Sage Journals: Discover world-class research

Abstract

Does the type of democratic regime matter for public evaluations of leaders? We argue two characteristics intrinsic to presidential and parliamentary regimes lead to divergent patterns of executive approval. For presidents, direct elections foster more personal leader-voter linkages; for prime ministers, dependence on the legislature for survival contributes to more institutionalized party systems. These two mechanisms should generate higher approval at the outset of a term—larger honeymoons— for presidents than for prime ministers, but also more rapid decline. Analyses of data from 40 countries produce evidence consistent with these constitutionally-based distinctions. Yet we uncover important within-regime differences. Within presidential systems, approval patterns vary along with paths to power—first-election versus re-election, and elected versus unelected. Within parliamentarism, honeymoons are greater for prime ministers overseeing single-party majoritarian governments. Study findings advance long-standing debates about the relative merits of presidential and parliamentary systems—particularly the tradeoff between democratic responsiveness and stability.

Keywords

executive approval presidentialism parliamentarism honeymoon macro public opinion political parties

Does democratic regime type shape citizens’ evaluations of political executives? In his famous article regarding the “perils of presidentialism,” Linz (1990) expressed concern that the “plebiscitarian” nature of presidential elections can ultimately put democratic stability at risk. The image or “aura” presidential candidates cultivate during electoral campaigns, he argued, can lead voters to over-estimate the winners’ ability to accomplish their goals once in office and, inevitably, to become disappointed when they fail. The potential disconnect between the public’s expectations and the reality of governing is further magnified by the winner-take-all nature of majoritarian elections which tempts presidents to conflate victory—however narrow—with a mandate to represent “the people as a whole” (Linz, 1990, pp. 60–61).

Parliamentary systems, Linz posited, are different. Expectations for prime ministers tend to be modest given the nature of their ascent to power and institutional position. Most obviously, they lack an independent popular mandate. Despite the rise of personalized politics in many parliamentary countries (Rahat & Kenig, 2018), prime ministers are typically members of parliament, selected to head the government as agents of their political party, and dependent on legislative majorities to stay in power. Moreover, as heads of government, prime ministers lack the symbolic premium of representing the state. These institutional and symbolic constraints limit parties’ incentives to cultivate direct links between prime ministerial candidates and voters, tempering popular expectations of government performance.

One implication of these arguments is that popular evaluations of presidents should differ from those of prime ministers. But the massive literature comparing presidential and parliamentary regimes largely ignores this implication despite its potential importance for politics in praxis. Executive popularity is a key source of political power. Honeymoons grant presidents and prime ministers the political cover to introduce major, and often costly, reforms (e.g., König & Wenzelburger, 2017; Stokes, 2001). Highly popular presidents more easily afford the often steep “cost of ruling” (Nannestad & Paldam, 2002; Wlezien, 2017), fare better in inter-branch bargaining (Becher & Christiansen, 2015; Martínez-Gallardo, 2014), and pass their agendas more efficiently (Calvo, 2007; Cheibub et al., 2004; Raile et al., 2011). Falling approval may trigger cabinet shuffles (Kam & Indriðason 2005; Martínez-Gallardo, 2014) and even risk the survival of presidents and prime ministers (Laver & Shepsle, 1998; Pérez-Liñán, 2007). But leaders can also leverage their public standing to extend their stay in office (Corrales, 2016; Schleiter & Tavits, 2016). And by forestalling corrective policies, extreme leader popularity may even precipitate financial crises (Herrera et al., 2020). In short, approval dynamics condition what prime ministers and presidents can achieve–overall and relative to each other.

Against this backdrop, we argue that two intrinsic characteristics can lead to divergent patterns of executive approval across regime types. First, echoing Linz (1990, 1994), we expect the incentives to conduct personalized electoral campaigns – which are greater under presidentialism – will encourage bandwagoning, cultivate higher expectations, and thus produce more robust post-election honeymoons for presidents than for prime ministers. Second, we expect the (in)dependence of executive and legislative tenure to shape executive support. In presidential systems, separation of survival reduces party system institutionalization and hampers the formation of strong partisan bonds. Where parties are less stable, we expect both larger honeymoons and more volatile leader support. By contrast, more durable partisan bonds developed in parliamentary systems should limit public bandwagoning and stabilize prime ministers’ approval throughout their tenure.

We test our expectations using time series data on executive popularity for 271 leaders in 40 countries: 19 parliamentary and 21 presidential. Results show that presidents, on average, enjoy greater honeymoons than prime ministers but incur higher costs of ruling and greater volatility in public support. As expected, these differences reflect, to some extent, how incentives to focus electoral campaigns more on individuals than on parties boost initial expectations for incoming leaders. While first-time executives in both regimes generally enjoy a post-election honeymoon, the phenomenon is more pronounced under presidentialism. Under parliamentarism, honeymoons are most substantial for first-time prime ministers who oversee single-party majorities. By contrast, reelected presidents and prime ministers do not enjoy significant honeymoons, suggesting personalism has one-off effects.

In line with our expectations, regime differences in approval ratings also reflect distinct patterns of party system institutionalization. The volatile nature of party competition characteristic of presidentialism is associated with more volatile approval ratings throughout the electoral cycle. Accounting for electoral volatility—our indicator of party system institutionalization—reveals patterns of approval that more closely approximate each other across the two regimes. Ancillary tests of these mechanisms within parliamentary and presidential regimes bolster these claims.

Study findings contribute to long-standing debates about the relative merits of presidential and parliamentary systems. First, the nature of elections shape approval dynamics in ways that Linz’s work implies (1990; 1994). Larger honeymoons stem in part from the incentives in presidentialism to run highly personalized campaigns, inflating popular expectations – especially when presidents win office for the first time. But our findings add important nuance to Linz’s claims by showing how re-election and variation across executives’ terms moderates what at first appear as regime-specific differences. Second, findings on party system stability are consistent with work arguing that many effects of regime choice are indirect (e.g., Cheibub et al., 2014; Mainwaring & Shugart, 1997). Third, our work sheds new light on the tradeoff between democratic responsiveness and stability. Presidents’ higher levels of initial political capital tend to erode quickly as performance fails to meet expectations and “costs of ruling” accrue, whereas prime ministers’ approval is less volatile. Hence, relative to parliamentarism, public opinion dynamics under presidentialism are more unstable but also allow for greater accountability between elections. Finally, our finding that personalism and volatility matter more for executives’ public standing in parliamentary systems further contextualizes debates about the scope for, and possible impact of, the “presidentialization” of politics in parliamentary regimes.

Executive Approval Patterns and Dynamics

Executive approval tends to follow a predictable cycle: it begins well above average but deteriorates over time before experiencing a slight end-of-term boost as new elections approach. These dynamics are evident across numerous presidential regimes over time despite, important differences in the strength of formal institutions, the presence of checks and balances, and the vigor of civil society and media independence. Indeed, this cyclical pattern has been well established in the United States (Gronke & Newman, 2003; Mueller, 1973; Stimson, 1976) and more recently across contemporary Latin America (Carlin et al., 2018), with some exceptions (Carlin & Martínez-Gallardo, 2019; Mueller, 1970). Re-elected presidents, however, rarely enjoy much of a second honeymoon (Brace & Hinckley, 1992).

Similar cyclical patterns exist in parliamentary systems. Müller and Louwerse (2020) find vote intentions for government parties fit the familiar, partially U-shaped pattern in a broad cross-section of parliamentary democracies.¹ Electoral support for government parties tends to start out high, decrease over the first half of the electoral cycle, and recover slightly as elections near. Employing either vote intention or approval data, several case studies find cyclical dynamics, including in Japan (Krauss & Nyblade, 2005), Canada (Johnston, 1999), Italy (Bellucci & Memoli, 2023), and the UK (Anderson, 1995a; Yantek, 1985). Evidence is mixed, however, for Denmark, Germany, the Netherlands (Anderson, 1995a, 1995b), and Spain (Romero-Vidal, Orriols, & Riera, 2023).

Despite the theoretical and political implications, few studies compare patterns of approval across constitutional regimes. One exception is Carlin et al. (2012), on which we build. Although it establishes some key cross-regime differences in cyclical dynamics, it does not assess why these differences exist or control for rival explanations. Relying on a more limited and unbalanced sample available at the time, the set of countries it employs is also overweighted toward presidentialism and excludes most parliamentary systems of Central and Eastern Europe, the Nordic countries, as well as Japan. With the Philippines as the only Asian case, their cross-regime comparisons are almost exclusively between developed parliamentary democracies in Western Europe and developing presidential systems in the Americas, limiting the external validity of their conclusions.² And, as their dataset ends in 2009, it excludes more than a decade of politics which may be consequential to electoral-cycle dynamics, especially as party systems and campaign dynamics in many parliamentary regimes become more volatile and personalized (see Müller & Louwerse, 2020).

