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.
=

# The Fragmentation Paradox: De-risking Trade and Global Safety

Authors:   
Discussant: None  
Video: https://www.youtube.com/watch?v=m3pg-N8Dv5E&t=2503s


## Talk (00:41:43 – 01:04:22)

[00:41:43] price of fragmentation to the fragmentation Paradox presented by Terry Maya welcome H we are also very happy to have you here of course um not as far long as away as our previous speaker  
[00:41:58] nonetheless um yes the floor is yours thank you very much you have 20 minutes you tell me when I'm overtime H yes I will be the annoying one with the cards yeah and if you'd like to be standing you just need to be sure that you're  
[00:42:11] standing close is to the microphone thank you that's fine I can definitely do that um thank you very much for uh having me again uh to that great uh great conference so um this paper joined  
[00:42:25] with um Isabel meon and Matas T is sort of combining elements of what some of you have seen this morning mimo's paper and and the paper we just uh uh seen  
[00:42:37] right now s and it's going to talk about um the fragmentation Paradox is what are the economic cost of of of the risking what the risking trade that that reducing trade uh between us and China  
[00:42:51] for instance but taking into account the endogenous probability of actually going at War of actually triggering the conflict and how those two things interact so um it's um it's a paper  
[00:43:04] about um fragmentation of global trade and uh it's high on the policy agenda I don't have to explain that here but uh the third bullet point is the most important is that most of the literature  
[00:43:17] as as long as we as We Know It uh has been looking at trade policy the risking is trade policy in an insecure way right you want to drisk so so there must be some risk but it takes the risk as given  
[00:43:30] it it doesn't it doesn't most of the literature doesn't take the endogenous probability of a conflict arising as something indogen as something that actually is going to be moved by by by by trade yet um uh actually um there if  
[00:43:45] you if you want a narrative about how trade might have precipitated uh or the fall of trade might have precipitates on Wars you can take the historical collapse of global trade um as mimo was was saying this  
[00:43:58] thing is actually ambiguous in the theory we're going to get back to uh to that and and we're going to take the two-way cation between uh between trade and geopolitics and in in that model in a in a sort of an extension of a model  
[00:44:13] we had with Philip Martin and um and Mattias tonig and apply it and show that the risking can actually paradoxically make the world less safe maybe paradoxically it can actually make the world less safe it will tend to do so  
[00:44:27] and and that's going to undo some of the welfare gains to be expected of their risking if you do risk it's because you expect something good out of it is that if you go at War you're going to be less dependent on the country you go at war with and that's increasing your welfare  
[00:44:42] that's reducing your GE economic losses but if at the same time you're increasing the risk of that bad thing happening you know there's something of a balance here and we're going to study here um the extent of the balance in a  
[00:44:56] try to quantify by that this is going to be a very simplistic model of um of uh of conflict uh mimo has uh has presented uh some of the elements already so I'm going to be a little bit uh short the the difference  
[00:45:10] here is that it's going to be a model where it's a diplomatic game of concessions between two countries that have to uh agree on how to share a pie and if they don't agree they go at war I'm going to have no aggressor here in  
[00:45:25] the other country they have to age if they don't agree they go at War there's going to be something so why should they do that ever if they're rational and and we believe that there's gigantic cost of of Wars gut has shown us with a paper  
[00:45:38] with Moritz that those huge effects of War so they should be rational and never go at War well the question is about the the whole thing is about imperfect information or incomplete information that and a symmetries of information is that you don't know actually my cost and  
[00:45:52] I don't know your cost and that's that's what's going to make it possible to actually go at War here is that when this uncertainty is large enough and the losses the opportunity cost of War are small enough there's going to be some uh actually situations where you you  
[00:46:07] actually go War the opportunity cost of War thing that you see in red here the ocw are going to going to be a foregone real consumption and we're going to use uh something that is now pretty pretty  
[00:46:21] classical in in in structural tra in in like sort of quantitative trade model we're going to take U an elaborated model with input output linkage that was just presented to you it's going to be actually a little bit  
[00:46:34] simpler in some dimensions and a little bit more complicated than others um and that's going to give me something about the opportunity cost of War which is the economic the pure thing that the consumer should care about which means  
[00:46:47] that a democracy or even an autocracy should care about maybe with different weights but they should care about that if they want to be reelected or if they don't want to be thrown out of power okay um and now we're going to put that  
[00:47:01] into into a model of um of War um I'm going to quantify three things that are new uh in a sense that that arise  
[00:47:14] because of the possibility of a war they I'm going to call them the endogenous geoeconomic factors and they all depend on those ocw um it's going to be the probability of escalation to war of course it's  
[00:47:28] going to be the exposed true cost of a war which means destruction plus trade all everything that that is going to be uh that is going to be lowering my my my real consumption if the war happens and  
