Governing for Growth? Economic Performance and Finance Minister Survival in Democracies and Autocracies

Abstract

Are the highest-ranking public officials rewarded for good performance? Using the case of finance ministers, we argue that this depends on the type of regime. Democratic leaders have incentives to retain high-performing finance ministers because they are accountable to the population through free and fair elections. In the absence of free and fair elections, autocrats face opposing incentives. While growth stabilizes the regime, a high-performing minister of finance can also be a potential rival. Moreover, autocratic finance ministers will, to a higher degree than democratic, be judged on their ability to enrich the leader rather than creating growth. As a result, autocrats reward finance ministers less for economic growth than democrats. Using data on over 2,800 finance ministers in the period 1966–2021, we demonstrate a close link between economic growth and the retention of finance ministers in democracies but not in autocracies. The findings indicate that democracies are better at fostering and rewarding competence at the highest level of political office, speaking to the merits of democracy relative to autocracy.

Introduction

For their nations to thrive, it is essential that politicians and high-ranking public officials govern in the people’s interest. Among them, finance ministers play a particularly powerful role. They shape the policy agenda, manage the government budget, and oversee economic policy-making (Armstrong et al. 2024; Blondel 1991a; Jochimsen and Thomasius 2014). Often described as the second most important person in government and the person responsible for the economy (Alexiadou et al. 2021), finance ministers are an ideal case for examining whether top officials are rewarded for good performance. Thus, this article investigates whether finance ministers are rewarded when the economy booms and held accountable when it plunges into crisis.

We show that the answer differs across regime types. In democracies, economic performance strongly predicts finance minister retention; in autocracies, it does not. We argue that this is due to key institutional differences. In democracies, but not in autocracies, the main threat leaders face is competitive elections that allow citizens to hold officials accountable. This difference shapes the incentives of leaders, giving those in democracies stronger reasons to reward or dismiss high-ranking officials based on performance.

Democratic leaders must win reelection to stay in power, making them accountable to the broader electorate. This creates a link between economic performance and finance minister retention for three reasons. First, a successful finance minister can boost public support (Buchs and Soguel 2022), making them an electoral asset. Second, strong economic performance signals ministerial competence, giving leaders an incentive to retain them to secure future gains (Huber and Martinez-Gallardo 2008). Third, replacing a poorly performing minister allows leaders to demonstrate accountability to voters and signal responsibility to investors (Alexiadou et al. 2021; Dewan and Dowding 2005).

In contrast, autocrats face opposing incentives. Without free and fair elections (Dahl 2008; Schumpeter 2013), the main threat to their rule comes from insiders seeking to outmaneuver them. Therefore, autocrats must be wary of subordinates who can act as potential competitors (Egorov and Sonin 2011; Svolik 2012; Zakharov 2016), especially those who can attract attention and build alternative power bases (Lee and Schuler 2020). Thus, if a minister of finance does their job “too well” and delivers high growth, the dictator may come to view them as a threat. ¹

Autocrats may also care less about economic growth than other priorities, leading to differing roles for finance ministers across regime types. Without electoral accountability, dictators often prioritize economic policies that reduce overall growth but benefit key allies, in order to maintain their ruling coalitions (De Mesquita et al. 2005). Moreover, some dictators hold views that contradict standard economic theory and have the power to enforce them, resulting in finance ministers prioritizing implementing the dictator’s agenda—even when it undermines growth.

However, even without popular suffrage, dictators have some incentives to retain a well-performing finance minister. While they must fear coups, they also face the risk of revolution or forced democratization (Svolik 2012). Economic growth can bolster public support, increase taxable income, and reduce the likelihood of regime collapse (Djuve et al. 2020; Wintrobe 2000).

As a result, being a finance minister in an autocracy is a “damned if you do, damned if you don’t” job; removal is just as likely whether the economy is booming or in crisis.

To test this theory, we combine data on finance minister survival in office from the WhoGov dataset (Nyrup and Bramwell 2020) with economic growth data from the Penn World Tables (Feenstra et al. 2015) and other datasets.

Using deviations from 10-year growth-trends for individual countries, we separate years into whether the economy is booming, normal, or in crisis (Brambor and Lindvall 2014; Krishnarajan 2019). To isolate the effect of economic performance on the removal of the finance minister from office from its broader effects on government stability, we exclude years during which the government leader changes from our analysis. We find that democratic finance ministers experience a significant decline in the likelihood of being removed when the economy is booming. They experience a decline of 4–5 percentage points when the economy is normal compared to when it is in a crisis and a decline of 10–11 percentage points when the economy is booming compared to a crisis year. Given that, on average, 30.7 percent of finance ministers are removed per year in democracies, these effects are substantial. In contrast, we find no association between economic performance and the likelihood of finance minister removal in autocracies.

We show that these results are robust across a wide range of robustness tests addressing endogeneity concerns, alternative estimators, alternative measures, and placebo tests. We further find that the relationship is stronger in parliamentary compared to presidential democracies and that there are no systematic differences between autocratic subtypes.

