A Theory Of Economic Coercion And Fragmentation

- Alright, well welcome to the EPWG and we're in the, the fancy Annenberg, which is always nice. So we are delighted to have Mateo Maori, who's from Stanford GSB, and he's gonna talk about a theory of economic coercion and fragmentation and he's, it's gonna be economics rules interrupt with questions whenever you want.

- Fantastic. Well thank you for having me here. In fact, this seminar is the first seminar I gave a Stanford when I joined in 2019. John Taylor immediately invited me to present and George Schultz was still alive then he was with us and he actually sat next to me. So it was really an honor. And then COVID hit about 20 days later, so it was a one and done, but now I'm back. So we'll talk about this paper which was written a while back, maybe like a, a year and a half, two, but now feels even more relevant than when we first wrote it. So we're gonna just jump right in. I used to have to somewhat motivate what this paper was about. That need for motivation has gone down a lot. So I'll do it very briefly. I got interested four or five years ago in the, in the area Geoeconomics, thinking about a Germanic countries using the strength that come from their finance and trade relationship to exert power abroad, either for economic goals or for geopolitical ones. And we wrote a first paper on the offense. In fact, that's the paper I present here. This paper is much more on the defense side. If these things are coming your way, how should you think about them? And in fact there's been a lot of policy emphasis recently on economic security and the risk of fragmentation. But what those objects are, why on earth do we need a government involved? Those are still open questions and to me when I started, I had a few things on my agenda. One was how should countries do it? First of all, is there a role for policy? And if there is a role for policy, what should it look like? And trying to think whether there was a deep trade off between economic security and the standard gains from trade arguments in economics that were familiar with, or whether the trade off was more accidental. You know, these two things in some scenarios can be in conflict but not always. And then of course, if I had good answers to try to not only provide a sense of what optimal policy look like, but hopefully one where we could use models to do measurement to start to get a sense of what's big, what's small, what does a counterfactual look like? And national security, I would argue is an area that is ripe for more quantitative analysis because you know, Khrushchev used to joke that you can make almost anything national security relevant, including buttons or soldiers because if, if they need to hold their pants 'cause they don't have buttons, how are they gonna shoot? So I think, you know in the debate today that rings a bell and economists if, if they're half decent, we should be able to at least get rid of details and say some stuff is definitely not big enough to be relevant. There's some stuff where there is debates and uncertainty. So that's a game plan. So what we're gonna do today together is I'm gonna present a model. I'll try to keep it simple. I know that the audience has economists and non-economist. The model is a little bit on the heavier side of economics because it's optimal policy, but I'll try to make it intuitive and, and then we'll take two measurements. So the key ingredients are going to be a collection of countries with a global input output matrix. So they're trading things to each other and coercion. I'm gonna, in this particular paper, I'll make it very simple. I'm making a threat of exclusion from things that the hegemon controls. You can think of many other threats. The threats not to buy, the threats not to transfer technology yet. I'm gonna make it simple. If you don't do what I say, I'm gonna cut you off from some inputs that you get, you get from people that I can control and anti coercion, it's going to set be a set of exempt policies that take place in the rest of the world when they know that they might be bullied later and they're trying to shape up their economies to potentially withstand undue influence. The key mechanism that I'm going to highlight is that exactly the same exact things that we have put forward for many years to generate the gains from trade, things like returns to scale and specializations are both the source of those gains from trade. But unfortunately also the the source of dependency. The simple argument is as me and John Cochrane, we have increasing returns to scale in the standard, you know, Krugman argument. We're gonna vote, decide to specialize in one thing or another, we're gonna scale up and trade with each other. That scaling up produces efficiency. Now if you think about that argument, the part that I didn't scale the the life that I didn't choose, it's a poor substitute for the dominant one that got scaled up. So if John can hold me up now I'm relying on a pretty bad alternative. And that tells you that the source of the gains from trade is the same as the source of dependency. And he makes the trade off in policy between integration and dependency and interesting one to study. And the, the main question is gonna be does the private sector choose optimally and we just have to live with these things or is there a reason to do, to do optimal policy? I'm gonna show you that in this context. A hegemon, a big country is going to have an incentive to hyper globalize is one to induce integration that goes well beyond what a global planner or benevolent planner would want to do in as long as it makes it even more central. So it's a bit of a drug dealer model. It wants to create an addiction to dealing with the hegemon because then the threats of excluding any one player is very powerful. On the other end, there is reasons to be concerned about policies of anti coercion. If each country, I'm gonna show you that if each country around the table starts to do anti coercion policy in un accordion in fashion, we end up overly se secure but quite poor. So we overdo it, we enter into a doom loop, but we all want to fragment. And the more we fragment, the more each one of us wants to do it. And it's very inefficient. And then I'll take you to the data and I'll show you least what's

- Big on what's one regiment and the rest of the world. Yeah,

- In this paper there's a single regiment and it's gonna be totally exogenous. We're working on papers with multiple hedge funds and I can give you a bit of a sense. Yeah. Because

- Even the broad sweep of history, trade blocks have been kind the dominant equilibrium for long periods of time, probably driven heavily by military considerations. Yes. So in any event

- Here I'm not gonna talk about, here's,

- Here's an issue of now are we breaking the safe blocks?

- Yes. And you will see exactly in the equations where there's a national force for coalition formation. I'm not gonna study it but I'll, I'll sort of talk a little bit more realistically about it because the game theory gets fast pretty quickly.

- The increasing returns seems I'm, I'm trying to hard to think of where that applies. I mean cars maybe, but there's still like five big car companies and BYD came along and clean. It's gonna clean box in no time so that there's one

- Global. So yeah it, it doesn't have to be increasing returns per se. That's why I had a specialization think of for example, network effects. Strategic complementaries, you want to be on a payment system because everybody else is on it. That's gonna work just as well.

- Yeah. That's different from increasing

- Returns.

