WEBVTT - How Economic Complexity Explains Which Countries Become Rich 0:00:10.240 --> 0:00:13.920 Hello, and welcome to another episode of the Odd Lots Podcast. 0:00:14.000 --> 0:00:16.360 I'm Joe Wisenthal and I'm Tracy Alloway. 0:00:16.680 --> 0:00:19.440 Tracy, I'm super late to everything. First of all, did 0:00:19.480 --> 0:00:20.120 you play Wordle? 0:00:21.600 --> 0:00:24.120 I did. I didn't get obsessed with it like some 0:00:24.200 --> 0:00:27.040 people did, but I think we were all fairly bored 0:00:27.280 --> 0:00:30.000 during that time period and on the lookout for any 0:00:30.040 --> 0:00:31.480 sort of entertainment. 0:00:31.080 --> 0:00:34.919 And it never clicked. But I'm always late to everything. 0:00:35.400 --> 0:00:37.320 Like a year ago, people were telling me it's like, oh, 0:00:37.320 --> 0:00:40.360 you gotta play this game trade all, and I did end, 0:00:40.600 --> 0:00:42.440 but like three or four months ago, I started really 0:00:42.440 --> 0:00:44.160 getting into it, and I think you're into it now too. 0:00:44.360 --> 0:00:44.600 Yeah. 0:00:44.680 --> 0:00:48.320 I think we both started playing around the same time, 0:00:48.600 --> 0:00:50.920 probably because we heard about it from the same person. 0:00:51.560 --> 0:00:53.880 It is a fun game, so for those who don't 0:00:53.960 --> 0:00:58.400 know it is like Wordle. It basically presents a sort 0:00:58.440 --> 0:01:03.840 of graphic schematic of an unnamed country's exports, and you 0:01:03.880 --> 0:01:05.920 have to try to guess what country it is. 0:01:06.520 --> 0:01:09.360 Yeah, exactly, And I'm not very good at geography, so 0:01:09.600 --> 0:01:12.720 often it'll say like, Okay, you're close, but it's like 0:01:12.720 --> 0:01:15.080 you have to go fifteen hundred kilometers to the northwest. 0:01:15.160 --> 0:01:17.399 I'm pretty bad at geography, so I'm not good. But 0:01:18.680 --> 0:01:22.440 as playing this game and watching the way the different 0:01:22.520 --> 0:01:26.000 shapes of different countries export mix, I feel like I've 0:01:26.080 --> 0:01:29.080 really started to learn things about the world. 0:01:29.480 --> 0:01:32.640 I have learned so much about the economy of Angola. 0:01:32.720 --> 0:01:33.000 Right. 0:01:33.080 --> 0:01:36.240 So you see a country and it's like eighty percent 0:01:36.319 --> 0:01:40.039 of their exports are like coffee in gold, right or 0:01:40.040 --> 0:01:42.160 something like that, and you're like, Okay, this is a 0:01:42.200 --> 0:01:45.680 relatively poor country. It has a lot of growth left 0:01:45.680 --> 0:01:49.200 to do. And then you see another country and it's 0:01:49.240 --> 0:01:56.120 like advanced circuits and medicine and hangars and bananas and 0:01:56.120 --> 0:01:57.520 all this stuff, and like, oh, and you start to 0:01:57.520 --> 0:01:59.960 see these shapes and these distributions. That's sort of like 0:02:00.280 --> 0:02:02.240 tell you things. It's like, Okay, I can guess maybe 0:02:02.240 --> 0:02:04.120 this is in Europe, or maybe they have a lot 0:02:04.120 --> 0:02:06.240 of things, maybe it's in Eastern Europe. You suddenly sort 0:02:06.280 --> 0:02:09.480 of learn like how rich countries goods exports really differ 0:02:10.040 --> 0:02:12.079 from poorer countries goods exports. 0:02:12.160 --> 0:02:15.120 Yeah, totally. There are three clues that it gives you. 0:02:15.200 --> 0:02:18.440 One is distance, as you just mentioned, and the other 0:02:18.600 --> 0:02:21.160 is total size of exports. So that gives you an 0:02:21.160 --> 0:02:24.680 indication of how big or small that economy is. And 0:02:24.720 --> 0:02:28.400 then that third thing is the nature of the exports, 0:02:28.400 --> 0:02:31.600 and it sort of divides them up into different categories, 0:02:31.840 --> 0:02:34.880 but it gives you a really good snapshot of a 0:02:34.919 --> 0:02:38.720 country's economy. And as you play the game, you start 0:02:38.720 --> 0:02:44.240 to recognize I guess certain economic export attributes. Yeah, go 0:02:44.560 --> 0:02:47.440 with certain types of countries or economies. 