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.
=

# Optimal Tariffs with Geopolitical Alignment

Authors:   
Discussant: None  
Video: https://www.youtube.com/watch?v=otRW31cJ3jw&t=0s


## Talk (00:00:00 – 00:22:05)

[00:00:00] Good morning everyone. Uh welcome to the session on trade and tariffs. We have a f we'll be starting our session with the first presentation by uh Johnston uh from Princeton and he will be presenting optimal tariffs with geopolitical  
[00:00:15] alignment. >> Okay. Thanks so much. Thanks so much for having me and to everybody for waking up early. Um we don't have a lot of time so let's go into it. So I don't have to convince anybody in this room that we're moving into a world which looks like after many years of cooperative trade  
[00:00:29] policy increasingly uncooperative and also in which the great powers of the world so here I'm mostly thinking about the US and China seem comfortable using tariffs as an instrument for you know geopolitical means. So what we're going to try to do in this paper is to  
[00:00:44] integrate these sort of geopolitical motives into an otherwise very standard and classical theory of optimal tariff setting. Um so you know the questions we want to answer are sort of what are the incentives of these great power countries in designing their optimal  
[00:00:58] tariffs when in addition to terms of trade they also care about the incentives they provide to small countries via carrots and sticks.  
[00:01:08] This type of theory needs a couple of sort of new ingredients. The first is we need some concept of geopolitical alignment in an economic model. So when I say alignment today, what I want you to think of is things like uh a small country may allow a large one to have a  
[00:01:21] military base in its country. Uh it may give up some autonomy over its domestic policies. Um or it may lend some legitimacy to a large regime. These are things that you know great powers of the world like. Um but there are things that  
[00:01:35] for small countries have both benefits and costs. So on one hand a small country may benefit in terms of military protection, but it may also lose in terms of autonomy.  
[00:01:46] The second key ingredient that we need today is the idea that a tariff can be not only an economics instrument but also a geopolitical one if it's used as a carrot or a stick. So I'll be more precise about exactly what we do in a minute but for now I want you to think  
[00:02:00] about um you know either a lone hegeimon or two competing great powers each of which offer one tariff rate to the countries that align with it geopolitically. So maybe a free trade agreement and on the other hand apply some other tariff rate to those that  
[00:02:14] don't align with it. So it provides some geopolitical incentives through its economic uh instrument. And what we want to understand in this world is you know are geopolitics a quantitatively relevant rationale for tariff setting?  
[00:02:28] Do they compete with sort of terms of trade things that we've understood before? And in a world of rising geopolitical tensions, how does, you know, this now theory that sort of endogenizes the geopolitical component of tariff setting, what does it predict about, for example, fragmentation when  
[00:02:41] geopolitical tensions rise? Okay, in the interest of time, I'm going to mostly skip the literature. You should think of this as take your classic optimal tariff setting, inverse elasticity rules, and such, and add one  
[00:02:54] twist to it, which is some geopolitics. All right? So the paper has a model which has two components an economic block and a geopolitical block. Um so let me walk through them now. The economic block is  
[00:03:08] basically a quasillinear armington model of trade. So the way it works is that there's many small countries a continuum of them some mass and within each of the small countries uh there's going to be some endowment of a quasi linear numer  
[00:03:22] and also some endowment of a differentiated uh unique product.  
[00:03:28] Then we're going to have some large countries. We're going to consider two different cases of the model. There's a unipolar case in which there's a single large country with some you know complimentary mass and then there's a bipolar case in which there's two large countries each with their own mass and the large countries economically  
[00:03:42] speaking are just going to be a bunch of little countries stuck together. So they're going to have some proportionately larger endowment of the quasillinear good and they're going to have proportionately more of these differentiated products that they can export to the rest of the world.  
[00:03:57] Preferences in this economy are going to be super simple and stylized. There's going to be homogeneous consumers in each economy and they have quasil linear preferences over this numer good and the differentiated varieties that they can import from all the other countries  
[00:04:11] around the world. All right. So that's sort of the economic block of the model.  
[00:04:15] Now let me tell you about the geopolitical block and then I'll explain how the two come together. So you know the key idea we need for a geopolitical model is that there's some notion of alignment. So for us in the unipolar version of the model that's going to mean that each small country has a  
[00:04:30] binary choice of does it align with the large country the hegeimon or not. And then in the bipolar version of the model there's going to be three choices.  
[00:04:37] Either you can align with one of the large countries call it home you can align with the second of the small countries call it foreign or you can align with neither. We're not going to allow them to align with both. That's just an assumption.  
