try to get through this
1:00:52relatively quickly so i can show you a
1:00:54little bit of models
1:00:55in the paper we do try to do some
1:00:57exercises to think about how big those
1:00:59effects are but i want to kind of get to
1:01:00the main results to show you
1:01:02what's different about this let me
1:01:05emphasize this one more time
1:01:06this is almost like a completely
1:01:08different way to come at the same
1:01:09problem
1:01:10because here i'm not going to try to
1:01:12specify what human capital is
1:01:14i'm not going to try to specify how they
1:01:16got it or what attributes of it are
1:01:18important
1:01:19i'm just going to say whatever they have
1:01:20they're going to take it to two
1:01:21different countries
1:01:23and that's going to let me to think
1:01:24about how big the z differences are
1:01:25that's what i mean by deductive
1:01:27it's like i'm going to deduce based on
1:01:29how much their wages change how
1:01:30different human capital must be
1:01:34you need the the kind of one tricky
1:01:36thing for this is that
1:01:37you need data where you know how much
1:01:39workers how much wages workers made
1:01:42before and after they migrate
1:01:44most data sets don't have that so here's
1:01:46a couple of data sets we found that did
1:01:48that's what we're going to use to
1:01:49quantify this and all we do in this
1:01:51paper
1:01:52is just convert wages to ppp adjusted
1:01:54wages
1:01:55and then we're just going to ask how
1:01:56much do ppp adjusted wages change
1:01:59relative to pppgd people worker gaps
1:02:02that's the notion of how much does your
1:02:03wage go up how much is the country
1:02:05matter
1:02:09and then what i've highlighted here is
1:02:11sort of the main results from that paper
1:02:13what we find is that workers from
1:02:15countries that are less than a sixteenth
1:02:17as rich as the united states so these
1:02:19are the very poorest countries
1:02:20immigrants come from to the united
1:02:21states when they move
1:02:22they get about three times higher wages
1:02:26the gdp gap between those countries in
1:02:28the united states on average is a factor
1:02:30of 32.
1:02:32so switching countries makes them richer
1:02:34right their wages go up by a factor of
1:02:36three but that's a small amount relative
1:02:38to the total gdp gap
1:02:40and so when you do it you'll find that
1:02:41that would imply that human capital
1:02:43accounts for about two-thirds of
1:02:44cross-country income differences
1:02:47you can do the same thing for different
1:02:49groups we have two different data sets
1:02:52but as a whole when we do this we
1:02:53consistently find that workers wages go
1:02:56up by
1:02:56a small amount relative to the gdp gap
1:03:00and that that would imply that human
1:03:01capital accounts for something like 60
1:03:04of income differences 50 to 60 percent
1:03:14so the deductive approach i think has
1:03:16strengths and weaknesses relative to the
1:03:18constructive approach it's exhaustive
1:03:19because it measures whatever human
1:03:21capital is wherever it comes from
1:03:23but it's not constructive and so
1:03:26somebody already asked me you know if
1:03:27education quality is so important how do
1:03:29i get it you can see that this is even
1:03:30worse now right this one is like if
1:03:32human capital is so important
1:03:33where do i get it from you know this
1:03:36would really not be very helpful for
1:03:37that
1:03:38it avoids double counting it does
1:03:41require these additional assumptions
1:03:42about migrants and i'm going to refer
1:03:44you to the paper if you want to ask
1:03:46questions you know
1:03:46that's it honestly it takes a whole
1:03:48seminar but i'm going to present the
1:03:49paper and really talk through all of
1:03:51that it's it's
1:03:52anyway what i like about it is that
1:03:54although it comes from a completely
1:03:56different direction you end up getting
1:03:57fairly similar
1:03:58results you get the human capital
1:03:59accounts for half to maybe 60 percent of
1:04:01income differences
1:04:03okay and so we have two different ways
1:04:05of getting to the same point
1:04:07both of us lead to think that human
1:04:08capital is really big i still think
1:04:10there's room to explore on the
1:04:12measurement side
1:04:13sort of my pet peeve with this whole
1:04:15literature is that
1:04:16we're really talking about measurable
1:04:18stuff we're not really talking much
1:04:19about skills
1:04:20i would really love to know more about
1:04:22what exact skills are missing in poor
1:04:24country like
1:04:25they don't go to school as long and the
1:04:26schools aren't as good so they don't
1:04:28know how to do
1:04:29what what specifically is it that they
1:04:31miss when they go to the job i've never
1:04:33had any sense that i know what that is i
1:04:34would love to know more about that
1:04:37anyway i think there's more to be done
1:04:39on the measurement side what i would
1:04:40like to do now let me see how am i doing
1:04:42on time
1:04:43yeah about what i expected what i'd like
1:04:45to do now is show you a little bit of
1:04:46models about human capital differences
1:04:48and the idea here is if you are
1:04:50convinced or at least intrigued to think
1:04:52that human capital
1:04:54might be vary a lot across countries and
1:04:56it might be a big piece of the
1:04:57development puzzle and also of the
1:05:00differences in across sectors during
1:05:02structural transformation
1:05:03you'd want to build that into your model
1:05:05how do you do that what is like the
1:05:07off-the-shelf simplest model you would
1:05:09start with
1:05:10so i'm going to give you a very simple
1:05:12version of these models
1:05:13and then give you references if you want
1:05:15to go and dig in and do the fuller
1:05:16bigger versions of them all
1:05:25all right so the goal here is mostly i
1:05:28first and foremost i want to illustrate
1:05:30the mechanisms that can help you explain
