Understanding Geoeconomics: How Global Economics Shape International Relations

Introduction to Geoeconomics

• Geoeconomics is defined as the practice where hegemonic nations, such as the United States and China, utilize existing trade and financial relationships to exert power abroad for either geopolitical or economic objectives. 25s
• The field of geoeconomics has experienced a significant increase in research attention over the last five years, particularly within the last two years. 8s
• The exertion of power through geoeconomic means can involve pressuring or "bullying" either foreign governments or private firms. 35s
• Key research questions in this field include identifying the sources of this power, determining how it is wielded, and establishing what constitutes optimal policy for nations like the U.S. or China. 42s

Economic Analysis of Geoeconomic Power

• While colloquial discussions of geoeconomics often rely on a zero-sum, mercantilist logic, economic analysis suggests that the global economy is not fixed, meaning the total economic "pie" can expand or shrink, potentially leaving all parties better or worse off. 1m0s
• A central area of study involves defining and measuring national security externalities and evaluating the potential trade-off between the gains from trade—such as specialization and economies of scale—and the risks of economic dependency. 1m25s
• Quantitative research in this field aims to perform policy counterfactuals, such as estimating the impact on the U.S. economy and specific sectors if the country were to be cut off from rare earth minerals. 1m45s
• The study of geoeconomics incorporates perspectives from political science to provide a broader conceptual framework beyond traditional economic models. 2m6s
• Empirical approaches to geoeconomics include the "sufficient statistics" method, which combines economic models with hard data, and the analysis of unstructured text to identify instances of one entity exerting pressure on another. 2m25s
• Current economic definitions of power, which often reduce the concept to market power or the ability to sell goods at a markup, are considered insufficient to capture the broader influence exerted by powerful countries or large corporations like Google and JP Morgan. 2m50s

Political Science Perspectives on Power

• Robert Dahl defined power in 1957 as the ability of actor A to influence actor B to perform an action that B would otherwise not choose to do 0s.
• Relational power is a concept in political science that focuses on the interaction between two actors while treating the broader environment as a fixed constant 15s.
• Susan Strange introduced the concept of structural power, which refers to an actor's capacity to shape the entire environment, such as by establishing rules or norms 26s.
• Political scientists have noted that the study of power within their field has historically been informal, prompting a need for more precise definitions that allow for theoretical modeling and empirical data analysis 48s.

Hirschman and the History of Economic Influence

• Albert Hirschman authored a notable book regarding Nazi Germany’s use of economic influence over Europe in the period leading up to World War II 1m15s.
• Hirschman’s early work challenged both mercantilism—the belief that nations only bargain over fixed resources—and the English free trade perspective, which assumed that free trade would naturally prevent imbalances 1m55s.
• Hirschman argued that trade inherently leads to imbalances, which subsequently result in power imbalances between nations 2m12s.
• The Herfindahl index, commonly used in modern economics to measure firm concentration and market power, was originally developed by Hirschman to measure the concentration of trading partners and identify power imbalances between sovereign states 2m25s.
• While the Herfindahl index was once significant in economics, its original application regarding trade power has largely been lost, and the study of these issues has shifted toward the field of international political economy 2m45s.

Theoretical Modeling of Geoeconomic Power

• A comprehensive reading list is provided, encompassing both economics and political science, to encourage deeper exploration of geoeconomics beyond standard economic training 0s.
• The research approach utilizes basic economic theory to construct models and frameworks for understanding geoeconomic questions 35s.
• Power is defined as the gap between the inside and outside options within the participation constraints of targeted entities 55s.
• The exertion of power is modeled as a costly action that forces agents to deviate from their privately optimal choices, effectively introducing wedges into their first-order conditions 1m15s.
• Structural power is conceptualized as the ability to manipulate an entire equilibrium in one's favor by influencing externalities within a general equilibrium framework 1m45s.
• A theoretical distinction is established between "micro" relational power and "macro" structural power based on how agents interact with equilibrium objects 2m5s.

Case Study: ASML and US Export Controls

• The Dutch lithography firm ASML serves as a recurring case study to illustrate these concepts, specifically regarding its role in producing advanced semiconductor machines for Chinese customers while utilizing US suppliers 3m5s.
• The US government's attempt to restrict ASML from selling machines to China is used as an example of an actor attempting to influence a firm's optimal input and output decisions 3m35s.
• The United States seeks to influence the actions of ASML, a Dutch firm, by imposing a wedge on its operations, such as increasing the perceived cost of selling machines to China or implementing quantity restrictions 0s.
• Because ASML is not a domestic firm, the United States cannot directly regulate, tax, or dictate export controls to the company, necessitating the use of inducements to influence its behavior 25s.
• The United States utilized negative inducements to gain compliance, threatening to invoke the foreign direct product rule against ASML 55s.
• Under this threat, ASML would be added to the entity list, which would prohibit any US person from doing business with the company, effectively acting as a "death sentence" for the firm 1m12s.
• Formally, this threat functions by significantly lowering the firm's value, forcing ASML to evaluate its participation constraint by comparing the value of its outside option—losing access to the US market—against the inside option of complying with US demands 1m35s.

