Economics and Epidemics: Evidence from an Estimated Spatial Econ-SIR Model

The Economic Feedback Loop of an Epidemic

Why do some economies recover swiftly from a pandemic while others remain suppressed for years? A new study from the Federal Reserve Board and the University of Pittsburgh suggests the answer lies in a hidden feedback loop. The researchers propose that when people fear infection, they voluntarily alter their economic lives. This creates a "slow burn" of both disease and recession.

The Bi-Directional Tradeoff

Managing a pandemic requires balancing two competing priorities: public health and economic stability. Traditionally, epidemiological models treated the economy as a static backdrop. Conversely, economic models often viewed health crises as external shocks.

The authors argue that this separation is a mistake. In reality, the two systems are deeply intertwined. When a virus spreads, individuals face a personal tradeoff. They weigh the economic benefit of participating in the marketplace against the physical cost of potentially contracting the disease. This creates a bi-directional interaction. High infection rates drive people to reduce economic activity through endogenous social distancing (voluntary behavior changes that slow transmission). This, in turn, slows the virus's spread. Understanding this loop is essential for predicting whether an epidemic will explode or settle into a long, protracted period of low-level transmission.

Beyond the Standard SIR Model

To capture this complexity, the authors develop an "Econ-SIR" model. This builds upon the foundational SIR framework—a mathematical model that tracks Susceptible (S), Infected (I), and Recovered (R) populations. In a standard SIR model, the rate of infection is usually treated as a fixed biological property.

The Econ-SIR model adds a layer of micro-founded economic decision-making. To ground these claims, the researchers use granular, daily data from 921 U.S. counties. This includes SafeGraph GPS data to track foot traffic and Homebase payroll records to monitor hours worked. Each susceptible agent evaluates their "economic need" ($z$) against the expected cost of infection ($\psi$). If the risk outweighs the benefit of a transaction, the agent chooses to stay home. The paper defines the rate of economic activity by susceptibles ($a_S$) as a function of this risk .

The authors extend this into a "Spatial Econ-SAIRD" model. This version accounts for asymptomatic carriers (A) who spread the virus unknowingly, mortality (D), and geographic mobility. By linking these datasets, the researchers estimate how biological and economic forces actually interacted during the COVID-19 pandemic.

Predicting the "Slow Burn"

The most striking result of the Econ-SIR framework is how it changes our understanding of epidemic trajectories. In a standard SIR model, the virus typically sweeps through a population until it hits a "herd immunity threshold." At that point, it vanishes. The authors report that including endogenous human behavior changes this fundamentally.

Instead of a single sharp peak, the model predicts two distinct phases. First, there is an initial period of exponential growth. However, as infections rise, the rising risk causes people to pull back from the economy. This creates a second phase: a "slow burn." In this regime, the effective reproduction rate ($R_e$) hovers near one. This leads to a protracted period where case levels and economic activity remain roughly constant for years .

The authors use this model to run counterfactual experiments. They find that under a "laissez-faire" approach (no government intervention), the U.S. could face approximately 1.25 million deaths over four years. During the first year, economic output would remain more than 10% below pre-pandemic levels . Interestingly, the study finds that the effectiveness of lockdowns depends heavily on the timeline of medical breakthroughs. If a vaccine is expected within two years, aggressive non-pharmaceutical interventions (NPIs) can dramatically reduce the total number of infections. This preserves long-term economic health. However, if no vaccine is on the horizon, lockdowns offer limited utility. They cannot change the eventual herd immunity threshold.

Redefining Policy Success

This research shifts the goalposts for pandemic policy. It suggests that the most effective way to manage an epidemic is to aim for an "eradication zone." This is a threshold of low enough transmission that community spread can be managed through testing and tracing.

The paper highlights a critical nuance. Mitigation measures that reduce transmission without halting all commerce—such as mask-wearing—can improve both health and economic outcomes. By lowering the cost of safe interaction, these measures allow the economy to function while preventing the "slow burn" that keeps markets suppressed.

Where the Edges Are

While powerful, the Econ-SAIRD framework has specific limitations. The model assumes that individuals gain permanent immunity after recovering. This is a premise the authors note is challenged by emerging evidence regarding reinfection. Additionally, the estimation process prioritizes fitting the dynamics of higher-population counties. This means the model may be less accurate for sparsely populated rural areas. Finally, the model's success relies on the "internal propagation mechanisms" triggered by the initial infection shock. It does not include a mechanism for the model to correct itself if it deviates from real-world data midway through a simulation.

Figures from the paper