Lack of research comparing approval dynamics is puzzling—especially in light of a venerable literature on comparative constitutional regimes that implies regime type should lead to divergent patterns of executive approval. We overcome previous theoretical and empirical limitations by connecting two key features of presidential and parliamentary regimes to executive approval and testing our expectations on a more comprehensive and balanced sample.

Constitutional Regimes and Executive Approval

We expect two regime particularities to generate different patterns of executive approval. First, as Linz (1990) noted, the direct popular election of the executive in presidential systems creates a “plebiscitarian” link between voters and leaders that is substantively different from the one established between voters in parliamentary systems and the prime minister. Second, the combination of direct/indirect elections with the mutual independence/dependence of the executive and legislative branches shape parties’ organizational strength, discipline, and centrality to the fate of executives (Samuels & Shugart, 2010). In parliamentary systems parties are strengthened by the requirement that prime ministers are supported—or at least not opposed—by a majority of the assembly to survive in office. In contrast, direct popular elections and fixed terms in presidential systems fuel political dynamics that weaken party institutionalization and boost electoral volatility.

Below we theorize how these parliamentary and presidential institutions shape the strategic environments in which leaders campaign, attain office, relate to their parties and voters, and govern—in short, the larger context in which citizens evaluate political executives.

Executive Selection, Personalism, and Executive Approval

Modes of executive selection generate distinct incentives for political actors. In presidential systems with popular elections, electoral campaigns typically center around the attributes of a single candidate. Conversely, indirect elections in parliamentary systems encourage campaigns focused on the platforms of political parties (Rahat & Sheafer, 2007). The resulting levels of electoral personalism—the degree to which individual actors, including the chief executive, are more prominent in the political process than parties or collectives—could be expected to translate into different patterns of executive approval. This should be especially clear at the beginning of a leader’s first term, when the election still looms large and hopes remain high.

In presidential systems, the direct election of a chief executive encourages parties to concentrate resources on presidential campaigns, shifts media focus to presidential contenders, and trains the public’s attention on candidates’ personal reputations rather than their parties’ platforms (Carey & Shugart, 1995; Samuels, 2002). Presidential campaigns are thus designed to highlight not only candidates’ policy agendas but also their unique personal attributes and political qualities. As a result, presidents rely more heavily on their personal charisma and ability to connect with voters to reach public office rather than working their way through traditional legislative and party hierarchies (Samuels & Shugart, 2010).

Candidate-centric campaigns are likely to create, in Linz’s words (1990, p. 53), “an aura, a self-image, and a set of popular expectations” that fuel higher-than-average levels of approval at the outset of a presidential term. According to Stimson (1976), this honeymoon is propelled by a public largely uninformed about public policy and, thus, prone to believe candidates’ campaign promises, regardless of their feasibility. The approval boost is fleeting, however. By expecting from leaders “more than can possibly be achieved under the best of circumstances,” the public “is always prone to great disappointment over what is actually achieved by merely mortal Presidents acting under less-than-ideal conditions” (Stimson, 1976, pp. 9–10). These early “naïve expectations” explain why new presidents pick up the support of a wide swath of voters, including many who did not vote for them (or did not vote at all) but are willing to jump on—and eventually off—the bandwagon (Stimson, 1976, p. 10; see also Brace & Hinckley, 1992; Castro Cornejo et al., 2022). This should be particularly true following elections that first bring presidents to office and less so for re-elections, in which the incumbent is a known commodity and expectations are more realistic.

By contrast, prime ministerial candidates do not run as individuals in a direct popular election but compete for the post indirectly from within their party. We expect this mode of executive selection to translate into lower levels of initial approval—honeymoons—for prime ministers compared to presidents for at least two reasons. First, indirect elections reduce the incentives of prime ministerial candidates to cultivate direct ties with voters. General election campaigns, in turn, tend to emphasize party links and platforms over the attributes of individual candidates (Kriesi, 2012). Second, citizens cannot split their ticket and vote for a prime ministerial candidate while voting against their party and, thus, are less likely to give initial approval to an executive from a party or coalition for which they did not vote (Samuels & Shugart, 2010).³

Beyond indirect elections, parliamentary systems produce less personalism than presidential systems in other ways. As heads of government and heads of state, presidents are imbued with a degree of national symbolism that prime ministers, mere heads of government, lack. Expectations for prime ministers are also more tempered: their government’s survival rests on maintaining the confidence of parliament, often requiring complex politics of coalition maintenance and granting them less room for maneuver than presidents, who serve fixed terms. Finally, turnout skews higher in parliamentary systems (Fumagalli & Narciso, 2012; Tavits, 2009), undermining prime ministers’ ability to leverage their personal appeal into initial support from non-voters.

Recent research, however, cautions against deterministic connections between constitutional regime and elite behavior, highlighting institutional variation both within and across presidential and parliamentary regimes (e.g., Cheibub et al., 2014). Intra-regime variation may be large enough to mitigate the incentives generated by different modes of executive selection. With respect to parliamentary systems, the office of the prime minister has grown more personalized as media-driven campaigns have become ubiquitous, often featuring televised leader debates and intensive social media campaigns (Poguntke & Webb, 2005; Webb & Poguntke, 2013). In turn, parties in parliamentary regimes face incentives to shift towards selecting candidates who can succeed in campaigns focused increasingly on individual personalities. Declining partisan attachments across parliamentary democracies contribute to this trend by reducing voters’ reliance on party cues and increasing their reliance on candidates’ personal traits (McAllister, 2007). In sum, although evidence of personalization in parliamentary regimes is mixed (Campus & Pasquino, 2006; Dowding, 2013; Kriesi, 2012; McAllister, 2007), we can expect approval dynamics in parliamentary regimes with high levels of personalism to approximate those of presidential regimes.

This discussion leads to a pair of testable claims:

H₁: On average, presidents in presidential systems begin their terms in office with higher levels of approval (stronger honeymoons) than prime ministers in parliamentary systems.

H₂: Differences in executives’ initial approval levels (honeymoons) between presidential and parliamentary systems diminish once we account for the extent of electoral personalism.

Executive-Legislative Survival, Party System Institutionalization, and Executive Approval

The first set of hypotheses focus on how presidents and prime ministers are selected. Variation in patterns of approval across constitutional systems also stems from the nature of executive-legislative dynamics and their effects on the party system. Whereas the executive and legislature are elected independently in presidential systems, in parliamentary systems their fates are linked. In parliamentary systems, the executive’s Prime ministers’ duration in office is endogenous to their public standing, a feature which may delay approval decline. Prime ministers must be supported, or at least not opposed, by a legislative majority—and if the government falls, members of parliament must face voters in new elections.⁴ The need to consolidate support around the prime minister creates incentives for party building and leads, on average, to more institutionalized parties in parliamentary regimes. These parties are typically characterized by strong ideological and programmatic linkages, ties to organized interest groups, the presence of leaders with extensive party experience, and strong organizational structures (see Mainwaring, 2018, p. 92). Dependence on their party further limits prime ministers’ incentives and ability to cultivate independent bases of support. For voters, party institutionalization strengthens partisan attachments and raises the psychic cost of withdrawing support from the prime minister.

Lacking a majority requirement for survival, presidential party systems tend to be less institutionalized, and legislative parties less cohesive and disciplined (Linz, 1994; Samuels & Shugart, 2010). Moreover, the direct election of presidents means popular candidates can win without the support of strong parties, making it more likely that outsiders will run in, and win, elections (Linz, 1994). Split-ticket voting, possible in most presidential systems, further allows presidents and legislatures to run on different platforms and cultivate unique bases of support (Samuels & Shugart, 2010). In short, presidential regimes tend to have weaker, less institutionalized party systems.

This has implications for executive approval. Weak parties, typical of presidential systems, create a bloc of floating voters (political independents and weak partisans) that can jump onto the executive’s bandwagon at the start of the term, contributing to larger honeymoons. An additional implication of weak party identification is that voters are more likely to switch party allegiances in the face of crises or shocks (Kayser & Wlezien, 2011; Lupu, 2015; Mainwaring, 2018), making job approval ratings volatile and potentially leading to rapid declines that can erase initial popularity advantages over prime ministers. In contrast, strong parties and durable partisan bonds, more common in parliamentary democracies, limit the magnitude of any honeymoon, but contribute to a stable minimum core of support throughout a leader’s term.

The consideration of the relationship of executives with their party and with the assembly motivates two additional hypotheses:

H₃: On average, approval for presidents declines more precipitously over their terms compared to approval for prime ministers.

H₄: Differences in approval dynamics between presidential and parliamentary systems diminish once we account for party system institutionalization.

Our discussion to this point has largely focused on initial approval ratings and their subsequent decline. Yet approval levels often rebound as executive terms wind down, either due to the prospects of a new term or their lame-duck status (Stimson, 1976). Indeed, Müller and Louwerse (2020) reveal a modest end-of-term rebound in parliamentary systems, especially under single-party government or when prime ministers control cabinet dissolution. Carlin et al. (2018) also identify a small end-of-term boost in approval for presidential systems. To the extent that the rebound is influenced by the nature of campaigns ahead of the next election, or by unanchored voters amidst weakly institutionalized parties, we should expect it to be larger for presidents than for prime ministers. This implies a final hypothesis:

H₅: On average, approval for presidents rebounds more strongly at the end of the term of the executive compared to approval for prime ministers.