[00:47:42] something that I think is new in the debates um you don't hear very often is that if you are to negotiate to avoid a war you got to make some concessions essentially you want to avoid Wars and you want to avoid Wars you come to the  
[00:47:56] table and you make concessions to try to avoid wars in the model there's a very clear uh uh prediction which is that those concessions are going to be based on the asymmetry of the opportunity cost of Wars if you got a lot to lose you're  
[00:48:10] going to have to pay me not to go at War because you got a lot to lose compared to me and the difference in thetic of war is going to be uh uh here here with the functional form assumption that we're going to make here is going to be  
[00:48:24] a very trivial uh uh uh thing and and and that's going to be one of the cost of the possibility of Wars is the how much I've got to pay you to avoid that okay so here we going to we're going to  
[00:48:38] we're going to typically uh um show you one case which is one um one application which is USA China with but but it's of course applicable to many many  
[00:48:52] things if you if this whole thing relies on the fact that you think that uh trade uh is is harmed by Wars if trade is not harmed by Wars of  
[00:49:06] course there's nothing to Der risk about and and forget about it um so here you can see two um one that is almost pure data on the left this is India Pakistan and you see this line this red line in  
[00:49:20] one is the gravity prediction a very simple gravity prediction over time okay so so essentially if the world if if India and Pakistan would be at normal trade they would be on this red line okay and you  
[00:49:34] can see that in 1948 they are 20 times above the normal well yes they just got out of being actually the same uh uh country under the common colonizer so they trade way too much but you see that  
[00:49:47] each of those red points is a war with intensity 5 which is the maximum more than 1,000 people uh uh military killed on the battlefield and you see that it goes down and down and down in 1965 there's a big war that  
[00:50:01] goes this thing goes now to 5% of what it should be and it still goes down it sort of never recovers that's one of the pair that um that actually goes at War the most often that's one example on the  
[00:50:15] right I gave you uh the glicken tayor uh estimation of world wars um and you can see that on impact uh the bilateral trade goes down by 85%  
[00:50:27] between two countries that go at War and the minus 12% that you see at the second gray bar is actually what uh what is costly in terms of trade not with the country you're at war with but with the  
[00:50:42] rest of the world so it's like a war is essentially killing trade 85% of it between uh the two countries and it's also destroying ports or there are sanctions or there are lots of stuff  
[00:50:55] that reduce trade by 12% and it goes essentially back to normal after 10 years 10 years of piece you go back to what gravity predicts which is what gravity uh uh works well and so it  
[00:51:07] enables us to to to to use that we're going to do something like that in a more elaborate model uh where you have a lot of input output linkages Etc but it's essentially going to be the same picture we're going to impose you see  
[00:51:21] this 85 and 12% we're going to impose in our counterfactuals that this is the cost of a war if you go at war between China and and and USA it's going to go down by 85% 12% with the rest of the world and we  
[00:51:33] compute you put cost of War about that so very fast uh telling you about the um the sequence of events so first the two countries have to have a dispute okay so they got a dispute about some resource  
[00:51:47] that is divisible okay so so they got a a pie to share they got to agree on how to share it essentially it can be on your soil on my soil in the middle whatever it's got to be something that we agree how to share and they got to  
[00:52:00] agree on a diplomatic protocol and and that's where mechanism design theorists uh uh help us a lot they negotiate Under The Chosen protocol and if they agree there's a transfer positive or negative  
[00:52:15] between between N&M otherwise they go at War as as I as I was already saying if they do go at War there's three type of cost that we can allow in this in this model first there's of course human  
[00:52:29] losses so so those are consumers workers so so it's of course affecting welfare there's economic damages this is called Alpha so there a tfp drop everything has got to be less efficient and there's those T Bill and  
[00:52:44] to mule that I just showed you before which is how much trade cost at valm equivalent uh is increased by by bilateral war and how much trade with the rest of the world is actually trade cost is actually  
[00:52:57] increased okay so that's that's what war is going to be about in our in our in our model the utility um uh cost of war is going to be this first term the difference in the log of the real  
[00:53:11] consumption between peace and War ocw and I'm going to show you ocw a lot after that and it's going to be this util which is the private information of how much my resource is actually  
[00:53:25] destroyed how much of my resource can I keep if I go at War and that's private okay this util is private how much you can you can you can view that as okay what is my probability of having a lot of damage and not you a lot of damage  
[00:53:39] what is my military superiority compared to you because that's going to affect of course the the the shrinkage between this this resource the the the sharing so so but that has to be separated from consumption so that's one thing here you  
[00:53:53] see it's additively separable from consumption in order to to to resolve the model um another thing is that we make I mean the model makes the Assumption the the the game theory part of it that um  
