introduction


21 presidents and prime ministers are among the millions of people who have tested positive for COVID-19. The pandemic’s onset boosted support for many world leaders, consistent with a “rally-’round-the flag” effect caused by national unity in the face of a common threat. But does personally contracting the novel coronavirus affect leaders’ popularity? If it raises national threat and patriotism, it could make leaders more popular -- a personal “rally-’round-the flag” effect. If, however, citizens take an infection as a signal of bad leadership, it may harm leaders’ public standing.


Our analysis of the high-profile cases of UK Prime Minister Boris Johnson, Brazilian President Jair Bolsonaro, and US President Donald Trump reveal modest popularity bumps in weeks that these leaders contracted COVID-19. However, such increases may be indistinguishable from the pandemic’s broader rally effects. To rule out the possibility that this conclusion reflects the shared characteristics among these cases, we look at three cases that differ in key ways from the first three: Russian Prime Minister Mikhail Mishustin, Polish President Andrzej Duda, and French President Emmanuel Macron. Similar patterns emerge even though polling has been sparser in these cases due to smaller industries, lack of time elapse, or both.


Our study relies on weekly estimates of net approval from the Executive Approval Project and health outputs from Our World in Data and Coronavirus (COVID-19) in the UK. Combining data across multiple firms helps distinguish changes in leader popularity from house effects and survey-related errors; weekly analysis reduces day-to-day noise.

boris johnson, prime minister of the united kingdom

Boris Johnson became the first major world leader to test positive for COVID-19 on March 27, 2020. He began the year with negative net approval -- a greater percentage of UK citizens disapproved of his work as PM than approved of it. Our graphic reflects the common wisdom that COVID-19’s arrival in early March 2020 produced a “rally-'round-the flag” effect. Did Johnson also receive a personal COVID rally?

Rallies-’round-the-flag are fueled by swelling patriotism, the opposition’s forbearance of criticism, or both in the face of national threats. COVID-19’s threat, in the form of new cases and deaths, grew immensely throughout March as did Johnson’s popularity. All three of these factors reached their first-wave peaks as Johnson exited the hospital. Yet we see evidence of a personal COVID rally. Johnson’s jump in net approval the week after testing positive remains even after controlling for these public health outcomes, according to regression analyses. How much of this jump is a personal COVID rally is unclear. Several events preceding Johnson’s diagnosis heightened national threat and patriotism -- key rally-’round-the-flag ingredients.

Eighteen presidents and prime ministers are among the millions of people who have tested positive for COVID-19. The pandemic’s onset boosted support for many world leaders, consistent with a “rally-’round-the flag” effect caused by national unity in the face of a common threat. But does personally contracting the novel coronavirus affect leaders’ popularity? If it raises national threat and patriotism, it could make leaders more popular -- a personal “rally-’round-the flag” effect. If, however, citizens take an infection as a signal of bad leadership, it may harm leaders’ public standing.


Our analysis of the high-profile cases of UK Prime Minister Boris Johnson, Brazilian President Jair Bolsonaro, and US President Donald Trump reveal modest popularity bumps in weeks that these leaders contracted COVID-19. However, such increases may be indistinguishable from the pandemic’s broader rally effects. To rule out the possibility that this conclusion reflects the shared characteristics among these cases, we look at three cases that differ in key ways from the first three: Russian Prime Minister Mikhail Mishustin, Polish President Andrzej Duda, and French President Emmanuel Macron. Similar patterns emerge even though polling has been sparser in these cases due to smaller industries, lack of time elapse, or both.


Our study relies on weekly estimates of net approval from the Executive Approval Project and health outputs from Our World in Data and Coronavirus (COVID-19) in the UK. Combining data across multiple firms helps distinguish changes in leader popularity from house effects and survey-related errors; weekly analysis reduces day-to-day noise.

mikhail mishustin, prime minister of russia

Vladimir Putin nominated Mikhail Mishustin to replace Dmitry Medvedev as Prime Minister on January 16, 2020, roughly two weeks before Russia saw its first case of COVID-19. Putin did not receive a traditional rally-’round-the-flag with the onset of COVID, according to Levada Center data. In fact, his approval fell to historic lows and only rose as Russia got the first wave of COVID under control.

Mishustin, was tasked with leading the government’s response. While more publicly respectful of the coronavirus than Putin, his policies to confront the virus is not aggressive, perhaps seeking to avoid panic. Nevertheless, Russians briefly rallied behind him, as our graphic shows. What of a personal rally?

Net approval of Mishustin ticks up 2 points immediately after testing positive on April 30. At the same time, new cases rose 31% and new deaths were up 24%. Regression analyses reveal that this hint of a personal COVID rally evaporates in the face of controls for health outcomes. So, like Johnson’s personal COVID rally, Mishustin’s personal rally was small and short-lived if it existed at all.


jair bolsonaro, president of brazil

After three positive tests for COVID-19 -- July 7, 16, and 21 -- Jair Bolsonaro announced a negative test on July 25. He was quite popular before the coronavirus reached Brazil in late March. Yet Bolsonaro’s popularity then nosedived before the advent of generous direct emergency economic aid in April raised his public support, according to our modeling. This reprieve was cut short by a political crisis ending in the resignations of popular justice and health ministers.

Unlike many leaders, Bolsonaro did not enjoy a general COVID rally. Indeed, new COVID-19 cases and deaths neither hurt or helped Bolsonaro’s public support in this period. Perhaps, by stubbornly downplaying the seriousness of the virus, Bolsonaro removed the key condition for the public to rally ‘round the flag: a stark national threat.