Mechanisms of Coercion and Strategic Sectors

• Micro power is defined as the maximum cost of private actions a firm is willing to accept before it becomes more beneficial to choose the outside option, while taking equilibrium aggregates as given 2m25s.
• The effectiveness of these threats depends on factors such as the ease of finding substitutes or the ability to develop alternative technologies, which determine how much the inside and outside options can be shifted 2m45s.
• A strategic sector is characterized by its ability to facilitate highly valuable threats 3m5s.
• Oil is not considered strategic if a specific variety can be easily replaced by substitutes, but it becomes strategic if a single entity, such as OPEC, controls the entire supply, forcing the use of inferior alternatives 3m15s.
• A hegemon can exploit a gap in power by bargaining with individual entities that track only their own private costs, while the hegemon benefits from the aggregate impact of changing the environment. 0s
• Strategic complementarity in social media allows a hegemon to influence users by encouraging them to close competing accounts, which increases the attractiveness of the hegemon's platform and diminishes the value of outside options. 15s
• Once users are consolidated on a single platform, the hegemon gains significant power because the outside options become unattractive due to a lack of other users, effectively creating a coordination addiction similar to a drug dealer model. 45s
• Strategic sectors are defined as those that enable a hegemon to manipulate the economy to its own advantage. 1m5s
• The justification for government intervention in these models differs from financial regulation; while financial regulation addresses coordination problems like fire sales, strategic sector intervention addresses situations where a large player is actively twisting the equilibrium in their favor. 1m15s

National Security and Infrastructure Risks

• Individual private agents may not account for national security implications when accepting bids from foreign entities, such as China's involvement in ports along the western seaboard of the U.S. and Latin America. 1m45s
• Because private agents fail to internalize the aggregate security risks of foreign control over strategic infrastructure, there is a high probability that government screening of these sectors is necessary. 2m5s
• The formal model for analyzing these dynamics involves $n$ countries and a set of productive sectors, where a sector is defined as an activity occurring in a specific location. 2m25s
• Production functions in this model incorporate a set of aggregates, a technique derived from Greenwald-Stiglitz, to capture externalities that affect the production function of the consumer. 2m50s

Economic Frameworks for Trade and Policy

• External economies of scale occur when the aggregate output of a sector increases the productivity of individual firms within that sector 0s.
• Strategic complementarity arises when the usage of an input across multiple sectors increases the productivity of that input, such as the network effects observed in shared payment systems 10s.
• The representative consumer model focuses on utility derived from consumption, though it can be adjusted to include geopolitical motives, such as preferences regarding the size of a foreign military sector or the relative output between countries 32s.
• While utility functions can incorporate arbitrary factors like pollution in climate change models, the primary focus remains on consumption-based objective functions to analyze policy games 1m5s.
• The budget constraint is determined by domestic factor income and firm profits, with the assumption that consumers own domestic firms 1m25s.
• Although cross-border ownership is possible, the model currently restricts ownership to domestic firms, excluding endogenous foreign direct investment (FDI) for the time being 1m45s.
• Market clearing conditions require that all produced goods are either consumed or used as intermediate inputs, and all factors of production must be fully utilized 2m15s.
• The model utilizes a global input-output matrix to analyze trade, with the central focus being the policy game played between nations 2m35s.

Hegemonic Bargaining and Coercion

• A hegemon is defined by its ability to threaten foreign entities with the loss of access to specific inputs, such as semiconductors, rare earths, or oil, if those entities do not comply with the hegemon's demands 2m55s.
• The hegemon may demand that other countries implement specific wedges in their first-order conditions or provide financial transfers 3m25s.
• Wedges imposed by the hegemon are designed to be revenue-neutral, meaning they are rebated to the agents, but because agents do not perceive these rebates as fixed, the wedges function effectively as quantity restrictions 3m35s.
• The analytical framework utilizes a primal approach where a hegemon dictates quantities to other nations through the implementation of wedges, leaving the rest of the world to either accept or reject these terms 0s.
• The hegemon possesses all bargaining power, allowing it to issue take-it-or-leave-it offers to maximize its own power 15s.
• Countries are assumed to have a full set of wedges available to apply to their domestic firms, with the assumption that domestic governments act dictatorially and face no participation constraints 32s.
• Domestic policy tools, such as industrial policy, export controls, and tariffs, are used by countries to manage their economies, which may serve purposes like providing public goods or maximizing positive technological externalities 55s.
• Countries anticipate future coercion from the hegemon and proactively shape their domestic economies to withstand potential threats, a practice currently referred to as anti-coercion or economic security policy 1m18s.
• Economic security policy involves strategic decisions, such as limiting dependency on a single foreign source for materials, to mitigate the impact of future hegemon intervention 1m35s.
• Foreign entities evaluate an inside option, which involves producing with a full set of inputs while accepting costly actions and transfers imposed by the hegemon, against an outside option, which involves producing with a restricted subset of goods without transfers or costly actions 1m55s.
• Firms maximize profits while perceiving wedges imposed by both the hegemon and their own domestic governments, treating these wedges as costs even though they are rebated in equilibrium 2m15s.
• The welfare functions for both the hegemon and other countries are based on maximizing the indirect utility of the consumer 2m45s.
• The hegemon’s wealth includes domestic firm profits, domestic factor payments, and transfers received from the rest of the world 2m55s.
• For foreign firms that accept the hegemon's contract, their welfare is represented by the value of the outside option plus domestic factor income, as the participation constraints are assumed to bind 3m5s.

Simplified Models and Equilibrium Analysis

• The model simplifies the analysis by setting the variables U and Node Z to zero and assuming linear consumer utility, which results in constant equilibrium prices. 0s
• By assuming constant prices, the model focuses exclusively on the variation of technology, represented by the variable zeds, and eliminates terms of trade manipulation. 25s
• The economy is modeled with the United States acting as a hegemon with a single sector that produces financial services linearly using local labor. 52s
• The rest of the world consists of identical small open economies that possess two sectors: a domestic sector producing local financial services and an intermediary that combines local and global varieties into a composite good. 1m5s
• The model incorporates external economies of scale in the domestic sector, where the productivity of the domestic alternative increases as the sector scales up. 1m35s
• Strategic complementarity is applied to the global sector, where it becomes more productive for an intermediary to use the global technology if other intermediaries in the same and other countries are also using it. 1m48s
• The messaging or payment system serves as a practical example of this framework, where the utility of a system increases as more countries adopt it, while domestic alternatives remain inefficient until they reach a sufficient scale. 2m6s
• A Constant Elasticity of Substitution (CES) aggregator is used to combine the domestic and global alternatives, with externalities embedded within the productivity functions of these two options. 2m35s
• The productivity of the global technology for intermediaries in country J is determined by the average usage of that technology across all other countries, with the strength of this externality governed by the parameter Kai. 2m55s
• The domestic economy of scale follows a similar functional form but depends solely on the usage of the technology by other intermediaries within the same country. 3m15s

