YT YouTube2Text

YouTube2Text → Transcripts → Supplemental lecture: Urbanization and development

Supplemental lecture: Urbanization and development — Transcript

by CEPR & VideoVox Economics · 13,914 words · 2,542 segments · language en · Watch on YouTube

Open in transcriber Summarize with AI Download SRT/VTT

Full transcript

0:28recording
0:53right let's go ahead and start good
0:56morning good afternoon good evening
0:58wherever you are in the world
1:00and welcome to all to this the the last
1:03actual uh lecture in the stag virtual
1:06course on key concepts in macro
1:08development
1:10today's lecture is a supplemental
1:12lecture we're really delighted to have
1:14klaus desmond here to talk about
1:16urbanization
1:17and development it's um wonderful of you
1:20to join klaus
1:22and this is our 20th lecture
1:25it is um really wonderful to see so many
1:28people
1:29who have stuck with us throughout and
1:31i'm delighted to welcome everyone
1:34um i'd like to give special thanks
1:36before we begin
1:38to the team from cepr which has made it
1:42possible to put together
1:43this lecture series and has done the
1:45work behind the scenes
1:47um keeping material up on the website
1:50we're
1:50loading the recordings onto youtube so
1:53i'd particularly like to
1:54single out mandy chan christy mcneil
1:58lauren waring
1:59and nadine clark who've done an amazing
2:02job and made this
2:03and made this work the
2:06the course is part of the steg program
2:08and as we've said many times before and
2:10at the risk of boring those of you who
2:12are now hearing this for the 20th time
2:14steg is a research program steg stands
2:17for structural transformation
2:18and economic growth it's a research
2:21program
2:22funded by the united kingdom's foreign
2:24commonwealth
2:25and development office as part of the uk
2:288 effort it's managed by cepr in london
2:32and we've just been really very
2:35fortunate to find
2:36colleagues across the profession who've
2:38been willing to contribute
2:40lectures to this series and to put a
2:43stock of knowledge out there that we
2:45hoped would
2:46lay a groundwork for doctoral level
2:48teaching
2:49on this topic which isn't really offered
2:51in any one institution
2:53and so um this lecture fits in perfectly
2:57the question of how
2:58urbanization relates to development is
3:00one that's obviously
3:01really fundamental to the processes of
3:04structural change and sectoral
3:06movements before i turn things over to
3:09klaus i just want to reiterate some of
3:10the ground rules which will also be
3:12familiar
3:13to all of you those of you in the
3:15virtual front row will control your own
3:17cameras and microphones
3:19klaus says you're more than willing to
3:21interrupt with questions
3:23as he goes along please feel free don't
3:25be shy
3:26for those of you who are not in the
3:28front row
3:29the way for you to ask questions is to
3:32pose them
3:32in the q a section so use that feature
3:37um klaus won't be able to see those
3:39questions but i and joe kaboski will do
3:42our best to answer as many of those
3:43questions as we can
3:45and the ones that we can't answer that
3:47are really hard and difficult
3:48we'll hand over to klaus and make him
3:50answer um
3:52and i think that's i think those are the
3:54key bits of information that we need
3:57so again without further ado i will turn
3:59things over to klaus
4:01and please go ahead and share your
4:03slides
4:04um for those watching the slides are all
4:07online i'll
4:08i will make sure that the link gets to
4:10you and
4:12please feel free to follow along klaus
4:14over to you
4:16thank you uh doug for the kind uh
4:18introduction it's uh
4:20it's a real pleasure to uh to be here
4:23today and uh to be able to talk to
4:26uh all of you about this important topic
4:29development and urbanization
4:33so i will divide
4:37my lecture today
4:40in two parts most of it
4:43i will dedicate to the discussion of
4:47one particular paper which i wrote
4:51a couple of years ago with a colleague
4:55with esteban rossi hamsburg
4:58on determinants
5:01of the city size distribution
5:04in particular in that paper we looked at
5:08the role of three important determinants
5:12in the city size distribution uh
5:15productivity or efficiency amenities
5:19and urban frictions and then we apply
5:23actually our methodology to look at the
5:26us and china
5:27and so you could ask yourself uh
5:30how is this related to uh
5:33to development well of course as you all
5:37know
5:38uh the development process uh is
5:41typically accompanied
5:42by um urbanization by people moving from
5:47rural areas to uh to cities
5:51and that leads to a number of challenges
5:54right so for example there's always this
5:57big
5:57question whether some of the world's
6:01mega cities are becoming
6:04too big right and this type of approach
6:07that i will
6:08be talking about to you today will be
6:10able to
6:12uh give you a framework
6:15uh to think about that question and to
6:17actually quantitatively
6:19assess that question so that's one
6:23reason why
6:25we think this is important another
6:27reason is that
6:29as you all know in many many countries
6:33uh governments are very worried about
6:37the spatial distribution
6:39of economic activity right so maybe
6:42development is too conscious
6:44concentrated in particular areas and
6:47they want to
6:48spread it over the territory and once
6:51again you can ask yourself
6:53is that a reasonable policy or is it not
6:57you know what are the welfare effects
6:59for example of
7:00a policy that would uh
7:03promote growth in for instance medium
7:08medium-sized or small-sized cities is
7:11that something that
7:12makes sense right so it may make sense
7:14from a territorial perspective
7:16does it also make sense from a welfare
7:18perspective so all these questions one
7:20one can ask by by applying the
7:23methodology
7:25i will be talking about today
7:28and then lastly maybe as as part of the
7:30motivation
7:32is that we think there's a lot of value
7:36in comparing urban systems
7:40uh across countries right so as i said
7:43our
7:44our initial analysis was done
7:47uh on the united states it's obviously
7:50not a developing country
7:51but you know you can learn a lot from
7:53say let me take
7:55this particular approach let me
7:58do the same exercise in china
8:01let me do the same exercise in mexico
8:04let me do the same
8:06exercise in nigeria and try to
8:08understand
8:09what are the similarities what are the
8:11differences
8:12what can we learn from one country
8:16for example might it be the case
8:19that china is slowly moving
8:22into in the direction of
8:26the urban system that we have in the u.s
8:28and if so
8:30what does that imply in terms of
8:33future gains or losses in terms of
8:38welfare right so there's lots of
8:40interesting questions here
8:41and so that's going to be the main part
8:44of my lecture and then towards the end
8:47of the lecture i will want to talk a
8:49little bit about a few
8:52other papers uh just to give you a
8:55flavor
8:56of the other work that has been done and
8:58that is being done
9:00in the area of
9:03urban economics and development or even
9:06more broadly
9:08what we could call spatial economics
9:11and development so the difference a bit
9:13between urban
9:15and spatial would be urban is mostly
9:18focusing
9:20on uh cities
9:23as the word itself says versus
9:26a more spatial economics approach a more
9:29economic geography approach
9:31would would be a little more
9:32comprehensive and take
9:34take the entire territory or the entire
9:38geography of a country including uh
9:40including say uh rural areas
9:44etc etc okay so without
9:48without uh holding you up further let me
9:52now turn to the model i will explain
9:55this in some detail and then show you
9:57some graphs and then move on
9:59okay so um so as i told you in the
10:02beginning
10:03we will be looking at three
10:06determinants of what makes
10:10a city an attractive place to live
10:14right so one determinant could simply be
10:17that
10:18the location has a high level of
10:20productivity
10:21so i can move to a city be highly
10:24productive
10:25earn a decent wage and that makes the
10:27city attractive
10:29a second reason could be attractive
10:32amenities
10:33right so people don't only live in
10:35cities for productive purposes
10:38they also live in cities for
10:43consumption purposes in fact it's
10:45something that that doug
10:46has written about and hopefully we will
10:48get to talking a bit about
10:50some of his work as well uh and then
10:53lastly um
10:56some cities are of course better than
10:59others
11:00in terms of managing
11:03what we call frictions but you can think
11:05of frictions as
11:06urban costs right and this is
11:08particularly important
11:10in developing countries so we all know
11:13that
11:13that cities are great places to be
11:15productive uh
11:17that density is a good thing uh it helps
11:20us
11:21it helps us be productive but on the
11:22other on the other hand
11:24it also leads to urban costs right
11:27cities
11:28tend to be congested cities tend to be
11:30polluted
11:31et cetera et cetera and some
11:34cities are better at managing that than
11:38others
11:38right um so so so that's those are the
11:42three determinants we're going to look
11:43at
11:44and then the second element of the
11:46approach is that
11:47we will not look at a single city
11:51but instead we will look at the entire
11:54system of cities what that means is if i
11:57take
11:57say the u.s or i take china i'm going to
12:00take all the cities in the u.s or
12:02take all the cities in china and what's
12:04important about that approach
12:06is that this is a general equilibrium
12:10approach meaning
12:11that if um
12:15one city is attractive
12:19for example that may be because the city
12:22itself
12:23has some good features but it also may
12:25be because other cities are less
12:26attractive
12:27right so it's a system where people uh
12:30can can move across space and so how
12:33well i do
12:35will obviously depend partly on how well
12:39others do and and i think that approach
12:41of a system of cities is important
12:43because very often when we think about
12:45urban policy
12:46urban policy is very often run by city
12:50governments
12:51right and so a city may say i want to do
12:53this
12:54to make my place more attractive but
12:57that doesn't necessarily mean
12:59that that is the best thing to do from
13:02a systemic uh point of view right and so
13:06taking into account uh the entire system
13:09of cities i think is
13:10is a useful is a useful approach and so
13:13once we have this model and once we can
13:15quantify
13:16we'll obviously be able to run some
13:18counterfactual exercises
13:20and the counter factual exercises will
13:22be a little bit
13:23uh along the lines of what i was um
13:27talking about in in the introduction i
13:30can for example look
13:31at would it make sense uh to uh
13:35support smaller cities would it make
13:37sense