0:02:47.560 --> 0:02:51.760 Right, so integrated circuits and palm oil, probably Southeast Asia, right, 0:02:51.919 --> 0:02:55.080 something like that. Anyway, this game that you and I 0:02:55.080 --> 0:02:58.080 have become obsessed with, it's sort of based on this 0:02:58.200 --> 0:03:02.760 work related to economic complexity, the Atlas of Economic Complexity, 0:03:03.080 --> 0:03:05.520 this sort of idea, and this is something that's come up. 0:03:05.520 --> 0:03:07.520 It came up in an episode we did with Henry 0:03:07.520 --> 0:03:10.720 Williams and David Os. It came up with the episode 0:03:10.720 --> 0:03:13.400 we did with Dan Wong and why some countries can 0:03:13.440 --> 0:03:17.040 develop airline industries and other countries can't. This idea that 0:03:17.560 --> 0:03:22.760 complexity of goods exports complexity is in itself a sort 0:03:22.800 --> 0:03:24.000 of predictor of wealth. 0:03:24.320 --> 0:03:28.720 Yeah, and maybe it's also a desirable model for countries 0:03:28.760 --> 0:03:31.320 to sort of aspire to, this idea that maybe they 0:03:31.320 --> 0:03:34.800 want to get away from simply producing a bunch of 0:03:34.840 --> 0:03:37.480 t shirts, So like fifty percent of their economy is 0:03:37.560 --> 0:03:40.160 T shirt exports or something like that. They want to 0:03:40.200 --> 0:03:43.680 get to a place where they have expertise across a 0:03:43.840 --> 0:03:47.880 broad and value added sort of realm of exports. 0:03:48.160 --> 0:03:50.440 Or maybe they export a lot of nickel and they 0:03:50.480 --> 0:03:53.000 want to be in like refined nickel, right, some nickel 0:03:53.040 --> 0:03:56.600 relatais the commodity just pure Anyway, this has sort of 0:03:56.640 --> 0:03:59.040 like really opened up a lot of these conversations playing 0:03:59.040 --> 0:04:01.880 like thinking about the world. So I'm really excited about 0:04:01.920 --> 0:04:04.640 our guest because our guest has done more work on 0:04:04.680 --> 0:04:08.160 this idea of economic complexity and why nations are able 0:04:08.200 --> 0:04:12.640 to develop complex, rich economies. He's also the creator of 0:04:12.720 --> 0:04:15.800 that at list of economic complexity. We're going to be 0:04:15.800 --> 0:04:18.560 speaking with Ricardo Housman. He's a professor at the Harvard 0:04:18.640 --> 0:04:22.640 Kennedy School and the founding director of the Harvard Growth Lab. 0:04:23.200 --> 0:04:26.240 All of my trade off friends are super excited about 0:04:26.279 --> 0:04:29.360 listening to this conversation. Doctor Husman. Thank you so much 0:04:29.400 --> 0:04:30.560 for coming on odd Lats. 0:04:31.640 --> 0:04:33.760 Oh, it's a pleasure to be with you. Thank you 0:04:33.800 --> 0:04:34.440 for inviting me. 0:04:34.640 --> 0:04:38.000 Absolutely. What is economic complexity? 0:04:38.720 --> 0:04:43.000 Economic complexity is an attempt to measure how much countries 0:04:43.480 --> 0:04:46.520 or places know what to do, so like I try 0:04:46.920 --> 0:04:51.280 to measure no. How Now, if you want to think 0:04:51.839 --> 0:04:54.640 about knowledge and say, well, and I know people who 0:04:54.760 --> 0:04:58.400 have a bachelor's degree, people who will hire high school dropouts, 0:04:58.480 --> 0:05:01.599 people who have a PhD. That sort of tells you 0:05:01.680 --> 0:05:05.760 how much a person knows. But if you ask yourself, 0:05:06.040 --> 0:05:09.760 how much does a society know, well, that would be different. 0:05:09.800 --> 0:05:12.520 That would not be characterized by the average, say the 0:05:12.640 --> 0:05:16.680 average number of years of schooling that the society has. 0:05:17.240 --> 0:05:21.200 Not A society that is full of just dentists will 0:05:21.279 --> 0:05:24.560 know less than a society that is half dentists and 0:05:24.640 --> 0:05:28.799 half lawyers, or a society that is a third dentist, 0:05:28.920 --> 0:05:32.760 a third lawyer, or third engineers. So in some sense 0:05:32.800 --> 0:05:36.760 you want to know how much the whole of society knows. 