col-xs-12
col-sm-12
col-md-12
col-lg-12
col-xs-12
col-sm-12
col-md-12
col-lg-12
videoType
dynamic-media
videoDmUrl
https://delivery-p136806-e1377785.adobeaemcloud.com/adobe/assets/urn:aaid:aem:98b64bed-6309-498f-92f8-dd2344a910ad/play?assetname=Trade-firm-econ-dev.mp4
keyFrameImage
timestamp

00:00 Good afternoon,

00:01 everyone.

00:01 Uh,

00:01 it's very,

00:02 uh,

00:02 uh,

00:03 good to kind of see you join the weekly seminar for the 2024 WDR

00:09 that focuses on economic growth in middle-income countries.

00:13 Uh,

00:13 I'm Sho McLal and with our

00:16 lead academic Ufuk Akcheed

00:18 and our TTL,

00:20 uh,

00:20 Joyce Abraham.

00:21 We are very

00:22 happy to welcome you to this seminar this afternoon

00:26 that has,

00:27 um,

00:28 Two of our background paper contributors,

00:31 Yong Hon Shim and uh Sina Attis,

00:35 who are going to tell us a little bit about an overview on the role of trade

00:41 and firms in economic development and they do a very nice

00:45 synthetic piece on that.

00:47 In addition,

00:48 Yung-hun and uh Sina will go deeper

00:52 into the South Korean experience.

00:55 And colleagues,

00:56 uh,

00:56 this is

00:57 the

00:57 first week of the next 3 weeks,

01:00 we'll be focusing on different

01:02 economic stories from South Korea,

01:04 so keep tuned on that.

01:06 So as,

01:07 uh,

01:07 words of introduction,

01:09 uh,

01:09 seen as a principal economist at the Board of Governors for the Federal Reserve.

01:14 And uh he does work,

01:16 a lot of work on looking at firm dynamics and looking at innovation,

01:21 competition

01:22 with uh a strong focus on quantitative analysis.

01:26 Jung-hen Shim uh is,

01:29 has just joined the IMF

01:31 uh research department,

01:33 but we'll still collaborate with him.

01:35 It's OK.

01:36 Uh,

01:36 so,

01:37 uh,

01:37 Jung-hen's been doing a lot of work

01:39 on macroeconomic growth and international trade.

01:43 And his research really looks at the role of technology adoption.

01:49 And innovation

01:50 over different stages of development.

01:52 And that's why Junghan your research is so central

01:56 to some of the ideas we are developing and putting forward in this year's WDR.

02:02 Uh,

02:03 Yong-han just completed his

02:04 PhD from the University of Chicago and Yuk was his

02:09 principal advisor.

02:10 So colleagues,

02:11 what I'll do is,

02:12 uh,

02:12 give you the floor.

02:13 You have about 40 minutes.

02:16 To

02:17 present

02:17 the key ideas from these uh background papers,

02:21 and then we'll open it up for a discussion.

02:24 Over to you.

02:28 Uh,

02:28 Xena,

02:28 let me show you the slide.

02:41 Uh,

02:42 so thank you very much.

02:43 Um,

02:44 I'm Sina Attesh from the board and happy

02:46 to talk about international trade and dyna uh,

02:48 from dynamism,

02:50 dynamics and economic development,

02:51 our joint work with Yonkun,

02:53 and the usual,

02:54 uh,

02:54 board and IMF disclaimers about,

02:56 uh,

02:57 about employees' work,

02:58 of course,

02:58 applies here.

03:00 So,

03:00 uh,

03:00 just briefly about the outline,

03:02 uh,

03:02 of today's talk.

03:04 So I'm gonna first start with,

03:05 um,

03:07 some,

03:07 uh,

03:07 key highlights from the literature on international,

03:10 like openness to trade or more broadly

03:12 international linkages and how they affect growth like the key channels.

03:16 Then I'm going to focus a little bit on the,

03:18 uh,

03:18 technological competition side like or foreign competition,

03:21 how it can foster.

03:23 Uh,

03:24 technology upgrading,

03:25 uh,

03:26 among domestic,

03:26 uh,

03:27 domestic firms and what type of policies can indeed,

03:30 uh,

03:31 support,

03:31 support this relationship.

03:33 And in light of this discussion,

03:34 Yongong will continue,

03:36 um,

03:37 with,

03:37 uh,

03:37 a refreshing look and like with no facts on the long-lasting development process

03:42 of South Korea

03:44 and

03:44 what type of industrial policies indeed supported,

03:47 uh,

03:47 supported this,

03:48 uh,

03:49 persistent,

03:49 uh,

03:50 development experience.

03:52 So with that,

03:54 let me

03:54 uh go into the key channels that are highlighted in the literature

03:59 on international trade and economic growth.

04:02 Uh,

04:02 there's a large body of work

04:04 that uses as a natural uh natural experiment,

04:08 the,

04:09 uh,

04:09 trade liberalization episodes in,

04:11 in Different countries.

04:13 Most of these

04:14 work,

04:14 uh,

04:15 relates to developing countries like the,

04:17 uh,

04:18 experiences of Latin American countries over 1991980s,

04:22 1990s,

04:24 and,

04:24 um,

04:25 the first channel that,

04:26 uh,

04:26 that is emphasized by this work is

04:29 The access to

04:31 better inputs,

04:31 so inputs that can be,

04:33 uh,

04:33 that can be imported from,

04:35 uh,

04:35 from abroad,

04:36 which can help,

04:37 uh,

04:38 firms improve their productivities.

04:40 Market size is a second channel,

04:42 uh,

04:43 in this,

04:43 in this literature,

04:45 so opportunity to access broader markets can,

04:48 uh,

04:48 induce firms indeed to improve their products and processes.

04:53 A third channel

04:55 is

04:56 the competition aspect

04:59 of trade liberalization on which I'm going to spend a little bit more time,

05:02 uh,

05:02 later,

05:02 so I'm skipping it right now.

05:04 But

05:05 a key highlight from,

05:07 uh,

05:08 from this body of work is indeed

05:10 that there's some heterogeneity,

05:12 uh,

05:12 in the effect of this liberalization episodes on,

05:15 on firm dynamics.

05:17 So firms

05:18 that are readier to begin with,

05:20 let's say that,

05:21 that have a high productivity or that are more and the intensive,

05:25 uh,

05:26 uh,

05:27 before,

05:27 before trade liberalization episodes

05:29 actually happen to gain more,

05:31 uh,

05:32 from,

05:32 uh,

05:32 from these episodes,

05:33 like,

05:34 uh,

05:34 an important heterogeneity that I'm gonna highlight later,

05:37 uh,

05:38 in my talk.

05:39 And another channel

05:41 on which there's vast and expans uh

05:44 expanding,

05:45 uh,

05:45 literature,

05:46 um.

05:48 Is the knowledge spillovers and technology transfers

05:52 across,

05:52 across countries.

05:54 Uh,

05:54 so

05:55 there's a body of work,

05:56 uh,

05:56 in this literature that,

05:58 uh,

05:58 compares and contrasts the importance of technological,

06:02 technology adoption versus,

06:03 uh,

06:04 standalone innovation

06:05 over different phases of,

06:07 uh,

06:08 economic development.

06:09 Indeed,

06:09 Yong Hong's work,

06:11 recent work,

06:11 uh,

06:12 sheds like,

06:13 uh,

06:14 Uh,

06:14 a new,

06:15 uh,

06:16 a fresh look,

06:17 uh,

06:17 on these dynamics

06:19 that he's gonna,

06:20 that he's gonna discuss,

06:21 uh,

06:21 in his part.

06:22 A related literature here is the,

06:24 uh,

06:25 cross-border flows of investment,

06:27 particularly foreign direct investment

06:29 and the associated,

06:30 uh,

06:31 technology transfers.

06:33 So a recent work here to emphasize,

06:35 I think is

06:36 uh by Chris Frans Rosen,

06:38 Shebnam Kalis Jan,

06:40 uh,

06:40 and,

06:40 and co-authors who highlight that not only the firms that

06:44 see FBI benefit from associated knowledge spillovers and technology transfers,

06:49 but also

06:50 other firms in,

06:51 in similar sectors operating

06:53 uh with similar technologies happen to,

06:56 uh,

06:56 benefit from these flows as well.

06:58 Um,

06:59 in,

06:59 in,

07:00 in recent and related work,

07:02 we also

07:03 highlight that

07:04 FDI flows can indeed

07:07 help

07:08 or alleviate negative implications of uh of

07:12 market concentration in a domestic economy.

07:14 We indeed document that.

07:15 In more concentrated sectors,

07:18 domestic firms,

07:19 uh,

07:20 benefit more from FDI flows in terms of their productivity growth.

07:24 And again,

07:24 not just the firms that receive those flows,

07:27 but

07:27 other firms in the same sector.

07:29 So

07:30 FDI flows and associated technology,

07:33 uh,

07:34 uh,

07:34 technology transfers

07:36 may,

07:36 uh,

07:37 help with business dynamisms in domestic,

07:39 um,

07:40 in the receiving or the,

07:41 in the host countries,

07:42 um,

07:43 uh.

07:44 Uh,

07:46 supporting business dynamism in more concentrated sectors.

07:49 So,

07:50 uh,

07:51 in another recent work,

07:52 actually,

07:53 we look at,

07:54 at,

07:54 at,

07:54 at the uh

07:56 similar relationship,

07:58 uh,

07:58 in cross-border flows that are going in the reverse direction that,

08:02 that are,

08:02 that is more,

08:03 uh,

08:03 or less explored in the literature.

08:06 So if you,

08:06 if you go to the next slide,

08:08 Yono.

08:09 So,

08:10 uh,

08:11 now we look

08:13 at

08:13 foreign firms' investments in

08:16 countries that are indeed at the technology frontier.

08:19 So we look at the transactions between foreign firms

08:22 that are investing in,

08:24 uh,

08:24 US startups that are,

08:25 uh,

08:26 at their,

08:26 at the frontier.

08:28 So,

08:29 um,

08:29 what we document here is that,

08:31 of course,

08:32 as might be expected,

08:33 firms that invest in those US startups,

08:35 increase patenting in,

08:37 in those technologies and also increase their citations

08:40 to the patent portfolio of those startups.

08:43 But again,

08:43 this is not just them.

08:45 Also,

08:46 other firms,

08:47 uh,

08:47 in the same sector with this investing foreign firm

08:51 increase their patenting in those technologies and again citations to the,

08:54 uh,

08:55 patent portfolio,

08:56 uh,

08:57 of the US,

08:58 US firms in that technology.

08:59 So

09:00 the knowledge spillovers that are coming from those transactions from these,

09:04 uh,

09:05 investments

09:05 abroad

09:06 actually is broader

09:07 and concern,

09:09 not just the firm that is investing.

09:11 And another interesting fact that we that we highlight

09:14 is that these effects are indeed larger and stronger

09:18 in sectors that rely more on basic technologies,

09:23 so technologies that require

09:25 a larger upfront investment in like in

09:29 in fundamental research in human capital,

09:33 and indeed

09:34 firms are apparently like

09:36 more laggard foreign firms invest in US dollars or

09:40 Firms that are at the technology frontier to

09:42 acquire knowledge and to build up capacity,

09:45 uh,

09:46 faster in such technologies and,

09:48 um,

09:50 WDI actually emphasizes a lot like the importance of human capital.