Global Planning and Decentralized Equilibrium

• To maintain a convex problem, specific assumptions are required regarding the functional forms of the model. 3m25s
• A global planner benchmark is established to maximize the sum of welfare across all countries, utilizing a primal approach to solve the problem. 3m45s
• In a symmetric country model, a global planner maximizes profits by calculating intermediary output and revenue minus costs, resulting in standard first-order conditions for the use of two instruments 0s.
• The global planner internalizes production externalities by taking derivatives over the technology levels, accounting for the fact that increasing the usage of a hegemon technology in one country enhances productivity for all others 35s.
• In a competitive environment with small countries, individual entities would treat technology as given and assume their actions do not affect the broader technology, effectively setting the derivative of the externality to zero 1m5s.
• When intermediaries in a decentralized economy face wedges on the usage of foreign inputs and domestic alternatives, their first-order conditions differ from the planner's because they do not account for the impact of their actions on the overall level of technology 1m20s.
• Intermediaries in a decentralized economy perceive the price of technology as the actual price plus a wedge 1m45s.
• The primal approach involves identifying a set of wedges that align the first-order conditions of private entities with the solution determined by the global planner 2m0s.
• By comparing the planner's first-order condition with the private intermediary's first-order condition, one can solve for the specific wedges required to implement the planner's optimal allocation 2m15s.
• The primal approach serves as a general tool for determining optimal policy by starting with the planner's dictated quantities and finding the wedges that induce those same outcomes in a decentralized equilibrium 2m55s.

Externalities and Nash Game Dynamics

• The final outcome of this analysis reveals that the planner recognizes the existence of two positive externalities 3m15s.
• Productivity for both domestic and global technologies increases with usage, justifying a subsidy strategy proportional to the strength of the externalities associated with each 0s.
• The application of this planning solution requires convex optimization to ensure unique solutions for intermediaries and the availability of complete instruments that allow a planner to dictate quantities in the allocation space 15s.
• In a benchmark scenario featuring infinitely many small open economies acting in a Nash game, each government fully subsidizes domestic technology because the associated externality is contained entirely within its borders 1m5s.
• Under the same Nash game conditions, governments neither tax nor subsidize global technology because they perceive themselves as too small to influence it, treating the global technology as a given 1m35s.
• A generic result in these planning problems is that externalities occurring across borders are typically undertaxed or undersubsidized in a Nash game, as individual actors fail to perceive the full scope of the externality 2m0s.

Hegemon Coercion and Participation Constraints

• The analysis of hegemon coercion involves a two-stage process: first, determining the optimal taxes imposed by the hegemon on foreign entities and the optimal transfers; second, determining the optimal domestic policies pursued by each country 2m35s.
• In the visual model of an intermediary within a country, constant prices result in a flat marginal revenue curve, while a production function that decreases with scale results in an increasing marginal cost curve 3m5s.
• When restrictions are placed on an intermediary preventing the use of foreign inputs, the marginal cost curve shifts to the left. 0s
• The extent of this shift depends on the Constant Elasticity of Substitution (CES) aggregator; if inputs are perfect substitutes, the curve does not shift, but if inputs are Cobb-Douglas, the inability to access foreign inputs prevents production entirely. 7s
• It is assumed that the hegemon designs the system so that the target country's participation constraint binds, meaning the hegemon extracts all possible surplus through transfers, leaving the target country only with the value of its outside option. 23s
• A fundamental conflict exists between the hegemon and target countries regarding the difference between the "inside option" (the value within the hegemon's system) and the "outside option" (the value available without the hegemon). 45s
• The hegemon prioritizes maximizing the gap between the inside and outside options, whereas the target country is concerned with the absolute level of its welfare. 53s
• While a hegemon may act positively by providing public goods or improving technology to raise the inside option, it may also choose to destroy value on the equilibrium path if doing so lowers the target's outside option even further. 1m15s
• This strategy, referred to as the "drug dealer model," involves making target countries dependent on the hegemon's system so that their outside options become significantly worse, thereby increasing the hegemon's power. 1m35s
• In formal modeling, the hegemon can use a primal approach to dictate the actions of the intermediary and solve for the necessary wedges to influence the inside option. 2m6s
• The hegemon's ability to dictate terms via wedges applies only to the inside option, as the target country retains autonomy over its outside option by refusing to engage with the hegemon. 2m18s

Hegemonic Strategy and Global Efficiency

• The primary objective of the hegemon is to maximize financial transfers from other countries, operating under the constraint of intermediary participation 0s.
• The model distinguishes between an "inside option," where the hegemon directly chooses the technology usage, and an "outside option," where countries independently choose their own economic path 15s.
• Transfers are determined by ensuring that participation constraints are binding, allowing these transfers to be substituted out of the mathematical problem 42s.
• When utilizing the inside option, the hegemon internalizes the fact that increased adoption of its technology makes that technology more attractive to all other participants 1m5s.
• The hegemon’s strategy regarding the domestic alternative involves a trade-off: scaling down the usage of domestic alternatives reduces the productivity of the inside option but simultaneously degrades the outside option for other countries 1m35s.
• The hegemon acts benevolently by encouraging the use of its global technology up to the same level a central planner would, aiming for global efficiency in technology adoption 2m15s.
• A divergence from the central planner occurs because the hegemon imposes taxes on domestic alternatives to eliminate the outside option, which is optimal for the hegemon but inefficient from a broader planning perspective 2m35s.
• In response to this dynamic, individual countries may implement domestic wedges to shape their economies and improve their outside options, thereby reducing their dependency on the hegemon and mitigating the impact of potential coercion 3m5s.

