Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=Transcript of the talk, from the video's captions. Auto-generated: speaker names in particular are unreliable.
=

# Foreign Political Risk and Technological Change

Authors:   
Discussant: None  
Video: https://www.youtube.com/watch?v=ix_kYmhkQKk&t=0s


## Talk (00:00:00 – 00:23:35)

[00:00:00] Um so thanks and welcome everybody. Uh thanks for being here to our first session on geopolitical spillovers. We are running a few minutes late so it's not all German punctuality here. It's I  
[00:00:13] hope that is okay. Um given you had to run all up all these flights of stairs.  
[00:00:18] Um we will now start with the um first um presentation by Jacob Moscona. I hope it pronounces it more or less correctly on foreign political risk and technological change. Uh you will have  
[00:00:32] 20 minutes then we will have a 10-minute discussion by B OA who is um on the on a Zoom call and I hope he can hear as well and then we'll have a 10-minute Q&A session. So Jacob,  
[00:00:46] >> thank you. Well, thanks all for being here. Thanks for including this paper in the program. I'm excited to present this paper foreign political risk and technological change joint with Joel Flynn and my woe at Yale and Antoine Levy at Berkeley.  
[00:00:58] I think they want to click on this. >> Yes. So, it's no surprise that political risk, international political tension can cause major damage to productivity and well-being. Both because it can escalate into actual conflict, but also  
[00:01:12] because it potentially threatens access to key inputs, resources, goods that are essential for production. And while there's growing interest in modeling how this political tension emerges and how restrictions to good things like trade sanctions emerge, there's relatively  
[00:01:25] little evidence about how firms and countries react to these threats impract. And one prominent possibility, which is what we study in this paper, is how innovation and technological progress reacts to foreign political threats, potentially reshaping and  
[00:01:40] determining their eventual economic. And there are a range of different examples of this. I'm going to start with one from the 2014 version of Elon Musk who when testifying before the US Senate justified major investments in US space  
[00:01:53] basically saying that there's this rising political threat in Putin and Russia and at the moment we're relying on the Russian government for all of our space shuttles and there's the potential that that access gets shut off as the US and Russia fall increasingly into into  
[00:02:07] conflict and so it's really important to invest in domestic capacity, domestic technology for you Fast forward 10 years. This came up already in the introduction, but in response now to rising political tension in the US,  
[00:02:22] political risk and uncertainty in the US and potential threats to access of European access to things like defense technology, there's a real push to invest in innovation and technology and defense industries in Europe. So this is what we're trying to study in this paper  
[00:02:35] in particular. Is it actually true systematically across the board that innovation reacts to foreign political risk? And if so, what are the mechanisms underlying that response? How can we sort of um explain it? And second, if so, sort of what are the broader  
[00:02:49] economic consequences of that reaction of technology to these foreign threats to production? So, in the paper, we start with a model that I'm not going to present today, but it's a model of sort of firm level exposure to changing foreign political risk and endogenous  
[00:03:02] innovation decisions and then sort of trade uh emerges in equilibrium. And the first key prediction is that greater exposure to foreign political risk increases innovation and that's especially going to be driven by firms with middle levels of productivity who  
[00:03:16] might you know access inputs from abroad when we're in this good state of the world but still want the ability to continue producing even when access to those inputs get get shut up and that's something we can actually test in the data and second maybe unsurprisingly is  
[00:03:29] that when this innovation response is greater that's going to reduce your reliance on those risky foreign goods so it's going to allow countries to better adapt to these changing exposure to foreign So what I'm going to start talking about  
[00:03:43] today is sort of how we take this question for the data. And so the thing we need to do is compile a whole range of different data sources. First on sector level data on technology development and investment and innovation. We're going to put together global data on patenting and R&D  
[00:03:57] investment link to the sectors that are actually making those investments and developing those technologies. And we we're going to link that with both patterns of trade and country byear measures of political risk and political firmware that we're going to use to construct a measure of each country and  
[00:04:11] sector exposure to foreign political risk. And I'll talk about this much more in detail in a second. And the main results are basically going to be three-fold all kind of conveying that foreign political risk plays a major role in shaping innovation, technology  
[00:04:25] development. First is that greater sector level exposure to foreign political risk substantially increases innovation, both investment and innovation and actually new patenting, new technology getting developed. That's going to be particularly strong when  
[00:04:38] risk emanates from geopolitical adversaries, countries where you might actually think this threat of cutting off access to goods or access to trade is most relevant or most likely to bite.  