[00:04:49] Now the small countries have preferences over alignment. These are, you know, for example, this trade-off between protection and autonomy that I was discussing in the introduction. And we're going to assume that for these small countries, uh, this can be either  
[00:05:02] a positive cost or a negative cost to benefit. And we're just going to represent that net cost by some ADA of JI for country I aligning with J.  
[00:05:12] There's going to be some distribution of this variable among the small countries in the unipolar case. And then in the case where there's two bipolar uh where there's two competing great powers, this will be some multi-dimensional distribution over small countries cost of aligning with each of the two  
[00:05:26] hgeimons. The large countries also value alignment. So in a case with a single hegeimon, we're just going to assume that they place some positive value on each of the people that they get each of the small countries that they get to  
[00:05:38] align with them. And in the case with uh two competing hedgeimons, we're going to assume that in addition to valuing when countries align with them, they also dislike according to this delta here uh how much small countries align with their rival.  
[00:05:53] All right. So the crux of the paper is sort of where these two economic and geopolitical blocks of the model come together.  
[00:06:02] So first let me tell you exactly what we do and then I'll explain sort of why we do it this way. So our approach is very simple and stylized. It's to assume that each large country has a single stick  
[00:06:14] and a single carrot. Its stick is that for all the countries that don't align with it, it applies a uniform MFN tariff rate to all of them. For all the countries, excuse all the countries that don't align with them, sorry, let me the the stick is that if you don't align  
[00:06:28] with me, I apply some MFN tariff to you and everybody else who doesn't align with me. If you do align with me, I give you a free trade agreement. Um, and then we're also going to allow each of the hedgeimons to decide, do they want to go about offering those free trade agreements or do they just want to apply  
[00:06:42] their MFN tariff to everyone? So there's a stick, there's a carrot. The stick and the carrots are the same for everyone.  
[00:06:48] Either you get a tariff, if you don't align with me, you get free trade if you do. Okay. So why do we do things this way? The first is we think from a theoretical perspective, this is sort of the simplest possible starting point.  
[00:06:59] It's going to facilitate a theoretical analysis that's straightforward because each of the hegeimons has essentially one you know variable that we need to take a first order condition over which is how large is its tariff that it applies to people who don't align with  
[00:07:12] it. The second is that you know from the perspective of the trade literature on uh you know international institutions this is sort of a minimal departure from the rules of the WTO. So we're maintaining here uh you know MFN that  
[00:07:26] you apply the same tariff to all the countries you apply tariffs to. We're maintaining article 24 which is that if I give you a deal it's uh all the way to a free trade agreement. And we're assuming that just one thing is breaking down which is cooperative tariff setting. So we're going to assume that  
[00:07:39] the two hegeimons uh you know set their sticks in a non-ooperative fashion not accounting for spillovers onto one another or the small countries. So, you know, let me sort of also address the elephant in the slide here, which is  
[00:07:53] this is not what the world looks like right now. So, the US is sort of picking off uh individual countries in the world right now and making different threats to each of them. So, I don't want to defend too much that this is how the actual world looks. What I want to say is just that I think right now we're in  
[00:08:07] a period of a lot of uncertainty about uh you know, in the next 5 years, 10 years, what will international institutions look like? Hopefully Bob is going to fix them all for us. Uh you know, what will trade regimes look like?  
[00:08:19] And we think this is sort of one of the plausible alternatives to consider. In the paper, we consider sort of a number of different possibilities for what are the constraints that the hegeimons are facing. Um, but for the talk today, I'm going to focus on this simple one where  
[00:08:31] you just have one carrot, one stick. All the small countries are just going to engage in free trade. They're not going to have their own tariff policies.  
[00:08:41] This is essentially optimal because they're small countries that don't have a lot of market power.  
[00:08:47] Finally, let me just explain the timing of the model. So, the way it works is that first the large countries are going to decide on whether they offer an MFN tariff and excuse me whether they offer free trade agreements and uh what their  
[00:09:00] MFN tariff is. Then small countries taking as given sort of what are their options are going to decide who are they going to align with and then finally that determines the tariffs that are facing each country. Everybody's going to interact economically and that  
[00:09:15] determines the economic component of their payoffs.  
[00:09:19] All right, so that's the model. Let me now sort of explain the theoretical part of our analysis.  
[00:09:26] Let's start with a unipolar case where there's a similar single hegeimon. So you can think of this as sort of maybe the US in the 1990s.  