Our article is the first to examine how economic performance affects finance minister turnover across regime types, ² and thereby makes several key contributions. First, it makes important contributions to the literature on comparing democracies and autocracies (Gerring et al. 2022). Only a fraction of this literature focuses on the government, and these studies largely treat the government as a unit instead of being made up of individuals. Second, there is growing literature that focuses on cabinet reshuffles and ministers (Huber and Martinez-Gallardo 2008). However, this literature has an empirical bias towards rich democracies and rarely compares across regimes, probably because the appropriate data was missing until recently (Helms and Vercesi 2022). This study shows that there are important differences in when autocracies and democracies terminate their most important ministers, highlighting the value of looking at countries outside the traditional scope of this literature. Third, there is large political economy literature on political responses to economic crises, including some that examine whether the personnel in charge is replaced and who is appointed in their place (Alexiadou and Gunaydin 2019; Armstrong et al. 2024). We add a new angle to this literature by looking at responses across regime types.

Our findings also have normative implications. One measure of regime quality is whether competent individuals are placed—and retained—in key positions (Fukuyama 2013; Rothstein 2011; Weber 1978). We argue and show that democratic institutions more effectively promote and retain high-performing finance ministers. This meritocratic edge helps explain why democracies are more likely to deliver prosperity and growth (Acemoglu et al. 2019; Knutsen 2021; Persson and Sjöstedt 2012).

Regime Types, Macro-Outcomes, and the Hiring and Firing of Public Officials

Democracies and autocracies are widely understood to produce different macro-level outcomes, from economic growth to human development (Gerring et al. 2012; Przeworski et al. 2000; Przeworski and Limongi 1993; Ross 2006). While early research produced mixed results, more recent studies find that democracy generally improves well-being and growth (e.g., Acemoglu et al. 2019; Gerring et al. 2022; Knutsen 2021).

But what explains these differences? One promising yet underexplored mechanism lies in how regimes manage elite personnel, specifically, whether competent ministers are more likely to be rewarded or retained in some regimes than others. Despite its importance, this question has not yet been studied directly. Below, we outline the main literature that informs our approach and to which we contribute.

First, a body of research has examined the effects of democracy on bureaucratic quality. Classical accounts warned that democratic competition might undermine professionalism by incentivizing clientelism, corruption, or short-term thinking (Fukuyama 2014; Weber 1978). More recent work suggests the opposite: that democratic institutions can foster a more meritocratic and effective public sector through transparency, accountability, and inclusiveness (Andersen and Cornell 2022; Bäck and Hadenius 2008; Charron and Lapuente 2010). However, this literature has focused on the bureaucracy rather than the internal dynamics of cabinets, leaving open how democratic accountability shapes the careers of political elites within government.

A second and growing literature takes ministers themselves as units of analysis. Researchers have examined patterns of ministerial selection and the characteristics of appointees, including gender (Heinzel et al. 2024; Krook and O’Brien 2012; Nyrup et al. 2023), education and professional background (Alexiadou 2022; Blondel 1991b), and social origins (Ricart-Huguet 2021; Smith and Martin 2017). This work emphasizes that governments are composed of individuals who can be replaced without regime change (Huber and Martinez-Gallardo 2008). However, much of this literature focuses on selection, not retention. Where retention is studied, it is so mostly within the context of larger cabinet reshuffles, and in the context of Western democracies (Helms and Vercesi 2022), even though some scholars have begun to look into how autocrats manage individuals in their cabinets (Bokobza et al. 2022; Goldring and Matthews 2021; Lee and Schuler 2020; Woldense 2022).

Third, there is rich literature on government and leadership turnover, showing that poor economic performance increases the risk of losing office (Djuve et al. 2020; Lewis-Beck and Paldam 2000; Lucardi 2019). However, these studies typically examine the survival of the regime or leader, rather than how performance shapes decisions at the ministerial level. A recent exception is Schmid and Aaskoven (2025), who show that fiscal deficits affect the survival of finance ministers in democracies but not in autocracies. Their argument centers on diverging regime preferences regarding government borrowing and the different supply and demand constraints autocracies and democracies face in credit markets. Our study, by contrast, foregrounds electoral accountability: we ask whether democratic leaders, facing competitive elections, are more likely than autocrats to reward or punish finance ministers based on overall economic performance, an outcome that, unlike fiscal deficits, is almost unambiguously desirable, making it a clearer basis for political accountability and a broader test of whether democratic institutions foster competence at the highest levels of government.

Taken together, these literature studies offer valuable insights into how leaders are selected, how they govern, and when they fall. But they leave a key question unanswered: how do regimes differ in how they retain or dismiss ministers in response to performance?

The Importance of Ministers of Finance

We fill this gap by studying whether finance ministers’ survival in office depends on economic performance—and whether this relationship varies between regime types. Finance ministers are well suited for examining questions of ministerial accountability and its potential role in explaining regime-level differences in macro-outcomes. Their portfolio directly targets one of the most studied outcomes in comparative politics, economic growth, and are generally considered influential enough to shape both economic policy and its outcomes (Armstrong et al. 2024; Blondel 1991a, 1991b). Thus, whether the incentives to deliver economic growth differ across regime types may well be part of why democracies and autocracies tend to diverge economically.