1:05:31these large human capital differences
1:05:34all right and the way you want to think
1:05:35about this is human capital as i've been
1:05:37emphasizing is an investment that people
1:05:39make
1:05:40people in rich countries are investing
1:05:41more that accounts for a lot of
1:05:44cross-country income differences but it
1:05:45doesn't explain it to explain it i would
1:05:47want to know
1:05:48why why do they invest more so what i
1:05:51want to do is give you some models
1:05:52i'm going to start with a benchmark
1:05:54model that can't explain it which is a
1:05:56useful result like oh if i do that i'm
1:05:57not going to get anywhere
1:05:59then how do i tweak that model to
1:06:00explain some of the differences
1:06:02all right this is the simplest way i
1:06:06could think of to write this model down
1:06:08so let's start with the firm it's a
1:06:10representative form so you can think of
1:06:11it as a bunch of firms if you want
1:06:13the firm is exactly what i showed you
1:06:15before nothing new here
1:06:17they operate this cobb douglas
1:06:19production function it's the same one i
1:06:20showed you
1:06:22i'm going to insist on the same two
1:06:23assumptions perfect substitution
1:06:25competitive labor markets
1:06:27that's going to give me the same wage
1:06:28equation i had before nothing new on the
1:06:31firm side
1:06:34what i want to do now is add a model of
1:06:36the worker
1:06:38really the person who's making decisions
1:06:41about investments
1:06:42and think about what would lead their
1:06:45decision about how much to invest to
1:06:46vary across countries
1:06:49so here's the simplest model this really
1:06:50is a dumbed down version of becker's
1:06:53book
1:06:54so let's have a consumer who lives for
1:06:56one unit of time
1:06:58and they're going to maximize utility
1:07:00and the utility is just i'm going to
1:07:02integrate over that period
1:07:04i have log preferences i have
1:07:06consumption of t
1:07:07at every instant t during that one unit
1:07:09of time
1:07:12and then they're going to decide their
1:07:14consumption profile
1:07:15the other choice they're going to get to
1:07:16make is how long they go to school
1:07:18before they start working
1:07:20so the simplest version possible of this
1:07:22is to say all right they're going to
1:07:23attend school for a length of time
1:07:25s schooling develops human capital so
1:07:28when they leave school they're gonna
1:07:30have human capital h of s
1:07:32the remaining unit of time one minus s
1:07:35they're going to go to work
1:07:36and they're gonna earn a wage then their
1:07:39budget constraint simply says that the
1:07:41integral
1:07:42of their consumption has to be equal to
1:07:44their income
1:07:45right the money to finance what they
1:07:47consume has to come out of their income
1:07:50and then their income piece by piece
1:07:53this is the z from the wage equation so
1:07:57literally i'm just copying the wage
1:07:58equation to start with
1:08:00right i'm plugging in their human
1:08:02capital that's a function of
1:08:03how long they go to school and then how
1:08:05much do they get to work well only one
1:08:07minus
1:08:08s of their life the rest of their time
1:08:10they spent in school
1:08:13if this is your model of the human
1:08:15capital choice you're going to get two
1:08:17main results this is this is well known
1:08:18from
1:08:19becker's book the first result you're
1:08:21going to get is that workers in this
1:08:23model or people really want to smooth
1:08:25their consumption over their life
1:08:27in this model i've assumed that people
1:08:28don't discount across the period
1:08:31or within the period so their
1:08:32consumption is exactly the same over
1:08:34their life they basically spend their
1:08:36income
1:08:36smoothly over the course of their life
1:08:41the second result that you're going to
1:08:42get is what's the optimal school choice
1:08:44so just take the first order condition
1:08:46and it really naturally gives you an
1:08:48equation that says the marginal benefit
1:08:50is equal to the marginal cost
1:08:52what's the marginal benefit of going to
1:08:54school well this first piece here
1:08:58is the wage you get for working one
1:09:01minus s periods
1:09:03the benefit says well i have a little
1:09:04more human capital if i go to school a
1:09:06little bit longer i'm going to have
1:09:07marginally more human capital that's
1:09:08going to marginally increase my earnings
1:09:12what's the cost of getting a little more
1:09:13schooling well this is the wage i could
1:09:16have earned during that instant
1:09:17right i could have made h of s instead
1:09:20but i spent that extra instant going to
1:09:22school and getting more human capital in
1:09:23the future
1:09:24that's the trade-off that a worker faces
1:09:26in this model should i get a little more
1:09:28human capital
1:09:29and get that benefit over the remaining
1:09:31one minus s
1:09:32or should i just go to work and enjoy
1:09:35extra earnings for that instant
1:09:39now what's really neat about this model
1:09:41uh is that if you look at it carefully
1:09:44the one minus alphas and the zcs cancel
1:09:47they're on both sides of the equation
1:09:50and so that tells me that the optimal
1:09:51school choice
1:09:52in this model is independent of z this
1:09:55is an idea that shows up in becker's
1:09:56book already i've always called this
1:09:58becker neutrality
1:10:01from our perspective this is a problem
1:10:05right we're trying to understand why
1:10:06people in different countries are
1:10:08investing different amounts in schooling
1:10:09and there's nothing about the country
1:10:11left
1:10:12so this is a theory that gets us nowhere
1:10:14i like it because it's a benchmark it
1:10:16says all right if i start with this
1:10:18model
1:10:18i can see that i'm going to go astray so
1:10:20i need to tweak something here i'm going
1:10:21to give you some ideas about what to
1:10:23tweak in the next few minutes