13:38to uh try to slow down
13:42growth in in bigger cities etc
13:45right and then as i said we will compare
13:47the us
13:48and china okay so the model
13:52we have a system of i
13:55cities each city will be characterized
13:58by a level of productivity
14:00by a level of amenities and then a level
14:03of what we call
14:04excessive frictions which reflect this
14:06efficiency of providing
14:08uh urban infrastructure that deals with
14:11that deals with congestion i'll have to
14:13be a little bit more specific about
14:15that third point in a in a second
14:18okay and then we'll have uh n
14:21agents or individuals or households and
14:24they can choose where they live and work
14:27and so the urban equilibrium will be one
14:30where
14:31utility uh equalizes across
14:35space so in terms of technology
14:39so we have goods that are being produced
14:42in these different cities
14:44as i said uh a city i will have a
14:46particular productivity
14:48ait and then production will depend on
14:52that productivity
14:53it will depend on capital and it will
14:55depend
14:56on on hours worked right
15:00capital will be freely mobile across pay
15:02so there will be a
15:03common a common interest rate in terms
15:06of
15:07preferences we will say that preferences
15:10depend
15:11on your consumption right it will also
15:14depend
15:16on your leisure right so i have one
15:18minus h
15:20where h is the hours you work the more
15:22you work
15:23the less utility you have so i have
15:25consumption i have leisure
15:27and then i have this gamma term in here
15:30gamma i
15:31i is a subscript for a particular city
15:34which will denote
15:35the city's amenities right and so we'll
15:38take that as a as a
15:40fixed characteristic in the exercise
15:43we're doing
15:44that just could be that you know if your
15:46city
15:47is one of the students i can see you
15:49have the golden gate bridge behind you
15:51if your city is located in the san
15:54francisco bay area
15:55well that's just a nice location that
15:58will add to the amenities
16:00of of that place right so so then
16:03the agent will obviously maximize
16:05utility subject to
16:07it to a budget constraint and without
16:09going into too much too many details
16:11here
16:12i just want to flag one thing in a
16:14budget constraint here
16:15or a few things one is that um
16:19that there will be a tax rate
16:22tau right which is levied on
16:26people's wages and that tax rate tau
16:30is actually what's going to finance
16:33the urban infrastructure right so people
16:36are going to be living
16:37in in this city they will need to be
16:39able to commute to work
16:41and the city will have to provide urban
16:42infrastructure
16:44call it roads or commuting
16:45infrastructure and that will cost
16:47something and that's what we're going to
16:49be calling it
16:51that's going to be financed by this tax
16:53style
16:55r will be referring to the fact that
16:58people live on on land and they need to
17:00pay rent
17:02and then t will refer to commuting but i
17:05will get
17:05i will get to that in in a second um
17:09okay so you can you can solve this uh
17:12this
17:12utility maximization problem and you get
17:15this first order condition
17:17right which obviously will depend uh on
17:20this
17:21tax tau and so the stack style will
17:25distort
17:26obviously your your your optimal
17:30decision
17:30of how much you want to how much you
17:33want to
17:34work right because it's a tax on labor
17:38and so sometimes in the literature
17:40rather than calling this
17:41tau a tax some people prefer to refer to
17:45tao
17:46as a labor wedge uh what's the
17:49difference between the two
17:51the the difference between the two is
17:53tao as a tax is obviously
17:55a very concrete concept you say okay
17:58there's a tax rate of say
18:00seven and a half percent and that's what
18:02i'm gonna call tao
18:03versus saying no no no tau is anything
18:07that kind of distorts my decision of
18:10how much how many hours i want to work
18:13and that could partly be
18:14taxes but it could partly be other
18:19maybe frictions or regulations or
18:23things that have to do with the context
18:26in which i live
18:27which which distort my decisions and so
18:30if you take that broader view then we
18:33would refer to tao as a labor wedge
18:35right and so what you have to understand
18:37is that when we call it the labor wage
18:39it's partly referring to taxes but it
18:41could partly be referring to other
18:43things
18:44and it's a bit of a black black box kind
18:47of approach
18:48where where we don't really want to be
18:51too specific of
18:52what other things might be in that towel
18:55right
18:56and so anyway that's just a bit of a of
18:59a side note here
19:00and then as i said agents will be
19:04freely mobile across cities and so
19:08utility will equalize across space
19:11okay uh so uh right
19:15for now i haven't really said anything
19:18yet about
19:19uh what it means to be a city the only
19:22thing i've said is there's these
19:23different
19:24points in space that have different
19:26productivity levels
19:27different amenity levels and different
19:29taxes and then people will allocate
19:31themselves
19:32across space until utility equalizes
19:35across all these points but what exactly
19:38does it mean
19:39here to be a city and can we give some
19:41more structure to what it means to be a
19:43city
19:44right and so a very standard model
19:47that is used in urban economics
19:50is called the monocentric city model
19:54and so the monocentric city model uh
19:57what it tries to
20:00what it tries to kind of conceptualize
20:02is the idea
20:04that uh cities are not only productive
20:08places but cities also
20:10cause some kind of congestion because
20:13people need to live
20:14somewhere and then they need to get to
20:16work right and the monocentric model
20:19the assumptions it makes is that
20:22production
20:23happens at the city center right call
20:26that sort of the central business
20:27district
20:28and then people live in concentric
20:32circles around that city center
20:36of course at different distances right
20:38they go away from the center
20:40we will assume that every individual
20:44occupies one unit of land and so as
20:48more people move to the city the city
20:51will grow
20:52and the distance from the furthest away
20:55resident to the city center
20:56will of course also grow and so then you
20:59face a trade-off
21:00right the people who are going to be
21:02living further away
21:04will have to commute longer distances
21:08right commuting is costly you can model
21:10this in different ways you can think of
21:12commuting being costly in terms of time
21:15we we modeled it here in this particular
21:18paper
21:19as commuting being costly in terms of
21:21goods
21:23and so the commuting cost td
21:26d being the distance from where you live
21:28to the city center
21:29will be kappa d right so some kind of
21:33a parameter times times the distance
21:36now as i said people live on on everyone
21:39lives on one unit of land
21:40and pays a particular rent rd
21:44which also changes with distance and we
21:46will normalize
21:48the price at the price of land
21:51at the city border sort of at the fringe
21:54of the city
21:55we will set that equal to zero you can
21:57set it equal to a
21:58concept it doesn't matter but we'll set
22:00that equal to zero
22:01so r d d bar where d bar is the
22:05the city border let's say right the fir
22:08the point furthest away from the city
22:10center
22:10we'll say that the rents over there are
22:12equal to zero and then we will
22:14make the assumption that all agents in a
22:17city
22:18can choose where they live if everybody
22:21can choose where they live
22:23then the sum of the rents that you pay
22:27plus the commuting costs that you pay
22:32that must equalize across people right
22:34and what that
22:35implies is that the person living
22:39furthers away from the city center
22:41will pay the highest commuting cost
22:43actually kappa d bar
22:45but will pay zero rents and then the
22:47person living closest to the city center
22:50will actually be paying essentially zero
22:53commuting costs right
22:54but will pay the highest rents and so
22:56you have a
22:57what they call a rent gradient and
23:00that's a very typical thing that we see
23:02in urban economics it's something which
23:04is very well studied the idea that
23:06rents tend to decrease
23:10with distance from the city center
23:14because it is compensating people
23:18for for longer commutes anyway and so
23:21we have the setup and so then you can of
23:23course if you say everybody lives
23:26on one unit of land if i have a
23:28particular population
23:29nit in the city right
23:32i can calculate how big the city should
23:35be i can calculate the radius of the
23:37city d
23:37bar once i know how many people live in
23:41the city and how big the city is
23:43i can actually also calculate the total
23:46miles traveled
23:47right how how what is the total amount
23:49of commuting
23:50that is being done in the city and
23:53what's interesting about that as you can
23:55see in this
23:55this expression here is that this is
23:58actually convex in in
24:02the number of people who live in the
24:04city and that gives you a little bit
24:06this idea that yes
24:07as cities grow maybe the productivity
24:10may go up
24:11but there's also you know congestion
24:13costs or these costs of of size
24:16are also increasing in in in in a convex
24:19way
24:20and that is actually important for the
24:22following reason
24:23if if if congestion wouldn't be um
24:27increasing in a convex way at the end
24:32everybody would want to live in the same
24:34city right which would be say the most
24:36productive city or the
24:37the city with the best productivity
24:39slash the best amenities
24:41right and so we of course know that in
24:44reality
24:45not everybody lives in the same city
24:47there's many cities across
24:49the territory and one one important
24:51reason for that of course is that
24:54there's a limit to how big a city can be
24:57right and so if you think of a city
24:59stretching a hundred miles for example
25:02but at some point the city becomes too
25:04big
25:05and it becomes too costly congestion
25:08becomes too costly to
25:09deal with and and so new cities would be
25:12formed
25:13okay um so that's kind of the setup
25:17then uh lastly uh
25:21so so we have the city we have this need
25:24to commute
25:25uh the government city government in
25:28this case
25:28will have to provide urban
25:31infrastructure
25:32transport infrastructure uh to uh
25:35to to have these people to have its
25:37residents
25:38uh commute and what we will assume is
25:41that
25:42the the cost of providing this transport
25:45infrastructure
25:46will depend on the total commuting cost
25:49which is this this expression here that
25:51you see at the end
25:52times the kappa and it will also depend
25:55on
25:56wages right of a worker
25:59or workers wage is actually w times h