09:54 So apparently there might be complementarities

09:56 coming from these type of investments,

09:58 uh,

09:59 learning

09:59 and acquiring frontier technology

10:02 through such investments and,

10:03 uh,

10:04 which can

10:05 help improve also

10:07 human capital and associated technologies,

10:09 uh,

10:10 in the,

10:11 in the foreign country.

10:13 So with that,

10:14 uh,

10:15 let me now switch to the,

10:16 uh,

10:17 final,

10:18 uh,

10:19 feature that I want to highlight that is associated with,

10:21 uh,

10:22 openness to trade,

10:23 uh,

10:24 the competition aspect,

10:25 which is,

10:26 uh,

10:26 the more contentious aspect,

10:28 uh,

10:28 of openness to trade.

10:30 So in recent work again with Ufuk Akjeed and Jean-Marie Implotti,

10:34 we

10:35 uh tackle indeed this,

10:37 uh this perennial question of how to manage

10:41 uh

10:42 intensified competition that is exerted by foreign

10:45 firms and what policies can help them.

10:47 Domestic firms

10:48 in such competitive,

10:49 uh,

10:50 competitive rates.

10:51 So whether tariffs or protecting these firms from competition is the,

10:55 uh,

10:55 better way to go,

10:56 or

10:58 whether supporting these firms,

10:59 these domestic firms through R&D subsidies

11:01 is more welfare enhancing.

11:04 And we look at these policies also

11:06 considering differences between the short run and long run,

11:09 like differentiating between different horizons of,

11:12 uh,

11:12 of policymakers.

11:15 So,

11:16 a key,

11:17 uh,

11:18 part

11:18 of our analysis or of the framework that we use in this,

11:22 in this work

11:23 is that

11:23 while it is encompassing,

11:25 like many features or aspects of openness to trade that I was highlighting,

11:30 uh,

11:31 while

11:33 While capturing these aspects,

11:34 it also

11:36 looks or like captures explicitly

11:39 the heterogeneous effects that competition

11:42 has on domestic firms.

11:44 So there's this,

11:45 there's of course the very important concern that keeps

11:49 Rising in policy discussions

11:50 that foreign competition from like better,

11:54 like exhorted by better firms that are at the technology frontier

11:57 can drive domestic firms,

11:59 uh,

12:01 can cause domestic firms to lose their market share or even exit the markets.

12:06 And indeed,

12:07 our model captures these dynamics,

12:09 but there's also another side of competition,

12:11 which is

12:12 competitive pressures can incentivize

12:14 some domestic firms

12:16 to improve their products and processes just to be able

12:19 to compete and or overcome their uh foreign rivals.

12:23 But of course,

12:24 which dynamic will play out

12:26 depends on the,

12:27 uh,

12:28 on the difference between the rivals,

12:29 just like in a,

12:30 in a race of,

12:31 in a running race,

12:32 right?

12:33 If the two competitors are close to each other,

12:35 there's the hope or potential of

12:37 overcoming your competitors,

12:39 just exerting a little bit more effort.

12:41 But if the difference between the 22 rivals is

12:45 very wide,

12:46 the laggard,

12:46 uh,

12:47 the laggard runner or the laggard firm

12:50 loses the hope of overcoming its rival,

12:53 capturing markets,

12:54 so it gets discouraged and relaxes its innovation effort.

12:58 So the framework,

12:59 as you see,

13:01 Relates technological differences to the dynamics of competition,

13:05 and those affect the dynamic incentives,

13:07 forward-looking investment incentives of,

13:10 uh,

13:10 of,

13:10 uh,

13:11 both domestic and foreign firms.

13:14 And this type of heterogeneity

13:16 indeed speaks very much to the extensive empirical work

13:19 that highlights these

13:21 different dynamics.

13:23 So just to highlight,

13:24 uh,

13:25 in the next slide,

13:25 just to highlight

13:27 how

13:28 uh this relationship between competition and innovation effort,

13:31 uh,

13:31 transpires in the model

13:33 on,

13:33 in the left panel.

13:34 I show again from the calibrated model,

13:37 the innovation effort of a generic firm

13:40 in the model

13:41 as a function of its position

13:44 relative to its foreign rivals.

13:45 So basically as a function of the technology gap between the two,

13:48 between the two firms.

13:49 As you clearly see,

13:50 there are two spikes.

13:52 Let me briefly explain,

13:53 uh,

13:53 what they mean.

13:55 On the left,

13:57 This concerns a domestic firm that is slightly

13:59 inferior in technology than its foreign rival,

14:02 but there are trade costs,

14:03 so importing those better goods is costlier

14:08 due to these trade costs.

14:09 So as a result,

14:10 this

14:11 domestic firm that is slightly inferior still has

14:14 some,

14:16 some food in the game.

14:18 So it can compete with the,

14:20 with the foreign rival.

14:21 So when their technologies are close to each other,

14:23 whilst accounting for the extra costs of importing,

14:27 the domestic firm exerts more innovative effort just to

14:31 be able to defend its domestic market against,

14:33 against,

14:34 um,

14:35 this foreign rival.

14:36 What happens around the uh rights spike

14:39 is basically,

14:41 it is about the

14:42 domestic firms that are,

14:44 uh,

14:44 that have better technology than their foreign rivals,

14:47 but they can't,

14:48 it is not enough to export to the other

14:50 markets because they also need to overcome this higher,

14:52 these,

14:53 um,

14:54 these higher costs associated with trade,

14:56 be it like iceberg costs or higher tariffs,

15:00 like,

15:00 uh,

15:01 tariffs,

15:02 um,

15:04 put in place by the,

15:05 by the other country.

15:07 So,

15:07 but if the domestic firm is

15:10 closing competition,

15:12 so if it can improve its product,

15:14 product quality a little bit more

15:16 and overcome

15:17 its rival and the uh trade costs,

15:20 it can

15:21 gain access to export markets.

15:22 So it intensifies its innovation efforts.

15:24 So as you see,

15:25 there are two regions where

15:28 Technologies get closed and competition stiffens,

15:31 and domestic firms are incentivized to,

15:34 to exert more innovative effort.

15:36 And interestingly,

15:37 we see a similar pattern in the data on the right,

15:39 uh,

15:39 on the right panel

15:41 to show that if you create a measure of technology gaps in the data

15:44 and then look

15:45 like the uh

15:48 Uh,

15:48 the,

15:49 uh,

15:50 innovate innovation in uh,

15:52 intensity of domestic firms,

15:55 in this case,

15:55 the US firms,

15:56 you see a similar pattern of,

15:58 uh,

15:59 double peaks like where uh innovation in,

16:01 uh,

16:02 intensity actually uh intensifies.

16:05 Uh,

16:06 so this is a reassuring and validating,

16:09 uh,

16:09 evidence for the mechanism that we are talking about here in this one.

16:13 So just to

16:15 Recap the policy implications of this model.

16:20 importantly,

16:20 there are two sides to policy.

16:22 Like there,

16:22 there are,

16:23 first,

16:23 the aesthetic effects and the dynamics,

16:25 dynamic effects.

16:26 So,

16:27 uh,

16:27 for instance,

16:28 consider

16:28 a unilateral increase,

16:30 uh,

16:31 like protectionist policies,

16:32 a unilateral increase in trade barriers on imports.

16:37 So in that case,

16:38 Domestic firms,

16:40 or dome domestic country can indeed

16:44 Potentially benefit from such policies,

16:47 but,

16:47 uh,

16:47 shifting production and profits

16:50 to home.

16:51 Like,

16:51 so protecting firms or increasing barriers

16:54 against imports,

16:55 uh,

16:56 means import becomes more costly

16:58 and production shifts to domestic firms.

17:00 But there are dynamic consequences.

17:02 Such policies can hurt forward-looking incentives of,

17:05 of such firms,

17:06 as I was showing.

17:07 If barriers are higher against imports,

17:10 so domestic firms are just shielded from competitive pressures,

17:13 and they don't need to exert

17:16 as much effort in improving their product qualities or processes,

17:20 so

17:20 they,

17:21 they,

17:21 they exert less effort to improve technologies.

17:24 So as a result,

17:25 what we find is

17:26 Profit shifting motives are actually

17:29 dominant

17:30 in terms of welfare only

17:32 when a few years ahead are considered.

17:35 So if the policy horizon is longer than that,

17:38 just beyond a few years,

17:40 or maybe unfortunately like longer than a,

17:44 longer than a presidential term.

17:46 Then the dominant,

17:48 uh,

17:49 effect is coming from the negative dynamic consequences

17:52 because firms are not exerting enough effort

17:55 as they don't feel much competitive pressure.

17:58 Indeed,

17:58 what we find is the optimal trade policy is to slash tariffs to zero

18:03 if the policy horizon in consideration is just beyond

18:06 a few,

18:07 uh,

18:07 a couple of,

18:08 a couple of years.

18:09 And the optimal subsidy,

18:12 sorry,

18:12 optimal policy

18:13 for a policymaker that is concerned over that

18:17 is concerned about the medium to longer term.

18:19 So a policymaker

18:22 who is non myopic,

18:23 the optimal policy is

18:25 using R&D subsidies.

18:26 So basically,

18:28 Supporting firms in the competitive rate instead of

18:31 killing competition and the associated incentives,

18:34 uh,

18:35 with higher barriers to trade.

18:36 And we also document

18:38 a policy complementarity.

18:39 So in a world that is more globalized,

18:43 defined by lower bilateral tariff rates,

18:45 there's less need to intervene via R&D subsidies just because markets.

18:49 And competitive pressures take care of,

18:52 uh,

18:52 appropriate incentives just by themselves,

18:54 uh,

18:55 forcing firms to,

18:56 to increase their innovative efforts.

18:59 But let me conclude on my last slide just with some policy.

19:04 Implications for developing countries

19:07 that are,

19:07 uh,

19:08 that are facing,

19:09 uh,

19:09 some challenges in their,

19:11 uh,

19:11 in their middle income base.

19:14 So of course,

19:15 every,

19:15 every economic economic analysis and model abstracts

19:18 from some considerations and of course,

19:19 there are valid considerations uh surrounding such policies like,

19:23 uh,

19:24 uh,

19:25 removing trade barriers,

19:26 for instance,

19:27 just outright

19:28 might entail other considerations like they,

19:31 they might generate great dislocations.

19:33 There might not the transition of the firms to innovate,

19:37 uh,

19:37 activities might not be as smooth.

19:40 But I think the importance of our policy implications are not just

19:43 the

19:44 exact uh policies per se,

19:46 but more about

19:48 what the goal or the uh direction of policies should be.

19:53 What we highlight is the protectionism.

19:56 Basically distorts

19:57 forward looking incentives of firms

19:59 and indeed

20:00 if

20:01 if a country after years of protectionist

20:03 policies wants to decrease these barriers,

20:06 it might find itself in a

20:08 worse position than it would be

20:10 in an alternative or counterfactual economy where

20:14 trade barriers were lower to begin with just because

20:17 firms

20:18 were shielded from competition and they were not

20:21 exerting enough effort in technology upgrading.

20:24 And indeed,

20:26 indeed history is bright with these examples,

20:28 right?