[00:09:34] To analyze this problem, let's start from the second stage where you have small countries deciding whether or not to align with the US. Well, their decision because of the quasi linear structure of the model is basically that if they align, they're going to get a  
[00:09:48] free trade agreement, face no tariffs from the US, and that means they're going to get a certain price Q for their exports. Alternatively, if they don't align, they're going to face tariffs from the US, and that means they're going to experience a lower price for their exports. They care about the value of their exports and so they're trading  
[00:10:03] off that economic consideration that they want to earn more money on selling their endowment against a geopolitical consideration which is that they have that exogenous cost remember that ADA hi of aligning with the home country and so  
[00:10:16] for a small country it's just about is your alignment cost greater or less than this economic incentive provided by tariffs so now let's think through what that means for the hedgeimon who's setting these tariffs what is their consider  
[00:10:30] eration when they think about their first order condition of raising their tariff a little bit higher or lower. So there's going to be essentially two effects. The first additive term in this expression uh is about the inframarginal countries that don't align with the  
[00:10:44] hegeimon. Um and for them the considerations are sort of the ones that arise in classical tariff theory. So you see this first term depends on one minus g. That's the number of countries that decide not to align with the hedgeimon and therefore the people to whom the  
[00:10:58] hedgeimon's tariff actually applies. And you'll see in the term multiplying that that there's sort of two different considerations. The first is what trade economists call the volume of trade or fiscal externality effect that says when I raise my tariffs I lose some imports I'm going to lose some import revenue on  
[00:11:13] those marginal units I'm not bringing in anymore. That's a fiscal externality that's bad. On the other hand, if I raise my tariff it's going to lower the world price of my imports. That's a terms of trade effect. That's good. And so in your classical tariff analysis, you would just, you know, set those two things equal to each other and that  
[00:11:27] would give you your standard optimal tariff formula. What's different here is that there's a third consideration represented by the second additive term on the right hand side of the expression. It says in addition when I raise my tariffs, I also am going to lower the world price received by folks  
[00:11:42] who don't align with me. And therefore, to the extent that there is some people who are on the margin of deciding whether or not to align, so that's represented by G prime, the density of the alignment cost distribution. To the extent that there are some folks on the margin, I'm going to bring them over.  
[00:11:56] Now, if I'm a hedgeimon, I like that because I have geopolitical preferences represented by beta here that I want more folks aligned with me. But it also may have economic costs because I'm giving up on being able to apply a tariff to those people. That's represented by the difference between  
[00:12:09] this SA and SN. That's the surplus of aligned versus non-aligned people. So these are sort of the three considerations that come together when the hedgeimon's deciding how high should their uh their MFN rate be.  
[00:12:21] So in the paper we characterize what the optimal MFN tariff looks like for this sort of hegeimon. And the result is that first if you care sufficiently much about geopolitics then you're going to offer carrots and sticks instead of just  
[00:12:34] applying a single MFA MFN rate to everyone. And then second conditional on offering it. Your optimal tariff is always higher in the presence of these geopolitical considerations because you have an additional reason to raise your tariff now to bring more people in.  
[00:12:48] You're going to use a higher tariff. The more you care about geopolitics. And then finally, it's not only how much you care about geopolitics that affects your optimal tariff from this geopolitical perspective, but also how much the countries who are you're trying to align with you care about geopolitics because  
[00:13:02] that determines how many of those countries there are on the margin versus the ones that are inframarginal. And so formally what we show is that uh your optimal tariff is increasing in the hazard rate of this alignment cost distribution.  
[00:13:14] Okay, so that's what's going on in the case with a single hegeimon.  
[00:13:19] Next, what we consider is the bipolar case where there these two competing great powers. So, in this case, the decision problem is a little bit more complex for these small countries. They have to decide first, do they want to align with home? That means they get a free trade agreement from the home  
[00:13:32] country but face tariffs from foreign. Second, do they want to align with foreign? Then they'll get a free trade agreement from foreign but face tariffs from home. Or they could align with nobody, face everybody's tariffs, but they don't have any alignment costs.  
[00:13:45] And so the way that we like to visualize their choices in this setting is in this figure I'm showing here where on the x-axis I have countries alignment costs with home. On the y- axis I have their alignment costs with foreign. And you see I've broken the figure into sort of  
[00:14:00] two different regions of countries that are going to decide to align with home align with foreign or remain unaligned. Intuitively if you have high alignment costs with everybody you're not going to align. If you have loan alignment costs with for with home but fairly high alignment costs with  
[00:14:13] foreign then you align with home and vice versa. And so what this means for optimal tariffs is that uh you know we arrive at an optimal tariff formula that's theoretically closely related to what we have in the bipolar case. But whereas in the excuse me what we had in  
[00:14:28] the unipolar case but whereas in the unipolar case there was a single margin of adjustment now there's three different margins represented by the three different borders in this figure along which my policy can theoretically change how people align and that means that there's sort of three different  
[00:14:42] hazard rates the hazard rates along each of the borders in this diagram that we have to factor into our analysis.  
