The Long Term Growth Model: Fundamentals, Extensions, and Applications

videoType

dynamic-media

videoDmUrl

keyFrameImage

timestamp

00:04 OK,

00:04 I'm gonna get going,

00:05 um.

00:07 With uh,

00:07 with our session today.

00:10 Um,

00:11 my name is Dion Filmer.

00:12 I'm the director of the World Bank's Development Research Group,

00:14 and it's my great pleasure here to welcome you all to

00:18 the March edition of our policy research talk series.

00:21 Uh,

00:22 as many of you know,

00:23 these talks give us an opportunity to present

00:25 work coming out of the World Bank's research department

00:28 with the goal of sharing the findings with colleagues inside and outside

00:31 of the department as well as outside of the World Bank.

00:34 Um,

00:35 I'd like to welcome our online audience,

00:37 um,

00:38 both on Webex as well as YouTube,

00:40 as well as our in,

00:41 in-room audience today.

00:43 Thank you for coming.

00:45 Um,

00:45 today I'm really glad to introduce my colleagues,

00:48 um,

00:49 Steven Pennings and Norman Loweza,

00:51 who will be talking about the project they

00:53 co-lead on the World Bank's long-term growth model.

00:57 And this is on the occasion of the launch of the book on the project,

01:00 actually,

01:01 here,

01:01 let me just hold up the book so you can see it.

01:04 Available at fine bookstores everywhere,

01:06 including the World Bank's online bookstore.

01:08 Um,

01:10 Um,

01:11 Stephen is a senior economist in the macro growth team of the research group here.

01:15 His research interests include fiscal policy,

01:17 especially fiscal transfers,

01:19 economic growth,

01:20 political economy,

01:21 exchange rate pass-through,

01:22 and monetary policy.

01:25 Norman

01:25 is the director of the Indicators Group within DEC,

01:28 the Development Economics Vice Presidency,

01:31 and he was formerly in the research group,

01:32 so we claim him as an honorary member.

01:35 He was a lead economist in our development research group and

01:38 managed the Asia hub of the research group based in Malaysia.

01:42 He was the director of the World Development Report 2014 on risk and opportunity,

01:46 managing risk for development.

01:48 His research has dealt with various areas of economic and social development,

01:52 including macroeconomic management,

01:54 economic and social development,

01:56 including macro

01:57 microeconomic flexibility,

01:59 private and public saving,

02:00 financial debt and stability,

02:02 natural disasters and crime and violence,

02:03 a broad range,

02:05 and a fascinating research,

02:07 uh,

02:07 uh,

02:08 history.

02:09 Um,

02:10 unfortunately,

02:11 our previously scheduled,

02:12 uh,

02:13 discussant,

02:13 uh,

02:13 Manuela Francisco had a last-minute conflict,

02:16 and she is unable to join us.

02:19 Thankfully,

02:20 we have Andrew Burns,

02:21 uh,

02:22 who has agreed to step in to be our discussion today.

02:25 Andrew is the World Bank's global lead for macroeconomic modeling

02:29 and works in the bank's macroeconomics and fiscal management global practice.

02:35 Um,

02:36 prior to joining the macroeconomics practice,

02:38 he worked for 10 years as the manager and lead

02:40 author of the World Bank's Global Economic Prospects Report,

02:43 which examines trends

02:45 in the global economy and how they affect developing countries.

02:49 Uh,

02:49 before joining the World Bank,

02:50 Andrew was the OECD at the,

02:52 with the OECD where he was the head of desk for Czech Republic,

02:55 Hungary and Poland and France,

02:58 um,

02:58 and also played a lead role in the OECD's job study overseeing the

03:01 evaluation of labor market policies in each

03:04 of the organization's 30 member countries.

03:08 So today I'll ask Norman and Steven,

03:10 and they're going to split the time between them,

03:12 to talk for approximately 40 minutes,

03:14 after which we'll hear from Andrew for about 10 to 15 minutes.

03:17 Uh,

03:17 we'll conclude the session with Q&A from the audience.

03:21 If you have a question,

03:22 please use the raised hand option in Webex,

03:25 or of course if you're in the room,

03:27 please just raise your hand.

03:29 Um,

03:29 or signal to me online in the chat that you have a question,

03:32 I'll call on you.

03:33 If you're following on YouTube,

03:34 uh,

03:35 please submit your question in the chat,

03:37 and that will get relayed to me.

03:39 As a reminder,

03:40 we are recording the sessions,

03:42 uh,

03:42 and please mute if you're not speaking.

03:45 So with that,

03:45 let me turn it over to Steven and Norman.

03:56 Good morning.

03:58 Good morning,

03:58 everyone.

03:59 Uh,

04:00 Steven and I are going to,

04:01 uh,

04:01 present

04:02 a brief introduction to the book that we have edited on the long-term growth,

04:07 uh,

04:08 model.

04:09 It is,

04:10 um,

04:10 uh,

04:11 a collection of articles that are

04:14 about the fundamentals,

04:15 extensions,

04:16 and applications of the project of the same

04:19 name.

04:21 So football fans.

04:23 I would like to play

04:24 like Messi.

04:27 And uh pop singers would like to sing to like Aretha Franklin.

04:34 For

04:36 economic growth,

04:37 countries would like to grow like Korea.

04:40 And why?

04:42 Because Korea's economic growth in the last six decades has been spectacular.

04:48 They

04:49 were able to achieve an annual growth rate of 7.2%,

04:53 almost 6% per capita.

04:56 With that

04:57 they increased

04:59 the.

05:00 Uh,

05:01 GDP per capita

05:03 by

05:04 a factor of 30

05:05 over the six decades

05:07 and,

05:07 uh,

05:08 having less than 10% of GDP of the US

05:12 by 1960,

05:14 uh,

05:14 by

05:15 2021

05:17 they were over 50% of that level.

05:21 So Korea is clearly the growth star.

05:25 Now,

05:26 why should uh countries

05:29 uh be so

05:31 interested in,

05:31 in economic growth?

05:33 Well,

05:33 because,

05:34 simply put,

05:35 economic growth is the foundation

05:37 of,

05:37 of,

05:38 uh,

05:39 social and economic development.

05:41 It uh generates jobs

05:43 and income,

05:45 especially for a labor force,

05:47 young labor force

05:48 entering the market.

05:50 It,

05:50 it fosters innovation and entrepreneurship.

05:54 Which in turn creates more growth.

05:57 Uh,

05:58 Especially when it's inclusive and sustainable,

06:01 it can promote political and social stability.

06:04 And it provides

06:07 services,

06:07 uh,

06:08 uh,

06:08 resources to the state

06:09 so that the state can in turn,

06:12 uh,

06:12 provide

06:13 pub uh,

06:14 public services and infrastructure.

06:18 Now,

06:19 it is

06:20 Therefore,

06:21 the key

06:22 economic growth,

06:23 the key to eliminating poverty

06:25 and providing for shared pros prosperity.

06:28 Now of course

06:30 this can be achieved especially when growth is inclusive,

06:34 when it's sustainable,

06:36 and when it's enduring.

06:39 Uh,

06:40 it's not a surprise,

06:41 therefore that economic growth is at the top

06:45 of our policy agendas around the world.

06:48 Now,

06:49 dreaming big on growth

06:51 is maybe a good thing.

06:54 Under certain conditions.

06:56 So yes,

06:57 dreaming big on growth is good if you are implementing a sound economic strategy.

07:04 That,

07:04 that one that will allow,

07:06 allow you to achieve a realistic growth rate.

07:10 But it may be problematic

07:12 if it leads you to poor decision making

07:15 and

07:16 social frustration.

07:18 For instance,

07:19 if you,

07:20 if you have expectations for growth

07:23 that are just very high

07:24 but unrealistic,

07:26 governments may,

07:27 may

07:28 go in a path of a very large public expenditures,

07:32 deficits,

07:33 and debt that they will not be able to,

07:36 to pay in the future.

07:38 It may also distort the plans

07:40 that they may have.

07:41 For instance,

07:42 if they think that growth is going to be very high,

07:45 then,

07:45 uh,

07:46 they might

07:47 want to,

07:47 um,

07:49 spend in luxuries rather than necessities or have their priorities all wrong.

07:56 Uh,

07:56 and when the,

07:57 uh,

07:58 these high expectations for growth are not met,

08:01 then social frustration

08:03 can ensue.

08:05 With that,

08:06 disappointment and social unrest are,

08:08 uh,

08:09 are unfortunately likely outcomes.

08:12 The government may

08:14 lose credibility if they just present this very high

08:18 expectations for growth.

08:20 So the origins of the long term growth model are rela are actually related

08:25 to this

08:26 uh as a reality

08:28 check.

08:29 We wanted to see

08:31 if the growth projections that many countries were making.

08:35 Uh,

08:36 actually,

08:37 uh,

08:37 were supported by a standard neoclassical growth model

08:43 here for those who know the literature associated

08:46 to the names of Solo and Swan.

08:50 But

08:52 After becoming a reality check,

08:54 we realized that we could do more

08:56 with this tool,

08:57 and then it became an,

08:59 uh an explanatory tool,

09:01 one that would allow us to produce alternative

09:04 growth scenarios

09:06 and also a way to understand the drivers

09:09 of,

09:09 uh,

09:10 of growth that are behind

09:12 those uh scenarios,

09:13 including

09:14 drivers related to policy.

09:18 So there was a a lot of interest in uh in the LTGM

09:22 and that generated in fact a demand for richer

09:26 models.

09:28 So

09:29 the

09:30 project expanded

09:31 in terms of the number of models that we were able to produce.

09:35 They were supported by academic papers

09:37 and we made a strong effort

09:40 to uh.

09:42 Implement each of the models that we produced with a toolkit.

09:46 So each of the models that we have,

09:49 um,

09:49 are,

09:50 uh,

09:51 carried in a tool kits that can be used

09:55 in a very simple way.

09:56 They are transparent,

09:58 they are,

09:59 as I said,

09:59 easy to use and you can download,

10:01 download all these from,

10:03 uh,

10:03 the project's website

10:05 which is given,

10:06 uh,

10:07 here in the slide.

10:09 So let me,

10:10 uh tell you about the,

10:11 the book structure

10:12 which it will also be the structure of our presentation

10:15 today.

10:17 So,

10:17 uh,

10:18 the,

10:18 the title of the book

10:19 tells you what this is about.

10:22 We have

10:23 a section on the fundamentals of the,

10:26 of the long-term growth

10:28 model.

10:29 Uh,

10:29 the first is uh a chapter on the,

10:32 the standard,

10:32 the basic model.

10:34 Then we have some extensions dealing with

10:37 public capital,

10:38 productivity growth,

10:39 and natural resources,

10:41 and we,

10:42 uh,

10:42 present a number of

10:44 of applications

10:45 of how this has been done

10:47 in,

10:47 uh,

10:47 in different contexts for different countries.

10:50 Uh,

10:50 in fact,

10:51 this is only a selection of the many applications that we have

10:55 conducted over the years.

10:57 Let me,

10:58 uh,

10:58 start actually from the end and tell you very briefly about uh these applications.

11:03 Uh,

11:04 we did one for Malaysia,

11:05 the basic,

11:06 uh.

11:07 Uh,

11:08 objective there is to study the transition

11:10 of Malaysia from a middle income to high income economy,

11:14 and we found that one of the main impediments of that transition

11:19 was,

11:19 uh,

11:20 the low female labor force participation.

11:23 And we had some recommendations regarding that particular

11:27 area

11:28 of improvement.

11:32 The second application is on Korea.

11:34 Well,

11:35 we,

11:35 I told you already that Korea is the,

11:38 the growth star around the world

11:41 and um we wanted to see.

11:43 If,

11:43 uh,

11:44 the model that we use now for projections

11:47 could actually be used also to explain what happened

11:50 in Korea in the past,

11:52 so this is,

11:52 uh,

11:53 rather than

11:54 a futuristic,

11:55 uh,

11:57 application is one

11:59 that

11:59 goes

12:00 into a retrospective.

12:02 To study the engines of growth.

12:05 And what we found

12:06 uh through that study is that

12:08 the

12:09 drivers of growth actually change over time.

12:11 In the 1960s,

12:13 we had labor and human capital being the main drivers.

12:16 In the 1970s,

12:17 it was mostly about capital accumulation,

12:20 and from the 18 1980s onwards,

12:24 it was productivity growth.

12:27 In the case of Bangladesh,

12:28 the,

12:29 the 3rd application here in the list,

12:31 uh,

12:32 we wanted to study if

12:33 the high growth,

12:35 uh,

12:35 rates that the country has been able to achieve would,

12:37 would,

12:38 uh,

12:38 continue,

12:39 uh,

12:39 in the future,

12:41 and,

12:41 uh,

12:41 the conclusion that we arrived was that

12:44 investment was not enough actually and that

12:47 the total factor productivity will have to increase

12:50 if the country

12:51 was to,

12:52 uh,

12:52 continue with high growth rates.

12:55 The Syria example

12:57 is uh is is very different.

12:59 The Syria example actually

13:01 looks at the devastation

13:03 that occurred because of the brutal war,

13:06 uh,

13:07 civil war in the country,

13:08 and

13:09 we did a study that you could regard maybe as science fiction here

13:13 of uh what would happen

13:15 after the war.

13:17 And what we did was to link these scenarios of of growth in Syria

13:23 with political settlements

13:25 that will drive

13:26 reconstruction assistance

13:28 and also voluntary repatriation

13:31 of people into Syria

13:33 and how the different political settlements

13:36 actually led to very different conclusions.

13:40 The last two applications on Egypt and Egypt and Sri Lanka

13:45 were,

13:45 uh,

13:46 are there also mostly for sentimental reasons you could say

13:50 because this is the way that we started

13:52 the project,

13:53 uh,

13:54 studying Egypt and,

13:54 uh,

13:55 Sri Lanka,

13:55 so we wanted to acknowledge those origins,

13:58 uh,

13:58 where we mostly use

14:00 the model

14:01 as a reality check.

14:04 OK,

14:04 so this was,

14:05 uh,

14:06 this is half the book,

14:07 but I,

14:07 I,

14:07 uh,

14:08 I think,

14:08 uh,

14:08 we can say that

14:09 it's,

14:10 uh,

14:10 it's covered and we invite you to,

14:11 to read the applications if you are,

14:13 if you're interested.

14:14 Let me briefly talk about fundamentals now.

14:18 So,

14:18 the,

14:19 uh,

14:20 as I said before,

14:21 we work with a neoclassical growth model.

14:25 It has three building blocks.

14:27 One,

14:27 a production function that

14:29 many of you are familiar with.

14:31 Uh,

14:31 it has

14:32 an

14:33 output being a,

14:35 uh,

14:35 being a,

14:36 uh,

14:36 a,

14:37 the result of combining productivity,

14:40 uh,

14:41 here.

14:42 Uh,

14:42 denominated by the letter A,

14:44 then capital

14:46 and uh labor augmented by human

14:49 capital.

14:51 And there are decreasing returns

14:53 to the factors of production,

14:55 capital

14:56 and labor,

14:56 that which makes it a neoclassical

14:59 growth model.

15:00 There is a the second building block is

15:03 a capital accumulation equation that tells us that capital

15:07 decreases with depreciation and increases with with investment.

15:11 And uh we want,

15:12 we were interested very much in the demographics uh and labor market forces

15:17 that drive growth

15:18 and we realized that of course,

15:20 uh in the production function what matters is the number of workers,

15:23 but uh

15:24 for uh welfare purposes we want to understand

15:27 uh

15:28 GDP as a ratio to

15:31 Uh,

15:31 to population,

15:32 so GDP per capita.

15:34 Uh,

15:35 of course,

15:35 uh,

15:35 the connection is related to two factors.

15:39 One

15:39 is,

15:40 uh,

15:40 the working age population.

15:44 Uh,

15:44 here represented by this ratio of W over NT

15:49 and N is population work,

15:51 W is working age population,

15:54 and the labor force participation that is L divided by W.

15:59 So

16:00 working age population,

16:01 something demographic,

16:03 population growth is a demographic force and labor

16:06 force participation in fact is a choice.

16:10 OK,

16:10 so,

16:10 uh,

16:10 if you do a little math here,

16:12 uh,

16:13 you will get a growth equation.

16:15 It's um.

16:17 An approximation,

16:18 but one that is very sensible

16:20 where you have that uh GDP growth is a function of productivity growth.

16:25 It's a function of things related to um

16:28 to people

16:29 like human capital,

16:31 demographics,

16:32 and labor force participation.

16:34 All of these multiplied by the returns

16:36 to labor,

16:37 uh beta in this case.

16:40 Then you have the capital segment.

16:43 This is uh investment multiplied by the marginal product of capital

16:48 and the depreciation component.

16:50 So this is uh a uh an,

16:52 an important equation that we use

16:55 in our projections.

16:57 You can see how all the factors that I have been discussing affect actually growth.

17:04 And uh

17:05 this is what we have in terms of how we

17:08 go from conditions and assumptions to outcomes in the model.

17:12 So we have some initial conditions and the model worked very well

17:16 because it relied on very simple

17:19 assumptions,

17:20 very simple initial conditions.

17:22 We just needed to know the capital

17:24 to output ratio.

17:26 A few parameters and also the capital labor shares

17:30 and the depreciation rate.

17:32 Once we,

17:33 once we have those conditions taken from reality,

17:36 and we didn't invent those,

17:38 we took them from the data,

17:39 then we applied a number of assumptions

17:42 on the future path of growth drivers.

17:45 I already told you what those growth drivers are

17:48 productivity,

17:48 human capital,

17:49 demographic growth rates,

17:51 and

17:52 importantly,

17:52 the investment rate.

17:54 So this is where

17:57 the art comes in,

17:58 what the right assumptions

18:00 should be,

18:01 and then we create with different assumptions scenarios,

18:05 and those scenarios

18:06 will lead us to the outcomes of interest.

18:10 The,

18:11 the most,

18:11 uh,

18:11 relevant outcome in this presentation is the growth rate,

18:15 but actually we explain in,

18:17 in one of the chapters there

18:19 how you can link

18:21 changes in growth rates to poverty rates as well.

18:24 And you can use um uh

18:26 some assumptions on the distribution of income in society and

18:30 if you do that,

18:32 then you can link changes in growth rates

18:35 with changes in poverty.

18:37 So that's the way it works and

18:39 uh I can just tell you 11 illustration of uh a lesson that we can derive

18:46 from this very simple model.

18:48 So

18:49 this is

18:50 a result

18:51 that investment-led growth by itself is not sustainable in the long run.

18:56 Why can we say that?

18:58 Why,

18:58 by looking at this equation?

19:00 Well,

19:01 we can say it because,

19:02 uh,

19:02 if we increase in fact this investment rate,

19:06 what's going to happen is that the capital output ratio

19:09 is going,

19:09 is going to increase,

19:11 and if you look at that equation where it says the marginal product of capital.

19:15 Uh,

19:16 capital output ratio is in the denominator,

19:18 which means that actually

19:20 there will be a decrease

19:22 in that,

19:22 uh,

19:22 in that factor.

19:24 What that means really,

19:25 uh,

19:26 in,

19:26 uh,

19:26 intuitive terms is that since there are decreasing returns to capital,

19:31 as you increase the capital output ratio,

19:33 then the strength of capital in producing growth goes down.

19:39 And that's something that,

19:41 that,

19:41 uh,

19:41 in the end will uh eliminate

19:44 uh the growth force that investment can have.

19:48 Therefore,

19:50 The policy advice is that if you want to achieve

19:53 growth uh in the long run,

19:55 then you cannot rely just on investment.

