Berger, Easterly, Nunn and Satyanath (2013) argue that CIA-backed political influence increased treated countries' imports from the United States by roughly a third. The claim is not that intervention opened markets in general, but that it redirected purchases toward American suppliers.
About the paper
Political influence operates through back channels: covert operations, diplomatic pressure, and secret funding. Berger, Easterly, Nunn and Satyanath (2013, American Economic Review) use declassified CIA records to observe when and where the United States installed or sustained friendly leaders during the Cold War.
The basic idea is simple. The CIA intervened for strategic and political reasons, not because economists were trying to boost exports. The authors then ask a narrower question: after a government became aligned with Washington, did that country start buying more from the United States?
The answer matters only because several pieces line up at once. Before intervention, there is no clear upward trend. After intervention, imports from the US rise, but total trade does not rise in the same way. The effect is also strongest in sectors where American firms were not obviously the best market choice, which makes political favoritism more plausible than simple market competition. The procurement results point in the same direction.
This is what American influence looked like.
The map shows countries in which the CIA installed, supported or helped sustain friendly leadership during the Cold War. Color intensity reflects the share of years each country spent under US-backed influence in the editorial dataset assembled for this project.
The geography of intervention was much broader than the handful of famous cases most people remember. The Cold War was not fought only in places like Iran, Guatemala, or Chile. It stretched across Latin America, the Middle East, Africa, and Asia, often through long-running relationships rather than single dramatic events.
Seen this way, the project begins less with trade and more with political reach. Before asking what intervention changed economically, it helps to see where the United States repeatedly tried to shape governments, alliances, and the direction of whole regimes.
Latin America was the most intensive arena.
Guatemala, Chile, Brazil, Nicaragua and their neighbors formed the densest cluster of sustained interventions. In this region, political influence often came with durable military and financial support, not merely one-off coups.
That matters for trade because the mechanism in the paper is gradual. Installed regimes do not alter procurement patterns instantly; relationships between state agencies, ministries and suppliers take time to form.
Guatemala spent 74% of the Cold War era under US-installed or US-supported governments in the project dataset.
The Middle East combined strategy and state demand.
Iran is the clearest case. After the 1953 coup, the Shah became a durable pro-Western ally for more than two decades. The intervention had obvious strategic significance, but it also sat in a region where state purchasing, oil and security relationships were deeply intertwined.
In settings like this, the procurement mechanism is easier to imagine concretely. If ministries, public agencies, or state-owned firms control a larger share of purchasing, a friendly regime can steer contracts toward American suppliers more directly than it could in a decentralized private market.
Asia shows why the question is about influence, not only coups.
South Korea, Taiwan, the Philippines, Thailand and South Vietnam were not all identical cases of regime overthrow. Some were long-run security relationships, others were overt or covert support for authoritarian consolidation.
That broader pattern helps motivate the paper's language of political influence. The economic question is whether governments aligned with Washington purchased differently once that alignment became politically binding.
This is what American influence looked like.
The map shows countries in which the CIA installed, supported or helped sustain friendly leadership during the Cold War. Color intensity reflects the share of years each country spent under US-backed influence in the editorial dataset assembled for this project.
The geography of intervention was much broader than the handful of famous cases most people remember. The Cold War was not fought only in places like Iran, Guatemala, or Chile. It stretched across Latin America, the Middle East, Africa, and Asia, often through long-running relationships rather than single dramatic events.
Seen this way, the project begins less with trade and more with political reach. Before asking what intervention changed economically, it helps to see where the United States repeatedly tried to shape governments, alliances, and the direction of whole regimes.
Latin America was the most intensive arena.
Guatemala, Chile, Brazil, Nicaragua and their neighbors formed the densest cluster of sustained interventions. In this region, political influence often came with durable military and financial support, not merely one-off coups.
That matters for trade because the mechanism in the paper is gradual. Installed regimes do not alter procurement patterns instantly; relationships between state agencies, ministries and suppliers take time to form.
Guatemala spent 74% of the Cold War era under US-installed or US-supported governments in the project dataset.
The Middle East combined strategy and state demand.
Iran is the clearest case. After the 1953 coup, the Shah became a durable pro-Western ally for more than two decades. The intervention had obvious strategic significance, but it also sat in a region where state purchasing, oil and security relationships were deeply intertwined.
In settings like this, the procurement mechanism is easier to imagine concretely. If ministries, public agencies, or state-owned firms control a larger share of purchasing, a friendly regime can steer contracts toward American suppliers more directly than it could in a decentralized private market.
Asia shows why the question is about influence, not only coups.