Although their exact responsibilities vary from country to country, finance ministers generally are in charge of government finances, economic policy, and financial regulation and are therefore the prime person responsible for handling the economy. Their importance is emphasized by the fact that the future leader often has a background as a finance minister (Blondel 1991b).

Whether they can influence economic performance is a more debated topic. Although finance ministers cannot fully control economic outcomes, they can plausibly shape them through multiple channels.

First, they can actively shape domestic policy making through economic, fiscal, and trade policies, which directly affect public and private investment, trade barriers, the competitiveness of domestic industries, and other policy outcomes that can affect long-term and short-term economic performance. This is corroborated by systematic studies that show the influence of finance ministers and their personal characteristics on outcomes such as fiscal deficits, government revenue and expenditure, and accounting practices (Alexiadou et al. 2021; Armstrong et al. 2022; Buchs and Soguel 2022; Chatagny 2015; Clémenceau and Soguel 2017; Jochimsen and Thomasius 2014; Moessinger 2014).

Second, finance ministers frequently represent their countries in international negotiations, where personal skill and networks can make a difference—especially in times of crisis (Kray and Haselhuhn 2007).

Third, finance ministers play a symbolic role. Their perceived competence (or lack thereof) can affect the credibility of a government’s economic policy and influence markets or investors (Alexiadou et al. 2021). As such, retaining or removing them in response to economic outcomes can serve as a meaningful political signal.

Accordingly, it is plausible that different incentive structures for finance ministers may lead to different performances across regime types and that this, in turn, can influence their larger economic trajectories.

Economic Growth and the Selection of Finance Ministers Across Regime Types

Regimes are complex systems with many differences. One key factor shaping leaders’ incentives when selecting top officials is whether they are accountable to citizens through elections. We conceptualize two ideal-type regimes: (1) democracies with free and fair elections, universal suffrage, freedom of expression, and a functioning opposition; and (2) autocracies, where elections are either absent, suffrage is limited, or contests are heavily skewed in favor of the incumbent. ³

To develop our argument, we assume, following De Mesquita et al. (2005), that leaders’ primary goal is to stay in power. This is a prerequisite for advancing secondary goals, such as enacting policies or enjoying the perks of office. Leaders must therefore manage their cabinets effectively, adapting to their institutional environment to maximize political survival.

Because electoral accountability differs across regime types, leaders face distinct incentives when responding to external pressures—particularly economic performance. In democracies, leaders must anticipate voter reactions to poor economic outcomes and are thus more likely to act in ways that signal competence, including selecting or replacing ministers accordingly. In autocracies, where leaders are insulated from popular pressures, such accountability is weaker or absent.

These divergent incentive structures have implications for key cabinet appointments—especially the finance minister, whose portfolio is closely tied to economic performance. Consequently, different political logics can shape the appointment, retention, and effectiveness of finance ministers across regime types, with broader consequences for economic outcomes.

In the following, we discuss in detail the different political logics of democracies and autocracies and how they lead to diverging expectations regarding the impact of economic growth on finance minister turnover in democracies and autocracies.

Economic Growth and Finance Ministers in Democracies

Democratic leaders must win competitive elections to stay in power. This gives them strong incentives to reward good economic performance and punish failure, especially by retaining or dismissing the minister of finance. We identify three reasons why democratic accountability ties the turnover of finance ministers to economic outcomes.

First, voters evaluate governments based on economic performance (Lewis-Beck and Paldam 2000), and the finance minister is often viewed as responsible for the economy. Even though growth is influenced by many factors, strong performance signals competence, which democratic leaders want to retain to improve future electoral outcomes. Voters tend to reward perceived competence and trust governments more when ministers are seen as capable (Devine et al. 2024).

Second, competent and successful finance ministers often enjoy public support and can become electoral assets for the leader (Buchs and Soguel 2022). Removing a popular minister may alienate voters or party elites, who can sanction the leader at the ballot box or in internal power struggles. In contrast to autocrats, democratic leaders are more exposed to such political costs.

Third, democratic leaders have strong incentives to sack finance ministers during economic crises. Firing key ministers can serve as a public signal of accountability and help shift blame (Dewan and Dowding 2005). Leaders may also nominate technocrats to demonstrate commitment to reform and reassure investors, potentially improving credibility and lowering borrowing costs (Alexiadou et al. 2021; Alexiadou and Gunaydin 2019). Crises may also prompt leaders to seek more competent replacements or to appoint women, who are perceived as “cleaner” and more compassionate, especially in times of corruption or volatility (Armstrong et al. 2022, 2024; Heinzel et al. 2024). These moves can reflect both scapegoating and sincere reform efforts. As economic crises dominate the political agenda, leaders face mounting pressure to act decisively (Singer 2011). ⁴ Finally, although democratic leaders may fear being challenged by political rivals, the personal risks are lower than in autocracies, where losing power can result in exile, imprisonment, or worse (Geddes et al. 2014).

In sum, democratic leaders are more likely to keep finance ministers when the economy is strong and replace them during downturns. This leads to the following expectation:

Hypothesis 1

In democracies, finance minister retention is related to economic performance, with better performance reducing the likelihood of removal.