19:58 It must be accompanied by

20:00 other sources

20:02 that drive growth like human capital,

20:04 productivity,

20:06 uh,

20:06 improvement in labor force participation,

20:08 etc.

20:10 so.

20:11 Generally speaking,

20:13 what we are advocating here for a broad-based growth strategy.

20:18 OK,

20:18 so I have in a just a few minutes to cover the fundamentals.

20:22 There's a lot more

20:23 in the book and I,

20:24 and I will invite you to,

20:26 to read,

20:26 uh,

20:27 what is behind.

20:28 So now I will discuss,

20:29 uh,

20:30 one of the extensions and then Steven

20:32 will discuss

20:33 two other extensions and also provide us with the conclusions.

20:38 So the is

20:41 the extension that I will discuss has to do with productivity growth.

20:45 Now this is the A

20:47 letter the a factor in uh the equation that I showed you before.

20:52 It happens to be very important,

20:53 but it's also

20:55 something of a black box,

20:57 a residual,

20:58 as,

20:58 uh,

20:59 Solo

21:00 said in,

21:00 in a,

21:00 in a,

21:01 in a in a famous quote,

21:03 uh,

21:03 the degree of our ignorance,

21:05 I think that's what he said more exactly.

21:07 Well,

21:08 but,

21:08 uh,

21:08 in fact,

21:09 that ignorance

21:10 can be,

21:11 uh,

21:12 circumvented

21:13 if we study

21:14 what the literature,

21:16 literature says about

21:17 the drivers of economic growth,

21:20 oh sorry,

21:20 of productivity growth.

21:23 So we wrote this paper with a colleague uh in the Malaysia

21:27 hub,

21:27 uh,

21:28 Young

21:28 Kim.

21:29 Uh,

21:30 we did a very large literature review

21:33 on the determinants of total factor productivity.

21:36 Uh,

21:37 we identified

21:38 these five areas,

21:40 innovation,

21:41 education,

21:42 market efficiency,

21:43 infrastructure,

21:44 and institutions

21:45 as drivers of total factor productivity.

21:48 We,

21:49 uh,

21:49 moreover,

21:49 examine,

21:50 uh,

21:51 what,

21:51 uh,

21:52 proxies could be used for each of these factors,

21:55 and what we did is we,

21:57 uh,

21:57 try to understand the link between total factor productivity growth

22:01 and that index that we had created.

22:04 We so we run this is a simplification of course but we run a a a regression

22:09 of the

22:10 average TFP growth,

22:12 um,

22:13 uh,

22:14 on

22:15 the this index of uh the drivers of of the of productivity

22:20 and uh the past level of TFP

22:24 just allow for some regression to the mean.

22:28 And what we found is that the higher the index,

22:31 the higher uh TFP growth would be with a diminishing effect.

22:37 So,

22:37 uh,

22:38 we then applied

22:39 this to a number of cases

22:41 and let me show you one of them.

22:43 I have spoken a lot about,

22:45 uh,

22:45 about Korea,

22:46 uh,

22:47 and,

22:47 uh,

22:48 what we decided to do is

22:50 see,

22:50 well,

22:51 what,

22:51 what would happen with TFP growth

22:54 if countries were to do exactly what Korea did

22:57 over the last three decades.

23:00 What would happen with with with uh the TFP growth of those countries

23:04 and we took averages for the different regions

23:07 and that's the result of one of the simulations

23:10 that we did

23:12 as you can see

23:13 the countries that were lagging behind the most

23:16 in uh sub-Saharan Africa and in MENA were able to achieve the highest increases

23:22 in uh TFP growth,

23:23 reaching up to 2%.

23:26 It doesn't seem like a lot,

23:27 but over time it builds

23:29 and actually it makes a difference,

23:31 and those countries that were not,

23:33 uh,

23:34 that had already increased,

23:35 uh,

23:35 TFP significantly like those,

23:38 uh,

23:38 neighboring countries in,

23:39 in,

23:40 um,

23:40 in,

23:40 um,

23:41 East Asia where did not benefit all that much

23:44 from,

23:44 uh,

23:45 copying Korea,

23:46 maybe because they had already done it.

23:49 We could do this also at the country level and in this case I show you

23:54 the uh my own country Peru and we did some uh

23:57 simulations for for Peru

23:59 uh in this case following Korea the world leader or Colombia,

24:04 the regional leader in in Latin America

24:07 and uh we were able to describe.

24:10 First,

24:10 what happened to this index of the drivers of TFP that's on the left hand side

24:16 of this graph

24:17 and then the effect on uh TFP growth rates

24:20 for uh for Peru.

24:24 OK,

24:24 so I have,

24:25 I have told you about uh productivity growth and now

24:28 I think it's the time for me to cede the floor

24:31 to Steven.

24:32 Thank you very much.

24:44 OK,

24:44 thank you very much,

24:45 uh,

24:46 Norman,

24:46 and thanks everybody for,

24:47 uh,

24:48 for attending.

24:50 So as Norman mentioned,

24:51 I'm gonna talk today about two extensions uh to the basic model,

24:55 and the first one of that,

24:56 it relates to public capital.

25:00 So the question we try to ask in the public capital extension of

25:03 the long-term growth model is what is

25:04 the quantitative effect of public investment,

25:07 that is

25:08 infrastructure on growth,

25:09 recognizing that

25:11 uh public public investment and infrastructure are not always exactly the same,

25:14 but in many cases they are.

25:16 So this is obviously a very important question

25:19 for development policy uh in general and in

25:22 particular for institutions like the World Bank,

25:25 for lending institutions.

25:27 So there's kind of two,

25:29 opposing views about this,

25:31 uh,

25:32 in the literature,

25:33 uh,

25:33 and in development policy.

25:35 So one is that,

25:36 um,

25:37 public investment and infrastructure investment have a really big effect on

25:41 growth and should be really at the core of a growth,

25:44 uh,

25:44 of a growth plan,

25:45 and,

25:45 um,

25:46 people refer,

25:47 for example,

25:48 to,

25:48 um,

25:49 China's uh very strong growth rates and their high rates of,

25:52 um,

25:53 Infrastructure investment,

25:55 large human needs,

25:56 and of course big infrastructure gaps.

25:58 So there's a World Bank report that

26:02 lists 700 million people without

26:05 safe drinking water and 1.2 billion without reliable electricity.

26:08 So these are obviously great needs.

26:11 And there's also some academic literature that also produces very

26:15 big estimates of the effect of public investment on growth,

26:18 and this,

26:18 this paper is an example of that.

26:21 Sort of the opposing view is that this is gonna have a small,

26:24 uh,

26:25 a small effect,

26:25 and,

26:26 and part of that,

26:27 um,

26:28 is based on a lot of the empirical evidence from literature that finds very,

26:33 very small impacts of higher public investment,

26:36 uh,

26:36 on growth,

26:36 and also questions some of the other large

26:39 impacts that they don't have causal validity,

26:41 um.

26:43 And

26:44 part of this discussion

26:45 is,

26:46 is about why,

26:47 why that would be the case,

26:48 and obviously corruption and politics

26:51 are key.

26:52 Um,

26:53 so Lant Pritchard has a very famous paper

26:56 where he argues that in many countries government investment spending

26:59 has created very little useful capital.

27:03 And uh Nack and Kiefer have a have a paper that argues that

27:07 in countries with poor governance,

27:09 um,

27:09 governments use public investment as a vehicle to increase their rent seeking.

27:13 So,

27:14 so these are sort of 22 opposing views,

27:16 um,

27:17 in the literature.

27:18 Um,

27:18 we're not trying to get into that debate exactly,

27:22 um,

27:22 but we,

27:23 we include,

27:24 um,

27:25 an effect on growth,

27:26 but we also take,

27:27 um,

27:28 some of these,

27:29 uh,

27:30 lack of efficiency seriously.

27:33 So perhaps a bit more generally in this

27:35 discussion about uh public investment and growth,

27:38 there's a lot of confusion in the literature.

27:41 Um,

27:43 beyond the size of,

27:44 of the effect,

27:45 there's confusion about levels versus changes,

27:48 the effects on the poorest countries,

27:50 and of course dynamics.

27:51 So,

27:52 so is it that levels or changes in public investment,

27:56 um,

27:57 Affects growth,

27:58 how do the effects vary across countries,

28:00 and how long do the effects last?

28:02 If we have a permanent increase in investment

28:04 in public infrastructure spending,

28:06 is that gonna boost growth forever?

28:08 Um,

28:09 so,

28:09 the LTGM public capital extension

28:11 is gonna answer all these questions in a very simple spreadsheet format,

28:15 and this is um joint work with Shamila

28:18 uh Tavadas.

28:21 So the LTGM,

28:22 uh,

28:22 public capital extension is,

28:23 is a very simple extension

28:26 of

28:27 the,

28:27 um,

28:27 basic framework that Norman presented.

28:30 So,

28:31 uh,

28:31 one can

28:32 see,

28:33 um,

28:35 If you start with

28:36 the um

28:37 the production function that that Norman showed,

28:41 um,

28:42 there is a total capital stock which is,

28:44 which is represented by K here.

28:47 And so what the public capital extension does,

28:50 uh,

28:50 for the most part is just to split this into two parts.

28:54 So,

28:54 um,

28:55 there's a government capital

28:57 which is mostly infrastructure,

28:59 we also call it public capital,

29:01 um,

29:01 and then there's also a private capital stock.

29:04 Um,

29:04 so that's the most important

29:06 change in the model.

29:08 Of course we also take into account,

29:10 um,

29:11 this potential lower efficiency.

29:13 Um,

29:14 so that's,

29:15 uh,

29:15 Pritchett and many others,

29:17 uh critique,

29:18 um,

29:19 that when you spend an extra dollar on,

29:21 on public investment in many countries,

29:24 you get less than an extra dollar of uh public capital stock.

29:28 And we actually produce a new infrastructure efficiency index,

29:31 um,

29:31 to measure that,

29:33 uh,

29:33 which includes electricity losses,

29:34 unpaved roads,

29:35 uh,

29:36 and water losses as well.

29:37 Um,

29:38 and the calibration of the model is based on estimates from the literature,

29:43 um,

29:43 and it builds in,

29:45 uh,

29:46 A pretty sizeable effect of public capital on growth,

29:50 so

29:50 more specifically this parameter

29:52 means that a 10% increase

29:55 in um the public capital stock is gonna boost GDP by 1 to 1.7%.

30:03 So what is the what is the quantitative effect of a

30:06 permanent increase in public investment on growth in the model?

30:09 So this is gonna vary across countries,

30:11 so we're gonna look at this

30:12 for representative countries of each um income group.

30:17 So I think the results can be summarized as saying it's,

30:19 it's helpful for growth,

30:21 um,

30:22 but it doesn't create a growth miracle.

30:25 So what we have here is we have kind of a um

30:28 an like an impulse response function where there's a

30:31 a

30:31 a

30:32 a one-time increase in public investment as a share of GDP

30:36 um by 1 by 1% point and uh that that is a permanent change,

30:41 and then this is the change in in growth.

30:45 So one can see that there's a a modest uh and temporary boost to growth.

30:50 So,

30:50 uh,

30:51 these different lines reflect,

30:53 uh,

30:53 typical countries

30:55 calibrated to,

30:56 um,

30:57 high income in blue,

30:58 upper middle income in yellow,

31:00 uh,

31:00 lower middle income in black,

31:01 and,

31:02 uh,

31:02 low income in red

31:03 countries,

31:04 um.

31:05 So,

31:06 uh,

31:06 one can see in the short-term that a 1% point of GDP increase in public,

31:11 uh,

31:11 investment,

31:12 uh,

31:13 across these,

31:13 these,

31:14 these different representative countries,

31:16 um,

31:16 boosts growth by about 0.15 to 0.2% points in the short run,

31:21 and then in the first few years.

31:23 Um,

31:23 that's excluding,

31:24 uh,

31:25 demand side multiplier effects,

31:26 supply side effects only.

31:29 Um,

31:29 however,

31:30 so,

31:30 so,

31:31 so I'd say that's a modest,

31:32 uh,

31:32 that's a modest boost in growth.

31:34 Um,

31:34 but it's also a temporary one,

31:36 so one can see that this,

31:37 this line in terms of the increasing growth relative to the baseline,

31:41 slopes down over time,

31:43 and so after 30 years,

31:44 the boost to growth is only 0.05 to 1.1% point.

31:49 Um,

31:50 and the reason for this,

31:51 um,

31:52 is that the

31:53 the size of the effect depends on the public capital to output ratio,

31:57 and,

31:58 uh,

31:58 there are diminishing marginal product of capital,

32:01 there's diminishing marginal product of capital over time.

32:03 When you increase public investment,

32:05 um,

32:06 that's going to increase the public capital stock relative

32:10 to,

32:10 uh,

32:11 faster than increasing GDP and so this,

32:14 uh,

32:14 ratio is going to increase,

32:16 um,

32:17 over time and reduce the effect on growth.

32:20 Um,

32:20 the other interesting result here is that,

32:23 uh,

32:23 for our sort of representative countries,

32:25 there's actually a pretty similar effect on,

32:27 on growth.

32:29 Um,

32:30 for,

32:30 uh,

32:30 you know,

32:31 for high income countries and low income countries,

32:33 it's actually not all that different.

32:34 Um,

32:35 and the reason for that is that this,

32:37 uh,

32:37 that on average within those income groups,

32:40 the,

32:40 um,

32:40 capital,

32:41 the public capital to output ratio is not actually all that different.

32:45 Um,

32:45 so that's a somewhat,

32:46 uh,

32:46 surprising.

32:47 Um,

32:48 surprising result and obviously very policy relevant.

32:52 So we produce many results,

32:54 uh,

32:56 the,

32:56 the model produces many results and I encourage you to read the chapter for a full,

33:00 uh,

33:01 for a full explanation,

33:02 but to summarize some of the main ones,

33:04 um.

33:06 So the effect of public uh investment

33:08 and private investment is gonna be heterogeneous

33:12 across countries and also over time.

33:14 So

33:14 even though those averages for typical high income,

33:17 middle income or low income countries are actually not all that different,

33:20 within each group there's a lot of heterogeneity.

33:22 Um,

33:23 and it's because that's gonna,

33:24 because that's because it depends on the

33:26 um shortages of public capital

33:29 relative to GDP and private capital relative to GDP.

33:32 And so those will depend on on country uh circumstances,

33:35 and that's why we need to have a tool like the public capital extension,

33:39 uh to calculate the effects on growth.

33:42 The second interesting finding is that in low income countries,

33:46 actually extra private investment has a a

33:48 bigger effect on growth than public investment,

33:51 which is something which is maybe a little bit surprising.

33:54 Um,

33:55 and the reason for that is just that the public,

33:58 uh,

33:58 that the that the private capital stock as a share of GDP

34:02 is really low in low income countries.

34:04 Um.

34:05 And to some extent that's intuitive because those

34:08 countries have a lack of access to credit,

34:10 they often have poor business environments,

34:12 um,

34:13 and so that really constrains,

34:14 uh,

34:15 private investment,

34:16 and so the policy,

34:18 um,

34:18 the,

34:19 the policy advice following that is to try

34:21 to increase private investment in those countries.

34:25 The third result is perhaps a little bit more technical,

34:27 but it is also surprising,

34:29 is that the level of efficiency um actually has no

34:32 no effect on the impact of public investment on growth.

34:36 So this is not our result,

34:37 it comes from a paper by,

34:39 uh,

34:39 an IMF working paper by Bergal,

34:41 and

34:42 it's,

34:42 it sort of stems back to the two forces that

34:46 are embedded in the opposing views I showed a couple of slides ago.

34:50 So this is a sort of simplified version of the

34:53 um equation Norman showed,

34:55 adapted to the public capital extension context,

34:58 and uh one can see that the sort of the first view,

35:01 Um,

35:02 if you have,

35:03 um,

35:04 low,

35:05 if you have low efficiency,

35:07 which is

35:07 a low,

35:08 um,

35:09 value of theta,

35:10 um,

35:12 You might have,

35:12 let's say in a particular country you might have

35:15 a billion dollars of public capital on the books,

35:19 um,

35:19 but in reality,

35:20 a lot of that public capital is very poorly constructed,

35:23 and so,

35:24 um,

35:25 the actual effective capital to public capital to output ratio is much lower,

35:30 and that increases the marginal productive capital.

35:32 Effectively it increases the need.

35:34 Um,

35:35 but then an offsetting force

35:37 that moves in the opposite direction

35:39 is that when you spend an extra million dollars on public,

35:42 uh,

35:42 investment in those countries,

35:44 you might only get,

35:45 uh,

35:46 $500,000 of useful capital because it's

35:49 poorly installed,

35:50 it's politically

35:51 allocated,

35:52 and so forth.

35:53 Um,

35:53 and so,

35:54 uh,

35:54 one can see that if the new public,

35:57 the efficiency of the new public investment and the old public capital are the same,

36:01 these effects cancel out.

36:04 But that's not to say that efficiency is not important,

36:07 um,

36:07 so improving the efficiency of public investment can

36:10 actually have a sizeable effect on growth,

36:12 and especially in countries that have high rates of public investment.

36:18 So we applied this,

36:19 um,

36:20 this model,

36:21 uh,

36:21 in,

36:21 in Malaysia,

36:22 which is uh an application in chapter 5.

36:26 so in Mala Malaysia is on the verge of

36:29 moving from middle income to high income status,

36:32 and policymakers were interested in knowing how

36:34 they could sustain growth in Malaysia.

36:36 And one of the,

36:37 uh,

36:37 things that came out of that discussion was

36:39 whether we could use higher public or private,

36:42 uh,

36:42 investment to the sustain sustain the growth rate.

36:44 And so this is uh,

36:45 the baseline we produced,

36:47 uh,

36:47 at the time.

36:49 Of the growth path in Malaysia

36:51 going out to 2050,

36:53 and one can see the declining rate.

36:56 And then we did some stimulations where we had a permanent increase of,

36:59 of investment,

37:00 we had a first a 1% point increase in

37:03 in public investment,

37:05 which led to a short run increase in growth of 0.15% points.

37:09 Uh,

37:09 we also increased private investment by 0.3 by 1% point of

37:14 GDP and that led to a bigger increase in growth.

37:17 So in Malaysia,

37:18 Malaysia was relatively

37:19 short of private investment,

37:22 and then we also tried a large,

37:23 an even larger effect which is 0.6% points.

37:25 So one can see that

37:26 that growth,

37:28 you know,

37:28 is,

37:28 um,

37:29 is going up,

37:30 um,

37:31 but,

37:33 What this,

37:34 these policies cannot do

37:36 is to sustain long-run growth,

37:39 um,

37:39 which is what the country was most interested in,

37:41 this is due to the diminishing marginal product

37:42 of capital that we've seen many times.

37:45 Um,

37:46 and as a result,

37:47 we needed to consider a combined reforms package,

37:50 uh,

37:50 of the one that Norman mentioned earlier,

37:52 taking into account,

37:53 for example,

37:54 higher TFP growth,

37:56 uh,

37:56 improved human capital,

37:57 and also,

37:58 uh,

37:58 labor force variables like female labor force participation,

38:01 and the details of that

38:03 are in the chapter.

38:05 So,

38:06 I've finished talking about public capital,

38:07 I'm gonna talk briefly about the natural resource extension,

38:10 which is the final

38:11 uh extension of the,

38:12 in the book.

38:14 Um,

38:18 OK.

38:22 I'm not

38:22 sure

38:25 What has happened here?

38:31 There was supposed to be a slide.

38:41 OK,

38:42 for some reason,

38:43 this slide is not loading.

38:46 Um,

38:48 But that's OK,

38:49 so I will have to tell you the.

