Search

We analyze the externalities that arise when social and economic interactions transmit infectious diseases such as COVID-19. Individually rational agents do not internalize that they impose infection externalities upon others when the disease is transmitted. In an SIR model calibrated to capture the main features of COVID-19 in the US economy, we show that private agents perceive the cost an additional infection to be around $80k whereas the social cost including infection externalities is more than three times higher, around $286k. This misvaluation has stark implications for how society ultimately overcomes the disease: for a population of individually rational agents, the precautionary behavior by the susceptible flattens the curve of infections, but the disease is not overcome until herd immunity is acquired. The resulting economic cost is high; an initial sharp decline in...

We study the impact of non-pharmaceutical policy interventions (NPIs) like “stay-at-home” orders on the spread of infectious disease. NPIs are associated with slower growth of Covid-19 cases. NPIs “spillover” into other jurisdictions. NPIs are not associated with significantly worse economic outcomes measured by job losses. Job losses have been no higher in US states that implemented “stay-at-home” during the Covid-19 pandemic than in states that did not have “stay-at-home”. All of these results demonstrate that the Covid-19 pandemic is a common economic and public health shock. The tradeoff between the economy and public health today depends strongly on what is happening elsewhere. This underscores the importance of coordinated economic and public health responses.

We use high-frequency indicators to analyze the economic impact of COVID-19 in Europe and the United States during the early phase of the pandemic. We document that European countries and U.S. states that experienced larger outbreaks also suffered larger economic losses. We also find that the heterogeneous impact of COVID-19 is mostly captured by observed changes in people’s mobility, while, so far, there is no robust evidence supporting additional impact from the adoption of non-pharmaceutical interventions. The deterioration of economic conditions preceded the introduction of these policies and a gradual recovery also started before formal reopening, highlighting the importance of voluntary social distancing, communication, and trust-building measures.

Understanding the immediate consequences of the COVID-19 pandemic on consumer behaviour is essential for informing the policy makers on the economic cost of strict measures, such as population lockdowns and business shutdowns. Yet, estimating the effect of the health shock on consumption, net of policy restrictions, is challenging because such measures affect consumer choices. South Korea is an interesting case because its policy response in the early stages of the pandemic did not involve such restrictive measures. We exploit this fact to study the consequences of the health shock on consumption. Because the intensity of the pandemic varied greatly across administrative regions, we are able to quantify the direct effect of the health shock on consumption at the epicentre of the pandemic and to compare it with that in locations initially spared from the virus. Further, we quantify...

Typical government responses to pandemics involve social distancing measures implemented to curb disease propagation. We evaluate the impact of state-mandated business closures in the context of the Covid-19 crisis in the US. Using state-level variations in the set of sectors defined as non-essential and forced to shut down, and geographic variations in industry composition, we estimate the effects of business closure decisions on firms’ market value, and on infection and death rates. We find that a 10 percentage point increase in the share of restricted labor is associated with a drop by 3 percentage points in April 2020 employment, a 1.87% drop in firms’ market value, and 0.15 and 0.011 percentage points lower Covid-19 infection and death rates, respectively. An extrapolation of these preliminary findings suggests that state-mandated business closures might have cost $700 billion...

We construct a quantitative model of an economy hit by an epidemic. People differ by age and skill, and choose occupations and whether to commute to work or work from home, to maximize their income and minimize their fear of infection. Occupations differ by wage, infection risk, and the productivity loss when working from home. By setting the model parameters to replicate the progression of COVID-19 in South Korea and the United Kingdom, we obtain three key results. First, government-imposed lock-downs may not present a clear trade-off between GDP and public health, as commonly believed, even though its immediate effect is to reduce GDP and infections by forcing people to work from home. A premature lifting of the lock-down raises GDP temporarily, but infections rise over the next months to a level at which many people choose to work from home, where they are less productive, driven...

The Scandinavian countries of Denmark, Iceland, Norway and Sweden have approached the first months of the 2020 novel coronavirus pandemic with a range of economic and health policies that have resulted in disparate outcomes. Though similar in behavioral norms and institutions, Denmark, Iceland and Norway chose a precautionary approach that formally shut down schools and businesses to protect human health, while Sweden took a Business-As-Usual (BAU) approach aimed at maintaining normal economic and social activities. Iceland and Denmark have further invested in testing, tracking and containing the disease. Economic costs of the pandemic and government fiscal and monetary interventions to reduce their impacts have been dramatic and similar across countries, while Sweden has had the most severe loss of life. Using a panel from the four countries since the beginning of the pandemic, we...

We analyse ‘stop-and-go’ containment policies which produce infection cycles as periods of tight lock-downs are followed by periods of falling infection rates, which then lead to a relaxation of containment measures, allowing cases to increase again until another lock-down is imposed. The policies followed by several European countries seem to fit this pattern. We show that ’stop-and-go’ should lead to lower medical costs than keeping infections at the midpoint between the highs and lows produced by ’stop-and-go’. Increasing the upper and reducing the lower limits of a stop-and-go policy by the same amount would lower the average medical load. But, increasing the upper and lowering the lower limit while keeping the geometric average constant would have the opposite impact. We also show that with economic costs proportional to containment, any path that brings infections...

In this paper, we examine the short-term consequences of COVID-19 and evaluate the impacts of stay-at-home orders on employment and wages in the United States. Guided by a pre-analysis plan, we document that COVID-19 increased the unemployment rate, decreased hours of work and labor force participation, especially for younger workers, non-white, not married and less-educated workers. We built four indexes (exposure to disease, proximity to coworkers, work remotely and critical workers) to study the impact of COVID-19. We find that workers that can work remotely are significantly less likely to have their labor market outcomes affected, while workers working in proximity to coworkers are more affected. The unemployment effects are significantly larger for states that implemented stay-at-home orders. Our estimates suggest that, as of early May, these policies increased unemployment by...

We use high-frequency Google search data, combined with data on the announcement dates of non-pharmaceutical interventions (NPIs) during the COVID-19 pandemic in U.S. states, to isolate the impact of NPIs on unemployment in an event-study framework. Exploiting the differential timing of the introduction of restaurant and bar limitations, non-essential business closures, stay-at-home orders, large-gatherings bans, school closures, and emergency declarations, we analyze how Google searches for claiming unemployment insurance (UI) varied from day to day and across states. We describe a set of assumptions under which proxy outcomes (e.g., Google searches) can be used to estimate the causal parameter of interest (e.g., share of UI claims caused by NPIs) when data on the outcome of interest (e.g., daily UI claims) are limited. Using this method, we quantify the share of overall growth in...