[00:54:08] Peace is par dominating war is that actually um this those the sum of the utias has to be uh negative and they can be negatively correlated um and they going to be uniformly distributed here  
[00:54:22] okay so they they're free to choose any type of protocol it's a very general thing so let me tell you let me show you immediately the results by contl so there's two type of things first they got to agree on a protocol and so here  
[00:54:37] we just borrow from contel and they say okay the the the simple mechanism here achieves the second best they first announce some utility cost of war that is going to be different from the true  
[00:54:50] utility cost of War then they if it's compatible if they don't overshoot jointly they agree on a transfer and then they uh they they they they produce and consume Etc okay if they don't agree  
[00:55:04] on transfer then they go at War okay and the optimal announcement that you see at the at the bottom once you've agreed on that you see the announcement is 2third of the of the true uh cost so they have  
[00:55:17] a tendency to strategically misreports which is that have over claim on the pie compared to what they really should which means that and and to to show you in a nutshell which means that if there is actually  
[00:55:31] not too much of a cost of war and a lot of uncertainty th those two guys having a tendency to overclaim will tend this thing to in some cases go actually at War and that's this this uh this red  
[00:55:44] part compared to the blue triangle in This Magnificent drawing by Maas tonig um the the the the theorist of the group um not quantitative theorist but qu  
[00:55:56] graphing theorist um and so the economics uh uh uh of of those things are are the following so that's essentially three equations plus a last one that joins them that that is  
[00:56:11] sufficient the probability of the appeasement is going to be a function of the squared but the important is that it's a sum of the opportunity cost of War okay the more there is to lose jointly the more we will tend to agree  
[00:56:24] in equilibrium but the peacekeeping cost the transfer is going to be a difference between those two opportunity cost of War okay so there's going to be this tension between the levels and the difference uh and and of course the  
[00:56:38] actual opportunity cost of War we can compute the true cost of War um is this this opportunity cost of War conditional War happening which means that if I go back to that super drawing um you see  
[00:56:53] they don't you don't go at War over the whole triangle you avoid the worst wars in equilibrium and so this is this is just conditional uh my my op Costa war and let me now summarize what is the  
[00:57:06] welfare on which which I'm going to compute um the expected welfare in the shadow of War so exante before knowing whether war will happen or not is that blue thing which is my essentially uh uh  
[00:57:21] real consumption in terms of peace and the red thing is the economy clause and you see that it's a weighted average of the cost of how much I've got to pay you to avoid war and and that's multiply by  
[00:57:35] S if we stay at peace and one minus s which is the true cost of War if we go at War so that's this thing which is which is essentially the geoeconomic loss and I can compute it so now let me take the five last minute to to actually  
[00:57:49] show you the uh empirical illustration um I'm not going to show you how we calculate the ocw but it's very it's a very similar model to what s just showed so you believe it and it's  
[00:58:02] beautiful and it works and I just show you some some some some some graphs instead um so this is the actually the TAA so that's the uh this uh um uh new Fantastic uh uh input output matrices  
[00:58:16] for the whole world of 60 countries I think this is the one that goes from 95 to 2020 the bilateral consumption the first left graph is how much us say consumes from China as a percent of its  
[00:58:30] overall consumption and by the way all those data sets have services in there which means that everything related to trade is going to be dampened like crazy right think about because all this  
[00:58:43] Services is mostly non-tradeable so all those mechanisms are based on trade so so you got to keep that in mind uh uh um okay so that's why it's only 1.4% of consumption here but you see the USA  
[00:58:57] goes up China doesn't depend very much uh uh uh on um on um it doesn't depend more than before and those two things cross which is going to uh bring me to  
[00:59:09] my uh next um next slide so on the left you see the opportunity cost of War for USA and China so in blue we got the USA and in red we got China over time this is the ones that we  
[00:59:24] compute so it's what would be the opportunity cost of War so the the the loss in real consumption if there was a full-blown war with 85% less trade 12% less trade with the rest of the world 3%  
[00:59:38] of loss of tfp between USA and China and you can see that those things cross again which is that at the first the USA actually loses less and China loses more and at the end because of that trade  
[00:59:53] Evolution it's the USA that would lose more in terms in economic terms okay and you can see also the dotted line for those of you who see the dotted line in Gray here um is the probability of Peace  
[01:00:06] so it goes up and down and up and down again as you can see that the opportunity cost the sum of those things goes up and down um um you you got some um some evolution of the S on the on the  
[01:00:20] middle I think it's the most important you got those geoeconomic costs let me remind you what they are those GE economic costs are those red terms okay so the TCW is the red graph the pkc is  
[01:00:34] the black graph here in the middle and the yellow is the weighted average so it means that essentially after 2010 USA has to pay China in terms of  