- Yeah,

- I, I put there increasing returns because it's the standard argument from trade for generating instant trade. Here's gonna be more general

- Question is trade on our planet or trade on some other planet

- We'll do one on our planet. Okay, so let's get started. So we have end countries since John has already started to question me, I'm gonna pick to be the heman and I'll pick the rest of us to be the end countries. You're not the only one that thinks of 'em that way. And each country has a set of productive sectors and a set of local factors Here, a productive sector is an activity taking place in one country. So think of like oil in Russia and oil in the US as being two different sectors that outputs might be quite close substitutes in the production function but a sector as a particular location. Each sector has a unit master firms and they're relatively straightforward. They use intermediate inputs from other firms potentially all around the world and they use the local factors. And to John's first question, you can see that function as a Z at the end. I'm gonna use what are familiar to economists externalities in reduced form like green wall STIGs, I'm gonna have a vector of aggregates that potentially affect the production function. So what are examples of that? An external economy scale is the more the sector produces, the more productive each firm is a strategic complementary in the use of an input is the more an input is widely shared across sectors, the more it is productive for anyone in the sector. Okay? Like a typical Facebook externality or payment system externality in general. I'm gonna carry them an abstract vector. And then in the examples I will specialize, okay? In each country as a representative consumer, the consumer is very straightforward. He cares about consumption potentially also cares directly about some geopolitics. I'm not gonna use it for today. So you can forget the second term of the utility function and it's a static model. So it's very straightforward. I have a consumption plan. My revenues are on the right side of the budget constraint. They are the profits of my firm and the factor income I embedded that you own your domestic firms. There is nothing wrong in this paper if I make you own some foreign firm. But what I'm not gonna do is do end dodging as acquisition. There's clearly gonna be an incentive to go and buy things around the world. I'm not gonna study that. Problem defines an indirect utility based on your wealth and the price vector. That's gonna be my welfare criteria. So it's gonna be very straightforward and markets have to clear everything that gets produced either gets consumed or gets used as an intermediate input and the factor market has to clear that's it. The economy is done, the rest is the policy game. So the policy game is what makes this paper very interesting at the end, everything I just told you happened, reduction and consumption in the middle is the offense. That was our first paper is take the world as given that is one big country exogenously the he him on John he has is special. He can do a couple of things. The first thing he can do is he can threaten foreign entities with a loss of input that he controls. What he asks in return is a set of costly actions. Namely it's gonna put wedges in everybody's first story conditions. Those in the in the model are gonna end up looking like quantity restrictions. It's telling a firm, I don't want you to sell to China or I don't want to input you to input from China. And you can also ask for direct transfers, which can be monetary or they can be political concessions if you want. Okay, so he offers a contract and I'm gonna make bargaining simple. It's gonna offer, I'll take it or leave it off. Okay? So I'm gonna give him all the bargaining power. The defense is the exempt period exempt. Every country is gonna make decisions on its domestic economy and the decisions here are gonna be a full set of wedges. So think of like tariffs, industrial policy, macroprudential policy, anything that we normally associate with the government over the medium run shaping the activity domestically. So one huge difference is every country can influence hiss own firms. John is gonna be able to go around the world and make offers to all the entities. Okay? These things are baked in. What we're gonna study is what what on earth happens. Okay? So let's look at this. The first thing, this comes from our first paper, it took us a long time in this literature to make concrete a notion of power, which also is very related to the notions in political science. In going back to hirshman. Now, what we found really helpful from a framework perspective were really two things to think of power as participation constraints and to think of costly actions as wedges in first order conditions. Okay? Once you have those two things, a lot of economic theories lot in, if you're a political scientist, they should remind you of the sort of definition of power by from Robert Dahl in the sixties. If A is power over B to the extent they can get B to do something he otherwise doesn't want to do. Okay? So let's look at it here. Each entity when it gets approached from by John in the middle is comparing an inside option to the outside option. The inside option is I produce unconstrained with the full set of inputs, but I have to take over some actions that the Hege and John is asking for. Those actions are gonna be privately costed to me and they're gonna be potentially, as I said at the bilateral level, like a very large set of space. And I might have to pay a transfer, I might have to give him some utility or some cash. What is my outside option? My outside option is I tell them that I'm rejecting the offer, he punishes me. I am losing access to a set of inputs. I can produce unconstrained with those set of inputs. So I don't have to, I do my private optimal given that set of inputs, but I lost a set of inputs. Okay? Very straightforward. Now if you think about the partial equilibrium, every one of us is gonna do that calculation. That's what I'm gonna call John's micropower. It's relational, it's take the world that's given. How much can he extract? What's gonna be more interesting to us is that when John asks for something, he's thinking about doing it with all of us contemporaneously and changing the aggregates, changing prices and the externalities. So he is manipulating the equilibrium in his favor. And I'm gonna call that one macro power. Let me just get to, I mean the objective functions are very straightforward. John maximizing his own utility as a hegemon, the utility of his consumer. He cares about the profits of his firms, the factor income in his country and getting transfers. The rest of us care about the opposite. We care about the profits of the firms that we retain, but of course we have to care about the profits. Net transfers. It's the part that we retain if we accept what he wants us to do. And potentially we have a set of firms that never deal with him and those go unconstrained. Okay? So they're very standard criteria for the purpose of today. I'm gonna start specializing so that we get close forms if you wanna see long expressions there in the paper, but I'm gonna skip them. I'm gonna make two assumptions. One is constant prices. Essentially I'm embedding linear utility in the consumers. I don't want to talk about terms of trade manipulation, I wanna talk entirely about manipulating the Zs. So to John's point, I'm, I'm, IM mean this planet. I'm not a huge fan of terms of trade manipulation. It's such an easy force to get in our models, but it just never quite rings a bell for me. Manipulating externalities. I think I, I see a lot more and I'm gonna tell you what incarnation they take and I said no, no direct political objectives. Pure simple economics on maximizing consumer utility. Okay, so here's how this paper works. So start with simple micro take affirm. I embed the constant prices. So the margin of revenue curve is constant at the price level. The production function at the micro is decreasing returns to scale. So the marginal cost curve is increasing. So the standard idea is the inside value of the firm. Undisturbed is the area between the two. Those are the profits. Okay? Now what is coercion? Coercion is a threat to exclude you from using some of the inputs. So in a weak sense, that's gonna shift the marginal cost card to the left. You're gonna do a little worse. It's gonna be more expensive to produce. Now how much is gonna depend on the availability of substitutes? Lots of other things that defines two areas. There's one area, the blue area, that's the outside option. That's where you retain. If the hegemon cuts you off, it's the best that you can do if you walk away from the hegemon. The orange area in the middle is the D, the gap between the inside option and the outside option. Now in the participation constraint, that's a sarus, that's a slack. If I gave you the threat and did nothing with it, that's the maximum value that the hegemon can extract. Before you tell me to that you're walking away, that immediately tells you what is the conflict between John and the rest of us. John cares about a difference. The difference between the inside and the outside option. We care about a level, we care about what we're left with in levels. Once this is all said and done and even if you change your bargaining protocol, that's not gonna change. Now that tells you that all l seql John is benevolent for him. Pushing the inside option, pushing the original car far out is good. It generates sarpa that you can extract. The problem is he is also perfectly happy with destroying value in equilibrium as long as it makes the marginal cost curve on the outside option shift farther to the left than the inside option. He's only caring about the distance. So it's gonna be perfectly happy to destroy value If it increases dependence, we don't care. We don't like that because we're left. Whatever we we have in levels, that's gonna be the main part that we're gonna exploit. Please go for it.

- So if the Heman says You can't buy my oil, he's worth, worth a worth. Yeah,

- Of course.