South Korea, Taiwan, the Philippines, Thailand and South Vietnam were not all identical cases of regime overthrow. Some were long-run security relationships, others were overt or covert support for authoritarian consolidation.
That broader pattern helps motivate the paper's language of political influence. The economic question is whether governments aligned with Washington purchased differently once that alignment became politically binding.
Intervention timeline
Main finding
The main estimate comes from a gravity-style panel regression that compares a country to its own earlier trade path while also netting out common year shocks and the usual determinants of bilateral trade. In plain language, the question is not whether aligned countries traded a lot in absolute terms, but whether imports from the United States rose once a government came under durable US-backed influence.
In the baseline specification, CIA intervention is associated with a coefficient of 0.293, which translates into roughly a 34% increase in a treated country's imports from the United States.
On its own, that number would not settle much. The interpretation depends on the surrounding checks: little evidence of pre-trends, no comparable rise in exports to the US, no broad import boom, and stronger effects where political favoritism should matter most.
That is why the next figure matters. Before looking at the size of the post-intervention coefficients, you want to know whether the pre-period is flat and whether the uplift appears only after intervention. If that timing holds, the 0.293 estimate reads less like noise and more like a change in sourcing toward American suppliers.
Before the plot
Read the event study as a before-and-after timing test. The points before intervention show whether countries that were later treated were already moving toward higher imports from the United States before the intervention happened. If those pre-intervention estimates are roughly flat, it becomes harder to say the result is just picking up an earlier trend that was already in motion.
The points after intervention ask the next question: do imports rise only once political alignment changes? That is the pattern you would want to see if the paper's story is right. Little movement before t=0 and a lasting shift upward afterward would fit the idea that intervention changed procurement and sourcing decisions rather than simply formalizing a change that had already begun.
The gravity model isolates the trade margin of interest.
International trade follows systematic patterns: larger economies trade more, and nearby economies trade more. The gravity model formalizes those regularities and is the standard empirical framework for bilateral trade.
The quantity of interest here is narrow: how much did a country's imports from the United States change after CIA intervention, net of country fixed effects, year effects, and the usual gravity controls?
The event study is useful because it turns that question into a sequence of before-and-after comparisons around intervention timing.
Before intervention, the series is comparatively flat.
The pre-treatment coefficients around t-3 to t-1 are centered near zero and do not suggest a clear upward trend. That is the key design check.
No event study can prove exogeneity on its own, but a flat pre-period makes the paper's interpretation much more credible. The US does not appear to have been selecting countries already drifting toward stronger import dependence on American suppliers.
The identifying argument rests less on the size of any one coefficient than on the absence of a visible pre-trend before intervention.
After intervention, imports from the US move upward.
From roughly t+2 onward, the coefficients turn positive and remain elevated. By t+5, the estimated effect is about a 28% increase relative to the counterfactual path.
That lag is substantively plausible. Political intervention may be immediate, but the commercial channel the paper emphasizes, especially procurement, would work through ministries, contracts and state purchasing decisions that take time to reshape.
The gravity model isolates the trade margin of interest.
International trade follows systematic patterns: larger economies trade more, and nearby economies trade more. The gravity model formalizes those regularities and is the standard empirical framework for bilateral trade.
The quantity of interest here is narrow: how much did a country's imports from the United States change after CIA intervention, net of country fixed effects, year effects, and the usual gravity controls?
The event study is useful because it turns that question into a sequence of before-and-after comparisons around intervention timing.
Before intervention, the series is comparatively flat.
The pre-treatment coefficients around t-3 to t-1 are centered near zero and do not suggest a clear upward trend. That is the key design check.
No event study can prove exogeneity on its own, but a flat pre-period makes the paper's interpretation much more credible. The US does not appear to have been selecting countries already drifting toward stronger import dependence on American suppliers.
The identifying argument rests less on the size of any one coefficient than on the absence of a visible pre-trend before intervention.
After intervention, imports from the US move upward.
From roughly t+2 onward, the coefficients turn positive and remain elevated. By t+5, the estimated effect is about a 28% increase relative to the counterfactual path.
That lag is substantively plausible. Political intervention may be immediate, but the commercial channel the paper emphasizes, especially procurement, would work through ministries, contracts and state purchasing decisions that take time to reshape.
Trade margins
This is where the economics becomes much clearer. The paper does not find a generic increase in trade after intervention. It finds a rise in one very specific margin: treated countries import more from the United States.
That distinction matters. If intervention simply stabilized economies or opened them to the world, we would expect broader import growth or a more symmetric pattern across trade margins. Instead, the movement is narrow and directional, which is exactly what makes the result economically interesting.