[01:00:49] diplomatic concessions in order to avoid war and it was the reverse before just because of their B asymmetry in the dependency now it it has to be that the USA has to  
[01:01:02] accommodate China on many things because of this bilateral dependency okay and that's and that's going growing over time now put yourself at the very end and let's uh think about the risking  
[01:01:16] so the final thing about the title of the paper I'm going to do a post 2018 the risking so I'm taking this situation in 2018 this is the world and now I'm going to  
[01:01:28] drisk which means that USA is going to increase its trade cost if you look only on the right of this graph you're going to increase this trade cost with respect  
[01:01:39] to China by 5 10 15 20 25% okay so each of those dots is the USA imposing trade cost of 10% or 20% to China as you see  
[01:01:53] and and then you can calculate the opportunity cost of War which means that now if I impose 20% tariff on China I trade less with China which means that if I go at war with China it's less it's less costly that's the whole Spirit of  
[01:02:05] the risking I think but that's only looking at the red uh at the red there's no laser this no yes maybe uh uh the on the right graph you can see the red which is the true cost of War if a war  
[01:02:20] happens okay and the thing is that of course the this thing goes down if if it if it if the war happens but s goes up in one minus s goes up so the war becomes more likely so there a there's a  
[01:02:34] cost here and I'm going to finish here which is that as you der risk you reduce the cost that it would the cost of paying China the cost of going at war with China if you go at war with China  
[01:02:48] but you make that event more likely which means that the welfare thing is is UN un signed but here we can say what it is we can say that there's actually uh  
[01:03:01] look at the right the right figure here this is a total welfare of day risking so the black line is the pure economic cost the yellow is what you gain in terms of the geomic losses so that's a  
[01:03:16] gain and so the red is the uh uh addition of the two and you can see that yes when you drisk with China you you actually will have a little bit less  
[01:03:29] losses than uh uh uh because of the reduction of Wars but the global balance is still a very strong negative uh picture it didn't have to be so the red could have been above zero could be  
[01:03:42] u-shaped if you go far enough but in our quantification the red is still pretty close to the black which means that the geoeconomic classes are not enough to actually reverse the story uh and I'll  
[01:03:55] finish here sorry for being uh a little bit over and L the floor to my discussion thank you very  
[01:04:02] [Applause] much  

## Discussion (01:04:22 – 01:16:45)

[01:04:22] okay okay okay great uh thank you so much for the opportunity to discuss this paper I learned a huge deal in about two different literatures um while reading  
[01:04:36] this paper and listening to the today I have to say it's a huge paper so it's really hard to fit it into 20 minutes to make a full sense of it um so I'll try to um overview it first and then give  
[01:04:48] some comments so we've seen uh first because of the first presentation in this session and in general just following uh the news that geopolitical tensions often push Global de globalization through sanctions through  
[01:05:01] trade Wars through du risking but the question is is could there be additional negative feedback loop from drisking to more geopolitical tensions and that negative feedback loop is exactly what this paper is trying to  
[01:05:16] theorize about and um calculate the probability of so the paper basically asks can the risking make countries more likely to go to war so this paper overall adds a lot of uh  
[01:05:28] more Nuance to you know theories of why we fight why conflicts happen and it explains it through um very nice structure um where uh why fight could be driven by De  
[01:05:41] risking so the the the huge paper uh consists of three parts uh where the first part is a bargaining model um where two countries choose this optimal  
[01:05:53] protocol uh which allows uh actually for um the authors to calculate the equilibrium of this model and then allows to sort of calculate three important components the probability that two countries will start fighting uh the cost for the two countries in the  
[01:06:08] time of peace and the cost for the two countries in the time of War so if you think about welfare of countries when there's no conflicts we just calculate welfare as a consumption value uh so welfare during peace and we don't think  
[01:06:22] about any geopolitical interactions but uh in this model you kind of add an additional component of welfare which is Ln in red that um changes changes the the uh  
[01:06:36] the usual calculation of of the country when it decides on trade policy or other policies and uh will depend on crucially on the probability that the war might happen and um will depend on its bargaining power with the other country  
[01:06:50] so that that is the Ln uh adding to normal utility utility function so this uh this component Ln um then leads us to the next big chunk of the  
[01:07:02] paper because Ln um as the authors show will depend on opportunity cost of War it's a quite Elegance it's just one parameter will determine all um pieces  
[01:07:14] of geopolitical factors all pieces that countries will take into consideration before going to war um and the the contribution of this paper and addition to showing this  
[01:07:27] elegant bargaining model is to calculate this opportunity cost of War um in um with a with a large trade model so this model will contain many elements uh that  
[01:07:41] countries will take into consideration when they go to war so the first element and there's no pointer but the first element is in this large equation is um economic damages because when you go to war you stop um uh producing at the same  