20:29 Like uh examples of import competition that Latin American countries or,

20:33 um,

20:33 like Turkey

20:35 over 1960s and 1970s,

20:37 leftist countries,

20:38 uh,

20:39 with firms that are dependent on state support and not ready to compete,

20:43 uh,

20:43 with foreign rivals once the,

20:45 uh,

20:46 uh,

20:46 trade barriers were,

20:47 uh,

20:48 were,

20:48 uh,

20:49 removed.

20:50 So as a result,

20:51 what we highlight is that the goal needs to be

20:54 to support firms to build up capacity to compete,

20:57 to upgrade their technologies and

21:00 to be ready to innovate in the future.

21:02 I think in this sense,

21:04 Korea's example is very telling that they have been steadily.

21:08 Increasing their,

21:09 uh,

21:09 their,

21:10 their import tariffs over the development stage,

21:13 while at the same time supporting their firms

21:15 to invest in technologies,

21:17 uh,

21:18 both via imitation and also innovation policies,

21:21 as Yongoon's recent work,

21:23 uh,

21:23 high,

21:24 highlights,

21:24 uh,

21:25 very clearly.

21:26 And with that,

21:27 let me,

21:27 uh,

21:28 leave the stage to Yong Hoon,

21:30 uh,

21:31 for,

21:31 uh,

21:31 for his talk on South Korean experience.

21:34 Thank you very much.

21:37 OK.

21:38 Thanks,

21:38 uh,

21:39 thanks for having me.

21:40 So I'm gonna present the second half of the presentation,

21:43 and I would like to zoom into a specific case,

21:45 which is South Korea.

21:47 So here,

21:48 we plot the GDP per capita of South Korea in the red line

21:52 with other different countries.

21:54 So here we can see that at the beginning in the 1950s,

21:57 Korea started as one of the lowest income countries

22:00 like China,

22:01 India,

22:02 or Vietnam.

22:03 And in the 1980s,

22:04 it caught up with some middle income countries like Brazil,

22:08 Mexico,

22:09 and Turkey.

22:10 But even after that,

22:11 Korea kept growing,

22:13 and now it is in the process of catching up with some of the high-income countries

22:17 like Japan,

22:18 France,

22:18 or the US.

22:20 And behind this development,

22:23 there are,

22:23 there were several industrial policies

22:26 during that development period.

22:28 So the first industrial policy was sector-specific policy.

22:32 Where the government was trying to support the heavy manufacturing sector.

22:36 So there are several papers documenting this policy,

22:39 and it turns out that this policy was effective

22:42 at transforming the country from

22:44 non-manufacturing

22:46 or light manufacturing intensive economy

22:49 to more heavy manufacturing intensive economy,

22:52 which is more export-oriented.

22:55 And also at the same time,

22:57 there was a trade policy

22:58 where as Sina mentioned briefly,

23:01 the government was reducing the import tariff

23:04 while the country was catching up with the frontier countries.

23:08 And the 3rd industrial policy was

23:10 somewhat implicit in a sense that

23:13 they were implicitly

23:14 picking the winners in a sense that

23:16 they look for the firms

23:18 and they

23:19 kind of point out the successful ones at the beginning

23:22 and then they keep subsidizing these large incumbents conditional on that

23:27 they're

23:28 exporting their goods to other countries.

23:31 So we have recent work,

23:33 uh,

23:33 documenting these patterns by looking at the

23:35 capital and labor wedge of these superstar firms

23:38 and study the impact of these,

23:40 uh,

23:41 uh,

23:41 policies on the increased,

23:43 uh,

23:44 market concentration over time.

23:46 And the last one is the technology policy,

23:49 where the government

23:50 first

23:51 started by subsidizing technology adoption,

23:54 and then they shifted toward innovation subsidy.

23:57 And this will be my focus on this today's presentation.

24:02 So here,

24:03 we

24:04 plot the technological growth of South Korea

24:07 with two different measures.

24:09 First one in the red line is the number of patents

24:12 per 1000 people.

24:14 So we can see at the beginning in the 1980s,

24:17 the number of patents per capita is almost zero,

24:20 but it rapidly grew over time,

24:23 and in 2020,

24:24 it is almost like 3 per 1000 people,

24:27 which is the uh one of the highest numbers in the world.

24:30 And the second measure is R&D intensity,

24:32 which is R&D expenditure over GDP in the blue line.

24:36 And we can see that

24:38 again in the 1980s,

24:40 R&D intensity is almost zero,

24:42 but it grew rapidly and now it is like 4%,

24:46 which is,

24:47 which is the second highest number

24:49 among the OECD countries.

24:52 So our question is,

24:53 how this country transformed from non-innovative country to innovative country.

24:59 And to see it more clearly,

25:01 let me zoom into a specific case of the firm,

25:04 which is Samsung Electronics.

25:07 So in 1938,

25:09 Samsung Electronics was founded.

25:12 Although Samsung Electronics is now nowadays one

25:14 of the largest companies in the world,

25:17 but at the beginning it started just as a small company

25:20 that made a noodle.

25:22 In 1969,

25:23 they decided to enter an electronics industry,

25:26 but of course,

25:27 they didn't have any technologies to develop any kinds of electronics products.

25:31 So to tackle this problem,

25:33 they decided to adopt

25:35 some technologies from foreign companies,

25:38 and they managed to sign a first technology transfer contract

25:42 with a Japanese company called Sanyo.

25:45 So under this contract,

25:46 Samsung pays an adoption fee to the Japanese company

25:50 and Sanos uh agrees to share the blueprints of the technologies

25:55 and provide know-how transfer in a sense that

25:58 they invite some of the Korean engineers to their factories

26:02 and teach them how to use certain types of

26:05 types of machines

26:06 or

26:07 teach them how to produce a black and white TV basically.

26:11 And this was a very successful adoption,

26:13 and Samsung Electronics was quite aggressive in adopting other technologies

26:19 um uh from the frontier companies.

26:22 But in 1980s,

26:23 it becomes harder and harder to adopt these technologies

26:27 for two reasons.

26:28 First,

26:29 uh,

26:29 these Japanese companies become more reluctant to share the technologies.

26:33 Why?

26:34 Because Samsung,

26:34 as Samsung becomes a,

26:36 a large company

26:37 and becomes a larger competitor to these Japanese companies,

26:41 these Japanese companies

26:43 don't want to share this technology with Samsung.

26:46 So sometimes they refuse to share the technology

26:49 or basically raise the price of the technology.

26:52 And the second reason is that

26:54 as Samsung grew,

26:56 there are not many things to learn from the Japanese companies.

27:00 Of course,

27:00 Japan was producing better TV

27:03 than Korean company,

27:05 but still they are in the same bandwidth,

27:07 so there are not,

27:08 there are not many things to learn from these Japanese companies.

27:12 So in,

27:13 at the end of like the 1980s,

27:15 Samsung decided to innovate their own technologies

27:19 by,

27:20 by investing a lot of money in R&D

27:22 and hiring some scientists from the university.

27:26 And then they managed to develop their own technologies

27:29 and their own products.

27:32 At the same time period,

27:33 the government was quite

27:35 active.

27:36 In subsidizing this technology investment.

27:40 In particular,

27:41 they started by subsidizing adoption

27:43 and then shifted toward innovation subsidy.

27:46 So here in the left-hand side,

27:47 we plot the adoption and innovation subsidy rate over time,

27:51 which is tax credit for the adoption fee

27:54 or R&D expenditure.

27:56 We can see that it started by subsidizing adoption at 30%,

27:59 but it gradually decreased over time,

28:02 while the innovation subsidies started a little bit later than that,

28:05 but it gradually increased over time.

28:09 So when we look at the aggregate data,

28:11 which is adoption expenditure shared,

28:13 adoption expenditure over adoption,

28:15 and R&D expenditure,

28:17 we can see that the firms

28:19 were transforming from adoption

28:21 toward innovation

28:23 as the country develops over time.

28:28 So,

28:28 uh,

28:29 let me introduce the data what we,

28:30 uh,

28:31 that we are using in that paper

28:33 and then present two empirical facts out of it.

28:36 So the data,

28:36 the,

28:37 the main data we are using is the technology transfer contracts between Korean

28:41 and foreign firms

28:42 from 1962 to 1993.

28:46 So here,

28:46 we went to the National Archives and we digitized

28:49 the official documents between the Korean and the foreign firms

28:54 where they are uh where they are um

28:57 uh trading the technologies.

29:00 So here we can capture the universe of these contracts because at that time period,

29:05 firms have to report all of their transactions involving foreign currencies.

29:09 And we can see 50% more than the majority

29:12 of the adoption contracts were with Japanese firms,

29:15 and the remaining 25% were with US firms,

29:18 and most of the contracts were in the manufacturing sector.

29:23 Using this data,

29:24 we merge this data with the firm balance sheet data so that we can see the

29:27 basic information of the firms and also the

29:30 patent data so that we can measure the innovation

29:33 at the firm level.

29:35 So here,

29:36 I'm gonna present two empirical facts.

29:38 First,

29:39 when the initial productivity gap between the

29:42 technology buyer and the seller is larger,

29:45 then the productivity growth after adoption

29:48 is larger.

29:50 So here in the y axis,

29:52 we plot the productivity growth after 5 years from adoption,

29:56 which is measured by sales for employment growth at the firm level.

30:01 And in the x axis,

30:02 we plot the productivity gap between these two firms,

30:05 which is measured by the ratio of sales per employment of the Korean firm

30:09 divided by the sales per employment of the foreign firm.

30:13 So we can see that

30:14 uh when the Korean firm has a much less productivity

30:18 than the foreign firm,

30:20 which means that they are lagging behind the foreign firms.

30:23 The productivity growth after adoption is larger.

30:27 However,

30:28 as the Korean firms are catching up with the foreign firms,

30:31 so that they have more or less similar productivity,

30:34 then the productivity growth after adoption

30:37 becomes much smaller.

30:40 On the other hand,

30:40 when we plot this same exercise

30:43 with the innovating firms,

30:45 we can see that the innovating,

30:46 uh,

30:47 the productivity growth after innovation

30:49 is basically flat

30:51 over this initial productivity gap.

30:54 Which means that,

30:55 uh.

30:56 The productivity growth after adoption

30:59 depends a lot on the initial productivity gap

31:02 compared with the innovation.

31:04 And also on top of that,

31:05 when the when the firms are lagging behind,

31:08 then the adoption brings larger productivity growth

31:12 than the innovation.

31:14 But on the other hand,

31:15 when the firm is catching up with the foreign firms,

31:18 innovation brings larger productivity gain than the adoption.

31:24 The second fact that I want to deliver here is that

31:27 non-addopters also get knowledge diffusion from these adaptive technologies.

31:33 So to see that,

31:34 let's suppose a simple situation where Samsung adopted technology from SAO.

31:40 Which is a Japanese firm,

31:41 but not from Panasonic.

31:44 And our question is,

31:46 do other firms

31:47 that did not directly adopt this technology

31:50 are also getting knowledge diffusion from this adaptive technology.

31:55 To see that,

31:56 we follow the innovation literature and we assume that

32:00 if patent A makes a citation to patent B,

32:03 it means that patent A is learning something from patent B

32:07 and is building on top of patent B.

32:10 And if that's the case,

32:12 then these other firms

32:13 should increase the patent citation

32:16 towards Sanu.