0:05:36.920 --> 0:05:40.680 And one of the important things about knowledge is that 0:05:40.880 --> 0:05:45.880 knowledge at the societal level has been exploding exponentially, but 0:05:46.040 --> 0:05:50.640 our mental capacity to know has not. So the way 0:05:50.720 --> 0:05:56.800 the economy has been adapting and adopting growing amounts of 0:05:56.920 --> 0:06:02.080 knowledge is by putting different bits of knowledge in different heads, 0:06:02.120 --> 0:06:04.880 sort of like parallel processing. You know, if you want 0:06:04.880 --> 0:06:07.200 to run a company, you will need somebody who knows 0:06:07.240 --> 0:06:11.799 about accounting, about finance, about marketing, about human resource management, 0:06:11.880 --> 0:06:17.160 about contracts, about taxes, about procurement, about engineering. So you 0:06:17.279 --> 0:06:20.760 want to have a lot of knowledge to run these things, 0:06:21.040 --> 0:06:23.800 but you cannot stuff that knowledge into a single head. 0:06:23.839 --> 0:06:26.799 You have to spread it into a bunch of heads, 0:06:27.680 --> 0:06:31.359 and then you have to bring those heads together back again. 0:06:31.480 --> 0:06:34.440 You have to kind of put humpty dumpty back together again. 0:06:35.480 --> 0:06:39.000 So the way in which a society grows is it 0:06:39.040 --> 0:06:42.040 grows its knowledge by putting different bits of knowledge in 0:06:42.080 --> 0:06:46.000 different heads and then by bringing those heads together. Now, 0:06:46.040 --> 0:06:50.320 if a society makes very simple things, it makes things 0:06:50.320 --> 0:06:53.040 that can be done by few people, because the knowledge 0:06:53.080 --> 0:06:55.760 that is needed to make one of those things, you know, 0:06:55.839 --> 0:06:58.640 fits in just a few heads. But if you're going 0:06:58.680 --> 0:07:02.080 to do stuff that requires not a lot of knowledge, 0:07:02.120 --> 0:07:05.080 you'll have to bring many, many more of these heads together. 0:07:05.400 --> 0:07:08.320 You'll have to network these brains together to make that thing. 0:07:08.880 --> 0:07:14.160 So complexity emerges as the consequence of distributed knowledge in society. 0:07:14.760 --> 0:07:17.760 You have different bits of people knowing different things, and 0:07:17.800 --> 0:07:20.680 then you have to bring those things together, and then 0:07:20.760 --> 0:07:25.760 these complex networks emerge from that process. So, in some sense, 0:07:25.800 --> 0:07:30.160 what is really driving growth is this growth of knowledge 0:07:30.200 --> 0:07:33.480 and the growth of using that knowledge, and consequently it's 0:07:33.520 --> 0:07:36.239 in this spread of different bits of knowledge in different 0:07:36.280 --> 0:07:39.920 heads and the ability to bring those heads together to 0:07:40.000 --> 0:07:42.840 make relatively long chains of brains. 0:07:43.360 --> 0:07:46.240 I have a bunch of questions already about comparing and 0:07:46.280 --> 0:07:51.880 contrasting measuring complexity versus the way economics has traditionally handled 0:07:51.880 --> 0:07:55.400 some of this. But maybe a step back question why 0:07:55.400 --> 0:07:58.640 did you decide to start looking into this? What is 0:07:58.680 --> 0:08:03.160 the benefit of looking at complexity within a particular society 0:08:03.280 --> 0:08:03.880 or economy. 0:08:04.800 --> 0:08:08.480 It goes back to the question and when Adam Smith 0:08:09.000 --> 0:08:11.440 asked himself what's the source of the wealth of nations? 0:08:11.680 --> 0:08:14.240 He said it was the division of labor? But why 0:08:14.280 --> 0:08:16.240 the hell would the division of labor matter? And he 0:08:16.360 --> 0:08:19.440 gives a thin factory example that if you split the 0:08:19.480 --> 0:08:21.520 work and this guy does the head and this guy 0:08:21.560 --> 0:08:25.440 does the body, et cetera, that increased productivity. That idea, 0:08:26.480 --> 0:08:29.200 I think has a kernel of what the story is. 0:08:29.800 --> 0:08:34.160 But this stuff becomes incredibly powerful when you're talking about 0:08:34.440 --> 0:08:38.280 knowledge driving the economy and driving society. So it's really 0:08:38.320 --> 0:08:43.079 about the division of knowledge that's what drives