Data Visualization by Jona Wilke
Fernandez-Villaverde, Mineyama and Song (2024) do two things at once: they turn fragmentation into a measurable index built from 16 indicators, and they trace what fragmentation shocks do to output, investment, and markets. The result is hard to dismiss as mood or metaphor. Fragmentation looks increasingly political in origin, and the economic costs are large enough to show up in macro data.
About the paper
The measurement problem comes first: fragmentation does not appear in one clean series. It shows up in tariffs, sanctions, capital flows, FDI, migration, patenting, geopolitical risk, energy uncertainty, and voting patterns in international institutions. Any single proxy catches only one slice of the process, which is why so much of the debate talks past itself.
The paper handles that by treating fragmentation as a latent variable. A dynamic hierarchical factor model extracts what 16 indicators share while still allowing trade, finance, mobility, and politics to move differently. That is the key analytical move: it separates a broad change in the world economy from noise in any one market.
The second contribution is causal. Using structural VARs and local projections on quarterly data for 89 economies, the authors estimate how fragmentation shocks affect GDP, industrial production, fixed investment, and stock prices. The pattern is not subtle: the shocks look contractionary, the damage persists, and emerging markets take the larger hit.
What did fragmentation look like during the Cold War?
The common factor compresses 16 noisy series into one index that captures the shared geopolitical setting behind trade, finance, mobility, and politics. During the Cold War that index is high, but it is also fairly steady, which fits a world divided into blocs but organized around a durable equilibrium.
Cross-border exchange still expanded within blocs, but the broader system was not politically open in the sense that marked the 1990s and early 2000s. The point is not that fragmentation was absent before 1989. It was normal.
Was the post-1989 decline more than a trade boom?
Yes. The collapse of the Soviet bloc, European integration, the WTO, and China's entry into global trade pushed the index down across multiple dimensions at once. What changed was not only tariffs or one financial series, but the broader expectation that cross-border integration was politically safe.
That is why the trough around 2007 matters. It marks peak confidence in an integrated world economy, not just unusually high openness in one sector.
The trough around 2007 is best read as peak confidence in integration.
Why has the common index kept rising since 2008?
The Global Financial Crisis, Crimea, the US-China trade war, COVID-19, and Russia's invasion of Ukraine all push the index upward. What matters is not any single episode, but the sequence: each shock lands on top of the last one, and the reversals are weak.
That persistence changes behavior. If firms begin to treat integration as reversible, they invest differently, hedge differently, and put less weight on the efficiency gains from deep cross-border specialization.
The issue is cumulative: integration no longer looks secure.
What did fragmentation look like during the Cold War?
The common factor compresses 16 noisy series into one index that captures the shared geopolitical setting behind trade, finance, mobility, and politics. During the Cold War that index is high, but it is also fairly steady, which fits a world divided into blocs but organized around a durable equilibrium.
Cross-border exchange still expanded within blocs, but the broader system was not politically open in the sense that marked the 1990s and early 2000s. The point is not that fragmentation was absent before 1989. It was normal.
Was the post-1989 decline more than a trade boom?
Yes. The collapse of the Soviet bloc, European integration, the WTO, and China's entry into global trade pushed the index down across multiple dimensions at once. What changed was not only tariffs or one financial series, but the broader expectation that cross-border integration was politically safe.
That is why the trough around 2007 matters. It marks peak confidence in an integrated world economy, not just unusually high openness in one sector.
The trough around 2007 is best read as peak confidence in integration.
Why has the common index kept rising since 2008?
The Global Financial Crisis, Crimea, the US-China trade war, COVID-19, and Russia's invasion of Ukraine all push the index upward. What matters is not any single episode, but the sequence: each shock lands on top of the last one, and the reversals are weak.
That persistence changes behavior. If firms begin to treat integration as reversible, they invest differently, hedge differently, and put less weight on the efficiency gains from deep cross-border specialization.
The issue is cumulative: integration no longer looks secure.
Key result
The central result is not just that the common index rises after 2008. It is that the political component pulls away from the rest and starts leading the process. Trade and finance still matter, but they increasingly look like transmission channels for political rupture rather than autonomous drivers.
That matters because market adjustment has limits. Firms can reroute some trade and capital; they cannot arbitrage away sanctions, export controls, strategic distrust, or bloc rivalry. Once politics leads, persistence becomes the main economic problem.
In the paper's estimates, the political index reaches 3.48 by 2023 Q4, while GDP losses in emerging markets are roughly three times as large as in advanced economies. Put differently: the shift is political at the source, but the costs travel through trade, finance, and investment.
Do the four dimensions move together?
Not exactly. The hierarchical model splits the common signal into trade, finance, mobility, and politics, which matters because it stops the analysis from treating every deterioration as the same kind of shock.
Trade often lags, finance moves quickly, and mobility has its own rhythm. The decomposition is not decorative. It shows where fragmentation starts and how it spreads.