26:02but then it will also depend
26:04on this uh this variable which we call
26:08git and this is what we will call the
26:11inefficiency or the frictions or these
26:13excessive frictions
26:14of of different city governments and so
26:17to be
26:18to be to be very clear what we kind of
26:21mean by g is the following
26:25g is not referring to the fact that
26:28bigger cities
26:29become more congested it's it's more the
26:32idea that
26:33if i take two cities of equal size
26:37right that are equally congested right
26:41in terms of how much commuting happens
26:43etc
26:44some city governments are better
26:47are more efficient at
26:51dealing with this than others they're
26:53they're more efficient at providing this
26:55infrastructure than others
26:56the one that's more efficient would have
26:59a higher g
27:00sorry would have a lower a lower g the
27:02one that's less efficient
27:04would have a higher g right and so
27:06that's why we in the paper we refer to
27:08this as excessive frictions
27:10because typically yes as cities grow
27:12frictions increase
27:14but but given those that increase in
27:17frictions
27:18some are bet some some city governments
27:20are better and some are worse
27:21at actually providing those those uh
27:25that infrastructure they're better or
27:26worse in terms of
27:28dealing with uh with those frictions
27:31okay so
27:32so that's kind of the idea and
27:36and and of course all these costs of
27:38providing transport infrastructure is
27:41then what will be
27:45financed by that labor
27:48tax towel okay um
27:51all right well um then of course
27:54characterization of equilibrium is
27:55actually very simple
27:56as i said you know people just allocate
27:59themselves across space
28:00until you tilt the equalizers and so
28:03what we really only need
28:04so so just to to sort of simplify a bit
28:08the only uh the only things i kind of
28:11need to know
28:13to solve this model is if i have
28:15information on
28:17productivity and on amenities
28:21right i have information on a i have
28:24information on gamma
28:25and i have information on this
28:28efficiency or these frictions with which
28:31governments supply infrastructure just
28:35git
28:36if i have a triplet if i have this
28:38triplet
28:39for all cities i then
28:43i can actually solve the model right i
28:45can me when
28:46by solving the model what i mean is i
28:49can tell you
28:50how many people will live in every
28:52single place
28:53right which is the main thing we're
28:55looking at because once i know that
28:57i can of course start looking at what
28:59does that mean in terms of
29:01city size distribution what does that
29:02mean in terms of
29:04welfare etc etc okay and of course
29:08as you can imagine the in in equilibrium
29:11we will find that more productive cities
29:13will be larger
29:15we will find that cities with better
29:18larger
29:18amenities are also larger and then we
29:21will find that cities with these
29:23higher excessive frictions higher git
29:26will be smaller right and so what we
29:30will do now
29:31is i will show you very briefly how we
29:34get
29:34data on these on on this triplet
29:39for all metropolitan areas in the u.s
29:42and then i can show you
29:43what we do in terms of exercise uh
29:47and then i will show you what we do
29:50for china and the comparison between
29:54the two will actually be i hope uh quite
29:57uh
29:57quite interesting okay uh
30:01are there any questions on
30:05the model part
30:11i was wondering a bit um how realistic
30:14it is to have these taxes be um
30:18being like on a city level
30:22yeah so so so uh you're actually quite
30:25right some of the more
30:27simple urban models would just say look
30:30let us just consider productivity and
30:33amenities as a two main characteristics
30:36that make cities different
30:39and that's kind of enough to understand
30:43most of the insights we can get from
30:46urban economics
30:48here as i said and and that's why
30:52you know that that's why your your
30:55question
30:56is useful um
30:59you can think of it as being a tax
31:02and then you can of course ask yourself
31:04yes well how how might how which part of
31:06these taxes are actually
31:07local taxes which part are state taxed
31:10which part of federal taxes
31:11it's obvious that part of the urban
31:14infrastructure
31:15will actually be let's say city roads
31:18right
31:18they typically are financed by
31:21property taxes which are local taxes
31:24right and so from that point of view we
31:26think
31:27our assumption is actually quite
31:29reasonable a lot of the
31:31local commuting infrastructure is
31:34actually financed
31:35by by local taxes uh
31:38but but again it's a bit of a broader
31:40concept
31:42uh you can go beyond the idea of
31:45of of strictly thinking of of it as
31:47being a tax
31:48and go to this idea of a labor wedge
31:52which would which would simply be
31:54telling you that
31:56some cities seem to be more efficient
31:59in terms of dealing with say
32:02congestion or dealing with urban
32:04infrastructure than others
32:06right um yeah so so those are the two
32:10parts of the
32:10of the answer but when we when we sort
32:13of
32:14get estimates of the style and when i i
32:18actually correlated those estimates of
32:20tao that are sort of predicted by the
32:21model
32:22and and i compared it to uh local taxes
32:26the correlation is actually quite high
32:28so so
32:29so it correlates well our notion of tau
32:33in the model actually correlates quite
32:35well with
32:37with the local taxes so that's that's
32:40what i can say about that for now
32:42um anything else any other quick
32:45question
32:45klaus i have a clarifying question for
32:47you in the
32:49q a from veronica goyal linking this
32:51back to the misallocation lectures that
32:53people have had
32:54yes if people relocate to a city because
32:57of
32:57amenities not productivity
33:00can that lead to misallocation in the
33:03aggregate
33:04how do you think about what's uh what's
33:06interesting of course is that
33:10when you have only one characteristic so
33:12suppose we say the only thing we have
33:15here
33:15is productivity then then
33:19then all these predictions are
33:20relatively straightforward right you
33:22want people to go to places
33:24that are more productive and leave
33:26places that are less productive and
33:27you're done
33:28over here it's a bit more complicated
33:31because
33:33not only do we have three determinants
33:36but
33:36the question is what are the
33:38correlations between them
33:40right is it typically the case that
33:42higher productive cities
33:43are also higher amenity cities that may
33:46be the case yes or maybe no
33:48and so uh
33:53so so so what what we can say is that
33:56if people move to cities with higher
33:59amenities
34:00that will of course be that that will
34:03improve
34:03their utility right and so it will
34:07improve welfare but but it may actually
34:10decrease
34:10uh it may decrease efficiency if it
34:14turns out that the higher amenity
34:16cities are cities with lower with lower
34:19efficiency
34:20right and so if you take if you take a
34:23view of
34:24misallocation as only referring to
34:27misallocation from the efficiency point
34:29of view
34:30versus if you take a view of
34:32misallocation
34:34from a welfare point of view right and
34:35so we're talking here about
34:37welfare and so yes it is possible that
34:41as you are maximizing welfare by letting
34:44people move
34:45where they want and they might be going
34:47to places with better amenities
34:49that that efficiency actually goes down
34:52that's possible
34:55okay sorry i also have a question about
34:58the
34:59labor tax yes
35:02i'm not sure whether i understand it
35:04right but uh
35:06in the production function it seems that
35:08the
35:09liberal text doesn't distort the
35:12production i think so the which is what
35:15it does is it it
35:16it affects uh if you go back here it
35:19affects
35:20the choice of the hours worked
35:24right and so you can kind of see this
35:26here from
35:27the tao entering into your into your
35:30budget constraints so so the tau will
35:33depend on the hours you work so it's not
35:35it's not a sort of it's not a tax that's
35:38fixed
35:38per person it's a tax that depends on
35:41how many hours you work
35:43and so when you think about the
35:45equalizing marginal
35:48marginal benefit and marginal cost
35:51higher labor taxes will
35:54push you towards working uh less
35:58hours which is something you can see in
36:00this first sort of condition here
36:02right and so so in that sense it is a
36:05it's a sort of a standard distortionary
36:09labor tax okay
36:14okay one more question um
36:18how about the
36:21the the key assumption is that the the
36:23cities are homogeneous right
36:29in what sense i mean are we looking at
36:31the structure
36:33and the factors that uh
36:36determine uh these
36:39cities are common to all is that is that
36:42is it clear to say yeah so so well so so
36:46uh so of course each city
36:49has a different so here as you can see
36:51this ait the gamma i t and the git are
36:54really the
36:55the three fundamental drivers and so the
36:58only thing is that the i
37:00so so the structure is the same yeah so
37:02the the model is of course the same
37:04model and what makes one city
37:06attractive or another it's the same
37:08types of determinants
37:09but but they do have an eye subscript
37:12meaning
37:13these are three characteristics which
37:16are
37:16city specific so different cities will
37:18obviously have different levels of these
37:20three
37:21of these three determinants but
37:23otherwise yes the structure
37:25the structure is the same for all the
37:27cities okay how do we account for
37:30endowment and maybe some cities are
37:33endowed with certain
37:34uh maybe resources um yeah so so that
37:38would actually in
37:39that would show up in our in a higher
37:42ait
37:43okay right so a city that is for example
37:46either well endowed with some natural
37:48resources or maybe
37:50uh located in a port in in a place
37:53that's
37:54that's easily accessible that that that
37:57would that would show up in the ait
37:59okay thank you okay all right
38:02and so then what do we do we have data
38:04on all these metropolitan areas in the
38:06u.s
38:07uh production we measure by gdp
38:11consumption we don't really have data on
38:14consumption by
38:16by msa by metropolitan area and so we
38:19actually use uh
38:20retail earnings and then adjust for
38:23national averages
38:25to get to get
38:28a measure of consumption at the city
38:30level and we add to that
38:33housing housing consumption
38:36capital we take actually u.s sector
38:39capital stocks
38:40and allocated to msas according to
38:44their shares and in sector learnings and
38:46then for hours work we use
38:48the current population survey um so how
38:51do we get ait
38:53ait we just get from this expression
38:56here so we have a measure for y