Economic Growth and Finance Ministers in Autocracies

Autocrats are not accountable through elections and face removal only through undemocratic means, mainly through elite coups or mass uprisings (Svolik 2012). In light of these two challenges, economic growth creates opposing incentives for the autocrat when it comes to retaining the minister of finance.

There are reasons why economic performance would make the finance minister more secure in an autocratic regime. Growth can stabilize the regime by legitimizing the government and reducing the risk of revolution (Gerschewski 2013; Wintrobe 2000), though the threat of revolt remains costly and less immediate than electoral punishment. Growth also increases tax revenue, enabling the dictator to co-opt elites, enrich themselves, or generally invest in stabilizing the regime (De Mesquita et al. 2005; Olson 1993), making stable finance ministers valuable to avoid disruption during booms. Poor performance, conversely, may prompt ministerial dismissal to signal accountability or restore competence.

Yet, despite this, there are also strong arguments for why economic growth will not work in the finance minister’s favor in the secretive world of authoritarian politics. We point to four ways in which the modality of political competition in autocracies works against highly performing finance ministers.

First, the loyalty–competence trade-off creates incentives that make autocrats more likely to purge well-performing and competent ministers. Since the main threat to their rule comes from within their own ruling coalition, autocrats may especially distrust highly capable ministers, as such individuals pose more potent threats to their position (Egorov and Sonin 2011; Svolik 2012). While not all dictators handle this dilemma like Muammar Gaddafi, who was known to fire ministers for being too talented (Ben-Ghiat 2020), many autocrats may be wary of particularly well-performing and popular underlings, especially in a position as prominent as that of the finance minister.

While this dilemma is most prominent for officials in the military, elite threats also come from important civilian leaders, such as prominent ministers (Bokobza et al. 2022; Lee and Schuler 2020). Although ministers of finance pose a lesser threat to the incumbent than, for example, the minister of defense, they can still become serious challengers. ⁵ When left unchecked, successful finance ministers can build independent support bases, prompting preemptive purges, as illustrated by the case of Anwar Ibrahim in Malaysia (Pepinsky 2009).

Second, purging carries risks, as it may be perceived as a power grab and provoke resistance from other elites (Nadporozhskii 2023; Svolik 2012). However, times of strong economic growth can give dictators the confidence and leverage to remove potential challengers more safely. Just as failed coups create opportunities for rapid consolidation (Bokobza et al. 2022; Timoneda et al. 2023), growth strengthens the leader’s position and reduces the risk of backlash. Thus, economic expansion can both elevate the finance minister as a threat and provide the dictator with a safer window to execute purges.

Third, since autocrats face little electoral pressure, they tend to evaluate finance ministers less on economic growth. Many prioritize policies that benefit the regime and its allies, sometimes at the expense of broader development (Burgis 2015; De Mesquita et al. 2005; Persson and Sjöstedt 2012; Van de Walle 2001). Dictators may also hold unconventional economic views and expect unquestioning loyalty. As a result, ministers might be judged on their ability to implement policies that do not necessarily promote growth, including managing budgets and finances to maintain elite support and obscure the true economic situation from international actors (Waterbury 1973). In contrast, democratic leaders are more accountable for economic outcomes through elections.

Finally, factors unrelated to economic outcomes—such as coup attempts, international pressures, and elite power shifts—often drive ministerial turnover (Bokobza et al. 2022; Kroeger 2020; Woldense 2022). Thus, performance tends to matter less in autocracies than in democracies, where voters hold leaders accountable.

Overall, this leads us to the following hypothesis:

Hypothesis 2

In autocracies, finance minister retention is unrelated to economic performance.

Research Design

Measuring Minister Removal

To measure minister removal, we use the WhoGov dataset (Nyrup and Bramwell 2020). WhoGov contains information on cabinet members for every July in the period 1966–2021 across 177 countries. The creators of WhoGov have singled out the minister of finance, which means that we can accurately identify the minister of finance in all country years. ⁶ Overall, WhoGov has information on 2,887 ministers of finance. For each minister, we create a binary variable indicating whether they are a part of the cabinet the following year (0) or not (1).

A limitation of using yearly minister data is that we cannot capture the exact timing of finance minister changes or multiple turnovers within a year. However, since our main independent variable—economic growth—is also measured annually, a more granular analysis would not be feasible. Moreover, our focus is on whether individual ministers are rewarded, not on general political instability. In Appendix A, we replicate our main results using the Armstrong et al. dataset (Armstrong et al. 2024), which includes monthly data. However, this dataset lacks full cabinet coverage and several variables needed for our later analyses, so we rely on WhoGov for our main results.

We cannot always determine why a finance minister left office—it could be due to voluntary resignation or unforeseen circumstances such as illness. However, given the prestige and power of the role, removal—regardless of subsequent reassignment—is typically a demotion. Cabinet positions offer significant spoils and influence (Meng 2019), and losing them signals a diminished status. It is therefore reasonable to assume that most departures are involuntary. Importantly, this would only bias our results if alternative exit reasons were systematically correlated with both regime type and economic growth.