1:10:25but i want you to understand the
1:10:26intuition for this because i think it's
1:10:27really powerful
1:10:29so what's going on here imagine i live
1:10:32in a country with a better
1:10:33government or better institutions or
1:10:35technology or whatever your favorite
1:10:37theory of tfp is
1:10:40i think most of us kind of intuitively
1:10:42would say ah well if there's a better
1:10:44technology it makes it more beneficial
1:10:46to invest in human capital right
1:10:48but if you look up here what this tells
1:10:50us is that yes the benefits of human
1:10:52capital investment go up
1:10:54right because i'm going to earn more off
1:10:56my human capital investment
1:10:58but in this model the other force that
1:10:59shows up is that if i live in a country
1:11:01with a higher tfp better technology
1:11:03whatever it is
1:11:04that also raises the cost because i
1:11:06could have earned higher wages during
1:11:08the time i spent in school
1:11:10and those two things exactly cancel
1:11:13okay and so the underlying idea here is
1:11:15that if
1:11:17time my time
1:11:20is the main investment i put into human
1:11:24capital
1:11:26then the opportunity cost of my time is
1:11:28proportional to z
1:11:29the benefit of my time is proportional
1:11:31to z that's not going to get me very far
1:11:35all right so now how can we tweak this
1:11:37and there's going to be
1:11:38i think i have time to do all three let
1:11:39me check my time one last time might
1:11:41have to cut one off there's three main
1:11:42ways that the literature is used maybe
1:11:44four
1:11:45to tweak this model so we can start to
1:11:47get some actions we can start to get
1:11:49cross-country differences in human
1:11:50capital
1:11:51and so if you're building a model of
1:11:52human capital i think you want to make
1:11:54sure you have some of these elements
1:11:56one is borrowing constraints
1:12:00go back and look at what i said maybe
1:12:02you didn't notice it
1:12:03it sneaks by people a lot what agents in
1:12:06this model want to do is they want to
1:12:08consume smoothly over their life
1:12:11including while they're in school where
1:12:13does the money for that come from
1:12:16well the way this model is written in
1:12:18the bar
1:12:19in the budget constraint you're allowed
1:12:22to borrow as much as you want
1:12:24there's perfect credit markets so the
1:12:27first thing you can do is say well
1:12:29i'm not so sure about that i think
1:12:32there's actually probably a lot of
1:12:33limits to the ability of children and
1:12:35their parents to borrow and finance
1:12:37these investments
1:12:39and i've listed here some of the major
1:12:41impediments you know in general i think
1:12:42most of us know that credit markets
1:12:44aren't perfect
1:12:45that it's hard for people to borrow
1:12:47large sums of money to finance all of
1:12:48their consumption and expenditures on
1:12:50schooling
1:12:52particularly since we don't let children
1:12:53write contracts
1:12:56we don't let them sign binding contracts
1:12:57right hi i'm 14 i'd like to borrow the
1:13:00money so i can go to heist to a really
1:13:01good high school to benefit myself over
1:13:03my life
1:13:04you know we don't let people sign a
1:13:06contract that agrees that they'll repay
1:13:07that
1:13:10you might be able to work around that if
1:13:11you think oh well the parents can invest
1:13:13on the behalf of the children
1:13:16but that requires that parents be really
1:13:17altruistic towards their children
1:13:19okay and so if you go and read becker
1:13:21and tones this is a beautiful paper in
1:13:23the journal of labor economics that just
1:13:24works through exactly what you would
1:13:26need
1:13:26in terms of altruism or credit markets
1:13:30to get the socially optimal investment
1:13:32and
1:13:33they're pretty tough restrictions
1:13:36for our purposes since we're trying to
1:13:37think about cross-country human capital
1:13:39differences the main idea you would want
1:13:41is
1:13:41maybe people in poor countries have a
1:13:43harder time borrowing
1:13:45and that seems like a pretty natural
1:13:46restriction if they have a harder time
1:13:48borrowing in poor countries
1:13:51because the credit markets aren't as
1:13:52good or because their parents can't
1:13:54borrow on their behalf
1:13:55or because you know they're they're
1:13:58they're near a subsistence constraint
1:13:59consumption is really low
1:14:01any of those types of mechanisms um then
1:14:04what you're going to get is that people
1:14:05in poor countries are going to have less
1:14:06schooling because they literally just
1:14:07can't borrow enough that's one force
1:14:09that you could put in the model
1:14:11and hello and um
1:14:15i think perhaps it's also impossible to
1:14:19hedge against unemployment after getting
1:14:22the education
1:14:26yes um or more generally there's a
1:14:30there's a literature um tom krebs
1:14:33is the big reference for this about
1:14:35human capital risk that you can't
1:14:36there's no way to hedge against risk we
1:14:38often talk about human capital as if
1:14:40go to college in the united states and
1:14:42you get sixty percent higher earnings
1:14:43but of course there's a huge thing
1:14:45yeah okay yeah yeah but there is no
1:14:48possibility of putting that in this
1:14:50model for example
1:14:52yeah no i mean
1:14:56like i said if you go read krebs uh
1:14:58krebs and write
1:14:59send me an email and i'll send you the
1:15:01exact papers people have put these into
1:15:02models nobody in macro development that
1:15:04i know of
1:15:06so there's a literature in the united
1:15:07states for example that says you know
1:15:09americans might not be willing to borrow
1:15:11for college because of this risk
1:15:13i don't know anybody who's taken that
1:15:15across countries yet i think that's kind
1:15:16of the current state of the literature
1:15:18that would be an interesting idea yeah
1:15:20because the and
1:15:21this is quite different if you look at