[00:14:48] Another question you could ask is in the presence of a competing you know US and China are tariffs complements or substitutes that's going to determine sort of what the reaction functions look like for each of these countries. Uh in the paper we characterize these reaction functions and show that there's many  
[00:15:02] theoretical channels through which if China uses a higher tariff the US might want to use a higher or lower tariff but quantitatively we find that those interactions are very small. Uh but now let me go to the quantification and tell you about those sorts of results.  
[00:15:16] All right, so in the last five minutes, let's talk a little bit about, you know, what remains to do and how we're going to quantify the model. So, so far we have this very stylized model that hopefully lays out some of the new trade-offs for a great power when  
[00:15:29] they're using their optimal tariff as a geopolitical instrument, but we haven't spoken yet to some of the motivating questions like, you know, are these alignment considerations quantitatively important? or if there's an increase in geopolitical preferences for these great  
[00:15:44] powers, how's that going to affect trade fragmentation? So to that end and with much humility, we're going to try to calibrate this very simple model, not with the idea of getting precise magnitudes, but with the idea of sort of understanding how qualitatively  
[00:15:57] important might these considerations be. Okay, so you know, we're going to calibrate the two different versions of this model to two distinct time periods in which we think they're relevant. So we're going to calibrate the unipolar model to the 1990s with the US as the  
[00:16:12] lone hedgeimon. We're going to calibrate the bipolar model to 2023 uh with the US and China as competing powers. Now the economic component of the calibration is going to be pretty standard. We're going to take an off-the-shelf trade elasticity and we're going to set the  
[00:16:25] ratio of quasillinear goods to differentiated goods and countries endowments in order to match uh you know existing estimates of the gains from trade integration. So the more interesting part of the calibration is  
[00:16:39] that you know we have to calibrate these some of these parameters that we as economists aren't so used to calibrating. So these are how much both the small and the large countries care about geopolitical alignment.  
[00:16:49] First we have to decide what alignment actually is. So we're going to follow a literature in uh political science and increasingly sort of geoeconomics and identify alignment with voting in the UN General Assembly. So for simplicity, we're just going to say you're aligned  
[00:17:03] with me. If you're in the top quartortile of folks who vote with me in the general assembly, you're aligned with the other hedgeimon if you're in the top quartortile voting with them. It turns out in the data those two things don't overlap.  
[00:17:13] So that tells us sort of in each of our periods that we consider how much of the time you'd have to vote with any given hedgeimon to be considered aligned with them.  
[00:17:23] Now, in order to plug this into the model, we need to know not only how many of your votes you'd have to change to reach the level of voting similarity that qualifies you as aligned with me, but also what economic value, what dollar value you place on having to  
[00:17:36] change those votes. And so the way that we do this part of our calibration is we take a votes per dollar constant from this literature on UN vote buying which is essentially tried to use uh you know foreign aid offers to figure out how much you'd have to offer countries say  
[00:17:50] as a fraction of their GDP in order to get them to vote with say the US. So given this constant and given a you know measurement of for each country how much how many votes they'd have to change to be in the top cile of voting with you we get a number for what's the cost of  
[00:18:03] aligning each small country. Of course, this is very speculative uh but I think the magnitude sort of passed the smell test. So, for example, for Malaysia, which is a country not aligned with China, but close to it, uh we estimate that bringing them to uh the top cile of  
[00:18:17] voting similarity with the US would require giving them three and a half% of their GDP a year. Now, you may not know whether that number sounds big or small.  
[00:18:24] One of the benefits of working with Gene and Elhanan is they have these great rolodexes of all the top uh political scientists and international relations scholars. And so they emailed their 10, you know, closest friends in those fields. And I'm here to tell you that the literature on uh international  
[00:18:38] relations thinks that's the right order of magnitude.  