To be clear, these differences across regimes are indirect features of institutional incentives, and exceptions exist. Parliamentary regimes can feature weak parties and exhibit different degrees of personalism; personalism also varies across presidential regimes as does the degree to which durable and institutionalized parties shape elections and governing. This suggests that there should not only be differences across regimes but variation within them and, further, that presidential and parliamentary regimes should converge when personalism or their party systems are similar.⁵

Baseline Executive Approval Dynamics in Presidential and Parliamentary Systems

Our main analytic goals are to establish systematic differences in approval dynamics between presidentialism and parliamentarism and to identify factors characteristic of each regime type that help explain these differences. In this section we focus on the first goal: comparing patterns of executive approval across regime types.⁵

While scholars have studied executive approval for over a half-century, lack of comparable data has impeded their ability to test expectations about patterns of approval across constitutional regimes. Conceptually, there is the question of who forms the executive. While in parliamentary systems the head of state (a president or monarch) and head of the government (a prime minister) are separate, in presidential systems these two functions are unified. Moreover, although our focus is on comparing parliamentary heads of government with individual presidents, in some parliamentary regimes the government is viewed as collective (that is, as including the cabinet, not just the prime minister). In practice, this means that while polls in presidential regimes seek evaluations of individual presidents, in parliamentary regimes surveys might assess leadership performance by tapping views about the prime minister and/or the government as a collective entity.

Fortunately, a worldwide explosion in the number of firms gauging executive approval in the past decades has enabled the Executive Approval Project to construct the Executive Approval Database (EAD). The EAD 2.01, released in 2019, includes quarterly executive approval measures for 19 parliamentary and 21 presidential regimes.⁶ The EAD employs Stimson’s (1991) dyads-ratio algorithm to construct a continuous latent executive approval time series based on input series collected from a variety of academic and commercial pollsters in each country. Key advantages of this approach include its ability to handle differing survey timing and frequency, missingness, and distinct question wordings. Additionally, the dyads-ratio algorithm estimates and helps correct house and sampling frame effects. Research suggests EAD estimates are comparable and reliable across a range of cases and time periods (Carlin et al., 2018; Carlin et al., 2023).

To compare levels and dynamics of executive approval across regime types, we first code every case in the dataset as presidential or parliamentary.⁷ We follow Elgie (2011) in labeling systems presidential if the head of government is directly elected. Overall coding decisions are relatively straightforward, with two exceptions. The first is Peru. At times scholars have viewed the Peruvian system as semi-presidential because the office of President of the Council of Ministers is sometimes referred to as primer(a) ministro(a). However, the political discourse and locus of executive power in Peru are clearly in the president’s hands. Similarly, South Korea has an office of Prime Minister for State Affairs. But, as in Peru, the holder is clearly subservient to the president. Both countries are coded as presidential.

Systems in which the head of state is not popularly elected are coded as parliamentary. These include countries with monarchs (e.g., Canada, Denmark) or indirectly elected presidents (e.g., Germany, Italy). We also label parliamentary those cases in which a directly elected head of state’s powers are constitutionally weak or not used in practice: Austria, Iceland, Ireland, North Macedonia, and Slovenia. Presidents in these cases are rarely involved in day-to-day politics, and elections to this office often go uncontested.

We begin our analysis by comparing overall executive approval dynamics across regime types. As a first cut, we use a semi-parametric approach—kernel-weighted local regression with 95% confidence intervals—to generate smoothed estimates of local average percent-positive approval rating (y-axis) for prime ministers and presidents between elections (x-axis) (Figure 1(a)).

Figure 1.

Approval dynamics by regime type. (a) Smoothed approval dynamics across regimes, 95% c.i. (b) Smoothed mean-centered approval dynamics across regimes, 95% c.i.

Consistent with our argument that approval dynamics vary by regime, Figure 1(a) displays stark differences in how citizens in parliamentary and presidential regimes evaluate their leaders. Presidents on average enjoy much higher approval ratings following an election—greater honeymoons—than prime ministers. Afterwards, they experience a more substantial drop in approval than parliamentary leaders. As the next election nears, presidents and prime ministers end their terms with roughly similar levels of approval. Cross-regime differences appear in sharper relief in Figure 1(b), which centers the series around the mean level of approval between elections. This controls for baseline support levels and is comparable to adding fixed effects by executive term.⁸

We must be careful, however, not to assume that these differences in executive approval dynamics are based on the presidential-parliamentary distinction without controlling for factors deemed to influence popular support which may also vary by regime type. Accordingly, we examine the robustness of these differences by regressing Approval on several additional factors.⁹ First, since the economy often influences executives’ job evaluations (Hellwig & Singer, 2023), we include Growth and Inflation, measured annually using the World Development Indicators. Second, to ensure our results are not driven by the degree of political democracy rather than the type of democratic regime, we use the annual Polyarchy score from the Varieties of Democracy dataset (V-Dem).^10,11 Since studies link presidents’ gender to approval levels (Carlin et al., 2020; Reyes-Housholder, 2020), and we have no evidence if the same is true for prime ministers, we add a Male leader dummy variable.

Models include a set of indicators to assess electoral cycles.¹² To capture honeymoon dynamics, we include dummy variables for the quarter when an executive election occurs (Election_t) as well as for the two subsequent quarters after the honeymoon (Election_t-1 and Election_t-2) to evaluate the decay of the direct honeymoon effect (Carlin et al., 2018). We also include dummy variables for the two quarters preceding an election (Election_t+1 and Election_t+2) to estimate any end-of-term election campaign effects.

We fit a series of parametric autoregressive distributed lag (ADL) models with random country-level intercepts to address panel heteroskedasticity. Panel unit-root tests reject consistent non-stationarity in the approval, growth, or inflation series. Thus, a panel ADL, equivalent to a general error correction model, is an appropriate specification and we report below mixed-effects models, estimated using OLS with country-level random intercepts. Models using estimators with panel-corrected standard errors or feasible generalized least squares with heteroskedastic consistent errors produce substantially identical results (see Table A3 in the appendix).

Results, reported in Table 1, align with our expectations regarding overall approval dynamics. Column I considers executive approval independent of regime type. The significant, positive coefficients for Election_t and Election_t-1 suggest that the “average” leader can expect to enjoy a honeymoon of around two quarters. The insignificant coefficient for regime type (Parliamentary System) implies that mean prime minister popularity is not reliably different from mean presidential popularity. The results also indicate a small campaign boost in approval ahead of elections (Election_t+1). Finally, in line with previous work (e.g., Cerda & Vergara, 2022), Growth_t consistently increases levels of approval, whereas Inflation_t does not.

Table 1.

Baseline and Dynamics of Approval in Presidential and Parliamentary Regimes.

	I	II
Approval_t-1	0.904**	0.903**
Approval_t-1	(0.006)	(0.006)
Male leader_t	0.304	0.315
Male leader_t	(0.282)	(0.280)
Growth_t	0.193*	0.199*
Growth_t	(0.085)	(0.085)
Growth_t-1	−0.137	−0.141
Growth_t-1	(0.085)	(0.085)
Inflation_t	0.001	0.001
Inflation_t	(0.001)	(0.001)
Inflation_t-1	−0.001	−0.001
Inflation_t-1	(0.001)	(0.001)
Polyarchy_t	−0.715	−0.732
Polyarchy_t	(0.599)	(0.596)
Parliamentary system_t (ref cat: Presidential_t)	−0.311	0.162
Parliamentary system_t (ref cat: Presidential_t)	(0.186)	(0.213)
Election_t	2.65**	2.982**
Election_t	(0.314)	(0.461)
Election_t-1	4.7**	7.055**
Election_t-1	(0.315)	(0.463)
Election_t-2	0.076	1.023*
Election_t-2	(0.315)	(0.464)
Election_t × parliamentary_t		−0.686
Election_t × parliamentary_t		(0.626)
Election_t-1 × parliamentary_t		−4.369**
Election_t-1 × parliamentary_t		(0.629)
Election_t-2 × parliamentary_t		−1.792**
Election_t-2 × parliamentary_t		(0.625)
Election_t+1	1.505**	1.668**
Election_t+1	(0.314)	(0.461)
Election_t+2	0.363	0.414
Election_t+2	(.313)	(0.457)
Election_t+1 × parliamentary_t		−0.345
Election_t+1 × parliamentary_t		(0.626)
Election_t+2 × parliamentary_t		−0.124
Election_t+2 × parliamentary_t		(0.624)
Constant	3.906**	3.703**
Constant	(.608)	(0.607)
Observations	4558	4558
Number of countries	40	40

Note. Cells report coefficients from autoregressive distributed lag estimates with random country-level intercepts and standard errors in parentheses. **p < .01, *p < .05, two-tailed test.