1:15:23the trend of what is happening in many
1:15:25developing countries
1:15:26in developing countries there is a
1:15:28growing uh commercialization of
1:15:30education
1:15:31especially at high school level people
1:15:33spend a lot of money
1:15:34acquiring education and and usually
1:15:38without credit
1:15:41and i i really have not really seen
1:15:45so much in terms of return to that uh
1:15:48such investment because this is an
1:15:50environment where you have
1:15:52very high unemployment rate and all of
1:15:54that uh
1:15:55so i i i i like to get the
1:15:58paper maybe that way it's a proposition
1:16:02that
1:16:02if maybe governments will be able to put
1:16:06in some better policy if it is possible
1:16:08to hedge against
1:16:09unemployment after making those
1:16:11investments in human capital
1:16:12yeah yeah if you send me an email my
1:16:15name is my email address
1:16:17basically my name at gmail i'm easy to
1:16:18find okay i'll send you that paper i
1:16:20want to say one thing real quickly
1:16:21though
1:16:23there's a distinction between two things
1:16:24one is just risk
1:16:26i don't know if i'll find a job i don't
1:16:28know if the job will pay i don't know if
1:16:30my degree is valuable there's a great
1:16:31literature on that
1:16:32there's a second phenomenon which is
1:16:34kind of strange in some developing
1:16:35countries where
1:16:37the college graduate unemployment rate
1:16:38might be 50
1:16:40yeah that's not risk
1:16:45that's like a misallocation i don't
1:16:47think anybody has i don't know i've not
1:16:48seen a good i've always been
1:16:50confused like crazy about that and i
1:16:52don't have any good models why that
1:16:54would be
1:16:55so anyway that would be so if somebody
1:16:57could explain to me just at a really
1:16:58basic level what's going on there i have
1:17:00no idea
1:17:00you you're right you're right i'm i'm
1:17:02worried about that too because that's
1:17:04that's actually that is the reality in
1:17:05many developing countries
1:17:07yeah thank you yeah hello thank you yep
1:17:11yeah i wanted to take on what my
1:17:13colleague just explained now with
1:17:14respect to
1:17:15uh the valid return of education
1:17:17especially
1:17:18territory institutions uh i think uh uh
1:17:21patronus also spoke on this but it
1:17:23should be
1:17:25for about the attire has been debunked
1:17:27that i can remember in the year 2000
1:17:30there was a conference head over there
1:17:31during the dental uh
1:17:33united nations ahead of kofi anna they
1:17:36have a conference i mean summit in
1:17:37senegal
1:17:39with respect to that of the
1:17:41international monetary fund and the
1:17:43world bank
1:17:44and that of the uh the other sister
1:17:46organization the state would give the
1:17:47whole
1:17:49then they have a conference specifically
1:17:51especially in africa most of the
1:17:53developing countries they
1:17:54they don't invest technically in
1:17:56education especially high in
1:17:58education like the other second rate and
1:18:00and and and other for
1:18:02the and other territory institutions
1:18:03simply because while they believe that
1:18:05they return to education
1:18:06it must grow lesser especially for high
1:18:08school and uh
1:18:10and university or college in college
1:18:11education but then
1:18:13most often judging from the what have
1:18:16been experienced in
1:18:17africa especially most of the african
1:18:19countries developing countries is that
1:18:21they invest most of things in the
1:18:22infrastructure development attack of
1:18:24social capital where rahima capital are
1:18:26left behind
1:18:27but in terms of purchasing other
1:18:29different mineral research i mean other
1:18:31different materials that sort of
1:18:32constructions
1:18:34and maybe bringing expertise down to
1:18:35come and walk to come and educate the
1:18:37people over there
1:18:38it is much more like great and expensive
1:18:40more than that of developing nations
1:18:42so if at all quality education has no
1:18:44more i mean return in terms of
1:18:46investment
1:18:47without an investment then technically
1:18:48speaking i'm not quite sure that there
1:18:50will be an advancement in what's in this
1:18:51african conscious because with all the
1:18:53progressions of human capital in a
1:18:55country it is difficult
1:18:56take this as a case study how you expect
1:18:58somebody to construct a
1:19:00bridge or breaches tonight and other
1:19:02different heroes without having the
1:19:04expertise
1:19:04so without advance knowledge of
1:19:06education would be quite more difficult
1:19:08so the quality of education is one of
1:19:09the major problems that we have in most
1:19:11of the african countries
1:19:12especially specifically most of the
1:19:15african country higher institutions they
1:19:17are embedded with political
1:19:19i mean infiltrations instead of
1:19:21institution to be independent of their
1:19:22own
1:19:23and crafting uh uh let's just say uh
1:19:26institutional policy based on
1:19:27state-of-the-art technology or based on
1:19:29state or other educations so the major
1:19:31problem is the institutions
1:19:32so institutional failures are one of the
1:19:34major urban problems
1:19:36that we have in most of the african
1:19:37country thanks
1:19:39thank you
1:19:42let me try to put two more mechanisms in
1:19:46and then i'll stop and take some
1:19:47questions i'm sorry i know i'm running a
1:19:48little late but i'll stay as late as
1:19:49anybody wants to ask me questions
1:19:53there's two other ideas that float out
1:19:54there one is to let
1:19:57another investment matter for human
1:19:59capital production
1:20:01and one that's been shown to work really
1:20:02well this this dates back to manueli and
1:20:04sasha
1:20:05is to let people buy some sort of goods
1:20:07to be used in the production of human
1:20:09capital
1:20:10and goods here you want to think about
1:20:11something that's produced so books