Fragmentation and Economic Security

• Nations are currently seeking to reduce their economic dependencies on major powers like the United States and China due to concerns regarding potential coercion 0s.
• Countries are actively attempting to build alternative systems, such as independent payment networks, to mitigate the risk of foreign powers leveraging existing dependencies against them 15s.
• A theoretical model demonstrates that banning dealings with a hegemon—by imposing infinite tariffs on global technology and subsidizing domestic alternatives—results in full economic fragmentation 25s.
• The effectiveness of an "outside option" for a country depends directly on the productivity of its domestic alternatives; if the outside option is entirely independent of the inside option, it creates the conditions for a complete ban on the hegemon's technology 55s.
• A hegemon typically prefers a higher level of globalization than a central planner would, as this centralization increases the hegemon's influence 1m35s.
• When a hegemon uses its bargaining power to make coercive "take it or leave it" offers that provide insufficient value to other nations, those nations may choose to exit the relationship entirely 1m50s.
• In a scenario of full fragmentation, the hegemon loses its power in equilibrium because it no longer maintains relationships with other countries 2m5s.
• The model suggests that a hegemon might be better off if it could commit to limited coercion or ensure that other nations retain some surplus, rather than engaging in constant bullying that leads to the loss of its own power 2m15s.

Multilateral Organizations and Commitment

• Perspectives on international organizations have shifted, moving away from the view that they exist solely to implement a planner's solution toward understanding them through the lens of these complex economic dynamics 2m35s.
• International relations can be modeled as a hegemon designing multilateral agreements, such as those involving the IMF or WTO, to constrain its own actions while serving its own interests 0s.
• These multilateral deals function as an offer where participants remain within a sphere of influence, accepting some level of bullying in exchange for retaining value, with the hegemon limiting its own power to incentivize participation 22s.
• Recent global trends represent a significant rethink of this traditional offer, moving away from previous models of engagement 42s.
• A theoretical framework for these relationships can be viewed as an equity contract where the hegemon commits to extracting only a fraction of the "inside option" value, leaving some benefit for the participant 55s.
• If the hegemon extracts nothing, it lacks incentive to participate; if it extracts everything, the participant loses trust; therefore, an equilibrium exists where the participant retains some power and extracts positive rents, preventing full fragmentation 1m15s.
• Multilateral organizations serve as mechanisms for limiting coercion, a field of study that remains under-researched, particularly regarding how commitment functions when two hegemons, such as the U.S. and China, compete 1m35s.

The Trade-off Between Efficiency and Security

• There is a fundamental, deep-seated trade-off between the gains from trade—driven by specialization and economies of scale—and economic security 2m15s.
• The standard argument for globalization assumes that markets are contestable and that there is no "hold-up" risk, meaning goods are guaranteed upon payment 2m35s.
• Specialization creates a vulnerability where a partner can withhold goods, leaving the other party with a poor "outside option" because they have lost the ability to produce the goods they previously outsourced 2m55s.
• The same forces that generate gains from trade simultaneously create the hold-up problem, making the trade-off between efficiency and security inherent rather than accidental 3m15s.
• Global economic shifts currently reflect a reassessment of this trade-off, with countries increasingly willing to sacrifice some gains from trade to achieve greater economic security 3m35s.
• Current anti-coercion policies are determined by countries acting in a Nash equilibrium, where each nation sets its policy while taking the policies of all other nations as given 4m5s.
• A "fragmentation doom loop" occurs when countries prioritize security by withdrawing from global technology markets, which subsequently reduces the attractiveness of that technology for other nations, leading to further collective withdrawal and amplification of the trend 0s.
• This cycle can result in an equilibrium where countries achieve excessive security at the cost of significant losses from trade, suggesting that coordinated policy would be more beneficial 0s.

Standard Economic Forces in Geoeconomics

• When removing simplifying assumptions—such as constant prices or specific utility functions—the resulting theoretical model becomes more complex and lacks closed-form solutions, but it remains useful for identifying standard economic forces versus new ones 25s.
• Standard economic forces include the propagation of subsidies or taxes through the input-output matrix, which affects aggregate technology levels and ultimately impacts firm profits or specific utility functions 55s.
• A government may impose export restrictions, such as taxing semiconductor exports to China, to influence downstream industries or shrink specific sectors, such as military applications, by leveraging the input-output matrix 1m25s.
• Terms of trade manipulation remains a standard economic force, where a country aims to increase the prices of its exports and decrease the prices of its imports, similar to the classic optimal tariff formula 2m5s.
• Pre-existing private distortions caused by existing wedges are considered a standard factor that can be ignored in this specific analysis 2m35s.
• The primary "new" element in this analysis is the concept of building power, which was previously excluded by assumptions such as linear utility, the absence of profit-affecting mechanisms for the hegemon, and the lack of specific utility terms 2m45s.

Building and Manipulating Power

• Power is defined by the gap between the value of an inside option and an outside option, which is manipulated by impacting specific activities that propagate through an equilibrium 0s.
• Actors manipulate activities by considering how their private costs spill over into the equilibrium, thereby affecting the inside and outside options of all other participants 25s.
• The same mechanism applies to price manipulation, where impacting an activity influences the vector of general equilibrium prices, which in turn alters the inside and outside options of others through the participation constraint 55s.
• The core components of building power include the participation constraint, wedges as costly actions, and externalities resulting from equilibrium amplification 1m20s.
• Optimal tariff strategies may conflict with power-building strategies, such as when an actor chooses to sell goods cheaply to prevent others from building alternatives, only to later extract power by threatening to exclude them 1m50s.
• Anti-coercion policy involves intervening in domestic activities not just for standard reasons like addressing externalities or price manipulation, but specifically to influence the future requests of a hegemon 2m35s.
• These future requests from a hegemon impact the profits of domestic firms, particularly regarding their outside options 2m55s.
• A sufficient statistics approach is used to apply the theoretical model to data, allowing for the calculation of counterfactuals 3m15s.
• Because the outside option is never realized in the theory, researchers use observable data from the inside option to infer what would occur if access to specific inputs, such as Chinese rare earths, were lost 3m35s.
• This methodology mirrors trade economics research regarding the cost of autarky, though it focuses on the loss of a specific collection of inputs rather than a total shutdown of trade 4m5s.