While the popularity of presidents and prime ministers does not differ on average, H₁ and H₃ focus, respectively, on starting approval levels and subsequent rates of decay. To test those hypotheses, the model in Column II parametrically evaluates how honeymoons differ between presidential and parliamentary regimes by interacting the election dummies with regime type. Results are consistent with our argument. Parameter estimates on the interaction terms Election_t-1 × Parliamentary and Election_t-2 × Parliamentary show that executives in presidential regimes tend to enjoy significantly larger honeymoons than their counterparts in parliamentary regimes. Presidents start out more popular than prime ministers, but the coefficient Parliamentary System_t is indistinguishable from zero, indicating no difference across regimes in average approval levels over the long run. Presidents thus lose their initial advantage over prime ministers as their support decays more quickly. Finally, with this model we can also test H₅. The coefficients on Election_t+1, Election_t+2 and their interaction with regime type indicate that leaders in both regimes see their popularity rebound to roughly the same extent, but only in the final quarter of their terms. The null effects on their interaction with regime type suggests their rebounds do not mediffer.

In sum, Table I highlights both similar overall approval patterns for presidential and parliamentary systems and key differences in the ways that leaders experience them. We find strong evidence for two of our hypotheses, as presidents more consistently begin their terms in office with higher levels of approval than prime ministers (H₁), and their approval declines more precipitously (H₃). However, while we observe modest end-of-term bounces, contrary to (H₅) we find no evidence that presidents experience a greater end-of-term bounce than prime ministers.

Unpacking Regime Effects: Personalism, Party Institutionalization, and Alternative Explanations

We now test theoretical expectations regarding two factors that we expect to be responsible for intra-regime differences in approval dynamics: personalism and party system institutionalization.

Personalism

We expect higher levels of personalism lead to more marked honeymoons in presidential compared to parliamentary systems (H₂). Measuring electoral personalism within and across regimes is complicated. Certain characteristics of political campaigns (e.g., length, access to televised debates, financing regulations, etc.) may cause personalism to vary across regimes. Unfortunately, there are no comparable over-time measures of campaign characteristics across multiple countries and regime types, inhibiting empirical tests.¹³ We therefore evaluate the potential importance of personalist elections by leveraging two related features: the role of first-time relative to re-elected executives, and the role of government type.

First, we argue personalism is higher when electoral institutions incentivize campaigns centered on individual candidates. Such campaigns can give candidates an aura of invincibility, extending their appeal beyond their party supporters and cultivating a sense of hope and expectation that translates into higher initial approval rates. But this exercise in “myth building,” to paraphrase public opinion pioneer Burns Roper (1969; quoted in Mueller, 1970, p. 20), is likely only possible when candidates run for, and win, office the first time. Incumbents running for re-election are known commodities: voters now have a clear sense of their (often tarnished) record and calibrate expectations accordingly. So, if the association between regime type and honeymoon size does indeed stem from a close personal connection between candidates and voters—the kind that facilitates an executive’s initial acclamation as a “providential figure” (O’Donnell, 1994 cited in Linz, 1994, p. 29)—then this effect should be larger for executives who are elected for the first time.

To test the contention that honeymoon effects differ depending on whether the leader is new to the office (or has been out of office for at least one term) or is continuing in office after a successful reelection, we estimate models similar to the ones in Column II in Table 1 but disaggregate election variables into First election and Re-election.¹⁴

Results presented in Table 2 provide strong evidence that honeymoons are mostly a function of politicians first winning office.¹⁵ The coefficients for First election_t, First election_t-1, and First election_t-2 capture approval dynamics for presidents who were elected for the first time (or newly elected after at least one term out of office) and are all positive and statistically significant, implying that new presidents’ approval increases for nine months following the election. Coefficients capturing the potential honeymoon of presidents in their second term (Re-election_t, Re-election_t-1, and Re-election_t-2), in contrast, provide little evidence of post-election boosts (see also Figure 2 below). In conclusion, presidential honeymoons are largely limited to newly elected presidents whose election, we argue, allows for the plebiscitarian, personalistic dynamics that Linz associated with presidential politics.

Table 2.

Dynamics of Approval in Presidential and Parliamentary Regimes: Election and Re-Election.

	III
	Coeff.	Std. Err
Approval_t-1	0.908**	(0.006)
Male leader	0.294	(0.277)
Growth_t	0.188**	(0.083)
Growth_t-1	−0.113	(0.083)
Inflation_t	0.000	(0.001)
Inflation_t-1	0.000	(0.001)
Polyarchy	−0.835	(0.588)
Parliamentary system (ref cat: Presidential)	0.202	(0.210)
First election_t	2.985**	(0.506)
Re-election_t	3.009**	(0.980)
First election_t-1	8.725**	(0.506)
Re-election_t-1	0.268	(0.999)
First election_t-2	1.719**	(0.506)
Re-election_t-2	−1.950*	(0.999)
First election_t × parliamentary	1.335*	(0.792)
Re-election_t × parliamentary	−2.352**	(1.123)
First election_t-1 × parliamentary	−2.988**	(0.795)
Re-election_t-1 × parliamentary	−0.024	(1.139)
First election_t-2 × parliamentary	−2.382**	(0.782)
Re-election_t-2 × parliamentary	1.052	(1.140)
Election_t+1	1.672**	(0.454)
Election_t+2	0.450	(0.451)
Election_t+1 × parliamentary	−0.378	(0.618)
Election_t+2 × parliamentary	−0.220	(0.616)
Constant	3.565**	(0.598)
Observations	4558
Number of countries	40

Note. Cells report coefficients from autoregressive distributed lag estimates with random country-level intercepts and standard errors in parentheses. **p < .01, *p < .05, two-tailed test.

Because the timing between election and inauguration can differ across countries or within countries across time, as can the length of coalition formation negotiations across governments, it can be misleading to infer substantive differences across regimes from the effect sizes of specific quarters after an election. Thus, Figure 2 displays the total direct effect (total marginal effect) of an election on the executive’s honeymoon by regime and election-type.¹⁶ The findings are generally supportive of H_2. For executives elected for the first time, average honeymoon approval, i.e., during the nine months post-election, is significantly higher than average baseline approval, i.e., excluding the honeymoon and two pre-election quarters. Presidents average approximately 13 percentage points higher approval following their first election, whereas prime ministers average about 9 points above the baseline. This four-point gap provides some evidence of the higher cost of ruling facing executives under presidentialism (c.f., Wlezien, 2017). Following a successful reelection, however, neither presidents nor prime ministers can reliably expect to enjoy a honeymoon of any substance.

Figure 2.

Approval dynamics by regime type and reelection. Note. Panels display model-based mean Approval (total marginal effect) of the election quarter and subsequent two quarters following an election (“honeymoon”) for first Elections and for Reelections versus the rest of each respective term (Baseline). Based on estimates from Table 2, with 90% c.i.

To further probe the relationship between personalism and approval dynamics, we compare single-party majorities to other governments within parliamentary systems.¹⁷ There are good reasons to expect higher electoral personalism in the former and, thus, larger honeymoons for the prime ministers who lead them. Where single-party majority governments are common, parties play a smaller role in government formation, and electoral campaigns tend to center on the candidate for prime minister (see Cross & Young, 2015). By contrast, where single-party majorities are less common, parties must build coalitions. Thus, attention during elections shifts from the attributes of individual candidates for prime minister to the party leaders involved in coalition negotiations (Langer & Sagarzazu, 2018). Elections in such contexts tend to be comparatively less personalized, as campaign messages do not revolve around candidates to head the government (e.g., Holtz-Bacha et al., 2014 comparing the United Kingdom and Germany).¹⁸

Figure 3 reports the total direct effect of first and subsequent elections for single-party majority and other government types across our parliamentary cases. As expected, the effect of elections on PM approval during the nine-month honeymoon period is substantially larger under majority governments than under all other types of parliamentary government. Again, this is true only for first elections. Strikingly, honeymoons for prime ministers leading single-party majority governments are essentially equal in magnitude to the average president’s honeymoon in Figure 2. When parliamentary elections are personalized, the post-election boost in approval ratings approximates approval for a directly-elected first-term president. Moreover, as with presidents, those high ratings quickly revert to baseline levels of popularity, indistinguishable from approval levels under other government types.

Figure 3.

Approval dynamics by majority and government status: Prime Ministers Only. Note. Panels display model-based mean of Approval (total marginal effect) for executives in parliamentary regimes for the election quarter and subsequent two quarters following an election (by type) and the mean for all but the final two quarters of the term (Baseline). Panels are separated by majority/minority status (rows) and single-party/coalition status (columns). Vertical bars report 90% confidence intervals. Based on models in Table A11 in the online appendix.