38:53 I'd have to tell you,

38:54 uh,

38:55 the idea here,

38:56 um,

38:57 because this slide is not loading for some reason.

38:59 Um,

38:59 so the natural resource extension,

39:01 uh,

39:02 is

39:03 motivated by the fact that many developing countries,

39:06 um,

39:07 have large natural resource sectors,

39:09 and if that natural resource sector is a very sizeable share of GDP,

39:14 then obviously,

39:15 um,

39:15 factors that affect,

39:17 that,

39:17 that natural resource sector are gonna be really important for long-run growth.

39:21 And also the model.

39:22 Is gonna be able to answer questions specific to natural resources,

39:27 so for example,

39:28 how do discoveries or resource depletion affect long-run growth,

39:32 how do commodity price shocks affect long-run growth,

39:35 and um the role of also the government's fiscal

39:39 rule and how they deal with those fiscal revenues.

39:42 Um,

39:43 and the,

39:44 the extension of the,

39:45 of the model is actually a very simple one.

39:48 So it um

39:49 it basically starts with the production function that that Norman showed

39:53 for the standard LTGM and then just adds an extra term,

39:57 which is um the GDP produced by the resource sector.

40:04 OK,

40:04 so now I think we have,

40:06 we're back again,

40:07 um.

40:09 So,

40:09 um,

40:10 this is the,

40:11 uh,

40:11 output of the resource sector,

40:14 um,

40:14 so I'm gonna work through an example of oil,

40:17 uh,

40:18 but,

40:18 you know,

40:19 in the model we have a spreadsheet with,

40:21 um,

40:22 over 10 commodities calibrated to over 50 countries.

40:26 Um,

40:27 so the output is going to be a function of,

40:30 um,

40:33 Uh,

40:34 it's gonna depend,

40:35 output of oil,

40:35 it's gonna depend on technology,

40:37 the reserves,

40:38 uh,

40:38 of oil,

40:39 and also the capital stock in the oil sector.

40:42 And so this is kind of an intuitive,

40:44 um,

40:45 function as one can see,

40:46 when,

40:47 um,

40:48 when the reserves are,

40:49 are depleted,

40:49 we're gonna need more oil capital or a better

40:53 technology to produce the same amount of oil,

40:55 and of course when reserves are depleted,

40:56 you can't produce any more oil.

41:00 Uh,

41:00 we're gonna keep track of,

41:01 um,

41:02 reserves based on production and,

41:05 uh,

41:05 exogenous discoveries.

41:06 So the reserves today are gonna be equal to the reserves last year,

41:10 less production plus discoveries.

41:12 Um,

41:13 we can allocate capital across sectors to equalize

41:16 the value of the marginal product of capital.

41:18 Governments and

41:20 crucially,

41:20 governments are gonna tax resource revenues

41:23 and can either invest those in the domestic economy,

41:26 Um,

41:26 or they can save them abroad,

41:28 or they can,

41:29 can also consume them,

41:30 we're not going to discuss that today.

41:31 Um,

41:32 and it's important to mention that the

41:34 model captures the long-run supply side effects,

41:37 mostly through investment of the commodity sector.

41:40 Um,

41:40 it doesn't capture short-run demand effects,

41:42 and those are really important in commodity,

41:44 uh,

41:44 exporting countries,

41:45 but they're not a long run,

41:47 uh,

41:48 issue,

41:48 they're more of a short run issue.

41:50 So I'm gonna talk about a a simple example,

41:53 um,

41:54 we had a calibrated model,

41:55 we calibrated our model to Angola

41:57 a few years ago,

41:59 um,

41:59 and we had a scenario where we had a 1010 year boom bust,

42:03 uh,

42:04 cycle in the oil price.

42:05 So

42:05 we started the simulation with oil prices at $50 a barrel,

42:08 they went up to $80 a barrel and they came back down again.

42:11 And,

42:12 um,

42:12 in Angola,

42:13 which is a very big oil big oil producer,

42:16 or should I say,

42:18 Uh,

42:18 oil is a very large share of the economy,

42:20 um,

42:21 the extra revenues generated from a shock like this

42:24 would be,

42:24 uh,

42:25 almost 25% of GDP.

42:27 Um,

42:27 and,

42:28 uh,

42:28 the vast majority of those accrue to the government.

42:32 Um,

42:32 and so one can immediately see

42:35 that

42:36 the,

42:37 um,

42:37 the fiscal response of the government is going to be crucial for long-run growth.

42:42 So,

42:42 uh,

42:42 in uh the red line we have a,

42:44 a model,

42:45 so this is this right-hand graph

42:47 is a total investment,

42:49 deviation from baseline,

42:50 and,

42:51 uh,

42:51 in the red line we have a,

42:53 um,

42:54 a case where the government is gonna invest those revenues,

42:57 uh,

42:57 in the domestic economy,

42:58 uh,

42:58 so we're gonna have a big spike in investment,

43:01 uh,

43:01 versus let's say,

43:02 in,

43:03 in the blue case,

43:04 if they save those those revenues,

43:06 uh,

43:07 Abroad,

43:08 um,

43:09 and this is gonna have a big,

43:10 big differences in the effect on growth,

43:12 um,

43:13 so,

43:14 the oil price shock is gonna have a big effect on gross domestic incomes,

43:18 um,

43:19 so that's basically the purchasing power of,

43:21 um,

43:21 domestic residents in,

43:22 in Angola,

43:24 um,

43:25 mostly due to this,

43:27 um,

43:27 big,

43:28 uh,

43:29 this big increase in oil revenues.

43:31 Um.

43:33 However,

43:34 uh,

43:35 oil prices do not directly affect,

43:37 um,

43:38 real potential GDP,

43:40 um,

43:41 and this is sort of another point of often point of confusion,

43:44 um,

43:45 And that's for the very simple reason that real GDP is a constant price measure.

43:49 Um,

43:51 so when oil prices go up,

43:52 most of that is exported,

43:53 but the real GDP is calculated with constant export prices.

43:58 Um,

43:58 so consequently,

44:00 the effect of um.

44:02 Uh,

44:02 commodity price shocks,

44:04 uh,

44:04 on long run growth

44:06 is gonna work,

44:07 uh,

44:07 indirectly,

44:08 mostly through the effects on investment.

44:11 Um,

44:12 so on the right hand side I have,

44:13 uh,

44:13 GDP,

44:14 uh,

44:14 relative to baseline,

44:16 and this,

44:16 uh,

44:17 red this,

44:18 uh,

44:18 black line is just

44:19 basically the increase in the size of,

44:21 of revenues,

44:22 approximately.

44:24 Um,

44:24 and one can see that if the government invests all these revenues,

44:28 it's gonna lead to a big increase in GDP.

44:31 Um,

44:31 but that increase in GDP lags

44:34 the,

44:35 uh,

44:35 boom-bust cycle somewhat,

44:37 um.

44:39 And uh the,

44:40 when commodity prices go up,

44:42 initially it has very little effect on GDP.

44:44 Um,

44:45 and then,

44:46 uh,

44:46 you know,

44:46 if it saves them abroad,

44:47 it has a much smaller impact.

44:49 Of course this excludes the short-run demand side effects,

44:52 which is gonna be really important in those countries.

44:56 OK,

44:56 so,

44:57 um,

44:57 this is our last slide.

44:59 Um,

44:59 so,

45:00 uh,

45:00 the LTGM book contains,

45:02 um,

45:03 many models and applications,

45:05 um,

45:06 and many,

45:07 um,

45:07 many policy,

45:08 uh,

45:09 policy relevant findings,

45:10 so,

45:11 uh,

45:11 we encourage you to read the book,

45:12 uh,

45:13 to cover all of those,

45:14 but we want to summarize,

45:15 uh,

45:15 four of those,

45:17 um.

45:18 So one of the first one is that

45:21 um although all um fundamental commod uh

45:25 although um

45:27 fundamental economic conditions are going to always uh affect growth,

45:31 um,

45:31 their relative importance varies across countries,

45:34 and,

45:34 and

45:35 as a result,

45:36 the growth performance constraints and opportunities

45:38 also vary with the country context.

45:40 The second one which I've mentioned

45:42 and Norman has mentioned several times,

45:44 is that the investment-led growth strategies are going

45:45 to be unsustainable in the long run.

45:47 Um,

45:49 investment-led growth can boost,

45:51 um,

45:51 GDP growth in the medium run,

45:54 um,

45:54 but even in that case it would need to be accompanied in most cases by,

45:58 uh,

45:58 increases in savings.

45:59 And finally,

46:00 high growth usually involves fast total factor productivity growth.

46:06 So this um this project has been,

46:08 um,

46:09 So Norman and I are coach ETLs of the project,

46:12 but we've got support from um,

46:14 from many people,

46:16 um,

46:17 and financial support as well,

46:18 so thank you to everybody who's put their work into this,

46:21 and uh you can download the book um from the QR code or also from our website.

46:26 Thank you very much.

46:30 Thank you,

46:31 Steven and,

46:31 and,

46:32 and Norman for that very clear presentation.

46:34 Um,

46:35 Uh,

46:36 we're gonna now turn over to,

46:37 to Andrew Burns,

46:38 who is online.

46:39 Um,

46:41 uh,

46:41 as I mentioned in the introduction,

46:42 Andrew has done his own fair share of,

46:44 of,

46:44 of,

46:45 of,

46:45 uh,

46:45 growth modeling.

46:46 Um,

46:47 so we,

46:48 we're very happy to,

46:49 uh,

46:49 welcome him today,

46:50 and again,

46:51 thank you for stepping in at the last minute,

46:52 Andrew,

46:53 we're really glad to have you with us.

46:56 Great.

46:56 Well,

46:57 thanks so much,

46:57 Dean.

46:57 It's uh my pleasure being here and thanks,

46:59 uh,

47:00 to,

47:00 um,

47:01 the team for such a great

47:03 presentation.

47:05 Um

47:06 As indicated,

47:07 I sort of come to,

47:09 uh,

47:10 reviewing this work,

47:11 uh,

47:11 somewhat late and so I apologize if my um

47:15 remarks are not as,

47:16 uh,

47:17 coherent and,

47:18 and well organized as uh was the presentation.

47:24 I think that

47:25 you know the the LTGM is uh exactly very much

47:34 exactly very much what the team I think somebody needs to say,

47:37 could you please mute yourself.

47:40 Great.

47:41 Um,

47:42 so,

47:43 You know,

47:44 the framework is exactly what it is.

47:45 It's a solo,

47:46 uh,

47:46 growth framework.

47:47 Um,

47:48 it's basically a very much a standard way of modeling the supply side in economics.

47:54 It's basically

47:56 the same supply side,

47:57 the same long run determinants of growth that we

47:59 have in our manager and mod models that we use

48:03 within the bank.

48:04 And it's the same,

48:05 um,

48:06 basic,

48:07 uh,

48:08 framework that you see in pretty much every,

48:10 uh,

48:11 economic model of either a long-term or short-term,

48:14 uh,

48:15 thing,

48:15 so

48:15 nothing wrong there.

48:17 It incorporates important extensions,

48:20 um,

48:20 from the standard solo swan model to deal

48:23 with the distinction between private and public,

48:25 uh,

48:27 um.

48:28 Investment,

48:29 uh,

48:30 it has some endogenous growth features for educational attainment

48:34 that wasn't discussed today,

48:36 but it's in the book and that's really,

48:37 uh,

48:39 interesting and useful.

48:39 It has,

48:40 um,

48:41 the resource extraction mod module which

48:45 helps to deal with,

48:46 um,

48:47 something that is sometimes,

48:49 uh,

48:49 in some of our clients a very tricky problem which is how do you deal with.

48:54 The implications of having uh a really important resource

48:59 and how it will affect growths going forward.

49:03 And it has an interesting,

49:05 uh,

49:06 simple sort of poverty extension and,

49:08 and as I said,

49:09 all of those things are,

49:10 or at least the first three of those things,

49:12 private capital accumulation,

49:14 demographic effects,

49:15 population growth,

49:16 and TFB,

49:17 those are very much a part of every model

49:20 and what um is.

49:23 Sort of the value added of this approach is the,

49:26 um,

49:27 the,

49:27 the endogenous growth from labor market policy,

49:31 um,

49:32 introducing the,

49:33 the quality and quantity of public sector investment into the overall equation

49:39 and uh looking at the changes in the real rents from natural resources.

49:45 Those kinds of effects um are not necessarily in a standard macro model,

49:50 although,

49:50 um,

49:51 like with the LTGM model,

49:53 we,

49:54 they are introduced in those instances in countries where

49:58 those are important effects.

50:02 I,

50:02 you know,

50:02 when I was trying to think about

50:05 how to describe the LTGM model,

50:07 I,

50:07 it,

50:08 it occurred to me that in some ways,

50:10 and I don't mean this in a negative,

50:12 uh,

50:12 way,

50:14 it's,

50:14 uh,

50:15 less a model than it is a product.

50:18 And,

50:19 or,

50:20 or a methodology.

50:22 It's a way of looking at the drivers of long-term growth

50:25 and trying to quantify the relative importance

50:29 and hinge discussions that are

50:33 very much uh taking place

50:36 in um the,

50:37 the very long term

50:39 into something that is uh relatively easily um.

50:45 Quantified and um understood.

50:49 So we see growth as a function of

50:52 productivity,

50:53 of capital accumulation,

50:55 of

50:56 um

50:57 labor force growth,

50:58 and then we can see,

50:59 OK,

51:00 what is the impact of doing something to any one of those determinants

51:04 on the long-term

51:06 uh growth rate

51:08 of the economy.

51:09 And I

51:11 Uh,

51:11 you know,

51:11 I think that I don't know if I'm gonna,

51:13 I'm gonna try and share my screen,

51:14 which is always risky effort here

51:17 just to show one of the slides that um the team had up.

51:24 Which is the,

51:25 basically the growth drivers,

51:27 right?

51:28 And so we have this basic equation,

51:30 which is at the heart of everything that's being done.

51:33 So growth is a function of TFP growth,

51:37 of the rate of growth of

51:40 uh human capital,

51:42 the rate of growth of demographics,

51:44 so the participation rates in the working-age population.

51:49 Um,

51:50 and,

51:51 um,

51:52 then it is a function of the investment to GDP ratio

51:58 and depreciation.

52:00 And that leads,

52:02 among other things,

52:04 just having that formulation,

52:05 which is a very standard um Cobb Douglas kind of formulation,

52:09 leads us to the result that

52:12 investment-led growth by itself

52:14 is not sustainable in the long run.

52:17 I actually,

52:18 I don't have a problem with the results,

52:21 but I do have a problem with um

52:24 how

52:25 the

52:27 um

52:28 That result is being.

52:31 Discussed and thought about in,

52:33 uh,

52:34 in the presentation and in some of the work that has been done with the model.

52:38 And I think it's kind of in some sense,

52:40 one of the challenges

52:42 um

52:42 that's uh there in the modeling system.

52:46 The result that in the long run,

52:49 um,

52:51 The capital,

52:52 the,

52:52 the investment rate

52:54 doesn't have an impact on the rate of growth of capital,

52:57 is a very standard or rate of growth of

53:00 potential output.

53:01 It's a very standard result from the Cobb Douglas,

53:05 um,

53:05 framework and it's effectively because

53:09 in the long run,

53:10 the rate of growth of the capital stock

53:12 will be equal to the rate of growth of GDP.

53:16 And that its fundamental determinants are going to be

53:20 the rate of growth of the labor force and the rate of growth of TFP

53:26 um as a function of the

53:29 size of the labor force in the overall

53:33 um.

53:34 Income share.

53:36 So,

53:36 but that's,

53:37 that,

53:37 that result that you cannot permanently change the rate of growth

53:43 of

53:43 the economy by changing the investment to GDP ratio

53:47 is

53:48 uh very much

53:50 part of the framework and a standard result.

53:54 What I object to is the

53:57 um

53:59 The formulation that that is not a sustainable source of growth,

54:05 and I don't object to the statement in itself because it's factually correct,

54:09 but I think in some sense it misses the point.

54:12 Because

54:12 it tends to make us think,

54:14 well,

54:15 if we cannot sustainably increase the rate of growth through

54:19 capital accumulation,

54:20 then capital accumulation

54:22 probably is not a good,

54:24 uh,

54:25 growth path or growth story.

54:27 And I think that's kind of illustrated here nicely in this chart,

54:30 um,

54:31 in a little less nicely in the,

54:32 in the Malaysian one.

54:34 We can see that the,

54:35 the impact on the growth rate is

54:38 relatively in this particular graph,

54:41 uh,

54:41 large initially and then diminishes over time.

54:44 But what that focus on the growth rate alone,

54:48 um,

54:51 obscures is the fact that

54:53 the level of GDP is higher

54:56 throughout this period,

54:57 and in fact,

54:58 the delta on the level of GDP is accumulating over time.

55:01 So when we were looking at,

55:03 so if you,

55:03 if you think about it,

55:04 the delta on GDP is going to be the integral or the,

55:08 the space underneath that change in the growth rate over time.

55:12 And if you look at,

55:14 for example,

55:15 this graph,

55:17 We're looking at a change in the growth rate of just a little bit under 1% point of GDP

55:24 that is sustained although diminishing over time for a period of 20 years.

55:30 That's effectively by my rough and ready calculations,

55:33 means that

55:34 in level terms,

55:36 at the end of the projection here.

55:39 The level of GDP in Malaysia is 20% higher

55:43 in,

55:43 on the purple line

55:45 than it is in uh the black line.

55:48 And

55:49 I think that all of us

55:52 involved in um

55:55 In development would say that some policy reform that

55:58 managed to increase permanently and in a sustainable fashion,

56:03 the level of GDP by 20%

56:05 is actually a really useful and helpful

56:09 policy.

56:10 It may not change the rate of growth.

56:14 In a,

56:14 in a,

56:14 in an important manner

56:17 way out in 2050,

56:18 but in the,

56:18 the,

56:19 the result of the interim effects

56:21 is actually sustained

56:23 long run and very important.

56:25 And so that's just one sort of comment that

56:28 I found um.

56:30 Worthy sharing with you here.

56:33 Um,

56:36 The other issue that

56:38 I have with the approach

56:40 is,

56:41 you know,

56:41 in some sense,

56:42 that,

56:43 that there's a trade-off and a trade-off between

56:46 the sort of simplicity of the model and,

56:48 and therefore it's easy to use and

56:51 then

56:52 the

56:54 Lack of complexity and therefore the missing of important interactions

56:58 that are going to be occurring in the economy.

57:00 And,

57:01 um,

57:02 that second effect,

57:04 I,

57:04 uh,

57:05 found not so much,

57:06 uh,

57:07 well illustrated in the book

57:09 as in uh another instance where the model was used to look at

57:14 the influence of resource,

57:15 um,

57:17 depletion,

57:18 uh,

57:19 in,

57:19 uh,

57:20 in a,

57:20 in an oil exporting economy.

57:23 And

57:24 just uh

57:25 briefly,

57:25 I sort of extracted it uh

57:28 from

57:29 the model.

57:30 I can just try and

57:31 pull it in here.

57:33 This is coming from the book rather than um

57:37 Rather than uh the presentation.

57:40 The sort of resource uh model or the standard model is that uh

57:46 the

57:47 GDP is equal to

57:49 TFP times um sort of quality augmented labor

57:54 uh times the capital stock raised to the share of each

57:58 in overall output and that

58:01 standard models applied to the non-oil sector

58:03 and then the oil GDP is given by a similar

58:09 function.

58:10 Andrew,

58:10 it's,

58:11 yeah.

58:12 What we're seeing on the screen is

58:13 still the application sustaining growth in Malaysia.

58:17 All right,

58:18 let me just do redo the share obviously share the application.

