Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable. ======================================================================== # Tapping into Scientific Talent: The Rise of Corporate and Public R&D in Japan's Chemical Industry Authors: Takuya Hiraiwa Discussant: None Video: https://www.youtube.com/watch?v=7LDwpGoPNFc&t=18972s ## Talk (05:16:12 – 05:34:43) [05:16:12] >> Okay. Yeah. Okay. Um, thank you so much for including my paper in the program. [05:16:17] I'm Takuya Higa from University of Maryland. So, um, I want to start my talk with this. So, our economy is relying on a broad range of imports from other countries. But on the other hand, we can easily lose access to those [05:16:31] imported item all of a sudden because of global tensions. I think of for example, China's export controls on the rare earth elements or how Russian invasion of Ukraine caused the global shortage of [05:16:43] key imported items. Events like these often created urgent need to create a new technologies to cope with the lack of imported product. uh and empirically we know that innovation efforts also do [05:16:56] get redirected where toward where the necessity arises. So that's a whole directive technical change of literature and uh it is exactly what we usually mean when we say necessity is a matter of invention. But when it comes to [05:17:11] whether we actually succeed in the immediate response to that necessity, the supply side really matters like Rosenberg described technologies cannot be developed in the party fashion. So [05:17:23] it's not like uh okay we face a huge demand and then okay so we now make our R&D investment that's how we get actually immediately automatically the required technology but that's not how [05:17:35] it works. uh what we actually need is the talent the supply of talent capable of generating uh new required technologies immediately when they face the do and emergence of demand and what [05:17:50] we need to care about is that how we actually deploy that talent in the long run so I want to start at the one concrete example so this is Jin Warren who actually is in the University of Minnesota who developed the world's [05:18:05] first rare Earth's free permanent magnets. His work actually got a lot of attention after 2010 rare earth crisis when the China imposer export control on the rare earth elements for the first [05:18:18] time and uh it is actually getting attention now because of grow continued global attention right now. So and the department of energy also uh launched the rare earth alternative R&D grant program after 2010 to deal with this and [05:18:32] then as a result this brought about more than 100 related patent uh since 2011 and then he also founded his own company to commercial his own technologies. But here is a point I want to emphasize. So [05:18:46] there are two point first uh he already did this basic research even before the crisis hit before the 2010. Uh he had been working on uh this as a scientific puzzle and then his key publications [05:19:00] about this research came out just right before the crisis hit. So the response to that necessity wasn't created out of nothing. uh it was tapping into the knowledge that had already been [05:19:14] accumulated before the crisis. Moreover, another important observation is the the spot for this line of work kept growing as well after even after the original crisis. So, and the it's uh it [05:19:29] eventually moved toward the uh beyond the original target towards different applications as further research was also carried out in the long run. So uh this paper asks two question related to this observation. First when the demand [05:19:44] for new technology is so urgent, how important is xanti admiration of scientific talent for respondence to that necessity and second how does that necessity affect long-term innovative [05:19:58] output of those talented individual. The purpose of this paper is to try to answer these question by utilizing a historical context of Japan's chemical [05:20:10] industry at the turn of 20th century. Uh I collected detailed individual level data of the census of all university educated engineers and scientists tracking their research outcomes and career histories. Okay, I will explain [05:20:24] the detail of the data in a moment. But why do we need to care about this context? First uh Japan had experienced a sudden import disruption during World War I period. Before the wartime, Japan had rely on the imports from Europe [05:20:38] mainly from Germany and the UK for key advanced chemicals. And once war hit um even though Japan didn't engage in the war by itself uh the import actually was [05:20:52] suddenly shut down and then there were actually no domestic technological base in Japan at that time in the chemistry field and this is a one of the most serious instances instances of the sudden necessity due to imported disruption. Here I want to just show how [05:21:06] severe this works. On the left you can see the share of the imported chemicals and on the right you can see the price index of those chemical uh import and domestic chemicals. Uh as you can see the both moved sharply right at the [05:21:20] onset of world one imported chemical craps while the price soared. But again the fundamental problem was Japan didn't have any domestic technological base at that time in chemistry field. So this created the urgent need to develop new [05:21:35] domestic technologies. Another reason why we this setting is pretty nice is Japan didn't have any robust chemical industry but on the talent side they already established the [05:21:48] frontier higher chemistry education system by the end of 19th century. Uh so this is uh Imperial University that were the first modern higher technical education system in Japan. uh the those [05:22:02] universities were training many chemists by the European standard and some of those chemists are already working on some basic research even before the chemical industry took off. So the important question is how did they [05:22:16] contribute when the sudden crisis hit when they actually got the important disruption that is exactly what I want to examine here. [05:22:24] So