1:20:14chalkboards desks computers paper pens
1:20:16etc
1:20:19if you do that you get a problem that
1:20:21looks like this now the budget
1:20:23constraint says that
1:20:24consumption over my life plus e that's
1:20:27my expenditures on the goods
1:20:29has to be equal to what i earn and on
1:20:32the right hand side the only thing
1:20:33that's changed is that my human capital
1:20:34depends on how long i spend in school
1:20:36and how much i earn
1:20:40if you specialize to the power human
1:20:42capital production function which in
1:20:44practice everybody does
1:20:46then you can show very simply i've got
1:20:47the derivations in the back i'm not
1:20:49going to do them today
1:20:50this does not fix the problem but how
1:20:52long you go to school
1:20:54is the same in all countries but
1:20:57people in richer countries or higher tfp
1:20:59countries
1:21:00will spend more on school and so the way
1:21:03the literature thinks about this is that
1:21:05that could give you more education
1:21:07quality right you have more books
1:21:08chalkboards better teachers or something
1:21:11like that
1:21:12that will give you more expenditures and
1:21:14then you can show that we'll also raise
1:21:15human capital
1:21:17okay so one trick you can do is just to
1:21:19say it's not only my time that i
1:21:20invest i also have to invest resources
1:21:23and the underlying intuition i'm not
1:21:24going to have time to go through all the
1:21:26math
1:21:26the underlying intuition here is simply
1:21:29that rich countries
1:21:30computers are cheap and so if you can
1:21:32use computers to buy education quality
1:21:35people in which countries will naturally
1:21:36buy more computers and they will get
1:21:38more education quality and they'll get
1:21:39more human capital
1:21:41this is probably the most widely used
1:21:43mechanism right now
1:21:45if you want a model that gets you
1:21:46cross-country human capital differences
1:21:48this is sort of what's most commonly
1:21:50used at the moment
1:21:54the last thing you can do is let the
1:21:55government matter
1:21:57through something that looks like public
1:21:59education and the way the public
1:22:01education works is imagine that the
1:22:03government also provides an input to
1:22:04human capital
1:22:06there's many ways to write this down and
1:22:08i'm running out of time so i'll just say
1:22:09quickly
1:22:10the basic idea here is that the
1:22:11government can matter either if it
1:22:12provides so much resources
1:22:14that it's basically you know getting
1:22:16families to invest more than they would
1:22:18have
1:22:18or if the government provides a
1:22:20complementary input to human capital
1:22:22production
1:22:24so there are a smaller set of papers
1:22:26that do this
1:22:28there are a few other there are a few
1:22:30other ideas out there and the one thing
1:22:32i want to say before i conclude
1:22:33i think the big open thing here most of
1:22:36this and most of what i've talked about
1:22:38and really most of the human capital
1:22:39literature in terms of modeling has been
1:22:41aggregate
1:22:42and so especially for a group like stag
1:22:44i think something that's really
1:22:45interesting is that
1:22:46these issues of skill bias technical
1:22:49change
1:22:50where what you the value of your human
1:22:53capital depends on the state of
1:22:54technology
1:22:55or also structural trans structural
1:22:58transformation
1:22:59right maybe the value of your human
1:23:01capital depends on what sectors are
1:23:02growing versus shrinking
1:23:04right maybe human capital is more
1:23:06valuable in a non-agricultural society
1:23:08than an agricultural society
1:23:09these are ideas that i think still could
1:23:11use to be explored more
1:23:12these are things that could be built
1:23:13into models of human capital across
1:23:15countries
1:23:17i think basically i should stop there
1:23:18i'm way over time as it is and i
1:23:20promised i would take questions
1:23:21so let me call it a conclusion those are
1:23:23the two themes
1:23:24somebody else talk
1:23:35uh i have a question about probably the
1:23:37very last point
1:23:38uh on structural change so i i didn't
1:23:41see much on the demand side
1:23:42uh meaning if there are no for instance
1:23:46let's say
1:23:47firms large firms like in western
1:23:49countries demanding for skills
1:23:52um i i see why it would be difficult for
1:23:56people in a developing country because
1:23:59there's no return to such investment you
1:24:02see
1:24:02so i wonder whether for this kind of
1:24:06models this is something that has been
1:24:07implemented if it's more like
1:24:09general equilibrium kind of thing
1:24:11structural change
1:24:13and so on and so forth um
1:24:20so i've seen two related ideas this is a
1:24:22great question i think it's
1:24:23something that's still under explored
1:24:25there's two related ideas that float out
1:24:27there
1:24:29one is you can write down a model where
1:24:30skilled and unskilled labor
1:24:32are imperfect substitutes like you would
1:24:35do in most labor
1:24:36models at this point and then you can
1:24:39have something that's called skill bias
1:24:40technical change right so that that
1:24:42benefits the skilled labor relative to
1:24:44the unskilled labor
1:24:46and then in that model there is an idea
1:24:48that whether you want to become skilled
1:24:50depends on how much skill bias technical
1:24:52change there has been
1:24:54and so i could point you to a few papers
1:24:55that are thinking about the world that
1:24:57way i think that's
1:25:00the most fully worked out model that
1:25:02i've seen that has these ideas
1:25:04a related thing that's out there i think
1:25:06many people have known and shown
1:25:08empirically that the value of skills
1:25:10like the return to a year of schooling
1:25:12is much higher in parts of services like
1:25:15business services
1:25:17there's a really high return to skill
1:25:19it's really low in agriculture