[00:18:42] Okay. So in the paper, we have some of these fun figures about what are the distributions of alignment costs on these two different time periods. In the interest of time, let me just move to uh the second part of the geopolitical calibration, which is how much do the  
[00:18:54] large countries uh value geopolitical alignment. Here we're going to do something else very straightforward and we're going to assume that uh the reason why countries have military spending is because there's other countries around the world that are not aligned with  
[00:19:09] them. And so we're going to say well one rough estimate is how valuable it would be of how valuable it would be for a hegeimon to get other countries aligned with it is that it would be able to reduce its military spending by whatever share of the military budget is  
[00:19:22] currently devoted to that country. So, uh, using US and Chinese military spending, we're able to back out numbers for how much the US and China value alignment. Uh, again, we think the magnitude sort of passed the smell test here. So, for example, the value to the  
[00:19:37] US of aligning South Korea is somewhat larger then, but on the same order of magnitude as the amount that the US actually spends on its troops deployed in South Korea.  
[00:19:47] Okay. So, in the last minute here, let me tell you what we actually find from this quantification. So first you know at the beginning I said we wanted to know are these alignment considerations important for trade policy or does terms of trade really dominate everything. We find that qualitative the the alignment  
[00:20:01] considerations are sort of first order. They're as important as uh you know the terms of trade considerations in the literature. Second uh suppose that rivalry increases uh you know China comes on the stage when we move from our hegemonic to our  
[00:20:16] bipolar calibration. What does that mean for non-ooperative tariffs?  
[00:20:21] First, I've already mentioned that strategic complimentarities are minimal.  
[00:20:24] So, the fact that China may raise its tariff a little bit doesn't affect how much the US wants to use for its tariff.  
[00:20:29] But we learned something while we were doing the calibration that is understood in the uh international relations literature, which is that at the same time as China has risen, uh preferences have changed a lot as represented by UN voting data. And in particular, UN voting has polarized a lot. So, that  
[00:20:44] means that uh there's fewer countries on the margin that the US can hope to bring in when it raises its tariff a little bit. And that means that we estimate there's much smaller uh you know optimal tariffs for geopolitical alignment in the you know current period than there might have been in the 1990s. Okay, I'm  
[00:20:58] out of time here. So in the paper we have many counterfactuals. What if everybody cares about uh you know alignment more? Well, that's going to raise tariffs. That's going to increase fragmentation. Uh but I'll leave the details for the paper. Um so you know  
[00:21:12] big picture here we think that non-ooperative trade policy is unfortunately the way of the future. We want to start incorporating geopolitical considerations into standard optimal tariff analysis. Uh you know this is of course just a first step but we hope that it's you know helpful in thinking  
[00:21:26] about some of the trade-offs involved and giving you a sense of that the magnitudes matter. Okay. Thanks.  
[00:21:37] >> Thank you. We now have Albert Martin uh discussing the paper.  
[00:21:54] Do I have to sit there? Can I >> um you can sit uh if you want because the pointer is not working. So you will be using whichever way is comfortable.  

## Discussion (00:22:05 – 00:32:45)

[00:22:05] >> Okay. So thank you very much for inviting me to discuss this uh nice paper. This is a paper that goes back to the theory of optimal tariffs. Just to remind you, it's based on terms of trade. So basically the idea is I put tariffs on imports. This pushes imports  
[00:22:19] towards the world market and this reduces their price in the international market. And so I have a trade-off. In a sense, I distort my quantities but I shift terms of trade in my favor. What the paper does is add terms of trade geopolitical consideration and ask do we want to use trade policy for  
[00:22:34] geopolitical reasons. So the framework is quite simple. John explained it very well. There's two blocks an economic block. So we have the world composed of a large egimon that has mass a continuum of small countries. All countries are endowed with a numerator good and with  
[00:22:47] varieties of differentiated goods in accordance to their size. And then we have preferences like the ones that you see here. We're are linear in the numerator and we value these varieties.  
[00:22:56] And to this they append a second geopolitical block. Basically countries can choose should I align with the edgimon or not. And the way in which it works is there's a cost of aligning with the edgimon a cost eta drawn from some distribution. So these heterogeneous across countries the edgeon in turn has  
[00:23:11] a utility beta for every country that aligns with it. Uh and in the multipolar world they also add uh this utility for every country that aligns with my rival.  
[00:23:20] Okay. And the timing is um the edge will announce tariffs. Countries will choose their alignment and then uh we'll trade and consume. Okay. So here is a graphical description of the world without geopolitical considerations.  
