By: Koichiro Ito, James M. Sallee, and Andrew Smith
Context
Policies are generally assessed for the impact they have domestically, ignoring the worldwide impact in an increasingly globalized society. For example, when a government subsidizes fuel-efficient cars, typical analysts only count the emissions saved within that country’s borders. But this assumption breaks down in a world where the same multinational car company designs and manufactures cars sold in dozens of countries.
To take the example further, if a policy in Japan pushes Honda to engineer a more fuel-efficient Civic for Japanese consumers, does Honda redesign just the Japanese Civic—or does the more efficient engine also end up in the American one? The answer depends on a straightforward economic logic: it can be cheaper to apply one design across all markets than to maintain separate product lines. This paper introduces a name for this phenomenon—attribute propagation—and rigorously measures how large it actually is.
Research Design
The authors exploit a natural experiment: Japan’s 2009 fuel-economy subsidy, which gave consumers a rebate of roughly $700–$1,500 when purchasing a car whose fuel economy exceeded a government-set target. The policy created strong incentives for firms to improve the fuel economy of models sold in Japan.
The key observation is that many models sold in Japan: The Civic, Accord, Corolla were also sold in the United States, while many other Japanese-brand models were sold only in the U.S. This creates a plausible treatment and control group within the American market: Models sold in both countries (potentially affected by the Japanese subsidy through attribute propagation) versus models sold only in the U.S. (unaffected). The authors compared how the fuel economy of these two groups evolved in the U.S. after 2009. They then built a model of multinational car markets to capture how competing vehicles adjust in response, and they simulate what would have happened had Japan never enacted the subsidy.
Findings
Japan’s subsidy significantly improved the fuel economy of cars sold in the United States.
The subsidy in Japan led to an 8.65 percent improvement in the fuel economy of affected vehicles sold in the United States even though the subsidy had no formal jurisdiction over American sales. As the chart below illustrates, those cars sold in both countries (treatment) and those sold just in the U.S. (control) tracked each other closely before 2009, then diverged sharply afterward.
The carbon benefits abroad are four times larger than those at home.
The U.S. market is much larger and Americans drive far more miles per year (11,21 8 vs. 3,206 annual miles per vehicle) compared to the Japanese markets. These two factors amplify the environmental impact of Japan’s policy in the U.S. The Japanese subsidy reduced carbon by 0.75 megatons per year domestically and by 2.27 megatons per year in the U.S., meaning roughly three-quarters of the policy’s total climate benefit occurred outside Japan’s borders. Put another way, the combined carbon reduction in Japan and the U.S. was four times larger than the carbon reduction in Japan alone.
Closing Take-Away
This study suggests policymakers might often underestimate the impact of their own regulations. By focusing only on domestic impacts, they miss a potentially larger benefit that travels through multinational firms’ product design choices. This paper provides the first rigorous evidence that these “attribute propagation” effects are not just theoretically plausible—they actually occur. In the case of this study, a domestic fuel-economy policy implemented in Japan generated more than five times its domestic climate benefit once global impacts are counted.
The implications extend well beyond automobiles. The same logic could apply wherever multinational firms sell standardized or near-standardized products: pharmaceuticals, electronics, appliances, chemical formulations, etc. Policies in major regulatory markets (the European Union’s chemical rules, FDA drug approval standards), may be quietly improving products worldwide without receiving any credit for it. Accounting for attribute propagation could fundamentally change cost-benefit analyses of environmental, safety, and technology regulations.