Updating Presidential Leverage: Comparing Measures of Leverage from Kennedy to Trump

Abstract

This article seeks to accomplish two things. First, I update the Index of Presidential Leverage (IPL) published in Presidential Leverage (Ponder 2018), which analyzed presidential action from Kennedy to the end of Obama’s first term. This article brings the analysis through the end of the Trump administration and replicates the book’s analysis of executive orders to determine if that measure still exhibits explanatory capacity. Second, I consider two alternative measures of presidential leverage that tap the same idea of the IPL, which is presidential approval in an institutional context (namely, public approval of Congress and public trust in Congress). The article employs these alternative measures in a preliminary attempt to determine the efficacy of the concept of presidential leverage across different measures. I find that the concept of presidential leverage operationalized in these various ways is not only robust, but as was the case in the original analysis, outperforms analyses using raw presidential approval in isolation. Thus, the concept of presidential leverage is shown to be applicable across various measurement strategies. The robustness of the presidential leverage measure is important because it undergirds the idea that presidential approval in context, rather than simply as a stand-alone indicator, provides nuance and explanatory power to the study of presidential action that may be missed otherwise.

Acknowledgments

An earlier version of this paper was presented at the 2019 annual meeting of the American Political Science Association. Thanks to Jordan Butcher, Amnon Cavari, Diane Heith, Doug Lemke, Matt Miles, Wendy Whitman-Cobb, and the three anonymous reviewers for excellent comments and suggestions.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Notes

1 See Ponder (2018), Chapter 2, for more details.

2 This section relies heavily on the analysis in Chapters 2 and 4 of Presidential Leverage.

3 The extraction employs Stimson’s WCALC program to calculate the coefficients. I explain this below. The WCALC software is available at https://stimson.web.unc.edu/software/.

4 Examples of presidents who exemplify and illustrate these levels are below.

5 Much of this section, through Obama’s first term, relies on Ponder (2018, Chapter 4).

6 Since leverage is equal to approval divided by trust, it is not surprising that leverage tracks strongly with each. Given the theoretical justifications, ceritus paribus, leverage increases as trust decreases (r= −0.53, p=0.000) and is positively associated with approval (r= 0.38, p=0.003).

7 This is also somewhat surprising given that, as noted in the discussion of the Bush 43 presidency, presidents tend to enjoy a rally around the flag effect, where their approval tends to increase, sometimes dramatically, in times of national turmoil, usually in response to a war (see Brody 1991). The length of that effect is based on patriotism and how long the president’s political opposition mutes its criticism (Hetherington and Nelson 2003). Examining the trends in the era of increasing polarization, from the latter part of the Bush 43 presidency through Obama and into the Trump administration, suggests while some rally effects exist, they are a diminished feature of American politics.

8 For a thorough exploration of why the public seems to “hate” Congress, even as it is the most “democratic” of the institutions, see Hibbing and Theiss-Morse (1995).

9 See a somewhat similar treatment of this general idea in James (2009, pages 59-60).

10 I examine executive orders here as only one component of unilateral action. Executive orders are perhaps, for better or worse, most closely associated with unilateral action. However, as other unilateral weapons are more closely studied, they could also be incorporated into leverage-oriented analyses. At any rate, the examination here of executive orders is in no way meant to imply that orders are the only or even necessarily the most important of the unilateral actins employed by presidents. They have marshaled an increasingly varied array of unilateral actions, such as foreign policy directives, executive agreements, memoranda, signing statements, and so forth. For overviews, see Howell (2005) and Rudalevige (2021b). For a thorough discussion of unilateral actions in the realm of foreign policy, see Peake (2023).

11 The differenced approval coefficient was −0.16, one-tail p=0.30. The lagged approval coefficient was −0.23, one-tail p=0.24.

12 For the model coinciding with the timeline for the congressional approval data, the variable for the differenced presidential approval variable was −0.48, one-tail p=0.15. For the lagged approval variable, the coefficient was −0.52, one-tail p=0.15, with an adjusted R²=0.36. For the model coinciding with congressional trust data, the differenced approval coefficient was −0.45, one-tail p=0.08. For the lagged approval variable, the coefficient was −0.44, one-tail p=0.11. Adjusted R²=0.31.

13 See Ponder (2018, Chapter 7) for a detailed discussion.

14 For an analysis of presidential leverage as opportunity and constraint in the realm of foreign policy, see Ponder and VanDenBerg (2024).