Production Functions and Input Substitutability

• The analysis of geoeconomics involves utilizing non-standard data sources to identify threats and interactions between actors, particularly in the context of artificial intelligence 0s.
• Production functions with decreasing returns to scale can be modeled using a Constant Elasticity of Substitution (CES) basket of intermediate inputs 35s.
• In a Cobb-Douglas production setting, where the elasticity of substitution (sigma) is equal to one, the loss of any single input variety results in a total loss of production 45s.
• Any production model more complementary than Cobb-Douglas will also result in a full loss of production if any individual input is lost 55s.
• The infinite cost of autarky in trade models is partially explained by the reliance on foreign inputs within a Cobb-Douglas framework 1m2s.
• Log profits for a producer are defined as a constant plus the log of the price index, meaning that changes in production conditions can be tracked through variations in this price index 1m20s.
• The impact of losing specific inputs is measured by the log ratio of the price index for the full set of inputs versus the price index for a restricted subset of inputs 1m45s.
• The severity of losing inputs is proportional to the expenditure share those inputs represented; losing a larger fraction of expenditure results in a greater negative impact on the producer 2m6s.
• An "hegemon" is defined as an actor capable of removing a specific fraction of inputs from a producer, with the resulting loss determined by the equilibrium expenditure share of those inputs 2m25s.
• Complex production models can be constructed by nesting CES functions to analyze how different layers of inputs affect final outcomes 2m45s.
• At the bottom level of a nested model, elasticities of substitution determine how easily a producer can switch between different foreign sources of a specific input 2m55s.
• In the case of oil, a high elasticity of substitution suggests that oil from different countries, such as Russia, Venezuela, and Canada, can serve as effective substitutes for one another 3m5s.
• If a producer is cut off from one variety of a highly substitutable input, the price index will increase only slightly as the producer reallocates to alternative sources 3m15s.

Measuring Domestic Share and Economic Dependency

• Economic modeling involves creating a layer for each sector that combines a basket of foreign varieties with domestic varieties to account for differences between them. 0s
• One reason for mixing domestic and foreign varieties is that they may not be perfectly substitutable, such as in consulting services where international projects may involve higher value-added services than domestic ones. 12s
• A second reason for distinguishing between domestic and foreign varieties is the security provided by the ability to produce goods domestically if a country is cut off from international trade. 35s
• The size of a domestic economy is a critical factor in security, as continental-sized economies like the United States, China, and Western Europe rely heavily on their domestic share. 42s
• The cost of autarky is generally lower for countries with large domestic economies that import very little, whereas it is significantly higher for countries like Singapore that rely heavily on imports and exports. 55s
• Measuring the domestic share of an economy is difficult, which can lead to policy mistakes because foreign trade data is often more readily available than domestic production data. 1m12s
• Because countries have historically maintained customs offices to track tariffs, data on manufacturing goods crossing borders is generally of high quality. 1m22s
• A potential policy oversight occurs when analysts focus heavily on import percentages from specific countries without accounting for the total domestic share of usage in that category, which may mitigate the actual dependency. 1m40s
• After creating a composite basket of domestic and foreign production, the model aggregates across industries, such as electronic equipment and cars. 2m5s
• Financial services are kept separate in the model because they are considered basic services necessary for making payments, with the assumption that little economic activity can occur without them. 2m12s
• The model utilizes nested Constant Elasticity of Substitution (CES) functions, which allow for the derivation of price indexes by starting from the outermost layer and nesting inward. 2m35s
• Economic power is defined as the percentage loss of value-added that can be inflicted on a target by restricting inputs. 2m48s
• Expenditure shares, represented by omegas, indicate the fraction of a target's expenditure allocated to specific baskets, such as the split between general goods and finance, or the breakdown of manufacturing into equipment and cars. 2m58s

Sufficient Statistics and Nonlinearity of Power

• The variable "big omega" represents the fraction of a basket of goods sourced from domestic production, where a value near zero indicates that almost all production is domestic 0s.
• The variable "little omega" measures the fraction of foreign varieties within a basket that is controlled by a hegemon, such as the United States or China 12s.
• While "little omega" receives significant attention in policy circles and the press—such as claims regarding the percentage of products sourced from China—domestic production shares are also critical to understanding economic aggregates 25s.
• Financial sectors are analyzed separately from other sectors because they are expected to significantly shift economic aggregates 42s.
• The framework utilizes "sufficient statistics" developed by Arkolakis, Costinot, and Rodríguez-Clare, which rely on measurable equilibrium expenditure shares 52s.
• Elasticities represent a challenging component of these models because they are difficult to estimate, particularly as data becomes more disaggregated, such as determining the elasticity of substitution between Nvidia and Huawei semiconductors 1m5s.
• Despite uncertainty regarding elasticity values, the framework helps constrain debates by clarifying what assumptions are necessary to produce specific numerical outcomes 1m25s.
• To simplify the model, the outer nest is assumed to be Cobb-Douglas, effectively setting the elasticity to one, based on the "idiot law of elasticities" 1m38s.
• The financial sector is treated as special, with a low elasticity of substitution in the 1.7 range, reflecting the difficulty of operating without the U.S. financial system 2m6s.
• The model incorporates an assumption from Feenstra and Weinstein that domestic varieties are less substitutable with the foreign basket than foreign varieties are with each other 2m22s.
• Analysis of the price index regarding the potential cutoff from a hegemon reveals two primary nonlinearities 2m45s.
• Losses in the model become infinite as "little omega" approaches one, representing a scenario where an entire basket is cut off from a hegemon, effectively making the cost of the good infinite 3m0s.
• The Constant Elasticity of Substitution (CES) framework provides significant insights into the nonlinearity of power within geoeconomics. 0s