[00:04:48] And finally, I'm going to show evidence that this technological response actually mediates the consequences of foreign political risk. In particular, this innovation response actually does reduce the reliant, you know, countries reliance on risky foreign input as foreign countries become more risky. But  
[00:05:03] at the same time, sort of interestingly to us, that really exacerbates the consequences of political risk for those countries that experience. So now you have this period of of political risk, political turmoil. Not only do you have to experience the direct effects of that, but now effectively the rest of  
[00:05:17] the world is innovating you out of the trade network. And so even if you emer reemerge from that episode of political risk, all this innovation has happened abroad that has sort of made you less relevant.  
[00:05:27] Okay, jumping into the data and measurement. So I already mentioned three or you know a couple sources of data we need. First we're going to put together data on innovation. The main source is going to be the universe of all patents filed in the United States  
[00:05:40] which we're going to um link to um so where we know information about the technology class of that patent the inventor particular which country the inventor is located in and a whole range of other sort of patent characteristics that that I'm sort of talk more about if  
[00:05:54] there are questions. We're going to have two units of analysis in the paper. The first where I'm going to show you a sort of brief case study is about uh critical minerals that also already came up in the introduction. So we're going to do one part of the analysis the level of critical minerals and we're going to  
[00:06:07] link patents to the individual minerals basically doing a keyword search to identify all the new technologies related to each of these minerals and then we're also going to do an analysis kind of across all traded sectors where we're going to again link patents and  
[00:06:20] now sectors using existing methods for for um linking patented technology classes the tradeable sector. We have other ways of measuring innovation if you don't like the patent data that as well if there are questions.  
[00:06:33] The second thing we need is a measure of country level time varying political risk which we take from the international country risk guides. This is going to be great because it's really longunning. So it goes back to the 1980s and it covers a really broad set of countries including things like  
[00:06:47] expropriation risk, conflict risk, risk of rising religious tension. That's going to be our main measure. We can also validate that in a couple different ways and replace that measure with other things that exist for shorter time periods or a smaller set of countries and everything look pretty pretty  
[00:07:00] similar. Finally, sort of for that piece of the analysis on on sort of allies versus enemies, we're going to put together a couple different ways of measuring political alliances across countries kind of at the country pair level, both from the coralates of war project, but also using sort of revealed  
[00:07:14] relationships across countries using voting data. And finally, it's going to be sort of um an ingredient a range of our measures is is trade data. So, at the sort of ndigit level, we're going to have sort of data on all trades.  
[00:07:28] Main thing we want to measure for this analysis is a measure of sort of political risk exposure. First the f first part of the analysis is just focused on the US. So in this case for a given sector and year. Oh sorry.  
[00:07:41] Should I keep going? Sorry. No no just yes.  
[00:07:44] >> Great. Okay. And that's basically going to be we're going to for for a given so focusing on the US say oh that's what in words we're going to sum across all foreign countries interacting the risk  
[00:07:59] level of that country. So the risk level of country C in time t with your that sector's exposure to that country. So in the case of minerals we're going to define each mineral's exposure to a given country is just the deposit share  
[00:08:13] of that mineral in that country. for across sort of all traded sectors, we're going to use a measure of sort of pre-p period uh imports from that country C in sector I to the sort of focal country in question.  
[00:08:26] And then uh yeah, so that's going to be our main measure of sort of time varying exposure of each sector uh to political to foreign political risk in each year.  
[00:08:34] Basically using sort of time varying measures of political risk across all foreign countries and some baseline measure of each sector's sort of exposure to that to that country. Um okay so first sort of warm up uh we're  
[00:08:48] going to really zoom into this topic of uh minerals and critical critical mineral risk. So again I already came up the introduction but minerals are have become central inputs to many modern technologies and sort of at the same  
[00:09:02] time are concentrated at least the um uh mining of these minerals are concentrated in many politically volatile regions. So the first question we're going to ask is sort of does mineral level innovation now I is going  
[00:09:14] to index minerals respond to foreign political risk using the following estimating equation which links innovation this is patenting about mineral I and time t to the political risk exposure of that mineral with mineral fixed effects capturing changes  
[00:09:29] in say demand for a given uh uh mineral because they could be differentially useful for different production processes and time fixed effects capturing the fact that you know a couple uh days ago when China threatened to restrict sort of global access to its  
[00:09:43] minerals. That's kind of a time level shock that could really affect risk across the board. I mentioned we're going to measure this political uh risk to minerals summing across countries waiting political risk in that country with the deposit share of that mineral  
[00:09:58] in that country and beta captures the response of innovation to form political risk.  