growth. It's 0:08:43.120 --> 0:08:46.640 the division of knowledge that allows the whole to know 0:08:46.840 --> 0:08:50.480 more than its parts. And so how would you go 0:08:50.600 --> 0:08:54.800 about measuring what a society knows how to do. Well, 0:08:54.880 --> 0:08:57.360 let's look at what they do, because if they do something, 0:08:57.360 --> 0:09:00.000 it means that they know how to do it right, 0:09:00.400 --> 0:09:02.440 So it's proof that they know how to do it. 0:09:02.480 --> 0:09:05.760 So we can look at what a society does to 0:09:05.840 --> 0:09:08.400 figure out what is it that they know how to do. 0:09:09.040 --> 0:09:11.079 So when you look at a country and say, okay, 0:09:11.120 --> 0:09:13.400 this country is only good at doing a few things, 0:09:13.400 --> 0:09:16.400 and this other country is good at doing many more things. 0:09:16.800 --> 0:09:19.280 And this country here is making things that seem to 0:09:19.320 --> 0:09:21.920 be simple things that can be done in small groups. 0:09:23.040 --> 0:09:25.840 This other country that has to do things that require 0:09:25.920 --> 0:09:29.640 this very broad network of brains coming together to make something. 0:09:30.160 --> 0:09:34.400 It tells you something about how those economies are managing knowledge. 0:09:35.080 --> 0:09:37.040 Tracy is sort of hinting at this, and I want 0:09:37.040 --> 0:09:38.840 to sort of drill down on this before moving on. 0:09:39.080 --> 0:09:41.320 You know, there are various traditional ways in which we 0:09:41.440 --> 0:09:45.320 measure economies. GDP is more or less a different way 0:09:45.360 --> 0:09:48.040 of like you just add up everyone's income and you say, okay, 0:09:48.040 --> 0:09:50.520 this country is richer than this country, and this country 0:09:50.600 --> 0:09:53.600 is richer than the next. Give us the basics. You 0:09:53.640 --> 0:09:56.680 know you have country A and country B. How do 0:09:56.720 --> 0:09:59.959 you like measure? Okay, this country is more comp more 0:10:00.120 --> 0:10:04.280 complex economy capable of more complexity than country be. 0:10:04.960 --> 0:10:08.320 So let me first say you can measure GDP and 0:10:08.440 --> 0:10:12.000 so GPS how much countries are able to make in 0:10:12.080 --> 0:10:14.560 terms of income. That doesn't tell you why they are 0:10:14.600 --> 0:10:17.840 able to make it. Economic complexity is trying to get 0:10:17.880 --> 0:10:20.600 at why is it that you are able to generate 0:10:20.640 --> 0:10:24.280 more income? What's underpinning that? And for that, we like 0:10:24.360 --> 0:10:27.920 to measure how much a society knows in our standard measure. 0:10:28.000 --> 0:10:30.960 We've now applied it to a whole different bunch of fields, 0:10:31.320 --> 0:10:34.880 not just in exports but other things. But we started 0:10:34.880 --> 0:10:37.480 with exports, and the reason why we started with exports 0:10:37.640 --> 0:10:40.160 is because we needed a data set that included all 0:10:40.160 --> 0:10:42.280 the countries in the world, so we could benchmark all 0:10:42.280 --> 0:10:45.800 the countries in the world with a standardized classification. And 0:10:45.920 --> 0:10:49.880 since international trade it involves different countries, they've all agreed 0:10:49.960 --> 0:10:54.640 on some common classification scheme, so it was expedient for us. 0:10:54.920 --> 0:10:58.600 But experts are also something that tells you whether a 0:10:58.600 --> 0:11:02.520 society is good enough at making something that it is 0:11:02.600 --> 0:11:04.880 able to sell abroad, So it's kind of like a 0:11:04.920 --> 0:11:08.520 litmus test that you're pretty good at making something. So 0:11:08.559 --> 0:11:11.920 I don't really care how much they make. I just 0:11:12.000 --> 0:11:14.160 care that they are able to make it, so that 0:11:14.160 --> 0:11:17.640 that knowledge is somewhere in that society. So the way 0:11:18.120 --> 0:11:21.319 you would think about calculating how much knowledge of society 0:11:21.360 --> 0:11:23.400 has would say, tell me how many different things that 0:11:23.440 --> 0:11:26.160 