Choke Points and Concentration of Trade

• Financial services exhibit high concentration, with 80% to 90% of imports in many bilateral relationships originating from the United States and its allies, such as Belgium, Luxembourg, Germany, and London. 12s
• China serves as a major manufacturing powerhouse, with many countries importing 20% to 30% of their goods from this single destination. 26s
• The functional form of these dependencies becomes particularly impactful when a single source accounts for a very high share of imports, such as 95%. 35s
• A "choke point" is defined by dependencies where there is a dominant player with almost no alternative, or where the elasticity of substitution is extremely low, making it technologically difficult to function without the imported good. 48s
• Sectors such as radar, advanced semiconductors, and finance exhibit characteristics that align with these definitions of dependency. 1m5s
• The relationship between dependency and power is highly nonlinear; power dissipates rapidly as soon as a country moves away from a state of extreme concentration. 1m23s
• Modern trade and macro theory provide a more sophisticated version of Albert Hirschman’s original concept of dependency, which focused on the concentration of trade among few partners. 1m33s
• While Hirschman’s approach involved summing the squares of trade shares, the current theoretical framework incorporates the elasticity of substitution to account for the technological difficulty of rearranging supply chains if a source is cut off. 1m45s
• Achieving safety does not require equal trade shares; it is sufficient to move away from the corners of extreme dependency. 2m6s
• Because power is nonlinear, it is not necessarily additive, meaning that if one entity loses power, it does not automatically accrue to another. 2m15s
• The value of controlling a specific node, such as Singapore in the context of financial services, is offensive; it allows a dominant player to push import shares closer to 100% when they already control the majority of the sector. 2m32s
• Controlling a small financial center, such as Singapore, can provide significant leverage for a nation like China, even if it does not immediately transform the country into a global financial superpower. 0s

Geoeconomic Power Dynamics and Data Limitations

• Geoeconomic power is not a zero-sum game; the loss of influence by the United States in a specific region does not necessarily result in an equivalent gain in power for China. 15s
• Small, open economies that rely heavily on trade with major powers are projected to experience the largest economic losses in the event of a geopolitical split. 35s
• The United States derives significant geoeconomic power from the financial sector, which, despite representing a small share of total gross output, is critical to the functioning of all other economic sectors. 53s
• China’s geoeconomic power is primarily rooted in manufacturing, which generally involves goods with higher elasticities of substitution, such as basic durable goods, making them easier to replace than financial services. 1m5s
• Rare earths serve as a notable exception to the general substitutability of Chinese manufacturing exports. 1m18s
• The presented economic calculations represent a medium-run perspective, as they allow for the re-optimization of sourcing and production inputs following a supply disruption. 1m35s
• Short-run models, which would assume an inability to adjust orders after a sudden cutoff, would result in significantly higher projected economic losses compared to the medium-run approach. 1m50s
• The medium-run calculations do not account for long-term shifts in equilibrium aggregates or fundamental changes to production functions. 2m6s
• The level of data aggregation significantly impacts the accuracy of geoeconomic modeling; using highly aggregate data can smooth over critical choke points, such as specific semiconductors or specialized processes. 2m25s
• Increasing the level of disaggregation to capture specific choke points often leads to a rapid decline in data quality and an increase in statistical noise. 2m45s

Challenges in Empirical Geoeconomic Research

• Most countries lack comprehensive micro-data for domestic production, making the domestic share of economic data difficult to track accurately 0s.
• Cross-border data for services is notoriously difficult to estimate because services are not typically subject to tariffs and are not tracked effectively 12s.
• Competition between the United States and China in sectors like information technology and artificial intelligence is difficult to measure because these areas are categorized as services 28s.
• As economic data is disaggregated, measuring elasticities becomes increasingly complex 45s.
• While it is common to assume that elasticities increase with disaggregation, this does not necessarily imply infinite losses, as products from different countries can serve as substitutes 55s.
• Input-output tables are necessary to understand indirect trade, such as Chinese content embedded in exports from Mexico or Vietnam to the United States 1m25s.
• Quantitative empirical research in geoeconomics is expected to grow as data quality improves 1m45s.
• Current models often fail to account for systemic adjustments, such as changes in exchange rates, wages, and interest rates, that would occur if a major economy like China were cut off 2m5s.
• While simplified formulas can provide a first-pass analysis, more complex computational models that calculate different equilibria are often required for accuracy 2m25s.
• Many powerful geoeconomic threats are never observed in data because the target entity complies with the demand to avoid the threatened consequence, such as the loss of access to U.S. financial markets 2m55s.
• The unpredictability of geoeconomic demands makes it difficult to prespecify or capture these pressures using hard data 3m30s.

Textual Analysis and Large Language Models

• Textual data is a valuable resource for research because its narrative and unstructured nature can capture complex information that is not easily categorized, such as the pressures faced by CEOs regarding international trade restrictions 0s.
• Analyzing text allows researchers to identify specific events or pressures—such as a company stopping trade financing for Iran or restricting sales to China—without needing to pre-specify categories in advance 0s.
• Recent advancements in artificial intelligence have significantly improved the ability to extract dense, complex information from large volumes of text 42s.
• Early methods of text analysis in economics included simple keyword searches, such as looking for the word "tariff," or more advanced bigram approaches that compare word combinations against baseline frequencies 42s.
• While human-led manual coding of documents allows for complex instruction-based extraction, it is impractical for large datasets, such as a collection of 1.2 million documents, due to the significant time investment required per document 1m35s.
• Large language models shift the research frontier by enabling the extraction of complex information with the efficiency of automated algorithms, effectively mimicking the capabilities of human coders 1m35s.
• The effectiveness of large language models is attributed to the cross-attention mechanism, a non-linear network that assigns weights to tokens even when they are far apart in a text, allowing the model to recognize relationships between distant words 2m6s.
• The research corpus consists of communications from CEOs and CFOs of public firms worldwide to their investors, a dataset frequently used in economic research and readily available through data providers 2m6s.
• Analyst reports are included in the research to supplement CEO communications, as CEOs may be cautious or guarded when discussing politically sensitive topics or government administrations 2m6s.
• Chinese CEOs may face different incentives than third-party analysts when reporting on company activities, as analysts may be more willing to disclose external pressures from foreign offices 0s.