Measuring Economic Performance

In our main analysis, we classify economic performance into three categories: economic crisis, economic boom, and normal economy. Following Krishnarajan (2019) and using the method from Brambor and Lindvall (2014), we identify sudden deviations from countries’ 10-year GDP per capita (GDP/PC) growth trends. Years with growth more than one standard deviation below the prior 10-year trend are coded as crisis years, while those more than one standard deviation above are coded as boom years. All remaining years are classified as normal. We use real GDP and population data from the Penn World Table (Feenstra et al. 2015), specifically expenditure-side real GDP at chained PPPs. ⁷

The trichotomy offers two main advantages over alternative approaches. First, it aligns with our theory: we do not expect minor growth fluctuations to influence ministerial turnover, but rather that sharp, sudden shifts up or down trigger political consequences. Second, it simplifies the analysis. The categorical measure avoids assumptions about functional form, is easy to model, and facilitates interpretation. In Appendix B, we show that our results are robust to alternative measures, including linear growth and a binary distinction between recessions and growth phases.

We focus on economic growth as our measure of performance, as it is arguably the most important indicator and offers better coverage—particularly in autocracies than alternatives like unemployment or inflation. Moreover, our theory specifically pertains to growth and may not extend directly to other economic indicators. ⁸

We rely on official growth figures rather than alternatives like light intensity data (Martinez 2022), since official statistics are those most visible to policymakers, markets, and the public, making them most relevant for testing our theory.

While autocratic regimes may systematically over-report growth (Martinez 2022), this concern is mitigated by our use of relative growth measures. Even if over-reporting occurs, it is likely consistent over time, allowing us to capture meaningful variation. Our binary recession/growth distinction in Appendix B further addresses this issue, as it is harder to manipulate the direction of growth statistics than their magnitude.

Distinguishing Between Regimes

Our theory treats democratic and autocratic regimes as a dichotomy based on contestation and participation (Dahl 2008). To distinguish regime types, we use the Lexical Index of Electoral Democracy (LIED) (Skaaning et al. 2015), which classifies regimes across seven levels based on up to six institutional criteria. We define democracies as those in level 6—systems with “minimally competitive, multiparty elections with universal suffrage for the legislature and the executive” (Skaaning et al. 2015, 1497); all others are coded as autocracies. Since LIED codes regime status at year-end while WhoGov measures cabinet composition in July, we lag the democracy score by 1 year.

Democracy measurement is inherently difficult, and borderline cases can lead to inconsistent coding across indices (Munck and Verkuilen 2002; Schmid 2025). To ensure our results are not dependent on a single measure, we replicate our analyses using two alternative dichotomous democracy indices: the Boix et al. (2013) classification (updated to 2020) and the V-Dem Polyarchy Index (Coppedge et al. 2022), dichotomized at 0.4 following Baltz et al. (2022). Results using both alternatives confirm our main findings (see Appendix C).

Econometric Set-Up

To estimate the effect of economic performance on ministers’ replacement rate, we separately analyze patterns of removal for finance ministers across autocracies and democracies, using linear probability models given by the following formula:

$R_{i,j,t}=\phi G_{i,j,t}+\omega F_{i,j,t}+\mathit{\beta }{\mathit{X}}_{i,j,t}+{\alpha }_j+{\lambda }_t+{\epsilon }_{i,j,t}$

(1)

for i = 1, ..., I ministers, j = 1, ..., J countries, and t = 1, ..., T years, where the outcome denotes the fate of the minister, R _i,j,t . As mentioned earlier, this variable is binary, where 1 denotes that the minister is removed from the cabinet, while 0 denotes that the person remains in the cabinet. G _i,j,t takes the value 1 in years with an economic boom and 0 in years of crises or normal economy. F _i,j,t takes the value 1 in years of normal economy and 0 in years with crises or booms. Consequently, ϕ Φ indicates the marginal effect of switching from an economic crisis to an economic boom on the likelihood that a finance minister is removed from office, and ϖ indicates the marginal effect of switching from an economic crisis to a normal economy. The error term is given by ε _i,j,t , and all standard errors are clustered by country.

One concern is that the observed associations stem from unobservable country-specific characteristics that affect both growth and cabinet instability. A second concern is that country-specific events, such as coups or revolutions, could affect both economic growth and government stability. Third, endogeneity could stem from reverse causality, as the relationship might be due to finance minister removal causing low growth, rather than the other way around. To mitigate such concerns, we have taken the following strategies.

In our base models, we include country fixed-effects, α _j , and a full set of year dummies, λ _t . Country fixed-effects are pivotal for identification as some countries may experience low growth rates and unstable governing coalitions due to country-specific, time-invariant factors such as the political culture and historical legacies. Meanwhile, we include year dummies to control for common yearly shocks such as the Great Recession in 2007–2009 and the dissolution of the Soviet Union in 1991. These models explicitly compare cabinet stability within regime types in a particular country, depending on whether the economy is in a boom, a crisis, or a normal state.

Next, we include an extensive set of relevant control variables in X _i,t (all lagged by 1 year unless otherwise stated).