[01:07:54] capacity and there you can model it as a reduction in productivity costs another very important cost of war is losses of factors so losses of human life and capital and that's um the component on the second line the first component on  
[01:08:08] the second line and you can also plug it in into the um calculation of opportunity cost of war and finally um the probably key component uh for the for this version of the paper is the  
[01:08:20] trade frictions uh so this uh trade frictions you lose trade when you start fighting with your um combatant and uh basically the how much  
[01:08:32] the loss of trade will um punish you in the case of you fighting um will determine your opportunity cost of War so how how much your combatant has  
[01:08:45] you on the hook um basically will um will will determine this opportunity cost of war and then in turn will determine the probability of war and determine your outcomes during peace time and War time so um with this in  
[01:08:59] mind let uh me just kind of summarize the the main kind of message of the paper so countries there's a tension between drisking and going  
[01:09:12] to uh do risking because of this o opportunity cost of War affecting probability of War when countries do risk it seems very rational and very optimal um if they feel like there's a conflict between those two countries  
[01:09:27] it's rational because it reduces your um peacetime transfer it reduces your um cost of maintaining peace it reduces your cost of actual War because you do risk you don't have to  
[01:09:40] you lose less when you stop trading with your um uh with the opponent but it has the opposite effect on the probability cost of war or probability of War because if it's War becomes less costly  
[01:09:54] it improves it increases the the likelihood that you will uh not find um an agreement and increases the importance of asymmetric information in this bargaining  
[01:10:07] game okay so um so this not internalizing the probability of escalation when countries the risk is basically the key takeaway so I would say um it's it U  
[01:10:22] raised many thoughts when I was reading this paper and I think it's it's it's an amazing paper it provides a lot of um uh intuition and structure in how we think about conflict is um for example so did  
[01:10:35] with this model in mind did did EU do the right thing by not der risking from Russia in 2014 so the war happened in 2022 but would the risking increase  
[01:10:47] already um the uh the likelihood of that war happening or make that war even um H happen even sooner that's the question or in the hindsight if if the probability of that were  
[01:11:00] happening was so certain then it would be optimal for this kind for EU to do risk already in advance and stop appeasing so again this model can calculate sort of the optimality of the risking uh in this scenario and many  
[01:11:14] other scenarios um okay so I will uh quickly go over my three comments I think uh you could uh consider them more like um suggestions when you write the paper just giving a little bit more context to  
[01:11:29] the reader who is an amateur in this field uh so opportunity cost of War crucially depends on on the level of Dr risking on it crucially depends on how much the other opponent has you on the  
[01:11:42] hook based on uh the trade um bilateral trade uh with with that opponent but the question is how much this factor is important when you um consider other  
[01:11:55] factors in the opportunity cost of War calculation uh for example loss of life and capital um they are important from the economic point of view and the paper considers that but they could be directly entering the utility function  
[01:12:08] of the representative agent and so kind of comparing uh the costs of war in the sense that you lose factors of production uh through the economic Channel yes okay there is a cost but there could be even bigger cost if you  
[01:12:22] just want to avoid uh loss of life and loss of your buildings in your country so um I wonder how the paper can kind of speak to additional additional kind of non-monetary costs non-economic uh costs  
[01:12:35] and uh second one and I think you mentioned it a little bit is when you have countries bargaining it's usually done by the leaders of these countries and if you think about a de Democratic country and an autocratic country the leader in the autocratic country will  
[01:12:48] not internalize all the economic um gains and losses uh in in the case of War and so if you put the weight of zero to the economic gains and losses and the weight of 100 to for example personal  
[01:13:02] risks uh that an autocratic leader will face then basically the the importance of trade and importance of the risking becomes quite small uh for at least for  
[01:13:13] that um for that autocratic leader so my I was wondering if it was possible to kind of calculate some scenarios where you have a symmetry between an autocrat and Democrat and uh see how much the  
[01:13:25] risking would um would uh matter there uh additionally so the paper builds up on a really nice and famous paper by Philip marang and the same  
[01:13:38] authors TI and Matias uh in 2018 that uses a simpler trade model to calculate opportunity cost of War um without any inut out linkages and then uh at least  
[01:13:51] in the draft that I I read I mean it's uh the new the new paper talks a little bit about input out linkages but I wanted a a little bit more intuition how that would um change welfare calculations how would would that change  
[01:14:04] probability of War I mean you say that yes the fact that world is now um more um more integrated in terms of the especially the value chains yes the opportunity cost of war is higher and so  
[01:14:16] the war is less likely but you know we talked about the asymmetric cost to China because of um uh asymmetric cost to China because it depends on uh Global value chains more so how if you do your  