- Why would he expose to

- It? Yes. So from a, from a technical perspective in this particular paper, because the firms are small, it doesn't matter and it's easy to embed some form of commitment from the losses of kicking out anybody or infinit or for the hegemon. They're large for that firm and he loses credibility on all the other firms. That's not a great answer. If you want to go deeper into this, there's two ways. One which we already did, which is do the repeat game and sustain it with the future ability to keep threatening you. The part that we never did, and I will explain why and I think it's actually empirically relevant, it's not a trivial assumption, is to think of a world that when I cut you the full general equilibrium changes that that would be technically very difficult 'cause I would've to recompute a lot of equilibrium. We're totally bypassing that by making the firm small. That's not in nos. For example, if you're thinking of cutting the whole of China, China, you're thinking of like a large ana equilibrium of effect and you would've to recompute the entire model. I think that my argument is that what you get out of this, it's surprisingly a lot for, for something that is much cleaner. But that's the thing.

- So in this model,

- In this model is very credible, is very easy. We see

- People doing things that hurt them in order to the reputation.

- That's exactly right.

- Said you used Hirshman. Is there any royalty or voice in these

- Models? Ah, very good. No, this is the early hirshman. This is Hirshman in 45. When he wrote national power before he became a more general social scientist. This is him as an economist. So this was him studying how Germany induces dependency building into World War ii. I'm not gonna talk a lot about like the political economy aspects. It's good. I kept it quite intentionally closer to the economics. 'cause I guess I had more to say about it. But it's a good question. Okay, so what is anti coercion policy? Well, it's the opposite. You know that this has been done to you. So you're gonna try to take a set of policies, example that shield you. What are those policies or those that make your outside option a little better? You're trying not to create that much dependence. You're building a home alternative or you're building a relationship with somebody else that can substitute in exposed if you get that off. So that's gonna be the form of anti coercion policy. Okay, good. Oh

- Please. That's like, is there like an intuition for a policy implication like for a United Nations that's kind of economic point. Oh

- Yes, there is one very big one. I'm gonna give you this response argument. You'll have to wait. There's one very big one, but if I tell you now, it just steals it under and I, I would rather give it to you once you see the model. But yes, here the, the short answer is yes here there's gonna be a lot of role for commitment and international organizations are a form of commitment. You will see that here I, I've embedded that John to be very powerful. But if he's not careful, we'll fighting him so hard that he would've been better off committing to a rule. Okay, good. Please. Does the model change if you impose a complete trade blockade? No, it's just a quantity issue. If you say that everything I ever buy is from you, that's gonna be switching me to other key on the outside option. Okay, good. So to make it concrete, I'm gonna specialize the model even farther. I'm gonna talk about finance, I'm gonna talk about the part of finance that used to be as borrowing as it gets, which is the infrastructure stuff. Like how do we make payments? How do we cast the assets? Assets. Clearly now it's nothing. But you know, it's quite interesting because the US has used it very aggressively to sanction of pressure other countries. So you can think of, you know, swift and disconnecting targeted entities. You can think of the sanctions on Russia, the pressure on HSBC to reveal some of the dealings they had with Huawei has been used very heavily. So I wanna understand what what makes it such a good coercion tool empirically, but also it's an area that has a very clear externality, urines, wt, which is the me. If you don't know what it is, probably everybody does. But WT is the mechanism that we use to send messages to do an international wire. It's a typical thing with a strategic complementary because you only want to be on it because everybody else is on it. So it's like X or Facebook or whatever. Good. Okay, I'm gonna take the previous economy and give you a very precise intermission. The US as a single sector, it has workers that act as bankers and provide financial services. The rest of the world is a collection of small open economies. They're all identical. They really only have two sectors that they care about. They have a exactly photocopy of the US sector a home. It's the home alternative. They could have used domestic financial services and they have an intermediary eye that puts together the global service and domestic ones and sells them to manufacturing and the rest of the economy. So the sector that is gonna be crucial for me is going to be the intermediary. It's not, it's gonna be nothing else that the person who makes the decision. How much do we rely on a home alternative versus the global system? And that's all it does. Now, where are the externalities baked in The home financial service is gonna have a local externality. Those payments are more efficient the more we use them. The global one as a strategic complementary across countries for each country is more efficient to use the global payment system. The more other countries are also using. Okay, so it's a local versus global and that's it. Okay. Oh please.

- Can you use swift just for international transactions or are you also thinking of using,