First, in the “Economic Controls”-model we include controls for the strength of a country’s economy and its trajectory by including the level of GDP per capita (lagged by 1 year), and the GDP per capita growth in t − 2 based on the Penn World Table (Feenstra et al. 2015). This is to ensure that any effects are not driven by the initial level of economic development, which tends to affect government stability, as well as prospects for relative economic growth, and to avoid conflating the effects of past and current growth.

Next, in the “Electoral Controls”-model we control for electoral confounders by including dummies for national executive and legislative elections from the NELDA dataset (Hyde and Marinov 2012). Thus, we account for the effects of election events both on government stability (Knutsen et al. 2017) and on economic outcomes (Chauvet and Collier 2014).

In the “Instability Controls”-model, we account for a range of controls related to general political instability, namely, (1) successful and attempted coups according to Powell and Thyne (2011), (2) the onset of intrastate, and the onset of interstate wars according to the UCDP/PRIO dataset (Davies et al. 2022; Gleditsch et al. 2002), and (3) the extent and frequency of public mass mobilization, such as demonstrations, strikes, and sit-ins using V-Dem’s v2cagenmob measure (Coppedge et al. 2022). We may fear that they disrupt the economy and can increase government instability. Moreover, we also include the share of high-ranking cabinet members (other than the finance minister and the leader) that are removed in year t to control for general government instability. We also control for whether the current cabinet consists of more than one party (according to WhoGov) to account for effects of coalition governments on economic performance and leaders’ freedom to remove the finance minister. Lastly, the instability controls include the base, squared, and cubic terms of ministers’ current years in office to account for time dependencies (Carter and Signorino 2010). Finally, we run a model that includes all controls at once.

Some of our control variables, specifically our not-lagged measures of instability, may be problematic to include, since they are potential mediating variables. To address this, we report the results of all our control sets separately. ⁹

Lastly, we run our models both including and excluding years where the leader changes. When we include leader-change years, we have two compounded effects of economic growth on minister removal: (1) the effect of economic growth on whether the leader (and the whole government) falls and (2) the effect of economic growth on whether an individual minister is removed. Poor economic performance is well documented to be related to an increase in the probability that a government falls in both democracies and autocracies (see, e.g., Djuve et al. 2020; Gasiorowski 1995; Lucardi 2019; Maeda 2010). Yet, we are mainly interested in the second effect, namely, the effect of economic performance on the removal of individual ministers and not the fall of the leader. We therefore mostly focus on the results where the years with leader changes are removed.

We use linear probability models (LPMs) with a split-sample approach as our main models. Alternative methods are available but come with trade-offs. An interaction model could be used instead of splitting the sample by regime type, but interpretation is more difficult—especially when both independent variables vary within units (Giesselmann and Schmidt-Catran 2022), as is the case here. Nonetheless, we replicate the analysis using interaction terms in Appendix F and find comparable results.

A limitation of the split-sample approach is that it does not allow for direct tests of coefficient equality across regime types. We therefore conduct formal equality tests, reported in Appendix F. We do not use survival models as our focus is on the immediate impact of economic performance on minister turnover rather than long-term tenure effects. However, Appendix F shows that our findings hold when using Cox proportional hazards models. We also replicate the results using logistic regression (Appendix D), which likewise confirms the robustness of our main conclusions.

Are Ministers of Finance Rewarded for Economic Growth?

Before moving on to the main findings, we look at the descriptive patterns. Figure 1 shows the probability of removal of finance ministers by economic performance, revealing three key points.

Figure 1.

Average removal rates of finance ministers in autocracies and democracies during economic booms, crises, and normal economies, including and excluding years of leader changes.

First, when we compare the removal rates with years including leader change, finance ministers are more likely to be removed in democracies compared to autocracies, reflecting more frequent leadership turnover and full cabinet replacement in democracies. Excluding leader changes, we find that the minister of finance is about equally likely to remain in office across regimes.

Second, including leader changes, we find a relationship between economic performance and the likelihood of finance minister removal across regime types. However, this association is much stronger in democracies compared to autocracies. For example, we see a decrease of almost ten percentage points in the likelihood of removal from crisis years to boom years in democracies, while the decrease only is three percentage points in autocracies.

Third, when excluding leader changes, the pattern disappears in autocracies but remains pronounced in democracies. Finance ministers in democracies are nearly 10 percentage points less likely to be removed during booms than crises, while the difference in autocracies is negligible, supporting our theoretical expectations.

Although the descriptive results are supportive of our argument, we also test our theory more systematically. The main results are shown in Figure 2. Since we are using linear probability models, the coefficient estimates can be directly interpreted as marginal effects on the probability of minister removal when the state of the economy changes from (1) economic crisis to normal economy and (2) from economic crisis to economic boom. Thus, if the effect size is −0.1, the minister experiences a 10 percentage point decrease in the likelihood of removal.

Figure 2.

Models show the marginal increase in the probability of being removed from cabinet during years with economic booms and normal economy relative to years of economic crises. The bars indicate 95 percent (thin) and 90 percent (bold) confidence intervals. The corresponding tables are reported in Appendix M. For autocracies N = 3,041 (excluding leader changes) and 3,438 (including leader changes), and N = 2,835 (excluding leader changes) and 3,539 observations (including leader changes) in democracies.

