Close shot of a vaccine being drawn from a vial into a syringe. The word COVID-19 floats in the background.

Research Monitor

February 15th 2021


The Pfizer-BioNTech vaccine proves its efficacy in Israel

In Israel, the world’s leader in the COVID-19 vaccine roll-out, a new study demonstrates the efficacy of the Pfizer-BioNTech vaccine at reducing new infections of COVID-19. The study was performed by Clalit Health Services, Israel’s largest health provider and which covers more than half of Israelis, with help from a team from Harvard University.

To understand the impact of the vaccine, the study compared new COVID-19 infections among 600,000 people that had already received both doses of the Pfizer-BioNTech vaccine with a control group who had not. The vaccinated group included 430,000 people between the ages of 16 and 59 years of age and 170,000 people over 60 years of age. The study was designed so that each vaccinated individual was compared with an unvaccinated individual with a closely-linked profile, including risk of infection, risk of developing serious illness and overall health.

Among the vaccinated group, there was a 94% drop in symptomatic cases of COVID-19. The same group were 92% less likely to develop a serious case of COVID-19. The data confirm the results of Pfizer-BioNTech’s Phase III clinical trials, which suggested that the vaccine had an efficacy near 95%. The results also suggest that the vaccine becomes increasingly effective from two weeks following the administration of the second vaccine dose.

The news follows results from the Weizmann Institute of Science in Rehovot, Israel. That institute tabulates national hospitalisation data and has confirmed that the sharp decline in hospitalisation with serious cases of COVID-19 already observed amongst those 60 years or older is now being seen among those 55 and older as well.

Mask fit as important as quality

A paper published in the journal PLOS ONE investigates the impact of the fit of COVID-19 masks on the protection they offer, both for the wearer and for others.

In the year since the pandemic began, the importance of wearing a face mask to limit the spread of SARS-CoV-2––the virus that causes COVID-19––has been well-demonstrated and has become broadly accepted. However, while much attention has been paid to the efficacy of the material a mask is made from, little has been paid to the role that a mask’s fit plays in limiting the spread of the virus. As a consequence, supply chain constraints have led many manufacturers to abandon both quantitative and qualitative testing of the fit of mask designs, implementing less rigorous subjective fit tests instead.

To plug this research gap, Eugenia O’Kelly, Anmol Arora, Sophia Pirog, James Ward and John Clarkson from the University of Cambridge looked at the role of fit in different types of masks and compared the efficacy of quantitative vs. subjective fit testing. To do this, seven participants first evaluated N95 and KN95 respirators by performing a subjective fit check, according to NHS self-assessment guidelines. The same participants then underwent quantitative fit testing of five different N95 respirators, a KN95 respirator, a surgical mask and cloth masks.

The results confirm that the fit of a mask is vital to its efficacy. The N95 respirators offered the highest protection of any of the options––but only when they fit well. Unfortunately, most N95 respirators failed to adequately fit the subjects. Otherwise, KN95 respirators, surgical masks and cloth masks all performed poorly on mask fit.


A new paper published in the American National Bureau of Economic Research (NBER) Working Paper Series investigates the impact on labour supply and employment growth of the COVID-19 stimulus boost to unemployment insurance in the United States.

In March 2020, the passage of the CARES Act––the United States’ initial COVID-19 stimulus package––introduced a $600 boost to weekly unemployment insurance benefits, called Federal Pandemic Unemployment Compensation (FPUC). This led to an unprecedented increase in the benefit replacement rate––the relative proportion of benefits to previous employment earnings––from roughly 48% of earnings on average to 145%. In traditional microeconomic theory, a change of this magnitude is be expected to disincentivise labour supply––people will be less motivated to find new work, so the thought goes––or job creation, as increased benefits put upward pressure on wages.

To investigate the extent to which this is the case, Arindrajit Dube from the University of Massachusetts Amherst took advantage of the abrupt expiration of FPUC payments in July 2020 to compare employment growth with and without the measure in place. To this end, Dube used high-frequency household census data to track employment amongst non-college educated workers in the United States. Because of different average earnings levels between workers in different states––meaning different average replacement levels––this data allowed for fine-grained analysis of the impacts of different relative benefits levels on local employment growth.

The analysis found no statistically significant relationship between benefits levels and labour supply or job creation, contradicting most existing estimates of labour-supply elasticity. This result, consistent with previous studies using different data sets, suggests that the COVID-19 pandemic in effect suspended normal behavioural characteristics.


A separate paper also published in the American National Bureau of Economic Research (NBER) Working Paper Series investigates the impact of the January 6th Capitol Riots in Washington D.C. on local risk-averting behaviour in its aftermath and subsequent community level spread of COVID-19.

On January 6th, the United States Capitol was besieged by densely packed and maskless rioters protesting the certification of Joe Biden’s election as the 46th President of the United States. In the riots’ aftermath, the Director of the Centres for Disease Control and Prevention (CDC) raised concerns that the event might in retrospect constitute a COVID-19 “surge event.”

To investigate whether this fear was born out, Dhaval Dave from Bentley University, Drew McNichols from the University of California San Diego and Joseph Sabia from the University of New Hampshire used anonymised cell-phone data to track the movements of rioters through Washington D.C. and then back to their respective counties. This information was then compared with local COVID-19 infection rates in the month following the Capitol siege.

The results were somewhat surprising. In the month following the siege, there was no obvious increase in COVID-19 infections in the District of Columbia. This was likely because of the virtual lockdown of Washington D.C. in the weeks leading up to President Joe Biden’s January 20th inauguration. The District of Columbia is also an overwhelmingly Democratic area, and so few of the Capitol rioters are thought to have been locals.

However, in the counties which supplied the largest inflow of rioters from elsewhere in the United States there was a significant growth in the rate of daily COVID-19 infections in the month following the Capitol siege. This suggests that the Capital riots were in fact a COVID-19 surge event, just not in Washington D.C.

We now publish a weekly newsletter to inform friends and supporters of the Global Commission’s progress and to provide updates when new content is published. Please sign up here:

Previous Research Monitor