- You can use them also domestically. Essentially what they really use is a bundle of the tool. So everything is on the intensive margin of how much of the tool are we using in the bundle. When you're on the outside option, because you're cut off, you can only use the domestic one. Let me, let me give you the precise answer, which you'll say it here. So what John is going after is in some sense what is the intermediary doing? So to, to those of you that are economists, it looks like a function that we all know it's a CS function. It's a constant elasticity production function. Let me sort of unwrap it. That's the intermediary in each country. It's taking as given the productivity of the systems and is deciding how much to use of the hegemon system X, I and J and how much to use of the domestic alternative x, I and H. Okay? And is bundling them together. Now the productivities of this tool are the two terms in red and blue. So if you look at the productivity of the red, it's the average of everybody else usage. So the more everybody else is using it, the more productivity is for me to use it with a parameter side that governs the extent of the external amount. So this kind of formulation is very convenient as been Newton growth theory. The domestic one in blue, you can see it's very similar with a different parameter IH, but it's only reflecting the domestic usage. It's telling me the more domestically the economy uses it, the more every single firm is gonna find it productive. And the things that really matter is the parameter sigma, let's say substitution. It tells me to what degree are the domestic and the foreign financial services or substitutes. So if you send Sigma to infinity, they're perfect substitutes. You're indifferent. Whether you use one or another as Sigma starts getting close to one, you're approaching cop Douglas. That's an interesting case 'cause it tells you that as soon as you lose the foreign services, the domestic ones are useless and anything with more complementarity of of cop Douglas is gonna have that fisure. So it's put really, you know, think of like when sigma gets a little too low, it gets you into a corner. All of a sudden if you don't have access to the foreign input, your domestic alternative is useless. Okay? And we're gonna, we're gonna return to this. This is a standard effect of all of these production functions, but I'm gonna use it quite happily. So it's good to point it out here. Okay, now take a pause. We're done with the model. So the rest is results. Okay, so first let's do two standard benchmarks. In international, we generally study two benchmarks almost all the time. One is a global planner, it's a planner that has access to all of the instruments in this case, in fact it's complete instruments. So it's as if he was choosing every input and level of production worldwide and is benevolent. Now what does benevolent mean in this context? It means that he puts weight on every country, but also I'm gonna choose those weights so that it doesn't have any redistributive motive. In principle, it gets no benefit from a transaction that moves wealth away from Valerie and to John. That doesn't change any of the EQU markets. Okay, that's pretty standard. Now in the general model, just so that you see the thoughts, what would that planner do? He would use a set of wedges through GP that maximize profitability here. Profitability, it's all value added, essentially GDP. So it's trying to make the pie as big as possible and it's doing it by affecting the externalities. Anything that has a positive externality, the planner subsidizes everything that has a negative externality. The planner taxes, that's a very standard result. The second set of equations shows you the close form solution for the particular specification I gave. What does that actually mean in that context? Here the planner wants to subsidize usage of both the global system and the domestic alternative. Both of them are positive externalities and so it just scales them up. Okay, that's totally standard, but it gives us a good benchmark to think about what the difference is. So the first thing I'm gonna do is I'm gonna put in a line, I'm gonna say global planner sees both positive externalities, gives them subsidies in proportion to the of the externality. The second benchmark that we normally study is what we call NASH policy. So there is no global planner. Every country is independently choosing domestic policy. So a full set of domestic wedges, potentially they, they can constitute tariffs, industrial policy, financial policy, just think of it as their complete instrument. So it's choosing the domestic allocations to production. But taking has given everybody's else policies. Okay? Now what does, what happens in that context? Well the top formula starts to be interesting. It tells you that it NASH policy, we get something relatively similar to the global planner, except you now only control one economy. So what are you doing? You're accepting subsidies thinking okay, if I put a subsidy or a tax on this activity ij, it will move the activity directly through the input output matrix of the entire world. That will affect all of the other country's choices potentially that will feed into the externalities. I only care about that fact in as long as those externalities feed back into my profits. So the only difference with the planner is the second part is that you only care about whether or not these externalities fall back into your profits. Okay? In the context of the economy that we just studied, I'm gonna also take the limit where I have infinitely many economies in the outside of the world. So there's more open economies, you can get a very sharp result, which is also a classic. Well the first wedge, it's identical to, sorry, the second word, the one on the right, it's identical to the planner choice. Why? Because the domestic externalities of the payment system occurs entirely within the country. The policymaker fully understands it and scales up the second, which is at zero comes from the following effect. What is standard in NASH policy is that you only see part of the externality if it's across borders here, by taking the end to infinity limit, I made you so small that you think you have no effect on the globals and therefore you never tax or subsidize, you just don't see it. So the more standard effect would've been you under tax a negative externality. Sorry, you yes, under tax a negative externality and you under subsidy a positive externality as long as as they're recording from cross border. Okay? So that's what I'm gonna call the, the non-cooperative outcome. I'm gonna stick them here. Okay? They give you two standard benchmarks. So what an economy like this looks like if we don't have a hegemon in the middle. Now let's look at what happens when I have a hegemon and when I have an hegemon plus antico urgent policy. So the first part is the offense. So cant have set their policies. We're in the middle and the hegemon is going around bullying people. What does he ask for? That's our, essentially our old paper is the theory of the offense. What does optimal coercion actually look like? Well, first we already said the participation constraint is there. So there are many places where you have no power. It's gonna be pretty obvious that if Valerie has a preferred outside option, if Sigma's very high, I got no power over Valerie. But to the extent that I have power, how do I spend it? What do I ask her to do? Well it tells you that I have two rationales, one of which is standard and one of which is totally absent from the standard theory. So the, the last part of that equation says domestic profit is pretty standard. Look, if you gimme a way to meddle, if you give John a way to meddle in our economies, one very standard motive for him is gonna be to tweak our policies so that they, we shape the equilibrium in its favor. If you think of the terms of trade, which are shut off, that's a classic one. The classic theory of an optimal tar phase. I put a tariff, you don't retaliate because you're gonna cut prices. I just manipulate the equilibrium in my favor. I manipulated one aggregate, which is the price vector here. I shut that off. But production externalities are identical. I will want you not to create a great alternative. It affects if it affects the profits from my firms. That's a typical industrial policy of it's just anti-competitive. I want to go and tell you stop doing things that compete with my firms. I want my firms to be monopolists that are identical to the optimal tariff theory. Formal, okay? They come from exactly the same place. The second component is totally absent from this textbook theory and it's about building power. Now what John is doing is asking me and Valerie to take actions thinking about how that feeds into both our inside and outside options and into everybody else. Inside and outside options is trying to look for actions that increase dependency. For example, it's gonna tell us, look, if you trade more with me and don't build alternatives and everybody else was relying on you, I'm very grateful it's not that expensive to you to do it, but it's terrible for everybody else. So you can see that here, the AMO is exploiting a gap between the private cost to me and Valerie of the actions we're being taken to take and the social value to him of how those actions reverberate on the equilibrium. And he's constantly checking what does he do to the inside option and what does he do to the outside option. Now if you make it very practical to the example I gave you on the payment system, what is the heman doing? Well the first thing he is doing is asking everybody to use his own system is incentivizing a usage of the global system. That's not that different from a partner. That's the part in which John is benevolent. It's saying, look, let's get everybody to use more of my system. It will make it productive. That will shift everybody inside options outwards. That's good for the world economy. The problem comes that he also cares about the outside option. On the outside option, John will say, I want you not to have alternative. I want you to scale down your own production. Why? Because by doing so, you're becoming overly reliant on my global system. So very much contrary to the planner or contrary to the NASH game, I'm actually discouraging the existence of alternative. So when the president says if the bricks build an alternative, I will punish them to hell. That's part of this strategy is I don't want you to create alternatives or the entire emphasis that we have on China building an alternative payment system and why that's bad for policy is because it manipulates the externalities. Because creates an alternative that a country like Russia could use if we shut them off.

- What extent is it? I'm, I'm making another alternative. So you are more vulnerable to my threats as opposed to making you not have an alternative. So that Valerie,

- Very good. It's it's both. And in fact, from your perspective, the second one is so much better.

- Yeah. 'cause you, if I I don't

- Internalize it

- Exactly. If I say you're more resistant by threats, if I say you make Valerie more resistant, then you don't care. That's

- Right. That's exactly right. And the way to think about, it's every time you ask me for something that I fully internalize, you're using up a lot of this luck in the participation constraint. You're using your power, you're spending it. But every time you ask me for something where the social value to you far exceeds my value because of the way it propagates through the economy, I don't care. I look at my social value, I, I still accept. But to you that might be very valuable.