58:22 No,

58:22 no,

58:22 no,

58:22 thank you,

58:23 um,

58:26 how the heck I hate.

58:27 Webex,

58:29 um.

58:31 Where is the unshare thing?

58:37 All right.

58:39 I think this is actually part of the slide that you were unable to prevent,

58:42 uh,

58:43 present.

58:43 So this is,

58:44 uh,

58:45 us

58:46 being doubly cursed here.

58:49 Anybody

58:50 wanna tell me how I can stop sharing?

58:54 Oh,

58:54 it must be up here.

58:55 Uh,

58:55 yep,

58:55 you stopped there.

58:57 I did.

58:58 Oh,

58:58 I guess so now I can try sharing again.

59:12 All right,

59:12 so this is the,

59:13 um,

59:13 you can see this now,

59:14 I hope.

59:15 Yes.

59:17 So this is the slide to which I was referring

59:19 and um basically the,

59:22 the sort of long run

59:24 resource enhanced model

59:26 um has for the non,

59:28 non-oil part of the economy,

59:29 the standard Cobb Douglas uh uh approach,

59:32 so that's why.

59:34 Oh there for non-oil,

59:36 and then the oil economy or the resource

59:39 economy is um a similar uh production function,

59:43 but um where resources are the

59:48 second important uh source of

59:51 um growth

59:53 and what the model says

59:55 pretty clearly,

59:56 if GDP is the sum of those two,

59:59 which I believe is how they're modeling it,

1:00:00 and that's indeed how we model it

1:00:03 ourselves.

1:00:04 Uh,

1:00:04 in,

1:00:05 in our models,

1:00:06 then,

1:00:06 uh,

1:00:07 as the,

1:00:08 uh,

1:00:09 amount of reserves declines,

1:00:11 then the contribution of the oil sector

1:00:15 to overall GDP growth and growth terms of rate of growth

1:00:19 declines.

1:00:21 But the problem with that is,

1:00:24 in my view anyway,

1:00:25 is that

1:00:27 it's really missing all of the compensating

1:00:33 adjustment mechanisms in the economy.

1:00:35 And if we think about

1:00:37 an oil economy,

1:00:39 the,

1:00:39 the classic

1:00:41 Presentation of it is that here's an economy

1:00:43 that suffers from Dutch disease,

1:00:45 and what is Dutch disease?

1:00:46 Dutch disease

1:00:47 is the effect of an overappreciated currency

1:00:50 reducing the competitiveness of the non-oil sector,

1:00:53 and as a result,

1:00:54 the non-oil sector

1:00:56 gets

1:00:57 squished in the overall activity in the economy.

1:01:02 What happens in real economies as the

1:01:05 resource sector disappears or its weight declines,

1:01:08 then the influence of the Dutch disease effect

1:01:11 declines at the same time

1:01:13 and therefore the rate of growth

1:01:16 of the non-oil parts of the economy expand.

1:01:20 And that whole adjustment mechanism is,

1:01:23 is not going to be here,

1:01:25 um,

1:01:25 just because we cannot articulate

1:01:28 in this model,

1:01:30 the influence of,

1:01:31 uh,

1:01:32 that,

1:01:32 um,

1:01:34 Exchange rate effect on the productivity of the non-oil sectors

1:01:39 and

1:01:40 that

1:01:41 effectively results,

1:01:43 I think,

1:01:43 in sometimes

1:01:45 um strong results being uh or strong conclusions being drawn from this modeling

1:01:51 that might not be um observed if we were using um

1:01:56 A model that was able to uh incorporate these effects.

1:02:00 Um,

1:02:02 So that's just kind of a,

1:02:03 uh,

1:02:04 I guess it's a

1:02:05 um.

1:02:07 A picky,

1:02:08 uh,

1:02:08 a picky point on my part in terms of the overall

1:02:12 approach.

1:02:13 I think that,

1:02:13 um.

1:02:16 It,

1:02:16 it's important to recognize that,

1:02:17 that,

1:02:18 that this approach or the,

1:02:19 the um

1:02:23 The,

1:02:24 the,

1:02:24 the LTGM model does have a lot of real benefits and it can help us

1:02:30 um

1:02:31 take a look at some of the,

1:02:32 the,

1:02:33 the,

1:02:33 the,

1:02:33 the impacts and the influence and think through

1:02:37 um

1:02:38 the long-term influence of different policies might sort of

1:02:42 I guess,

1:02:43 takeaway is that when we do it,

1:02:45 that we want to be careful that one,

1:02:48 we're thinking about the levels,

1:02:50 and I think that's one of the challenges of

1:02:53 the,

1:02:53 the long-term growth model.

1:02:55 It is focusing on the growth rates and long-term growth rates,

1:02:59 and as a result,

1:03:00 we might be

1:03:02 um

1:03:03 not giving due attention to what's happening to the level

1:03:06 of activity because it's that level of activity that ultimately

1:03:10 is driving poverty reduction and it's super important.

1:03:14 Um,

1:03:15 and the second thing is that,

1:03:17 um,

1:03:18 because we're missing the price mechanisms,

1:03:20 because we're missing important,

1:03:22 um,

1:03:23 adjustment mechanisms in the economy,

1:03:26 we might also be missing,

1:03:28 um,

1:03:29 important,

1:03:30 maybe their 2nd or 3rd-order effects,

1:03:32 but sometimes those 2nd or 3rd-order effects,

1:03:35 uh,

1:03:35 can be,

1:03:36 um,

1:03:37 dominant.

1:03:38 So,

1:03:38 um,

1:03:39 bottom line,

1:03:40 um,

1:03:40 I think that it's,

1:03:42 it's a great tool,

1:03:43 and I think the,

1:03:44 the proof of the pudding is the fact that it's been used,

1:03:47 uh,

1:03:47 in,

1:03:47 in many instances.

1:03:50 Um,

1:03:52 I think one of the big challenges for the,

1:03:55 for the model,

1:03:56 and that's true for any model

1:03:58 is while it can tell us

1:04:00 um what would happen if we improved our index of TFP from

1:04:06 5 to 0 points to,

1:04:08 to 7,

1:04:10 it can't really tell us how to do that,

1:04:12 and ultimately,

1:04:13 that's the development challenge.

1:04:14 So,

1:04:15 Uh,

1:04:16 reiterating the point that productivity is,

1:04:19 uh,

1:04:20 and productivity growth

1:04:21 is the most important,

1:04:23 um,

1:04:24 uh,

1:04:24 sort of long-term driver

1:04:26 of growth.

1:04:28 Absolutely.

1:04:29 How do you achieve it?

1:04:30 I'm not sure the model is really helping us,

1:04:33 um,

1:04:34 very much in that regard.

1:04:36 Um,

1:04:39 So I think that was it.

1:04:40 The only last um remark that I,

1:04:42 that I wanted to make was that

1:04:44 I thought in regards to this,

1:04:47 how do we achieve it question,

1:04:50 that it would be useful um

1:04:52 for the model to be,

1:04:55 look at some of the sort of uh drivers of productivity that are um

1:05:02 Laid out in the CEM 2.0

1:05:05 tool which sort of

1:05:07 does the same kind of exercise,

1:05:10 but really tries to look at which are the particular

1:05:16 Uh,

1:05:16 drivers of productivity growth,

1:05:18 drivers of participation rate in individual countries that are

1:05:22 weak

1:05:23 versus those that are relatively strong,

1:05:26 and then use that in combination with this tool to motivate the discussion.

1:05:31 But by and large,

1:05:33 I think it's an impressive body of work.

1:05:35 Um,

1:05:36 I was thinking as I was reading it that it may be the long-term growth model

1:05:40 because it looks at growth in the long run,

1:05:43 but it's also the model that's been around,

1:05:45 uh,

1:05:45 quite a while in the bank and has had a long history,

1:05:48 and I think that,

1:05:49 uh,

1:05:49 both Norman and,

1:05:50 uh,

1:05:51 Steven are to be congratulated for their contributions that they've made,

1:05:54 uh,

1:05:55 throughout the bank,

1:05:56 uh,

1:05:57 via this tool.

1:05:58 Thanks.

1:06:00 Thank you,

1:06:00 Andrew,

1:06:01 for,

1:06:01 for,

1:06:01 for those,

1:06:02 uh,

1:06:02 those comments.

1:06:03 Before opening up,

1:06:04 um,

1:06:05 uh,

1:06:05 I'm gonna turn back to,

1:06:06 to,

1:06:06 to Norman and Steven.

1:06:07 Just a reminder,

1:06:08 if you have questions in the room,

1:06:10 raise your hands when we open it up here.

1:06:12 Uh,

1:06:12 I have a couple of questions from YouTube that I've gotten,

1:06:15 um,

1:06:16 if you have a question and you're on the Webex,

1:06:18 please raise your hand or just signal in the chat.

1:06:21 That you'd like to ask a question and I'll call on you.

1:06:24 So,

1:06:24 uh,

1:06:25 over to,

1:06:25 to Norman and,

1:06:26 and Steven to,

1:06:27 to react a little bit to Andrew's

1:06:29 comments.

1:06:30 And then Andrew,

1:06:30 I'll also come back to you sort of towards the end of the Q&A if you have any

1:06:34 reflections,

1:06:35 or,

1:06:35 or feel free to jump in anytime.

1:06:40 Well,

1:06:41 Andrew,

1:06:41 thank you very much for,

1:06:42 uh,

1:06:42 for your comments,

1:06:44 uh,

1:06:45 and,

1:06:45 and for your very kind words,

1:06:47 uh,

1:06:48 towards,

1:06:49 towards us and towards the project.

1:06:51 Uh,

1:06:52 actually,

1:06:53 I have to say also that we have collaborated with you over many years

1:06:58 and received your feedback,

1:07:00 and for that we are also very,

1:07:02 very grateful.

1:07:03 Uh,

1:07:03 we,

1:07:04 um,

1:07:04 initiated this project as a joint

1:07:08 Product with uh with MTI

1:07:10 and uh I believe that we have

1:07:13 benefited a lot from the interaction with operations that is every single

1:07:18 uh

1:07:19 extension,

1:07:20 every single application was driven by

1:07:23 the interest from our MTI uh colleagues and from our colleagues in operations.

1:07:30 Uh,

1:07:30 so that that level of interaction has been uh very beneficial for us,

1:07:35 and I hope that,

1:07:36 uh,

1:07:36 we,

1:07:36 uh Steven and I hope that that

1:07:38 that has been also been beneficial for,

1:07:40 for you guys.

1:07:42 Uh,

1:07:42 you pointed out several,

1:07:44 uh,

1:07:44 limitations,

1:07:45 and,

1:07:45 and we agree with most of them.

1:07:47 Uh,

1:07:47 I,

1:07:48 I want to add that,

1:07:49 uh,

1:07:50 we see this model as complementary to the work

1:07:53 that you,

1:07:54 Andrew,

1:07:54 and all the people in MTI in the Global Prospects,

1:07:57 um,

1:07:58 Uh,

1:07:59 group,

1:08:00 uh,

1:08:00 do in terms of,

1:08:01 uh,

1:08:02 short-term forecasting.

1:08:03 Uh,

1:08:04 our,

1:08:04 our work is not really for,

1:08:06 for

1:08:07 the forecasting of,

1:08:09 of growth in the,

1:08:10 in the next 235 years.

1:08:13 That is very much dependent on international conditions,

1:08:16 for instance,

1:08:17 uh,

1:08:18 dependent on the demand effects

1:08:20 which we don't have in the,

1:08:22 in the in the model because we

1:08:24 Uh,

1:08:24 are actually assessing potential growth rates that,

1:08:28 that depend on this equilibrium between demand and supply,

1:08:31 and therefore

1:08:32 everything is driven by the production,

1:08:35 uh,

1:08:35 function of the,

1:08:36 of the economy.

1:08:37 So this is uh a limitation that we,

1:08:39 uh,

1:08:40 recognize that we don't consider demand effects and that we don't,

1:08:44 uh,

1:08:45 consider short run and cycle

1:08:47 cycles in the economy.

1:08:48 But,

1:08:48 uh,

1:08:49 if you regard this as complementary to the work that is done.

1:08:53 Uh,

1:08:53 in other parts of the bank,

1:08:55 then uh you see that,

1:08:56 uh,

1:08:57 uh,

1:08:57 rather than a limitation,

1:08:58 it's just a distribution of tasks

1:09:01 among different groups in,

1:09:02 in,

1:09:02 in the bank,

1:09:03 which we welcome very much.

1:09:06 This is the first reaction.

1:09:08 The second reaction is that

1:09:10 uh we we agree with you that uh the

1:09:13 that although investment

1:09:16 does not in the long run lead to uh uh to to growth,

1:09:21 uh,

1:09:21 that's what happens in neoclassical growth models where they are

1:09:24 decreasing returns to the factors of uh of production,

1:09:27 uh,

1:09:28 that the level of GDP,

1:09:30 uh,

1:09:30 is in fact,

1:09:31 um,

1:09:32 Uh,

1:09:33 permanently,

1:09:34 uh,

1:09:34 affected,

1:09:35 and that it will benefit from any level of,

1:09:38 of,

1:09:38 uh,

1:09:39 of growth.

1:09:40 That is undeniable.

1:09:41 And,

1:09:41 and that is something that we should,

1:09:43 uh,

1:09:43 uh,

1:09:44 recognize maybe more,

1:09:45 more explicitly in,

1:09:46 in the,

1:09:47 in the presentation that presentations that we make.

1:09:50 Uh,

1:09:51 uh,

1:09:52 the only thing here to point out is that for many

1:09:55 Uh,

1:09:56 for many policymakers,

1:09:58 for many countries,

1:09:59 the challenge is not only to increase the level of income,

1:10:03 but actually to increase growth rates

1:10:05 because that is indeed what's going to have a

1:10:09 a more permanent effect on development.

1:10:12 Um,

1:10:13 it,

1:10:13 it is a,

1:10:14 a level of how

1:10:16 dynamic the economy is,

1:10:18 and it's that,

1:10:19 um,

1:10:20 that dynamism

1:10:21 which actually leads to development and that's

1:10:24 why we emphasize growth rather than,

1:10:26 than levels.

1:10:27 So we recognize the point that you are making,

1:10:30 of course.

1:10:32 On,

1:10:32 uh,

1:10:32 just the last issue that I will,

1:10:33 uh,

1:10:34 address is,

1:10:35 is on productivity,

1:10:36 productivity growth.

1:10:37 We do have 11 of the chapters dealing with

1:10:41 the drivers of productivity,

1:10:43 uh,

1:10:44 and,

1:10:45 uh,

1:10:45 what is missing,

1:10:47 uh,

1:10:47 possibly there

1:10:48 is that,

1:10:49 uh,

1:10:50 some of the productivity components,

1:10:52 uh,

1:10:53 are actually uh can be embedded in,

1:10:56 um,

1:10:57 in the other factors of production.

1:10:59 So,

1:10:59 for instance,

1:11:00 uh,

1:11:00 when we talk about infrastructure as being a driver

1:11:03 of productivity,

1:11:04 uh,

1:11:05 this could come in the form of some capital.

1:11:08 Uh,

1:11:08 when we do,

1:11:09 we say innovation,

1:11:10 this could also come in the form of investment of new types.

1:11:14 Uh,

1:11:15 and,

1:11:15 uh,

1:11:16 we go through the simple way of uh of

1:11:19 adding that to growth through uh

1:11:22 total factor productivity,

1:11:24 but if some of that is embedded in capital,

1:11:28 then uh that will have an effect as well.

1:11:29 That's something that we don't,

1:11:31 don't consider,

1:11:32 but it is an interesting question that that we

1:11:36 we could address in the future.

1:11:37 But I will stop at this and,

1:11:38 and thank you very much,

1:11:39 Andrew again for

1:11:41 not just your comments now but for the collaboration that we have had over the years.

1:11:47 Uh,

1:11:48 yes,

1:11:48 thank you,

1:11:48 um,

1:11:49 very much for the comments,

1:11:50 um,

1:11:51 and to reiterate what Norman said,

1:11:52 we agree with,

1:11:53 uh,

1:11:54 almost

1:11:54 all of them.

1:11:55 Um,

1:11:56 to add a couple of other,

1:11:58 uh,

1:11:58 more detailed some of one of Norman's comments,

1:12:01 uh,

1:12:01 on the level,

1:12:02 um,

1:12:03 of,

1:12:04 uh,

1:12:05 Of,

1:12:05 uh,

1:12:06 the effect of capital on,

1:12:08 on the level of GDP we completely agree with that,

1:12:10 and

1:12:11 I think that

1:12:13 some of the discussion,

1:12:14 I,

1:12:14 I guess depends on

1:12:16 the country context.

1:12:17 So,

1:12:18 as Norman mentioned on the first slide,

1:12:21 you know,

1:12:22 a lot of the countries that sort of come to us

1:12:25 have these very big ambitions,

1:12:28 um,

1:12:29 And so the only way they can achieve those

1:12:31 ambitions is by growing fast for a long time,

1:12:34 and um we,

1:12:35 so we agree that

1:12:37 it,

1:12:37 you know,

1:12:38 the,

1:12:38 these investment investment might increase the,

1:12:40 the level of growth,

1:12:41 but in order to achieve those ambitions,

1:12:43 which is sort of where the conversation starts,

1:12:45 um,

1:12:46 it needs to be,

1:12:47 it needs to be higher growth

1:12:49 for the long-term,

1:12:50 um.

1:12:51 Uh,

1:12:51 so on the,

1:12:52 the second point,

1:12:53 uh,

1:12:53 on,

1:12:54 um,

1:12:54 Dutch disease.

1:12:56 Um,

1:12:57 so yes,

1:12:58 we agree that the model,

1:13:00 um,

1:13:00 misses some of those,

1:13:01 um,

1:13:02 some of those effects,

1:13:03 and,

1:13:03 and Dutch disease is a particularly tricky one,

1:13:07 to include in this model because it doesn't have a demand side,

1:13:10 um.

1:13:12 You could have um

1:13:14 some other things,

1:13:15 you know,

1:13:15 related like maybe a learning by doing or or other things,

1:13:19 but

1:13:19 the difficulty with a lot of those um those

1:13:21 mechanisms is that they're very hard to calibrate.

1:13:24 Um

1:13:26 So,

1:13:26 but,

1:13:26 uh,

1:13:26 the point is,

1:13:27 is well taken,

1:13:28 um,

1:13:29 and I think we,

1:13:29 we need to,

1:13:30 we do need to be careful with how we uh discuss the findings of the model,

1:13:36 uh,

1:13:36 in a policy,

1:13:37 um,

1:13:38 in a policy context.

1:13:39 And also thank you for mentioning the CEM 2.0 tool,

1:13:42 we will,

1:13:42 we will definitely have a look at that.

1:13:46 Thanks,

1:13:46 Norman and Steven.

1:13:47 Um,

1:13:47 Andrew,

1:13:47 before coming back to you,

1:13:48 I'm gonna open it up a little bit to collect some other questions.

1:13:52 Um,

1:13:52 we have two people online,

1:13:53 and,

1:13:54 and forgive me for,

1:13:55 for how I'm gonna pronounce this.

1:13:57 We have,

1:13:57 uh,

1:13:58 Zelo and then we have Angela.

1:14:00 Let's start with Zelo.

1:14:01 Do you wanna unmute yourself and ask your question?

1:14:05 Yes,

1:14:06 indeed.

1:14:06 Can you hear me well?

1:14:07 We can indeed.

1:14:10 Uh,

1:14:10 that's great.

1:14:10 Thank you very much,

1:14:11 uh,

1:14:11 uh,

1:14:12 Norman and Steven,

1:14:13 uh,

1:14:14 and also for,

1:14:14 uh,

1:14:15 uh,

1:14:16 very important comments by Andy,

1:14:18 um,

1:14:18 and a great presentation.