32:18 Compare with the Panasonic firm,

32:20 a Panasonic pattern,

32:21 which is a plus b firm.

32:23 So the basic idea is to plot the difference of citation between the seller

32:28 and the plausible firm

32:30 around the first technology adoption.

32:33 So this is the result,

32:35 and we can see that

32:36 after several years from the first technology adoption,

32:39 we can see that the other firms

32:41 that did not directly adopt this technology

32:45 also increased patent citation

32:47 to this adaptive technology,

32:49 which means that

32:51 when other firms are adopting the technology,

32:54 I can also learn something from this adapted technology,

32:57 which implies that there is some

32:59 knowledge spillover from these adoption activities.

33:03 So those are basically two facts,

33:05 and let me explain the takeaways.

33:08 So the technology adoption can bring larger productivity gain

33:12 when the gap between the uh the firm and the frontier firm is larger.

33:17 And adoption also brings knowledge diffusion,

33:21 which can potentially cause underinvestment problem of the firms because

33:25 these are private firms do not internalize this knowledge diffusion.

33:30 So subsidizing adoption

33:32 can be very effective at the early stages of development.

33:36 However,

33:37 as we saw in the uh empirical fact,

33:40 the gain from adoption will diminish

33:42 as the country catches up

33:44 with the frontier countries.

33:46 So we have this basic idea and we quantify the policy implications

33:51 using two contrite growth models with endogenous adoption

33:54 and innovation decisions.

33:58 So here we

33:59 came back to the uh

34:01 the adoption and innovation subsidy of South Korea,

34:04 and then we put it

34:05 in the model

34:06 and then we're trying to compare the actual policy with no subsidies.

34:11 Or what if we just subsidize adoption and never switch to innovation?

34:16 Or lastly,

34:16 what if we just started by subsidizing innovation

34:20 and never subsidize adoption?

34:22 So basically,

34:23 we do some kind of simulation using the model.

34:27 And then here's the result.

34:29 So here's the case when we subsidize only adoption

34:33 for the whole period

34:34 with the same uh subsidy rate.

34:37 And this is the,

34:38 in the right-hand side,

34:40 this is the GDP relative to the no subsidies case.

34:44 So here we can see at the beginning,

34:46 GDP is

34:47 a little bit smaller than the no subsidies case.

34:50 Because

34:51 we are using uh more labor

34:54 into adoption activities.

34:56 Other than,

34:57 uh,

34:58 uh,

34:58 instead of production.

35:01 But

35:02 in the beginning,

35:03 it has a much higher growth rate.

35:05 So GDP already caught up with the no subsidy case

35:09 and it is much larger than the no subsidy case.

35:13 More importantly

35:15 This deviation from the no subsidy case

35:18 does not grow exponentially,

35:20 which implies that

35:22 subsidizing only adoption

35:24 does not increase long run growth rate.

35:27 This is because,

35:28 first,

35:28 the gain from,

35:29 uh,

35:30 productivity,

35:31 the gain from adoption

35:32 will diminish

35:35 as the,

35:35 as the,

35:36 as the Korean firms are catching up with the Japanese firms.

35:40 And also subsidizing adoption

35:42 will,

35:43 will,

35:44 will make the Korean firms stuck at the adoption phase

35:47 in a sense that

35:48 since the adoption is cheaper than the market price.

35:51 Even if they can

35:53 fully catch up the Japanese firms

35:54 and leapfrog the Japanese firms by innovating their own technologies,

35:59 they're just keep waiting

36:01 for the Japanese firms to innovate,

36:03 and they are just adapting and following the Japanese firms' technology

36:07 whenever the Japanese firms innovate their own thing.

36:10 So it it it actually.

36:13 Can reduce the long-run growth rate.

36:16 On the other hand,

36:17 when we subsidize only R&D or innovation,

36:20 at the beginning,

36:21 it has a much smaller GDP than the adoption subsidy only case,

36:26 because at the beginning,

36:27 adoption is much more effective than the innovation,

36:30 but subsidizing innovation instead adoption

36:32 can be distorted.

36:34 But in the end,

36:35 it has a higher long run growth rate.

36:38 So in the end,

36:38 after like 40 years,

36:40 it can uh catch up with the adoption subsidy case.

36:44 Now we

36:46 plug in the actual policy.

36:49 Since the actual policy gradually shifted

36:51 toward innovation subsidy from adoption subsidy.

36:55 At the beginning,

36:55 it closely followed the adoption subsidy case,

36:58 and as the

36:59 government switched to innovation subsidy,

37:02 the long run growth rate is also higher than the adoption subsidy case.

37:07 So in the end,

37:08 when we calculate the welfare compared with the no subsidy case,

37:11 we can see that the actual policy was the

37:13 most effective compared with the two time variant policies.

37:20 So I think uh I can conclude here.

37:22 So the,

37:22 uh,

37:23 the main message we want that we want to deliver here is that the policies

37:27 that were effective for developed countries

37:30 may not

37:31 work really well for developing countries.

37:33 And this state-dependent policy that started with

37:36 adoption subsidy and shifted toward innovation subsidy

37:40 was indeed very effective

37:42 in South Korea's development.

37:44 Thank you very much.

37:48 Thank you very much,

37:48 Yanha and uh Xina,

37:50 not only for the very uh useful presentation,

37:54 but sticking to time.

37:56 folk,

37:56 if you're connected,

37:57 would you like to provide some broader context before we open up to comments

38:02 or questions?

38:03 Mhm.

38:04 Uh,

38:04 uh,

38:04 absolutely.

38:05 So,

38:05 uh,

38:06 thanks,

38:06 thanks,

38:07 Jung and thanks,

38:07 Xina,

38:08 for,

38:08 for,

38:08 for this,

38:09 uh,

38:10 very nice presentation.

38:12 Uh,

38:13 so I,

38:14 you know,

38:14 the,

38:14 the main,

38:15 the main takeaways that I,

38:17 uh,

38:17 that I see in,

38:18 in,

38:19 in their presentation is that,

38:21 uh,

38:22 first,

38:23 the,

38:24 the optimal growth strategy is clearly,

38:26 uh,

38:27 stage dependent.

38:28 And,

38:29 uh,

38:30 so,

38:30 as a result,

38:31 you know,

38:31 not only,

38:32 not only the,

38:33 the firms have to be dynamic along the development stages.

38:37 But also the policies,

38:38 the optimal policies,

38:40 the policymakers have to be dynamic as well.

38:42 And I think,

38:43 you know,

38:44 uh,

38:44 here,

38:45 uh,

38:45 Yong Hon is presenting obviously the,

38:47 the optimal policies,

38:49 uh,

38:49 along the way,

38:50 but there's also a major political economy problem,

38:53 you know,

38:53 changing the policies

38:55 is,

38:56 uh,

38:56 is,

38:56 is,

38:57 is,

38:57 is not easy,

38:58 uh,

38:58 uh,

38:59 and,

39:00 you know,

39:00 uh,

39:01 I think that's,

39:01 that's an important message.

39:02 The second one.

39:04 Uh,

39:04 uh,

39:05 is that

39:06 what is good for the society is not necessarily good for,

39:10 uh,

39:10 policymakers.

39:11 I think that's also a very important message that's coming out in the sense that,

39:15 uh,

39:15 if policymakers' horizon is,

39:17 is,

39:17 you'd say,

39:18 5 years,

39:19 which is the short run,

39:20 or at most 8 years,

39:22 in that case,

39:23 protectionist policies can be uh

39:26 uh optimal from a policymaker's point of view.

39:29 But if you care about the longer run of the economy,

39:33 Uh,

39:34 of course,

39:34 uh,

39:35 boosting the technology is much more important,

39:37 and to boost the technology,

39:39 international spillovers and the competition channel,

39:42 uh,

39:42 coming from

39:43 international trade is extremely critical.

39:45 So that's again,

39:47 I,

39:47 I just kept hearing tensions,

39:49 political economy in the background from this presentation,

39:52 which was very useful.

39:54 Uh,

39:55 um,

39:56 yeah,

39:56 so let me stop here and I'm,

39:57 I'm,

39:57 I'm curious about the,

39:59 uh,

39:59 the audience,

40:00 uh,

40:00 feedback to you.

40:01 Thanks for you raise important questions

40:04 that in the current environment where trade

40:07 restrictions,

40:08 at least in technology access to certain countries are being limited,

40:12 and what happens to their growth prospects.

40:14 OK,

40:15 colleagues,

40:15 uh,

40:15 uh,

40:16 please,

40:16 uh,

40:16 uh,

40:17 raise your hand and I'll turn to you,

40:18 uh,

40:19 if you have questions.

40:26 Everyone's very,

40:27 uh,

40:27 reserved,

40:28 I think,

40:28 today,

40:29 right?

40:29 So,

40:30 uh.

40:37 Questions,

40:37 uh,

40:38 I don't see folks with questions,

40:40 so,

40:40 uh.

40:44 Junghon,

40:44 I think one of the key issues going back to your simulation

40:48 is that you're saying this switch from

40:51 imitation subsidy to innovation subsidy was timed pretty well in Korea.

40:56 So what was it really that the Korean

40:58 government was doing in terms of gathering information,

41:02 able to,

41:03 to be able to,

41:04 you know,

41:05 figure this

41:06 switch out.

41:07 And second,

41:08 uh,

41:09 I think a lot of the debate.

41:11 If you look at

41:13 the Fabrizio Zindiboti's papers,

41:17 is that earlier on when you're imitating

41:19 a lot of bad practices and conglomerates can get you there.

41:23 But then when you're innovating,

41:25 you need to sort of

41:27 really get your competitive environment really well.

41:29 So how did,

41:30 what were the things happening in Korea when the switches were going on?

41:34 I see.

41:35 So,

41:36 uh,

41:36 basically,

41:38 so when there's,

41:39 there's,

41:39 there's like a big change over this time and also like this time.

41:44 That was quite

41:46 closer to the president's change.

41:49 So

41:50 as Wuu already mentioned,

41:51 there was like political constraint in a sense that

41:54 uh

41:55 uh during the incumbent presidency,

41:58 it is kind of harder to change the policy,

42:00 but after there's a new,

42:02 new president,

42:02 then it is relatively easier

42:04 to change the policy,

42:06 right?

42:06 And I think the main information that they were using is just the

42:10 GDP or like a very aggregate measure in terms of

42:13 like uh compare the GDP with the Japan or the US

42:17 and then try to figure out,

42:19 oh now we are,

42:20 we have to move on to innovation phase or something like that.

42:24 And

42:26 And I think

42:27 the second

42:29 question was,

42:30 uh,

42:30 OK,

42:31 so it's,

42:31 uh,

42:31 about the Zillipoti paper.

42:33 So there was also like an interesting paper,

42:36 another interesting paper

42:37 that talks about uh Asian financial crisis,

42:40 uh,

42:40 in 1997.

42:42 So after that,

42:43 there was like a reform,

42:44 like a lot of reform from,

42:46 uh,

42:46 from the fund,

42:47 and

42:49 When you see the entry rate or entry rate

42:51 in the patent data,

42:53 the entry rate went up after the financial crisis,

42:57 which

42:58 I believe that it was uh helping the dynamism of the economy was

43:03 going in the right direction.