- Like in the early years of the us, the UK did not want the US trading independently with other places. Yes. And you're saying that's not so much to support the UK zone thing, it's just we want the other places to be more dependent on,

- I mean I, I give generally two classic examples are, I mean at least in recent policy. One is why do we put so much emphasis on semiconductors? I mean we don't care so much about like what A SML does. We care about the fact that if you don't have those machines through the input output or amplification, we're potentially affecting a lot of the economy, including the military usage or the development of i amic conductors. Now to A SML, when you approach them, that's clearly bad news. But they're looking at their private value of not selling the machine. That's a fraction of the value to the US of China potentially not being able to produce high-end semiconductors. That's a typical example. It also gives you, you will see it in a second, a, a big conflict between the government and its own firms. Think of Nvidia. The way I think about Nvidia is Nvidia would like to sell semiconductors to everybody at a high markup. That's what maximizes profits. What is the US government optimal policy in these papers is sell to everybody at a much lower markup and threaten China for exclusion, right? You flood the market every, it's very hard to build an alternative. 'cause now Nvidia chips are so damn competitive. They're essentially being priced of perfect competition. It's very, for the shareholders of Nvidia, it might be great for the national interest because it makes it very hard for China to build. So that's why I, I like this theory. Once you get them in standard economic terms, you can start making sense of what the channels are. And then the question is, are some of these things big or small? So can we, we come back

- To the time horizon you're thinking of here across a lot of this stuff. Obviously the input output matrix is endogenous and people anticipate some of these things, right? I mean, China's been thinking about the US in dependence on the US for a long time and by we for certainly have for a while back in the nineties it was on anybody's radar screen. So that changes and who can produce what, what's, what substitutes actually eventually can be placed over what period of time is, is not indifferent to what these threats are.

- Those are great questions. So the answer is we allow for a lot of that. So lemme tell you what we allow for and what we don't allow for. So first future coercion is gonna be fully anticipated. So here you're, you know, example, you wanna start building industries and do all that. We'll see that in a second, second even exposed once I cut you off. This isn't a short run model. I let you reoptimize all your relationships. So if you lose the relationship with me, you're not stuck with a pre, you know, this is not, the Leman goes down and you can't find another bank here. You can, you can go to the other banks that you have relationship with and even scale them up. What it is that I'm mean. So in that sense, I would think of this as a medium run model. What it is that we're not allowing, the things that we're not allowing is that once you cut me off, I'm not allowing the domestic government to reoptimize policy. So the, the sense in which the policies are exempted is that they have to be inva to the inside and the outside option. Now when is that good or bad? It's great for thinking about the inside and outside option. 'cause otherwise you get into crazy things where I promise you the moon on the outside option just to get you not to accept a contract. What is bad is clearly over time there is probably more that we could allow including rebuilding relationships that we're not gonna study. Okay? Okay, good. Now let's think about the example part. What does each country knowing that that's what John is gonna ask us in the middle, what do we do? And here I cooked up the model to deliver an extreme here. What the countries do is they totally ban dealing with the hegemon. They put an infinite tax on using the hegemon input and efficiently subsidize domestic production. So it's a full fragmentation equilibrium. Now the reason why I did it comes up to, to the question that he asked at the beginning, and you will see in a second. Now in general, in the general theory, that doesn't have to happen. What tends to happen is that I'm trading off between the gains from integration and dependency and I'm gonna generally have a tendency to do less integration whenever it makes me dependent. So as long as integration doesn't generate any dependency, that's good for both of us. But if it helps you twist the knife later, Exenta will want to walk away and I will want to start moving away from full integration. Now why did I cook it up this way? Which gives you also a sense of how different the outcomes are from the other two benchmarks. Well, first is because it kept the home alternative constant and it gave me all the actions on what do you do about globalization. So it makes a very compelling case for how different the world looks under the different scenarios, but also because of the following slide. So I'm the typical economist and I, by presenting this paper, I realized that in this one I am median, a lot of my colleagues feel the same. If you had asked me how do you think as an economist about international organizations, I always taught them as an internation of a global planner. There a mechanism to get to a global planner implementation or maybe to a second best. But I, I would've start with the objective function is the one of a global planner. And then there are constraints on the instruments and implementation. I've totally changed my mind on this and I changed my mind in a way that it brought me a lot closer to the political science. And I'm, you know, I'm mischaracterizing these things, but as a caricature it's helpful. The political scientists have always had a view that the organizations are instead an commission of the hemen. That the IMF and the World Bank are two blocks from the White House for a reason. And if Zambia designed the IMF, it will look different from the way we designed it in the us. So they always had a view that the liberal order is not the absence of these issues, it's a particular incarnation of these issues. And that stuck with me in writing these papers. Why? And this goes to the question that was asked. Imagine, you know, the model, I cooked it up for a reason like this. John has all the power, it moves second exogenously, it can make, I can see that this light resonated with you or maybe you want it as a permanent record or how wrong I am. It's one of the two.

- No, it was like light bulb.

- Yeah. So many, so many of us at that moment were exposed is obvious. And I never thought about it this way. And so if you think about the way I, I rigged it. John has all the power. He moves second, he can bully all of us. But the problem is we're so scared of him. That ex example, we decide to walk away in equilibrium. He has no power in the simple incarnation I just gave you, he gets no revenue at all because we walk away. Now that's clearly a crazy example, but he makes the point. Now think of a world where John ante commits to limiting extraction. It's a rule that only binds him and yet it's beneficial to him. It's a form of commitment because once he has that rule, a few of us will come in and decide that it's okay to deal with him. As we come in, it becomes more attractive to everybody else. Alludes them into and equilibrium is gonna extract some revenue. And you know, to give you a very simple incarnation of this, you can do it formally. You can say, look, what if instead of extracting entire surplus, he says, I'm only gonna extract a fraction of the inside option and leave you the rest. That lets him to generate positive revenue. He extracts actually concessions in equilibrium because all of us, all of us are lu in. So, and I think that this is important. I mean for a second, let me step away from the confines of the narrow paper and just be clear. These are my personal opinions, not that, not necessarily the science, but it also led me to evaluate the messaging that we've been given to the general public on international organizations. So if we sell them as they're doing international, some globalist force, they're good for everybody else. It generates what we see now, which is a lot of resentment towards some of these multilateral institutions. And some of what these institutions did, by all means might have been too favorable to everybody else. But there is a bit of a narrative that comes from the fact that as economists, the first thing we do is we put them as the global planner. And instead there is an alternative one to think about, which is these are constraints that we impose to lu in the rest of the world to deal with us and then maximizes our power. That it's a form of commitment. So he changes the narrative quite a bit in a way that I don't think it's in even away from, you know, the confines of the equations. But it's a topic for a broader discussion that I, I'm not sure I want to have today.

- And on the record, I think if you spend any time in the US government, CEA or the treasury or any place like that, you come much closer to that point of view that we can't, we don't always get our way. But basically we're these organizations were set up in part for us to be able to deal with the again, absolutely that comes up in practice. In practice what seems to have happened though they've actually been used for the great benefit of the rest of the world in general. And so the question is empirically with all the reductions in tariffs till recently in yeah, that's

- Why I said for me it's was

- Much more of a national, national treatment. You don't have cross-border firms, but national treatment firms, all these kinds of rules are really giving up a lot for a heman.