1:25:22so i think there's probably an interplay
1:25:24there which is that as an economy
1:25:25experiences structural transformation as
1:25:27agriculture shrinks
1:25:28and the skilled service sector grows
1:25:30there's probably a rising demand
1:25:32for for education and that might lead
1:25:34workers to get more education
1:25:36um joe do you have the full model of
1:25:38that would you say i don't know he's
1:25:40probably not listening but
1:25:41joe has the closest model i've seen to
1:25:43that but i don't think he has all of
1:25:44that fully worked out i think there's
1:25:45still some room for that idea
1:25:48yeah i agree that there's room i
1:25:51actually there was a
1:25:52related question in the q a and i
1:25:53answered i think that there's an area
1:25:55for research
1:25:58yeah i'm i'm really i mean i'm surprised
1:26:00i haven't done it i'm surprised nobody's
1:26:02done it i'm surprised you haven't done i
1:26:03mean
1:26:04it's been kind of floating in the work
1:26:06but i have not seen anybody fully work
1:26:08it out i think it would be great
1:26:13thank you very much i mean i don't have
1:26:15a full model
1:26:16part of my research is about
1:26:18liberalization of services
1:26:20and how this uh fosters human capital
1:26:23accumulation because
1:26:25this attracts uh new firms
1:26:28uh from foreign firms or private sector
1:26:31firms
1:26:32and this raises the returns to education
1:26:36i mean i've not seen anybody even build
1:26:39that in i agree that seems like a
1:26:40natural starting point right like we
1:26:42know
1:26:43you can just go get i've done this you
1:26:45can go get minster wage equations and
1:26:46just show that schooling is twice as
1:26:48valuable in some of the service sectors
1:26:50than it is in agriculture
1:26:52then it's obvious that i mean that seems
1:26:54like something you could just build into
1:26:55the model
1:26:56pretty easily that at least right that
1:26:58as service sectors grow because of
1:27:00liberalization structures that's what
1:27:01you're talking about right yeah yeah
1:27:03that seems like a great
1:27:04starting point at least i i hope people
1:27:06take these ideas up
1:27:09alicia you had a question about degrees
1:27:11versus years of school
1:27:14yes thank you professor and thank you
1:27:17for your lecture i have a question
1:27:19and in the first approach we treat the
1:27:22workers with and without a degree
1:27:24basically the same but in real life with
1:27:27like
1:27:28because firms only if
1:27:31some workers with certain degree or
1:27:34higher
1:27:35which could cause their wage differences
1:27:37and their
1:27:38human capital accumulation in the future
1:27:40so i think like the
1:27:42degree might matters in in a way i'm
1:27:46wondering
1:27:47do we have evidence in literature shows
1:27:50that
1:27:51the years of schooling matters more than
1:27:53the
1:27:54certain degrees yep great that's a great
1:27:57question so
1:27:59i'll give you two answers one is if you
1:28:01want to just read about this a little
1:28:02bit
1:28:03it has the weirdest name in the world
1:28:04the literature is called the literature
1:28:06on sheepskin
1:28:07effects i'm so sorry i don't know why
1:28:10they decided that was a good idea but if
1:28:11you search for the phrase sheepskin
1:28:14effects like literally skin of sheep
1:28:16you're going to find the papers you're
1:28:17looking for
1:28:20and that's the literature which asks
1:28:22exactly the question you have in mind
1:28:24which is
1:28:24if i know you went to 16 years of school
1:28:27but you didn't complete college versus
1:28:29whether you did does that still make a
1:28:31difference for your wages and the answer
1:28:32is yes
1:28:32the answer is it makes a pretty big
1:28:34difference and so people have
1:28:35investigated that in a wide variety of
1:28:37ways
1:28:39and then they've talked about whether
1:28:40that has something to do with selection
1:28:42or signaling there's a whole literature
1:28:43which is
1:28:45honestly totally different from what i
1:28:46talked about today at a practical level
1:28:49with people in macro development the
1:28:50papers i'm talking about today including
1:28:52my own
1:28:54really in most countries data sets they
1:28:56don't ask how long you went to school
1:28:57they ask what degree you have like do
1:28:59you have a college degree
1:29:00and then what most people including me
1:29:02do is we say well college takes 16 years
1:29:04and we just write 16 years of schooling
1:29:06and then we run the linear wage equation
1:29:10okay so the idea that you need a certain
1:29:13degree or that this degree is more
1:29:15valuable than that one that's just
1:29:18ignored that's the that's the that's the
1:29:20best phrase i can give you for it
1:29:23thank you you're welcome the honest
1:29:25truth is
1:29:29signaling is really tough for human
1:29:30capital people right
1:29:33yeah and when i say degree not only
1:29:36like the degrees also could be the
1:29:40graduate
1:29:41high school and the middle school like
1:29:44the all matters
1:29:45especially for developing countries yeah
1:29:48yep
1:29:49yep i agree completely and like i said
1:29:51it's something that
1:29:55i don't think anybody's really taken
1:29:57into account yet i've not seen it
1:29:59todd there's another great question in
1:30:01the q a q a a becker type question from
1:30:03emmanuel leclaire when we think about
1:30:06borrowing constraints is there also view
1:30:08that a firm may not want to invest in
1:30:10developing technology
1:30:12that supports high skilled individuals
1:30:14because human capital is intangible so
1:30:16not collateralizable
1:30:19and imperfect financial markets may lead
1:30:21firms to invest in technology
1:30:23that promotes the use of
1:30:24collateralizable capital instead
1:30:27yep that's a fantastic question um
1:30:33you it's such a good question it's tough
1:30:37to answer
1:30:39my hunch is yes