1:14:20 Um,

1:14:21 um,

1:14:22 I,

1:14:22 I work on,

1:14:22 on Egypt,

1:14:23 by the way,

1:14:23 uh,

1:14:24 uh,

1:14:24 Middle East,

1:14:24 and,

1:14:25 and,

1:14:25 and,

1:14:25 uh,

1:14:25 we're in the middle of,

1:14:26 uh,

1:14:27 putting together the,

1:14:28 uh,

1:14:28 uh,

1:14:28 actually

1:14:29 preparing a,

1:14:30 uh,

1:14:30 a country economic memorandum which is

1:14:32 actually focusing on long term growth issues.

1:14:35 So this is very much.

1:14:36 on our minds and we'd very much like to kind of look

1:14:40 at the Egypt application and review it and revisit it in this.

1:14:46 So I'd just like to add a comment

1:14:49 on

1:14:50 that in the case of Egypt I think it's particularly useful

1:14:54 because Egypt is.

1:14:56 One of those cases where

1:14:58 you know indeed we do have we do see these kind of fundamental imbalances

1:15:04 in the drivers of long term growth,

1:15:06 you know,

1:15:07 for example,

1:15:08 the

1:15:09 female labor force participation is

1:15:13 about 15%,

1:15:15 if that,

1:15:17 right,

1:15:17 extremely low,

1:15:18 and it actually has been declining.

1:15:21 Right?

1:15:21 So,

1:15:22 so you have the,

1:15:23 uh,

1:15:23 the LA,

1:15:24 uh,

1:15:24 you know,

1:15:25 very much underutilized,

1:15:26 you know,

1:15:27 for long term growth in Egypt.

1:15:30 Private investment is also a low,

1:15:32 reflecting the kind of difficulties in the

1:15:34 investment climate and the regulatory environment,

1:15:37 you

1:15:38 know,

1:15:38 but also

1:15:40 the kind of

1:15:42 role of the state

1:15:45 and the large footprint of the state owned enterprises that

1:15:50 kind of also take up the space for private sector activity and private investments.

1:15:56 So

1:15:58 you know,

1:15:59 is not really growing

1:16:02 as fast as it should.

1:16:03 You know.

1:16:04 The private investments are not as high as they should be.

1:16:08 Long term productivity growth

1:16:10 also hasn't been high,

1:16:11 so what's left is basically these very large

1:16:15 public investments

1:16:16 and

1:16:18 ongoing,

1:16:18 but also the future public investments plan.

1:16:22 In Egypt that we see,

1:16:24 and I think it's a very legitimate question is this sustainable?

1:16:28 And I think the macro team in MTI thinks it's not,

1:16:32 and these other drivers of growth must

1:16:37 be kind of revisited and reactivated in one way or the other.

1:16:42 How that is to be done,

1:16:43 obviously,

1:16:43 as Andy says,

1:16:44 is easier said than done.

1:16:46 But

1:16:47 so I think that in the case of Egypt,

1:16:49 in the case of these countries where you

1:16:51 do have these very large and obvious imbalances between

1:16:55 drivers of growth,

1:16:56 I think the long term growth model can point out some of the important

1:17:00 policy directions that

1:17:01 that

1:17:03 that can inform the policy dialogue.

1:17:04 So I very much welcome that.

1:17:09 I would quibble to give other people a chance as well,

1:17:12 could you keep your comments.

1:17:14 Just one more minute.

1:17:15 Yeah,

1:17:15 please,

1:17:15 thank you.

1:17:16 So that,

1:17:17 that's,

1:17:17 that's the,

1:17:18 uh,

1:17:18 uh,

1:17:18 so that that was the,

1:17:19 the main comment on the productivity was that,

1:17:22 um,

1:17:22 you know,

1:17:23 I,

1:17:23 I would quibble a little bit with the,

1:17:25 with the notion that the literature is completely

1:17:28 kind of unclear about the aggregate

1:17:31 contribution of public investments to growth.

1:17:34 You know,

1:17:35 my own work with Johannes on South Africa data

1:17:40 shows that if you actually use disaggregated data on infrastructure

1:17:46 down to lower levels of infrastructure sectors,

1:17:50 then you actually get much richer estimates and more robust effects than

1:17:54 if you just rely on the very aggregate figures in cross country

1:17:59 estimates.

1:17:59 Thank you very much.

1:18:00 Over.

1:18:01 Thanks,

1:18:02 and sorry for cutting off there.

1:18:04 Um,

1:18:04 just a sense,

1:18:05 are there questions in the room?

1:18:06 Just could you raise your hand if you have a question?

1:18:07 OK,

1:18:07 we have one question over here,

1:18:09 2 questions.

1:18:10 OK,

1:18:10 I'm gonna complete the questions.

1:18:11 There's another question online by Angela.

1:18:14 Angela,

1:18:14 do you mind coming in and also just keeping it brief so that we can sort of do a full,

1:18:19 uh,

1:18:19 round of the table.

1:18:21 Go

1:18:22 ahead.

1:18:22 Yeah,

1:18:22 sure,

1:18:23 uh,

1:18:23 thanks so much.

1:18:24 I hope you can hear me.

1:18:25 It's,

1:18:26 I'm in Jakarta and it's like 10:19,

1:18:28 so I won't switch on my camera,

1:18:30 sorry.

1:18:30 So I'm Angela.

1:18:31 I work on Indonesia and also currently working on a CEM for Fiji.

1:18:37 I have two quick questions.

1:18:38 One is,

1:18:39 um,

1:18:40 so I noticed that

1:18:41 you,

1:18:42 you released this LTGMTFP

1:18:45 code extension,

1:18:47 but it's a simple Excel sheet that sort of allows the user to put in

1:18:51 a lot of different

1:18:53 Uh proposed determinants of TFP I guess,

1:18:57 education and innovation and lots of sort of uh variables there.

1:19:01 Uh,

1:19:01 but I,

1:19:02 I was wondering if you can come in a little bit on,

1:19:04 on how

1:19:05 one can sort of

1:19:07 feed that into the broader sort of,

1:19:10 uh,

1:19:10 uh baseline LTE.

1:19:12 EGM,

1:19:13 uh,

1:19:14 uh,

1:19:14 uh,

1:19:14 a 2 because

1:19:15 you can do that and then you sort of see where TFP is going,

1:19:18 but it's,

1:19:18 it's,

1:19:19 it's not super clear how then you can

1:19:21 take those numbers and

1:19:23 plug it into the model so they can say,

1:19:24 OK,

1:19:25 if this happens to education,

1:19:26 then this is what's gonna happen to growth.

1:19:28 Um,

1:19:29 when Andrew was speaking,

1:19:30 I thought that would be

1:19:32 Perhaps that was a some sort of a solution to

1:19:35 sort of the user going into sort of deeper detail.

1:19:38 What should you do to increase TFP and then

1:19:41 when TFP increases to that,

1:19:42 what would happen to growth.

1:19:43 So if you can comment on those linkages,

1:19:45 that would be great.

1:19:47 The second question is,

1:19:48 um,

1:19:49 on,

1:19:49 there's lots of

1:19:51 other tools uh by MTI.

1:19:53 That use a CP Douglas as well,

1:19:55 identification that are more backward looking.

1:19:58 So you're looking at

1:19:59 how has TFP been growing in the past and how has,

1:20:03 you know,

1:20:03 uh,

1:20:04 labor and human capital been growing in the past,

1:20:07 um,

1:20:08 and it would be nice to hear if you thought about how,

1:20:11 how to link those uh sort of

1:20:14 two stories so that,

1:20:15 you know,

1:20:16 you can use these tools for the

1:20:18 Backward looking of how the economy has been progressing over time

1:20:22 because those are also sort of long term looking but from,

1:20:25 you know,

1:20:25 sort of backward looking

1:20:26 and then use this to sort of forward looking and if you thought about how to

1:20:30 sort of take the results from,

1:20:32 from those sort of sharply decomposition type of tools

1:20:36 uh and link it to what the LTGM is able to give you.

1:20:40 Uh let me stop here.

1:20:41 Thanks.

1:20:42 Thanks,

1:20:42 Angela.

1:20:43 I'm gonna come to Stti and then art,

1:20:44 but I'm just gonna relate two questions we got from,

1:20:47 from YouTube.

1:20:48 One is asking whether the,

1:20:50 the female labor force participation could be parameterized in the model,

1:20:54 relating to another comment,

1:20:56 and then another question related to,

1:20:57 uh,

1:20:58 can this model be applied to sectoral growth,

1:21:01 for example,

1:21:01 agriculture alone,

1:21:02 and how would one think about that?

1:21:04 Um,

1:21:05 then there is,

1:21:06 um,

1:21:08 OK,

1:21:08 let's go to Stutti and then art.

1:21:10 St.

1:21:14 Thank you,

1:21:15 great presentation.

1:21:16 Uh,

1:21:16 my question is about to the extent to which both in the book and in your,

1:21:21 uh,

1:21:21 ongoing work and policy dialogue with governments,

1:21:25 how do you make the link between

1:21:27 clearly the questions you're getting are very policy oriented,

1:21:31 uh,

1:21:32 powerful leaders,

1:21:33 uh,

1:21:33 at the top,

1:21:34 wanting to do something about growth.

1:21:36 How do you link the policy dialogue.

1:21:39 from this very sort of growth driver accounting exercise to the micro foundations,

1:21:46 to what extent are public basic public economic insights brought in about

1:21:51 the role of government versus markets

1:21:54 and

1:21:55 very simple policies,

1:21:57 but the most fundamental of security of property rights,

1:22:00 lower

1:22:01 Of transaction costs,

1:22:03 you know,

1:22:03 one of the

1:22:05 empirically quantifiable results that have come up in

1:22:08 the macro literature on growth is how trust,

1:22:11 you know,

1:22:12 this very intangible concept of trust

1:22:15 matters so empirically significantly for growth.

1:22:19 So I'm just wondering in that policy dialogue along with your messages about

1:22:24 Just relying on investment is not sustainable.

1:22:27 Is there scope to think

1:22:30 with more micro foundations public economics approach to public policy?

1:22:35 Thank you.

1:22:36 Thanks to the Art

1:22:41 Uh,

1:22:41 thanks,

1:22:41 Dean.

1:22:42 Thanks,

1:22:42 Steven and Norman.

1:22:43 I,

1:22:43 you know,

1:22:43 watched this take shape over quite a few years and it's great to see,

1:22:47 uh,

1:22:47 it's great to see the book out.

1:22:48 Uh,

1:22:49 three quick questions.

1:22:50 The one is,

1:22:51 um,

1:22:52 have you thought about doing a sort of retrospective forecasting exercise?

1:22:57 So imagine that you guys have designed this model in 1980 and

1:23:01 ask how you do over the 20 and 40-year growth horizons.

1:23:05 Um,

1:23:06 so the projections that come out of the model.

1:23:08 So I think that would really help to both.

1:23:11 Strengthen the messages.

1:23:13 In those dimensions where,

1:23:14 you know,

1:23:15 you,

1:23:15 uh,

1:23:17 where the model does well in,

1:23:18 in sort of projecting the future,

1:23:20 um,

1:23:21 uh,

1:23:21 and also would,

1:23:22 would sort of point out to,

1:23:23 to the caveats.

1:23:25 Um,

1:23:25 second one,

1:23:26 building a little bit on the ques second question,

1:23:28 building on the question that came online.

1:23:30 Not just

1:23:30 would you apply this model at the sector level,

1:23:32 but have you thought about doing a sort of

1:23:35 uh sectoral version of the model in which,

1:23:38 you know,

1:23:38 you can capture

1:23:39 intersectoral factor reallocation its contribution to growth,

1:23:43 which quantitatively is a,

1:23:44 is a big deal.

1:23:46 And the last question,

1:23:47 which is not really unfair,

1:23:48 it's a little unfair and maybe it's more a question for Andrew than for you guys,

1:23:51 but

1:23:52 why isn't this the only tool

1:23:54 that we're using in CEMs for long-term growth questions,

1:23:57 right?

1:23:57 Cause there's a lot of

1:23:58 proliferation of tools.

1:23:59 What are the barriers,

1:24:00 you know,

1:24:01 what do we need to do to ensure that it's,

1:24:03 it's used even more widely.

1:24:06 Thanks,

1:24:07 thanks everyone.

1:24:07 So,

1:24:08 uh,

1:24:08 we have 5 minutes left,

1:24:09 so don't feel like you have to answer

1:24:12 everything extensively.

1:24:13 If you could take sort of 3 minutes and then we can have 2 minutes,

1:24:16 uh,

1:24:17 with,

1:24:17 with Andrew at the end,

1:24:18 uh,

1:24:19 over to you.

1:24:23 All right.

1:24:25 Thank you very much.

1:24:26 Uh,

1:24:27 on Serco,

1:24:28 actually,

1:24:28 the,

1:24:30 I will just say that we have now,

1:24:32 I am now uh in the

1:24:35 director of the Global indicators Group.

1:24:37 We do lots of work on the the business climate.

1:24:40 And also on um

1:24:42 women's economic empowerment and that will certainly

1:24:46 have an effect on female labor force participation

1:24:49 and the productivity of the,

1:24:51 of the economy.

1:24:52 So I would uh

1:24:53 direct you to the work that we do in that regard in in the Global indicators group.

1:24:58 And I think that we can have a,

1:25:00 a,

1:25:00 a close look at what you are doing in your CEM

1:25:03 and see how the two,

1:25:05 these,

1:25:05 uh,

1:25:06 two forces on female female labor force participation in business

1:25:09 business climate can affect,

1:25:11 uh,

1:25:11 growth,

1:25:12 uh,

1:25:12 prospects for uh for Egypt.

1:25:14 And I think,

1:25:15 uh,

1:25:15 your,

1:25:16 the instincts that your,

1:25:18 uh,

1:25:19 colleagues in MTI saying about uh the effects,

1:25:22 the limited effects of public

1:25:25 investment are,

1:25:26 are right.

1:25:27 Angela,

1:25:28 uh,

1:25:28 we,

1:25:29 we,

1:25:29 we may want to,

1:25:30 uh,

1:25:30 uh,

1:25:31 uh,

1:25:31 to speak uh bilaterally on how you could,

1:25:34 uh,

1:25:35 use the,

1:25:36 uh,

1:25:36 the total factor productivity extension for your Indonesia work.

1:25:40 I think it,

1:25:40 it can,

1:25:41 it can allow you to,

1:25:42 to do it if you

1:25:44 have data for what we call the,

1:25:47 the proxies,

1:25:48 uh,

1:25:48 of,

1:25:49 um,

1:25:49 of the,

1:25:50 of these,

1:25:50 uh,

1:25:50 five

1:25:52 areas that affect productivity.

1:25:55 Uh,

1:25:56 well,

1:25:56 clearly the female labor force can be parameterized and

1:25:59 that's something that we do in the model.

1:26:01 Uh,

1:26:01 on the sectoral growth,

1:26:03 actually that we had an extension that we are

1:26:06 preparing on sectoral growth,

1:26:08 uh,

1:26:08 that has three different,

1:26:09 uh,

1:26:10 sectors in the economy,

1:26:12 uh,

1:26:12 and that we want to see,

1:26:14 uh,

1:26:15 in fact,

1:26:15 that we would,

1:26:16 there would be uh productivity gains of going from one sector to another.

1:26:20 Uh,

1:26:21 and,

1:26:21 uh,

1:26:21 we have not included it here because it's not,

1:26:24 uh,

1:26:25 ready,

1:26:25 but it is in uh,

1:26:26 in the pipeline of,

1:26:27 uh,

1:26:27 of work

1:26:28 that that we intend to,

1:26:30 to do.

1:26:30 The complication with that is that we couldn't do it in Excel,

1:26:33 actually,

1:26:34 as that was our,

1:26:35 uh,

1:26:36 our goal,

1:26:36 to always have very simple,

1:26:38 uh,

1:26:38 a very simple toolkits.

1:26:40 We have to do it in,

1:26:40 in MATLAB.

1:26:42 Uh,

1:26:42 Suti,

1:26:43 finally,

1:26:43 on,

1:26:43 on,

1:26:44 uh,

1:26:44 policy dialogue,

1:26:45 I think that this is something that we rely a lot on,

1:26:49 uh,

1:26:49 on our colleagues and our,

1:26:51 uh,

1:26:51 our

1:26:52 colleagues in,

1:26:52 uh,

1:26:53 in operations

1:26:54 to help us understand,

1:26:56 uh,

1:26:56 the main concerns of,

1:26:57 of policy makers

1:26:59 and then how we can transform those those

1:27:01 concerns into things that we can actually handle,

1:27:04 uh,

1:27:04 in,

1:27:05 in the model.

1:27:05 Some of the things that you mentioned are

1:27:07 in fact included in our productivity assessment.

1:27:11 I

1:27:13 OK,

1:27:14 so I'll just add um some,

1:27:16 some minor uh,

1:27:17 comments,

1:27:18 so,

1:27:18 uh,

1:27:19 the,

1:27:19 the second uh Webex comment um.

1:27:22 So on the TFP extension um feeds a path of,

1:27:27 produces a path of TFP which can be fed into the main LTGM.

1:27:31 So that's a,

1:27:32 so they,

1:27:32 they,

1:27:32 they can work together,

1:27:33 you just link one spreadsheet to another.

1:27:36 Um,

1:27:37 the relation to historical,

1:27:38 uh,

1:27:40 relation to some other tools,

1:27:41 so I think they're quite complementary,

1:27:43 um,

1:27:43 and uh tools like the growth accounting tool,

1:27:47 can be used to uh estimate historical TFP

1:27:50 which can guide some of our future assumptions.

1:27:52 Um,

1:27:52 I'm happy to talk more bilaterally about that.

1:27:56 For the uh the YouTube uh comment,

1:27:58 female labor force participation is something you can,

1:28:00 you have complete,

1:28:02 Uh,

1:28:02 flexibility in the tool about how you can calibrate,

1:28:05 uh,

1:28:06 future versions,

1:28:07 future,

1:28:08 uh,

1:28:08 values of that,

1:28:10 um,

1:28:11 to compliment Norman's question,

1:28:13 Norman's response to Soodi's question,

1:28:16 um,

1:28:17 so

1:28:18 it's true we also,

1:28:19 as Norman says,

1:28:19 rely a lot on the country economists there,

1:28:22 but I think the role of the LTGM is

1:28:24 to provide

1:28:26 sort of a first step of breaking down a big problem.

1:28:29 So the government sort of has this big goal of growth and then

1:28:35 we sort of don't even

1:28:36 know where to start to discuss that and

1:28:38 the LTGM helps to break this down into different

1:28:41 growth drivers and then we can have those discussions

1:28:45 on the more micro determinants.

1:28:47 Um,

1:28:48 uh,

1:28:48 in terms of art's question about the retrospective,

1:28:51 uh,

1:28:51 forecasting exercise,

1:28:53 so I'm hoping to do

1:28:54 one of those just before I retire,

1:28:57 um,

1:28:57 as it's a long-term growth model.

1:28:58 No,

1:28:59 the serious,

1:28:59 the serious question is,

1:29:01 uh,

1:29:02 so

1:29:02 the career chapter actually does this.

1:29:06 Um,

1:29:06 so the career chapter,

1:29:07 uh,

1:29:08 calibrates the model at various points in time

1:29:11 and looks at what growth would be

1:29:14 if we applied those in the future,

1:29:15 and,

1:29:16 um,

1:29:18 So it has um the the the result is

1:29:21 that it,

1:29:22 you know,

1:29:22 it is,

1:29:23 if you make good assumptions,

1:29:24 you get good results.