43:05 So it was a crisis,

43:06 so it was a bad thing basically,

43:08 but it has some upside

43:10 from that crisis.

43:14 So,

43:15 so then if you look at the firms right now,

43:17 besides

43:18 your example,

43:19 which is very detailed on Samsung,

43:21 what is the general

43:23 marketplace for coming up with new ideas and innovations?

43:27 Is it largely

43:29 in a model where the largest firms

43:32 have R&D capabilities and technical capabilities,

43:35 or do you also see in the Korean market that a lot of change is being driven by these,

43:40 you know,

43:41 new entrants who are growing rapidly?

43:44 I see,

43:45 but still,

43:45 I mean,

43:46 even after this crisis,

43:47 still,

43:47 uh,

43:48 from the data,

43:49 what I see is the leading firm in this innovation is incumbent and large firms.

43:54 So for example,

43:55 like Samsung,

43:55 LG,

43:56 like Hyundai,

43:57 like,

43:57 uh,

43:57 are driving this innovation.

43:59 But the context is a little bit different from the US,

44:02 uh,

44:03 uh,

44:03 example in a sense that

44:05 they are all exporting firms,

44:06 so they are competing in the global market.

44:09 So even if they are concentrated in a domestic market,

44:12 they are not

44:13 dominating in the global market.

44:14 So in that sense,

44:15 it is kind of,

44:16 it can still be innovative

44:18 in that sense.

44:27 Colleagues,

44:27 any

44:28 questions from uh

44:31 Well,

44:31 you wanna come in with your question?

44:34 So far.

44:36 Just a,

44:37 just a quick um question slash clarification.

44:41 On,

44:42 on,

44:42 on this slide,

44:43 your,

44:44 your,

44:44 your model,

44:45 uh,

44:46 you know,

44:46 uh,

44:46 you have a small caption there saying government

44:49 spending is financed by a lump sum tax.

44:51 So your model is assuming no dead weight loss from taxation,

44:56 correct?

44:57 So

44:58 so

44:58 my,

44:59 my,

44:59 my question to you,

45:00 if you can clarify that,

45:01 but also reflect in the policy world,

45:04 when,

45:04 when we look at this,

45:05 um,

45:06 sort of.

45:07 Relative to no subsidy scenario,

45:09 these subsidies are being financed by distortive taxes

45:13 in,

45:13 in the real world.

45:14 So

45:14 the results will all,

45:16 you know,

45:17 at least,

45:18 uh,

45:18 you know,

45:18 will be lower than,

45:19 than,

45:20 than what you have in your chart

45:22 or,

45:22 or could be even different,

45:23 right?

45:24 Yeah,

45:24 I think that's a very good point.

45:26 So here to simplify the things,

45:28 we assume that they are collecting tax by lump sum tax.

45:31 But we have this additional parameter

45:33 that governs the inefficiency of uh imposing lump sum tax.

45:37 So now parameter is,

45:39 uh,

45:40 is 2,

45:41 from the literature,

45:42 which means that we have to uh collect $2

45:46 to use $1 for this,

45:48 uh,

45:48 this subsidy program.

45:50 So it is actually even larger

45:53 effect.

45:54 If we assume that there's no inefficient,

45:56 inefficiency,

45:57 uh,

45:58 at all

45:59 from the lump sum tax,

46:00 as you mentioned,

46:01 but we are adjusting the parameter to be more realistic.

46:07 Can I just add one more,

46:09 uh,

46:10 comment here,

46:10 uh,

46:11 to Yong-hun's answer.

46:13 So Jaffar,

46:14 you're,

46:14 you're,

46:14 you're

46:16 bringing an excellent point.

46:17 Uh,

46:18 this class of models that Yong-hun is building,

46:21 if you make

46:22 the policy

46:23 stage dependent,

46:25 the

46:27 That the model will

46:28 ask

46:29 you to have a lower tax on larger firms,

46:34 and this is because

46:35 to incentivize firms to be large,

46:38 you are promising some additional return if

46:42 you want to make the taxes distortionary.

46:43 You can,

46:46 in some sense put some backload of carrots

46:48 so that you can also use the tax policy

46:51 to create an additional incentive for firms to grow.

46:54 Right,

46:54 so,

46:55 uh,

46:55 first through the subsidy policy,

46:57 you're subsidizing imitation,

46:58 so that's good for the country,

46:59 but if you want to also exploit variation in tax policy,

47:03 these models will tell you that

47:05 lower the taxes as firms are growing,

47:07 then there will be a trickle down of incentives.

47:10 So if being large is,

47:11 is good,

47:12 then all the smaller firms will also try to be uh good.

47:15 So

47:16 these models,

47:16 these forward-looking uh uh competition models,

47:19 uh,

47:20 have that flavor.

47:21 So that's something that I wanted to clarify.

47:24 Uh,

47:25 one additional thing,

47:26 uh,

47:26 uh,

47:26 that I would like to also highlight is that,

47:29 uh,

47:29 in,

47:29 in Sinna's remark,

47:31 so he,

47:32 he made a point that's very,

47:33 I think,

47:34 fundamental for the WDR which is,

47:37 uh,

47:38 in,

47:38 in,

47:38 in,

47:38 in Sinna's presentation,

47:40 he was giving an example of an event

47:42 study where foreigners are investing in US startups,

47:46 but especially in basic fields,

47:48 in more technical fields.

47:50 So that already shows that in real life,

47:53 uh,

47:54 uh,

47:54 countries,

47:55 when they are not able to deliver innovation by themselves,

47:58 especially in technical fields that are requiring

48:00 more basic knowledge like academic knowledge,

48:02 for instance,

48:03 which they are not good at,

48:05 then they try to go and,

48:06 and source it at the,

48:07 uh,

48:07 uh,

48:08 at uh directly from the US.

48:10 So that interaction,

48:11 that complementarity is also very interesting.

48:14 Another

48:15 They invest but they us

48:18 which uh looks like exactly the message that,

48:20 uh,

48:20 you know,

48:21 the WDR is already,

48:23 uh,

48:23 uh,

48:23 building on.

48:25 Getting connected to the world,

48:26 especially in technical fields,

48:28 is more valuable for the country.

48:34 Thanks,

48:35 uh,

48:36 the folk,

48:36 uh.

48:40 Uh,

48:40 may I come in here?

48:41 Maybe,

48:41 uh,

48:42 you have a question,

48:43 right?

48:43 Please come in.

48:44 Yeah,

48:45 thank you.

48:45 I think actually it's a good uh segue.

48:48 Now,

48:48 I was wondering because you,

48:50 you,

48:50 you know,

48:51 by what you just said,

48:52 but also

48:53 in the presentation,

48:55 uh,

48:56 when giving the example of um

48:59 uh in

49:00 sort of the technological exchange,

49:02 uh,

49:03 with Japan,

49:03 you mentioned that uh engineers from Korea.

49:07 would go to Japan.

49:08 So I was wondering,

49:10 uh,

49:10 both in reality but also in your model,

49:13 how does the human capital development of Korea,

49:18 uh,

49:18 its path of human capital development enabled

49:22 these,

49:22 um,

49:23 technological upgrading strategy.

49:25 Thanks.

49:27 I see.

49:27 I think uh that human capital was extremely important in that period,

49:32 but just one thing is that we didn't have really good data to distinguish the policy,

49:37 so we only can look at the outcome of the education,

49:41 for example,

49:41 like a college entrance rate or graduation rate

49:45 is,

49:45 uh,

49:45 was much higher than other middle-income countries,

49:48 so that can uh reduce the

49:51 cost for innovation or adoption because uh smarter people can also innovate,

49:57 like,

49:57 uh,

49:57 I guess better and,

49:58 you know,

49:59 like adapt also better.

50:01 But here,

50:01 uh,

50:03 I'm,

50:03 I,

50:03 I can clearly see that uh there is a complementary

50:06 between the education policy and this innovation and adoption policy,

50:10 but here,

50:11 we didn't have really good data,

50:12 so that's why we just,

50:13 uh,

50:14 assume that this is constant over time.

50:27 But Rufuki had also looked carefully at the human

50:30 capital upgrading and technical skill development in Korea,

50:33 right?

50:34 So maybe you want to say something on this.

50:36 No,

50:37 that's,

50:37 uh,

50:37 that's absolutely right.

50:38 So in the,

50:39 in the,

50:39 in the main analytical framework

50:41 for the WDR,

50:43 uh,

50:43 there is already a,

50:45 uh,

50:46 a very clear message coming out,

50:48 which is,

50:48 of course,

50:49 uh,

50:50 in order to innovate,

50:52 you need to have good engineers,

50:53 educated engineers.

50:55 Uh,

50:55 if you have good engineers,

50:56 then your firms can become more innovative.

50:59 Uh,

50:59 but in the beginning,

51:00 you don't have that,

51:01 obviously.

51:02 So as a result,

51:03 you know,

51:03 either you can have an organic.

51:06 Uh,

51:07 education reform and have your human capital pool grown organically over time,

51:12 but that's going to be a long-lasting process.

51:15 So probably it will take 1015 years to improve the education pool in the country

51:21 because they need to also get technical training and also work experience.

51:25 An alternative is directly

51:27 to,

51:27 you know,

51:28 borrow engineers from uh uh.

51:32 Or uh utilize the diaspora.

51:35 Uh,

51:35 so those two things can also strengthen the talent pool,

51:38 uh,

51:39 relatively quickly,

51:40 and that can create a synergy between firms who are trying

51:43 to innovate versus the human capital that exists in the country,

51:47 and the,

51:48 the two can,

51:48 can,

51:49 and can,

51:50 uh,

51:50 deliver much more valuable output.

51:52 And that's exactly the main result that's emerging also from the

51:56 analytical framework that we are adopting for the World Development report.

52:01 That was a great point,

52:03 ER.

52:06 Right,

52:06 colleagues,

52:07 any other comments?

52:09 Please take the,

52:11 take advantage of our speakers and if any questions on,

52:16 especially on the trade issues

52:18 from Sina's review,

52:19 please come on in.

52:29 All right.

52:30 Uh,

52:30 it looks like we have no further questions.

52:32 So colleagues,

52:32 thank you very much for your time.

52:35 Uh,

52:35 Aung Hen and Sina,

52:36 thank you very much for joining us and presenting your work.

52:39 We are putting out the background paper on the WDR web page,

52:43 so you should be able to access that.

52:46 Uh,

52:46 this recording and slides will also be

52:49 be available.

52:50 We'll send out the announcement for next week's seminar soon,

52:54 so look forward to seeing you next Wednesday.