The base models-specification includes both country- and year-fixed effects, but no further control variables. Here, we find that finance ministers in democracies are 4.3 percentage points less likely to be removed when the economy is normal relative to when it is in crisis. Moreover, when the economy is booming, the minister of finance is 10.4 percentage points less likely to be removed relative to crisis years. Both of these coefficients are significant at the 95 percent level.

Moving to finance ministers in autocracies, we find neither a statistically nor substantially significant effect of economic performance on removal. Neither switching from economic crisis to normal economy nor switching from economic crisis to economic booms reduces the likelihood that autocratic finance ministers are removed from cabinet.

The results are robust to different model specifications. This includes controlling for additional economic controls (GDP per capita and previous years’ growth rate), as well as electoral controls (national executive elections and legislative elections), and when controlling for other events that indicate political instability (single-party government, time in office, removal of other high-ranking ministers, successful coup attempts, failed coup attempts, mass mobilizations, the onset of interstate, and the onset of intrastate conflict). Finally, the results hold when we include all these controls in the same model.

In sum, the results in Figure 2 provide robust evidence in favor of our two hypotheses, namely, that finance ministers in democracies are more likely to stay in the government when the economy is growing (Hypothesis 1), while finance ministers in autocracies are no more likely to remain in office when the economy is performing well relative to when it is in crisis (Hypothesis 2). That we obtain these results when excluding leader changes further highlights that this effect is distinct from effects of economic growth on government stability more broadly.

However, another question is whether the coefficients are significantly different from each other across regime types. We show that they are significantly different from each other using three different methods in Appendix F.

First, we use pooled analyses with and without country fixed-effects, where we include both regime types in the same analysis and interact regime type with the measure of economic performance. In both models, we find that the interaction is significant, meaning that the effect is different for democracies compared to autocracies.

Second, we rely on randomization inference to assess and reject that the differences between coefficients for the democratic and autocratic subsamples are due to chance. Here, we randomly split the sample into two sub-samples 10,000 times and calculate the difference between the coefficients in randomly split samples to assess how likely a random split produces a difference in coefficients that is as large as the difference between democracies and autocracies.

Third, we bootstrap the sample 10,000 times, split each bootstrap into democracies and autocracies, run our main models without leader change years, and calculate how robust the difference in coefficients is to random changes to the sample.

Thus, in all three tests for equivalence, the pooled analyses with interactions, the randomization analysis, and the bootstrap, we reach the same conclusion; democratic finance ministers are more likely to remain in office when the economy is performing well compared to when it is in crisis, while this is not the case for their autocratic colleagues, and the difference between the two is statistically significant.

Testing the Robustness of the Main Results

To test the robustness of our findings, we run a range of alternative specifications, none of which changes our substantial conclusions. First, we use logit models instead of LPMs in Appendix D, and Cox proportional hazard models in Appendix E.

Next, in Appendix C, we distinguish between democracies and autocracies using the binary BMR-index (Miller et al. 2022) and the V-Dem Polyarchy Index (Coppedge et al. 2022) with a cut-off at 0.4.

Furthermore, we test the construct validity of our main variables. In Appendix N, we use an alternative dependent variable where we include both removals from the cabinet and demotions where the minister of finance is allocated to a less prestigious portfolio within the cabinet. In Appendix B we use alternative measures of economic performance, specifically (1) a binary measure comparing years with negative and positive growth, (2) growth as a continuous predictor, (3) variations of our trifecta with varying deviations from the growth trend and the length of the growth trend, (4) the log of truncated inflation rates, and (5) a binary measure comparing years with inflation rates exceeding 5 percent to years at or below 5 percent inflation. In Appendix A, we use the new dataset of Armstrong et al. on finance ministers (Armstrong et al. 2024).

Lastly, in Appendix O, we show that our main finding is robust to controlling for government net borrowing, confirming that the boom effect is not driven by fiscal deficits, as the findings of Schmid and Aaskoven (2025) could lead one to expect.

Further Analyses

We conduct three additional analyses to support our theory.

First, our theory refers exclusively to ministers of finance. The patterns in our main results should therefore not be found for other types of ministers who have no responsibility for the economy. Figure 3 replicates our analysis with defense ministers, and in Appendix G we do the same for foreign ministers. We find no systematic patterns for either of these ministers. This strengthens our theory’s credibility, as the unique pattern for ministers of finance highlights their distinct role in economic governance and indicates that our results are not driven by more general differences in how autocratic and democratic leaders manage their cabinet in relation to economic performance.

Figure 3.

Models show the marginal increase in the probability of being removed from cabinet during years with economic booms and normal economy relative to years of economic crises for the minister of defense. The bars indicate 95 percent (thin) and 90 percent (bold) confidence intervals. The corresponding tables are reported in Appendix M.