1:29:25 The challenge for Korea's case is that some of the growth drivers change over time.

1:29:30 Um,

1:29:32 So I think that's all for me.

1:29:34 Thank you very much for your comments.

1:29:36 Thanks.

1:29:36 We are out of time,

1:29:37 but Andrew,

1:29:37 I want to give you just a chance to come in if you had a burning two-hander.

1:29:42 Sure,

1:29:43 just a,

1:29:43 just a real quick uh thing because we're talking about the long-term growth model.

1:29:48 I just want to emphasize that levels matter.

1:29:52 Um,

1:29:52 the reason we're interested in growth is because it increases the levels

1:29:58 and that often

1:29:59 what we're seeing in terms of sustained growth over the long run is in fact

1:30:06 the result of a series of

1:30:09 level changes occurring time and time and time after again,

1:30:14 uh,

1:30:14 either to TFP

1:30:16 or

1:30:17 to

1:30:18 To labor inputs or investment rates

1:30:22 and so just a plea that as we're working forward that we

1:30:26 keep in the back of mind and as we're using the tool

1:30:29 that what we're really after is increasing the level

1:30:32 and we achieve that through

1:30:34 growth as opposed to having growth as the objective itself.

1:30:40 Thanks,

1:30:40 Andrew.

1:30:40 So,

1:30:41 in,

1:30:41 in closing,

1:30:42 let me just thank you,

1:30:42 Andrew,

1:30:43 first again for your comments and,

1:30:45 and,

1:30:45 and reactions and again for stepping in at the last minute.

1:30:48 Uh,

1:30:48 I wanna thank all of you for participating online and in the room.

1:30:51 Uh,

1:30:52 I thought this was an incredibly rich discussion.

1:30:53 I wanna thank

1:30:54 Steven and Norman for a really good,

1:30:57 clear presentation and for engaging on the,

1:30:59 on the questions,

1:31:00 and also to congratulate you on,

1:31:02 I mean,

1:31:02 it's not a capstone.

1:31:03 I guess it's a milestone in the project of bringing together where things are now,

1:31:07 and it's already clear that this is moving forward with additional extensions

1:31:11 and I think we're all excited to see

1:31:13 where this goes next.

1:31:14 So congratulations on the road here and,

1:31:16 and looking

1:31:17 forward to engaging,

1:31:19 uh,

1:31:19 going forward.

1:31:19 So please join me in,

1:31:20 in thanking and congratulating them.