Methodology for LLM-Based Data Extraction

• Research conducted with Antonio Copa utilizes Large Language Models (LLMs) as classifiers by providing them with detailed, multi-page prompts that function as specific instructions 15s.
• The process of using LLMs for data analysis involves transforming text into structured datasets, often consisting of binary variables such as whether a firm is affected by a tariff or if that impact is positive or negative 45s.
• Some LLM outputs remain less structured, such as identifying specific products impacted by export controls, though the majority of the classification remains hard data 58s.
• A primary consideration for researchers using LLMs is the distinction between closed-weights and open-weights models 1m8s.
• Closed-weights models, such as ChatGPT, are proprietary, meaning the underlying algorithms and weights are hidden, which prevents researchers from ensuring full replicability if the provider updates the model in the background 1m15s.
• Open-weights models allow researchers to download the weights, run them on local infrastructure, and ensure deterministic results, mirroring the replicability standards of traditional software like Python, Stata, or MATLAB 1m40s.
• Beyond replicability, open-weights models offer the advantage of allowing researchers to fine-tune or modify the models to suit specific needs 2m5s.
• The primary disadvantages of open-weights models include higher computational infrastructure requirements and a more cumbersome setup process compared to simple API calls 2m15s.
• While computational costs for running open-weights models have decreased significantly over the last three years, the landscape remains highly volatile, with trends in model accessibility shifting rapidly between open and closed architectures in both the United States and China 2m25s.

Econometric Analysis and Firm-Level Evidence

• Econometric analysis of generated data requires addressing concerns regarding narrow replicability, potential biases, and non-classical measurement errors. 0s
• Recent research in econometrics is actively focused on developing methods to handle these data challenges. 0s
• ASML serves as a case study for analyzing geoeconomic pressure, where the firm faces pressure from the United States and the Netherlands to limit sales of extreme ultraviolet and deep ultraviolet products to China. 15s
• Algorithms can process public transcripts from firm executives to classify the nature of geopolitical pressure, the parties involved, and the resulting negative impact on the firm. 15s
• Researchers can extract granular data from these transcripts, such as whether a firm changed suppliers, increased prices, or reduced inputs, across numerous dimensions. 42s
• A significant advantage of using text-based analysis in geoeconomics is that global trade is dominated by multinational corporations, which are disproportionately likely to be public firms that produce accessible text. 55s
• Textual data can be collected from multiple points along a supply chain—including suppliers, the firm itself, and customers—allowing researchers to evaluate the broader economic consequences of specific actions, such as the cost of restricted access for Chinese customers. 1m20s
• Aggregate data on export controls shows a significant upward trend in recent years regarding the number of publicly listed firms mentioning such controls. 1m45s
• While early export control trends were driven largely by United States pressure on China regarding semiconductors, more recent data shows an increase in Chinese firms pressuring the United States through export controls on rare earth materials. 1m45s

Trends in Sanctions and Tariff Policies

• Simpler methods, such as bigrams, may be sufficient for identifying aggregate trends when a policy involves a specific, commonly used keyword like "tariff." 2m6s
• Export controls are complex to define, and while artificial intelligence can assist in phrasing them, manual analysis of biograms remains a viable alternative for identifying specific terminology 0s.
• Sanctions analysis reveals two primary historical episodes: the measures imposed on Russia following the invasion of Crimea and the war in Ukraine, and the sanctions imposed by the United States on Huawei and ZTE during the Trump administration 18s.
• Tariff analysis highlights the broad application of measures during both Trump administrations, which affected a wide range of global partners 33s.
• Geoeconomic pressure can be traced by identifying which sectors are being utilized by specific nations to exert influence, such as the United States targeting China via the semiconductor supply chain and China targeting the United States through rare earths 42s.
• Textual analysis of firm reports can reveal third-party effects, such as Indian or Chinese firms reporting positive outcomes from U.S. or European sanctions on Russian oil due to the availability of cheaper energy 1m5s.
• Theoretical models of economic pressure suggest that maximum leverage occurs when a target relies entirely on a single source for a resource or when the elasticity of substitution follows a Cobb-Douglas function 1m23s.
• Empirical data comparing U.S. and Chinese economic interactions indicates that simple models often fail to capture all relevant variables, necessitating the inclusion of additional factors 1m36s.
• Provisional AI analysis suggests that U.S. policymakers are generally targeting the correct sectors to exert pressure on China, aligning with the strategy of maintaining a "high fence on a small garden" rather than applying broad, indiscriminate controls 1m52s.
• Specific sectors like semiconductors and aircraft are frequently targeted for export controls, whereas sectors like automobiles have not been utilized by the United States against China 2m15s.
• In contrast to the targeted approach of export controls, tariff policies have historically been applied broadly across many countries, regardless of whether those nations were significant customers or had viable alternatives 2m32s.
• Economic theory distinguishes between threats to restrict sales, such as export controls, and threats to restrict purchases, which function similarly to the implementation of tariffs or full trade bans 2m48s.

Market Elasticity and Firm Responses to Coercion

• The theory of a tariff suggests that if a large market cuts off an exporter, the exporter must lower prices to find new customers, which may lead the exporter to offer concessions to the market imposing the tariff 0s.
• The ability of an exporter to find alternative customers depends on market elasticity; if a market is large, the exporter faces pressure to lower prices to maintain the same sales volume 0s.
• Russia serves as an example of this theory, as its inability to sell oil to Western Europe and the U.S. forces it to rely on smaller markets like India and China, reportedly resulting in a 30% discount on its exports 25s.
• Research plans involve analyzing sectors where the model predicts that finding alternative customers should be easy to determine if firms actually report that being cut off from a market had little impact 52s.
• For sectors where the U.S. represents a small portion of a firm's customer base, such as 10%, firms can likely reallocate their sales quickly, whereas firms selling specialized products like oversized SUVs may face greater difficulties 1m20s.
• When the U.S. implements export controls, Chinese firms report negative impacts on profits and respond by increasing domestic research and development (R&D) to replace U.S. semiconductors and avoid future supply disruptions 1m45s.
• U.S. firms also face high costs when cut off from large markets like China, as seen with Nvidia, which has responded by investing in R&D to develop chips that comply with U.S. export control regulations while still meeting Chinese demand 2m15s.
• Analysis of U.S. firms shows that while the majority report being negatively affected by U.S. tariffs, a subset of firms reports being positively affected 2m45s.