- That that's why I, I really think of it as a theoretical statement that changed my perspective rather than quantity statement of what might have happened.

- Original argument for the entire approach of the US after World War II is not, it didn't set up this international order out of the goodness of its heart. I mean, it was clearly in order to support us and and very explicitly by, by Truman and others to support US power and in order to, you know, win this competition with the Soviet Union. I mean there was that, there was, there's no question about that is the strategic goal and that's how it was sold to the US than

- That is

- Absolutely

- By, by no means, you know, my view, it's a view that is very common, for example, in the political science. I don't think it's, I think it's fair to say that in economics that's not the way we approach it. And I think that, that I've totally changed my mind in writing this papers, that that's something that it's, it's a missed opportunity for us to explore. Okay, now in the last part of the talk, I think I have until 1 45, right?

- Yes. - Okay, fantastic. I gave you high level theory. For some of you it's too simple. For some of you is too complicated, but it's somewhere there. I think the main advantage of that theory is to make it formal, to make it precise, to give a definition of power, a definition of the optimal policy. Why is there a role for government intervention and all of this. But I, I wanted to go farther than simply saying, okay, look this formalizes some of these concepts, if it's useful, it has to be useful also in terms of quantifying things. And I would argue that quantifying things in this space is crucial because lots of things you can cut makes no difference. Lots of things that get claimed to be a national interest but is are important when you look carefully, eh, you would have to work hard to make sure that these are big. It's also particularly crucial to have a model because most of the most powerful threats are gonna be of the equilibrium path. You know, if the US tells a European bank do X or we will disconnect you from accessing the US system, that's a death sentence. They will comply on the equilibrium path. All you will ever see is the actions. So models are very good at that counter fashional. And in particular in this one I'm gonna use very heavily advances in trade theory that essentially use simple summary statistics to draw things about the counteral liquidity. Okay? If you're a trade person, this is entirely obvious. If you're not a trade person, there is no chance. You'll understand it as I explain it in the next few minutes, but I will try to make it intuitive. Okay? So think of the big production function that we started with and specialize it in particular, specializing in layers. So at the bottom number four is all of the varieties of a product that you buy all around the world. All of the chips, all of the varieties of oil, okay? Think of a narrow basket, you're getting it from many, many companies. The first thing you do is you take that input and combine it with a domestic variety. So you take all foreign oil and combine it with domestic oil, all foreign chips and domestic chip, okay? And the domestic one is gonna play a large role. And very often we forget it even in our models. Then you're taking these baskets that are potential very different goods and you're starting to put them together so you're mixing oil, steel and something else. So start making goods. And at the end, because we like abstractions, you make some comite good that the consumer consumes, okay? So it's what's called a nest step production function. It's very helpful because it's, it simplifies many, many sectors into something that we can get a handle on in the aggregate. Okay? That's what it does. Okay? So if you're a trade economist, you take that one and you do a lot of algebra and what comes out is a formula that tells you that is very simple to measure. What is the cost of not having access to any of the inputs? And one of you asked me if I a full blo blockade, A full blockade is just financial archy, you don't have access to any of the foreign basket. Now what does that formula look like? It looks like a mess like this, but it's beautiful. Why? Because in red such you have expenditure share, you go into the data and you check how much do we buy of this particular goods from every country around the world? How much do we produce domestically? Those are not in some counter fashional world. Those are in the world we live in. Okay? So you get them from the data in blue, you have nasty parameters. They're elasticities. They tell you how difficult it is to move from one thing to another to produce, not to produce with oil from Russia, but to produce from oil from Venezuela. How substitutable are they? I call them nasty because they're very hard to measure. Those are not in the data. Okay? Now to to, to give you a much closer intuition, suppose that I specialize the economy even far. And I say, look, there's only two things. There is financial services and all other goods and services. Then the formula collapses even more. And we get to this. Now bear with me because this is the, the nasty part. Everything else is a lot more intuitive, okay? And I'm gonna take it in parts. So there is one part that I want to emphasize because it has to do with the nonlinear nature of power. Think about the inner basket, what you buy from the foreigners, how expensive it is to lose access to what the hegemon controls. It's a highly nonlinear function, but for good reasons. The first one tells you that if in equilibrium you're spending 99% of your expenditures on, think that on things that the hegemon controls, it must be really bad for UOL SQL to try to rearrange. 'cause you're trying to move 99% of your expenditures into the 1% alternative. So you're gonna move the prices massively against you effectively. The second thing that happens is the one I mentioned before, if you look at the parameter sigma, when it gets very, very close to one, things become called Douglas, it tells you that if you lose any one variety, all of the rest doesn't help you. So it, that's a case of like more complimentary and called Douglas blows up the index. But it tells you that, for example, if sigma is 10 and omega is 0.2, the losses of not dealing with the US are tiny. Let me repeat it. Like these days we throw around the world shock points a lot. A shock point for an economist is a sector that has two characteristics. An omega very close to one. The avan controls almost a hundred percent of the worldwide expenditures. You put them into a corner and a sigma very close to one. There is very few alternatives and the dose conditions, that formula blows up. The threat becomes very powerful. As soon as you move away from those two, it dissipates very quickly. And to those of you that are trade here, that's one of the reasons why it's so hard to generate large gains from trade is that with, with substitution effects and lots of other things, it's hard to get big numbers. Numbers get very big if you corner me or if it's really essential. Flip aside, it tells you that to believe that something that's really important for national security, you have to pass one of those two tests. If you're not in that space, getting high numbers is gonna be very difficult. Now I, what I did in the, at the bottom is I collected the data I, I went in the omegas are observable. I looked at every bilateral product byproduct. These are millions of products between the US and the rest of the world on China and the rest of the world. And what I'm plotting here is a frequency distribution is a kernel's mood density. So let me give you an idea what the numbers mean. If you look at the China, the fact that China is a large manufacturer and exporter of manufacturing goods tells you that for many countries and many product combinations, 20% of all the foreign purchases fall on China. That's pretty big for a country. The US alone would be a fraction of that. When you look at the western world, that's a US coalition. This was before the carbon speech. I'm not gonna get into that. I shouldn't have said that. It includes Australia, western Europe, lots of other countries. So that gives it a big component. If you actually did the US alone, that will shrink it towards zero a lot faster. But when you look at finance, the US is truly special. If you look at financial services, the fraction that gets spent on the US Western Europe, it's overwhelming. That's really where you see the choke point. Very often, those expenditure shares are in the I eighties, I nineties. Essentially, if you can cut off the basic US financial system, there's very low alternative ants. So if you think about why the US uses finance so much, that's where it comes from. And if you aggregate these numbers, let me show you what they look like. If you go to the, the whole formula and you plug them in, think of the US first. So those numbers are essentially percentage of value added of the target that you can make collapse if you shut them off. So you have a lot of power over Mexico. Well that's pretty obvious. There's a lot of integration with Mexico. So in goods and services, and then you have Singapore, Colombia, you can see the countries. A lot of it comes from finance. This is the US alone, okay. China, almost all of the power comes from manufacturing. One exception Singapore, which is of course the offshore financial center for China. So they have a lot of financial services in China. If you widen the coalition, if you count, not the US cutting you off, but all of Western Europe, Canada, Australia cutting you off together, then the numbers jump up a lot, right? 'cause now you're pushing a lot more into the corner. You're starting to kick in the ality. But you can see that the one that really accelerates the fastest is finance. Because by the time all those country cuts you off, you have very, very little, very low alternatives for China. Of course that doesn't make a huge difference because the countries, I mean putting in the coalitions are relatively small in terms of their footprint worldwide. On, on, on trade and finance. So it, you know, somebody asked me at the beginning, are you talk, did I talk about coalitions? And I said, I wouldn't with this one exception, which is entirely baked in, I didn't provide a theory of why Canada and Europe will get along with the US and impose the sanctions. I said, suppose that they did. Let me quantify it. It's pretty interesting 'cause it reminds you of how important, if you want to sanction a country or you want to threaten something, how important the coalition really is. If you're relying on these big nonlinearities, having Europe not go along in sanctioning Russia, it's a total disaster, right? Both in terms of manufacturing, but especially in terms of finance. Swift is a Belgium cooperative, Euro nuclear, it's in Belgium, London is a huge financial center. So the nonlinearity of power is again, something that is pretty obvious once you see it. But once I realize that, it also changed my mind on lots of things. And I'll tell you too and then I'll, I'll, I'll stop and take both questions.