Why is financial fragmentation the most cyclical layer?
Cross-border capital flows and FDI surged during the integration era, then reversed sharply after 2008. Because finance reprices quickly, this dimension tracks the common factor more closely than trade or mobility.
That is why geopolitical shocks often show up first in capital allocation. Finance is the fastest bridge from political tension to macroeconomic cost.
Finance is where geopolitical shocks become macroeconomic constraints first.
Why does the political dimension now lead?
Since 2008 the political factor rises much more sharply than the others. Sanctions, conflict risk, energy insecurity, and diplomatic distance are no longer background conditions. They drive the process.
Once politics leads, the economic implications are harder to unwind. Relative prices can adjust; political distrust tends to linger, which makes long-run expectations more fragile.
Political fragmentation matters because it makes economic fragmentation harder to reverse.
Do the four dimensions move together?
Not exactly. The hierarchical model splits the common signal into trade, finance, mobility, and politics, which matters because it stops the analysis from treating every deterioration as the same kind of shock.
Trade often lags, finance moves quickly, and mobility has its own rhythm. The decomposition is not decorative. It shows where fragmentation starts and how it spreads.
Why is financial fragmentation the most cyclical layer?
Cross-border capital flows and FDI surged during the integration era, then reversed sharply after 2008. Because finance reprices quickly, this dimension tracks the common factor more closely than trade or mobility.
That is why geopolitical shocks often show up first in capital allocation. Finance is the fastest bridge from political tension to macroeconomic cost.
Finance is where geopolitical shocks become macroeconomic constraints first.
Why does the political dimension now lead?
Since 2008 the political factor rises much more sharply than the others. Sanctions, conflict risk, energy insecurity, and diplomatic distance are no longer background conditions. They drive the process.
Once politics leads, the economic implications are harder to unwind. Relative prices can adjust; political distrust tends to linger, which makes long-run expectations more fragile.
Political fragmentation matters because it makes economic fragmentation harder to reverse.
Why does fragmentation look like a bloc story?
Because the world economy is not fragmenting uniformly. The relevant question is not whether openness falls everywhere at once, but which relationships are breaking and which political networks absorb the shock.
Before 2008 the blocs move broadly together; afterward their paths separate. That is what a geopolitical break looks like in the data.
Why is the China-Russia bloc diverging fastest?
Crimea, the US-China trade war, and the post-2022 sanctions regime all deepen that split. The result is not a simple fall in global integration, but a reallocation of integration across rival political networks.
For firms, the decision problem changes with it. Efficiency still matters, but now it is efficiency conditional on bloc exposure.
Fragmentation changes the geography of market access.
Why are the global costs asymmetric?
The paper finds that shocks linked to the US-EU bloc have immediate global GDP effects, whereas China-Russia shocks transmit more weakly. That asymmetry reflects the current architecture of finance, payments, and reserve-currency power.
Fragmentation is costly not just because blocs are separating, but because the bloc closest to the core of the system is part of the rupture.
Where the rupture happens matters as much as the rupture itself.
Why does fragmentation look like a bloc story?
Because the world economy is not fragmenting uniformly. The relevant question is not whether openness falls everywhere at once, but which relationships are breaking and which political networks absorb the shock.
Before 2008 the blocs move broadly together; afterward their paths separate. That is what a geopolitical break looks like in the data.
Why is the China-Russia bloc diverging fastest?
Crimea, the US-China trade war, and the post-2022 sanctions regime all deepen that split. The result is not a simple fall in global integration, but a reallocation of integration across rival political networks.
For firms, the decision problem changes with it. Efficiency still matters, but now it is efficiency conditional on bloc exposure.
Fragmentation changes the geography of market access.
Why are the global costs asymmetric?
The paper finds that shocks linked to the US-EU bloc have immediate global GDP effects, whereas China-Russia shocks transmit more weakly. That asymmetry reflects the current architecture of finance, payments, and reserve-currency power.
Fragmentation is costly not just because blocs are separating, but because the bloc closest to the core of the system is part of the rupture.
Where the rupture happens matters as much as the rupture itself.
How do fragmentation shocks reach the real economy?
The structural VARs show that a one-standard-deviation fragmentation shock pushes down stock prices, industrial production, fixed investment, and GDP per capita. The mechanism is broad: expectations weaken, financing tightens, and real activity slows.
That breadth is the point. Fragmentation behaves less like a tariff change and more like a regime shock to the expected returns on openness.
Why does GDP stay below its pre-shock path?
GDP per capita keeps deteriorating for roughly two years and does not return quickly to its pre-shock path. That persistence is the main macro result.
A plausible reading is that fragmentation lowers the level of efficiency by weakening specialization, slowing technology diffusion, and discouraging cross-border complementarities that are hard to rebuild.
This looks like a level loss, not a short-lived disturbance.
Why is investment the main amplification channel?
