Over the past three decades, the Roman political market has provided a compelling lens through which to observe the dynamics of democratic transformation in Italy’s capital—a city characterized by electoral volatility, rising abstention, and the progressive fragmentation of party competition.

As both the political and symbolic center of the country, Rome encapsulates the structural tensions of Italian democracy: the erosion of traditional party loyalties, the proliferation of civic and populist movements, and the reconfiguration of voter behavior in a context of growing informational complexity.

Following the market metaphor long established in political science (Downs, 1957; Wolinetz, 1979; Sartori, 1976; Mair, 1997), the political field can be understood as a competitive marketplace of ideas and representation. Parties and candidates behave as adaptive organizations seeking to maximize support, while voters act as rational consumers who evaluate the available “products” based on ideological distance, perceived benefits, and cognitive costs.

When the informational or psychological costs of participation outweigh perceived benefits, abstention becomes a rational form of exit from the political market (Hirschman, 1970).

In this framework, electoral behavior is dynamic rather than static—driven by how citizens process political information and interpret signals in an increasingly diversified and fluid environment.

This process of voter transformation has been described by Norris (2002) and Dalton (2018) as the rise of cognitive mobilization and partisan dealignment. Voters have become more informed, autonomous, and capable of independent judgment, yet simultaneously less anchored to traditional ideological or partisan identities. As citizens gain access to more information and perceive greater choice, they rely less on stable cues and more on situational evaluations. The result is a form of electoral complexity in which individual rationality coexists with systemic unpredictability—an ideal context for analyzing political entropy.

To conceptualize this complexity, the present study employs the notion of entropy, a concept originally derived from thermodynamics and adapted by Shannon (1948) in information theory. Entropy measures the degree of uncertainty or dispersion within a system: when votes are evenly distributed among many actors, entropy rises, reflecting fragmentation and uncertainty; when votes concentrate around fewer parties, entropy declines, signaling order and predictability.

In political systems, as shown by Väyrynen (1972) and Coleman (1995), entropy captures the underlying informational coherence of the political market—how citizens distribute their support across an expanding or contracting array of options.

The Roman case offers an ideal empirical laboratory for applying this framework. Between 1993 and 2023, Rome experienced 29 electoral cycles across four levels—local, regional, national, and European—involving over 2.3 million voters per cycle and more than 500 competing lists. This sequence allows for a longitudinal and multi-level analysis of how entropy evolves as a function of both political supply and voter behavior. The study draws on Byeon’s (2005) systemic model of entropy as a balance between adaptability and political support, and on Ferreira and Dionisio’s (2008) interpretation of entropy as a correlate of voter disaffection and abstention. In this view, entropy embodies both structural and behavioral dimensions of political change: it reflects how the fragmentation of the offer generates uncertainty, and how citizens react by either dispersing or withdrawing their participation.

Ultimately, the goal of this paper is to capture the systemic dynamics of voting behavior in relation to the structure of the political offer. By treating entropy as both an outcome and a driver of political disaffection, the study seeks to illuminate the adaptive rhythms of democratic competition in Rome. Entropy emerges as a synthetic, underutilized measure capable of revealing the informational and behavioral complexity of modern democracies—highlighting how political systems oscillate between dispersion and re-stabilization, uncertainty and order, as they continuously adjust to the pressures of pluralism and change.

This paper constitutes a first attempt to approach the political market through a complexity-based perspective. While this perspective provides a useful lens for interpreting electoral dynamics, it also entails inherent limitations, as it relies on an analytical framework that remains relatively underdeveloped and only marginally applied in empirical political research. The study should therefore be regarded as an initial contribution, intended to be refined and extended in future work.

This section connects the study to the broader political science literature on electoral abstention, showing how traditional explanations of nonvoting—from rational choice and representation crises to cultural disaffection—can be reframed in entropic terms, as manifestations of systemic dispersion of participatory energy.

In recent decades, political science has devoted increasing attention to the question of voter abstention, treating it not merely as a numerical drop in turnout but as a structural symptom of changing relationships between citizens, institutions, and political culture. Abstention is no longer seen as an episodic deviation from democratic norms but as a long-term expression of transformation in political engagement and trust.

Already in the early 2000s, Corbetta and Cavazza (2009) laid the groundwork for this shift, pointing out the methodological stagnation of previous studies that focused too narrowly on aggregate data and ignored the individual motivations behind electoral behavior. They argued that declining participation stemmed not from apathy but from a rationalization of electoral choice: nonvoting as a conscious, deliberate act reflecting a weakening bond between citizens and the political system.