38:58for k and for h so we get ait that way
39:02the labor wedge of course we need some
39:05assumptions on what
39:06size and theta etc but we have a measure
39:09of
39:10of c of h and of a so we can back out
39:14tau then the excessive frictions git we
39:18actually take the govern
39:19government budget constraint uh we take
39:22logs
39:23and then we run this regression here and
39:26the residual is actually
39:28what we call uh uh is the log of git so
39:31we can back out git this way
39:33and then the manatee levels we it is
39:36kind of a residual
39:38right so we we use the model
39:42and then
39:45determine the amenity levels gamma i t
39:50so that the model generated
39:53distribution of people across cities
39:55actually matches
39:57the one that we see in the data
40:00okay all right so let me
40:04and then we also have a way of getting
40:07getting an estimate of kappa right kappa
40:10is if you remember that's the cost of
40:13the cost of commuting per per mile or
40:16per kilometer
40:18traveled okay so so with all of this
40:21we can you know once we have the as i
40:22said once you have these triplets for
40:24all the msas
40:26and you have parameter values you can
40:28just uh
40:29allocate people across cities and and
40:31then you can start doing some exercises
40:33okay that's that's the nature of what
40:36we're doing here uh
40:38here's one graph just sort of
40:40introductory
40:41uh to because you'll see a lot of these
40:43types of graphs
40:44so these types of graphs you may have
40:47seen in other
40:48lectures or in in your career in
40:51economics
40:52this type of graph to represent size
40:55distributions
40:56you know farm size distribution or
40:58whatever here it's the
41:00city size distribution and so what we
41:02have on the
41:04x-axis the horizontal axis is the
41:08log of population and what i have on my
41:10y-axis
41:12is the log of the probability
41:16of being bigger than the population on
41:18the on the on the x-axis
41:20so that's a little bit of a of a
41:22mouthful
41:23uh but but the way to maybe easier for
41:26me to say is
41:27the way to interpret these these graphs
41:30are
41:30if the graphs are flatter right
41:34say the red one or the orange one
41:36[Music]
41:37then you have more dispersion
41:40in the city size distribution you know
41:43on the top
41:44the top kind of the smaller cities right
41:46so i have more small
41:47cities and i have more large cities
41:51right so it's a little more dispersed if
41:54if it's a little steeper right then
41:57there's less dispersion meaning
42:01uh the city sizes are of more
42:04equal equal size and so
42:10excuse me just as a as an example here
42:13for instance if you compare
42:16the red the red
42:21the red curve to the blue one
42:25the red one is for a lower level of
42:28commuting cost
42:29than the blue one so think about what
42:31that means
42:32intuitively if if i increase commuting
42:36costs
42:37who is that going to be bad for well
42:39that's typically going to be bad
42:41for cities where there's a lot of
42:42commuting right so that will be
42:45bad let's say for the bigger cities
42:47right and so
42:48higher kappa would tend to make
42:52the bigger cities smaller and the
42:54smaller cities bigger
42:56right that explains why the blue or the
42:59light blue curve
43:00is kind of steeper than
43:04the red one so the red one is for lower
43:07commuting costs
43:09it's not as costly to be big so i see a
43:11more dispersed city sizes
43:13distribution for higher commuting cost
43:16i see a less dispersed city size
43:20distribution
43:21okay and of course then you can look at
43:25utility and welfare etc but let me move
43:27on
43:28in the in the interest of time here so
43:30so here's the first
43:32exercise where what we do
43:35is the following we say
43:39what would happen to the city size
43:43distribution
43:45if for example we eliminated
43:50all productivity differences across
43:52space
43:53right or what would happen to the city
43:56size distribution
43:58if we eliminated all amenity differences
44:01across space
44:03and so just just to explain a little bit
44:05better
44:06the blue curve in all four pictures
44:10is the actual city size distribution
44:13right the ones that are colored in
44:16in another color not blue are counter
44:19factual city size distributions
44:21so let me for example uh
44:24call your attention to the one on the
44:26top right
44:28so the green curve is
44:31the city size distribution that i would
44:34get
44:35if all cities had the same
44:39efficiency level in the u.s right
44:42the purple one here at the bottom left
44:45is
44:45what would the city size distribution
44:47look like if all cities had the same
44:49amenity levels and and the black one is
44:53what would the city size distribution
44:54look like if all
44:56cities had the same level
44:59of of excessive friction so what you see
45:03is a couple of things for example if you
45:05look at efficiency
45:07yeah you know the city sizes region
45:08doesn't look that different
45:10there's a little bit uh
45:14you know there's maybe a little bit this
45:16is some more smaller cities
45:19but on the top part of the distribution
45:21you don't really see a very big
45:22difference
45:23in terms of utility equalizing
45:27efficiency of cross space
45:29now you know the utility goes from 10 to
45:3210.12 so
45:33there's a little bit of a improvement in
45:36utility but again it's not a whole lot
45:38right so the the welfare effects is
45:42a kind of an increase by 1.2 percent so
45:44the welfare effects of equalizing
45:47efficiency across space in america is
45:49actually not that big
45:51although you do get a lot of
45:53reallocation
45:54so the reallocation here is measured by
45:58uh the the increase in the population
46:01and the cities that are growing
46:03divided by the total population so you
46:06do get a lot of reallocation
46:08a lot of people moving across space but
46:10the city size distribution is not
46:12changing very much
46:13and overall utility is not changing very
46:16much either
46:17and so you look at these utility numbers
46:19here and that's true for efficiency
46:22it's true for equalizing amenities
46:25across space it's true for equalizing uh
46:29[Music]
46:30excessive frictions across space but but
46:32in all three cases the reallocation is
46:35relatively big
46:36welfare effects are relatively small
46:38city size distribution doesn't change
46:40very much maybe except in this
46:42particular case
46:43where you get a much less dispersed
46:47city size distribution so that's an
46:49interesting
46:50first observation right
46:53you can of course do many things with
46:56this
46:58rather than looking at the the the
47:01entire distribution
47:02you could say let me look at
47:05certain particular cities what happens
47:08to them
47:09right so for instance let me look at
47:12efficiency here
47:13if you equalize efficiency across space
47:17look at the city size distribution you
47:19don't see that much happening
47:20but but when you look at the data for
47:22individual cities
47:24the effects are actually quite big yeah
47:26well not surprising the most efficient
47:28cities will obviously lose
47:30that's what happens when you equalize
47:32efficiency across space of course others
47:34that are less efficient will gain
47:36right and so uh at the individual city
47:39level
47:40you see big changes that's of course
47:42what then explains
47:44the big reallocation across space you
47:47know there wouldn't be bigger
47:48reallocation across space
47:50if if there's no big effects at the
47:53local level
47:54right uh the same is true with amenities
47:57so if you equalize amenities across
47:59space
48:00the east coast what gained
48:03what does that mean well that means that
48:05these places don't have very good
48:06amenities
48:07right because if they get the average
48:09amenity they'll do better
48:11the west coast loses but that would
48:14tell you that they actually have good
48:15amenities if you give them the
48:17average amenities they will end up doing
48:19worse
48:22right so so there's lots of interesting
48:24if you're interested in particular
48:26cities
48:26there's lots of interesting insights uh
48:30and there's lots as i said lots of
48:31reallocation but not a whole lot of
48:34welfare gains and you can look at this
48:37in maps
48:38for example one interesting thing is
48:40that if you eliminate
48:42or if you make all these frictions equal
48:45across
48:46cities the places that are gonna gain
48:49are the brown
48:50the brown areas so you see a lot of the
48:52cities actually in the rust belt
48:54uh would be gaining a population
48:58if they had the same level of frictions
49:00of the rest of the country i thought
49:02that was kind of an interesting
49:04and interesting finding so lots of lots
49:07of
49:07cool stuff you can do with this um
49:11but let me let me know for the for the
49:15remaining 10 minutes before i sort of
49:17switch to the second part of the lecture
49:19let me actually uh talk a bit about
49:23uh china okay um
49:26so so we've seen this exercise for the
49:29u.s
49:30let's do the same exercise for china
49:32right
49:34well a couple of interesting
49:37differences let me focus on the one of
49:41productivity then i'll talk a bit about
49:43amenities as well but
49:44let's look at the top right graph over
49:48here
49:48so once again the blue curve
49:52is the actual observed
49:55city size distribution the green one
49:58here
49:59is how would the city size distribution
50:02look like
50:03in china if everyone had the same
50:07level of productivity okay so the first
50:11thing you notice
50:12is that
50:15uh as you can see this curve has become
50:17flatter right
50:18flatter remember means more dispersion
50:21the bigger cities are actually becoming
50:23bigger the smaller cities are becoming
50:25smaller
50:26right that's interesting uh
50:30because you know we've talked a bit
50:31about uh mega cities
50:34are mega cities desirable are they not
50:36desirable etc
50:38here we're kind of finding hey you know
50:40if everybody had the same efficiency
50:41level
50:42uh we would see even bigger cities than
50:45the ones we see today
50:47in china and the cities are as you know
50:48already very big in china
50:50right and so if anything we'll see
50:54more big cities and bigger big cities in
50:56china
50:57if there is a efficiency efficiency
51:00equalizes across space
51:01the other very interesting thing here is
51:04remember in the case of the us
51:06we had a welfare increase of 1.2 percent
51:12uh when you equalize efficiency across
51:14base here you have a wealthier increase
51:16of 46 percent right or 47 percent
51:21uh log point but let's let's just call
51:24it 47 percent here
51:26huge difference huge welfare gains
51:29from equalizing efficiency across space
51:31so so that of course
51:32ties in i think probably with
51:36some of the discussions you've had on
51:38misallocation
51:40in your in your other lectures and so
51:43what is very interesting to me about