Economic Impacts of Tariffs on Firms

• A tariff is defined as a combination of two policies: a tax on consumers and a subsidy for domestic producers 3m15s.
• Economists often view tariffs unfavorably because they function as two combined policies, whereas separate instruments could target specific margins without creating unintended consequences, such as taxing consumers while subsidizing domestic producers 0s.
• Data analysis of US firms shows that those reporting positive effects from tariffs experience increased profit margins, while those reporting negative effects see a decrease in profit margins 18s.
• Firms negatively affected by tariffs are significantly more likely to report facing higher input prices, particularly in scenarios where foreign suppliers do not lower their prices to compensate for the tax 35s.
• While firms negatively affected by tariffs often pass some of these costs to consumers by increasing their own sales prices, the difference in price increases between positively and negatively affected firms is relatively muted 55s.
• A domestic steel mill that does not rely on imported steel can be positively affected by tariffs because its competitors, who do import steel, face higher input costs and raise their prices 1m12s.
• A domestic producer not facing higher input costs has the option to either capture market share by maintaining lower prices or increase their markup, effectively functioning as a domestic producer subsidy 1m30s.
• Research indicates that firms reporting positive effects from tariffs are the only ones planning to increase domestic investment 1m45s.

Methodological Challenges in Qualitative Data

• Systematic analysis of firm-level data serves as an ongoing project to test economic theories beyond traditional hard data, though this approach faces specific methodological challenges 1m55s.
• A primary limitation of using qualitative text data is the lack of precise quantitative information, as CEOs typically describe plans to change prices without specifying the magnitude or the exact rate of pass-through 2m15s.
• Data used in economic research often contains significant noise and bias that cannot be dismissed as simple measurement error 0s.
• Because the noise in large datasets is substantial, researchers cannot rely on standard techniques designed for classic measurement error 25s.
• Structural techniques exist to characterize and correct bias, though these methods are difficult to implement 45s.
• Informal methods, such as running multiple models to test for result reliability, serve as robustness checks but lack standardized criteria for determining if the exploration is sufficiently extensive 55s.
• Unlike human intelligence, which often demonstrates generalizability across different tasks, AI models can perform highly sophisticated functions while simultaneously failing on trivial dimensions 1m20s.
• AI models can sometimes outperform human researchers by identifying nuances in language that humans might overlook 1m50s.
• An example of AI sophistication occurred when a model correctly distinguished between "tariffs" as government-imposed import taxes and "tariffs" as general fee schedules for services like mobile phones or Netflix, whereas a simpler bigram analysis failed to make this distinction 2m5s.
• AI models can struggle with specific, nuanced phrasing, such as identifying the context of "increasing sanctions" or differentiating between formal export controls and export impediments caused by climate or safety legislation 2m45s.
• The ways in which AI models fail are often unpredictable, contrasting sharply with their high performance in other areas 3m5s.

Student Research and Future Directions

• Students have a significant opportunity to generate their own data for research papers, a practice that is transforming the field of economics 0s.
• Historical research methods, such as the paper-based surveys conducted by Schiller in the 1980s, have evolved significantly due to the scaling of online survey technology by the 2010s 0s.
• Researchers like Stephanie Stanchva have utilized modern technology to conduct surveys, and the use of AI as a classifier now allows for the generation of data from unstructured video and text 0s.
• Data regarding research on tariffs, export controls, and sanctions has been made available online, with plans to continue sharing such data publicly 42s.
• While the cost of conducting this type of research was initially very high, it has decreased significantly and is now considered relatively affordable, even for students 42s.
• Students are encouraged to utilize publicly available data sources, such as the congressional record and earnings calls, rather than relying solely on proprietary datasets 42s.
• Much of this research can be performed on a standard laptop, provided the researcher asks good questions and implements effective methods 42s.
• The competition between the United States and China and the resulting rearrangement of the world order are expected to be highly consequential topics that will shape economics for the next 20 years 2m6s.
• There is a growing research agenda in this field, utilizing tools from macroeconomics, trade, and political economy 2m6s.
• Current research efforts include exploring the dynamics of multiple hegemons, as opposed to the single-hegemon models previously discussed 2m6s.
• Future research interests include the political economy of conflicts between firms and their respective governments, with Nvidia cited as an example of such a conflict 2m6s.
• A conflict exists between the interests of firms like Nvidia, which aim to maximize profits by selling to all markets, and the United States government, which seeks to restrict sales to China for national security reasons 0s.
• The tension between a firm's national security value and its shareholder value creates incentives for corporate lobbying 15s.
• While some argue that China holds an advantage in geoeconomics due to its ability to dictate actions to state-owned enterprises, domestic political systems in other countries may impose participation constraints on what governments can demand from private firms 25s.
• Excessive government intervention or the commandeering of private technology for national security purposes can diminish the long-term incentives for domestic innovation 55s.
• Geoeconomics is distinct from soft power, such as cultural influence, and hard power, such as military threats or intervention 1m25s.
• Foreign Direct Investment (FDI) can serve as a mechanism for gaining power over foreign entities, allowing a country to dictate the actions of a firm by owning its equity rather than relying on external pressure or bullying 2m15s.
• Current research relies on structural models to infer the outcomes of economic restrictions, but there is a need for reduced-form methods to provide cleaner, more empirical evidence of these causal relationships 2m45s.
• Economists need to develop reliable models to evaluate industry claims for subsidies based on national interest, as most trade is not strategically sensitive and does not pose a national security risk 3m10s.
• A reading list has been provided for independent review during the summer period. 0s