showAllTimestamps

no

transcript

OK, I'm gonna get going, um. With uh, with our session today. Um, my name is Dion Filmer. I'm the director of the World Bank's Development Research Group, and it's my great pleasure here to welcome you all to the March edition of our policy research talk series. Uh, as many of you know, these talks give us an opportunity to present work coming out of the World Bank's research department with the goal of sharing the findings with colleagues inside and outside of the department as well as outside of the World Bank. Um, I'd like to welcome our online audience, um, both on Webex as well as YouTube, as well as our in, in-room audience today. Thank you for coming. Um, today I'm really glad to introduce my colleagues, um, Steven Pennings and Norman Loweza, who will be talking about the project they co-lead on the World Bank's long-term growth model. And this is on the occasion of the launch of the book on the project, actually, here, let me just hold up the book so you can see it. Available at fine bookstores everywhere, including the World Bank's online bookstore. Um, Um, Stephen is a senior economist in the macro growth team of the research group here. His research interests include fiscal policy, especially fiscal transfers, economic growth, political economy, exchange rate pass-through, and monetary policy. Norman is the director of the Indicators Group within DEC, the Development Economics Vice Presidency, and he was formerly in the research group, so we claim him as an honorary member. He was a lead economist in our development research group and managed the Asia hub of the research group based in Malaysia. He was the director of the World Development Report 2014 on risk and opportunity, managing risk for development. His research has dealt with various areas of economic and social development, including macroeconomic management, economic and social development, including macro microeconomic flexibility, private and public saving, financial debt and stability, natural disasters and crime and violence, a broad range, and a fascinating research, uh, uh, history. Um, unfortunately, our previously scheduled, uh, discussant, uh, Manuela Francisco had a last-minute conflict, and she is unable to join us. Thankfully, we have Andrew Burns, uh, who has agreed to step in to be our discussion today. Andrew is the World Bank's global lead for macroeconomic modeling and works in the bank's macroeconomics and fiscal management global practice. Um, prior to joining the macroeconomics practice, he worked for 10 years as the manager and lead author of the World Bank's Global Economic Prospects Report, which examines trends in the global economy and how they affect developing countries. Uh, before joining the World Bank, Andrew was the OECD at the, with the OECD where he was the head of desk for Czech Republic, Hungary and Poland and France, um, and also played a lead role in the OECD's job study overseeing the evaluation of labor market policies in each of the organization's 30 member countries. So today I'll ask Norman and Steven, and they're going to split the time between them, to talk for approximately 40 minutes, after which we'll hear from Andrew for about 10 to 15 minutes. Uh, we'll conclude the session with Q&A from the audience. If you have a question, please use the raised hand option in Webex, or of course if you're in the room, please just raise your hand. Um, or signal to me online in the chat that you have a question, I'll call on you. If you're following on YouTube, uh, please submit your question in the chat, and that will get relayed to me. As a reminder, we are recording the sessions, uh, and please mute if you're not speaking. So with that, let me turn it over to Steven and Norman. Good morning. Good morning, everyone. Uh, Steven and I are going to, uh, present a brief introduction to the book that we have edited on the long-term growth, uh, model. It is, um, uh, a collection of articles that are about the fundamentals, extensions, and applications of the project of the same name. So football fans. I would like to play like Messi. And uh pop singers would like to sing to like Aretha Franklin. For economic growth, countries would like to grow like Korea. And why? Because Korea's economic growth in the last six decades has been spectacular. They were able to achieve an annual growth rate of 7.2%, almost 6% per capita. With that they increased the. Uh, GDP per capita by a factor of 30 over the six decades and, uh, having less than 10% of GDP of the US by 1960, uh, by 2021 they were over 50% of that level. So Korea is clearly the growth star. Now, why should uh countries uh be so interested in, in economic growth? Well, because, simply put, economic growth is the foundation of, of, uh, social and economic development. It uh generates jobs and income, especially for a labor force, young labor force entering the market. It, it fosters innovation and entrepreneurship. Which in turn creates more growth. Uh, Especially when it's inclusive and sustainable, it can promote political and social stability. And it provides services, uh, uh, resources to the state so that the state can in turn, uh, provide pub uh, public services and infrastructure. Now, it is Therefore, the key economic growth, the key to eliminating poverty and providing for shared pros prosperity. Now of course this can be achieved especially when growth is inclusive, when it's sustainable, and when it's enduring. Uh, it's not a surprise, therefore that economic growth is at the top of our policy agendas around the world. Now, dreaming big on growth is maybe a good thing. Under certain conditions. So yes, dreaming big on growth is good if you are implementing a sound economic strategy. That, that one that will allow, allow you to achieve a realistic growth rate. But it may be problematic if it leads you to poor decision making and social frustration. For instance, if you, if you have expectations for growth that are just very high but unrealistic, governments may, may go in a path of a very large public expenditures, deficits, and debt that they will not be able to, to pay in the future. It may also distort the plans that they may have. For instance, if they think that growth is going to be very high, then, uh, they might want to, um, spend in luxuries rather than necessities or have their priorities all wrong. Uh, and when the, uh, these high expectations for growth are not met, then social frustration can ensue. With that, disappointment and social unrest are, uh, are unfortunately likely outcomes. The government may lose credibility if they just present this very high expectations for growth. So the origins of the long term growth model are rela are actually related to this uh as a reality check. We wanted to see if the growth projections that many countries were making. Uh, actually, uh, were supported by a standard neoclassical growth model here for those who know the literature associated to the names of Solo and Swan. But After becoming a reality check, we realized that we could do more with this tool, and then it became an, uh an explanatory tool, one that would allow us to produce alternative growth scenarios and also a way to understand the drivers of, uh, of growth that are behind those uh scenarios, including drivers related to policy. So there was a a lot of interest in uh in the LTGM and that generated in fact a demand for richer models. So the project expanded in terms of the number of models that we were able to produce. They were supported by academic papers and we made a strong effort to uh. Implement each of the models that we produced with a toolkit. So each of the models that we have, um, are, uh, carried in a tool kits that can be used in a very simple way. They are transparent, they are, as I said, easy to use and you can download, download all these from, uh, the project's website which is given, uh, here in the slide. So let me, uh tell you about the, the book structure which it will also be the structure of our presentation today. So, uh, the, the title of the book tells you what this is about. We have a section on the fundamentals of the, of the long-term growth model. Uh, the first is uh a chapter on the, the standard, the basic model. Then we have some extensions dealing with public capital, productivity growth, and natural resources, and we, uh, present a number of of applications of how this has been done in, uh, in different contexts for different countries. Uh, in fact, this is only a selection of the many applications that we have conducted over the years. Let me, uh, start actually from the end and tell you very briefly about uh these applications. Uh, we did one for Malaysia, the basic, uh. Uh, objective there is to study the transition of Malaysia from a middle income to high income economy, and we found that one of the main impediments of that transition was, uh, the low female labor force participation. And we had some recommendations regarding that particular area of improvement. The second application is on Korea. Well, we, I told you already that Korea is the, the growth star around the world and um we wanted to see. If, uh, the model that we use now for projections could actually be used also to explain what happened in Korea in the past, so this is, uh, rather than a futuristic, uh, application is one that goes into a retrospective. To study the engines of growth. And what we found uh through that study is that the drivers of growth actually change over time. In the 1960s, we had labor and human capital being the main drivers. In the 1970s, it was mostly about capital accumulation, and from the 18 1980s onwards, it was productivity growth. In the case of Bangladesh, the, the 3rd application here in the list, uh, we wanted to study if the high growth, uh, rates that the country has been able to achieve would, would, uh, continue, uh, in the future, and, uh, the conclusion that we arrived was that investment was not enough actually and that the total factor productivity will have to increase if the country was to, uh, continue with high growth rates. The Syria example is uh is is very different. The Syria example actually looks at the devastation that occurred because of the brutal war, uh, civil war in the country, and we did a study that you could regard maybe as science fiction here of uh what would happen after the war. And what we did was to link these scenarios of of growth in Syria with political settlements that will drive reconstruction assistance and also voluntary repatriation of people into Syria and how the different political settlements actually led to very different conclusions. The last two applications on Egypt and Egypt and Sri Lanka were, uh, are there also mostly for sentimental reasons you could say because this is the way that we started the project, uh, studying Egypt and, uh, Sri Lanka, so we wanted to acknowledge those origins, uh, where we mostly use the model as a reality check. OK, so this was, uh, this is half the book, but I, I, uh, I think, uh, we can say that it's, uh, it's covered and we invite you to, to read the applications if you are, if you're interested. Let me briefly talk about fundamentals now. So, the, uh, as I said before, we work with a neoclassical growth model. It has three building blocks. One, a production function that many of you are familiar with. Uh, it has an output being a, uh, being a, uh, a, the result of combining productivity, uh, here. Uh, denominated by the letter A, then capital and uh labor augmented by human capital. And there are decreasing returns to the factors of production, capital and labor, that which makes it a neoclassical growth model. There is a the second building block is a capital accumulation equation that tells us that capital decreases with depreciation and increases with with investment. And uh we want, we were interested very much in the demographics uh and labor market forces that drive growth and we realized that of course, uh in the production function what matters is the number of workers, but uh for uh welfare purposes we want to understand uh GDP as a ratio to Uh, to population, so GDP per capita. Uh, of course, uh, the connection is related to two factors. One is, uh, the working age population. Uh, here represented by this ratio of W over NT and N is population work, W is working age population, and the labor force participation that is L divided by W. So working age population, something demographic, population growth is a demographic force and labor force participation in fact is a choice. OK, so, uh, if you do a little math here, uh, you will get a growth equation. It's um. An approximation, but one that is very sensible where you have that uh GDP growth is a function of productivity growth. It's a function of things related to um to people like human capital, demographics, and labor force participation. All of these multiplied by the returns to labor, uh beta in this case. Then you have the capital segment. This is uh investment multiplied by the marginal product of capital and the depreciation component. So this is uh a uh an, an important equation that we use in our projections. You can see how all the factors that I have been discussing affect actually growth. And uh this is what we have in terms of how we go from conditions and assumptions to outcomes in the model. So we have some initial conditions and the model worked very well because it relied on very simple assumptions, very simple initial conditions. We just needed to know the capital to output ratio. A few parameters and also the capital labor shares and the depreciation rate. Once we, once we have those conditions taken from reality, and we didn't invent those, we took them from the data, then we applied a number of assumptions on the future path of growth drivers. I already told you what those growth drivers are productivity, human capital, demographic growth rates, and importantly, the investment rate. So this is where the art comes in, what the right assumptions should be, and then we create with different assumptions scenarios, and those scenarios will lead us to the outcomes of interest. The, the most, uh, relevant outcome in this presentation is the growth rate, but actually we explain in, in one of the chapters there how you can link changes in growth rates to poverty rates as well. And you can use um uh some assumptions on the distribution of income in society and if you do that, then you can link changes in growth rates with changes in poverty. So that's the way it works and uh I can just tell you 11 illustration of uh a lesson that we can derive from this very simple model. So this is a result that investment-led growth by itself is not sustainable in the long run. Why can we say that? Why, by looking at this equation? Well, we can say it because, uh, if we increase in fact this investment rate, what's going to happen is that the capital output ratio is going, is going to increase, and if you look at that equation where it says the marginal product of capital. Uh, capital output ratio is in the denominator, which means that actually there will be a decrease in that, uh, in that factor. What that means really, uh, in, uh, intuitive terms is that since there are decreasing returns to capital, as you increase the capital output ratio, then the strength of capital in producing growth goes down. And that's something that, that, uh, in the end will uh eliminate uh the growth force that investment can have. Therefore, The policy advice is that if you want to achieve growth uh in the long run, then you cannot rely just on investment. It must be accompanied by other sources that drive growth like human capital, productivity, uh, improvement in labor force participation, etc. so. Generally speaking, what we are advocating here for a broad-based growth strategy. OK, so I have in a just a few minutes to cover the fundamentals. There's a lot more in the book and I, and I will invite you to, to read, uh, what is behind. So now I will discuss, uh, one of the extensions and then Steven will discuss two other extensions and also provide us with the conclusions. So the is the extension that I will discuss has to do with productivity growth. Now this is the A letter the a factor in uh the equation that I showed you before. It happens to be very important, but it's also something of a black box, a residual, as, uh, Solo said in, in a, in a, in a in a famous quote, uh, the degree of our ignorance, I think that's what he said more exactly. Well, but, uh, in fact, that ignorance can be, uh, circumvented if we study what the literature, literature says about the drivers of economic growth, oh sorry, of productivity growth. So we wrote this paper with a colleague uh in the Malaysia hub, uh, Young Kim. Uh, we did a very large literature review on the determinants of total factor productivity. Uh, we identified these five areas, innovation, education, market efficiency, infrastructure, and institutions as drivers of total factor productivity. We, uh, moreover, examine, uh, what, uh, proxies could be used for each of these factors, and what we did is we, uh, try to understand the link between total factor productivity growth and that index that we had created. We so we run this is a simplification of course but we run a a a regression of the average TFP growth, um, uh, on the this index of uh the drivers of of the of productivity and uh the past level of TFP just allow for some regression to the mean. And what we found is that the higher the index, the higher uh TFP growth would be with a diminishing effect. So, uh, we then applied this to a number of cases and let me show you one of them. I have spoken a lot about, uh, about Korea, uh, and, uh, what we decided to do is see, well, what, what would happen with TFP growth if countries were to do exactly what Korea did over the last three decades. What would happen with with with uh the TFP growth of those countries and we took averages for the different regions and that's the result of one of the simulations that we did as you can see the countries that were lagging behind the most in uh sub-Saharan Africa and in MENA were able to achieve the highest increases in uh TFP growth, reaching up to 2%. It doesn't seem like a lot, but over time it builds and actually it makes a difference, and those countries that were not, uh, that had already increased, uh, TFP significantly like those, uh, neighboring countries in, in, um, in, um, East Asia where did not benefit all that much from, uh, copying Korea, maybe because they had already done it. We could do this also at the country level and in this case I show you the uh my own country Peru and we did some uh simulations for for Peru uh in this case following Korea the world leader or Colombia, the regional leader in in Latin America and uh we were able to describe. First, what happened to this index of the drivers of TFP that's on the left hand side of this graph and then the effect on uh TFP growth rates for uh for Peru. OK, so I have, I have told you about uh productivity growth and now I think it's the time for me to cede the floor to Steven. Thank you very much. OK, thank you very much, uh, Norman, and thanks everybody for, uh, for attending. So as Norman mentioned, I'm gonna talk today about two extensions uh to the basic model, and the first one of that, it relates to public capital. So the question we try to ask in the public capital extension of the long-term growth model is what is the quantitative effect of public investment, that is infrastructure on growth, recognizing that uh public public investment and infrastructure are not always exactly the same, but in many cases they are. So this is obviously a very important question for development policy uh in general and in particular for institutions like the World Bank, for lending institutions. So there's kind of two, opposing views about this, uh, in the literature, uh, and in development policy. So one is that, um, public investment and infrastructure investment have a really big effect on growth and should be really at the core of a growth, uh, of a growth plan, and, um, people refer, for example, to, um, China's uh very strong growth rates and their high rates of, um, Infrastructure investment, large human needs, and of course big infrastructure gaps. So there's a World Bank report that lists 700 million people without safe drinking water and 1.2 billion without reliable electricity. So these are obviously great needs. And there's also some academic literature that also produces very big estimates of the effect of public investment on growth, and this, this paper is an example of that. Sort of the opposing view is that this is gonna have a small, uh, a small effect, and, and part of that, um, is based on a lot of the empirical evidence from literature that finds very, very small impacts of higher public investment, uh, on growth, and also questions some of the other large impacts that they don't have causal validity, um. And part of this discussion is, is about why, why that would be the case, and obviously corruption and politics are key. Um, so Lant Pritchard has a very famous paper where he argues that in many countries government investment spending has created very little useful capital. And uh Nack and Kiefer have a have a paper that argues that in countries with poor governance, um, governments use public investment as a vehicle to increase their rent seeking. So, so these are sort of 22 opposing views, um, in the literature. Um, we're not trying to get into that debate exactly, um, but we, we include, um, an effect on growth, but we also take, um, some of these, uh, lack of efficiency seriously. So perhaps a bit more generally in this discussion about uh public investment and growth, there's a lot of confusion in the literature. Um, beyond the size of, of the effect, there's confusion about levels versus changes, the effects on the poorest countries, and of course dynamics. So, so is it that levels or changes in public investment, um, Affects growth, how do the effects vary across countries, and how long do the effects last? If we have a permanent increase in investment in public infrastructure spending, is that gonna boost growth forever? Um, so, the LTGM public capital extension is gonna answer all these questions in a very simple spreadsheet format, and this is um joint work with Shamila uh Tavadas. So the LTGM, uh, public capital extension is, is a very simple extension of the, um, basic framework that Norman presented. So, uh, one can see, um, If you start with the um the production function that that Norman showed, um, there is a total capital stock which is, which is represented by K here. And so what the public capital extension does, uh, for the most part is just to split this into two parts. So, um, there's a government capital which is mostly infrastructure, we also call it public capital, um, and then there's also a private capital stock. Um, so that's the most important change in the model. Of course we also take into account, um, this potential lower efficiency. Um, so that's, uh, Pritchett and many others, uh critique, um, that when you spend an extra dollar on, on public investment in many countries, you get less than an extra dollar of uh public capital stock. And we actually produce a new infrastructure efficiency index, um, to measure that, uh, which includes electricity losses, unpaved roads, uh, and water losses as well. Um, and the calibration of the model is based on estimates from the literature, um, and it builds in, uh, A pretty sizeable effect of public capital on growth, so more specifically this parameter means that a 10% increase in um the public capital stock is gonna boost GDP by 1 to 1.7%. So what is the what is the quantitative effect of a permanent increase in public investment on growth in the model? So this is gonna vary across countries, so we're gonna look at this for representative countries of each um income group. So I think the results can be summarized as saying it's, it's helpful for growth, um, but it doesn't create a growth miracle. So what we have here is we have kind of a um an like an impulse response function where there's a a a a one-time increase in public investment as a share of GDP um by 1 by 1% point and uh that that is a permanent change, and then this is the change in in growth. So one can see that there's a a modest uh and temporary boost to growth. So, uh, these different lines reflect, uh, typical countries calibrated to, um, high income in blue, upper middle income in yellow, uh, lower middle income in black, and, uh, low income in red countries, um. So, uh, one can see in the short-term that a 1% point of GDP increase in public, uh, investment, uh, across these, these, these different representative countries, um, boosts growth by about 0.15 to 0.2% points in the short run, and then in the first few years. Um, that's excluding, uh, demand side multiplier effects, supply side effects only. Um, however, so, so, so I'd say that's a modest, uh, that's a modest boost in growth. Um, but it's also a temporary one, so one can see that this, this line in terms of the increasing growth relative to the baseline, slopes down over time, and so after 30 years, the boost to growth is only 0.05 to 1.1% point. Um, and the reason for this, um, is that the the size of the effect depends on the public capital to output ratio, and, uh, there are diminishing marginal product of capital, there's diminishing marginal product of capital over time. When you increase public investment, um, that's going to increase the public capital stock relative to, uh, faster than increasing GDP and so this, uh, ratio is going to increase, um, over time and reduce the effect on growth. Um, the other interesting result here is that, uh, for our sort of representative countries, there's actually a pretty similar effect on, on growth. Um, for, uh, you know, for high income countries and low income countries, it's actually not all that different. Um, and the reason for that is that this, uh, that on average within those income groups, the, um, capital, the public capital to output ratio is not actually all that different. Um, so that's a somewhat, uh, surprising. Um, surprising result and obviously very policy relevant. So we produce many results, uh, the, the model produces many results and I encourage you to read the chapter for a full, uh, for a full explanation, but to summarize some of the main ones, um. So the effect of public uh investment and private investment is gonna be heterogeneous across countries and also over time. So even though those averages for typical high income, middle income or low income countries are actually not all that different, within each group there's a lot of heterogeneity. Um, and it's because that's gonna, because that's because it depends on the um shortages of public capital relative to GDP and private capital relative to GDP. And so those will depend on on country uh circumstances, and that's why we need to have a tool like the public capital extension, uh to calculate the effects on growth. The second interesting finding is that in low income countries, actually extra private investment has a a bigger effect on growth than public investment, which is something which is maybe a little bit surprising. Um, and the reason for that is just that the public, uh, that the that the private capital stock as a share of GDP is really low in low income countries. Um. And to some extent that's intuitive because those countries have a lack of access to credit, they often have poor business environments, um, and so that really constrains, uh, private investment, and so the policy, um, the, the policy advice following that is to try to increase private investment in those countries. The third result is perhaps a little bit more technical, but it is also surprising, is that the level of efficiency um actually has no no effect on the impact of public investment on growth. So this is not our result, it comes from a paper by, uh, an IMF working paper by Bergal, and it's, it sort of stems back to the two forces that are embedded in the opposing views I showed a couple of slides ago. So this is a sort of simplified version of the um equation Norman showed, adapted to the public capital extension context, and uh one can see that the sort of the first view, Um, if you have, um, low, if you have low efficiency, which is a low, um, value of theta, um, You might have, let's say in a particular country you might have a billion dollars of public capital on the books, um, but in reality, a lot of that public capital is very poorly constructed, and so, um, the actual effective capital to public capital to output ratio is much lower, and that increases the marginal productive capital. Effectively it increases the need. Um, but then an offsetting force that moves in the opposite direction is that when you spend an extra million dollars on public, uh, investment in those countries, you might only get, uh, $500,000 of useful capital because it's poorly installed, it's politically allocated, and so forth. Um, and so, uh, one can see that if the new public, the efficiency of the new public investment and the old public capital are the same, these effects cancel out. But that's not to say that efficiency is not important, um, so improving the efficiency of public investment can actually have a sizeable effect on growth, and especially in countries that have high rates of public investment. So we applied this, um, this model, uh, in, in Malaysia, which is uh an application in chapter 5. so in Mala Malaysia is on the verge of moving from middle income to high income status, and policymakers were interested in knowing how they could sustain growth in Malaysia. And one of the, uh, things that came out of that discussion was whether we could use higher public or private, uh, investment to the sustain sustain the growth rate. And so this is uh, the baseline we produced, uh, at the time. Of the growth path in Malaysia going out to 2050, and one can see the declining rate. And then we did some stimulations where we had a permanent increase of, of investment, we had a first a 1% point increase in in public investment, which led to a short run increase in growth of 0.15% points. Uh, we also increased private investment by 0.3 by 1% point of GDP and that led to a bigger increase in growth. So in Malaysia, Malaysia was relatively short of private investment, and then we also tried a large, an even larger effect which is 0.6% points. So one can see that that growth, you know, is, um, is going up, um, but, What this, these policies cannot do is to sustain long-run growth, um, which is what the country was most interested in, this is due to the diminishing marginal product of capital that we've seen many times. Um, and as a result, we needed to consider a combined reforms package, uh, of the one that Norman mentioned earlier, taking into account, for example, higher TFP growth, uh, improved human capital, and also, uh, labor force variables like female labor force participation, and the details of that are in the chapter. So, I've finished talking about public capital, I'm gonna talk briefly about the natural resource extension, which is the final uh extension of the, in the book. Um, OK. I'm not sure What has happened here? There was supposed to be a slide. OK, for some reason, this slide is not loading. Um, But that's OK, so I will have to tell you the. I'd have to tell you, uh, the idea here, um, because this slide is not loading for some reason. Um, so the natural resource extension, uh, is motivated by the fact that many developing countries, um, have large natural resource sectors, and if that natural resource sector is a very sizeable share of GDP, then obviously, um, factors that affect, that, that natural resource sector are gonna be really important for long-run growth. And also the model. Is gonna be able to answer questions specific to natural resources, so for example, how do discoveries or resource depletion affect long-run growth, how do commodity price shocks affect long-run growth, and um the role of also the government's fiscal rule and how they deal with those fiscal revenues. Um, and the, the extension of the, of the model is actually a very simple one. So it um it basically starts with the production function that that Norman showed for the standard LTGM and then just adds an extra term, which is um the GDP produced by the resource sector. OK, so now I think we have, we're back again, um. So, um, this is the, uh, output of the resource sector, um, so I'm gonna work through an example of oil, uh, but, you know, in the model we have a spreadsheet with, um, over 10 commodities calibrated to over 50 countries. Um, so the output is going to be a function of, um, Uh, it's gonna depend, output of oil, it's gonna depend on technology, the reserves, uh, of oil, and also the capital stock in the oil sector. And so this is kind of an intuitive, um, function as one can see, when, um, when the reserves are, are depleted, we're gonna need more oil capital or a better technology to produce the same amount of oil, and of course when reserves are depleted, you can't produce any more oil. Uh, we're gonna keep track of, um, reserves based on production and, uh, exogenous discoveries. So the reserves today are gonna be equal to the reserves last year, less production plus discoveries. Um, we can allocate capital across sectors to equalize the value of the marginal product of capital. Governments and crucially, governments are gonna tax resource revenues and can either invest those in the domestic economy, Um, or they can save them abroad, or they can, can also consume them, we're not going to discuss that today. Um, and it's important to mention that the model captures the long-run supply side effects, mostly through investment of the commodity sector. Um, it doesn't capture short-run demand effects, and those are really important in commodity, uh, exporting countries, but they're not a long run, uh, issue, they're more of a short run issue. So I'm gonna talk about a a simple example, um, we had a calibrated model, we calibrated our model to Angola a few years ago, um, and we had a scenario where we had a 1010 year boom bust, uh, cycle in the oil price. So we started the simulation with oil prices at $50 a barrel, they went up to $80 a barrel and they came back down again. And, um, in Angola, which is a very big oil big oil producer, or should I say, Uh, oil is a very large share of the economy, um, the extra revenues generated from a shock like this would be, uh, almost 25% of GDP. Um, and, uh, the vast majority of those accrue to the government. Um, and so one can immediately see that the, um, the fiscal response of the government is going to be crucial for long-run growth. So, uh, in uh the red line we have a, a model, so this is this right-hand graph is a total investment, deviation from baseline, and, uh, in the red line we have a, um, a case where the government is gonna invest those revenues, uh, in the domestic economy, uh, so we're gonna have a big spike in investment, uh, versus let's say, in, in the blue case, if they save those those revenues, uh, Abroad, um, and this is gonna have a big, big differences in the effect on growth, um, so, the oil price shock is gonna have a big effect on gross domestic incomes, um, so that's basically the purchasing power of, um, domestic residents in, in Angola, um, mostly due to this, um, big, uh, this big increase in oil revenues. Um. However, uh, oil prices do not directly affect, um, real potential GDP, um, and this is sort of another point of often point of confusion, um, And that's for the very simple reason that real GDP is a constant price measure. Um, so when oil prices go up, most of that is exported, but the real GDP is calculated with constant export prices. Um, so consequently, the effect of um. Uh, commodity price shocks, uh, on long run growth is gonna work, uh, indirectly, mostly through the effects on investment. Um, so on the right hand side I have, uh, GDP, uh, relative to baseline, and this, uh, red this, uh, black line is just basically the increase in the size of, of revenues, approximately. Um, and one can see that if the government invests all these revenues, it's gonna lead to a big increase in GDP. Um, but that increase in GDP lags the, uh, boom-bust cycle somewhat, um. And uh the, when commodity prices go up, initially it has very little effect on GDP. Um, and then, uh, you know, if it saves them abroad, it has a much smaller impact. Of course this excludes the short-run demand side effects, which is gonna be really important in those countries. OK, so, um, this is our last slide. Um, so, uh, the LTGM book contains, um, many models and applications, um, and many, um, many policy, uh, policy relevant findings, so, uh, we encourage you to read the book, uh, to cover all of those, but we want to summarize, uh, four of those, um. So one of the first one is that um although all um fundamental commod uh although um fundamental