[01:14:30] welfare calculations with final goods trade and with IO trade and compare the two that would be really nice and a very last uh comment very quickly so uh the paper and not the presentation used one more example on um calculating the  
[01:14:45] probability of war and that was the case between Russia and Ukraine and uh here I mean first um the conflict started not in 22 but in 2014 14 and it looks like even in the data that the paper shows the trade actually  
[01:14:59] was increasing before 2014 so between 2010 2014 and so but the the conf still conflict still happened so um here I mean kind of could you speak a little  
[01:15:11] bit about like why uh well why this uh didn't play out the same way as the model would predict and second I think to describe the conflict in 2022 after a long period of the  
[01:15:24] risking between Russ R and Ukraine it's important to consider also de coupling between Russia and the collective West because Russia would probably take into consideration the loss of trade with EU much more deciding to go to war than  
[01:15:37] just loss of trade with Ukraine and so I think that would um kind of be very very interesting to to to think about that and finally I mean is um loss of trade with Russia was that an important  
[01:15:50] component for Ukraine to uh stop accepting the whatever was justed to it on the bargaining table by Russia or there were other factors um that are unrelated to trade it would be really helpful to the reader but overall it's  
[01:16:03] an amazing paper um one can learn a lot reading it and I suggest everyone to to read it and um think about it thank you thank you very much do you want to  
[01:16:20] react to the discussion I mean you can also stay there wherever you want to sit um I just got uh I just was told that the next point on the agenda was moved  
[01:16:32] to at 3:45 instead of 3:30 so we have a lot of time to answer questions now and there's no rush for the coffee that will be plenty of time do you want to answer first and  

## Q&A (01:16:45 – 01:35:48)

[01:16:45] then collect questions no of course okay um so yes there might be a there might be something about for the first question about the the the weight  
[01:16:57] um uh of a loss of life directly into utility that we we could definitely do that I think it's related to that question about democracies versus autocracies and sort of the  
[01:17:12] weights right now it's not in the model per se but it's there's two questions is whether you want a symmetry in the sort of the alphas and and and the gamas even though they  
[01:17:25] enter the opportunity cost of War right now and not directly the utility still M you might view that as an implicit as an implicit cost you could have Alpha times the perception of how much it matter I mean or you can have it directly I think  
[01:17:40] we could have that without too much of a problem one thing that would really I think be more difficult uh would is that an autocracy is more um uncertain so  
[01:17:54] it's a it's like probably you got a more uncertainty and more a symmetry between autocracy democracy than than two democracies in terms of the thing or  
[01:18:06] maybe not but it could be that it's much harder to predict or and and and somehow the autocrats want it that way it's harder to predict what would be the cost uh and so that would be that would be changing the  
[01:18:21] ETA with respect to the uh 2008 paper calculation of opportunity cost of War so first this original paper was was only Computing the probity of escalation it was not doing the welfare second of  
[01:18:35] the opportunity cost of War um the io uh the input output thing is going to be really important because if your if your partner is hurt in terms of alpha in  
[01:18:48] terms of its tfp it's going to feed back in a very magnified way because it's his TF p is going to hurt you because you you use this as inputs much much bigger  
[01:19:01] extent than the direct so there can be some magnification effect in that in that effect also in the old paper we wen't doing General equilibrium we were doing this this thing without nominal wage adjustment and without production  
[01:19:15] adjustment so that that also actually changes things quite a lot because those are massive effects in terms of gdps in terms of TF p and and trade so so having General equilibrium is much more was a  
[01:19:29] challenge at the time but it's not anymore so that's that's the good news about the technology Russia Ukraine yeah we need to do that seriously that graph I mean one thing for sure is that the Russia Ukraine trade was actually  
[01:19:42] the bilateral was decreasing but Russia was trading more with the rest of the world which acts as an insurance as we explained in that in that very first paper and makes the conflict more likely that makes  
[01:19:56] escalation more likely but the rest of the World includes EU in a way so in the side of Ukraine but but if you if you think about the estimates of the third  
[01:20:09] party uh trade sanctions in a sense those are relatively minor compared to the bilateral so here we took the most extreme which is 12% but in our original  
[01:20:20] paper it was not even 5% so so so your global trade with the rest of the world when you fight one country goes down very little with the rest of the world so so so  
[01:20:34] that's okay thank you and should we open the floor now to questions from the audience and we would collect questions if that's okay for you okay let's start with the ones who have an ask questions yes so  
[01:20:48] yeah hello thank you um my name is Jonathan Pak I'm a journalist at Ure active I realized this is not a very academic question but we are at a ministry here so I want to ask it would you say that drisking like the EU is  
[01:21:02] currently doing Visa V China for instance is a mistake would you conclude [Music] that uh Jesse shreer Columbia this is a really interesting paper uh and one  