51:45this is
51:47you know at some level it shouldn't
51:49maybe surprise you
51:51right and the reason why maybe this
51:52shouldn't surprise you
51:54is well as with many other
51:58emerging economies
52:01you know the difference in efficiency or
52:04in productivity across space
52:06in china is way way way way bigger
52:09than the difference in efficiency across
52:12space in the us
52:14right in mature economies yes of course
52:16there's more efficient
52:17places and less efficient places in the
52:19u.s but
52:20the differences are not that big in
52:22china these differences are an order of
52:24magnitude bigger they're
52:26large right you take the you take the
52:28most advanced
52:30cities on on the eastern seaboard in
52:34china
52:34well these cities are very similar to
52:38cities in very advanced economies you go
52:40all the way
52:41to western china and you will find
52:44cities that are still very far behind in
52:47terms of in terms of development
52:49and so that's interesting to observe and
52:52so in that sense yes obviously there's a
52:54lot of misallocation
52:55it's not real misallocation because the
52:57efficiency is whatever it is
52:59but but there's misallocation in the
53:01sense that you know
53:02the efficiency hasn't spread to all
53:04these places yet if it would
53:07then then you would see these big you
53:09would see this big
53:11you would see this big welfare welfare
53:15gains
53:16the other thing that's interesting about
53:17this counterfactual exercise
53:19is that of course if you say let me uh
53:23sort of fast forward a couple of decades
53:26right maybe 20 30 40 years
53:29we kind of know where china is heading
53:32obviously
53:33because of technology diffusion across
53:36space
53:37it's very clear that these very
53:40large efficiency differences across
53:43cities in china
53:44will become smaller right that's the
53:46standard thing that happens
53:49with space and development development
53:52typically in the early stages is very
53:54uneven across
53:55space but then eventually the efficiency
53:58diffuses across space
54:00making efficiency more equal right and
54:03so
54:04we we if we had i mean of course we
54:07don't know but if you had to make a
54:08prediction
54:09it's very likely that the spatial
54:11differences in efficiency will become
54:13smaller
54:13and if that's true then it's very likely
54:17that we'll see something like this
54:18happening in china
54:20and that the potential welfare gains
54:23from that
54:24are actually pretty large and what's
54:26interesting about this exercise is that
54:28when we say we're making efficiency
54:30equal across space
54:31it's not that we're giving everybody the
54:34best efficiency in the country
54:35we're just equalizing it to the average
54:38right so we're not
54:39increasing we're not increasing
54:42average efficiency right the population
54:44weighted efficiency is the same
54:47as as as as the one in the data
54:50right we're just allowing it to be more
54:53equally spread across across
54:57the geography right and so we we thought
55:00that was very interesting because it
55:01it's really very very different from
55:04what we saw
55:06in in the case of the us and again big
55:08reallocation et cetera et cetera
55:11okay so uh
55:14a few more things uh before
55:17i finish with this uh part
55:21of of the lecture um
55:25is is the following i just want to give
55:27you a flavor of
55:28what other types of exercises could you
55:31do
55:32with with this uh with this framework
55:35so uh one thing we did is
55:39we looked at uh
55:42a policy or we compared a
55:46policy that would target uh
55:50lagging cities meaning cities that are
55:53behind
55:53in terms of either efficiency amenities
55:55or frictions
55:56versus a policy that would target
56:00either medium-sized or small cities
56:04and and try to look at what would give
56:07us
56:08uh the greater or the bigger welfare
56:11effects
56:11and the reason why we wanted to look at
56:13that is because
56:16china for a while at least
56:19had this policy which i mentioned here
56:21at the top in in
56:22italics of controlling the big cities
56:26moderating the development of
56:28medium-sized cities
56:29and encouraging the growth of small
56:31cities so you can say okay
56:32what would happen if we actually gave
56:35incentives for small cities to grow
56:38would that be a good thing
56:40so so here's what you see for instance
56:42if you look at
56:44the first row of the table this would be
56:48what happens if i take the worst cities
56:52in terms of either efficiency amenities
56:54or frictions and i improve
56:56let's say their efficiency by 20
56:59okay or i improve their amenities by 20
57:03what are the welfare effects in the
57:05aggregate
57:07right to get a 5 welfare increase you're
57:09getting 11 welfare increase
57:11in amenities versus if you target the
57:14same type of policy
57:16but you look at the smallest cities so
57:18you you
57:19increase the smallest city improve small
57:21cities by 20
57:23you typically get smaller effects
57:27so anyway what this means is you could
57:30use
57:31a framework like this to uh to look at
57:34different policies
57:36right it doesn't make sense to target
57:38the smaller cities
57:39right or or or again does it make sense
57:43to try to slow down
57:45growth in big cities is that is that a
57:47reasonable policy
57:48from from the welfare point of view um
57:51[Music]
57:53so so that's that's kind of the
57:55framework here we actually in
57:57in a in in another paper also compared
58:01uh this exercise or we implemented this
58:04exercise for
58:06from mexico and in mexico
58:09you find something which is somehow
58:11somewhere in between
58:12china and the us but actually much more
58:15close
58:16to the us than to china right so we
58:18don't find
58:19a particularly high welfare effects in
58:22mexico either they're a little bit
58:23bigger than in the us and so
58:25yes it's somewhat in between but but
58:28it's much closer to the us than
58:30than to china but anyway the good news
58:33is
58:34if you're interested in this type of
58:35approach the data requirements
58:39to do these types of exercises are
58:43relatively mild you don't need a whole
58:46lot of data
58:48and as i said i think there's a lot to
58:50learn
58:51from uh doing this type of analysis
58:55in different countries right we learned
58:58so much
59:00by doing this for china right
59:03for in some sense i feel that this paper
59:07became really interesting when we did
59:09the china exercise
59:10right just because the comparison uh
59:13taught us so many interesting things
59:15and again china is obviously just one
59:17country it's an important country
59:19and it's an interesting country but
59:20there's other countries out there
59:22you know there's an urbanization is
59:24happening everywhere
59:26and it's not happening in the same way
59:28everywhere right and so
59:30you know maybe it is interesting to do
59:32some of these exercises in
59:34in some african countries that are going
59:36through a process of
59:38of of urbanization or you can do this
59:41for other
59:42south asian or south american countries
59:45again there's a lot to learn from doing
59:48this
59:49for for countries that are going through
59:51a process of urbanization
59:53in some sense there's probably more to
59:56learn from those countries than
59:57from doing this exercise for france
1:00:01or for italy where things are fairly
1:00:04stable and so you can probably guess
1:00:06more or less what you will find
1:00:09okay the other
1:00:12is there any quick question before i
1:00:14move on
1:00:15and talk very briefly about a few other
1:00:17papers
1:00:23okay uh i i was
1:00:26thinking about uh either way we can
1:00:29introduce like
1:00:30uh friction related to uh human capital
1:00:34institutions like only people with
1:00:37some minimum level of an educational
1:00:40admin
1:00:42are allowed to be in a place as opposed
1:00:44to
1:00:45people who are lord uh
1:00:48other do they create more productive
1:00:52cities or yeah so
1:00:56so i so you're asking about whether
1:00:59there's any role in these models for
1:01:01human capital right yeah um
1:01:04yes there is i mean we we obviously in
1:01:07the paper that i have presented
1:01:09i haven't said anything about human
1:01:11capital but there is a large
1:01:14and growing literature in in urban
1:01:17economics
1:01:18that has been looking at what they call
1:01:22sorting right and it's typically sorting
1:01:25of people with different skills to
1:01:28different types of cities
1:01:30right so you're absolutely right that
1:01:32cities are
1:01:33different not just because of some kind
1:01:35of exogenous level of productivity
1:01:38they're also different because they
1:01:39attract different type of people
1:01:42and there's growing interest in that so
1:01:44for a while if you go back say 50
1:01:4660 yards the typical urban
1:01:50the the urban models that were looking
1:01:52at city specialization
1:01:54uh they were they would typically look
1:01:58at
1:01:58sexual specialization they would say oh
1:02:00that city
1:02:01produces cars and the other city
1:02:03produces furniture and the third city
1:02:05produces something else
1:02:07now there's much more interest
1:02:10in not really what they call sexual
1:02:12specialization but
1:02:14a functional specialization and sorting
1:02:17of
1:02:18people of different skills to different
1:02:21cities
1:02:22right so what makes a city an attractive
1:02:26place
1:02:27for someone who's highly educated right
1:02:31and it's a very important question and
1:02:32city governments are obviously super
1:02:34interested in that because they all want
1:02:36to track those types of people
1:02:38maybe they are looking for urban
1:02:39amenities right maybe
1:02:41they're looking for like-minded people
1:02:44and because of that you get these you
1:02:47get these phenomena
1:02:48of sorting right so people people want
1:02:50to live in a place
1:02:52partly the sorting happens because of
1:02:54productive reasons but partly the
1:02:55sorting also happens because of
1:02:57uh utility and consumption reasons right
1:03:00there's
1:03:01people may just want to live in a
1:03:02particular area or particular city or
1:03:04particular neighborhood
1:03:06because they find like-minded people
1:03:08right and so
1:03:09uh being attractive for human capital is
1:03:12is interesting again in this particular
1:03:15paper
1:03:16we we didn't consider a heterogeneity
1:03:18and skill
1:03:19but that's something that could be
1:03:21introduced and as i said there's
1:03:22there's lots of interest in urban
1:03:24economics in those in those types of
1:03:26questions
1:03:28thank you yeah thank you a really quick
1:03:31one for you
1:03:32um in in these models does zipf's law
1:03:35hold
1:03:35for city size distribution um
1:03:39that's a that's uh that's an interesting
1:03:41question so
1:03:42so yes it i mean the zips law holds
1:03:46as you know for for typically for for