[00:10:04] So even in this sort of specific case study where we can do a really good job linking innovations to individual minerals and tracking how that responds to risk, we see a big response even at the annual frequency of patenting to  
[00:10:18] mineral risk. If we aggregate everything the decade level trying to capture the fact that this innovation response can take time and maybe especially responsive to longer run changes in risk get an even sort of larger um estimate by about sort of 50%. And then  
[00:10:33] importantly, there's sort of no pre-existing trend. So you're not reacting to sort of uh that there isn't a relationship between patenting and future changes in political risk, which suggests both that sort of the firms doing this innovation aren't anticipating this risk, but also that  
[00:10:47] there's no sort of pre-existing trend in this relationship. And the effect is large. So a one standard deviation change in political risk leads to about a 30% increase in patenting as we measure it. And things look similar if you citation weight these patents. You do a range of different techniques for  
[00:11:00] measuring actually the quality of Next question is so that's one specific area of production one really important sort of set of inputs but does this generalize does a similar relationship  
[00:11:12] exist across all traded sectors so basically to answer that question we go back to the same estimating equation except now I indexes all nakes six-digit sectors and because we don't have this sort of convenient geographically fixed  
[00:11:27] deposit share to weight foreign political risk by in this sum we just use the import share from uh each country uh to the US at baseline and sort of sort of sum across across  
[00:11:41] all countries here. So when PR is higher or increases it means the political risk in the countries on which you rely for that particular sector goes up.  
[00:11:51] Okay. And then again sort of beta in this regression is going to capture the effect of foreign political risk on on innovation.  
[00:11:59] So this is like a big table. First, ignore panel A, focus on panel B for a second. This is the one measuring sort of just focusing on that fixed pre- period import shares in the measure. You see an increase in the total number of patents. Actually a pretty similar  
[00:12:13] coefficient estimate implying a pretty similar magnitude to the previous result. So about a one standard deviation increases a change in political risk increases innovation by about 30%.  
[00:12:23] Similar if you sort of use a PPML specification instead of OS.  
[00:12:27] uh similar result if you sort of citation weight or value weight these patents. So valuating basically looking at the value the private value of each patent using the change in the stock market return of of firms around around  
[00:12:40] filing uh or other measures of sort of patent importance um driven by uh in large part sort of new firm entry as compared to an increase in innovation with of existing firms. Although sort of  
[00:12:53] this distinction kind of is not is not so clear across different sort of specifications but but across the board across all sectors an increase in political risk associated with more innovation in that sector. Again larger  
[00:13:07] effects when you go up to the decade level thinking that you know innovation takes time to react and maybe more responsive to longer run changes in political risk and sort of no evidence at all of anticipation or or pre-existing trends in this relationship. But really this political  
[00:13:21] risk change occurs and then innovation react.  
[00:13:26] We do sort of a range of additional results sort of probing what's underlying these results in the paper.  
[00:13:31] Again, I'm happy to talk more of their questions in the Q&A. But first, sort of consistent with that model that I didn't explain in detail but sort of describe.  
[00:13:40] This is really driven by sort of middle productivity firms might be especially responsive to changes in foreign political risk. The way to think about it is really productive firms are not really relying on foreign sourcing. for a given good much in general to begin with. And really unproductive firms are  
[00:13:54] kind of have no hope of developing new technology and sort of onoring production. But it's these middle productivity firms that do this sort of insurance innovation is what we call it where that really seems to happen. We find a lot of evidence that it's driven by the private sector and that this isn't really driven by sort of  
[00:14:09] government investment or government response. It's really private incentives sort of sort of pushing this in a variety of ways. And those results that I just showed you are probably underestimates because there's some evidence of sort of spillover effects across the production network. So if upstream sectors get hit but you see  
[00:14:23] some evidence of innovation. >> Okay. How am I doing on time?  