Fixed investment falls much more than GDP. Firms respond to fragmentation not only by producing less, but by cutting the capital spending that would have supported future growth.
Once investment falls, the shock compounds. Lower spending today means lower productive capacity tomorrow, which is why even a politically temporary break can leave durable economic damage.
Investment is where geopolitical tension becomes medium-run growth damage.
How do fragmentation shocks reach the real economy?
The structural VARs show that a one-standard-deviation fragmentation shock pushes down stock prices, industrial production, fixed investment, and GDP per capita. The mechanism is broad: expectations weaken, financing tightens, and real activity slows.
That breadth is the point. Fragmentation behaves less like a tariff change and more like a regime shock to the expected returns on openness.
Why does GDP stay below its pre-shock path?
GDP per capita keeps deteriorating for roughly two years and does not return quickly to its pre-shock path. That persistence is the main macro result.
A plausible reading is that fragmentation lowers the level of efficiency by weakening specialization, slowing technology diffusion, and discouraging cross-border complementarities that are hard to rebuild.
This looks like a level loss, not a short-lived disturbance.
Why is investment the main amplification channel?
Fixed investment falls much more than GDP. Firms respond to fragmentation not only by producing less, but by cutting the capital spending that would have supported future growth.
Once investment falls, the shock compounds. Lower spending today means lower productive capacity tomorrow, which is why even a politically temporary break can leave durable economic damage.
Investment is where geopolitical tension becomes medium-run growth damage.
Do the GDP effects survive different identification strategies?
The paper does not rely on one identification scheme. Cholesky ordering, narrative restrictions, and external controls all point in the same direction: fragmentation is contractionary.
That consistency matters because the claim is ambitious. If the sign and ranking survive across setups, the result is much harder to dismiss as a quirk of specification.
Who bears the largest losses?
Emerging markets lose roughly three times as much GDP as advanced economies. That is not incidental. It reflects greater dependence on external demand, imported technology, and foreign financing.
The distributional point is hard to miss: the countries least able to shape geopolitical rivalry are often the ones most exposed to its macroeconomic costs.
The periphery pays for shocks generated by the core.
Where are the losses concentrated?
Manufacturing, trade-related services, and finance suffer the most. Public administration and other domestically oriented sectors move much less. That is exactly what one would expect if fragmentation works by damaging cross-border networks.
Sectoral concentration helps on interpretation. The shock lands where trade, finance, and specialization are most deeply embedded in production, which is harder to reconcile with a story of generic pessimism.
The most global sectors are the first to show the cost.
Why is reintegration slower than fragmentation?
The impulse responses show a final asymmetry: the costs of fragmentation arrive within quarters, while the gains from lower fragmentation emerge more slowly.
That matters because political shocks can destroy an integrated equilibrium faster than markets can rebuild it.
Fragmentation moves faster than reintegration.
Do the GDP effects survive different identification strategies?
The paper does not rely on one identification scheme. Cholesky ordering, narrative restrictions, and external controls all point in the same direction: fragmentation is contractionary.
That consistency matters because the claim is ambitious. If the sign and ranking survive across setups, the result is much harder to dismiss as a quirk of specification.
Who bears the largest losses?
Emerging markets lose roughly three times as much GDP as advanced economies. That is not incidental. It reflects greater dependence on external demand, imported technology, and foreign financing.
The distributional point is hard to miss: the countries least able to shape geopolitical rivalry are often the ones most exposed to its macroeconomic costs.
The periphery pays for shocks generated by the core.
Where are the losses concentrated?
Manufacturing, trade-related services, and finance suffer the most. Public administration and other domestically oriented sectors move much less. That is exactly what one would expect if fragmentation works by damaging cross-border networks.
Sectoral concentration helps on interpretation. The shock lands where trade, finance, and specialization are most deeply embedded in production, which is harder to reconcile with a story of generic pessimism.
The most global sectors are the first to show the cost.
Why is reintegration slower than fragmentation?
The impulse responses show a final asymmetry: the costs of fragmentation arrive within quarters, while the gains from lower fragmentation emerge more slowly.
That matters because political shocks can destroy an integrated equilibrium faster than markets can rebuild it.
Fragmentation moves faster than reintegration.
Key takeaways
The paper's first contribution is measurement. Fragmentation is treated as a common index built from multiple indicators, not inferred from one series at a time.
The second contribution is causal. Higher fragmentation lowers economic activity, and the estimated losses are larger in emerging economies than in advanced ones.
The final point is about distribution and timing. The sectors most tied to global markets are hit harder, and the damage from fragmentation arrives faster than the gains from lower fragmentation.
Fernandez-Villaverde, Jesus, Tomohide Mineyama, and Dongho Song. “Are We Fragmented Yet? Measuring Geopolitical Fragmentation and Its Causal Effect.” NBER Working Paper 32638 (2024).
Visualization by Jona Wilke