One-way or a two-way interaction model?

-

Why is it interesting?

The interaction between the epidemic and the macro economy can be modelled in different ways. This comparison shows the main differences between a one-way and a two-way interaction epidemic macro model. In the one-way interaction model (Atkeson 2020), the epidemic has an impact on the economy but agents cannot influence the evolution of the epidemic through their decisions. In contrast, the Eichenbaum, Rebelo and Trabandt (2021) model features a two-way interaction between the epidemic and the macro economy.

What to do on the Epi-MMB?

  • Models: ERT_21_Epi, A_20

  • Shocks: Initial infections of 0.1% of the population

  • Variables: Susceptibles, Infected, Recovered


What is interesting?

The impulse responses to an initial infections shock where 0.1 percent of the population gets infected significantly differ for the two models. In the ERT_21_Epi model, agents decide to cut back consumption and hours worked to mitigate the risk of becoming infected. Omitting this motive in the model leads to a much faster and more severe spread of the virus. The infection curve is significantly flatter in the two-way interaction model, suggesting that the epidemic will last longer but will also be milder. The number of recovered and deceased people* is consequently also much higher in the A_20 model with almost 90 percent of the population, compared to approximately 54 percent in ERT_21_Epi.

Most models on the Epi-MMB implement a two-way interaction between the epidemic and the macro economy, as this provides greater scope for additional research, for example, regarding optimal policy responses.


*The A_20 model only reports the sum of recovered and deceased people, which can be seen in the chart. Since the number of deaths is relatively small in the ERT_21_Epi model, it does not make much of a difference for the illustration.


Reference: Eichenbaum, Rebelo and Trabandt (2021), Atkeson (2020)

Comments

Post a Comment

Popular posts from this blog

Competitive equilibrium versus social planner?

-
Image
What are differences between competitive equilibrium and social planner solutions in pandemic macroeconomic models? Why is it interesting? The comparison shows the results of the pandemic shock for commonly defined macroeconomic aggregates and epidemic variables between a model with competitive equilibrium (MY_21) and a model with social planner solutions (BCG_21). What to do on the EPI-MMB? Models: MY_21, BCG_21 Shocks: Model specific shock: initial infections of < 0.001% of the population for MY_21 and initial infections of 0.1% the population for BCG_21 Variables: Consumption, Investment, Labour, Output, Susceptibles, Infected, Recovered, Deaths What is interesting? Notably, MY_21 predicts a deeper recession and a milder pandemic spread than BCG_21. One major contributing factor is the difference in equilibrium types between the two models. In BCG_21, individual agents make economic decisions while taking into account different infection rates and evolutions of the pandemic acros
Read more

Heterogeneity in SIR-macro?

-
Image
 Eichenbaum et al. (2020) developed a “SIR-macro” model which is often used as a basis in the recent highly developed literature on covid pandemic. The model embeds a simple economy with the epidemiological SIR part based on Kermack and McKendrick (1927). In reality, the pandemic spread with heterogeneity - including age, region and occupation. Appropriate policy responses should take these heterogeneities into consideration. What to do on the epi-mmb? Model: ERT_21_NK, GP_20, KUX_20, CF_20 Shocks: Initial infections of 0.1% of the population Variables: Consumption, Labour, Output, Susceptibles, Infected, Deaths These models are based on ERT_21_NK, but slightly differ in several dimensions. GP_20 incorporates age-specific socio-economic interactions to study the role of demographic factors in the COVID-19 epidemic evolution, KUX_20 features heterogeneous production sectors of consumption goods, and CF_20 added multiple locations and travel cost to the model to study the optimal impleme
Read more