What about a personal COVID rally? Bolsonaro’s popularity reached a 2-month high the week after he contracted the virus. But our graphic shows this formed part of a 5-week upward trend, which halted at that point and only resumed after Bolsonaro tested negative two weeks later. Statistical analyses that control, respectively, for new cases, new deaths, and the stringency of COVID policy reject the notion that Bolsonaro’s coronavirus infection rallied the public to his side.

As with Boris Johnson and Mikhail Mishustin, any effects of contracting COVID-19 on Jair Bolsonaro’s popularity were likely limited. At the same time, and similar to Johnson and Mishustin, his viral infection did no lasting damage to Bolsonaro’s mass popularity.

donald trump, president of the united states of america

Donald Trump became the ninth sitting president or prime minister to announce testing positive for COVID-19 on October 2, 2020. It had no discernible impact on his popularity.

Public support for Donald Trump in early 2020 was fairly typical of his support throughout his presidency. By our estimates, roughly 14% more Americans disapproved of how he handled his job than approved. Growing angst over the coronavirus -- embodied in case and death counts -- moved net approval to highs not seen since he took office in January 2017. Yet compared to his counterparts around the globe, Trump’s general COVID-19 rally was rather weak.


Visible and statistical inspection of the data turn up no evidence that Trump received a personal COVID-contraction rally. Our figure shows that weekly estimates of net approval barely flinch during this time. Daily estimates from FiveThirtyEight and Morning Consult tell a similar story. Any movement observed in the immediate aftermath of Trump contracting COVID was certainly minimal and probably zero.

andrzej duda, president of poland

Andrzej Duda was re-elected amidst Poland’s first wave of the coronavirus pandemic. Directly ahead of June’s first-round election Duda became the first foreign leader to visit Donald Trump since most of Europe issued lockdown orders. Battling razor-thin polling margins ahead of a second-round presidential election in July, Duda campaigned against mandatory COVID-19 vaccines. On October 24, 2020 he contracted the coronavirus.

Available polling data in Poland is currently too sparse to get a precise weekly estimate of a personal COVID rally. However, polling from TNS Kantar and gated reports from CBOS show Duda’s net approval falling from October to November. While Duda is the only leader to lose support after testing positive for COVID, we caution against attributing this to his infection. It could very well reflect the social eruption following the passage of Duda’s strict abortion law on October 22--two days prior to his positive COVID diagnosis. A more confident conclusion requires more data and time.

emmanuel macron, president of france

Though Emmanuel Macron arguably showed more respect for the coronavirus than most of the leaders we have studied, he tested positive for it on December 17, 2020. Initially sluggish in his response, by June Macron had beaten back France’s first-wave of the coronavirus.

As with Poland’s Duda, we require more time and data to estimate and analyze the effects on Macron’s popularity with confidence. Daily readings from Morning Consult indicate a robust generally COVID rally and modest personal COVID bounce. In the six days between reporting his diagnosis to December 23, Macron’s net approval improved 6 points from -31 to -25. Without proper controls we cannot say for sure whether this can be attributed to the event of Macron’s infection, to something else, or whether it is simply noise. Regardless, it fits the pattern of personal COVID rallies tending to be small if they happen at all.

why personal covid-19 rallies are likely to always be small

Brazil and Poland are highly polarized both politically and socially. Political polarization data are missing for the United States but it is assuredly high, suggesting it would resemble Brazil and Poland. Yet UK, Russia, and France are not terribly polarized on either score. Since our six cases greatly differ in their degree of societal and political polarization, they are unlikely culprits for the pattern of unimpressive personal COVID rallies we observe.

We can also rule out populist rhetoric as the driving force. Macron’s personal COVID rally was also weak at best even though his La République En Marche! scores significantly lower than any of the other leaders’ parties. Andrzej Duda has been politically independent since 2015, but his erstwhile Law and Justice (PiS) party, with whom he governs, outscores all parties under analysis for populist rhetoric. Bolsonaro’s personalist party ranks second, followed by Trump’s Republican Party and Johnson’s Conservatives. Mishustin, a technocrat, is not a member of Putin’s United Russia--a party whose strongly populist rhetoric is roughly on par with Johnson’s Conservatives.

These six cases also vary qualitatively in terms of regime type and constitutional structure. The UK is a constitutional monarchy with a parliamentary government. Brazil and the USA combine presidentialism and federalism. Russia is nominally a federalism but largely functions in a unitary manner and is firmly authoritarian; Poland is unitary and centralized; France is a unitary and decentralized state. Russia, Poland, and France feature dual-executive or semi-presidential systems. Within this form of government, Mishustin is an appointed head of government and Duda and Macron are directly elected heads of state.

Given the great cross-case variation in these potential explanatory factors, meager personal COVID rallies may well be expected in most circumstances. Thus we can conclude with some confidence that the public is unlikely to rally behind leaders who contract COVID-19. If they do, the rally is likely to be small and short-lived. Johnson’s relatively strong personal rally could be explained by many factors, including the fact that it occurred early in the pandemic. Rarely should we expect public support to fall when a leader becomes infected. Only Poland displayed this pattern and we lack the proper controls to understand whether this was due to Duda contracting the coronavirus or to broader health outcomes. While we await more evidence to bolster our conclusions, leaders who contract COVID do not appear to witness a meaningful change in how citizens evaluate them.