On 25 August 2017, one day before Hurricane Harvey struck Texas, President Trump pardoned a former sheriff accused of racial profiling and issued a ban against transgender soldiers in the military. On 14 June 2018, the day of the inauguration of the 2018 FIFA World Cup that Russia was hosting, the Russian government announced a rise in the retirement age and an increase in value added tax. On 13 July 1994, the Italian government of Silvio Berlusconi passed an emergency decree that freed hundreds of politicians with pending corruption charges – on the day Italy qualified for the final of the FIFA World Cup.
These anecdotes raise a suspicion: could these apparent coincidences be due to strategic behaviour? In other words, do politicians intentionally time controversial policies to coincide with important events that distract the media and the public, so as to minimize negative publicity? While it has been documented that corporate announcements (DellaVigna and Pollet 2009) and military operations (Durante and Zhuravskaya 2018) employ strategic timing, evidence whether politicians use similar tactics is lacking.
We examine this question by studying the timing of executive orders (EOs) issued by US presidents over the past 40 years (Djourelova and Durante 2019). Presidential EOs provide an ideal setting to study strategic behaviour, because the White House has complete discretion over their timing and content.
EOs do not require the approval of Congress and are, in fact, often used by presidents to push measures that Congress opposes. Because of this, they can generate controversy, with the president being accused of overstepping his constitutional prerogatives. Studies have shown that these concerns are shared by voters and can reduce the president’s popularity (Christenson and Kriner 2017, Reeves and Rogowski 2018).
The political cost of EOs is however ultimately determined by how much media attention they attract. In turn, this depends on what other important events happen at the same time, and compete with EOs for news time (Eisensee and Strömberg 2007). Taking this into account, a president planning to issue a controversial EO may have an incentive to make it coincide with 'distracting' events.
Evidence of strategic timing
To test whether EOs are strategically timed to the news cycle, we examine the relationship between the probability that an EO is signed on a given day, and a daily measure of 'news pressure'. This measures the time devoted to the top three stories in the prime-time newscasts of US TV networks (Eisensee and Strömberg 2007, Durante and Zhuravskaya 2018), and captures the presence of other important news - unrelated to EOs and their content - that can crowd out the news coverage of EOs.
We find that EOs are more likely to be signed on the eve of days when the news is dominated by other events. The positive relationship is depicted in Figure 1, where we plot the frequency of EOs against next-day news pressure. Crucially, this finding applies only to periods of divided government – when the president and Congress belong to opposite parties. Indeed, in these times, unilateral presidential action is more likely to attract criticism due to the presence of a hostile Congress.
Figure 1 Upper panel: Frequency of EOs by quintile of next-day news pressure. Lower panel: Non-parametric regression of signing of EOs on next-day news pressure. Sample: periods of divided government.
Source: Djourelova and Durante (2019).
We then explore what types of EOs, and what types of news, drive this relationship. If the concurrence of EOs and other important news is due to strategic timing, the effect would be more pronounced for EOs that are ex ante more likely to make the news and generate controversy, i.e. EOs that carry a stronger incentive to 'conceal'. We find that the effect is statistically significant (Figure 2) only for:
- EOs on topics other than routine government operations, arguably the least contentious and newsworthy category
- EOs important enough to be covered by the Associated Press newswire
- EOs on issues over which the president and Congress have disagreed the most in the preceding months.
We find, instead, no evidence of strategic timing for EOs announced to the press in advance, which the administration clearly would have no interest in concealing.
Figure 2 Heterogeneity by type of EO
Source: Djourelova and Durante (2019).
Furthermore, if the documented pattern in the timing of EOs is a result of a deliberate forward-looking strategy, then EOs should coincide only with news events that could have been anticipated (political or sporting events, for example), but not with those that were completely unexpected (earthquakes or terrorist attacks). To test this prediction, we apply text analysis techniques to the content of each news segment and distinguish between news containing more words associated with anticipation, and those containing more words associated with surprise. Figure 3 shows the most frequent words occurring in the resulting news categories.
Figure 3 Words in news segments associated with anticipation (left) and surprise (right)
Source: Djourelova and Durante (2019).
We find that the positive relationship between the timing of EOs and news pressure is entirely driven by news associated with anticipation, while there is no correlation with surprising news. This conclusion is further supported by a placebo test exploiting the occurrence of actual unexpected events: earthquakes, terror attacks, and mass shootings. Despite attracting significant attention in the news, the timing of EOs is unrelated to these types of events.
Finally, we ask why EOs may be timed to news on the following rather than the same day. That is, what aspects of next-day coverage may be especially harmful for the president. Video analysis of hundreds of news reports on EOs shows that next-day news are more likely to feature the reactions of Congress – which, during periods of divided government, is on average critical of the president’s conduct.
Limiting the media's watchdog role
A well-functioning media informs citizens about government actions and holds policymakers accountable. Yet, public attention is limited, and sophisticated politicians can reduce scrutiny of their actions by timing announcements to moments when the press and public are distracted. We present evidence that politicians at the highest level may engage in strategic behavior and shed light on the conditions that make such conduct more likely. Our research suggests that, even if the media is not politically biased, the strategic behaviour of politicians can limit its role as a watchdog. Ultimately, this can undermine political accountability.
Christenson, D P and D L Kriner (2017), "Mobilizing the public against the president: Congress and the political costs of unilateral action", American Journal of Political Science 61(4): 769-785.
DellaVigna, S and J Pollet (2009), "Investor inattention and Friday earnings announcements", The Journal of Finance 64(2): 709-749.
Djourelova, M and R Durante (2019), "Media Attention and Strategic Timing in Politics: Evidence from US Presidential Executive Orders", CEPR Discussion Paper 13961.
Durante, R and E Zhuravskaya (2018), "Attack when the world is not watching? US media and the Israeli-Palestinian conflict", Journal of Political Economy 126(3): 1085-1133.
Eisensee, T and D Strömberg (2007), "News droughts, news floods, and US disaster relief", Quarterly Journal of Economics 122(2): 693-728.
Reeves, A and J C Rogowski (2018), "The public cost of unilateral action", American Journal of Political Science 62(2): 424-440.