52:56 Thank you.

showAllTimestamps
no
transcript
Good afternoon, everyone. Uh, it's very, uh, uh, good to kind of see you join the weekly seminar for the 2024 WDR that focuses on economic growth in middle-income countries. Uh, I'm Sho McLal and with our lead academic Ufuk Akcheed and our TTL, uh, Joyce Abraham. We are very happy to welcome you to this seminar this afternoon that has, um, Two of our background paper contributors, Yong Hon Shim and uh Sina Attis, who are going to tell us a little bit about an overview on the role of trade and firms in economic development and they do a very nice synthetic piece on that. In addition, Yung-hun and uh Sina will go deeper into the South Korean experience. And colleagues, uh, this is the first week of the next 3 weeks, we'll be focusing on different economic stories from South Korea, so keep tuned on that. So as, uh, words of introduction, uh, seen as a principal economist at the Board of Governors for the Federal Reserve. And uh he does work, a lot of work on looking at firm dynamics and looking at innovation, competition with uh a strong focus on quantitative analysis. Jung-hen Shim uh is, has just joined the IMF uh research department, but we'll still collaborate with him. It's OK. Uh, so, uh, Jung-hen's been doing a lot of work on macroeconomic growth and international trade. And his research really looks at the role of technology adoption. And innovation over different stages of development. And that's why Junghan your research is so central to some of the ideas we are developing and putting forward in this year's WDR. Uh, Yong-han just completed his PhD from the University of Chicago and Yuk was his principal advisor. So colleagues, what I'll do is, uh, give you the floor. You have about 40 minutes. To present the key ideas from these uh background papers, and then we'll open it up for a discussion. Over to you. Uh, Xena, let me show you the slide. Uh, so thank you very much. Um, I'm Sina Attesh from the board and happy to talk about international trade and dyna uh, from dynamism, dynamics and economic development, our joint work with Yonkun, and the usual, uh, board and IMF disclaimers about, uh, about employees' work, of course, applies here. So, uh, just briefly about the outline, uh, of today's talk. So I'm gonna first start with, um, some, uh, key highlights from the literature on international, like openness to trade or more broadly international linkages and how they affect growth like the key channels. Then I'm going to focus a little bit on the, uh, technological competition side like or foreign competition, how it can foster. Uh, technology upgrading, uh, among domestic, uh, domestic firms and what type of policies can indeed, uh, support, support this relationship. And in light of this discussion, Yongong will continue, um, with, uh, a refreshing look and like with no facts on the long-lasting development process of South Korea and what type of industrial policies indeed supported, uh, supported this, uh, persistent, uh, development experience. So with that, let me uh go into the key channels that are highlighted in the literature on international trade and economic growth. Uh, there's a large body of work that uses as a natural uh natural experiment, the, uh, trade liberalization episodes in, in Different countries. Most of these work, uh, relates to developing countries like the, uh, experiences of Latin American countries over 1991980s, 1990s, and, um, the first channel that, uh, that is emphasized by this work is The access to better inputs, so inputs that can be, uh, that can be imported from, uh, from abroad, which can help, uh, firms improve their productivities. Market size is a second channel, uh, in this, in this literature, so opportunity to access broader markets can, uh, induce firms indeed to improve their products and processes. A third channel is the competition aspect of trade liberalization on which I'm going to spend a little bit more time, uh, later, so I'm skipping it right now. But a key highlight from, uh, from this body of work is indeed that there's some heterogeneity, uh, in the effect of this liberalization episodes on, on firm dynamics. So firms that are readier to begin with, let's say that, that have a high productivity or that are more and the intensive, uh, uh, before, before trade liberalization episodes actually happen to gain more, uh, from, uh, from these episodes, like, uh, an important heterogeneity that I'm gonna highlight later, uh, in my talk. And another channel on which there's vast and expans uh expanding, uh, literature, um. Is the knowledge spillovers and technology transfers across, across countries. Uh, so there's a body of work, uh, in this literature that, uh, compares and contrasts the importance of technological, technology adoption versus, uh, standalone innovation over different phases of, uh, economic development. Indeed, Yong Hong's work, recent work, uh, sheds like, uh, Uh, a new, uh, a fresh look, uh, on these dynamics that he's gonna, that he's gonna discuss, uh, in his part. A related literature here is the, uh, cross-border flows of investment, particularly foreign direct investment and the associated, uh, technology transfers. So a recent work here to emphasize, I think is uh by Chris Frans Rosen, Shebnam Kalis Jan, uh, and, and co-authors who highlight that not only the firms that see FBI benefit from associated knowledge spillovers and technology transfers, but also other firms in, in similar sectors operating uh with similar technologies happen to, uh, benefit from these flows as well. Um, in, in, in recent and related work, we also highlight that FDI flows can indeed help or alleviate negative implications of uh of market concentration in a domestic economy. We indeed document that. In more concentrated sectors, domestic firms, uh, benefit more from FDI flows in terms of their productivity growth. And again, not just the firms that receive those flows, but other firms in the same sector. So FDI flows and associated technology, uh, uh, technology transfers may, uh, help with business dynamisms in domestic, um, in the receiving or the, in the host countries, um, uh. Uh, supporting business dynamism in more concentrated sectors. So, uh, in another recent work, actually, we look at, at, at, at the uh similar relationship, uh, in cross-border flows that are going in the reverse direction that, that are, that is more, uh, or less explored in the literature. So if you, if you go to the next slide, Yono. So, uh, now we look at foreign firms' investments in countries that are indeed at the technology frontier. So we look at the transactions between foreign firms that are investing in, uh, US startups that are, uh, at their, at the frontier. So, um, what we document here is that, of course, as might be expected, firms that invest in those US startups, increase patenting in, in those technologies and also increase their citations to the patent portfolio of those startups. But again, this is not just them. Also, other firms, uh, in the same sector with this investing foreign firm increase their patenting in those technologies and again citations to the, uh, patent portfolio, uh, of the US, US firms in that technology. So the knowledge spillovers that are coming from those transactions from these, uh, investments abroad actually is broader and concern, not just the firm that is investing. And another interesting fact that we that we highlight is that these effects are indeed larger and stronger in sectors that rely more on basic technologies, so technologies that require a larger upfront investment in like in in fundamental research in human capital, and indeed firms are apparently like more laggard foreign firms invest in US dollars or Firms that are at the technology frontier to acquire knowledge and to build up capacity, uh, faster in such technologies and, um, WDI actually emphasizes a lot like the importance of human capital. So apparently there might be complementarities coming from these type of investments, uh, learning and acquiring frontier technology through such investments and, uh, which can help improve also human capital and associated technologies, uh, in the, in the foreign country. So with that, uh, let me now switch to the, uh, final, uh, feature that I want to highlight that is associated with, uh, openness to trade, uh, the competition aspect, which is, uh, the more contentious aspect, uh, of openness to trade. So in recent work again with Ufuk Akjeed and Jean-Marie Implotti, we uh tackle indeed this, uh this perennial question of how to manage uh intensified competition that is exerted by foreign firms and what policies can help them. Domestic firms in such competitive, uh, competitive rates. So whether tariffs or protecting these firms from competition is the, uh, better way to go, or whether supporting these firms, these domestic firms through R&D subsidies is more welfare enhancing. And we look at these policies also considering differences between the short run and long run, like differentiating between different horizons of, uh, of policymakers. So, a key, uh, part of our analysis or of the framework that we use in this, in this work is that while it is encompassing, like many features or aspects of openness to trade that I was highlighting, uh, while While capturing these aspects, it also looks or like captures explicitly the heterogeneous effects that competition has on domestic firms. So there's this, there's of course the very important concern that keeps Rising in policy discussions that foreign competition from like better, like exhorted by better firms that are at the technology frontier can drive domestic firms, uh, can cause domestic firms to lose their market share or even exit the markets. And indeed, our model captures these dynamics, but there's also another side of competition, which is competitive pressures can incentivize some domestic firms to improve their products and processes just to be able to compete and or overcome their uh foreign rivals. But of course, which dynamic will play out depends on the, uh, on the difference between the rivals, just like in a, in a race of, in a running race, right? If the two competitors are close to each other, there's the hope or potential of overcoming your competitors, just exerting a little bit more effort. But if the difference between the 22 rivals is very wide, the laggard, uh, the laggard runner or the laggard firm loses the hope of overcoming its rival, capturing markets, so it gets discouraged and relaxes its innovation effort. So the framework, as you see, Relates technological differences to the dynamics of competition, and those affect the dynamic incentives, forward-looking investment incentives of, uh, of, uh, both domestic and foreign firms. And this type of heterogeneity indeed speaks very much to the extensive empirical work that highlights these different dynamics. So just to highlight, uh, in the next slide, just to highlight how uh this relationship between competition and innovation effort, uh, transpires in the model on, in the left panel. I show again from the calibrated model, the innovation effort of a generic firm in the model as a function of its position relative to its foreign rivals. So basically as a function of the technology gap between the two, between the two firms. As you clearly see, there are two spikes. Let me briefly explain, uh, what they mean. On the left, This concerns a domestic firm that is slightly inferior in technology than its foreign rival, but there are trade costs, so importing those better goods is costlier due to these trade costs. So as a result, this domestic firm that is slightly inferior still has some, some food in the game. So it can compete with the, with the foreign rival. So when their technologies are close to each other, whilst accounting for the extra costs of importing, the domestic firm exerts more innovative effort just to be able to defend its domestic market against, against, um, this foreign rival. What happens around the uh rights spike is basically, it is about the domestic firms that are, uh, that have better technology than their foreign rivals, but they can't, it is not enough to export to the other markets because they also need to overcome this higher, these, um, these higher costs associated with trade, be it like iceberg costs or higher tariffs, like, uh, tariffs, um, put in place by the, by the other country. So, but if the domestic firm is closing competition, so if it can improve its product, product quality a little bit more and overcome its rival and the uh trade costs, it can gain access to export markets. So it intensifies its innovation efforts. So as you see, there are two regions where Technologies get closed and competition stiffens, and domestic firms are incentivized to, to exert more innovative effort. And interestingly, we see a similar pattern in the data on the right, uh, on the right panel to show that if you create a measure of technology gaps in the data and then look like the uh Uh, the, uh, innovate innovation in uh, intensity of domestic firms, in this case, the US firms, you see a similar pattern of, uh, double peaks like where uh innovation in, uh, intensity actually uh intensifies. Uh, so this is a reassuring and validating, uh, evidence for the mechanism that we are talking about here in this one. So just to Recap the policy implications of this model. importantly, there are two sides to policy. Like there, there are, first, the aesthetic effects and the dynamics, dynamic effects. So, uh, for instance, consider a unilateral increase, uh, like protectionist policies, a unilateral increase in trade barriers on imports. So in that case, Domestic firms, or dome domestic country can indeed Potentially benefit from such policies, but, uh, shifting production and profits to home. Like, so protecting firms or increasing barriers against imports, uh, means import becomes more costly and production shifts to domestic firms. But there are dynamic consequences. Such policies can hurt forward-looking incentives of, of such firms, as I was showing. If barriers are higher against imports, so domestic firms are just shielded from competitive pressures, and they don't need to exert as much effort in improving their product qualities or processes, so they, they, they exert less effort to improve technologies. So as a result, what we find is Profit shifting motives are actually dominant in terms of welfare only when a few years ahead are considered. So if the policy horizon is longer than that, just beyond a few years, or maybe unfortunately like longer than a, longer than a presidential term. Then the dominant, uh, effect is coming from the negative dynamic consequences because firms are not exerting enough effort as they don't feel much competitive pressure. Indeed, what we find is the optimal trade policy is to slash tariffs to zero if the policy