uh let me quickly go through the data structure. So the backbone is the imperial university registry which provides a list of universal university graduate from science and engineering departments. Uh this provides the name [05:22:39] graduate names and divisions and graduation years and conveniently they are listed in the order of their academic ranking. So effectively we can utilize a measure of their academic [05:22:52] performance and then link those data with imperial university alumni surveys uh which report each graduate employer and address on an annual basis. So [05:23:03] combining these it uh I obtained the full panel of the this graduate employer employee matched data set and then for measuring the innovative output I utilize the patents and publication [05:23:17] records. So patents data include all the patents uh granted during this period the publication data include the articles published uh in the two of Japan's top chemistry journal at that [05:23:29] moment. And finally some complimentary data set uh which include uh some accounting and information of chemistry chemistry related companies um the risk of universe of research laboratories and [05:23:42] institutions in Japan at that time and then try the statistic and some income or earning information for some chemistry graduate student. [05:23:51] So uh let me begin with the big picture first. Uh this is showing the patent productivity. So the number of patents per graduate over time separatory for chemistry fields other high-tech fields [05:24:05] related to the second industrial revolution technologies and the non-hightech fields before the war chemist patenting by chemistry graduate uh looks pretty similar to other high-tech fields but uh like at the [05:24:18] onset of world war chemistry patenting shot up in response and moreover even though imports from Europe came back Right after the end of the world war uh chemistry patent difference between the chemistry patenting and other field [05:24:33] didn't cross chemistry graduate kept patenting even at much higher rate even after the war. But this is just the average response uh to this kind of import disruption. So the question is how much the pre-wall accumulation of [05:24:48] scientific knowledge contributed to this average trend. [05:24:52] So to examine this I first split chemistry field graduate into two groups based on their pre-war scientific publications whether or not they had a publication in the top chemistry journal [05:25:05] before 1914. Uh so I use this as a measure for uh their whether they had been engaged in the frontier scientific research in chemistry field before the war and then [05:25:18] I estimate the individual difference in differences regressions using pre914 graduate and compare the graduate from chemistry divisions with other graduate especially limiting the control [05:25:31] comparison groups to other high-tech divisions to try to pass out the general trend associated with the secondary industrial liberation technologies and here oh [05:25:43] okay here's a uh main result uh for chemistry graduate uh with without any publication before the war there are basically not nothing going on here so um there wime [05:25:57] coefficient and post war coefficient are very close to zero and insignificant but the chemistry graded with pre-war publication the picture is completely different So they actually increased the patenting significantly during the war [05:26:10] time. The vector is actually larger than their average patent rate and then moreover they also kept uh increasing patenting after the end of the war. Uh and lastly doubling relative to the wartime effect. So you can this trend [05:26:25] clearly in uh the event study plot here. So the black line is the effect for those with fabrication while the light line shows the uh those without pre fabrications. Uh so the main conclusion [05:26:39] here is that uh uh we need actually the people who have been engaged in the scientific activity before crisis hit to respond to the crisis. So in other words, the equally educated chemistry graduate who didn't have any experience [05:26:53] on the scientific work couldn't meaningfully create a response to the this import disruption crisis. And here uh quickly a few things I've checked the first you may wonder obviously uh whether this just preo publication just [05:27:06] proxy for the general ability or academic status but fortunately we have the data on the academic ranking and whether they went on the doctoral degree uh in the future. So and the the main [05:27:18] point of the uh results is publication record still matters even after controlling for uh accounting for these factors. So pre scientific activities reflect not just the general academic [05:27:31] ability or academic status but also something different. I also look at the whether increasing in patent came from uh new graduate uh who enter into the market after the war time after after [05:27:46] the onset of the world war and the short answer is uh in this case incumbent patent productivity was generally higher than them and then they contributed more to the response to the crisis. [05:28:00] Okay. But the remaining question is why did we see the persistent productivity difference uh between the different field because think about it uh import [05:28:13] from Europe now came back after the war and um effectively there was no longer urgent need to create new technology in chemistry field but invention burst didn't go away. So that brings me to the [05:28:27] second question which is what sparted long-term boost of the invention. Uh what I found here is that the war also triggered the emergence of long live [05:28:40] research assets. Uh firms or the government also started to building their own research oriented organizations such as corporate laboratories or public chemical research institute. On the right you can see the [05:28:53] number of corporate laboratories. uh on the right you can see the number of public chemical research institute both types of the institutions uh take took off right uh during wartime or right after the onset of world war one and [05:29:07] then the important