- This is, this is sort please this kind of quite similar in spirit to Bagwell Staiger.

- Yes. - That looking at multilateral versus unilateral action and coalition formation ly in their case customs unions.

- Absolutely. Yeah. So let me mention a couple of things that that have to do with this. So first one part that this illuminated for me is that the difference between power, relevance and macro relevance. Let me give you the following thought experiment. Okay. Which will be familiar to most economists in the room. China today accounts for 3% of the world financial transactions in their own currency depending on exactly how you count. Okay? Supposedly, I give you a scenario where 10 years from now they're 10% of the transactions. The US is still in the high fifties and sixties. Okay? So as a macro economist, our immediate answer is no difference to the world EQU for pass through of exchange rates. Macro monetary policy, probably not a big deal. Probably, you know, not unlikely that it might happen. We're not talking going neck to neck. Think about the power implication. For us power, that's an absolute disaster because moving away from controlling 95% of the world financial system to 85, it's a lot worse than moving from 85 to 75 and so on. It's so nonlinear that it's really the choke point that matter. If they're 10% a country, the size of Russia, that 10% is an ocean of transaction for them to carry through. So in the terms of these formulas, it tells you that once you're moving away from that being very, very close to one countries can rebalance without much trouble. So the power relevance is very different. And I think instead it leads to the not thinking about this this way and thinking instead about the macro relevance leads to the wrong policy conclusion. 'cause it leaves you with some comfort, which I'm sure you've all read in the newspapers or in the policy awards, that no matter what, we're still gonna be very dominant as the dollar. That might very well be true. But for power exertion, that's a huge difference. Well before it becomes macro relevant or fiscal relevant,

- Really important point. Because most of the discussion has been, if the dollar loses its dominant position, it will occur slowly.

- Yes.

- So people, and you're saying, well, before we start worry, we think we should worry about it. We should for these.

- Yes. So people look, as you're very familiar with it, people phrase that discussion in terms of shares. Yeah. And they're thinking that that would be for neck to neck, like 50 50 shares. But in fact the real problem for power is when you go from 90 10 to, yeah, we're 80 $20,

- 88% on one side or the other of every four. Now

- The good news is that flip the argument, it also tells you that if you're, for example, a middle power, if you're Europe, Canada and you're looking to do economic security, this also tells you that you can get a lot of security with very little fragmentation. You don't, you know, and this goes to Hirshman, so I'm not sure how many of you know, but the HIRSHMAN index, the hirshman infer index that it's used to measure concentration, particularly in IO, was originally written in the book in 45 to measure dependencies on foreign countries. So it was about exactly those. But it in some sense it makes the right point of concentration, but also the wrong point because it gives you the sense that in order for there to be no problem, you want to get to equal shares. That would be a tremendous amount of fragmentation. I mean, if we start requiring the idea that in policy economic security means equal shares, that's a disaster. I mean we would've to rearrange the world economy a lot. Instead this tells you no, no, no. It's really about moving very specific industries that are big, very relevant in the geometric and the satisfied these conditions away from the corner. So you can get a lot of economic security for relatively little fragmentation, at least in terms of the shares in the data. So to me, change, again, the way I look at the data,

- Some of what's going on too is, is security without changing the shares at all. If you build a say payment system that you don't actually use, but you can use in the of sanctions, that's what sort of what China's up to. It's sort of like a stockpile, a stockpile of rare earth. You don't actually use it now, but it gives you the ability to

- Substitution. Yes. In particular, that's the big reason. So you can see, you see that omega IGR, that's a share of domestic expenditure. It's very important because what is the first alternative? The first alternative is you go domestic. I mean if you coming up first, I look for somewhere else, but ultimately domestic you can never cut off. And so it, it's not subject to the coalition, it's not subject to anything. That's my first alternative and I can scale it up. So the only question which goes into the externality is how easy it is to scale up quickly. If it's an industry where it doesn't really matter, I can do it exposed, then the hege has no power over you. And that's what why I said I'm not doing a short term model. I'm gonna allow you to fully scale up the problem. Only of course, if it's an industry where in some sense the productivity, the knowing how to over operate, it really depends on having operated at scale for a while. If it's an industry like that, then we can win. You can see the debates today, like if we think of manufacturing, well we don't have to have a manufacturing sector. If in a potential confrontation we can switch to it as we did in World War ii. If instead you feel that there is a lot of externalities and learning by doing, by opening manufacturing and that we couldn't actually scale it up quickly on a whim, then you have to create that alternative example. That's exactly what that equation says.