horizon in consideration is just beyond a few, uh, a couple of, a couple of years. And the optimal subsidy, sorry, optimal policy for a policymaker that is concerned over that is concerned about the medium to longer term. So a policymaker who is non myopic, the optimal policy is using R&D subsidies. So basically, Supporting firms in the competitive rate instead of killing competition and the associated incentives, uh, with higher barriers to trade. And we also document a policy complementarity. So in a world that is more globalized, defined by lower bilateral tariff rates, there's less need to intervene via R&D subsidies just because markets. And competitive pressures take care of, uh, appropriate incentives just by themselves, uh, forcing firms to, to increase their innovative efforts. But let me conclude on my last slide just with some policy. Implications for developing countries that are, uh, that are facing, uh, some challenges in their, uh, in their middle income base. So of course, every, every economic economic analysis and model abstracts from some considerations and of course, there are valid considerations uh surrounding such policies like, uh, uh, removing trade barriers, for instance, just outright might entail other considerations like they, they might generate great dislocations. There might not the transition of the firms to innovate, uh, activities might not be as smooth. But I think the importance of our policy implications are not just the exact uh policies per se, but more about what the goal or the uh direction of policies should be. What we highlight is the protectionism. Basically distorts forward looking incentives of firms and indeed if if a country after years of protectionist policies wants to decrease these barriers, it might find itself in a worse position than it would be in an alternative or counterfactual economy where trade barriers were lower to begin with just because firms were shielded from competition and they were not exerting enough effort in technology upgrading. And indeed, indeed history is bright with these examples, right? Like uh examples of import competition that Latin American countries or, um, like Turkey over 1960s and 1970s, leftist countries, uh, with firms that are dependent on state support and not ready to compete, uh, with foreign rivals once the, uh, uh, trade barriers were, uh, were, uh, removed. So as a result, what we highlight is that the goal needs to be to support firms to build up capacity to compete, to upgrade their technologies and to be ready to innovate in the future. I think in this sense, Korea's example is very telling that they have been steadily. Increasing their, uh, their, their import tariffs over the development stage, while at the same time supporting their firms to invest in technologies, uh, both via imitation and also innovation policies, as Yongoon's recent work, uh, high, highlights, uh, very clearly. And with that, let me, uh, leave the stage to Yong Hoon, uh, for, uh, for his talk on South Korean experience. Thank you very much. OK. Thanks, uh, thanks for having me. So I'm gonna present the second half of the presentation, and I would like to zoom into a specific case, which is South Korea. So here, we plot the GDP per capita of South Korea in the red line with other different countries. So here we can see that at the beginning in the 1950s, Korea started as one of the lowest income countries like China, India, or Vietnam. And in the 1980s, it caught up with some middle income countries like Brazil, Mexico, and Turkey. But even after that, Korea kept growing, and now it is in the process of catching up with some of the high-income countries like Japan, France, or the US. And behind this development, there are, there were several industrial policies during that development period. So the first industrial policy was sector-specific policy. Where the government was trying to support the heavy manufacturing sector. So there are several papers documenting this policy, and it turns out that this policy was effective at transforming the country from non-manufacturing or light manufacturing intensive economy to more heavy manufacturing intensive economy, which is more export-oriented. And also at the same time, there was a trade policy where as Sina mentioned briefly, the government was reducing the import tariff while the country was catching up with the frontier countries. And the 3rd industrial policy was somewhat implicit in a sense that they were implicitly picking the winners in a sense that they look for the firms and they kind of point out the successful ones at the beginning and then they keep subsidizing these large incumbents conditional on that they're exporting their goods to other countries. So we have recent work, uh, documenting these patterns by looking at the capital and labor wedge of these superstar firms and study the impact of these, uh, uh, policies on the increased, uh, market concentration over time. And the last one is the technology policy, where the government first started by subsidizing technology adoption, and then they shifted toward innovation subsidy. And this will be my focus on this today's presentation. So here, we plot the technological growth of South Korea with two different measures. First one in the red line is the number of patents per 1000 people. So we can see at the beginning in the 1980s, the number of patents per capita is almost zero, but it rapidly grew over time, and in 2020, it is almost like 3 per 1000 people, which is the uh one of the highest numbers in the world. And the second measure is R&D intensity, which is R&D expenditure over GDP in the blue line. And we can see that again in the 1980s, R&D intensity is almost zero, but it grew rapidly and now it is like 4%, which is, which is the second highest number among the OECD countries. So our question is, how this country transformed from non-innovative country to innovative country. And to see it more clearly, let me zoom into a specific case of the firm, which is Samsung Electronics. So in 1938, Samsung Electronics was founded. Although Samsung Electronics is now nowadays one of the largest companies in the world, but at the beginning it started just as a small company that made a noodle. In 1969, they decided to enter an electronics industry, but of course, they didn't have any technologies to develop any kinds of electronics products. So to tackle this problem, they decided to adopt some technologies from foreign companies, and they managed to sign a first technology transfer contract with a Japanese company called Sanyo. So under this contract, Samsung pays an adoption fee to the Japanese company and Sanos uh agrees to share the blueprints of the technologies and provide know-how transfer in a sense that they invite some of the Korean engineers to their factories and teach them how to use certain types of types of machines or teach them how to produce a black and white TV basically. And this was a very successful adoption, and Samsung Electronics was quite aggressive in adopting other technologies um uh from the frontier companies. But in 1980s, it becomes harder and harder to adopt these technologies for two reasons. First, uh, these Japanese companies become more reluctant to share the technologies. Why? Because Samsung, as Samsung becomes a, a large company and becomes a larger competitor to these Japanese companies, these Japanese companies don't want to share this technology with Samsung. So sometimes they refuse to share the technology or basically raise the price of the technology. And the second reason is that as Samsung grew, there are not many things to learn from the Japanese companies. Of course, Japan was producing better TV than Korean company, but still they are in the same bandwidth, so there are not, there are not many things to learn from these Japanese companies. So in, at the end of like the 1980s, Samsung decided to innovate their own technologies by, by investing a lot of money in R&D and hiring some scientists from the university. And then they managed to develop their own technologies and their own products. At the same time period, the government was quite active. In subsidizing this technology investment. In particular, they started by subsidizing adoption and then shifted toward innovation subsidy. So here in the left-hand side, we plot the adoption and innovation subsidy rate over time, which is tax credit for the adoption fee or R&D expenditure. We can see that it started by subsidizing adoption at 30%, but it gradually decreased over time, while the innovation subsidies started a little bit later than that, but it gradually increased over time. So when we look at the aggregate data, which is adoption expenditure shared, adoption expenditure over adoption, and R&D expenditure, we can see that the firms were transforming from adoption toward innovation as the country develops over time. So, uh, let me introduce the data what we, uh, that we are using in that paper and then present two empirical facts out of it. So the data, the, the main data we are using is the technology transfer contracts between Korean and foreign firms from 1962 to 1993. So here, we went to the National Archives and we digitized the official documents between the Korean and the foreign firms where they are uh where they are um uh trading the technologies. So here we can capture the universe of these contracts because at that time period, firms have to report all of their transactions involving foreign currencies. And we can see 50% more than the majority of the adoption contracts were with Japanese firms, and the remaining 25% were with US firms, and most of the contracts were in the manufacturing sector. Using this data, we merge this data with the firm balance sheet data so that we can see the basic information of the firms and also the patent data so that we can measure the innovation at the firm level. So here, I'm gonna present two empirical facts. First, when the initial productivity gap between the technology buyer and the seller is larger, then the productivity growth after adoption is larger. So here in the y axis, we plot the productivity growth after 5 years from adoption, which is measured by sales for employment growth at the firm level. And in the x axis, we plot the productivity gap between these two firms, which is measured by the ratio of sales per employment of the Korean firm divided by the sales per employment of the foreign firm. So we can see that uh when the Korean firm has a much less productivity than the foreign firm, which means that they are lagging behind the foreign firms. The productivity growth after adoption is larger. However, as the Korean firms are catching up with the foreign firms, so that they have more or less similar productivity, then the productivity growth after adoption becomes much smaller. On the other hand, when we plot this same exercise with the innovating firms, we can see that the innovating, uh, the productivity growth after innovation is basically flat over this initial productivity gap. Which means that, uh. The productivity growth after adoption depends a lot on the initial productivity gap compared with the innovation. And also on top of that, when the when the firms are lagging behind, then the adoption brings larger productivity growth than the innovation. But on the other hand, when the firm is catching up with the foreign firms, innovation brings larger productivity gain than the adoption. The second fact that I want to deliver here is that non-addopters also get knowledge diffusion from these adaptive technologies. So to see that, let's suppose a simple situation where Samsung adopted technology from SAO. Which is a Japanese firm, but not from Panasonic. And our question is, do other firms that did not directly adopt this technology are also getting knowledge diffusion from this adaptive technology. To see that, we follow the innovation literature and we assume that if patent A makes a citation to patent B, it means that patent A is learning something from patent B and is building on top of patent B. And if that's the case, then these other firms should increase the patent citation towards Sanu. Compare with the Panasonic firm, a Panasonic pattern, which is a plus b firm. So the basic idea is to plot the difference of citation between the seller and the plausible firm around the first technology adoption. So this is the result, and we can see that after several years from the first technology adoption, we can see that the other firms that did not directly adopt this technology also increased patent citation to this adaptive technology, which means that when other firms are adopting the technology, I can also learn something from this adapted technology, which implies that there is some knowledge spillover from these adoption activities. So those are basically two facts, and let me explain the takeaways. So the technology adoption can bring larger productivity gain when the gap between the uh the firm and the frontier firm is larger. And adoption also brings knowledge diffusion, which can potentially cause underinvestment problem of the firms because these are private firms do not internalize this knowledge diffusion. So subsidizing adoption can be very effective at the early stages of development. However, as we saw in the uh empirical fact, the gain from adoption will diminish as the country catches up with the frontier countries. So we have this basic idea and we quantify the policy implications using two contrite growth models with endogenous adoption and innovation decisions. So here we came back to the uh the adoption and innovation subsidy of South Korea, and then we put it in the model and then we're trying to compare the actual policy with no subsidies. Or what if we just subsidize adoption and never switch to innovation? Or lastly, what if we just started by subsidizing innovation and never subsidize adoption? So basically, we do some kind of simulation using the model. And then here's the result. So here's the case when we subsidize only adoption for the whole period with the same uh subsidy rate. And this is the, in the right-hand side, this is the GDP relative to the no subsidies case. So here we can see at the beginning, GDP is a little bit smaller than the no subsidies case. Because we are using uh more labor into adoption activities. Other than, uh, uh, instead of production. But in the beginning, it has a much higher growth rate. So GDP already caught up with the no subsidy case and it is much larger than the no subsidy case. More importantly This deviation from the no subsidy case does not grow exponentially, which implies that subsidizing only adoption does not increase long run growth rate. This is because, first, the gain from, uh, productivity, the gain from adoption will diminish as the, as the, as the Korean firms are catching up with the Japanese firms. And