[00:14:28] >> Um we have six minutes. >> Oh great. Okay. Um so the next thing we can do is we can zoom out to the glo that was all US. Zoom out to the global level and say does the same relationship exist at the global level where now we're running the exact same regression  
[00:14:42] except we're basically stacking across all countries. So now it's not just INT, it's ITC where C indexes countries and political risk we can now measure at the sector uh country year level where now we're just exploiting the fact that  
[00:14:55] different countries are relying on other on different parts of the world for sourcing in each sector. The advantage of this is we can look across sort of all countries and start thinking about headers need but also we can include things in the specification like sector  
[00:15:08] year fixed effects just capturing you know aggregate changes across sectors that you might worry sort of spuriously correlate with stuff in the previous regression. Here we can fully absorb that fully absorb country specific time trends and just baseline differences in innovation across sectors and countries  
[00:15:22] in general only exploiting country specific changes in political risk in the specific countries from which you source each sector.  
[00:15:30] And again we find this positive effect of political risk on on innovation including sort of various importance weighted measures of innovation. You know again same story no pre-existing trends bigger effects over the long run etc.  
[00:15:43] More interestingly now because we have this full set of countries we can really try to understand which types of trade relationships are driving and sort of one hypothesis you might have is that it's you know on the one hand it could be driven by sort of any  
[00:15:56] any sort of risk of of things going haywire in foreign country. On the other hand it might be stronger when political risk emerges among nonallies because they're really the ones that might put restrictive policies in place. They were really the ones where a trade breakdown may truly be possible in response to  
[00:16:11] political turmoil in those countries. Right? So we can basically estimate the same regression as before except now measure this political risk exposure separately in countries that are your allies and countries that are your non-allies. Basically just by taking  
[00:16:26] this sum over using those definitions either from UN voting data or from the coralates of war project. countries that are we're going to define as your as your allies and then include these two things political risk for non-allies and political risk for allies separately in  
[00:16:39] the regression. And the answer is it's basically all driven by a political risk for non-allies. So what really matters is when there's an increase in political turmoil, political tension in countries that you're you don't have strong political relationships with, that's  
[00:16:53] really where innovation emerges and there's actually a zero effect in in the allied countries.  
[00:17:00] So I'm going to skip this for time, but one thing that you can see is that what drives this wedge is really the sources of political risk where you think the government actually is in has control over this. So things like conflict, whether that foreign country is your ally or enemy, if there's an increase in  
[00:17:15] conflict, that might sort of mess up your ability to source stuff regardless and that that country government doesn't have much control mediating the consequences of that. So you don't see a big difference between allies and enemies there. It's really these sources of political risk like democratic backsliding, military and politics etc.  
[00:17:30] that drive this wedge between allies and enemies and that sort of makes a lot of sense actually.  
[00:17:36] One potential mechanism which I'm not going to have time to explain in enough detail that drives this relationship is that when sort of political turmoil emerges in your enemy countries that's really where you expect breakdown of trade to emerge because you  
[00:17:50] might expect policy to emerge restricting trade either because your country starts restricting trade with those countries things like the sanctions and trade restrictions or because they start doing the same like Elon Musk's example of Putin in the introduction like Putin might start  
[00:18:05] restricting our access to to um to rockets. So we can test that we can test whether the actual probability of a trade trade restriction emerging uh between uh two countries in a given year  
[00:18:19] is in response to political risk is different if they're enemies uh or not.  
[00:18:26] And the answer is that the probability of me putting a restriction on a trading partner or them putting a trade restriction on me in response to political risk uh in their country, they both go up. So that's kind of consistent  
[00:18:40] with this innovation result being stronger, anticipating the fact that these enemy relationships are really where trade breakdown is more likely to emerge. And you can see that there are a range of reasons why that trade breakdown might be more likely. One reason is policy. And you actually see  
[00:18:53] that in the in the trade policy data. I think I have like three minutes probably. So I want to kind of leave you with the fact that sort of in this innovation isn't just sort of like great we can count patents more patents emerge  
[00:19:07] but this is actually reshaping the sort of economic and trade consequences of political risks. So the first thing I want to show you is that there's big differences around the world in the extent to which this innovation response emerges. In particular, we can run the same sort of baseline regression, but  
[00:19:22] separate markets that are doing relatively little innovation at baseline and markets where there's a lot of innovation going on at baseline. And this response is really all concentrated in these high innovation markets both in terms of total patenting response and citation weighted patenting suggesting  
[00:19:36] that markets or countries vary substantially in their ability to respond to these form shot innovation.  