1:03:49pretty much
1:03:49all the city sizes and size distribution
1:03:52so
1:03:53so yes typically it holds where
1:03:55sometimes it breaks down a little bit
1:03:58is at the very upper end you know so
1:04:01there's some if you look say at the top
1:04:03three or top four cities across the
1:04:06in different countries it's not obvious
1:04:08that zips law
1:04:09holds there you know you have some
1:04:12countries where the
1:04:13where the biggest city really has say
1:04:15half of the population
1:04:17and then the ones that kind of fall
1:04:19below that
1:04:20uh don't quite uh you know so so the
1:04:24at the very upper tail it doesn't
1:04:26doesn't necess doesn't necessarily hold
1:04:28ziploc
1:04:28but but let's say in in the model that
1:04:30we present
1:04:31there's nothing that that generates
1:04:34zips law i'm just more talking about
1:04:37does it hold in the data
1:04:39right there's there's nothing in the
1:04:40model because the model is not really a
1:04:42model of
1:04:43of of dynamics right so so
1:04:47there's lots of interesting models that
1:04:49urban models that
1:04:51generate zips law you know they sort of
1:04:53have a model where
1:04:54there's particular shocks that happen
1:04:56and these shocks over time
1:04:58lead to a city-sized distribution that
1:05:01that actually satisfies
1:05:03zip's law this this is a very
1:05:05interesting literature as well and
1:05:07we should probably let you move ahead
1:05:08here yes let me move ahead and
1:05:11i just want to in the second part in the
1:05:13remaining 10 or 12 minutes that i have
1:05:16i want to mention a couple of other
1:05:18issues in urban
1:05:20economics and development just to
1:05:21broaden the discussion a little bit
1:05:23okay so uh
1:05:27so let me start with a handbook chapter
1:05:30that i wrote with vern henderson a
1:05:34couple of years ago
1:05:35where we kind of look at the question of
1:05:38what do we know
1:05:39about the spatial distribution of
1:05:41population
1:05:42and how that changes with development
1:05:46and so one approach that we'll be very
1:05:49familiar with you
1:05:50for you because you've studied a
1:05:52structural transformation in so much
1:05:54detail
1:05:55over the past weeks and months
1:05:59is there there's there's models where we
1:06:02say okay
1:06:03if the income elasticity for food is
1:06:06less than one
1:06:07then of course development will lead to
1:06:11industrialization
1:06:12right and when development leads to
1:06:15industrialization
1:06:17then kind of that also implies
1:06:20urbanization
1:06:21to the extent that industry tends to be
1:06:23more urbanized than agriculture
1:06:25right of course another approach which
1:06:28you're also familiar with from your
1:06:29classes on
1:06:31structural transformation is you can
1:06:34make the assumption that the elasticity
1:06:36between food and non-food is less than
1:06:38one
1:06:38in that case industrialization
1:06:42results from higher productivity growth
1:06:45and agriculture
1:06:46and so uh that kind of uh
1:06:50that kind of means that to have an to
1:06:54have
1:06:54industrialization or to have an
1:06:56industrial revolution if you wish
1:06:58you need to first have an agricultural
1:07:00revolution
1:07:01right and so you get you get
1:07:03productivity growth in agriculture that
1:07:05moves people into dust industry an
1:07:08industry
1:07:09is simply more urbanized than
1:07:11agriculture and that's partly because
1:07:14of course industry is less land
1:07:16intensive right
1:07:18so it tends to be more concentrated and
1:07:21secondly it may also be more
1:07:23concentrated because there's more
1:07:26benefits
1:07:27from proximity in industry than there is
1:07:30in in agriculture okay
1:07:34now if you move a little bit forward and
1:07:36say what has happened in
1:07:38say the last decades in the u.s and in
1:07:41some other
1:07:42developed countries what we see is that
1:07:46although industry yes urbanized
1:07:50in the early time period if you look
1:07:53more recently
1:07:54we see growing spatial dispersion of
1:07:58manufacturing
1:07:59manufacturing is still relatively
1:08:01concentrated but it's becoming less
1:08:02concentrated across space across
1:08:04over time whereas with services we're
1:08:07seeing the opposite
1:08:08right services typically used to be sort
1:08:11of the industry as you know which was
1:08:13non-tradable and you know wherever
1:08:14people live
1:08:16where wherever people live you have
1:08:17services so services extremely spread
1:08:19across space
1:08:20but we see growing spatial concentration
1:08:24of services and again that that may be
1:08:27related to land intensity
1:08:30which i said was lower in manufacturing
1:08:32than in agriculture but it's of course
1:08:34even lower
1:08:35in services than it is in industry and
1:08:38so
1:08:38services are displacing manufacturing
1:08:41in cities and then manufacturing is
1:08:43becoming geographically more
1:08:46dispersed it may also have something to
1:08:48do with the life cycle of industries
1:08:51right as industry matures
1:08:54uh there is less need to be
1:08:58close to each other because closeness is
1:09:00particularly important in the
1:09:02when you're learning right when when
1:09:05when technology
1:09:06is new and there's a lot of innovation
1:09:08there's a lot of benefits
1:09:09from proximity that maybe that's more
1:09:12true now in some service industries than
1:09:14it is in
1:09:15manufacturing industries and so these
1:09:17manufacturing industries
1:09:18no longer need to be spatially
1:09:20concentrated
1:09:21and are dispersing versus services that
1:09:24are becoming more
1:09:26concentrated and that's a similar
1:09:28pattern you know we looked at this in
1:09:30the u.s and we've looked at this in
1:09:32europe
1:09:32other people have looked at this in
1:09:34other countries for example in korea
1:09:36that's something that's very well
1:09:37documented so when industrialization
1:09:39happened in korea
1:09:41seoul was the main manufacturing center
1:09:44uh in the last decades that's no longer
1:09:46true
1:09:47what we've seen is industry moving out
1:09:50of seoul
1:09:50into into uh other areas
1:09:54um okay now let me uh
1:09:57switch quickly to something else here i
1:10:00just want to uh
1:10:02very briefly discuss a paper by by doug
1:10:05and uh remy genwa bandits fall route
1:10:09who looked at this at this very
1:10:12interesting observation
1:10:13right where you know as i said in the
1:10:15previous slides
1:10:17the the standard the standard view has
1:10:20been
1:10:21development leads to industrialization
1:10:23industrialization
1:10:25implies urbanization what they
1:10:27documented in a paper is that
1:10:30there are many places in africa
1:10:34and in the middle east where we see huge
1:10:37increases in urbanization
1:10:39without seeing big increases in
1:10:41industrialization
1:10:42and and and the way they explain this is
1:10:45that what really matters for
1:10:48urbanization is growing income right
1:10:51and growing income does not necessarily
1:10:54have to come from industrialization
1:10:56growing income could also come for
1:10:57instance from having a
1:10:59higher income from natural resources and
1:11:01so the sort of idea
1:11:03is that okay we have natural resources
1:11:06say
1:11:06gas or oil our income goes up
1:11:11we're not industrializing but as our
1:11:13income improves
1:11:14and grows we want to spend our income on
1:11:17stuff
1:11:18right and we want to live in a city uh
1:11:21to enjoy the city
1:11:22to be able to spend money in the city
1:11:24and so this is kind of the idea that
1:11:27when we think about cities and
1:11:29development
1:11:30we shouldn't think only about cities and
1:11:33development
1:11:35cities being productive centers
1:11:38cities are also consumption centers
1:11:40right and
1:11:41and that's true in some of these places
1:11:43in africa in the middle east
1:11:44it's it's actually also true in some of
1:11:46uh
1:11:47some of the developed countries right i
1:11:49mean there's there are many
1:11:51people some of the rich people in the
1:11:53world
1:11:54who want to move to say a city like
1:11:56london
1:11:57not to be productive they just want to
1:11:59live in london to enjoy london
1:12:01right of course london is a very
1:12:02productive city right because
1:12:04lots of people are working there but
1:12:06there's a growing uh
1:12:08share of the population or just there to
1:12:11enjoy the place
1:12:12right and so if you have a lot of money
1:12:14you also want to spend it where can you
1:12:15spend it well you can spend it in
1:12:17in some of these uh big big cities
1:12:21okay i'm i'm i'm slowly going to um
1:12:25be wrapping up here uh but i wanted to
1:12:28mention
1:12:29just a few more things uh another really
1:12:32interesting paper
1:12:33uh recent paper on development
1:12:37and and space is this paper by henderson
1:12:41and others
1:12:42in the qje so what they look is what are
1:12:45the
1:12:45what they look at is what are the
1:12:47determinants of the global distribution
1:12:49of economic activity
1:12:51right and so they look at the entire
1:12:53planet and they look
1:12:54they use satellite lights right to
1:12:56measure the level of development
1:12:58or the level of the distribution of
1:13:00economic activity across space
1:13:02and they say let's kind of look at two
1:13:05groups of determinants
1:13:07let's think about the permanence that
1:13:09are important for agriculture
1:13:11so maybe the soil quality or the
1:13:14temperature or whatever or the
1:13:15precipitation so you look at certain
1:13:18factors or determinants that are
1:13:19important for agriculture productivity
1:13:22and then they look at certain
1:13:23determinants that are important for
1:13:25trade
1:13:26meaning access how well am i connected
1:13:30to the rest of the world maybe am i a
1:13:32port am i on a big river
1:13:34that that facilitates trade and they
1:13:36come up
1:13:37with a very then they run a bunch of
1:13:39regressions but they come up with a very
1:13:40puzzling finding
1:13:42which is that it turns out
1:13:45that agriculture so this group of
1:13:47variables which we could call
1:13:49agricultural determinants explains much
1:13:51more of the variation
1:13:54in the spatial distribution of economic
1:13:57activity
1:13:58in developed countries than in
1:13:59developing countries
1:14:01it's puzzling in the sense that you
1:14:03would say well in developed countries
1:14:05nobody is doing agriculture anymore
1:14:07how come agriculture seems to determine
1:14:09where people live
1:14:10and i go to developing countries but
1:14:12agriculture is still very important
1:14:14and there it matters a lot less than in
1:14:16developed countries
1:14:17and so the explanation they come up with
1:14:20is that
1:14:21in the early developers let's say
1:14:24today's advanced economies
1:14:26structural transformation out of
1:14:29agriculture into industry
1:14:31happened at a time when transportation
1:14:34costs were still very high
