The wearing of face masks by the general public has been a heated policy issue during the COVID-19 pandemic, with national health authorities and the World Health Organization giving inconsistent or contradictory recommendations over time, ranging from ‘not recommended’ to ‘mandatory’. ‘Conspiracy theories’ and misinformation surrounding mask wearing abound in social or other media, fuelled by some people’s perception that mask mandates constitute restrictions on individual freedoms. Protests or altercations related to mask requirements are reported frequently.
Even though there is now significant agreement in the medical literature that respiratory transmission of COVID-19 is the dominant vector (Meyerowitz et al. 2020), and many clinical studies have shown that face masks reduce the transmission of COVID-19 and similar illnesses (Chu et al. 2020, Prather et al. 2020, Leung et al. 2020, Greenhalgh et al. 2020), a face mask mandate may not be effective in practice if it fails to increase the prevalence of mask wearing (compliance), or if it leads to increased contact due to a false sense of security.
An added challenge is disentangling the impact of mask mandates from that of other policies, behaviour, and other factors (Chernozhukov et al. 2020, Mitze et al. 2020). Given the absence of large-scale randomised controlled trials (Howard et al. 2020), quantitative observational studies are essential for informing health policy and public opinion.
We evaluate the impact of face mask mandates and other public policies on COVID-19 cases in Canada (Karaivanov et al. 2020). The Canadian data have the advantage of allowing two complementary approaches to identify the policy effects and distinguish them from other factors.
We first observe that indoor face mask mandates were implemented across 34 districts (‘public health units’) in Ontario at different dates, staggered over two months. This is because the provincial authorities allowed regional health agencies or municipalities to decide on mask mandates independently. Figure 1 displays the gradual introduction of mask mandates across the public health units in Ontario – Canada’s most populous province with roughly 40% (15 million) of the country’s population – during our study period.
Figure 1 Timing of mask mandates in Ontario
Notes: There are a total of 34 public health units in Ontario. Mandatory indoor masks were introduced between June 12 and August 17. See Karaivanov et al. (2020) for the exact date for each public health unit and more details.
We compare public health units that adopted mask mandates early to those that adopted mask mandates later and estimate the effect of mask mandates on the spread of COVID-19. The main advantage is that we can use differences in the time of mask-mandate implementation in relatively small areas, holding potentially confounding province-level policies or events constant.
We also evaluate the impact of mask mandates and non-pharmaceutical interventions in Canada as a whole, using variation in the policy timing and intensity over time in the country’s ten provinces. The advantage of this second approach is that we can analyse the potential impact of not just mask mandates, but also other interventions that vary little or not at all within Ontario (e.g. regulations on businesses and gatherings, school closures, travel, and long-term care). In addition, the province-level data allow us to study both the closing-down period (March–April) and the gradual reopening period (May–August), providing evidence from both the imposition and the relaxation of policies.
Face mask mandates and COVID-19 cases in Canada
Figure 2 displays the average weekly case growth in Ontario public health units with mask mandates (blue) or without mask mandates (red). On average, the public health units that had a mask mandate two weeks prior exhibit lower case growth than the public health units without a mask mandate two weeks prior. The two-week lag represents the approximate delay between policy implementation (with the resulting changes in mask wearing or other behaviour) and its effect on new registered cases, accounting for the COVID-19 incubation period, time between symptom onset and testing, and test turnaround time.
Figure 2 Mask mandates and case growth in Ontario (raw data)
Notes: The figure plots the average log weekly case growth ∆log(Weekly Cases) in the public health units with mask mandate (blue) vs without (red) mask mandate 14 days prior.
Figure 2 plots the raw data, so it omits any other factors that could potentially cause differences in the observed growth rates of COVID-19 cases. We formally estimate the impact of mask mandates by controlling for other policy measures, changes in other behaviour (proxied by Google mobility location data), information (previous case totals and case growth), the growth rate of tests, and public health unit characteristics (fixed effects).
Our estimates, from a range of empirical specifications, imply that two weeks after implementation, mask mandates are associated, on average, with a reduction of 29 to 37 log points in the weekly case growth rate, which can be interpreted as a 25%–31% weekly reduction in new cases relative to the no-mandate trend.
As an example, suppose weekly cases in weeks 1, 2, and 3 are 100, 150, and 225, respectively (50% weekly case growth). Our estimates imply that holding all else equal, a mask mandate implemented at the beginning of week 1 would reduce week 3 cases from 225 to around 160, with further reductions relative to trend in the next several weeks. We find corroborating evidence in the province-level analysis, with an estimated 36%–46% weekly reduction, depending on the empirical specification (see Karaivanov et al. 2020 for details).
The effectiveness of any public policy critically depends on how it affects peoples’ behaviour. We used survey data on mask usage in Canada to examine whether mask mandates actually increase mask use. The survey asked how often people wear a face mask outside their homes. Figure 3 displays the differences in self-reported mask usage between Canadian provinces with mask mandates and provinces without mask mandates, ranging from six weeks before to five weeks after the introduction of the mask mandate.
Figure 3 Mask mandates and mask usage
Notes: The data source is YouGov COVID-19 Public Monitor. The outcome is a binary variable taking value 1 if the respondent answered ‘Always’ to the question ‘Thinking about the last 7 days, how often have you worn a face mask outside your home?’ ‘0’ is the mandate implementation date and ‘-1’ is the reference point.
We find an increase of roughly 20 percentage points in mask usage, on average, as soon as a mask mandate is implemented, and this effect persists over the following weeks. These results show that mask mandates enjoy broad compliance in Canada and help confirm the mechanism for the significant impact mask mandates have on the spread of COVID-19.
