An empirical analysis of COVID-19 response: comparison of US with the G7

ABSTRACT

We compare the US policy response to COVID-19 with its G7 counterparts between March and September 2020. The G7 countries, while economically and ideologically aligned, have instituted vastly different policies to mitigate the spread of the disease with varying degrees of compliance. To quantify the effect of policy responses on the spread of infections, we estimate beta for each country which is the slope coefficient of daily new cases in each country regressed against world new cases. First, we test for structural breaks in daily data for world new cases using the Bai Perron method which endogenously determines break points. We obtain five break dates that allow us to divide the time period into six windows and estimate betas separately for each window. Next, we rank the G7 countries based on their beta values for each window. Our empirical findings suggest that countries that eased their lockdown measures moderately while enforcing nationwide mask mandate and comprehensive contact tracing generally performed better in mitigating the spread of new infections. Furthermore, countries with higher degree of compliance saw improvement in their rankings. US was ranked mostly in the bottom half of the G7 group but not always the worst.

KEYWORDS:

• COVID-19
• structural break
• G7
• stringency index

JEL CLASSIFICATION:

• C22
• H11
• I18

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1. The Gauss code for Bai-Perron test is downloaded from the website of P. Perron. Codes are available upon request.

2. Bai and Perron methodology also provide a partial structural break model where coefficients of a selected number of variables are allowed to change keeping some coefficients same throughout the sample period (See Bai and Perron (1998) for details.)

3. There is compelling scientific evidence pointing to a faster spread of the B.1.1.7 variant compared to the original SARS-CoV-2. We decided to stay away from combining different variants that will impact our calculation of beta values.

4. Note that the number of confirmed new cases is always lower than actual number of new cases on any given day primarily due to limited testing.

5. The world data is based on 207 countries on this site.

6. There is a time lag between policy implementation and its outcome of typically two weeks.

7. Coronavirus Cases Surge in France as People Return to Schools, Offices – WSJ.

8. Respondents are asked if they have worn a face mask outside their home. The response options ranged from ‘Always’ or ‘Frequently’ to ‘Not at all’ (i.e. complete non-compliance). https://coviddatahub.com