[00:19:42] That's one point. But the second point is that we can actually exploit that variation and say well if we're interested in sort of the trade consequences of well let me let me let me go here.  
[00:19:54] The question is sort of does this innovation response actually shape the trade consequences of political risk. So are countries actually able to reduce reliance on risky sort of sourcing country partners. The ideal but impossible experiment that you might want to say is so to compare the effect  
[00:20:09] of a political risk shock in a world where all this innovation is actually taking place and a world in which you shut all that innovation response and you can't do that. What we can do is exploit the variation that I showed you in the previous slide that high  
[00:20:22] innovation markets are more elastic to foreign political risk than low innovation ones. So we can construct a measure of each export export market exposure to this foreign political risk response. based on whether they tend to  
[00:20:36] export to high innovation versus low innovation markets at baseline.  
[00:20:42] And the hypothesis is that increased exposure to innovative export markets would increase the effect of political risk on exports. And so we can kind of test sort of this question not by this ideal experiment but by exploiting  
[00:20:55] heterogeneous exposure to this foreign innovation response.  
[00:20:59] Perfect. So what we need to do first is measure for each uh country and sector your exposure to foreign innovation and that's going to be basically whether at baseline you're exporting to markets on average that are doing at the exporting  
[00:21:14] to sort of country sector pairs that are doing more versus less innovation and then we regress how do my exports I'm a country sector in how do my exports respond to foreign political  
[00:21:27] risk and how does that elasticity change when I'm exporting to countries that are doing relatively more innovation and our hypothesis if all this innovative response is reshaping these sort of trade consequences of political risk is  
[00:21:40] that fi is negative why well I'm exporting to countries I get this political risk shot that reduces my productivity reduces my capacity to export if they're also sort of innovating me out of the trade network that effect should be larger and more  
[00:21:54] persistent because then even if my political risk shock ends they've now built capacity innovation they're less reliant on my exports and so rather than show you the table I can just show you this in graph form so this is that  
[00:22:07] heterogeneous effect there's a larger and persistently uh negative effect of my exports on foreign countries especially when I export to foreign countries that are doing a lot of  
[00:22:18] innovation so that's a evidence that um uh this innovation is actually sort of reducing reliance of these innovating markets on risky exporters but it's also kind of saying one explanation for why  
[00:22:31] even temporary or transitory increases in political risk and tension can have long run effects on my export performance because now the rest of the world has effectively innovated away their reliance or need for me in the trade network and that's sort of what  
[00:22:45] this is showing and the these effects of the this innovation effect is actually large even relative to the direct effect of political risk but I think I'm I'm in stoppage time but so hopefully I've convinced you that like thinking about innovation is really important for trying to understand the overall  
[00:22:59] economic consequences of political risk, political tension. I've showed you sort of evidence of that in in a variety of ways and I think there's, you know, this is only sort of a first paper trying to understand all of this and there are a range of additional um directions you can go but I'll I'll leave it at that and really looking forward to the  
[00:23:13] discussion and also any questions you have.  
[00:23:16] >> Thanks so much. Um we will now >> so we now stop sharing and RF it would be great if you can share your slides and I hope this looks  
[00:23:30] >> I think this should work now. >> Yes.  

## Discussion (00:23:35 – 00:35:55)

[00:23:35] >> Excellent. Yeah. Thank you very much for inviting me to discuss the paper and also thank you for allowing me to participate virtually. Uh unfortunately I couldn't make it uh due to a last minute issue. Um so I think Jacob did a  
[00:23:50] great job presenting the paper so there's not much uh so I don't want to spend much time um summarizing it but just to give you a the paper in a nutshell let's say so the question that they're asking in this paper is how foreign political risk shapes domestic  
[00:24:03] innovation and trade and the mechanism that they're exploring is uh firms innovating preemptively effectively to hedge against possible disruption of foreign suppliers something that they call insurance innovation As Jacob  
[00:24:18] mentioned, then they provide um evidence at various levels, but effectively uh what they show is that sectors that are more exposed to political risk end up filing more patents, especially when uh  
[00:24:32] the political risk is associated with a trading partner that is considered a nonally. And the implication or particularly interesting implication I think is that geopolitical risks drives directed technical change that reduces  
[00:24:46] interdependence. So in some sense it's a private sector channel of global fragmentation.  