1:14:36and and that means that at that time
1:14:39cities
1:14:40tended to be located in agriculture
1:14:43regions right
1:14:44so just sort of go back 200 or 300 years
1:14:47right and think about it in the
1:14:49following way
1:14:51where did cities grow cities grew in
1:14:53places where it was agricultural surplus
1:14:56right so that there was enough to eat
1:14:58for everyone
1:14:59and some people could start doing other
1:15:01stuff so they moved into cities
1:15:04to dedicate themselves to other
1:15:06productive activities
1:15:07but of course once they were in these
1:15:09cities they still needed to eat
1:15:10right and so where did they get the food
1:15:12from they sourced that food
1:15:14locally right and
1:15:18and they had to source it locally
1:15:20because transportation costs were still
1:15:22very high
1:15:23and so the point is that in advanced
1:15:26economies
1:15:27cities tended to emerge in places that
1:15:30had a productive agricultural hinterland
1:15:33versus in less developed countries or
1:15:36later developers if you wish
1:15:38where trade was already relatively cheap
1:15:42that wasn't really needed and so cities
1:15:44could emerge
1:15:45in places that didn't necessarily have
1:15:47have have an agricultural hinterland
1:15:50let me wrap up with one last thought
1:15:53here
1:15:54uh related to a space and
1:15:57sector of specialization etc um
1:16:00i will just take one minute i realize
1:16:03i'm going a little bit over time here
1:16:05but i've been working in the last years
1:16:07quite a bit
1:16:09on models uh spatial models
1:16:13of uh climate change right and so
1:16:17why is that interesting in the context
1:16:19of
1:16:20this course i think well of course
1:16:22because
1:16:24on the one hand you need to think about
1:16:26space
1:16:27because different places are going to be
1:16:29affected differently
1:16:31by global warming but you also need to
1:16:34think about sectors
1:16:35because obviously agriculture depending
1:16:38on where you are
1:16:39may be much more affected by global
1:16:41warming than other sectors
1:16:43right and so we will see big
1:16:46shifts in where different things
1:16:50are going to be produced
1:16:53because of because of climate shocks
1:16:56right and so we can look at
1:16:57you know different maps and we can do
1:16:59different things
1:17:00with these types of models on you know
1:17:03here this last thing here is
1:17:05looking at uh in the earth 2200
1:17:10where when agric cultural output
1:17:12happened this is part b
1:17:15in in a world without climate change
1:17:17versus in a world with climate change
1:17:19which would be c
1:17:21so so that's very interesting i think as
1:17:23well sort of thinking about
1:17:24not just urban and space but think
1:17:28urban space plus sectors right
1:17:30agriculture versus other sectors
1:17:32how that might be impacted by climate
1:17:35change
1:17:36i'm going to stop here i went a little
1:17:38bit over time and i'm happy to take
1:17:41any questions you might have
1:17:45fantastic thanks a ton klaus um
1:17:48let me open things up to the front row
1:17:50first
1:17:54hi close thanks for the presentation
1:17:57really interesting
1:17:58um going back to the that first model
1:18:01where you have the ait
1:18:02term for cities i guess how would you
1:18:06think about making it endogenous because
1:18:08i could imagine so
1:18:09actually one thing we we did do and i
1:18:12didn't mention
1:18:13uh we did a number of exercises where we
1:18:17endogenized aid in the following sense
1:18:21as you know in in urban economics and in
1:18:23economic geography
1:18:26there is a lot of evidence that
1:18:29ait is made up of two components
1:18:33one is kind of exogenously given
1:18:37maybe relates to one of the questions we
1:18:39had before
1:18:40you know maybe you know you have better
1:18:42amenities so you have
1:18:43you have better endowments or you're in
1:18:46a more favorable location
1:18:48we could think of that as being
1:18:50exogenous and then
1:18:51an endogenous component that is related
1:18:56to size right so bigger cities
1:19:00uh tend to be more efficient because
1:19:03they're bigger
1:19:04right and there's a lot of estimates out
1:19:06there and typically the estimates
1:19:08range somewhere between an elasticity of
1:19:11somewhere between
1:19:13point zero two and point zero six
1:19:16meaning
1:19:17when you double a a city in size
1:19:20the productivity would go up by
1:19:23somewhere between
1:19:24two and six percent so so that's the way
1:19:26in which we endogenize
1:19:28the ait you could of course go to a
1:19:31fully
1:19:32dynamic model in which you also
1:19:35think about innovation and incentives to
1:19:38innovate
1:19:39and how that may be different across the
1:19:42cities because different cities have
1:19:44different markets
1:19:45and if i'm in a bigger market i may have
1:19:47a
1:19:48greater incentive to uh to engage in
1:19:50local innovation
1:19:52uh that is actually this last couple of
1:19:55slides that i showed you
1:19:56these models with climate change uh
1:19:58actually also
1:20:00those are fully dynamic and look at at
1:20:03the
1:20:03incentives to innovate in in the in the
1:20:06paper on urban
1:20:07with the ait there we endogenize
1:20:11simply in terms of bringing in city size
1:20:15as an additional determinant of city
1:20:18productivity
1:20:21thank you yeah
1:20:24that was a question from stephen barlow
1:20:27in the q a
1:20:28really kind of a complex of questions
1:20:31about
1:20:32covin and um
1:20:35the extent to which we may be seeing a
1:20:38change in the pattern of people working
1:20:39in city centers
1:20:41how much does that motivate a rethinking
1:20:44of the standard urban model
1:20:47so that that's a really fascinating
1:20:50question
1:20:51and and now we'll answer it maybe in two
1:20:53different ways so
1:20:55um i was recently reading a paper by
1:20:59nick bloom and some other people at
1:21:01stanford
1:21:02about you know will will work at home
1:21:06will that stay or will it not stay and i
1:21:08think they come up with some kind of
1:21:10number that
1:21:11probably uh 20 percent uh
1:21:14they will 20 of work will be done from
1:21:17home uh
1:21:18now the very important question when we
1:21:20say 20
1:21:21just to say a number right suppose we
1:21:23say 20 will will be done from home
1:21:26it's very different if you say
1:21:29every individual will say work
1:21:33four days in the office and one day at
1:21:36home
1:21:37versus saying 80 of the individuals
1:21:41will work five days a week in the office
1:21:43and twenty percent of the individuals
1:21:45were work
1:21:46all five days from home uh if you take
1:21:49those two
1:21:50uh cases they would have very very
1:21:53different implications
1:21:54for uh for city sizes right and for the
1:21:58urban model so
1:21:59if if i take the second case which i
1:22:02sort of feel is a more reasonable one
1:22:03but i'm not 100 sure
1:22:05if you sort of say we'll move to a model
1:22:07where
1:22:08many people will work one day a week or
1:22:11maybe two days a week from home
1:22:14then you could imagine that cities may
1:22:16actually become a lot larger
1:22:18right because the easiest way of
1:22:20modeling that would simply to say look
1:22:22my commuting costs are going to go down
1:22:24by 20 or my 40 percent
1:22:26and you know if i need to commute every
1:22:28day an hour in
1:22:30an hour out that's too much i mean the
1:22:33average commute in america is about
1:22:3530 minutes in and 30 minutes out it
1:22:38seems like
1:22:39more than 30 minutes people don't have a
1:22:40lot of patience for that
1:22:42but but if you only have to do this
1:22:43three days a week
1:22:45you may be perfectly happy with 45 60
1:22:47minutes
1:22:48it's okay right if you do this three
1:22:49days a week and so cities could actually
1:22:52become
1:22:52a lot larger because the urban costs are
1:22:56going down because you
1:22:57don't need to do it every eve every day
1:23:00versus
1:23:01the other model where you say no no this
1:23:04is going to be 20 of people who will
1:23:06never go into the office
1:23:08well then of course those those 20
1:23:10become completely footloose
1:23:12they could be anywhere right and we've
1:23:14always had some
1:23:15part of the labor force you know people
1:23:17working you know as
1:23:18telephone uh how i answered telephones
1:23:20and customer service
1:23:22they may just be living in a small city
1:23:24somewhere in uh
1:23:26in anywhere really right and they don't
1:23:28need
1:23:29they don't need any of that urban urban
1:23:32they don't need any of that urban
1:23:33environment so
1:23:34so depending on where we're heading the
1:23:38the the the the effects may be quite
1:23:41different
1:23:42another another small
1:23:45thing related to covet is
1:23:48there is some evidence
1:23:52that there is again a growing demand
1:23:56for living in the suburbs
1:23:59because there's growing demand for space
1:24:02right and so that may also of course
1:24:05change the structure of cities
1:24:07right in the following sense in our
1:24:10model
1:24:10i made the assumption that every
1:24:13individual lives on one unit of land
1:24:16that's obviously a simplifying
1:24:17assumption right it's true that
1:24:20as you move away from the city center
1:24:23people tend to live on bigger and bigger
1:24:25plots of land
1:24:26right if you bring that into the model
1:24:29and if you start thinking that
1:24:31the demand for space there's a shock to
1:24:34the demand for space
1:24:36because i'm going to be working more at
1:24:37home for example right
1:24:39i need some more space if i'm going to
1:24:41need a home office
1:24:43that may change that rent gradient and
1:24:46that may change
1:24:49the compactness of cities and there may
1:24:51be some effects
1:24:52there may be some effects there as well
1:24:55another
1:24:56related and i'll stop here because i
1:24:57realize there may be some other
1:24:58questions but
1:24:59another really interesting uh question
1:25:02at least to me is
1:25:04what is going to happen to these
1:25:07agglomeration economies right
1:25:10so so you know if you look at
1:25:14agglomeration economies as i said these
1:25:16are the conglomerations you get from
1:25:17density
1:25:18but but we as economists we think of
1:25:20those economies as being external
1:25:23right they're externalities they're
1:25:25external to the individual
1:25:26they may even be external to the firm so
1:25:29if you allow
1:25:30individuals to choose whether they go to
1:25:33work
1:25:34or not go to work or even if you allow
1:25:36firms to tell people what they should do
1:25:39right those decision makers are not
1:25:42fully internalizing
1:25:44these agglomeration economies so there
1:25:47may be some
1:25:49there may be some bad surprises in terms
1:25:52of productivity if
1:25:54if this is taken to an extreme the
1:25:56example would be
1:25:57suppose we now all change our
1:26:00preferences
1:26:00and we all want to stay at home always