1:30:42really when we think about these issues
1:30:47you know this is this is
1:30:50all right let me talk away from the
1:30:51literature now let's throw the
1:30:52literature away for a minute and i'll
1:30:53come back in a second
1:30:55i think most of us would agree that the
1:30:58value of
1:30:59investing in specific technologies
1:31:02plants organizational structures depends
1:31:04very much on the types of skills workers
1:31:06have
1:31:07the types of skills workers are willing
1:31:09to get depends on the plant
1:31:10organizational structure and technology
1:31:12that firms are going to invest in
1:31:14they're complements and if either one of
1:31:15those is constrained it's going to
1:31:17affect both margins that's kind of the
1:31:18idea of the question the way i would see
1:31:20it i think that's exactly right
1:31:23ben jones has a paper where he kind of
1:31:25like talks about this stuff
1:31:27i don't think he has a full quantitative
1:31:29model of it though he kind of
1:31:31hints at it i don't think anybody's
1:31:34really worked it out
1:31:37now coming back to the literature you'll
1:31:40notice my answer to a lot of these
1:31:41things are the same
1:31:43a lot of the really best questions come
1:31:45back to the same thing which is
1:31:49here once you write down this production
1:31:52function that stuff can't show up
1:31:55so what would be really interesting if
1:31:57you wanted to do that i can kind of give
1:31:58you a road map of how to do it you would
1:32:00want to write down again the model where
1:32:02skilled and unskilled labor are like two
1:32:03distinct types
1:32:05and they're imperfect substitutes then
1:32:07you want to have skill bias technical
1:32:09change but you want firms to invest in
1:32:11that and you want the workers to invest
1:32:12in the skills
1:32:14and then you constrain one of them with
1:32:16a borrowing constraint that's a model
1:32:17where you could talk about all of this
1:32:19you'll notice that's not this model this
1:32:20model would not even let you begin to
1:32:22have that conversation
1:32:23so i think that's a fantastic idea that
1:32:25would be really interesting to read
1:32:30okay unfortunately we are um out of time
1:32:34and um todd is willing to stay here but
1:32:37i know uh
1:32:38mandy and kirsty uh it's late in the
1:32:40evening for them over in the uk
1:32:42um so i want to thank todd i want to
1:32:45thank you all for participating
1:32:47um again the todd's talk and his slides
1:32:50are available on the stegg website
1:32:53and tomorrow uh 4 p.m london time we
1:32:56have bertolt herrendorf who will start
1:32:58talking about
1:32:58um structural transformation at the
1:33:01level of sectors
1:33:02um so let's uh let's i've already done
1:33:06it thank you
1:33:07very much todd and uh we'll see you all
1:33:09tomorrow hopefully
1:33:10thank you bye-bye um hello todd
1:33:14hi yeah uh there is an interesting thank
1:33:17you very
1:33:18much for the lecture it's quite
1:33:21inspiring uh there's an interesting
1:33:23perspective here in
1:33:24in germany uh they they have
1:33:27well structured um
1:33:31uh how do i put it now a kind of labor
1:33:35labor market kind of um the
1:33:38the year of schooling is structured in
1:33:40such a way that when
1:33:42a child gets to a particular level it
1:33:44can choose to
1:33:45whether move for a degree or to go for
1:33:48what they call a kind of training
1:33:49what they call house building and um
1:33:53i'm not sure if there's any study on
1:33:54these so far but there have been a lot
1:33:57of concern with respect to that because
1:33:59those who actually go for the house
1:34:02bidding system
1:34:03they tend to progress faster in their
1:34:06career because many of the organizations
1:34:07just need
1:34:08the only job experience
1:34:12you guys know so after like a two or
1:34:14three year of training
1:34:16you go right on into the organization uh
1:34:19if it is a mechanical electrical or
1:34:21whatever depending on the industry
1:34:23and then you get um on-the-job training
1:34:26now
1:34:26how do we factor in for this variation
1:34:29in in some of this model
1:34:30because at the end of the day the guy
1:34:33grows on the job
1:34:35and we i think they reward more
1:34:38on years of experience and and that's
1:34:40why i particularly like that model the
1:34:43the uh mindsets wage equation because it
1:34:46captures both the
1:34:48years of schooling and as well as the
1:34:58experience
1:35:00yes and that's why i think you know
1:35:04we were excited to uh reincorporate
1:35:06experience back into this
1:35:10so at that level you know i think it's
1:35:12taken into account and this
1:35:13what i mean by that is that if what
1:35:16we're interested in is just measuring
1:35:17human capital and making sure we
1:35:19measure experience yeah into human
1:35:22capital
1:35:22that i feel good about you know that
1:35:24that is captured by by these papers
1:35:29there's a second effect which is around
1:35:32the world countries structure their
1:35:34educational systems very differently
1:35:36very differently yeah in some countries
1:35:39people they call this tracking
1:35:40you know when you start to break
1:35:42students into different groups and some
1:35:43of them are labeled the college group
1:35:44and some are labeled the work group
1:35:47some countries do that very early some
1:35:49countries do it very late
1:35:51some countries set up exams so that
1:35:53students study like crazy when they're
1:35:5515 years old in america you don't do
1:35:56anything when you're 15 years old
1:35:58yeah at least i didn't um
1:36:03and i don't think anyone has yet figured
1:36:05out the importance of those institutions
1:36:07you know i'm saying like
1:36:08the fact that it affects the experience
1:36:10profile we've measured that that's great
1:36:12i feel good about that or i don't know
1:36:14as good as i do about anything
1:36:16um
1:36:19but the underlying importance of the