also subsidizing adoption will, will, will make the Korean firms stuck at the adoption phase in a sense that since the adoption is cheaper than the market price. Even if they can fully catch up the Japanese firms and leapfrog the Japanese firms by innovating their own technologies, they're just keep waiting for the Japanese firms to innovate, and they are just adapting and following the Japanese firms' technology whenever the Japanese firms innovate their own thing. So it it it actually. Can reduce the long-run growth rate. On the other hand, when we subsidize only R&D or innovation, at the beginning, it has a much smaller GDP than the adoption subsidy only case, because at the beginning, adoption is much more effective than the innovation, but subsidizing innovation instead adoption can be distorted. But in the end, it has a higher long run growth rate. So in the end, after like 40 years, it can uh catch up with the adoption subsidy case. Now we plug in the actual policy. Since the actual policy gradually shifted toward innovation subsidy from adoption subsidy. At the beginning, it closely followed the adoption subsidy case, and as the government switched to innovation subsidy, the long run growth rate is also higher than the adoption subsidy case. So in the end, when we calculate the welfare compared with the no subsidy case, we can see that the actual policy was the most effective compared with the two time variant policies. So I think uh I can conclude here. So the, uh, the main message we want that we want to deliver here is that the policies that were effective for developed countries may not work really well for developing countries. And this state-dependent policy that started with adoption subsidy and shifted toward innovation subsidy was indeed very effective in South Korea's development. Thank you very much. Thank you very much, Yanha and uh Xina, not only for the very uh useful presentation, but sticking to time. folk, if you're connected, would you like to provide some broader context before we open up to comments or questions? Mhm. Uh, uh, absolutely. So, uh, thanks, thanks, Jung and thanks, Xina, for, for, for this, uh, very nice presentation. Uh, so I, you know, the, the main, the main takeaways that I, uh, that I see in, in, in their presentation is that, uh, first, the, the optimal growth strategy is clearly, uh, stage dependent. And, uh, so, as a result, you know, not only, not only the, the firms have to be dynamic along the development stages. But also the policies, the optimal policies, the policymakers have to be dynamic as well. And I think, you know, uh, here, uh, Yong Hon is presenting obviously the, the optimal policies, uh, along the way, but there's also a major political economy problem, you know, changing the policies is, uh, is, is, is, is not easy, uh, uh, and, you know, uh, I think that's, that's an important message. The second one. Uh, uh, is that what is good for the society is not necessarily good for, uh, policymakers. I think that's also a very important message that's coming out in the sense that, uh, if policymakers' horizon is, is, you'd say, 5 years, which is the short run, or at most 8 years, in that case, protectionist policies can be uh uh optimal from a policymaker's point of view. But if you care about the longer run of the economy, Uh, of course, uh, boosting the technology is much more important, and to boost the technology, international spillovers and the competition channel, uh, coming from international trade is extremely critical. So that's again, I, I just kept hearing tensions, political economy in the background from this presentation, which was very useful. Uh, um, yeah, so let me stop here and I'm, I'm, I'm curious about the, uh, the audience, uh, feedback to you. Thanks for you raise important questions that in the current environment where trade restrictions, at least in technology access to certain countries are being limited, and what happens to their growth prospects. OK, colleagues, uh, uh, please, uh, uh, raise your hand and I'll turn to you, uh, if you have questions. Everyone's very, uh, reserved, I think, today, right? So, uh. Questions, uh, I don't see folks with questions, so, uh. Junghon, I think one of the key issues going back to your simulation is that you're saying this switch from imitation subsidy to innovation subsidy was timed pretty well in Korea. So what was it really that the Korean government was doing in terms of gathering information, able to, to be able to, you know, figure this switch out. And second, uh, I think a lot of the debate. If you look at the Fabrizio Zindiboti's papers, is that earlier on when you're imitating a lot of bad practices and conglomerates can get you there. But then when you're innovating, you need to sort of really get your competitive environment really well. So how did, what were the things happening in Korea when the switches were going on? I see. So, uh, basically, so when there's, there's, there's like a big change over this time and also like this time. That was quite closer to the president's change. So as Wuu already mentioned, there was like political constraint in a sense that uh uh during the incumbent presidency, it is kind of harder to change the policy, but after there's a new, new president, then it is relatively easier to change the policy, right? And I think the main information that they were using is just the GDP or like a very aggregate measure in terms of like uh compare the GDP with the Japan or the US and then try to figure out, oh now we are, we have to move on to innovation phase or something like that. And And I think the second question was, uh, OK, so it's, uh, about the Zillipoti paper. So there was also like an interesting paper, another interesting paper that talks about uh Asian financial crisis, uh, in 1997. So after that, there was like a reform, like a lot of reform from, uh, from the fund, and When you see the entry rate or entry rate in the patent data, the entry rate went up after the financial crisis, which I believe that it was uh helping the dynamism of the economy was going in the right direction. So it was a crisis, so it was a bad thing basically, but it has some upside from that crisis. So, so then if you look at the firms right now, besides your example, which is very detailed on Samsung, what is the general marketplace for coming up with new ideas and innovations? Is it largely in a model where the largest firms have R&D capabilities and technical capabilities, or do you also see in the Korean market that a lot of change is being driven by these, you know, new entrants who are growing rapidly? I see, but still, I mean, even after this crisis, still, uh, from the data, what I see is the leading firm in this innovation is incumbent and large firms. So for example, like Samsung, LG, like Hyundai, like, uh, are driving this innovation. But the context is a little bit different from the US, uh, uh, example in a sense that they are all exporting firms, so they are competing in the global market. So even if they are concentrated in a domestic market, they are not dominating in the global market. So in that sense, it is kind of, it can still be innovative in that sense. Colleagues, any questions from uh Well, you wanna come in with your question? So far. Just a, just a quick um question slash clarification. On, on, on this slide, your, your, your model, uh, you know, uh, you have a small caption there saying government spending is financed by a lump sum tax. So your model is assuming no dead weight loss from taxation, correct? So so my, my, my question to you, if you can clarify that, but also reflect in the policy world, when, when we look at this, um, sort of. Relative to no subsidy scenario, these subsidies are being financed by distortive taxes in, in the real world. So the results will all, you know, at least, uh, you know, will be lower than, than, than what you have in your chart or, or could be even different, right? Yeah, I think that's a very good point. So here to simplify the things, we assume that they are collecting tax by lump sum tax. But we have this additional parameter that governs the inefficiency of uh imposing lump sum tax. So now parameter is, uh, is 2, from the literature, which means that we have to uh collect $2 to use $1 for this, uh, this subsidy program. So it is actually even larger effect. If we assume that there's no inefficient, inefficiency, uh, at all from the lump sum tax, as you mentioned, but we are adjusting the parameter to be more realistic. Can I just add one more, uh, comment here, uh, to Yong-hun's answer. So Jaffar, you're, you're, you're bringing an excellent point. Uh, this class of models that Yong-hun is building, if you make the policy stage dependent, the That the model will ask you to have a lower tax on larger firms, and this is because to incentivize firms to be large, you are promising some additional return if you want to make the taxes distortionary. You can, in some sense put some backload of carrots so that you can also use the tax policy to create an additional incentive for firms to grow. Right, so, uh, first through the subsidy policy, you're subsidizing imitation, so that's good for the country, but if you want to also exploit variation in tax policy, these models will tell you that lower the taxes as firms are growing, then there will be a trickle down of incentives. So if being large is, is good, then all the smaller firms will also try to be uh good. So these models, these forward-looking uh uh competition models, uh, have that flavor. So that's something that I wanted to clarify. Uh, one additional thing, uh, uh, that I would like to also highlight is that, uh, in, in Sinna's remark, so he, he made a point that's very, I think, fundamental for the WDR which is, uh, in, in, in, in Sinna's presentation, he was giving an example of an event study where foreigners are investing in US startups, but especially in basic fields, in more technical fields. So that already shows that in real life, uh, uh, countries, when they are not able to deliver innovation by themselves, especially in technical fields that are requiring more basic knowledge like academic knowledge, for instance, which they are not good at, then they try to go and, and source it at the, uh, uh, at uh directly from the US. So that interaction, that complementarity is also very interesting. Another They invest but they us which uh looks like exactly the message that, uh, you know, the WDR is already, uh, uh, building on. Getting connected to the world, especially in technical fields, is more valuable for the country. Thanks, uh, the folk, uh. Uh, may I come in here? Maybe, uh, you have a question, right? Please come in. Yeah, thank you. I think actually it's a good uh segue. Now, I was wondering because you, you, you know, by what you just said, but also in the presentation, uh, when giving the example of um uh in sort of the technological exchange, uh, with Japan, you mentioned that uh engineers from Korea. would go to Japan. So I was wondering, uh, both in reality but also in your model, how does the human capital development of Korea, uh, its path of human capital development enabled these, um, technological upgrading strategy. Thanks. I see. I think uh that human capital was extremely important in that period, but just one thing is that we didn't have really good data to distinguish the policy, so we only can look at the outcome of the education, for example, like a college entrance rate or graduation rate is, uh, was much higher than other middle-income countries, so that can uh reduce the cost for innovation or adoption because uh smarter people can also innovate, like, uh, I guess better and, you know, like adapt also better. But here, uh, I'm, I, I can clearly see that uh there is a complementary between the education policy and this innovation and adoption policy, but here, we didn't have really good data, so that's why we just, uh, assume that this is constant over time. But Rufuki had also looked carefully at the human capital upgrading and technical skill development in Korea, right? So maybe you want to say something on this. No, that's, uh, that's absolutely right. So in the, in the, in the main analytical framework for the WDR, uh, there is already a, uh, a very clear message coming out, which is, of course, uh, in order to innovate, you need to have good engineers, educated engineers. Uh, if you have good engineers, then your firms can become more innovative. Uh, but in the beginning, you don't have that, obviously. So as a result, you know, either you can have an organic. Uh, education reform and have your human capital pool grown organically over time, but that's going to be a long-lasting process. So probably it will take 1015 years to improve the education pool in the country because they need to also get technical training and also work experience. An alternative is directly to, you know, borrow engineers from uh uh. Or uh utilize the diaspora. Uh, so those two things can also strengthen the talent pool, uh, relatively quickly, and that can create a synergy between firms who are trying to innovate versus the human capital that exists in the country, and the, the two can, can, and can, uh, deliver much more valuable output. And that's exactly the main result that's emerging also from the analytical framework that we are adopting for the World Development report. That was a great point, ER. Right, colleagues, any other comments? Please take the, take advantage of our speakers and if any questions on, especially on the trade issues from Sina's review, please come on in. All right. Uh, it looks like we have no further questions. So colleagues, thank you very much for your time. Uh, Aung Hen and Sina, thank you very much for joining us and presenting your work. We are putting out the background paper on the WDR web page, so you should be able to access that. Uh, this recording and slides will also be be available. We'll send out the announcement for next week's seminar soon, so look forward to seeing you next Wednesday. Thank you.
showAllTranscripts
no
duration
PT52M57S
scene7File
worldbank/Trade-firm-econ-dev
scene7Domain
https://worldbank.scene7.com/
scene7FileAvs
worldbank/Trade-firm-econ-dev-AVS
title
Trade-firm-econ-dev
description
Trade-firm-econ-dev
showTimestampAndTranscript
yes
col-xs-12
col-sm-12
col-md-3
col-lg-3
col-xs-12
col-sm-12
col-md-7
col-lg-7
  • add-style
  • lp-body-content
In this WDR 2024 Seminar Sina Teoman Ateş (Principal Economist at the Board of Governors of the Federal Reserve System, International Finance Division) and Younghun Shim (Economist, Research Department of the International Monetary Fund) discuss "Trade, Firms, and Economic Development" with chair Somik Lall (Director, World Development Report 2024, World Bank Group).
lp-heading-top-medium
lp-heading-bottom-medium
col-xs-12
col-sm-12
col-md-2
col-lg-2