are they able to make? You're mentioning angola, or they 0:11:26.200 --> 0:11:30.000 make mostly oils essentially that's their thing, or do they 0:11:30.000 --> 0:11:34.000 do many things? So the diversity of their export basket 0:11:34.080 --> 0:11:36.360 is kind of like a first cut, right, But you 0:11:36.400 --> 0:11:41.120 would say, well, but products differ in how knowledge intensive 0:11:41.160 --> 0:11:43.720 they are and how difficult they are to make. So 0:11:44.040 --> 0:11:46.800 we found a trick on how to measure how difficult 0:11:46.800 --> 0:11:49.640 it is to make a product by simply asking the 0:11:49.720 --> 0:11:52.760 question how many countries are able to make this product. 0:11:53.200 --> 0:11:56.000 So if you talk about raw wood, many many countries 0:11:56.040 --> 0:11:59.360 export raw wood. If you tell me about microscopes or 0:11:59.480 --> 0:12:02.720 X raymage, very few countries are able to export those things. 0:12:02.960 --> 0:12:06.240 So that tells you something already about how difficult it 0:12:06.320 --> 0:12:08.960 is to make these things, right, So we call that 0:12:09.000 --> 0:12:11.960 the ubiquity of a product. And then you can ask 0:12:12.000 --> 0:12:17.559 yourself the question, Okay, on average, how ubiquitous are the 0:12:17.600 --> 0:12:21.400 products that this country makes? That is this country making 0:12:21.520 --> 0:12:24.840 mostly things that are simple to do that everybody knows 0:12:24.880 --> 0:12:26.920 how to do, or or they make things that are 0:12:26.960 --> 0:12:29.240 hard to do, things that are done in few places. 0:12:29.880 --> 0:12:32.439 So that's kind of like a different cut out of 0:12:32.440 --> 0:12:36.360 the data. And you simply repeat this process an infinite 0:12:36.480 --> 0:12:40.160 number of times and it generates an algorithm that ranks 0:12:40.240 --> 0:12:43.440 both the countries in terms of how complex they are 0:12:43.880 --> 0:12:46.480 and the products in terms of how complex they are. 0:12:46.960 --> 0:12:50.160 So essentially, it's an operation on this matrix if you 0:12:50.200 --> 0:12:53.360 want that relates countries to the products that they make. 0:12:53.520 --> 0:12:57.240 It is an eigenvector of that matrix. Essentially, if you 0:12:57.240 --> 0:13:00.640 want to get fancy at the mouth, but it captures 0:13:00.679 --> 0:13:02.760 this idea how many things are we able to do 0:13:02.840 --> 0:13:05.360 and how complicated it is to do those things. 0:13:21.960 --> 0:13:26.800 So is it possible to have a good economic outcome 0:13:27.280 --> 0:13:31.040 with low complexity? Or I guess another way of saying 0:13:31.080 --> 0:13:35.480 this is is complexity as synonym for development or wealth here? 0:13:35.840 --> 0:13:39.000 Or capturing something different? Could you have a country that 0:13:39.040 --> 0:13:43.679 scores low on complexity but is still a relatively good 0:13:43.840 --> 0:13:46.400 place to live from an economic perspective. 0:13:47.360 --> 0:13:52.560 So one major thing that economic complexity does not capture 0:13:52.800 --> 0:13:57.680 is natural resource wealth. Okay, so when you make a graphs, say, 0:13:57.720 --> 0:14:00.800 of the relationship between how complex you are are and 0:14:00.880 --> 0:14:04.520 how rich you are, the outliers tend to be countries 0:14:04.600 --> 0:14:08.880 that are natural resource rich. So they get their income 0:14:08.960 --> 0:14:11.880 not so much by what they know how to do, 0:14:12.080 --> 0:14:15.720 but from the natural resource wealth that they happen to have. 0:14:16.120 --> 0:14:19.440 So a place like the United Arab Emirates or Saudi 0:14:19.480 --> 0:14:24.320 Arabia perhaps might score low on complexity, but they have 0:14:24.360 --> 0:14:27.480 a ton of oil wealth and so they're still quite 0:14:27.520 --> 0:14:30.600 prosperous on a GDP per capita basis that sort of 0:14:30.600 --> 0:14:32.520 thing exactly. 