[01:21:16] thing I'd be I'd be curious to hear is in the context of the framework is there what what would the characteristics of a good be where say the us or China were to buy you know for a given amount of dollars they would increase trade together what would be the  
[01:21:30] characteristics of a good that would have the most effect on say raising the opportunity cost of War thank you and then yes I saw a hand over there uh Pinger from tubing uh thanks a  
[01:21:44] lot this is really obviously very important and exciting exciting work bringing also endogenizing this this this Choice uh my question is rather uh specific about about this opportunity cost of war and how you empirically  
[01:21:59] discipline that I understand you use gravity in a sense but it's always seemed to be a um a country specific variable although we would probably expect that it's kind of bilateral right so and and since you picked that example of China us here the cost opportunity  
[01:22:14] cost of War of a full-blown War would would kind of be in infinite right so anihilation um so so how do you do that and and then I think how quite something probably depends on that number right  
[01:22:27] yeah and please apologize to the left side of this room I can't really see everyone but maybe there's also a female participant having a question no okay then we start with the three first  
[01:22:41] questions so um the the the question is do the risk the risking a mistake um really depends on the precise case so so for instance here I I I showed what  
[01:22:55] our model is is is saying for USA China might be quite different actually for Europe China because the relationships are not asymmetric as exactly the same  
[01:23:08] they're not maybe on the same Goods uh and so and so the the picture might might be very different we we didn't do it yet so so I I should say this is the very first uh thing we do with the full  
[01:23:22] quantitative apparatus the the thing that you saw with Russia Ukraine was a very very um sort of a reduced form of that so so we're going to obviously test  
[01:23:34] other and and say a mistake here I mean defining a mistake is that you got to believe a lot the model right is that this probability of going at war is what's going to drive whether it's a  
[01:23:48] mistake or not because if it's increasing the probability of War too much yeah for sure sure it's going to be a mistake so so how this probability of War adjusts uh uh is is really key so so take with a grain of salt but here for  
[01:24:03] USA China for sure the economic costs are way dominating the geoeconomic gains that's that's what this uh this thing is is telling  
[01:24:15] us um on the um what so so yes we can we can that's next on the the you know to-do list which is look for what would be the  
[01:24:29] risking Goods that would have the so here bak and fari type of uh theoretical findings of which are the goods that are the most critical in a sense that those  
[01:24:41] domar weights about the global influence uh really are going to be essentially it's going to be the good it's going to be a a weighted average of the goods where you have super high depend  
[01:24:55] dependence um and very low substitute stability okay so those are the things that will essentially uh uh be be the critical the critical thing um but and to respond to  
[01:25:10] the to the asked question fed into this Global General equilibrium uh computation which which renders everything pretty complicated so how is this opport cost of work computed so so  
[01:25:23] very practically um we take the Matrix of trade flows The Matrix of gdps and the model is then forget about the input output linkages for a minute the model is telling us  
[01:25:37] that a change in trade costs and a change in productivity you got a new Matrix of trade flows a new Vector of GDP and that's the new state of the world okay so that's that's what we do in trade with those quantitative trade  
[01:25:50] model now how to do that for a war well we put those numbers which are of course rough estimates uh for the precise us China  
[01:26:03] pair taken in 2018 and we recompute so for instance it's a little bit like what what you were having the US China asymmetric tariff so USA is cutting uh  
[01:26:18] there's no retaliation but it still has an effect so it makes USA not nomal wage increase because you export more including to yourself so that your nominal wage increase the nominal wage of China in the first round decreases  
[01:26:33] but that makes China a better competitor everywhere because its wage decrease so it's increasing its market share everywhere in the world so it's affecting everybody so the wage of everybody's moving and the trade flows are moving and when it stops moving  
[01:26:47] you're done and you get the opportunity cost of war and that and and we extract the welfare cost for all the country countries in the world actually and I just showed you one here sorry a long I've seen at least two  
[01:27:01] more questions on this side uh yes yeah this is Hamid Fus from University of Rochester a very nice paper so I was wondering like should we really interpret the war as a military  
[01:27:14] war or can we relabel everything as a a trade War meaning that if like some trade war begins would would that like continue more trade War because I'm asking this because in that trade War  
[01:27:28] you wouldn't expect like maybe necessarily productivity losses or like like uh trade cost increasing with like third party countries this is why I'm  
[01:27:39] asking like can we relabel everything or not yeah I I'll be a little bit  
[01:27:50] provocative mhm so um I follow you on the outcomes of the model given the utility structure you use right so uh leaders care for the struction of of people and capital for  
[01:28:04] trade and output but unfortunately leaders don't always care for this right so it's a little bit the points you also made as discussant and um I mean we know that  