In summary, these results demonstrate how personalism drives executive approval dynamics in two ways. First, executive approval tends to rise most dramatically following the most personalized—first—elections in both presidential and parliamentary regimes. Second, differences between majority and other government types suggest personalization of elections magnifies these dynamics within parliamentary regimes. Although we are limited to testing our argument indirectly, the evidence presented here suggests that honeymoons are at least partially a function of the personalized relationships that direct elections cultivate. Unpacking the effects of first and subsequent elections also provides insight into why we find no support for H₅—which draws on research on the cyclical dynamics of approval to posit an end-of-term bump. Namely, any cyclical effects, either across or within regimes, are dwarfed by these “re-election” effects.

Party System Strength

Our second argument is that lower levels of party system institutionalization in presidential systems contribute to greater initial honeymoons and, subsequently, to sharper declines in approval as the executive’s term progresses. If this is the case, we should find (per H₄) that differences in approval dynamics between regimes diminish once we account for variations in institutionalization. To assess this proposition, we use a common operational measure of party system institutionalization, electoral volatility. We adopt and update Mainwaring et al.’s (2017) measure of total volatility in votes for parties between national legislative elections. Though distinct, presidential and legislative electoral volatility are linked (Mainwaring et al., 2017), and legislative volatility allows us to compare countries with and without direct elections of the executive. As expected, on average, presidential systems in our sample have substantially higher volatility (mean = 35.43) than parliamentary systems (mean = 14.15), indicating lower party system institutionalization in the former. Overall, parliamentary cases have low volatility with few countries displaying levels of volatility similar to the mean presidential system (see Figure A7 in appendix for histograms of volatility by regime type).¹⁹

To test the claim that cross-regime differences in executive approval dynamics are, in part, a function of electoral volatility, we again turn to a parametric model. To test whether honeymoon bumps are larger in more volatile party systems, we interact electoral volatility and dummy variables for the first three quarters that leaders are in office. We also interact regime type and the post-electoral quarters for first elections and re-elections to see if including party-system institutionalization reduces the unexplained gap between presidential and prime ministerial approval early in the term observed in Figure 2. For efficiency and clarity, we only include one-quarter lags of growth and inflation and exclude interactions between regime type and the pre-election, end-of term dummies (statistically insignificant in Table 1 and in Table A3).

Table 3 reports parameter estimates. Model IV, which includes all executives, yields three main takeaways. First, overall, executives who govern in less institutionalized (more volatile) party systems appear to receive larger honeymoon bumps in their first and second quarters than executives in more institutionalized party systems (Model IV). Second, these effects are restricted to the first term; second-term honeymoon bumps tend to be small and short-lived regardless of the degree of party institutionalization. Finally, statistically insignificant coefficients on interactions between regime type (Parliamentary) and the first three quarters of an executive’s term suggest that presidential and prime ministerial honeymoons do not meaningfully differ once we account for our hypothesized variables: electoral volatility (proxy for party system institutionalization) and re-election (proxy for personalism).²⁰

Table 3.

Approval, Elections, and Electoral Volatility in Presidential and Parliamentary Regimes.

	IV	V	VI
	All Executives	Presidential	Parliamentary
Approval_t-1	0.907**	0.904**	0.899**
Approval_t-1	(0.007)	(0.009	(0.01)
Male leader	0.163	1.282**	−1.014**
Male leader	(0.296)	(0.464)	(0.377)
Growth_t-1	0.0533	0.037	0.0948*
Growth_t-1	(0.028)	(0.041)	(0.037)
Inflation_t-1	<-0.000	<-0.000	<-0.048
Inflation_t-1	(<0.000)	(<0.000)	(<0.032)
Polyarchy	−1.226	−1.309	−0.291
Polyarchy	(0.677)	(0.829)	(1.448)
Parliamentary system	−0.045
Parliamentary system	(0.216)
Volatility	−0.006	−0.007	−0.009
Volatility	(.005)	(.006)	(.012)
First election_t	3.383**	3.435**	4.503**
First election_t	(0.835)	(0.972)	(0.936)
Re-election_t	3.120*	3.322	0.609
Re-election_t	(1.580)	(1.832)	(0.951)
Election (first)_t-1	5.845**	5.510**	5.764**
Election (first)_t-1	(0.835)	(0.972)	(0.936)
Reelection_t-1	0.217	−0.387	1.538
Reelection_t-1	(1.580)	(1.832)	(0.951)
Election (first)_t-2	0.0557	0.509	−2.536**
Election (first)_t-2	(0.843)	(0.984)	(0.925)
Reelection_t-2	−2.507	−1.695	−2.472*
Reelection_t-2	(1.620)	(1.867)	(0.991)
First election_t × parliamentary	1.238
First election_t × parliamentary	(0.866)
Re-election_t × Parliamentary	−2.307
Re-election_t × Parliamentary	(1.442)
First election_t-1 × parliamentary	−1.456
First election_t-1 × parliamentary	(0.867)
Re-election_t-1 × parliamentary	−0.016
Re-election_t-1 × parliamentary	(1.442)
First election_t-₂ × parliamentary	−1.521
First election_t-₂ × parliamentary	(0.861)
Re-election_t-2 × parliamentary	1.521
Re-election_t-2 × parliamentary	(1.481)
First election_t × volatility	−0.013	−0.016	−0.001
First election_t × volatility	(0.02)	(0.024)	(0.048)
Re-election_t × volatility	−0.008	−0.009	0.013
Re-election_t × volatility	(0.031)	(0.037)	(0.075)
First election_t-1 × Volatility	0.081**	0.091**	<-0.000
First election_t-1 × Volatility	(0.02)	(0.024)	(0.048)
Re-election_t-1 × volatility	0.00759	0.025	−0.114
Re-election_t-1 × volatility	(0.031)	(0.037)	(0.075)
First election_t-₂ × volatility	0.052*	0.039	0.124**
First election_t-₂ × volatility	(0.02)	(0.024)	(0.047)
Re-election_t-2 × volatility	0.011	−0.007	0.152
Re-election_t-2 × volatility	(0.031)	(0.037)	(0.081)
Election_t+1	1.657**	2.005**	1.382**
Election_t+1	(0.32)	(0.531)	(0.373)
Election_t+2	0.398	0.569	0.218
Election_t+2	(0.32)	(0.535)	(0.369)
Constant	4.364**	3.541**	5.064**
Constant	(0.741)	(0.95)	(1.483)
Observations	4291	2221	2070
Number of countries	40	21	19

Note. ADL models estimated with country-level random intercepts. Standard errors in parentheses.

**p < .01, *p < .05, two-tailed test.

The role of party system institutionalization is summarized in Figure 4. Comparing average honeymoon effect size across regime types, the discrepancy in honeymoons between first-term presidents and first-term prime ministers observed in Figure 2 disappears when we introduce the interaction with volatility (based on Table 3, Column IV). Again, consistent with our expectations (H₄), these results provide evidence that, overall, the differences in approval dynamics in Figure 1(a) and (b) are at least in part a function of party system institutionalization. Once we consider the effects of both volatility and reelection, approval dynamics across presidential and parliamentary systems look similar.

Figure 4.

Approval dynamics by regime-type and reelection, accounting for volatility. Note. Panels display model-based mean of Approval (total marginal effect) of the election quarter and subsequent two quarters following for first Elections and for Reelections compared to the mean level for the rest of the term (Baseline). Based on estimates from Table 3, Column VI, with 90% c.i. Using the natural log of volatility has no substantive effect, see Figure A5 in the appendix.

Not all presidential regimes and parliamentary regimes are equal, however, and if party systems shape the public’s reaction to new leaders, we should see differences within regime types as well. Models V and VI examine approval dynamics separately for presidents and prime ministers. Results in Model V confirm that weaker party systems grant presidents greater honeymoons. Results for prime ministers in Model VI show similar effects of volatility, though its conditioning effect attains statistical significance on the second election lag (First election_t-₂ × Volatility). Overall, honeymoons for presidents with strong party systems are very similar to those for the average prime minister, while honeymoons for prime ministers with weaker party systems more closely approximate those of the average president.

These results also clarify the implications of the analysis in the previous sections. Table 2 and Figure 3 show approval tends to start higher and fall more following personalist elections. However, once we also consider party system strength, in the form of electoral volatility, we can largely account for the stark approval dynamics observed in Figure 1. Hence, personalism and party system institutionalization go a long way towards understanding the different approval dynamics that presidents and prime ministers enjoy.

Alternative Explanations: Age of Democracy and Unelected Leaders

We assess the strength of our results in several ways. First, since party system institutionalization increases with the age of democracy, we run models including this variable (Mainwaring & Torcal, 2006; Mainwaring & Zoco, 2007). This also allows us to consider Cheibub’s (2007) argument that differences in democratic survival between regime types are not primarily linked to institutions (Linz, 1994) but to the fact that presidentialism occurs in challenging political environments, such as Latin America during the Cold War. Further, if newer democracies are less compatible with democratic survival and less conducive to party building, we should see higher levels of approval and less fluctuation in more established democracies. Our measure is the number of years since a democratic regime change, as indicated by a three-point change in the Polity score. Regression results reveal no evidence that leaders in older democracies reap consistently higher support from the public or enjoy larger or smaller honeymoons (see Table A6 in the appendix). Age of democracy neither conditions the level or dynamics of public support for the executive nor does it alter the differences between presidential and parliamentary regimes.