[00:24:52] So my overall assessment is that this is a this is a really fantastic paper. I really enjoyed reading it. It's it's novel. It's creative. It's well executed. Um I think the idea is really um elegant and and of course also  
[00:25:06] timely. Um what I found particularly appealing is that it really opens a new channel linking geopolitical geopolitics to innovation and trade which is anticipatory which is private which is  
[00:25:18] endogenous and in my view a major step beyond the policydriven fragmentation literature which is of course also interesting and important but it's just nice to see this um uh kind of additional layer that this paper puts on  
[00:25:32] top of it. Then I also think there's a pretty strong bridge between theory and empirics. Jacob didn't present the model and the model is relatively simple yet it gives insights that Jacob also alluded to in his presentation  
[00:25:44] particularly that the main response is coming from um the farms that are in the middle uh that are um particularly exposed to this political risk but at the same time also have a fighting  
[00:25:58] chance of innovating themselves out of it. Um and then I think I already mentioned it but uh it's it's really a new lens on on fragmentation. It shows that decoupling can emerge without new tariffs particularly through forms  
[00:26:12] riskdriven R&D choices which adds micro foundations to the debates on geoeconomic resilient.  
[00:26:20] Um I have a a number of questions that I would ask the authors but to be clear it's those are not questions I would all expect to be answered in this in this paper. As Jacob said it's a first attempt uh I think to to to look at  
[00:26:34] these issues but I think it raises a number of interesting questions. So one question that came to my mind is well just a quantitative one how large is the channel that they are that they are highlighting. It's a compelling channel.  
[00:26:47] I think I'm convinced after reading the paper that it's also a significant channel but would be interesting to see how big of a driver of geoeconomic fragmentation it is relative to the more classic uh first moment uh trade shocks.  
[00:27:02] Um, another thing I was thinking is that of course innovation is one response to deal with this political risk, but there's other responses. For example, you could diversify um your sourcing uh structure. You could you could diversify  
[00:27:16] not just from one you could start sourcing not just from one supplier but for multiple suppliers. There's also another uh obvious response which is to um stockpile so to increase your your  
[00:27:28] inventory and and I'd be just interested to understand under what conditions firms choose innovation as their response um and under what conditions they choose other mitigation mechanisms or whether maybe there's some  
[00:27:41] complimentarity um between them and then the last point which um I think is a question that I have more broadly for this for this choice. I would really like to understand whether there's a whether there's a market  
[00:27:56] failure and and whether these uh this insurance innovation if you will um for farms uh that farms are undertaking as uh is is socially um optimal because of course this then determines the role of policy.  
[00:28:10] I also have some additional reflections. Um I'm triggered by this paper but it's not so much comments on the paper but I anyways wanted to wanted to share this as part of my discussion. Uh the first point I wanted to make is that this  
[00:28:25] paper I think really um again shows the value of having a rules based uh trading system because it shows that it's not just um it's not just the realization of  
[00:28:38] uh of of of trade policy shocks uh that matter but it's also the risk of these trade policy shocks uh themselves. It's also the um the uncertainty which means that it's not just important to have a a  
[00:28:52] trading system that um keeps tariffs low that keeps trade costs low but it's also important to have a trading system um that gives you predictability and I think that's one of the uh traditional strengths of a rules based trading  
[00:29:07] system at least as I as I see it. Uh the second point and I already mentioned it but but let me maybe elaborate a little bit. I think it's super important in this whole area to um understand what the market failures are because if you  
[00:29:21] um look at the policy world my sense a little bit that as soon as there's a problem uh policy makers seem to think that they are part of the part of the solution and as a result there's a rush  
[00:29:34] to resilience by governments but often I think without a clear understanding of what the what the market failures are and and one thing that that that I'm that I'm thinking about that uh and that that's obviously a point that other  
[00:29:47] people have have also made is that firms already have a strong incentive to um build resilience because firms are also um uh suffering if supply relationships break down for example. So it's not  
[00:30:00] clear um that there's a that the priority at least is not clear that there's a there's a market failure. One particularly compelling argument that I've heard um for there being a market failure is the argument that um um  
[00:30:13] basically there's a there's a limit to how high prices can be um in times of disruption. So say for example that you had invested in making respirators domestically during co and then the  
[00:30:27] trade show comes and you can't import these respirators anymore. the equilibrium price would be extremely high, but presumably there'd be political and ethical limits on uh the price you could actually sell them for.  