0:14:32.600 --> 0:14:36.000 So, or you could say, tell me, controlling for their 0:14:36.080 --> 0:14:39.360 natural resource wealth, they have a lot of complexity or not. 0:14:40.200 --> 0:14:41.800 And let me tell you a few of the things 0:14:41.840 --> 0:14:45.120 we have established. The first one is that the complexity 0:14:45.240 --> 0:14:48.720 correlates very highly with how rich you are, and if 0:14:48.720 --> 0:14:52.760 you control for your natural resource wealth, it correlates even better. 0:14:52.840 --> 0:14:56.560 So there's a very strong relationship between economic complexity and 0:14:56.640 --> 0:15:01.120 natural resource and your GDP per capita in level. So 0:15:01.160 --> 0:15:04.280 that's a very strong relationship. But more importantly than that, 0:15:05.560 --> 0:15:10.239 countries that are more complex than you would have expected 0:15:10.280 --> 0:15:13.800 them to be given their income level tend to grow 0:15:13.840 --> 0:15:17.360 faster in the future, and countries that have relatively low 0:15:17.400 --> 0:15:21.720 complexity relatives to their income level tend to grow less 0:15:21.800 --> 0:15:26.040 in the future. So your economic complexity relative to your 0:15:26.040 --> 0:15:28.880 income level is a predictor of how fast you will 0:15:28.920 --> 0:15:31.680 be able to grow. And typically when we look at 0:15:31.720 --> 0:15:34.360 how good these predictions are, they tend to be best 0:15:34.680 --> 0:15:37.640 at a horizon of about ten years, So it tells 0:15:37.680 --> 0:15:40.760 you something about what your next decade is likely to 0:15:40.880 --> 0:15:41.280 look like. 0:15:41.920 --> 0:15:45.680 What's an example of maybe something in history in which 0:15:45.720 --> 0:15:50.560 at some point a country scored significantly high complexity and 0:15:50.960 --> 0:15:53.480 then did it over the next several years. It's like, yeah, 0:15:53.480 --> 0:15:56.000 this country really boomed. What's an example that stands. 0:15:55.680 --> 0:15:59.360 Out of this, Well, there are many examples, but suppose 0:15:59.440 --> 0:16:02.160 if you have take the picture in two thousand and eight, 0:16:02.440 --> 0:16:07.400 two outliers were India and Greece. India had extremely low 0:16:07.440 --> 0:16:11.200 income levels for its level of complexity, and Greece had 0:16:11.240 --> 0:16:14.880 extremely high income levels for their level of complexity. And 0:16:14.960 --> 0:16:18.400 what happened is that India has been the fastest growing 0:16:18.480 --> 0:16:22.080 large country in the world since then, and Greece collapsed. 0:16:22.680 --> 0:16:28.040 So we also see countries that increase their complexity initially 0:16:28.080 --> 0:16:31.200 by moving into a new set of products. In the 0:16:31.240 --> 0:16:34.640 case of Thailand, they first moved into garments and spent 0:16:34.800 --> 0:16:37.680 a whole decade adding a bunch of different garments to 0:16:37.720 --> 0:16:41.040 their export baskets. Suddenly they got into electronics and then 0:16:41.080 --> 0:16:43.760 started to add a bunch of electronics to their export basket, 0:16:43.800 --> 0:16:46.720 and then they got into cars and machines and so on. 0:16:47.120 --> 0:16:51.040 So they have been increasing their complexity and have very 0:16:51.560 --> 0:16:54.880 sustained high growth rates in the process. In general, what 0:16:54.960 --> 0:16:58.400 we find is that only about a fifth of the 0:16:58.480 --> 0:17:02.280 countries that were poorer than the US, say, in nineteen 0:17:02.360 --> 0:17:06.960 seventy have really caught up narrowed the gap with the US. Okay, 0:17:07.280 --> 0:17:12.720 since nineteen seventy onwards, only twenty percent of countries narrowed 0:17:12.760 --> 0:17:15.720 their income gap with the US. Those twenty percent of 0:17:15.760 --> 0:17:19.560 countries that narrow their income gap with the US increase 0:17:19.840 --> 0:17:24.199 their complexity very significantly. The other eighty percent did not. 0:17:25.119 --> 0:17:29.479 So I would tell you that sustained growth implies this 0:17:29.640 --> 0:17:35.800 process of absorbing knowledge, distributing in a your society, mobilizing 0:17:35.840 --> 0:17:38.800 that knowledge to make more things and more complex things. 