- Stockpiling is,

- Yes, stockpiling works like that because it gives you a domestic alternative. You spend time on it. So,

- Oh please. I know you, you mentioned briefly this is in some other work, but could you say something about international ownership of like international financial markets? So if China owns some of Nvidia, it seems like this provides some insurance value, but then on the flip side, maybe the US wants to prevent this and

- Say something like that. I gave you too optimistic of an interpretation. I meant it could be for another paper, not one that we've written, but, but I can tell, I, I can speak to the forces. So first of all, formally the LaGrange multipliers on the participation constraints are essentially your shopping list. They tell you in equilibrium, how valuable would it be to have an extra unit of power on that particular sector. So the one the guys you want to buy are defer of the LaGrange multiplier is very high. Because I think of buying as, I don't have to bully you anymore. I can tell you, I can dictate if I, if the US were to buy A SML, we could simply dictate if it was government ownership that they can't sell to China. You own it, you control it. So I really think of FDI in particular as a controlled stake, as bypassing the participation constraint and just directly dictating the outcomes. It also has a more subtle effect though because it makes you care about the profits of the firm. So in some sense, ownership makes the firm domestic because I ultimately, I'm the residual claimant to its profits. So now I, I care about their profits, not just because they move the inside option and increase my power, but also because they get paid to me for the foreign firm in this model. I care because it's good for you and I can extract some of it, but I don't care because I get the profits. So it really also aligns the incentives if you can do it. Very good. Okay, the last thing I will mention, I'm gonna skip these things. Just as an example of the nonlinearity of power, I wouldn't trust the numbers too much, but it's sort of fantasy. I computed the fraction of foreign financial services that Russia was spending on the Western coalition. The one that eventually sanctioned it back in 2005 in three. That's a green line. So you can see that in unsuspecting times in 2005, or at least for I believe unsuspecting times, 94% of all foreign financial services that Russia buy are coming from the Western coalition. After the early invasion of Crimea, first of all there are some sanctions, but also Russia starts to shift away and starts to build up benefits. So it goes from somewhere around 95 or 94 down to the mid eighties. It, it's not a huge movement of the share. Now the red doted line plotted on the other side tells you the financial power that the Western coalition would've under those conditions on Russia. And you can see a power gets cut by 50% for a 10% movement in the share. And it's precisely because I started the model very, very close to the nonlinearity where the shed is so, so high that now you have a lot of power. Or similarly, if you were to measure power on Iran right now, it would tell you very little. 'cause we've exhausted, there's no relationship. So there's nothing more that I can threaten you with. You're not really that dependent on me. So that that, that's a bit of the logic. Okay, I'm gonna conclude here. So hopefully this was interesting to you. I think this paper has tried to do several things. The main one was to provide an optimal, a theory of optimal economic security policy, but as a byproduct of it, to provide a a clean definition for what do we mean by power? What are the incentives, what should the government get involved? And also thinking about the value of committing not to do too much coercion, the value of rules. And eventually to give you a list, a little bit of a flavor for how do you take it to the data. I want to be clear, those estimates are clearly uncertain. We're not great at measuring elasticities. You ask 10 different economies are gonna have widely different numbers. For me personally, it really pushed me to confront what do I have to believe for these things to be big? And it reminded me in a lot of cases things aren't big because the supply chain is too diversified. There are too many alternatives or because you think the substitutions are very high. I think that that's a useful exercise for to discipline. If you apply to semiconductors, for example, try to do a thought exercise of how many things do you need to believe for export controls of semiconductors to be massively effective and do you believe them after you look at them? Or what are the empirical uncertainties? I think that that's really the value added of them of doing it all together. Lemme stop here and take more questions. If there are any

- In the model, it's all about the, the geoeconomics of power and, and you know, but the other aspect of of power is the, the armed forces and you know, and everybody's rational in this model. But if there were irrational, if it were multi period or something you, there was, you know, possibility that somebody's irrational. You have to build up your armed forces' power as well. And you know, the Trump administration is currently using the, the geo, the the power that has financially and with tariffs and things to extract what they say are resources. You know, that should have been paid by the Europeans and others for, you know, for the, the safety that we were providing. So how would you modify the model to,

- I mean we intentionally in this whole agenda, try to focus on the economics for really two different reasons. One that is more acceptable scientifically, which is I think it, it was data that was missing. There are plenty, you know, there's lots of good work on what I'm gonna define clearly hard power like military interventions. I felt this middle area of we're bullying each other with economics. There was a lot less. The second part which is much less defensible is we just have better ideas, skillset that, you know, I don't have particularly good ideas on the military, but when we put them together we have thought about whether they're compliments and substitutes. And one nice way to think about this is if you think about enforcement, the question is, is the military a substitute or a compliment to ergonomics? And I come to believe that in many circumstances there could be compliments because in some sense, you know, not in this paper, particularly in an earlier paper, a lot of the power comes from bundling together many different threats. You think that these are separate economic activities, but the hegemon is gonna tell you it's all one deal. And so if the military gives you power on one area and geoeconomics give you power on another, the combination is much more likely to generate these big spikes. And so the power is an additive. It's much more than additive. And that's a sense in which I, I lose this being I think cut that as compliments, but there's a lot more worth it on that that we, we only scratched.

- Kind of similar to like your eye-opening realization earlier about how the IMF actually is and their incentives. Does the US have any incentive to convince people, I guess in maybe in like a Barer Gordon sense that they are the type not to try and they're not trying to extract surplus from other countries, but they are trying to Yeah, so nicely set up the world order and would that change anything if they were trying?

- Lemme give you a few answers, some of which I have worked on particularly in the earlier papers. So if you think of like a repeat, so one of our earlier papers on the offense was repeated games. And one of the things that becomes interesting in that world is moving the inside option versus moving the outside option. So here I only made you a negative bully, you are, you're making my outside option worse by threatening me. I suppose that I could do things like moving the inside option. For example, suppose that you have enforcement issues in your company, it'd be great to build infrastructure, but surely will get ated so it doesn't get built. And I say, well I'm a, I'm a country with a large military, a big operation there is, if you don't go along with me, we will punish you so severely that you can actually sustain the project. Now that moves the inside option. Now in this paper they look relatively similar, but in a repeated game they're not, which gets to what you have in mind. Why? Because if you repeat the game, the inside option also increases the continuation value of the game because I think that you're gonna keep doing it with me and that's a very powerful incentive for me to be aligned with you and take your offer. If instead you try to lower the outside option on the one period, it was roughly just the same. But the problem is that that lowers the values tomorrow because I also know that you're gonna lower my values tomorrow that fits back into a low continuation value dealing with you and pushes me not to accept your contract. So in general, positive inducements are better and negative inducements in this context. The problem is the cost of the positive inducements. I gave you one that was sort of cheap because it was about enforcement. The typical positive inducement, inducement is a Bri or you know, a loan or things like that. But those are fiscal costs, so you have to be careful with those. And in that, in those models, we didn't do reputation models, we did simple sps like sub sub game, perfect equia, but it's all about like, if I deviate one, you're never gonna have to ask me again. So it has an element of what you have in mind, although it gets complicated pretty quickly.