Second, we explore timing in Appendix H. We run models looking at the association between economic performance measured in year t − 3 to t + 3 and the removal of the finance minister. For democracies, we only find a significant association between the economic performance measured in t and t + 1 and the removal of the finance minister. The association in t + 1 is, at first, puzzling but can be explained by two factors. First, WhoGov captures the cabinet in July every year, meaning that our dependent variable captures whether a finance minister is removed between July in year t and July in t + 1, and we are therefore using a dependent variable with half a year of lead. This offers some protection against reverse causality, since half of the time period is in the following year, and the removal of the finance minister in the last half of the year hardly can impact growth much. In addition, this also means that economic growth in t + 1 can impact the removal of the finance minister in year t (since half of the time period is in year t + 1). Moreover, and returning to the issue of reverse causality, namely, where the removal of the minister of finance leads to adverse economic outcomes, we would expect to see this in t + 1 and t + 2, rather than in year t. In fact, we do find some indication that this is the case, both in autocracies and democracies. This suggests that economic performance is only related to the survival of finance ministers in democracies, while finance minister removal may be related to adverse economic performance in both democracies and autocracies.

Third, we investigate whether our findings differ between subtypes of democracies (in Appendix I) and autocracies (in Appendix J).

For democracies, we use the classification by Anckar and Fredriksson (2019) to distinguish parliamentary and presidential systems (including semi-presidential systems). ¹⁰ We find that economic performance is related to the removal of finance ministers in both types of democratic systems but that the relationship is stronger in parliamentary systems compared to presidential systems. This is unsurprising since parliamentary systems have more direct accountability mechanisms and less diffusion of power, causing the leader to be more sensitive to negative events. Cabinet members require a parliamentary majority in parliamentary systems, while presidents are freer to appoint cabinet members. Furthermore, presidents are elected for a fixed term and are often faced with term limits, while prime ministers in parliamentary systems can lose their majority between elections and are more likely to face irregular elections (Clark et al. 2024; Gerring et al. 2009). In addition, cabinet members are much more likely to be elected politicians themselves in parliamentary systems compared to presidential systems, implying that their personal popularity is of greater importance (Nyrup et al. 2025).

For autocracies, we divide our sample into monarchies, party-based, personalist, and military regimes using the classification created by Geddes et al. (2014). We find no systematic relationship between economic performance and the removal of ministers in any type of autocracy. Moreover, we split autocracies into closed and electoral autocracies, that is, autocracies holding regular multi-party elections but failing to meet the minimal standards of democracy according to LIED (Skaaning et al. 2015). We find no relationship between economic performance and finance minister survival in electoral autocracies. Hence, elections and political openness without democracy do not appear to offer the same scope conditions as democracy.

Conclusion

We demonstrate a close link between economic growth and the retention of finance ministers by the chief executive in democracies but not in autocracies, and show that this link in democracies is distinct from broader effects of economic growth on the survival of the government as a whole. This implies that finance ministers in democracies are, to a higher degree than their autocratic counterparts, masters of their own fate. If they prove themselves competent and deliver on economic growth, they are more likely to remain in office. This finding gives important insights into the management of top officials. Democratic institutions incentivize leaders to retain and reward seemingly well-performing top officials. In addition, electoral institutions nudge the leader to show accountability when things go poorly. This is good news for the citizens of democracies, since it gives top officials strong incentives to perform well and implement policies that increase economic output.

The absence of free and fair elections changes the incentive structure for autocrats, meaning that it is not in their interest to reward top officials for good performance, at least in the economic area. Autocrats are wary of potential competitors, who can use their success to build alternative power bases and challenge the leader. Furthermore, autocratic leaders can use the strong economy to purge from a position of strength, and thereby solidify their rule in the long run. Moreover, autocratic finance ministers may not have as strong incentives in the first place to help the leader build a strong economy. Instead, they may be better off helping the leader enrich herself and by implementing the dictator’s favored policies independently of their economic impact. This entails that the officials in charge of steering the economy in an autocracy have little reason to take a professional interest in the citizens’ economic well-being and conduct their job in a manner that benefits the public at large. Ultimately, these mechanisms can help explain why autocracies on average have more unstable and worse growth than their democratic counterparts.

The study also opens up avenues for future research. The study is the first to investigate the effect of economic performance on the removal of ministers between types of regime. Yet, do the results travel to other domains? For example, are democratic health ministers, in comparison to their autocratic colleagues, rewarded for handling the COVID-19 pandemic well? If so, this would strengthen the case that democracies are better at selecting competent people for top positions. In addition, the personal characteristics of the finance minister could play a role. Is the relationship between economic growth and survival moderated by the educational level of the finance minister, or do we see another type of finance minister being appointed following a crisis in democracies compared to autocracies? In addition, we only briefly discuss how the relationship differs between sub-types of regimes, and future research could more systematically investigate this. Although this study has shown that democrats and autocrats manage their cabinet differently in response to the performance of the economy, we are still far from fully understanding the comparative politics of the appointment of public officials.

Supplemental Material

Supplemental Material - Governing for Growth? Economic Performance and Finance Minister Survival in Democracies and Autocracies

Supplemental Material for Governing for Growth? Economic Performance and Finance Minister Survival in Democracies and Autocracies by Jonas Willibald Schmid and Jacob Nyrup in Political Research Quarterly.