We also consider whether differences across regime types dissipate when we compare elected and non-elected chief executives. We have argued that differences in approval dynamics between presidential and parliamentary regime are shaped by the ties voters and leaders cultivate during the selection of the chief executive. If this is the case, we might expect distinct approval patterns for executives who reach power via elections and those who do not—whether presidents who assume the office after the death, resignation, or impeachment of their predecessor or prime ministers who follow a resignation or a change in party leadership.

Once again, our expectations differ by regime type. If indirect elections reduce personalism, then approval patterns for prime ministers should vary based on whether they came to power through elections or not. In fact, non-electoral changes in leadership are quite common in parliamentary systems and considered fairly routine by voters. By contrast, high levels of personalism in presidential systems should make differences between directly elected and non-elected presidents more acute. Moreover, given fixed terms, removing a president from office prior to the end of his or her term is rare and can lead to a political or national crisis, facilitating bandwagoning. Indeed, while 16% of parliamentary governments in our sample are headed by leaders not selected following an election, in presidential systems the rate is only 6%.

To assess these implications, we plot local regression curves for presidents and prime minister, separating them by how they came to power. Note that, unlike previous plots, the x-axis here is not time between elections, since non-elected leaders do not arrive in office following elections, but rather marks the time between taking office and leaving it. Results displayed in Figure 5 are consistent with our expectations. In parliamentary regimes, prime ministers whose ascent to power does not follow a general election have approval patterns similar to those who do. Near the end of the term, however, unelected prime ministers seem to maintain slightly higher approval ratings. This might reflect redoubled campaign efforts to legitimize and maintain their power in the upcoming election.

Figure 5.

Mean-centered approval dynamics by regime type and path to power. Note. Solid lines display report mean-centered approval with 95% confidence intervals for elected executives; dashed lines represent the same for unelected executives. Horizontal axes display proportion of time between executive enters and exists office.

In contrast, differences in approval patterns between elected and unelected presidents are stark. Unelected presidents experience a large honeymoon bounce and a quick collapse before a modest pre-election bounce. The even higher degree of volatility in these cases further suggests that the dynamics of presidential approval are unconstrained by the institutional buffers of parliamentarism. A sharp difference in how unelected leaders in presidential and parliamentary systems are evaluated provides additional evidence that the plebiscitarian link to voters and less intimate leader-party relations encourage the larger honeymoons, more precipitous declines, and the overall greater levels of approval volatility we find in presidential countries.

Conclusion

Over three decades ago, Juan Linz pondered how the choice of constitutional regimes might make the wave of new democracies more sustainable. Numerous studies have assessed many of his arguments, but until now none have rigorously examined Linz’s intuitions regarding executive approval—a proxy for political capital and a leading indicator for instability. Our study illuminates how regime type structures politics by developing and testing theoretical arguments linking differences in executive approval patterns to distinguishing features of presidential and parliamentary regimes.

Our findings underscore the indirect effects of regime type via the incentive structure that characterizes each set of institutions. Direct elections that bring new leaders to power—more frequent under presidentialism lead to greater personalism; the dependence of prime ministers on partisan majorities—a signature of parliamentarism—contributes to party system institutionalization. Notably, these features of political competition have downstream effects on executive approval: presidential approval tends to be higher at the outset of a term but also more volatile, declining more steeply than approval for prime ministers. Personalism and party system institutionalization remain robust predictors of cross-regime differences even after controlling for a range of potential confounders. In short, patterns of public approval are closely linked to differences in methods for selecting and maintaining the executive.

By underlining the importance of these fundamental, regime-defining differences, our findings contribute to long-standing debates about the relative merits of presidential and parliamentary systems—particularly the tradeoff between democratic responsiveness and stability. Presidentialism’s critics decry its poor record of democratic stability (Linz, 1994; disputed, among others, by Hicken et al., 2022); its defenders argue it bolsters responsiveness by endowing an easily identifiable political executive with a mandate (Mainwaring & Shugart, 1997). Parliamentarism, by contrast, is less responsive but compensates with greater stability (Powell, 2000). Our results reflect this tradeoff. Although direct elections grant robust honeymoons for presidents, weaker partisan attachments make public support more fleeting. Conversely, prime ministers have more modest honeymoons, but their support shrinks more gradually over time. Overall, we provide new evidence that, relative to parliamentarism, executive approval dynamics under presidentialism at once reflect greater instability and provide for greater inter-election accountability.

We also demonstrate that as parliamentary regimes assume key characteristics of presidentialism, their approval dynamics approximate those of presidential regimes. Though high levels of personalism and electoral volatility are not typical of parliamentary systems, prime minister approval dynamics in such contexts begin to resemble those under presidentialism. Thus, if the trends that encourage personalism across all democracies continue or accelerate (Frantz et al., 2021; Mezey, 2013; Poguntke & Webb, 2018), we can expect even more candidate-centric campaigns and less-institutionalized party systems in parliamentary democracies and, with these changes, more presidential-like approval patterns. Although we identify factors that make approval dynamics in parliamentary regimes look more “presidential,” we do not uncover factors that make approval in presidential regimes look more “parliamentarian.” Therefore, our findings imply that basic institutional differences across regimes should limit the convergence of approval patterns.

Additionally, our study reveals new insights into what drives approval dynamics. For instance, while honeymoons differ across regimes as expected, the differential is dwarfed by the gap between first-time and re-elected executives in either regime. We, similarly, identify an end-of-term bounce for executives (Carlin et al., 2018; Müller & Louwerse, 2020; Stimson, 1976) but detect no regime differences. We do, however, reveal major regime differences for leaders who assume office outside the “normal” electoral process. The transfer of power to a new prime minister between elections, which happens for around 1 in 6 of the prime ministers in our sample, rarely precipitates a break with “normal” approval dynamics. In contrast, presidents assume office outside elections infrequently (around 1 in 17 in our sample) and typically do so in times of crisis; their initial approval tends to be higher but also declines more sharply, probably because they lack a strong electoral or partisan mandate. So, while we do not dispute the centrality of elections in shaping approval dynamics, our analyses suggest areas open to further theoretical refinement.

Finally, our study points to some important questions for future research. Does greater approval volatility in presidential systems reflect a stronger transmission between events and executive responsiveness (Hellwig & Samuels, 2008)? Can less volatile approval patterns help explain why parliamentarism operates as more of a coordination device, improving a variety of political and socio-economic outcomes (Gerring et al., 2009; see also McManus & Ozkan, 2018)? Answers to these questions can further inform our broader understanding of regime types and governance.

Supplemental Material

Supplemental Material - Executive Approval Dynamics in Presidential and Parliamentary Democratic Regimes

Supplemental Material for Executive Approval Dynamics in Presidential and Parliamentary Democratic Regimes by Cecilia Martínez-Gallardo, Gregory J. Love, Jonathan Hartlyn, Ryan E. Carlin, Timothy Hellwig, and Matthew M. Singer in Comparative Political Studies

Footnotes

Acknowledgments

The authors would like to thank Juan Bogliaccini, Aníbal Pérez-Liñan, Rosario Queirolo, Sarah Shair-Rosenfeld, Mariano Torcal and participants at the Universidad Nacional de San Martín, Universidad de la República, APSA, and WAPOR-LATAM for their comments.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the Center for Human Rigths and democracy at Georgia State University (Seed Grant), Reckford Professorship at UNC-Chapel Hill.

ORCID iDs

Cecilia Martínez-Gallardo

Gregory J. Love

Jonathan Hartlyn

Ryan E. Carlin

Timothy Hellwig

Matthew M. Singer

Data Availability Statement

Authors will deposit research data in a data repository that issues a persistent identifier (DOI) and provide link(s) in the final manuscript. Replication data and code (Love et al. (2024)) is deposited in the Comparative Political Studies Dataverse at the Harvard Dataverse Repository. DOI: 10.7910/DVN/VNOSXU.

Supplemental Material

Supplemental material for this article is available online.

Notes

Author Biographies

Cecilia Martínez-Gallardo is an Associate Professor of Political Science at the University of North Carolina at Chapel Hill. Her research interests are in the study of executive politics in Latin America, including the dynamics of government formation and executive approval. She is a core member of the Executive Approval Project and the Chapel Hill Expert Survey-Latin America.

Gregory J. Love is Professor of Political Science at the University of Mississippi. His research is focused on the intersection of political institutions and mass behavior, often in Latin America. He is also a core member of the Executive Approval Project.

Jonathan Hartlyn is the Kenneth J. Reckford Professor of Political Science at the University of North Carolina at Chapel Hill. His research and teaching interests are in comparative politics, especially of Latin America, with particular focus on questions of democratization, political behavior, political institutions, and state-society relations. He is a core member of the Executive Approval Project and the Chapel Hill Expert Survey-Latin America.