[00:30:39] And that in a way is a market failure that could um make that that could allow you to make a case for government intervention um that that would require um some public policy push to invest in missions. But I just think it's  
[00:30:53] important to think these these things through. And and the last point, even a more broader point, is that um I'm I'm concerned that the Zentire policy debate um is always focused on strategic  
[00:31:08] independence. So the goal always seems to be to become independent of um politically uh risky um trading partners and and here we see also that in this paper we see that farms kind of also thinking in the same direction. And of  
[00:31:23] course I can absolutely um understand this but from a trade perspective I think what it means is that you end up investing in your weaknesses because typically there's a reason why you import something uh from abroad and and I'm just wondering what happened to this  
[00:31:37] idea of strategic interdependence that I think was guiding uh trade policym for quite some time. If you think about European integration for example, it was all about making France and Germany um interdependent and precisely because  
[00:31:52] they were political um foes and not political allies. So the logic was entirely different. Um so so you wanted to be dependent on each other because you didn't get along politically and uh  
[00:32:07] the reason I find this attractive is because it would mean investing in your strengths. So rather than making sure that you're not overly dependent on a trading partner, you kind of try to make sure that there's a balance of of dependence. Of course, I'm not arguing  
[00:32:21] that this is something that is going to naturally arise as the as as the less fair equilibrium, but I think from a policy perspective, uh it's something that we should not forget.  
[00:32:33] Thanks. >> Thank you so much. um R OA and we would give you a very short time to to reply and then we will open for question.  
[00:32:47] >> No. Yeah. Uh >> and please use the microphone either the hand microphone or Thanks actually. But is it uh I I think it wasn't uh whatever I can >> I hope that Ralph listen was able to listen to your presentation before.  
[00:33:00] >> Yes. Yes. Yes. Oh, perfect. >> Uh just three three I mean first of all thank you so much for all those comments. super interesting and give us a lot to like think about and sort of keep keep working toward. Um there's three things I wanted to say. So first I  
[00:33:13] think this question of this relationship between sort of for lack of a better word adaptation via innovation and adaptation via sort of reallocation or resourcing is super interesting and those two things could be substitutes.  
[00:33:26] This is something that also comes up in other work we've done on adaptation to climate change where similarly on the one hand you could change your you think about agriculture like you can change your technology by developing seeds that are more heat resistant and keep increasing your ability to produce as the sort of you know heat apocalypse  
[00:33:40] comes or you could reallocate from parts of the world that remain productive or becoming increasingly productive and those two things could fight against each other. Your capacity to your incentives to innovate could be lower if you can easily reallocate and vice versa. And of course that decision too  
[00:33:54] interacts with international political ties and the extent to which you actually want to or can rely on other countries for your food production and this kind of idea of independence versus interdependence. I think I think but there's like no research on that at least also in climate about the relationship between those two  
[00:34:08] mechanisms and I think it would be super interesting. Second, this question about market failures and like the extent to which you know the private market directs the directs innovation in sort of the right way is also super interesting and another area that is hugely important and about which there's  
[00:34:23] almost no evidence with one exception I can make a plug for the the AA lecture from a couple years ago written by the adviser to two of the four co-authors on the on the paper Jonasu which called something like does the market get the direction of  
[00:34:38] innovation right trying to make a first attempt that answering that question in a couple different sectors and kind of a plug for studying this but there really hasn't been much empirical work trying to understand that question and it's it's incredibly important. Um the third  
[00:34:51] point I want to make is is related to this last piece about uh independence versus interdependence and I think one thing you can see from our results is that even small moves that governments make in either direction can have these cascading effects as the private sector  
[00:35:06] responds to those changes in policy. So you can see from our paper, one of the reasons that private firms respond to foreign political risk is because of this anticipation of restricting trade policy which increases their incentives to innovate and use domestically which  
[00:35:20] further kind of you know ex increases this this uh decoupling or or interdependence. And so that's an example where even you know relatively small moves toward independence on the trade policy side can have much larger effects because of the private sector  
[00:35:35] innovation response to those policy changes. So we show that in that direction, but you could also imagine the same thing going in the opposite direction. And so I think thinking about these links between government policy changes in either direction, but then how that induces sort of a private  
[00:35:48] sector response as incentives change is also really sort of interesting and important for thinking about these these two directions. All right. Sorry. And now I question.  