1:26:03right
1:26:05you know it's a very big question what
1:26:07would happen to these agglomeration
1:26:09economies
1:26:10right and i think that's a fascinating
1:26:12question and and
1:26:14and it's not obvious that that the
1:26:16so-called decentralized economy
1:26:19will bring us to a place that's
1:26:23uh that's good we don't really know
1:26:25what's going to happen to that
1:26:29any other questions yeah i wanted to ask
1:26:33you a quick question
1:26:34um actually it's it's sort of true so
1:26:37one was a more general question because
1:26:39you were saying in on the first part as
1:26:41well that
1:26:42yeah you don't even need that much data
1:26:44to make great analysis using that simple
1:26:46model
1:26:47but um this said is there any like very
1:26:50innovative or interesting data source
1:26:53for um
1:26:54yeah in the context of urbanization you
1:26:56would like
1:26:57be able to point out to us um yeah and
1:27:00the second
1:27:01and the second question i have was
1:27:02referring also to this first model
1:27:05yeah where it's not entirely clear to me
1:27:07how um
1:27:08maybe different yeah costs of
1:27:12of actual goods in the city might enter
1:27:15like
1:27:15how differences in prices and not just
1:27:17trends
1:27:19enter into that model and how they
1:27:21matter
1:27:23and thank you very much for this awesome
1:27:24presentation yes thank you
1:27:27so your first question is a fantastic
1:27:30question
1:27:31and i don't have a very good answer but
1:27:34what i
1:27:35do what i what i what i can say
1:27:38is that the world of
1:27:41spatial data is really
1:27:44exploding right and so there's
1:27:48more and more information available
1:27:52at the local level and and people are
1:27:55using uh
1:27:57very creative ways of
1:28:00getting at these data right so so of
1:28:02course the one example is using
1:28:04satellite data to
1:28:06to measure uh local economic activity
1:28:10but there's also people who have looked
1:28:12at
1:28:14at more who've looked at for example
1:28:19data on parking lots of
1:28:23shopping malls to try to understand the
1:28:26retail activity at the local level you
1:28:28know how full are those parking lots and
1:28:30how
1:28:31what does that tell me about uh about uh
1:28:35the economic cycle so so so there's more
1:28:37and more data
1:28:39available at uh at the local level
1:28:42uh and and it's really an exploding
1:28:45field so i think
1:28:46uh doing stuff uh doing spatial
1:28:49economics in general is is
1:28:52is interesting because there's lots of
1:28:53of course interesting questions but
1:28:55it's also promising this promising field
1:28:58because
1:28:59the data are just uh getting better and
1:29:02better
1:29:03uh over time right and so i think it
1:29:06it still requires some doses of
1:29:10creativity but when you kind of look
1:29:13around
1:29:14and you look at how people are creative
1:29:16it's really quite
1:29:18quite amazing there was recently
1:29:21a paper i don't remember where it was
1:29:23published
1:29:24by some economists uh
1:29:28looking at how we can use uh historical
1:29:31spatial data to
1:29:33think about some of the big questions of
1:29:35economic history
1:29:37right and so people are using all kinds
1:29:39of map
1:29:40technologies and all kinds of ways of
1:29:43trying to get the information at the
1:29:46local level
1:29:47you know even going back to say 19 18th
1:29:49century
1:29:51so it's a it's a very it's it's a very
1:29:53uh
1:29:55it's a very exciting uh it's a very
1:29:57exciting field
1:29:59to to be in uh your second question on
1:30:03on prices in general i think there's two
1:30:07things here one is
1:30:08uh in general uh we all know that
1:30:13in bigger cities there is more
1:30:15competition for land
1:30:17and so land prices go up and that
1:30:19translates into
1:30:20higher housing prices but it just
1:30:24typically also translated to higher
1:30:26prices in general
1:30:28and so a lot of urban economists
1:30:31kind of simplify things and just focus
1:30:34on
1:30:35on house prices as a kind of proxy a
1:30:39kind of all-encompassing proxy
1:30:41of measuring uh the cost of density that
1:30:44that goes to through the lamp price uh
1:30:48but but but you're right with that not
1:30:50everything is said
1:30:51right there's also evidence for instance
1:30:53that bigger cities
1:30:56uh uh have more competition
1:30:59right so because they have more
1:31:00competition the prices of certain things
1:31:03may actually also be lower
1:31:04right so land prices are higher but it
1:31:07that doesn't mean that the price of
1:31:08every single
1:31:09uh product you buy is necessarily higher
1:31:12so
1:31:13so there's also fascinating literature i
1:31:15think on
1:31:16on size of market size of local market
1:31:19degree of competition how that feeds
1:31:22into prices how that feeds into uh
1:31:24innovation etc etc
1:31:31any uh any other questions
1:31:34i have a quick one for you i think it's
1:31:36a quick one from
1:31:37somash mature in the in the q a do we
1:31:40have some better indices
1:31:42than um then
1:31:46indices of spatial dispersion so um
1:31:48obviously hirschmann her selector
1:31:50um but what kind of indices do we have
1:31:54of dispersion among manufacturing
1:31:56services that
1:31:57are useful looking at locations
1:32:01so i would say two things so if you just
1:32:03want to have
1:32:04one statistic
1:32:08yes so you know you'll use the the
1:32:11herschleifer index you can look at uh
1:32:14and of course the variance or the
1:32:16coefficient of variation
1:32:18or the difference between the 80th and
1:32:20the 20th percentile so
1:32:22all the sort of standard all all the
1:32:25sort of standard measures
1:32:27but i'll but but what is true as well is
1:32:29that
1:32:30sometimes it is actually useful to look
1:32:33at the entire distribution
1:32:35right and so you may sometimes be
1:32:37surprised you may think well
1:32:38everything is said by increasing or
1:32:40decreasing dispersion
1:32:42but when you look at the entire
1:32:44distribution uh
1:32:46you may see some things you you didn't
1:32:50expect
1:32:50right so it may very well be that that
1:32:53for example again that
1:32:55that maybe these medium-sized places are
1:32:58uh
1:32:59are becoming uh are doing particularly
1:33:02well
1:33:02and that the the both the biggest and
1:33:05the smallest and
1:33:06the medium small are not doing that well
1:33:08so so i would always
1:33:10recommend to the extent possible rather
1:33:13than immediately going
1:33:15to a summary statistic like the ones we
1:33:19mentioned
1:33:20that i would i would plot the entire
1:33:22distribution
1:33:23just to see hey is there anything else
1:33:25that's interesting so you could for
1:33:26example have a situation where the tail
1:33:29is becoming a lot fatter or where the
1:33:31tail is not become
1:33:32where the tail is becoming thinner maybe
1:33:34maybe the action is over there
1:33:36right and then maybe looking at
1:33:38dispersion or concentration is not
1:33:40really the right way of
1:33:42looking at the data right and so that's
1:33:44that would be my
1:33:45that would be my recommendation
1:33:49i think i should probably in the
1:33:51interest of time there's
1:33:52there's still always last questions
1:33:55coming in but i think in the interest of
1:33:57time we should probably close things out
1:33:59and so i'd like to thank you klaus and
1:34:02i'd also like to turn things over
1:34:04to my my partner in crime joe kaboski
1:34:07to close out today's lecture and and
1:34:09really i guess to close out the whole
1:34:11set of lectures so my thanks klaus and
1:34:14maybe others can join me first
1:34:16before i turn it over to joe maybe those
1:34:19in the front row can do what we've been
1:34:20doing and unmute
1:34:21and let's give klaus uh a round of um
1:34:25real as close to real applause as we can
1:34:32yeah thank you all it was it was it was
1:34:35really fun and i
1:34:36i enjoyed it and very good questions too
1:34:38i i always worry in these types of
1:34:41online settings where there's lots of
1:34:43people uh
1:34:45how do you get to a good level of
1:34:48engagement and interaction and i'm very
1:34:50happy that this part this part just
1:34:53worked really well so thank you
1:34:55thanks and and joe let me turn things
1:34:57over to you
1:34:58yeah that's a great segue into uh what i
1:35:01wanted to say
1:35:01i mean the first is that this is the
1:35:03last uh lecture of the course
1:35:05um uh klaus was the headliner cleanup
1:35:09batter and um so i wanted to thank the
1:35:12cepr staff
1:35:14especially for you know everything that
1:35:16they've done
1:35:17making this possible uh that's um
1:35:21mandy chan kirsty mcneil mcneil
1:35:25lauren waring and nadine clark um
1:35:28and people you know all through uh cepr
1:35:33um this course has really been a
1:35:36grand experiment uh not only for cepr
1:35:40but i think for the profession
1:35:42i know that um next week i'm meeting
1:35:45with people for bread
1:35:46to try to figure out a similar type of a
1:35:50course
1:35:50so that we can reach more people um
1:35:53i it's in in our minds in doug's and
1:35:57in my mind it's been a tremendous
1:35:58success i've learned a ton
1:36:00uh just from sitting in on these
1:36:01lectures um
1:36:04but we want to learn more and we'd like
1:36:06to get feedback from the students
1:36:08so we will be sending out uh and and
1:36:10maybe as well from the
1:36:12lecturers uh we will be sending out
1:36:15an evaluation uh i think
1:36:18both uh maybe with an anonymous part and
1:36:21a uh identified part because
1:36:25we'd also like to sell the idea and so
1:36:27if we can get some nice quotes
1:36:30with names attached that would be
1:36:32wonderful
1:36:34but we also want to learn how to improve
1:36:35the course
1:36:37are there ways to like klaus said are
1:36:39there ways to improve
1:36:40engagement um we we need to find out uh
1:36:44you know
1:36:45where things went right and where they
1:36:46went wrong so i just want to thank you
1:36:48all
1:36:49of course it's you know always the
1:36:51people that hang out until the very end
1:36:52are
1:36:53only a fraction of the overall people um
1:36:56but i want to thank you guys especially
1:36:59and
1:36:59uh please give us feedback so we can
1:37:01learn from this
1:37:03thanks klaus as well bye-bye
1:37:12thanks everyone we look forward to
1:37:14hearing from you all
1:37:15and to staying in touch so many many
1:37:18thanks
1:37:20all the students all the people who've
1:37:21been attending it's
1:37:23it's uh it's been great to get to know
1:37:25many of you over the past weeks too
1:37:28so look forward to staying in touch
1:37:33bye everyone
1:38:34you

About this transcript

This page contains the full transcript of Supplemental lecture: Urbanization and development by CEPR & VideoVox Economics, generated from the public captions YouTube serves with the video. The transcript has 13,914 words across 2,542 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.