04. July 2020
COVID-19 and the Rally Round the Flag Effect
Alternative title: The Rally Rona Flag Effect
The purpose of this post is to expand on this interesting tweet I saw: https://twitter.com/leonardocarella/status/1243648800872443904
What is the “rally round the flag effect”?
In the US, it’s basically an event that draws Americans together to support the President. John E. Mueller said such events must 1) be international, 2) involve the United States and the President, and 3) is specific, dramatic, and sharply focused in his 1970 paper Presidential Popularity from Truman to Johnson.
He said that domestic events, like riots, scandals, or strikes, don’t count because they exacerbate internal divisions. Events that don’t involve the US don’t count either because they appear irrelevant to the public. He said that events that are gradual also don’t count because they don’t command public attention and their impacts on public attitudes are likely diffused.
Examples of this effect in US history include the Cuban Missile Crisis and the Bay of Pigs. Notice that the first event had a “good” outcome and the second had a “bad” outcome. Mueller found that the public reacted in about the same way in terms of their approval for the President.
Mueller also found that time after the rally round the flag event and approval rating of the President was negatively correlated, i.e. as time increases, approval rating decreases. This happens even for successful events. Marc J. Hetherington and Michael Nelson saw that happen for the Cuban Missile Crisis, Operation Desert Storm, and the September 11 Attacks in their 2003 paper Anatomy of a Rally Effect: George W. Bush and the War on Terrorism. Page 3 of the PDF I linked has a nice graph that shows approval ratings dropping back down to pre-event levels after a time.
Does this effect hold for other countries?
Most of the research of this effect has been done on the United States, but there is evidence that this effect occurs in other countries in North America and Europe too.
Brian Lai and Dan Reiter suggested that in the United Kingdom, rallies are more likely to happen when there is a direct and intense threat to national interests in their 2005 paper Rally ‘Round the Union Jack? Public Opinion and the Use of Force in the United Kingdom, 1948–2001.
Anna Yudina noted that President Vladimir Putin’s popularity rating went up after the 2014 Invasion of Crimea in her 2019 paper The Rally ‘Round the Flag Effect In Russia: How an International Crisis Turns Regime Opponents Into Regime Supporters (this is a download link, by the way). She found similarities between Russia and the United States in the types of people who were most likely to rally, like being nationalistic. However, she saw the rally effect was more intense and more prolonged in Russia than the US.
Dieuwertje Kuijpers looked at 10 advanced liberal democracies (Canada, Denmark, Germany, Italy, the Netherlands, Norway, Spain, Sweden, United Kingdom, and the United States) and found that the non-US countries experienced rallying effects for “nonpeacekeeping missions” and subsequent erosion in support as the mission dragged on in her 2019 paper Rally around All the Flags: The Effect of Military Casualties on Incumbent Popularity in Ten Countries 1990–2014.
Does the COVID-19 pandemic count as rallying event?
If we go by Mueller’s definition, I would say yes. If you look at this Johns Hopkins map and the consider that some people call it the Chinese coronavirus (libs owned), it’s obviously international. It involves countries and their heads of states, as those heads of states have to determine policy to combat the outbreak. And it is an intense and sharp event because the large number of cases and how exponentially the virus spread.
Approval ratings of heads of states and COVID-19
I will compile a table of approval ratings because disapproval numbers are sometimes not given, that’s how the original tweet did it, and Mueller found that the approval and disapproval ratings had a strong negative correlation anyway.
You can also look at this poll conducted by Morning Consult, which looks at various world leaders net approval ratings and how they’ve been affected by the pandemic, which I found after I spent 2 hours looking for and translating polls of most of the leaders below 🙃. More data never hurts though!
Amateur analysis
Of everyone listed, only Abe, Bolsonaro, and Obrador saw their approval ratings not improve from just before WHO declared a pandemic to now. There’s no margin of error given for those three, but, gun to my head, I would guess it’s around +-3% since Morning Consult says they’ve conducted “between an average of 315 and 616 daily interviews per country”. Even if it’s less, all three are very close to a positive approval change.
Bolsonaro and Obrador are still above 50% approval, so that’s nice for them.
If you look Abe in the Morning Consult graph, you can see that his net approval around February 26, 2020 is -30. That’s the day before Abe asked to close down all schools because Japan started seeing cases earlier than all the other countries on the list. If that is used as the starting point, Abe’s net approval went up to -23 (his raw approval numbers likely went up too), which looks a rally round the flag effect.
It does look like Trump is experiencing a rally round the flag effect on his approval rating, to my dismay. Even FiveThirtyEight’s adjusted approval rating average confirms this (side note: this tool is much, much more useful than looking at single polls or even simply averaging out all polls. Every country should have a similar one).
But perhaps I shouldn’t be dismayed. For one, recall that rally round the flag can happen to both “good” and “bad” events. Trump has fucked up and continues to fuck up the response to this pandemic. I think Trump’s improved approval ratings doesn’t necessarily mean that the American public won’t see what we’re seeing, if they haven’t already.
Secondly, the thing about rally round the flag is that approval ratings eventually go down, unless you’re Putin. I’m sure Trump would love to be like Putin, but he is not because we live in a liberal democracy, thank God. I’m fairly confident that Trump’s approval rating will fall back down to the low-mid 40s.
Third, Trump’s rally round the flag effect (among his many other things) is quite small compared to that of other world leaders. I think that shows that the American public is extremely polarized on Trump and/or that people who don’t like him really, really don’t like him.
In addition, if you look at question 6 of this Monmouth poll, you’ll see that respondents' approval rating of their state governors is 72%. Perhaps Americans are rallying around their respective state flags rather than the American flag.
Fifth, averaging national polls after the WHO declared COVID-19 to have reached pandemic status, Joe Biden, who almost certainly will be the Democratic nominee, leads Donald Trump 47.28 to 41.71. People really hate Trump’s guts, it seems like.