When it comes to designing taxes and subsidies to fix social problems, traditional economics says policymakers only need to know the size of the harm—set the tax equal to the damage, and you're done. But a new working paper from the Federal Reserve Bank of Boston shows that approach breaks down when policies face high implementation costs, require distortionary taxes to fund them, or create significant distributional consequences. The research demonstrates this failure through two case studies: London's congestion pricing, where extremely high fixed administrative costs justified a four-decade shift from measuring traffic harm to predicting traffic reduction, and Boston's proposed housing subsidies, where the analysis reveals a large ratio of total subsidized units to marginal induced units, indicating high costs per additional unit actually built.
The paper presents a framework showing when policymakers should focus on measuring externalities versus predicting quantity responses to interventions. According to the report, the optimal tax or subsidy deviates from the externality size when taxes create deadweight losses or when distributional consequences matter, making it lower for subsidies and higher for taxes than the externality alone would suggest. For Boston's housing subsidies specifically, the framework reveals that these subsidies become cost-effective primarily when they generate secondary benefits such as neighborhood revitalization or additional tax revenues from neighboring properties, which reduce the distortionary taxation. In the London congestion pricing case, the study documents how policy discussions evolved from the externality-focused 1964 Smeed Report—which analyzed traffic harms—toward quantity-focused analysis over four decades, driven by the need to justify the pricing system through significant traffic reduction given extremely high fixed administrative costs.
The report finds that "the traditional Pigouvian approach of focusing exclusively on externality measurement is insufficient when policies involve substantial implementation costs, require distortionary taxation for funding, or create significant distributional consequences." The authors write that when implementation costs are low and distributional concerns are negligible, policymakers should focus on externality magnitudes, but conversely, when taxes impose social costs or distributional issues are significant, understanding quantity responses becomes more valuable. According to the research, policymakers and economists must invest resources in understanding both the magnitudes of externalities and quantity responses to interventions, particularly in estimating the ratio of total affected units to marginal units induced by policy interventions.
The disconnect between textbook economics and real-world policy becomes clear in how practitioners actually make decisions. The report explains that policy discussions about congestion pricing focus heavily on predicting traffic reduction rather than measuring traffic harms alone, while housing subsidy decisions center on how many units will be built rather than on housing's social benefits. This happens because real policies don't operate in the frictionless world of economic theory—they require expensive infrastructure to administer, they're funded by taxes that themselves distort the economy, and they redistribute resources in ways that create political winners and losers. When a congestion pricing system costs hundreds of millions to set up, knowing that traffic congestion causes $50 million in annual harm isn't enough—you need to know whether the system will actually reduce traffic enough to justify those upfront costs. Similarly, when housing subsidies are funded by distortionary taxes and primarily benefit certain income groups, understanding how many additional units each dollar actually induces becomes critical to evaluating whether the policy makes economic sense.
The findings suggest that strengthening public support for welfare-improving Pigouvian policies requires more comprehensive empirical work that documents both price and quantity effects, rather than relying primarily on externality estimates to guide policy design. For practical policy evaluation, the report concludes that cost-benefit analyses should emphasize how many units of behavioral change each dollar of intervention actually purchases, not merely the value of eliminating the externality. The bottom line: when implementation isn't free and politics matter, measuring the problem is only half the battle—you also need to know whether your solution will actually work.