In Figure 4, we perform two hypothetical policy simulation exercises (so-called counterfactuals) to further illustrate and quantify the estimated impact of face masks in Canada. The first simulation assumes that, contrary to reality, face masks were required in all Canadian provinces as of 7 July, the date on which Ontario’s largest public health unit, Toronto, and Canada’s capital, Ottawa, adopted mask mandates. The second counterfactual simulation considers the opposite scenario, assuming instead that mask mandates were not adopted in any Canadian province by mid-August (see Karaivanov et al. 2020 for details on how the graphs were created).
The left panel of Figure 4 shows that nationwide adoption of mask mandates in early July is predicted to reduce total cases per week in the country by 700 cases, on average, as of 13 August. The right panel of Figure 4 shows that the failure to adopt any mask mandates would have instead led to a large increase (more than 1,500 per week) in new cases Canada-wide by mid-August.
Figure 4 Mask policy simulations
Notes: The left panel assumes that mask mandates were adopted in all provinces on 7 July (the adoption date in Toronto and Ottawa). The right panel assumes that mask mandates were not adopted in any province. We use the mask estimate (-0.376) from column (4) of Table 1 in Karaivanov et al. (2020). The counterfactual mean value (the green or red solid lines) and confidence bands (shaded areas) are displayed.
The role of other policy measures
We also evaluate the impact of other non-pharmaceutical interventions on COVID-19 case growth in Canada, including regulations on businesses and gatherings, school closures, travel and self-isolation, and long-term care homes. We exploit the differences in implementation date and intensity over time of these policies in the country’s ten provinces. We do this for the whole data sample (March–August), as well as separately for the ‘closing’ period (March–April) and the ‘reopening’ period (May–August).
We find that the most stringent restrictions on businesses and gatherings observed in the data are associated with a weekly decrease of 48%–57% in new cases, relative to the trend in the absence of restrictions. The business/gathering regulations estimates are, however, noisier than our estimates for mask mandates and not always statistically significant. School closures and travel restrictions are associated with a large decrease in weekly case growth in the closing period.
Mask mandates were introduced in Canada at a time when most other policy measures were relaxed, as part of the economy’s reopening. We find that reduced restrictions on businesses or gatherings (including retail, restaurants, and bars) in the reopening period are positively associated with subsequent COVID-19 case growth – a factor that, together with schools reopening in September, can offset in magnitude and obscure the health benefits of mask mandates.
Mask mandates also work in Germany and the US
Our results for Canada are consistent with two other recent observational studies on the effects of mask mandates. Chernozhukov et al. (2020) use differences in the timing of mask mandates for employees in public-facing businesses across US states and find that these mandates are associated with a reduction of 9–10 percentage points in the weekly growth rate of COVID-19 cases. Our estimates are larger, partly because we analyse universal mandatory mask wearing. Similarly, Mitze et al. (2020) study the city of Jena and regions in Germany that adopted mask mandates before the rest of the country and estimate that requiring the public to wear face masks reduced the daily growth rate of COVID-19 cases by about 40%.
Our results suggest that mandating wearing face masks in public can be an effective policy to slow the spread of COVID-19. Some may cite the currently growing cases in Europe, certain US states and Canadian provinces with mask mandates as ‘evidence’ against this conclusion. It is important to remember that the reduction effect we estimate is relative to the absence of mask mandates and not absolute: although we find a substantial effect of requiring masks in indoor public spaces, it may be far from enough to prevent an increase in new infections on its own.
Nevertheless, a mask mandate can be an important element of a suite of policies that keeps COVID-19 transmission at a manageable level, especially given the policy’s relatively low cost to the economy.
Authors’ note: The data used in this research are publicly available at github.com/C19-SFU-Econ.
Chernozhukov, V, H Kasahara and P Schrimpf (2020), “Causal impact of masks, policies, behavior on early COVID-19 pandemic in the US”, medRxiv pre-print.
Chu, D, E Akl, S Duda, K Solo, S Yaacoub and H Schünemann (2020), “Physical distancing, face masks, and eye protection to prevent person-to-person transmission of SARS-CoV-2 and COVID-19: A systematic review and meta-analysis”, Lancet 395: 1973–87.
Greenhalgh, T, M B Schmid, T Czypionka, D Bassler and L Gruer (2020), “Face masks for the public during the COVID-19 crisis”, BMJ 369: m1435.
Howard, J, A Huang, Z Li, Z Tufekci, V Zdimal, H-M van der Westhuizen, A von Delft, A Price, L Fridman, L-H Tang, V Tang, G L Watson, C E Bax, R Shaikh, F Questier, D Hernandez, L F Chu, C M Ramirez and A W Rimoin (2020), “Face masks against COVID-19: An evidence review”, Preprints 2020.
Karaivanov, A, S Lu, H Shigeoka, C Chen and S Pamplona (2020), “Face masks, public policies and slowing the spread of COVID-19: Evidence from Canada”, medRxiv pre-print
Leung, N, D Chu, E Shiu, K Chan, J McDevitt, B Hau, H Yen, Y Li, D Ip, J Peiris, W Seto, G Leung, D Milton and B Cowling (2020), “Respiratory virus shedding in exhaled breath and efficacy of face masks”, Nature Medicine 26: 676–80.
Meyerowitz, E, A Richterman, R Gandhi and P Sax (2020), “Transmission of SARS-CoV-2: A review of viral, host, and environmental factors”, Annals of Internal Medicine, Reviews.
Mitze, T, R Kosfeld, J Rode and K Wälde (2020), “Face masks considerably reduce COVID-19 cases in Germany”, medRxiv pre-print.
Prather, K, C Wang, and R Schooley (2020), “Reducing transmission of SARS-CoV-2”, Science 368(6498): 1422–4.