1:36:21institutions i don't think we have a
1:36:22good handle on that
1:36:24that's the truth and if you look by the
1:36:26way if you look at
1:36:27germany germany is a really weird
1:36:29country when you do these calculations
1:36:31barrow and lee tell us that germans have
1:36:33very low education levels
1:36:35because people aren't in schooling very
1:36:37long they go into these vocational
1:36:38training
1:36:39i can't say the german word house
1:36:41whatever yeah
1:36:43but then if you look at their experience
1:36:44profiles they're extremely steep germany
1:36:46has the steepest experience profiles of
1:36:48any country in the world
1:36:50so they're a very unusual country in
1:36:52this sense right very little education
1:36:54and education human capital but more
1:36:55experienced human capital
1:36:57it would be nice to have a theory of
1:36:58that but i don't think anyone does
1:37:00yeah absolutely and i think so to a very
1:37:02large extent
1:37:03maybe and i saw one of the graphs you
1:37:06you
1:37:07presented where the return on education
1:37:10is actually higher
1:37:11even than that of america that of us and
1:37:14all of that and
1:37:15and i i think it's also partly because
1:37:17they have this graded
1:37:19returns definitely if you have a
1:37:22bachelor's degree you don't have the
1:37:24same amount as someone who has a
1:37:26master's degree
1:37:28and then if you have a phd so they've
1:37:30got this graded
1:37:31system that that is actually uh used in
1:37:34in in in how they uh calculate
1:37:37their thing but i don't know and but
1:37:41for you which do you think is the best
1:37:43model
1:37:44for a developing country to approach you
1:37:46know
1:37:47uh adopt for example
1:37:50i think the truth is that there is so
1:37:52little research that's been done on this
1:37:54that i don't even know where to start
1:37:57okay i'm somebody who's really reluctant
1:37:59to make policy prescriptions
1:38:01you know even in cases where i think
1:38:02there's been a lot of work done but here
1:38:04there's been almost no work i
1:38:07like i wouldn't even know what paper to
1:38:09recommend to you to read so that you
1:38:11could start
1:38:12this is this is this is totally
1:38:13uncharted area as far as i know
1:38:15i think this is probably sorry i think
1:38:18this is probably a moment
1:38:19to make clear the distinction between
1:38:21the academic research
1:38:22and the policy conversation which is
1:38:25important
1:38:26and which we hope to inform but where
1:38:30the decisions that get made are often
1:38:33um are often based on much more than the
1:38:36academic evidence
1:38:37right political political forces come to
1:38:39bear
1:38:41it's not clear that the policy choices
1:38:44should be made entirely based on
1:38:46on what the academic evidence suggests
1:38:51because political processes are there
1:38:53for a reason
1:38:54but i think it's a really important and
1:38:56interesting set of questions that you
1:38:57raise about what policy should do
1:38:59i'd love to continue that conversation
1:39:01but i i am conscious that here in
1:39:04in the uk as in other places we're
1:39:06keeping our colleagues at cepr
1:39:09um past the end of their work day and i
1:39:12think we should
1:39:13wind things up i want to thank i want to
1:39:16thank
1:39:16todd i also want to thank the members of
1:39:19the panel the
1:39:20participants who've been here um you've
1:39:22all done just a fantastic job it's great
1:39:24to have these conversations
1:39:26and um and i think
1:39:30one of the things we need to reflect on
1:39:32collectively is this
1:39:33relationship between what we do as
1:39:35researchers what we do as academic
1:39:37economists
1:39:38and and the real deep challenges of
1:39:42making policy
1:39:43um one of the things that the stag
1:39:46program is supposed to do
1:39:47is to support research that speaks to
1:39:50policy
1:39:51and we've all to be very honest we all
1:39:53struggle with what that means
1:39:54i think many of us like todd um feel
1:39:57very uncomfortable
1:39:59with the awareness that what we have to
1:40:01say is only
1:40:03only begins to get to the edges of the
1:40:06big policy questions
1:40:07and i think what what you've raised are
1:40:11abdullahi are really the central
1:40:13questions that policymakers struggle
1:40:14with
1:40:15how should we design our systems in any
1:40:18event
1:40:18i'd like to thank those of us who are
1:40:20still around i would encourage you if
1:40:22you can to unmute your
1:40:23your microphones and let's give todd an
1:40:25actual um
1:40:27old-fashioned out loud round of applause
1:40:30um
1:40:31i just think we want to thank you for
1:40:32taking the time out of your day and for
1:40:34the effort that went into preparing the
1:40:36lecture todd um we'll see you all
1:40:38tomorrow tomorrow
1:40:41todd thanks thanks a ton thank you for
1:40:44the opportunity thanks for that
1:40:46thanks a lot bye everybody bye

About this transcript

This page contains the full transcript of Supplemental Lecture: Human capital in developing countries by CEPR & VideoVox Economics, generated from the public captions YouTube serves with the video. The transcript has 18,649 words across 3,152 segments, with the original timestamps preserved so you can click any line to jump to that moment in the embedded player.

What you can do with it

Use the transcript to take notes, quote the speaker, build a study guide, generate a summary with ChatGPT or Claude via the YouTube Summary tool, or export it as a timed subtitle file with YouTube to SRT. You can also re-open it in the transcriber to choose a caption track available on the video.

Free YouTube transcript tool

YouTube2Text is a free YouTube transcript generator — no signup, no daily limit. Paste any YouTube link and get the full transcript instantly, with timestamps, click-to-jump, caption language selection, AI prompts for ChatGPT, Claude, and Gemini, and exports to TXT, SRT, VTT, or Markdown.