thing is those organization didn't go away even after the end of the war and meanwhile talent also moved into those newly emerged organizations so on the graph on the right side so you can see the share of [05:29:22] the graduate in niche sector over time and this middle line green line actually shows the research sector and you can see the research sector growing right after the onset of world war and this [05:29:35] happened in two ways. The first existing organization also started to build their own research laboratory and move people into the research organizations or second incumbent workers also change their jobs to move into the research at [05:29:50] other organizations. But anyway crisis restruation of scientific talent to response to the crisis matching people with newly emerging long-term research asset. So this is a key interactions what I can [05:30:04] hear find here. So uh I did the same exercise as before but uh now I split the individual by whether they end up with in the research jobs after 1915 and [05:30:19] then on the left you can see the effect for those without no pre-wab without pre-wall fabrication and if they move into the research sector uh yeah they slightly improve the pattern right after [05:30:33] the war but this came quite Wait. On the other hand, uh on the right side, the for those with pre-war uh publications, if they didn't move to the research sector, there's nothing actually going [05:30:47] on. So, they didn't produce anything. But if they did move to research sector, uh their effect is actually significant and much more than other people. So, uh I want to be clear that this is just a [05:31:00] uh result of selection. So and it's kind of the description of where you can find the postwar invention booth right but still it's informative about mechanism. [05:31:09] So we can find the postwar invention booth only where the examant scientific talent are matched with the complimentary research asset that emerged after the war. So uh finally I also look at the direction of invention [05:31:24] in the long run because in the short run it's obvious that it's uh directed toward uh import substitution purpose right but the long run so we are curious about whether it's just a continuation of world one follow on patent or [05:31:38] invention the short answer is actually not so I also divide the technology subourses and groups into whether uh any patents are filed during wartime. So the [05:31:51] this uh black green line actually show the wartime technologies. So especially at the very final level this wartime technology uh is uh even slightly decline after the wartime uh and then [05:32:05] those technology area uh that have little to do with the wartime technology actually exploded after uh 1920s. [05:32:13] So uh in the long run so the direction of the research actually branched out into the different areas and this while the short run actually they are all all actually import substitution. Okay, finally let me uh describe how I [05:32:28] actually think about this because uh this pattern doesn't easily come out of the standard models of demand to innovation like a directive technical change or typical standard infant [05:32:40] industry argument or import substitution story because I mean that predict uh the increase in the patenting in the uh area that experience a demand shock and they should follow the same technology area in the future. [05:32:55] But that is not exactly what we are seeing here right uh so um I actually proposed some kind of model inspired by the calendar and the connearch over the browning p so obviously I don't [05:33:10] have enough time to explain all the detail of the model but the main idea key idea is the combination of extern knowledge uh scientific knowledge plus the world research experience works as a [05:33:23] springboard for the new directions of technologies temporary water and premium induced the formation of knowledge as in direct technical change but only those who had worked on related scientific [05:33:36] puzzles uh could meaningfully respond to the world one crisis so that is consistent with my first finding but the successful wartime research reduced the uncertainty of subsequent research even [05:33:50] in the other other field and then that also expanded research possibility frontier beyond the wartime technology area and encourage further research in the subsequent period. Okay. So that's a [05:34:03] that's all and then so main per main uh message from this is that uh okay so to deal with sudden uh crisis that happened uh caused by the import disruption. So it's important to have the extent human [05:34:16] capital and to obtain some long-term response. So we need to support by the long research in emergence of long research organization or organization innovations. So without knowled we can really meaningful respond to that [05:34:30] crisis. But what actually bring a long-term effect is actually organizational response. Okay. Thank you so much. ## Q&A (05:34:43 – 05:35:58) [05:34:43] I think we probably only have time for one question. Uh >> so in what directions did they go they [05:34:54] patent in fields that basically now with the English or the Germans from >> yeah um do you have a sense >> yeah I think so uh Japanese patent [05:35:08] system actually also works as same as the US patent system so they actually need to recognize some no in that it's distinct ative from the Germany or English technology. So it's kind of more [05:35:19] domestic new technology area. So yeah >> the new stuff and also patent in the US. [05:35:29] >> Yeah, some of the patents. So they are patenting globally. So I can also observe that in the I mean global whether uh Google patent so they actually have the information about whether they are patent in the other countries. Yeah, some good patent actually patented in other area but the [05:35:43] pretty minority of them actually work. I think it's around the 10% of the patenting. So in that sense, we can actually measure the quality but uh yeah I didn't do some cord as here. Yeah, we [05:35:53] can actually do that. Yeah. >> Okay. Uh we have a 20 minute break and we'll come back at 250.