[01:28:17] the single best indicator historically for countries not going to war is their boss democrac racies right because if there are democracies the utility factors which you site play a role for the decision-  
[01:28:31] making MH and if they don't or at least not to the extent they do in democracies then the likelihood of war is much higher so I guess you you accept that uh not being covered in the model but I think it would potentially I know it's  
[01:28:46] very very difficult but it would be interesting to enlarge the utility uh factors which affect these decisions in order to um to capture these because unfortunately the problem which we face  
[01:29:00] as as I would say the global West yeah with Russia and China is that we're not sure the other side is really following the Assumption of utility which you site right and therefore the payoffs for the  
[01:29:14] risking are also different to what you predict in your model right they might be there might be much higher because the risk of that war being started from the other side is much  
[01:29:26] High okay yes verge of thank you on the verge of  
[01:29:43] going to war and then think about the economic determinance of that but A Simple Theory you want you kind of would like to explain the world to um to you right and and I'm grappling with um whether the model can explain peace  
[01:29:57] because most countries do not go to war with each other right and if I understand correctly the probability of war is determined by the difference in the opportunity cost of War so if I think of a big country like the us or  
[01:30:11] China and then Visa V almost any small country uh wouldn't the model predict that they go to war because for the bigger country the opportunity cost is  
[01:30:23] uh much lower or or did I not understand correctly here do you want to take on more questions or move other questions to the coffee  
[01:30:37] break it's you're the boss we have um okay I will say let's move other questions to the coffee break then you have some time to response and I can respond to those ones at least perfect thank you  
[01:30:52] um okay so is war military war or a trade War I think in order so yes you can interpret that as as as a trade War I think um with the  
[01:31:05] kvat that it's going to be I mean it you would need to have a trade model where the tfp which is here like a complete  
[01:31:18] residual I mean there is in the fact that there's input output linkages is going to make this the the country less productive because the cost are of the inputs is growing but not the tfp which  
[01:31:32] is really okay you destroy capital and and Etc so I think you would you would need to cut cut this this channel for this to be a to be a to be a trade war and of  
[01:31:45] course the loss of uh workers um but yes you can you can have this thing as a more General conflict that the P military um but the fact that  
[01:31:57] the dis risking is so linked to the geopolitical risk I think makes it the obvious uh the obvious application but you can say the dispute about China and  
[01:32:10] um and U and the USA is about um not agreeing on what is the good level of trade and and and there there you can say okay I'm going to put to 100% of  
[01:32:25] everything you you you do and that's that's the that's the the threat um so the fact that the leaders in democracies and autocracies have uh  
[01:32:38] different weights I think in in the model the way it can be handled right now um is is this um ulda which is the uncertain cost of a war okay so this the  
[01:32:52] the pure economy and then there's this additive thing which is the there's a resource and there's this uncertain amount of the resource they going to lose and you can view this resource and it's a utility of  
[01:33:06] the decision maker right so you can view this thing as having a this uncertain part as having a gigantic uh share and you can even make it like for um autocratic leader it might actually um  
[01:33:21] this utila might be negative it looks it's good to be a to be a chief of war and it to to look strong Etc so that I don't think there's fundamentally an impossibility to treat that within that  
[01:33:35] within that setup um we would need to empirically think about um how to how to maybe estimate this type of uh this type of uncertainty and this type of uh  
[01:33:49] non-economic gains and losses in autocracy and democracy because yes this this there this this Democratic peace hypothesis is like there's never been it's it's a little bit more uh restrictive than that there's never been  
[01:34:02] a war of intensity five the one I've showed you between two full democracies so it's it's the the political scientists have found this thing there's never been a war of of full intensity  
[01:34:16] between two full democracies there's been a lot of Wars between democracies um a lot of bad things happening between the um and the last uh uh question um can  
[01:34:30] you uh explain a dispute between asymmetric countries um uh yes I think I think here the again the US this this this part that is uncertain in this  
[01:34:43] thing the economic losses would be very um unequal which means that in a sense a small country has to give way to the USA diplomatically right and but we  
[01:34:57] see that the small countries that depend a lot on the USA uh have because because they're they got bilateral uh asymmetric thing they obey to the USA when they asked to I mean and and same thing for  
[01:35:11] France I mean I'm not targeting the USA so I think we see that okay so now I think we need to wrap up the session otherwise we have a war for  
[01:35:24] the cake um thank you very much for being here thank you for everyone participating after the lunch thank [Music] you yeah just quick announcement the ban  
[01:35:39] at hams prize will start at 3:45 in the other big room um yes this will be the next point on the agenda