In geoeconomic terms, this is a distortion of allocation. Political access appears to have shifted demand toward American suppliers without expanding overall trade proportionately.
Approximate percent changes with 95% confidence intervals
The pattern is economically important because it narrows the interpretation. Political influence appears to reallocate demand toward US suppliers rather than raising trade across all margins.
The remaining question is why the shift favored US firms.
Once we know the effect is concentrated in imports from the US, the next task is to decide whether this looks like ordinary liberalization or politically mediated favoritism.
The comparative-advantage test addresses exactly that question. If intervention simply made economies more open, American gains should be strongest in sectors where US producers were already globally competitive.
The gains are largest where the US is least competitive.
The estimated slope is negative. As US comparative advantage rises, the intervention effect falls. The strongest gains are in low-RCA sectors, exactly where American firms would be least likely to outperform on price or efficiency alone.
That is why this otherwise abstract chart matters. A simple market-opening story would predict the opposite pattern. The actual pattern is more consistent with purchases being steered toward American suppliers for political reasons, especially in sectors where economic fundamentals alone would not deliver the same outcome.
The RCA result is a support test, not the headline result. Its value is that it helps distinguish political favoritism from generic liberalization.
The mechanism is most plausible in government demand.
Once the sector pattern is established, procurement becomes the natural next question. Governments can influence what ministries and public agencies buy; they cannot directly command private consumers to favor foreign suppliers.
The interaction estimates point in that direction. The effect is weak when the government share is negligible and becomes much larger in settings where public demand is more important.
Read cautiously, the procurement evidence makes the political-favoritism interpretation institutionally credible.
The remaining question is why the shift favored US firms.
Once we know the effect is concentrated in imports from the US, the next task is to decide whether this looks like ordinary liberalization or politically mediated favoritism.
The comparative-advantage test addresses exactly that question. If intervention simply made economies more open, American gains should be strongest in sectors where US producers were already globally competitive.
The gains are largest where the US is least competitive.
The estimated slope is negative. As US comparative advantage rises, the intervention effect falls. The strongest gains are in low-RCA sectors, exactly where American firms would be least likely to outperform on price or efficiency alone.
That is why this otherwise abstract chart matters. A simple market-opening story would predict the opposite pattern. The actual pattern is more consistent with purchases being steered toward American suppliers for political reasons, especially in sectors where economic fundamentals alone would not deliver the same outcome.
The RCA result is a support test, not the headline result. Its value is that it helps distinguish political favoritism from generic liberalization.
The mechanism is most plausible in government demand.
Once the sector pattern is established, procurement becomes the natural next question. Governments can influence what ministries and public agencies buy; they cannot directly command private consumers to favor foreign suppliers.
The interaction estimates point in that direction. The effect is weak when the government share is negligible and becomes much larger in settings where public demand is more important.
Read cautiously, the procurement evidence makes the political-favoritism interpretation institutionally credible.
Procurement test
The procurement channel is not just a narrative flourish. In the paper's interaction specification, the estimated intervention effect rises with the government share of demand. At zero government share, the effect is close to nil; at higher public shares, it becomes economically meaningful.
That does not prove every extra dollar flowed through ministries, but it is difficult to reconcile with a story in which intervention simply made the whole economy more open in a neutral, market-wide way.
Why it matters
The broader economic importance of the paper is not the size of one coefficient. It is the demonstration that political influence can alter the pattern of trade even when comparative advantage points elsewhere.
That matters for how we think about the international economy. Trade flows are often treated as the outcome of prices, technology and institutions. This paper shows that geopolitical leverage can also reweight who wins market access.
In that sense, the project is about more than the Cold War. It is an early and unusually clean example of geoeconomics: state power reshaping commercial outcomes through political channels rather than through open protectionism alone.
The evidence is consistent with a form of commercial imperialism. After CIA intervention, treated countries imported more from the United States, and the pattern is hardest to explain with a simple market-opening story.
The interpretation becomes convincing because the supporting pieces line up: flat pre-trends, concentration in low-RCA sectors, asymmetry across trade margins, and stronger effects where public procurement matters more.
The paper does not show that intervention increased trade in general. It shows a more specific pattern: aligned regimes appear to have shifted purchases toward American suppliers, plausibly through channels tied to state demand.
Berger, Daniel, William Easterly, Nathan Nunn, and Shanker Satyanath. “Commercial Imperialism? Political Influence and Trade During the Cold War.” American Economic Review 103, no. 2 (2013): 863-96.
Visualization by Jona Wilke