Ryan E. Carlin is Professor of Political Science and Director of the Center for Human Rights at Georgia State University. His research foci include comparative political behavior, public opinion, and political economy. He is a core member of the Executive Approval Project.

Timothy Hellwig is Professor of Political Science at the University at Buffalo, SUNY. His interests are in policy responsiveness, political parties, and the electoral consequences of economic change. He is co-editor of Economics and Politics Revisted (2023), and coauthor of Democracy Under Siege: Parties, Voters, and Elections After the Great Recession (2020). He is a core team member of the Executive Approval Project.

Matthew Singer is the Alan R. Bennett Honors Professor of Political Science at the University of Connecticut. His research focuses on political accountability. He is a core team member of the Executive Approval Project.

References

Aguiar-Conraria

Fernandes

Magalhães

(2023). The economy and executive approval in a semi-presidential regime: The case of Portugal. In Hellwig

Singer

M. M.

(Eds.), Economics and politics revisited: Executive approval and the new calculus of support (pp. 108–130). Oxford University Press.

Anderson

(1995a). Party systems and the dynamics of government support: Britain and Germany, 1960–1990. European Journal of Political Research, 27(1), 93–118. https://doi.org/10.1111/j.1475-6765.1995.tb00631.x

Anderson

(1995b). The dynamics of public support for coalition governments. Comparative Political Studies, 28(3), 350–383. https://doi.org/10.1177/0010414095028003002

Becher

Christiansen

F. J.

(2015). Dissolution threats and legislative bargaining. American Journal of Political Science, 59(3), 641–655.

Bellucci

Memoli

(2023). Government popularity in Italy: From valence to positional economic accountability. In Hellwig

Singer

M. M.

(Eds.), Economics and politics revisited: Executive approval and the new calculus of support (pp. 155–180). Oxford University Press.

Bormann

N. C.

Golder

(2013). Democratic electoral systems around the world, 1946–2011. Electoral Studies, 32(2), 360–369.

Brace

Hinckley

(1992). Follow the leader: Opinion polls and the modern presidents. Basic Books.

Calvo

(2007). The responsive legislature: Public opinion and law making in a highly disciplined legislature. British Journal of Political Science, 37(2), 263–280. https://doi.org/10.1017/s0007123407000130

Campus

Pasquino

(2006). Leadership in Italy: The changing role of leaders in elections and in government. Journal of Contemporary European Studies, 14(1), 25–40. https://doi.org/10.1080/14782800600617912

10.

Carey

J. M.

Shugart

M. S.

(1995). Incentives to cultivate a personal vote: A rank ordering of electoral formulas. Electoral Studies, 14(4), 417–439. https://doi.org/10.1016/0261-3794(94)00035-2

11.

Carlin

R. E.

Carreras

Love

G. J.

(2020). Presidents’ sex and popularity: Baselines, dynamics, and policy performance. British Journal of Political Science, 50(4), 1359–1379. https://doi.org/10.1017/s0007123418000364

12.

Carlin

R. E.

Hartlyn

Hellwig

Love

G. J.

Martínez-Gallardo

Singer

M. M.

(2018). Public support for Latin American presidents: The cyclical model in comparative perspective. Research and Politics, 5(3), 205316801878769. https://doi.org/10.1177/2053168018787690

13.

Carlin

R. E.

Hartlyn

Hellwig

Love

G. J.

Martınez-Gallardo

Singer

(2023). The Executive Approval Database. Conceptual and Empirical Bases. In Hellwig

Singer

M. M.

(Eds.), Economics and politics revisited: Executive approval and the new calculus of support (pp. 155–180). Oxford University Press.

14.

Carlin

R. E.

Martínez-Gallardo

(2019). Outliers of presidential approval: Dynamics, levels, and rates. Revista Latinoamericana de Opinión Pública, 8(2), 7–27.

15.

Carlin

R. E.

Martinez-Gallardo

Hartlyn

(2012). Executive approval under alternative democratic regime types. In Chalmers

Mainwaring

(Eds.), Problems confronting contemporary democracies: Essays in honor of alfred stepan (pp. 203–226). University of Notre Dame Press.

16.

Castro Cornejo

Beltrán

Ley

Galicia

(2022). From election day to presidential approval: Partisanship and the honeymoon period in Mexico. Electoral Studies, 75(1), 102438. https://doi.org/10.1016/j.electstud.2022.102438

17.

Cerda

Vergara

(2022). Economic growth and political approval ratings: Evidence from Latin America. Political Behavior, 45(4), 1735–1758. https://doi.org/10.1007/s11109-022-09798-y

18.

Cheibub

J. A.

(2007). Presidentialism, parliamentarism, and democracy. Cambridge University Press.

19.

Cheibub

J. A.

Elkins

Ginsburg

(2014). Beyond presidentialism and parliamentarism. British Journal of Political Science, 44(3), 515–544. https://doi.org/10.1017/s000712341300032x

20.

Cheibub

J. A.

Przeworski

Saiegh

S. M.

(2004). Government coalitions and legislative success under presidentialism and parliamentarism. British Journal of Political Science, 34(4), 565–587. https://doi.org/10.1017/s0007123404000195

21.

Corrales

(2016). Can anyone stop the president? Power asymmetries and term limits in Latin America, 1984–2016. Latin American Politics and Society, 58(2), 3–25. https://doi.org/10.1111/j.1548-2456.2016.00308.x

22.

Cross

Young

(2015). Personalization of campaigns in an SMP system: The Canadian case. Electoral Studies, 39, 306–315.

23.

Dowding

(2013). The prime ministerialisation of the British prime minister. Parliamentary Affairs, 66(3), 617–635. https://doi.org/10.1093/pa/gss007

24.

Elgie

(2011). Semi-presidentialism: Subtypes and democratic performance. Oxford University Press.

25.

Frantz

Kendall-Taylor

Wright

(2022). Personalist ruling parties in democracies. Democratization, 29(5), 918–938.

26.

Frantz

Kendall-Taylor

Nietsche

Wright

(2021). How personalist politics is changing democracies. Journal of Democracy, 32(3), 94–108. https://doi.org/10.1353/jod.2021.0036

27.

Fumagalli

Narciso

(2012). Political institutions, voter turnout, and policy outcomes. European Journal of Political Economy, 28(2), 162–173. https://doi.org/10.1016/j.ejpoleco.2011.10.002

28.

Gerring

Thacker

S. C.

Moreno

(2009). Are parliamentary systems better? Comparative Political Studies, 42(3), 327–359. https://doi.org/10.1177/0010414008325573

29.

Gronke

Newman

(2003). FDR to Clinton, Mueller to ? A field essay on presidential approval. Political Research Quarterly, 56(4), 501–512. https://doi.org/10.1177/106591290305600411

30.

Grossman

Guinaudeau

(2023). Cost of ruling above all else: Explaining presidential popularity in France. In Hellwig

Singer

M. M.

(Eds.), Economics and politics revisited: Executive approval and the new calculus of support (pp. 155–180). Oxford University Press.

31.

Hellwig

Samuels

(2008). Electoral accountability and the variety of democratic regimes. British Journal of Political Science, 38(1), 65–90. https://doi.org/10.1017/s0007123408000045

32.

Hellwig

Singer

M. M.

(Eds.). (2023). Economics and politics revisited: Executive approval and the new calculus of support (pp. 181–203). Oxford University Press.

33.

Herrera

Ordoñez

Trebesch

(2020). Political booms, financial crises. Journal of Political Economy, 128(2), 507–543. https://doi.org/10.1086/704544

34.

Hicken

Baltz

Vasselai

(2022). Political institutions and democracy. In Coppedge

Edgell

A. B.

Knutsen

C. H.

Lindberg

S. I.

(Eds.), Why democracies develop and decline (pp. 161–184). Cambridge University Press.

35.

Hochstetler

Samuels

(2011). Crisis and rapid reequilibration: The consequences of presidential challenge and failure in Latin America. Comparative Politics, 43(2), 127–145. https://doi.org/10.5129/001041511793931834

36.

Holtz-Bacha Chrisna Ana

Ines Langer

Susanne

Merkle.

(2014). “The personalizaon of polics incomparave perspecve: Campaign coverage in Germany and the United Kingdom.” European Journal of Communication 29.2: 153–170.

37.

Horowitz

D. L.

(1990). Presidents vs. parliaments: Comparing democratic systems. Journal of Democracy, 1(4), 73–79. https://doi.org/10.1353/jod.1990.0056

38.

Johnston

(1999). Business cycles, political cycles and the popularity of Canadian governments, 1974–1998. Canadian Journal of Political Science, 32(3), 499–520. https://doi.org/10.1017/s0008423900013949

39.

Kam

Indriđason

(2005). The timing of cabinet reshuffles in five Westminster parliamentary systems. Legislative Studies Quarterly, 30(3), 327–363. https://doi.org/10.3162/036298005x201581

40.