[00:23:33] What you see here on the vertical axis is tariffs. On the horizontal axis you see this parameters. on one end these countries that are kind of close to me in preferences. The others are very far from me. Absent geopolitical considerations in this world, I want to put one tariff which is the one that you see there. It's the tariff that  
[00:23:48] maximizes uh terms of trade uh gains. Okay. And then they say, well, now if what changes if I add geopolitical considerations, uh it's a it's a bit restrictive what I can do. I can either offer you zero tariffs in the simplest version of the  
[00:24:02] model or the MFN tariff. And the question is what will happen to the MFN tariff relative to the one I just showed you when I add geopolitical considerations? And the key insight of the paper is that this tariff will go up. Why will it go up? Well, it's very simple. Uh imagine I put myself at the  
[00:24:17] point where the tariff maximizes terms of trade consideration. Now I add geopolitical consideration. Well, if I increase it, I have second order losses on terms of trade but first order gains in alignment. And so at the end, the tariff ends up looking like this.  
[00:24:29] Without geopolitical considerations, you have a flat tariff and now we will have a zero tariff, a high tariff, and countries will sort themselves out.  
[00:24:35] Those that are closer to me will take the free trade and those that are farther from me in deta will go for the high tariff. Okay, so this is the paper in a nutshell. Then they go to a world with two edgons. The same forces are at work terms of trade, geopolitical, it's  
[00:24:49] very similar, but it's a bit more complicated. And so it's it's not clear which way the model goes. So they go quantitative. And there the key insight is that um these tariffs say between the US and China may actually be strategic  
[00:25:02] um substitutes as opposed to strategic uh compliments. Okay. So uh this is the paper in a nutshell. Great paper. Okay.  
[00:25:13] So simple model definitely timely. So I enjoyed it a lot. I'm not going to pick on the model on specific aspects of the model. I want to mention a few kind of thoughts that I had while reading the paper that may be useful for them and and for the rest of us. Okay. The first  
[00:25:27] one is a bit of a cheap shot which is the way we model geopolitics. Uh once again this is not aimed specifically at this paper but for us going forward how do we think about the strategic objectives geopolitical and the geopolitical interaction. Their model is  
[00:25:41] essentially one of voting. Countries have heterogeneous preferences etas of voting with the edgeon and the edgeon doesn't care. the more people I bring the same. I value you all the same.  
[00:25:50] Okay? And that's fine as a start. But if you think about geopolitical interaction, it's a lot more uh nuanced.  
[00:25:56] So think of the US. I have many dimensions of strategic interaction. I want to contain China. I want to uh guarantee supply of some inputs. Maybe I want to have a foot in the Arctic once the ice goes away. And this means I deal with very different groups of countries.  
[00:26:09] I value them differently. And I'm bringing this up because we need to think about it, but also because trade policy seems quite coarse. this uniform tariff for everyone seems like course when I'm dealing with countries in different dimensions. Okay. So one thought that I had while reading the  
[00:26:23] paper that brings me to my first comment on the model is that is trade policy the right uh instrument for this? Okay. So uh uh Gman and Helpman had this great paper showing us how you can use tariffs to distribute domestically for political economy reasons. This paper is related  
[00:26:37] but it's about distributing across countries in some sense. I think it's very welcome. But one thing that I was thinking about reading the paper is why use tariffs? Why don't I just pay countries directly for theiras to bring them in my sphere? And one thing they could do very easily is look okay  
[00:26:52] imagine I allowed for site payments. How would they be combined with tariffs if at all? I suspect that the difference will depend on the model of the WTO that you consider basically how much can I discriminate with tariffs that makes tariffs very unappealing if I cannot  
[00:27:06] discriminate at all and also on the distribution of this G parameter. But more broadly, they could shed light on, you know, in which kinds of circumstances we're likely to see trade policy versus payments used for geopolitical reasons and which kinds of may prefer to use trade policy or site  
[00:27:20] payments. So, China notoriously is well known for using FDI concessional loans.  
[00:27:26] This seems to be a tool that they prefer and this model could shed light as to why some countries would prefer one and not the other. Okay. The other one he maybe John touched on it and this related to maybe what Bob will talk  
[00:27:39] about later but is what does this do to the geopolitical to the rules based system of of trade. Okay. Um this is a world where geopolitical tensions are potentially strong the ones that they mention but at the same time  
[00:27:53] countries are sticking to MFN. They're very well behaved. And you could ask yourself well once we live in this world do we value MFN at all or do we want to throw it out the window? And actually the model can be useful I think can be used to answer this question because on the one hand of course if now I live in  
[00:28:08] a world where I want to give differential tariffs to different people to bring them into my sphere this makes WTO very costly but of course my rival can do exactly the same and so it's not obvious whether on the one hand I want to reinforce these rules or break away  
[00:28:23] from them and if I want to break away from them what are the countries that are most more likely to break away from it first and this will depend I suspect on parameters the distribution of fetas, the betas um and the deltas.  