economic conditions are going to always uh affect growth, um, their relative importance varies across countries, and, and as a result, the growth performance constraints and opportunities also vary with the country context. The second one which I've mentioned and Norman has mentioned several times, is that the investment-led growth strategies are going to be unsustainable in the long run. Um, investment-led growth can boost, um, GDP growth in the medium run, um, but even in that case it would need to be accompanied in most cases by, uh, increases in savings. And finally, high growth usually involves fast total factor productivity growth. So this um this project has been, um, So Norman and I are coach ETLs of the project, but we've got support from um, from many people, um, and financial support as well, so thank you to everybody who's put their work into this, and uh you can download the book um from the QR code or also from our website. Thank you very much. Thank you, Steven and, and, and Norman for that very clear presentation. Um, Uh, we're gonna now turn over to, to Andrew Burns, who is online. Um, uh, as I mentioned in the introduction, Andrew has done his own fair share of, of, of, of, uh, growth modeling. Um, so we, we're very happy to, uh, welcome him today, and again, thank you for stepping in at the last minute, Andrew, we're really glad to have you with us. Great. Well, thanks so much, Dean. It's uh my pleasure being here and thanks, uh, to, um, the team for such a great presentation. Um As indicated, I sort of come to, uh, reviewing this work, uh, somewhat late and so I apologize if my um remarks are not as, uh, coherent and, and well organized as uh was the presentation. I think that you know the the LTGM is uh exactly very much exactly very much what the team I think somebody needs to say, could you please mute yourself. Great. Um, so, You know, the framework is exactly what it is. It's a solo, uh, growth framework. Um, it's basically a very much a standard way of modeling the supply side in economics. It's basically the same supply side, the same long run determinants of growth that we have in our manager and mod models that we use within the bank. And it's the same, um, basic, uh, framework that you see in pretty much every, uh, economic model of either a long-term or short-term, uh, thing, so nothing wrong there. It incorporates important extensions, um, from the standard solo swan model to deal with the distinction between private and public, uh, um. Investment, uh, it has some endogenous growth features for educational attainment that wasn't discussed today, but it's in the book and that's really, uh, interesting and useful. It has, um, the resource extraction mod module which helps to deal with, um, something that is sometimes, uh, in some of our clients a very tricky problem which is how do you deal with. The implications of having uh a really important resource and how it will affect growths going forward. And it has an interesting, uh, simple sort of poverty extension and, and as I said, all of those things are, or at least the first three of those things, private capital accumulation, demographic effects, population growth, and TFB, those are very much a part of every model and what um is. Sort of the value added of this approach is the, um, the, the endogenous growth from labor market policy, um, introducing the, the quality and quantity of public sector investment into the overall equation and uh looking at the changes in the real rents from natural resources. Those kinds of effects um are not necessarily in a standard macro model, although, um, like with the LTGM model, we, they are introduced in those instances in countries where those are important effects. I, you know, when I was trying to think about how to describe the LTGM model, I, it, it occurred to me that in some ways, and I don't mean this in a negative, uh, way, it's, uh, less a model than it is a product. And, or, or a methodology. It's a way of looking at the drivers of long-term growth and trying to quantify the relative importance and hinge discussions that are very much uh taking place in um the, the very long term into something that is uh relatively easily um. Quantified and um understood. So we see growth as a function of productivity, of capital accumulation, of um labor force growth, and then we can see, OK, what is the impact of doing something to any one of those determinants on the long-term uh growth rate of the economy. And I Uh, you know, I think that I don't know if I'm gonna, I'm gonna try and share my screen, which is always risky effort here just to show one of the slides that um the team had up. Which is the, basically the growth drivers, right? And so we have this basic equation, which is at the heart of everything that's being done. So growth is a function of TFP growth, of the rate of growth of uh human capital, the rate of growth of demographics, so the participation rates in the working-age population. Um, and, um, then it is a function of the investment to GDP ratio and depreciation. And that leads, among other things, just having that formulation, which is a very standard um Cobb Douglas kind of formulation, leads us to the result that investment-led growth by itself is not sustainable in the long run. I actually, I don't have a problem with the results, but I do have a problem with um how the um That result is being. Discussed and thought about in, uh, in the presentation and in some of the work that has been done with the model. And I think it's kind of in some sense, one of the challenges um that's uh there in the modeling system. The result that in the long run, um, The capital, the, the investment rate doesn't have an impact on the rate of growth of capital, is a very standard or rate of growth of potential output. It's a very standard result from the Cobb Douglas, um, framework and it's effectively because in the long run, the rate of growth of the capital stock will be equal to the rate of growth of GDP. And that its fundamental determinants are going to be the rate of growth of the labor force and the rate of growth of TFP um as a function of the size of the labor force in the overall um. Income share. So, but that's, that, that result that you cannot permanently change the rate of growth of the economy by changing the investment to GDP ratio is uh very much part of the framework and a standard result. What I object to is the um The formulation that that is not a sustainable source of growth, and I don't object to the statement in itself because it's factually correct, but I think in some sense it misses the point. Because it tends to make us think, well, if we cannot sustainably increase the rate of growth through capital accumulation, then capital accumulation probably is not a good, uh, growth path or growth story. And I think that's kind of illustrated here nicely in this chart, um, in a little less nicely in the, in the Malaysian one. We can see that the, the impact on the growth rate is relatively in this particular graph, uh, large initially and then diminishes over time. But what that focus on the growth rate alone, um, obscures is the fact that the level of GDP is higher throughout this period, and in fact, the delta on the level of GDP is accumulating over time. So when we were looking at, so if you, if you think about it, the delta on GDP is going to be the integral or the, the space underneath that change in the growth rate over time. And if you look at, for example, this graph, We're looking at a change in the growth rate of just a little bit under 1% point of GDP that is sustained although diminishing over time for a period of 20 years. That's effectively by my rough and ready calculations, means that in level terms, at the end of the projection here. The level of GDP in Malaysia is 20% higher in, on the purple line than it is in uh the black line. And I think that all of us involved in um In development would say that some policy reform that managed to increase permanently and in a sustainable fashion, the level of GDP by 20% is actually a really useful and helpful policy. It may not change the rate of growth. In a, in a, in an important manner way out in 2050, but in the, the, the result of the interim effects is actually sustained long run and very important. And so that's just one sort of comment that I found um. Worthy sharing with you here. Um, The other issue that I have with the approach is, you know, in some sense, that, that there's a trade-off and a trade-off between the sort of simplicity of the model and, and therefore it's easy to use and then the Lack of complexity and therefore the missing of important interactions that are going to be occurring in the economy. And, um, that second effect, I, uh, found not so much, uh, well illustrated in the book as in uh another instance where the model was used to look at the influence of resource, um, depletion, uh, in, uh, in a, in an oil exporting economy. And just uh briefly, I sort of extracted it uh from the model. I can just try and pull it in here. This is coming from the book rather than um Rather than uh the presentation. The sort of resource uh model or the standard model is that uh the GDP is equal to TFP times um sort of quality augmented labor uh times the capital stock raised to the share of each in overall output and that standard models applied to the non-oil sector and then the oil GDP is given by a similar function. Andrew, it's, yeah. What we're seeing on the screen is still the application sustaining growth in Malaysia. All right, let me just do redo the share obviously share the application. No, no, no, thank you, um, how the heck I hate. Webex, um. Where is the unshare thing? All right. I think this is actually part of the slide that you were unable to prevent, uh, present. So this is, uh, us being doubly cursed here. Anybody wanna tell me how I can stop sharing? Oh, it must be up here. Uh, yep, you stopped there. I did. Oh, I guess so now I can try sharing again. All right, so this is the, um, you can see this now, I hope. Yes. So this is the slide to which I was referring and um basically the, the sort of long run resource enhanced model um has for the non, non-oil part of the economy, the standard Cobb Douglas uh uh approach, so that's why. Oh there for non-oil, and then the oil economy or the resource economy is um a similar uh production function, but um where resources are the second important uh source of um growth and what the model says pretty clearly, if GDP is the sum of those two, which I believe is how they're modeling it, and that's indeed how we model it ourselves. Uh, in, in our models, then, uh, as the, uh, amount of reserves declines, then the contribution of the oil sector to overall GDP growth and growth terms of rate of growth declines. But the problem with that is, in my view anyway, is that it's really missing all of the compensating adjustment mechanisms in the economy. And if we think about an oil economy, the, the classic Presentation of it is that here's an economy that suffers from Dutch disease, and what is Dutch disease? Dutch disease is the effect of an overappreciated currency reducing the competitiveness of the non-oil sector, and as a result, the non-oil sector gets squished in the overall activity in the economy. What happens in real economies as the resource sector disappears or its weight declines, then the influence of the Dutch disease effect declines at the same time and therefore the rate of growth of the non-oil parts of the economy expand. And that whole adjustment mechanism is, is not going to be here, um, just because we cannot articulate in this model, the influence of, uh, that, um, Exchange rate effect on the productivity of the non-oil sectors and that effectively results, I think, in sometimes um strong results being uh or strong conclusions being drawn from this modeling that might not be um observed if we were using um A model that was able to uh incorporate these effects. Um, So that's just kind of a, uh, I guess it's a um. A picky, uh, a picky point on my part in terms of the overall approach. I think that, um. It, it's important to recognize that, that, that this approach or the, the um The, the, the LTGM model does have a lot of real benefits and it can help us um take a look at some of the, the, the, the, the impacts and the influence and think through um the long-term influence of different policies might sort of I guess, takeaway is that when we do it, that we want to be careful that one, we're thinking about the levels, and I think that's one of the challenges of the, the long-term growth model. It is focusing on the growth rates and long-term growth rates, and as a result, we might be um not giving due attention to what's happening to the level of activity because it's that level of activity that ultimately is driving poverty reduction and it's super important. Um, and the second thing is that, um, because we're missing the price mechanisms, because we're missing important, um, adjustment mechanisms in the economy, we might also be missing, um, important, maybe their 2nd or 3rd-order effects, but sometimes those 2nd or 3rd-order effects, uh, can be, um, dominant. So, um, bottom line, um, I think that it's, it's a great tool, and I think the, the proof of the pudding is the fact that it's been used, uh, in, in many instances. Um, I think one of the big challenges for the, for the model, and that's true for any model is while it can tell us um what would happen if we improved our index of TFP from 5 to 0 points to, to 7, it can't really tell us how to do that, and ultimately, that's the development challenge. So, Uh, reiterating the point that productivity is, uh, and productivity growth is the most important, um, uh, sort of long-term driver of growth. Absolutely. How do you achieve it? I'm not sure the model is really helping us, um, very much in that regard. Um, So I think that was it. The only last um remark that I, that I wanted to make was that I thought in regards to this, how do we achieve it question, that it would be useful um for the model to be, look at some of the sort of uh drivers of productivity that are um Laid out in the CEM 2.0 tool which sort of does the same kind of exercise, but really tries to look at which are the particular Uh, drivers of productivity growth, drivers of participation rate in individual countries that are weak versus those that are relatively strong, and then use that in combination with this tool to motivate the discussion. But by and large, I think it's an impressive body of work. Um, I was thinking as I was reading it that it may be the long-term growth model because it looks at growth in the long run, but it's also the model that's been around, uh, quite a while in the bank and has had a long history, and I think that, uh, both Norman and, uh, Steven are to be congratulated for their contributions that they've made, uh, throughout the bank, uh, via this tool. Thanks. Thank you, Andrew, for, for, for those, uh, those comments. Before opening up, um, uh, I'm gonna turn back to, to, to Norman and Steven. Just a reminder, if you have questions in the room, raise your hands when we open it up here. Uh, I have a couple of questions from YouTube that I've gotten, um, if you have a question and you're on the Webex, please raise your hand or just signal in the chat. That you'd like to ask a question and I'll call on you. So, uh, over to, to Norman and, and Steven to, to react a little bit to Andrew's comments. And then Andrew, I'll also come back to you sort of towards the end of the Q&A if you have any reflections, or, or feel free to jump in anytime. Well, Andrew, thank you very much for, uh, for your comments, uh, and, and for your very kind words, uh, towards, towards us and towards the project. Uh, actually, I have to say also that we have collaborated with you over many years and received your feedback, and for that we are also very, very grateful. Uh, we, um, initiated this project as a joint Product with uh with MTI and uh I believe that we have benefited a lot from the interaction with operations that is every single uh extension, every single application was driven by the interest from our MTI uh colleagues and from our colleagues in operations. Uh, so that that level of interaction has been uh very beneficial for us, and I hope that, uh, we, uh Steven and I hope that that that has been also been beneficial for, for you guys. Uh, you pointed out several, uh, limitations, and, and we agree with most of them. Uh, I, I want to add that, uh, we see this model as complementary to the work that you, Andrew, and all the people in MTI in the Global Prospects, um, Uh, group, uh, do in terms of, uh, short-term forecasting. Uh, our, our work is not really for, for the forecasting of, of growth in the, in the next 235 years. That is very much dependent on international conditions, for instance, uh, dependent on the demand effects which we don't have in the, in the in the model because we Uh, are actually assessing potential growth rates that, that depend on this equilibrium between demand and supply, and therefore everything is driven by the production, uh, function of the, of the economy. So this is uh a limitation that we, uh, recognize that we don't consider demand effects and that we don't, uh, consider short run and cycle cycles in the economy. But, uh, if you regard this as complementary to the work that is done. Uh, in other parts of the bank, then uh you see that, uh, uh, rather than a limitation, it's just a distribution of tasks among different groups in, in, in the bank, which we welcome very much. This is the first reaction. The second reaction is that uh we we agree with you that uh the that although investment does not in the long run lead to uh uh to to growth, uh, that's what happens in neoclassical growth models where they are decreasing returns to the factors of uh of production, uh, that the level of GDP, uh, is in fact, um, Uh, permanently, uh, affected, and that it will benefit from any level of, of, uh, of growth. That is undeniable. And, and that is something that we should, uh, uh, recognize maybe more, more explicitly in, in the, in the presentation that presentations that we make. Uh, uh, the only thing here to point out is that for many Uh, for many policymakers, for many countries, the challenge is not only to increase the level of income, but actually to increase growth rates because that is indeed what's going to have a a more permanent effect on development. Um, it, it is a, a level of how dynamic the economy is, and it's that, um, that dynamism which actually leads to development and that's why we emphasize growth rather than, than levels. So we recognize the point that you are making, of course. On, uh, just the last issue that I will, uh, address is, is on productivity, productivity growth. We do have 11 of the chapters dealing with the drivers of productivity, uh, and, uh, what is missing, uh, possibly there is that, uh, some of the productivity components, uh, are actually uh can be embedded in, um, in the other factors of production. So, for instance, uh, when we talk about infrastructure as being a driver of productivity, uh, this could come in the form of some capital. Uh, when we do, we say innovation, this could also come in the form of investment of new types. Uh, and, uh, we go through the simple way of uh of adding that to growth through uh total factor productivity, but if some of that is embedded in capital, then uh that will have an effect as well. That's something that we don't, don't consider, but it is an interesting question that that we we could address in the future. But I will stop at this and, and thank you very much, Andrew again for not just your comments now but for the collaboration that we have had over the years. Uh, yes, thank you, um, very much for the comments, um, and to reiterate what Norman said, we agree with, uh, almost all of them. Um, to add a couple of other, uh, more detailed some of one of Norman's comments, uh, on the level, um, of, uh, Of, uh, the effect of capital on, on the level of GDP we completely agree with that, and I think that some of the discussion, I, I guess depends on the country context. So, as Norman mentioned on the first slide, you know, a lot of the countries that sort of come to us have these very big ambitions, um, And so the only way they can achieve those ambitions is by growing fast for a long time, and um we, so we agree that it, you know, the, these investment investment might increase the, the level of growth, but in order to achieve those ambitions, which is sort of where the conversation starts, um, it needs to be, it needs to be higher growth for the long-term, um. Uh, so on the, the second point, uh, on, um, Dutch disease. Um, so yes, we agree that the model, um, misses some of those, um, some of those effects, and, and Dutch disease is a particularly tricky one, to include in this model because it doesn't have a demand side, um. You could have um some other things, you know, related like maybe a learning by doing or or other things, but the difficulty with a lot of those um those mechanisms is that they're very hard to calibrate. Um So, but, uh, the point is, is well taken, um, and I think we, we need to, we do need to be careful with how we uh discuss the findings of the model, uh, in a policy, um, in a policy context. And also thank you for mentioning the CEM 2.0 tool, we will, we will definitely have a look at that. Thanks, Norman and Steven. Um, Andrew, before coming back to you, I'm gonna open it up a little bit to collect some other questions. Um, we have two people online, and, and forgive me for, for how I'm gonna pronounce this. We have, uh, Zelo and then we have Angela. Let's start with Zelo. Do you wanna unmute yourself and ask your question? Yes, indeed. Can you hear me well? We can indeed. Uh, that's great. Thank you very much, uh, uh, Norman and Steven, uh, and also for, uh, uh, very important comments by Andy, um, and a great presentation. Um, um, I, I work on, on Egypt, by the way, uh, uh, Middle East, and, and, and, uh, we're in the middle of, uh, putting together the, uh, uh, actually preparing a, uh, a country economic memorandum which is actually focusing on long term growth issues. So this is very much. on our minds and we'd very much like to kind of look at the Egypt application and review it and revisit it in this. So I'd just like to add a comment on that in the case of Egypt I think it's particularly useful because Egypt is. One of those cases where you know indeed we do have we do see these kind of fundamental imbalances in the drivers of long term growth, you know, for example, the female labor force participation is about 15%, if that, right, extremely low, and it actually has been declining. Right? So, so you have the, uh, the LA, uh, you know, very much underutilized, you know, for long term growth in Egypt. Private investment is also a low, reflecting the kind of difficulties in the investment climate and the regulatory environment, you know, but also the kind of role of the state and the large footprint of the state owned enterprises that kind of also take up the space for private sector activity and private investments. So you know, is not really growing as fast as it should. You know. The private investments are not as high as they should be. Long term productivity growth also hasn't been high, so what's left is basically these very large public investments and ongoing, but also the future public investments plan. In Egypt that we see, and I think it's a very legitimate question is this sustainable? And I think the macro team in MTI thinks it's not, and these other drivers of growth must be kind of revisited and reactivated in one way or the other. How that is to be done, obviously, as Andy says, is easier said than done. But so I think that in the case of Egypt, in the case of these countries where you do have these very large and obvious imbalances between drivers of growth, I think the long term growth model can point out some of the important policy directions that that that can inform the policy dialogue. So I very much welcome that. I would quibble to give other people a chance as well, could you keep your comments. Just one more minute. Yeah, please, thank you. So that, that's, that's the, uh, uh, so that that was the, the main comment on the productivity was that, um, you know, I, I would quibble a little bit with the, with the notion that the literature is completely kind of unclear about the aggregate contribution of public investments to growth. You know, my own work with Johannes on South Africa data shows that if you actually use disaggregated data on infrastructure down to lower levels of infrastructure sectors, then you actually get much richer estimates and more robust effects than if you just rely on the very aggregate figures in cross country estimates. Thank you very much. Over. Thanks, and sorry for cutting off there. Um, just a sense, are there questions in the room? Just could you raise your hand if you have a question? OK, we have one question over here, 2 questions. OK, I'm gonna complete the questions. There's another question online by Angela. Angela, do you mind coming in and also just keeping it brief so that we can sort of do a full, uh, round of the table. Go ahead. Yeah, sure, uh, thanks so much. I hope you can hear me. It's, I'm in Jakarta and it's like 10:19, so I won't switch on my camera, sorry. So I'm Angela. I work on Indonesia and also currently working on a CEM for Fiji. I have two quick questions. One is, um, so I noticed that you, you released this LTGMTFP code extension, but it's a simple Excel sheet that sort of allows the user to put in a lot of different Uh proposed determinants of TFP I guess, education and innovation and lots of sort of uh variables there. Uh, but I, I was wondering if you can come in a little bit on, on how one can sort of feed that into the broader sort of, uh, uh baseline LTE. EGM, uh, uh, uh, a 2 because you can do that and then you sort of see where TFP is going, but it's, it's, it's not super clear how then you can take those numbers and plug it into the model so they can say, OK, if this happens to education, then this is what's gonna happen to growth. Um, when Andrew was speaking, I thought that would be Perhaps that was a some sort of a solution to sort of the user going into sort of deeper detail. What should you do to increase TFP and then when TFP increases to that, what would happen to growth. So if you can comment on those linkages, that would be great. The second question is, um, on, there's lots of other tools uh by MTI. That use a CP Douglas as well, identification that are more backward looking. So you're looking at how has TFP been growing in the past and how has, you know, uh, labor and human capital been growing in the past, um, and it would be nice to hear if you thought about how, how to link those uh sort of two stories so that, you know, you can use these tools for the Backward looking of how the economy has been progressing over time because those are also sort of long term looking but from, you know, sort of backward looking and then use this to sort of forward looking and if you thought about how to sort of take the results from, from those sort of sharply decomposition type of tools uh and link it to what the LTGM is able to give you. Uh let me stop here. Thanks. Thanks, Angela. I'm gonna come to Stti and then art, but I'm just gonna relate two questions we got from, from YouTube. One is asking whether the, the female labor force participation could be parameterized in the model, relating to another comment, and then another question related to, uh, can this model be applied to sectoral growth, for example, agriculture alone, and how would one think about that? Um, then there is, um, OK, let's go to Stutti and then art. St. Thank you, great presentation. Uh, my question is about to the extent to which both in the book and in your, uh, ongoing work and policy dialogue with governments, how do you make the link between clearly the questions you're getting are very policy oriented, uh, powerful leaders, uh, at the top, wanting to do something about growth. How do you link the policy dialogue. from this very sort of growth driver accounting exercise to the micro foundations, to what extent are public basic public economic insights brought in about the role of government versus markets and very simple policies, but the most fundamental of security of property rights, lower Of transaction costs, you know, one of the empirically quantifiable results that have come up in the macro literature on growth is how trust, you know, this very intangible concept of trust matters so empirically significantly for growth. So I'm just wondering in that policy dialogue along with your messages about Just relying on investment is not sustainable. Is there scope to think with more micro foundations public economics approach to public policy? Thank you. Thanks to the Art Uh, thanks, Dean. Thanks, Steven and Norman. I, you know, watched this take shape over quite a few years and it's great to see, uh, it's great to see the book out. Uh, three quick questions. The one is, um, have you thought about doing a sort of retrospective forecasting exercise? So imagine that you guys have designed this model in 1980 and ask how you do over the 20 and 40-year growth horizons. Um, so the projections that come out of the model. So I think that would really help to both. Strengthen the messages. In those dimensions where, you know, you, uh, where the model does well in, in sort of projecting the future, um, uh, and also would, would sort of point out to, to the caveats. Um, second one, building a little bit on the ques second question, building on the question that came online. Not just would you apply this model at the sector level, but have you thought about doing a sort of uh sectoral version of the model in which, you know, you can capture intersectoral factor reallocation its contribution to growth, which quantitatively is a, is a big deal. And the last question, which is not really unfair, it's a little unfair and maybe it's more a question for Andrew than for you guys, but why isn't this the only tool that we're using in CEMs for long-term growth questions, right? Cause there's a lot of proliferation of tools. What are the barriers, you know, what do we need to do to ensure that it's, it's used even more widely. Thanks, thanks everyone. So, uh, we have 5 minutes left, so don't feel like you have to answer everything extensively. If you could take sort of 3 minutes and then we can have 2 minutes, uh, with, with Andrew at the end, uh, over to you. All right. Thank you very much. Uh, on Serco, actually, the, I will just say that we have now, I am now uh in the director of the Global indicators Group. We do lots of work on the the business climate. And also on um women's economic empowerment and that will certainly have an effect on female labor force participation and the productivity of the, of the economy. So I would uh direct you to the work that we do in that regard in in the Global indicators group. And I think that we can have a, a, a close look at what you are doing in your CEM and see how the two, these, uh, two forces on female female labor force participation in business business climate can affect, uh, growth, uh, prospects for uh for Egypt. And I think, uh, your, the instincts that your, uh, colleagues in MTI saying about uh the effects, the limited effects of public investment are, are right. Angela, uh, we, we, we may want to, uh, uh, uh, to speak uh bilaterally on how you could, uh, use the, uh, the total factor productivity extension for your Indonesia work. I think it, it can, it can allow you to, to do it if you have data for what we call the, the proxies, uh, of, um, of the, of these, uh, five areas that affect productivity. Uh, well, clearly the female labor force can be parameterized and that's something that we do in the model. Uh, on the sectoral growth, actually that we had an extension that we are preparing on sectoral growth, uh, that has three different, uh, sectors in the economy, uh, and that we want to see, uh, in fact, that we would, there would be uh productivity gains of going from one sector to another. Uh, and, uh, we have not included it here because it's not, uh, ready, but it is in uh, in the pipeline of, uh, of work that that we intend to, to do. The complication with that is that we couldn't do it in Excel, actually, as that was our, uh, our goal, to always have very simple, uh, a very simple toolkits. We have to do it in, in MATLAB. Uh, Suti, finally, on, on, uh, policy dialogue, I think that this is something that we rely a lot on, uh, on our colleagues and our, uh, our colleagues in, uh, in operations to help us understand, uh, the main concerns of, of policy makers and then how we can transform those those concerns into things that we can actually handle, uh, in, in the model. Some of the things that you mentioned are in fact included in our productivity assessment. I OK, so I'll just add um some, some minor uh, comments, so, uh, the, the second uh Webex comment um. So on the TFP extension um feeds a path of, produces a path of TFP which can be fed into the main LTGM. So that's a, so they, they, they can work together, you just link one spreadsheet to another. Um, the relation to historical, uh, relation to some other tools, so I think they're quite complementary, um, and uh tools like the growth accounting tool, can be used to uh estimate historical TFP which can guide some of our future assumptions. Um, I'm happy to talk more bilaterally about that. For the uh the YouTube uh comment, female labor force participation is something you can, you have complete, Uh, flexibility in the tool about how you can calibrate, uh, future versions, future, uh, values of that, um, to compliment Norman's question, Norman's response to Soodi's question, um, so it's true we also, as Norman says, rely a lot on the country economists there, but I think the role of the LTGM is to provide sort of a first step of breaking down a big problem. So the government sort of has this big goal of growth and then we sort of don't even know where to start to discuss that and the LTGM helps to break this down into different growth drivers and then we can have those discussions on the more micro determinants. Um, uh, in terms of art's question about the retrospective, uh, forecasting exercise, so I'm hoping to do one of those just before I retire, um, as it's a long-term growth model. No, the serious, the serious question is, uh, so the career chapter actually does this. Um, so the career chapter, uh, calibrates the model at various points in time and looks at what growth would be if we applied those in the future, and, um, So it has um the the the result is that it, you know, it is, if you make good assumptions, you get good results. The challenge for Korea's case is that some of the growth drivers change over time. Um, So I think that's all for me. Thank you very much for your comments. Thanks. We are out of time, but Andrew, I want to give you just a chance to come in if you had a burning two-hander. Sure, just a, just a real quick uh thing because we're talking about the long-term growth model. I just want to emphasize that levels matter. Um, the reason we're interested in growth is because it increases the levels and that often what we're seeing in terms of sustained growth over the long run is in fact the result of a series of level changes occurring time and time and time after again, uh, either to TFP or to To labor inputs or investment rates and so just a plea that as we're working forward that we keep in the back of mind and as we're using the tool that what we're really after is increasing the level and we achieve that through growth as opposed to having growth as the objective itself. Thanks, Andrew. So, in, in closing, let me just thank you, Andrew, first again for your comments and, and, and reactions and again for stepping in at the last minute. Uh, I wanna thank all of you for participating online and in the room. Uh, I thought this was an incredibly rich discussion. I wanna thank Steven and Norman for a really good, clear presentation and for engaging on the, on the questions, and also to congratulate you on, I mean, it's not a capstone. I guess it's a milestone in the project of bringing together where things are now, and it's already clear that this is moving forward with additional extensions and I think we're all excited to see where this goes next. So congratulations on the road here and, and looking forward to engaging, uh, going forward. So please join me in, in thanking and congratulating them.

showAllTranscripts

no

duration

PT1H31M26S

scene7File

worldbank/LTGM-PolicyResearchTalk-March2023

scene7Domain

scene7FileAvs

worldbank/LTGM-PolicyResearchTalk-March2023-AVS

title

LTGM PolicyResearchTalk March2023

description

LTGM PolicyResearchTalk March2023

showTimestampAndTranscript

yes