0:17:39.080 --> 0:17:42.080 Because you get more complex. Things are essentially things that 0:17:42.119 --> 0:17:45.920 require more knowledge, and things that require more knowledge require 0:17:46.040 --> 0:17:49.840 deeper networks of humans collaborating, whether it's in a single 0:17:49.920 --> 0:17:52.040 firm or in a longer value chain. 0:17:52.359 --> 0:17:54.720 Can you talk a little bit more about how you 0:17:54.800 --> 0:17:59.760 build complexity and diffuse that knowledge within a society, because 0:17:59.760 --> 0:18:03.159 I am imagine if you're a developing country, maybe you 0:18:03.280 --> 0:18:07.840 find that you have a competitive advantage in one type 0:18:07.840 --> 0:18:09.159 of thing. I'm going to go back to the T 0:18:09.280 --> 0:18:12.560 shirt example. You can make T shirts cheaper than anyone 0:18:12.640 --> 0:18:15.520 else and more efficiently, I guess, And then I would 0:18:15.520 --> 0:18:20.760 imagine the temptation is to just stick with that specialization 0:18:21.440 --> 0:18:24.480 and just do the thing that you are currently really 0:18:24.480 --> 0:18:27.439 good at. So how do countries actually break out of 0:18:27.480 --> 0:18:31.560 that dynamic and start developing expertise in other areas. 0:18:31.920 --> 0:18:35.560 There are essentially three mechanisms that I would like to mention. 0:18:36.480 --> 0:18:40.800 The first one is that countries tend to move from 0:18:40.840 --> 0:18:43.879 where they're good at to what I like to call 0:18:44.280 --> 0:18:49.720 or Stuart Kaufman coin the phrase the adjacent possible. When 0:18:49.720 --> 0:18:53.160 we look at countries adding products to their export basket, 0:18:53.240 --> 0:18:57.440 those products tend to be cognitively near. If you want 0:18:57.560 --> 0:19:01.280 the products that they were making before. And one of 0:19:01.320 --> 0:19:05.600 the contributions we've made is we've developed a technique to 0:19:05.800 --> 0:19:11.159 measure this cognitive proximity for all the products in the world. 0:19:11.240 --> 0:19:14.280 And you can locate every country in the world and 0:19:14.400 --> 0:19:19.400 find out what's in their adjacent possible So countries tend 0:19:19.440 --> 0:19:22.000 to move from the things that they are currently good 0:19:22.040 --> 0:19:24.639 at to things that are in their adjacent possible And 0:19:24.720 --> 0:19:27.520 we call this cognitive map of the products of the world. 0:19:27.520 --> 0:19:30.800 We call it the product space. And this product space 0:19:30.880 --> 0:19:34.040 is very heterogeneous. There are some parts of the product 0:19:34.040 --> 0:19:37.080 space there where you have products that are tightly connected 0:19:37.080 --> 0:19:39.040 to each other. So if you know how to make 0:19:39.080 --> 0:19:41.640 one kind of product is kind of like easy to move. 0:19:41.840 --> 0:19:44.959 You have a rich adjacent possible. You have many ways 0:19:45.000 --> 0:19:48.800 of reorganizing that knowledge to make other things. We like 0:19:48.840 --> 0:19:52.560 to use the metaphor that products are like trees, and 0:19:52.680 --> 0:19:55.880 firms are like monkeys. They live on trees, they exploit 0:19:56.000 --> 0:19:59.680 certain trees. So the product space is like the map 0:19:59.680 --> 0:20:03.160 of the forests. And so you can, by the way, 0:20:03.160 --> 0:20:05.280 if you go to the Atlas and Economic complexity, we 0:20:05.320 --> 0:20:08.160 have the we choose a country, we have the product space. 0:20:08.200 --> 0:20:12.800 We will tell you where in that forest, does this 0:20:12.960 --> 0:20:15.760 country have its monkeys, and then it can tell you 0:20:15.800 --> 0:20:20.240 know which trees are close to those monkeys, and it 0:20:20.280 --> 0:20:23.000 can tell you so you know, what are other characteristics 0:20:23.040 --> 0:20:26.240 of those trees that might make it sexier or less 0:20:26.240 --> 0:20:29.520 sexy to move in that direction. Okay, so the first 0:20:29.520 --> 0:20:32.159 thing I want to say is that countries tend to 0:20:32.400 --> 0:20:36.359 diversify by moving to their adjacent possible depending on where 0:20:36.359 --> 0:20:40.000 they