[00:28:37] Okay, so these are my first comments. You know, what are the objectives? Is this the right instrument? And can we use this model to think about the world trade system going forward? Then let me see a little bit about the numerical exercise. Not because I'm big on  
[00:28:50] calibration, but uh because they do something interesting and they have some interesting results that I I would like John to to comment a bit more on. The the first thing is that they find that when we want go from the uniolar world of the 90s to the multipolar world of  
[00:29:04] the 2020s, the US should have declined tariffs significantly. And the reason is quite simple. They find in the model that if we go to the 90s, a lot of countries of mass was close to the US.  
[00:29:13] So there I have an incentive to raise this punishment to bring them into my camp. But if I look at preferences in the 2020s, they appear to be biodal.  
[00:29:21] Many countries are too close to China. I'm not going to attract them anyway with this policy. So I should have reduced tariffs. Is this consistent with what we saw in the data? A second point is that one reading of the paper is that geopolitics doesn't matter much at all.  
[00:29:35] Why why am I telling you this? I would like John to comment because in the paper it's true. You raise the MFN tariff a lot but you also offer many other countries zero tariffs. When you average all this out actually the effect on trade is quite small. If you go from zero geopolitical considerations to very  
[00:29:50] strong geopolitical considerations, trade falls by about one one and a half percent. So it's not that much of an effect. On the other hand, maybe that's a good thing. You know, the IMF and others have been telling us world trade is not falling, but it's being fragmented into camps. And this seems to be would be one reading of this. What  
[00:30:04] they're finding would be consistent with this. And then let me add one last comment on the calibration, which is the way they calibrate this ETA this this utility of aligning with the edgeon.  
[00:30:15] That's not an easy parameter to calibrate. The way they do it is they go to UN voting and they say, "Oh, there's this literature that tells us that countries that vote with the US receive like 40% more foreign aid. So, they're being bought off." And this is a measure of how costly it is to how much I have  
[00:30:29] to pay them to come with me. But this goes back to my previous comment. You know, this that means we live in a world where alliances are being bought through transfers. And so, somehow these transfers are interacting already with trade policy. I would like to know how  
[00:30:41] you think about that. Okay. And one last comment is I want to do some self-promotion. It won't be completely uh useless but this is a model where uh alliance is  
[00:30:55] completely bilateral. What do I mean by that? If I go to the edgeon my gains from doing so are independent of what other countries are doing. Okay. So we have this related paper with Fernando Broner Josephine Meyer and and Kristoff Kristoff Treves  
[00:31:09] which is you know very different but similar in some regards. It's a paper where um alignment in policies is multilateral. Alignment in policies increases gains from trade. Everyone including the edgeon chooses their  
[00:31:22] alignment. And the difference with this paper is that in that world whenever I go closer to the edgeon I align with the edgeon erase the benefits of everyone else of aligning with the edgeon as well. And so this generates strategic complimentarities multiplicity and so  
[00:31:36] on. Now why is this interesting? when I read John's paper, why did I find a connection interesting is you know we find historical evidence that's consistent with our mechanism but more broadly you know we have a narrative there of the multipolar world which goes a little bit as follows when China is  
[00:31:51] small it will align with the US because it's too small to do something very different it won't attract countries but one China becomes large enough to assert itself then it may choose a different path why because it understands that it's so large that it will draw countries with it now one thing that  
[00:32:05] they find when they parameterize this distribution of ETAs is that in the 90s they used to be by unimodal. Countries were close to the US. In the 2000s part of the reasons for which they find what they find is that it's become biod  
[00:32:19] according to our worldview that could be endogenous. It's not by chance that we went to a biodel distribution of fetas is that maybe once China became large enough and China asserts itself well lots of countries starts aligning with China. Uh and it's part of the same uh  
[00:32:34] process in a sense. Okay. So overall uh great paper. Thank you.  

## Q&A (00:32:45 – 00:38:59)

[00:32:45] >> Should we collect some questions? >> Uh I would like to open the floor for questions. Um yeah.  