## Q&A (00:35:55 – 00:41:40)

[00:35:55] >> Yeah. So um I would suggest we take uh three questions, collect three questions and then uh you will have some minutes to to answer them and um my colleague will come around with the microphone.  
[00:36:06] I see two questions. >> It should be on. Yeah, >> I thought it was.  
[00:36:17] >> Uh, so you say this is this is really coming from the private sector and I wonder just as more of a clarifying question. I mean when the DoD gives money in the 80s to General Dynamics or now to Palunteer and those are private companies but they're clearly being  
[00:36:31] directed by the DoD for example. And so um are you able to see in the patent data that this somehow is like completely independent of the government?  
[00:36:42] >> Yeah. So so there are a couple different so I think our ability to do that is imperfect but we can do that in a couple different ways. First is that >> I collect >> oh should I collect?  
[00:36:49] >> It's fine if you if you keep it short then we can do it.  
[00:36:52] >> So so so first the first cut you can do is you can look at private firms versus the patents and nonprivate firms and sort of it's all driven by private firms but that doesn't capture the second piece right >> the second thing we can do is in the patent data you can identify patents that that uh were funded by national  
[00:37:07] government like government interest patents and so you can see it's also driven by sort of well it's um slightly weaker if anything for these government interest patents but certainly really strongly there for these non-government interest >> yeah it's a lesser it's a little bit noisier there fewer but yeah it's a  
[00:37:21] little bit there too the third thing you can do is you can see that yeah there's no there's no evidence of a difference between the two the third thing you can see is if you just look across sectors as at the ones that sort of the US government has at least identified as critical as compared to the ones that it  
[00:37:35] is not identified as critical. You would imagine they concentrate their investments in these critical sectors.  
[00:37:40] It's anything a little bit stronger for these non-critical sectors as compared to it's not statistically distinguishable, but across the board it really doesn't seem to be driven by anything that looks like public investment. It's and it's sort of as wrong with these things that we really see it could only really be driven by by  
[00:37:54] private decisions. >> Can we ask the second two questions? Two people raising their hands before But we can hear you. Um  
[00:38:14] I was wondering these are maybe something you said in the presentation but clarify enemies.  
[00:38:25] Was it was it binary that you were only like enemy friends? change over time and then another another maybe question on exposure you said as well I'm sorry but  
[00:38:40] how did you manage exposure country or >> and I think there was a question right behind you of the Great. Uh so uh so the first point so so  
[00:39:19] how we measure what I showed you what I showed was using a binary classification of allied just kind of a median split but across three different definitions of how you sort of define allies and enemies. one using the coral to war  
[00:39:32] project uh one using the UN voting data and I'm sort of using sort of a a measure based on sort of like all uh military non military alliances between countries things look pretty similar if you use continuous measure it's just like it's easier to say like you know  
[00:39:46] two coefficients one ally one enemy but like if you want to do a continuous interaction it's not going to make much of a difference um in the interpretation changing over time so we're using a fixed definition and that's in part because variation in at least these measures over It's much smaller compared  
[00:40:01] to just the cross-sectional variation and it ends up sort of not messing like the numbers end up being that different but just sort of we want to say we don't want any way in which if you know these relationship are evolving in part in response to political risk  
[00:40:15] and innovation as well. We don't want to incorporate that into our measure. So we're just using sort of a fixed definition but but that's in part it's not going to make much um direction versus levels. I think we would interpret a lot of this kind of direction in part because That's what  
[00:40:30] we're able to identify because we're always going to have a a time fixed effect or a country time fixed effect in the regression which is going to be fully absorbing sort of in the on the US sort of how the level of innovation is changing over time the amount of patenting changing time and then the  
[00:40:45] country sector version how innovation uh country level is changing time. If we wanted to like identify the the sort of level change you need kind of a more complicated model of that intercept and how that's changing over time. That's  
[00:40:58] tricky in the data because overall things increasing over time for lots of reasons and and so I wouldn't know exactly how to how to discipline that model I'd be convinced by but you need more structure to say something about  
[00:41:12] overall overall >> so thanks so much to um all the people asking questions also thanks to Osa and of course um to Jacob and I saw there are many more questions many more hands  
[00:41:25] raised um feel free to discuss over coffee we have to move on and we will now hear um thanks um hand this over to uh we will now have the presentation by