Karp

J. A.

Banducci

S. A.

(2008). Political efficacy and participation in twenty-seven democracies: How electoral systems shape political behaviour. British journal of political science, 38(2), 311–334.

41.

Kayser

M. A.

(2005). Who surfs, who manipulates? The determinants of opportunistic election timing and electorally motivated economic intervention. American Political Science Review, 99(1), 17–27. https://doi.org/10.1017/s0003055405051464

42.

Kayser

M. A.

Wlezien

(2011). Performance pressure: Patterns of partisanship and the economic vote. European Journal of Political Research, 50(3), 365–394. https://doi.org/10.1111/j.1475-6765.2010.01934.x

43.

König

P. D.

Wenzelburger

(2017). Honeymoon in the crisis: A comparative analysis of the strategic timing of austerity policies and their effect on government popularity in three countries. Comparative European Politics, 15(6), 991–1015. https://doi.org/10.1057/cep.2016.1

44.

Krauss

E. S.

Nyblade

(2005). ‘Presidentialization’ in Japan? The prime minister, media and elections in Japan. British Journal of Political Science, 35(2), 357–368. https://doi.org/10.1017/s0007123405000190

45.

Kriesi

(2012). Personalization of national election campaigns. Party Politics, 18(6), 825–844. https://doi.org/10.1177/1354068810389643

46.

Langer Ana

Ines

Iñaki Sagarzazu . (2018). “Bring back the party: personalisaon, the media and coalionpolitics.” West European Politics 41.2: 472–495.

47.

Laver

Shepsle

K. A.

(1998). Events, equilibria, and government survival. American Journal of Political Science, 42(1), 28–54. https://doi.org/10.2307/2991746

48.

Linz

J. J.

(1990). The perils of presidentialism. Journal of Democracy, 1(1), 51–69. https://doi.org/10.1353/jod.2005.0026

49.

Linz

J. J.

(1994). Presidential or parliamentary democracy: Does it make a difference? In Linz

J. J.

Valenzuela

(Eds.), The failure of presidential democracy (pp. 3–87). Johns Hopkins University Press.

50.

Love

G. J.

Martinez-Gallardo

Hartlyn

Carlin

R. E.

Hellwig

Singer

M. M.

(2024). Replication data for: Executive approval dynamics in presidential and parliamentary democratic regimes. Harvard Dataverse, V1. https://doi.org/10.7910/DVN/VNOSXU

51.

Lupu

(2015). Party polarization and mass partisanship: A comparative perspective. Political Behavior, 37(2), 331–356. https://doi.org/10.1007/s11109-014-9279-z

52.

Mainwaring

(2018). Party system institutionalization, predictability, and democracy. In Mainwaring

(Ed.), Party systems in Latin America (pp. 71–101). Cambridge University Press.

53.

Mainwaring

Gervasoni

España-Najera

(2017). Extra- and within-system electoral volatility. Party Politics, 23(6), 623–635. https://doi.org/10.1177/1354068815625229

54.

Mainwaring

Shugart

M. S.

(1997). Juan Linz, presidentialism, and democracy: A critical appraisal. Comparative Politics, 29(4), 449–471. https://doi.org/10.2307/422014

55.

Mainwaring

Torcal

(2006). Party system institutionalization and party system theory after the third wave of democratization. In Katz

R. S.

Crotty

W. J.

(Eds.), Handbook of party politics (pp. 204–227). Sage.

56.

Mainwaring

Zoco

(2007). Political sequences and the stabilization of interparty competition: Electoral volatility in old and new democracies. Party Politics, 13(2), 155–178. https://doi.org/10.1177/1354068807073852

57.

Marsteintredet

(2014). Explaining variation of executive instability in presidential regimes: Presidential interruptions in Latin America. International Political Science Review, 35(2), 173–194. https://doi.org/10.1177/0192512112459548

58.

Martínez-Gallardo

(2014). Designing cabinets: Presidential politics and ministerial instability. Journal of Politics in Latin America, 6(2), 3–38.

59.

McAllister

(2007). The personalization of politics. The Oxford handbook of political behavior. Oxford University Press.

60.

McManus

Ozkan

F. G.

(2018). Who does better for the economy? Presidents versus parliamentary democracies. Public Choice, 176(3–4), 361–387. https://doi.org/10.1007/s11127-018-0552-2

61.

Mezey

M. L.

(2013). Presidentialism: Power in comparative perspective. Lynne Rienner.

62.

Mueller

(1970). Presidential popularity from Truman to Johnson. American Political Science Review, 64(1), 18–34. https://doi.org/10.2307/1955610

63.

Mueller

(1973). War, presidents, and public opinion. John Wiley and Sons.

64.

Müller

Louwerse

(2020). The electoral cycle effect in parliamentary democracies. Political Science Research and Methods, 8(4), 795–802. https://doi.org/10.1017/psrm.2018.47

65.

Nannestad

Paldam

(2002). The cost of ruling: A foundation stone for two theories. In Dorussen

Taylor

(Eds.), Economic voting (pp. 17–44). Routledge.

66.

Pérez-Liñán

(2007). Presidential impeachment and the new political instability in Latin America. Cambridge University Press.

67.

Poguntke

Webb

P. D.

(2005). The presidentialization of politics: A comparative study of modern democracies. Oxford University Press.

68.

Poguntke

Webb

P. D.

(2018). Presidentialization, personalization and populism: The hollowing out of party government. In Cross

W. P.

Katz

R. S.

Pruysers

(Eds.), The personalization of democratic politics and the challenge for political parties (pp. 181–196). Rowman and Littlefield.

69.

Powell

G. B.

Jr. (2000). Elections as instruments of democracy. Yale University Press.

70.

Rahat

Kenig

(2018). From party politics to personalized politics? Party change and political personalization in democracies. Oxford University Press.

71.

Rahat

Sheafer

(2007). The personalization(s) of politics: Israel, 1949–2003. Political Communication, 24(1), 65–80. https://doi.org/10.1080/10584600601128739

72.

Raile

E. D.

Pereira

Power

T. J.

(2011). The executive toolbox: Building legislative support in a multiparty presidential regime. Political Research Quarterly, 64(2), 323–334. https://doi.org/10.1177/1065912909355711

73.

Reyes-Housholder

(2020). A theory of gender’s role on presidential approval ratings in corrupt times. Political Research Quarterly, 73(3), 540–555. https://doi.org/10.1177/1065912919838626

74.

Romero-Vidal

OrriolsRiera

Ll.P.

(2023). Moving beyond the economy: Executive approval in Spain. In Hellwig

Singer

M. M.

(Eds.), Economics and politics revisited: Executive approval and the new calculus of support (pp. 181–203). Oxford University Press.

75.

Samuels

(2002). Presidentialized parties: The separation of powers and party organization and behavior. Comparative Political Studies, 35(4), 461–483. https://doi.org/10.1177/0010414002035004004

76.

Samuels

Hellwig

(2010). Elections and accountability for the economy: A conceptual and empirical reassessment. Journal of Elections, Public Opinion, and Parties, 20(4), 393–419. https://doi.org/10.1080/17457289.2010.490723

77.

Samuels

D. J.

Shugart

M. S.

(2010). Presidents, parties, and prime ministers: How the separation of powers affects party organization and behavior. Cambridge University Press.

78.

Schleiter

Tavits

(2016). The electoral benefits of opportunistic election timing. The Journal of Politics, 78(3), 836–850. https://doi.org/10.1086/685447

79.

Stimson

J. A.

(1976). Public support for American presidents: A cyclical model. Public Opinion Quarterly, 40(1), 1–21. https://doi.org/10.1086/268264

80.

Stimson

J. A.

(1991). Public opinion in America: Moods, cycles, and swings. Westview Press.

81.

Stokes

S. C.

(2001). Mandates and democracy: Neoliberalism by surprise in Latin America. Cambridge University Press.

82.

Tavits

(2009). Direct presidential elections and turnout in parliamentary contests. Political Research Quarterly, 62(1), 42–54. https://doi.org/10.1177/1065912908317026

83.

Walther

Hellström

(2019). The verdict in the polls: How government stability is affected by popular support. West European Politics, 42(3), 593–617. https://doi.org/10.1080/01402382.2018.1490598

84.

Webb

Poguntke

(2013). The presidentialisation of politics thesis defended. Parliamentary Affairs, 66(3), 646–654. https://doi.org/10.1093/pa/gss059

85.

Wlezien

(2017). Policy (mis) representation and the cost of ruling: US presidential elections in comparative perspective. Comparative Political Studies, 50(6), 711–738. https://doi.org/10.1177/0010414015626446

86.

Yantek

(1985). Government popularity in Great Britain under conditions of economic decline. Political Studies, 33(3), 467–483. https://doi.org/10.1111/j.1467-9248.1985.tb01156.x

Supplementary Material

Please find the following supplemental material available below.

For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.

For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.

0.00 MB

0.53 MB