[00:33:07] Maybe while we're waiting, I'll reply to a few of Alberto's points and then we can take questions. Um, so first, thanks so much for the discussion. This was great. I'm just going to reply to a couple things. One, you emphasized a lot this idea of, you know, there might be multiple instruments like what if you  
[00:33:21] have transfers that you can do as an alternative carrot. I think you're spot on and as you point out, that's even implicit in our calibration of this vote buying literature. So we have an extension uh where we consider the case where a hedgeimon can offer not only  
[00:33:35] carrots and sticks through trade but also they can give you a transfer uh if you align with them and what we find there is that in general the hedgeimon wants to use both uh the reason is that at the margin when it starts from its  
[00:33:48] optimal tariff you might think it would only use transfers because in globally they're more efficient but from the hegeimon's perspective when it deviates a tiny amount from its optimal tariff that has only a second order cost from it but creates incentives for other countries to first order. And so in  
[00:34:02] general, it's going to want to use both. But I I really like your idea that we should think more about that. And this question of which countries should use, you know, transfers and which should use tariffs would be super interesting. And maybe we can speak to US versus China. I love that. Um,  
[00:34:17] you know, there's some of these other questions about um how valuable is maintaining MFN in this sphere? You're right. It's sort of not clear and it depends a lot on uh the ways that countries are competing here as you say you know trade globally may become more  
[00:34:31] open because countries now offer free trade agreements to the people who align with them. Um it's also going to depend on uh or how open you get it's going to depend on the rules. For example, if countries were allowed to offer um you know different carrots and sticks to  
[00:34:46] each country individually then as they care more about geopolitics well their stick is always the same. They can always threaten anyone with otterarchy, but when they want to sweeten the deal more, they're going to offer larger and larger trade subsidies. In that world, you might think actually you'll get more trade the more we care about  
[00:35:00] geopolitics. Um, but so I I agree there's a lot to think about in that space. Maybe let me stop there, take some questions.  
[00:35:07] >> So quick question. Um, cool paper. I was just thinking about whether this model can help us think also about what's going on today with the US increasing tariffs on allies and the fact that sometimes it looks like the countries  
[00:35:22] are paying to participate in the US uh in sphere of influence rather the rather than the other way around.  
[00:35:33] >> Do we have any more questions? >> Yeah, sorry John. Um, so I guess to push back a bit on your answer to Alberto, the reason that you get that effect is because you've got  
[00:35:47] one market where you can pay, you know, where you can use payments, which is the geopolitical market, and you've got one market where you've got to do standard monopolist extraction. if you could  
[00:36:00] extract lumpsum transfers in both markets, you would I think always just use non-distortionary side payments. And so like it I I would be interested to push you a little bit there on once you start  
[00:36:14] integrating these these tools into the geopolitical market, to what extent do they start also integrating into the economic market?  
[00:36:28] Yep. I want to follow up on Rodrigo's question. Uh great presentation, great discussion. Uh but I also agree that um that if one goes and reads Steven Myron's uh blueprint, uh he's very  
[00:36:42] explicit about saying uh we are going to make sure the countries don't retaliate against our tariffs by threatening to withdraw security arrangements. And in some sense that that does kind of turn things on its head relative to your  
[00:36:54] paper I think in the sense that uh that the US is basically using uh using the the threat of of problems with security as a as a cudgel to raise tariffs against its allies and keep them from  
[00:37:08] raising terrorists against it. So that I think if I think that was Rodrigo's question and I I totally feel the same way. uh maybe your paper actually has that element in some way or maybe it could have it but it would be an interesting thing to think about.  
[00:37:20] >> Yeah. So thanks so much for these questions. Um you know I so I the way I interpret the question from both Rodrigo and Bob is that uh you know the what I'm calling a  
[00:37:33] geopolitical action by uh these small countries of I want you to align with me may actually be an economic policy by those small countries like I want you to not use tariffs on me or I want you to  
[00:37:46] make a big investment in my country. Now how are the results going to change for that? I think it basically depends on how complicated of a model you're using.  
[00:37:55] So in our setting, we have all these quasi linearity assumptions so that my decision about the tariff I'm going to apply to you isn't going to interact economically with the decision of that the benefit that I get from you using a low tariff on me is going to be  
[00:38:09] separable from the cost and benefits of me using a large tariff on you. And so I think our results would go through if instead of geopolitical alignment that I'm trying to coersse you into, it's your tariffs on me or some transfers that you make me. But you know in general equilibrium in a more  
[00:38:23] complicated model those two things would be related. It would be interesting for us to think through that more. Um Chris great point. I don't have a great response but I'd love to talk more with you about it and think about you know depending on the ways that your trade  
[00:38:38] and transfer policies are allowed to be interrelated. What do you actually end up using? Thanks.  
[00:38:47] >> Thank you so much. >> Thanks everybody.  
[00:38:58] We'll move on to the next presentation for today um by uh Robert Ster um on geopolitics and the world trading system.