IN 1958 Paul Newman and Joan Collins starred in “Rally Round The Flag, Boys!”, a cold-war comedy about a small American town that gets chosen as the site of a new military base to house guided missiles. Twelve years later, John Mueller, a political scientist, adopted the same phrase in an academic-journal article analysing what factors caused presidents’ approval ratings to rise and fall. According to Mr Mueller, presidential popularity is influenced by four factors: how long a president has served, the rate of economic growth (or contraction), the presence of a war, and other significant events in foreign affairs. The last of these, he said, can cause a “rally-round-the-flag” effect on presidents’ ratings.
With America now supporting Ukraine and punishing Russia for its invasion, is Joe Biden due for his own rally? If so, how large a bounce should be expected—and is he on track to get it?
To answer these questions, The Economist compiled a dataset of every presidential-approval poll conducted since 1943. We used it to calculate a rolling average of approval polls for each leader. Our model uses statistical techniques from machine-learning to control for factors which could affect the polling average. Among other things, we let our computers determine how much weight to give recent polls relative to older ones, and to calculate whether aggregates for modern presidents should be computed differently from those for presidents in the more distant past; a dearth of polling on Franklin D. Roosevelt and Harry Truman, the first presidents to be subject to approval surveys, for example, can add too much noise to a model that attempts to capture the popularity of recent leaders, such as Mr Biden or Donald Trump.
The resulting dataset quantifies each president’s approval rating on each day they served in office. We used it to measure the size of “rally-round-the-flag” effects after they became embroiled in foreign conflicts or faced significant threats to Americans at home. For example, in the month after George H.W. Bush announced the Gulf war against Iraq in 1991, his rating soared by more than 20 percentage points, according to our average (see chart). Jimmy Carter’s popularity increased by 22 points in the two months following the onset of the Iran hostage crisis, when militants held over 50 Americans in Tehran in the aftermath of the Islamic Revolution.
Over nine historical rally-round-the-flag events identified by The Economist, the average president’s rating has increased by just over six points in the six subsequent weeks. But there is a large amount of variation within that. Assuming Mr Biden’s response to Vladimir Putin’s invasion of Ukraine qualifies as such an event, his expected bounce would range from three to ten points within the 80% confidence interval of predicted changes.
But political polarisation may yet dampen any real jump in Mr Biden’s ratings. As the share of Americans who identify with one party over the other has increased, impacts on approval ratings from foreign crises, and even domestic politics, have decreased in magnitude. After the military raid that killed Osama bin Laden, for example, Barack Obama’s popularity rose by only 6-7 points, compared to the 20- and 30-point boosts for other leaders and events. And when the World Health Organisation announced it was officially calling covid-19 a pandemic in 2020, Mr Trump’s approval rating grew by only 3-4 points.
Mr Biden would welcome any improvement he can get. His approval rating today is 43%. That is an increase of roughly two points since February 23rd, the day before Mr Putin launched his war. That could be early evidence of a bounce—but it is not yet fully outside the range of random variations in polls. Only one poll, from Marist College, conducted for NPR and PBS NewsHour, has shown a sizeable bounce for the president. It is also possible that a few good news cycles for him have caused Democrats to start enthusiastically responding to canvassers’ solicitations, artificially boosting the apparent share of Americans who like Mr Biden. Such “differential partisan non-response” is one reason polls overestimated support for Mr Biden in the election in 2020.
Still, online polling conducted for The Economist by YouGov suggests the president has done what the people want on Ukraine. Both Democrats and Republicans support America’s various sanctions on Russia, such as those on large companies, billionaires and state resources. Majorities of both groups also support shipping weapons to Ukrainian fighters. (Many also go further than the administration has done and support admitting Ukraine into NATO.) But just because Americans have rallied against Mr Putin does not mean they will stand by Mr Biden. Historically, most rally-round-the-flag effects have significantly faded in about four months. With America’s mid-term congressional elections still eight months away, a surer plan for political success would be to tackle inflation, bring in popular policies and keep the number of new covid-19 cases down.
Our recent coverage of the Ukraine crisis can be found here