started, and not every country starts with the same 0:20:40.040 --> 0:20:42.399 deck of cards. They don't start with their monkeys in 0:20:42.440 --> 0:20:45.760 the same place. Some countries start with their monkeys in 0:20:45.960 --> 0:20:51.399 very very promising parts of the product space because their 0:20:51.600 --> 0:20:54.480 trees are very closely connected to each other there, so 0:20:54.760 --> 0:20:57.760 it's easy for the monkeys to jump from tree to tree, 0:20:58.040 --> 0:21:00.720 and other parts of the product space are very sparse 0:21:01.080 --> 0:21:03.199 with their trees are very far from each other, so 0:21:03.240 --> 0:21:06.080 it's hard for those monkeys to move. Okay, So that's 0:21:06.359 --> 0:21:12.199 mechanism one, move towards the adjacent possible. Mechanism two is 0:21:12.240 --> 0:21:15.119 that you have to solve this chicken and neck problem. 0:21:15.760 --> 0:21:18.359 You don't know how to do the things you don't do, 0:21:19.640 --> 0:21:21.600 but you need to know how to do things to 0:21:21.680 --> 0:21:25.080 start doing things you are not doing before. So you 0:21:25.240 --> 0:21:29.000 need watchmakers to make watches. But how do you become 0:21:29.000 --> 0:21:33.360 a watchmaker in a country that doesn't make watches. How 0:21:33.400 --> 0:21:36.880 do you solve this chicken and egg problem. We think 0:21:36.920 --> 0:21:39.600 that this solving the chicken and eck problem is the 0:21:39.680 --> 0:21:43.640 thing that forces countries into moving just to the adjacent 0:21:43.680 --> 0:21:47.040 possible because it's hard for them to solve too many 0:21:47.080 --> 0:21:49.440 of these chicken and egg problems at the same time. 0:21:49.920 --> 0:21:53.120 But one way to accelerate the solution of these chicken 0:21:53.160 --> 0:21:57.720 and eck problems is to bring watchmakers from outside your country. 0:21:58.160 --> 0:22:00.960 That is, you may not have what makers in your country, 0:22:01.000 --> 0:22:03.520 but maybe you start with a group of Swiss watchmakers 0:22:03.640 --> 0:22:06.360 and they'll train the next generation of watchmakers and now 0:22:06.400 --> 0:22:11.760 suddenly you do watchmaking right. So migration plays an outsized 0:22:11.880 --> 0:22:16.280 role in diversification because you need to add knowledge that 0:22:16.400 --> 0:22:18.959 was not in the system before. So if you can 0:22:19.000 --> 0:22:21.399 attract people that had knowledge that was not in the 0:22:21.440 --> 0:22:25.280 country and you can engage them having worked there and 0:22:25.440 --> 0:22:29.000 have that knowledge spread, that seems to be very important. 0:22:29.280 --> 0:22:33.240 There's a very nice story about Bangladesh here where you know, 0:22:33.240 --> 0:22:36.200 if you look at Bangladesh in the ATHLETs of economic complexity, 0:22:36.200 --> 0:22:38.679 you'll tell you that ninety percent of their exports are 0:22:38.760 --> 0:22:42.640 garments and that they all started in the eighties. These 0:22:42.680 --> 0:22:45.800 exports of garments, Well, what underpinned that was a company 0:22:46.320 --> 0:22:50.200 which was called Desh and that company sent one hundred 0:22:50.240 --> 0:22:52.840 and twenty six of its workers for six months training 0:22:52.880 --> 0:22:56.359 program in Korea. Because the program that company was created 0:22:56.400 --> 0:23:00.280 by Dai Wu. So these guys went to Korea. Yeah, 0:23:00.359 --> 0:23:04.160 trained in Korea, came back, started the company and started 0:23:04.200 --> 0:23:08.480 to produce. Fifty six of those people left the company 0:23:08.960 --> 0:23:13.200 to create their own startups, and those fifty six children 0:23:13.400 --> 0:23:17.480 of this company DSH are the core of the export 0:23:17.680 --> 0:23:21.679 industry of garments in Bangladesh. So in some sense you 0:23:21.760 --> 0:23:25.600 have to infect the system with knowledge and assure that 0:23:25.640 --> 0:23:28.240 the mechanisms are going to allow that knowledge to spread. 0:23:29.080 --> 0:23:32.199 Tracy prefaced the question. She talked about a country that 0:23:32.240 --> 0:23:34.200