From the 2000s onwards, research increasingly interpreted abstention as a rational and strategic decision. Cerruto (2012) demonstrated that voters often weigh the personal costs and benefits of participation, choosing to abstain when the perceived stakes of an election are low. This approach, aligned with economic voting theory, suggests that citizens are more likely to turn out in first-order elections (national contests) than in second-order ones (local, regional, or European), where the perceived impact of their vote is lower.

More recent work has framed abstention within the broader context of political disaffection and representation crises. Biorcio and Natale (2024) identify three phases in its evolution: a first, physiological stage in the First Republic, when abstention was marginal and often logistical; a second, linked to the breakdown of mass parties and political subcultures in the 1990s; and a third, current stage, in which abstention has become structural and cross-cutting, affecting even politically informed and civically engaged citizens. In this latest phase, not voting is no longer a sign of indifference, but rather a deliberate act of disengagement rooted in disillusionment with how democracy functions.

A complementary perspective comes from Legnante and Segatti’s (2001) concept of the intermittent abstainer, describing voters who alternate between participation and nonparticipation depending on context and motivation. This pattern captures the growing fluidity of electoral behavior: voting has lost its ritual dimension and become a strategic and reversible choice, reflecting a more autonomous yet skeptical electorate.

An institutional approach, advanced by Fruncillo (2023), explores the limited effect of electoral reforms on turnout. Measures such as early, postal, or electronic voting may facilitate access, but they rarely reverse the overall decline in participation. Fruncillo argues that the roots of abstention are primarily cultural rather than technical, linked to the erosion of civic duty and collective trust.

Taken together, this literature suggests that abstention today is not simply a withdrawal from politics but a distinct form of political behavior—the outcome of rational calculation, institutional change, and civic disillusionment. It can thus be interpreted as a manifestation of democratic entropy, a diffusion of participatory energy within increasingly fragmented political systems.

This paragraph complements the framework of the analysis by outlining the institutional architecture of the Roman political field. Through a systematic overview of the electoral laws governing local, regional, national, and European contests.

The evolution of Italy’s electoral systems between 1993 and 2023 has deeply shaped the dynamics of the Roman political market. Across local, regional, national, and European elections, each reform has redefined the rules of competition, the distribution of political power, and ultimately the structure of voter choice—contributing to growing fragmentation and uncertainty in the system.

At the local level, the watershed came with Law 81/1993, which introduced the direct election of mayors and a two-round majority system, replacing pure proportional representation. This marked the beginning of the personalization of local politics and the rise of a bipolar contest structure. In Rome, its impact was immediate: the 1993 election between Francesco Rutelli and Gianfranco Fini symbolized the start of a new era of media-driven leadership and direct executive accountability. While this reform strengthened visibility and political competition, it also weakened partisan stability and increased vote dispersion—a key driver of political entropy at the municipal level.

At the regional level, the Tatarella Law (Law 43/1995) established a mixed system combining proportional representation with a majority premium and the direct election of the regional president. Later constitutional reforms in 1999 and 2001 reinforced regional autonomy, while the Lazio Region further adjusted its rules through Regional Laws 2/2005 and 10/2017. The 2017 reform abolished the majority bonus and introduced a limit of two consecutive presidential terms. Despite these changes, the underlying logic remained one of governability over representativeness, creating systems that tend to favor executive stability but reduce pluralism.

At the national level, three major electoral reforms defined the period. The Mattarellum (Law 276/1993) introduced a semi-majoritarian model: 75% of seats filled through single-member districts and 25% via proportional allocation. This fostered bipolar coalitions and enhanced candidate-centered politics. The Calderoli’s Law (Law 270/2005), by contrast, reinstated a proportional system with closed lists and a majority bonus, significantly reducing voter influence over candidate selection—and was later ruled partly unconstitutional. The current Rosatellum (Law 165/2017) reintroduced a mixed structure, with roughly 61% of seats distributed proportionally and 37% through first-past-the-post contests, and a 3% national threshold. This ongoing reform cycle reflects the chronic instability of Italy’s institutional engineering: a system constantly oscillating between proportional fairness and majoritarian efficiency, never quite achieving either.

The European elections, in contrast, have remained relatively stable since Law 18/1979, amended in 2000 to include a 4% threshold. The use of a pure proportional system with multi-member constituencies ensures pluralism but contributes to a sense of distance between voters and European institutions, reflected in persistently lower turnout levels.

Overall, the Italian electoral framework over these thirty years reveals a patchwork of rules that pull in different directions. The coexistence of proportional, mixed, and majoritarian logics across levels of governance has fostered a complex and fragmented political environment. In Rome, this institutional pluralism has amplified vote volatility and systemic uncertainty—a fertile ground for observing entropy in political competition, where instability, dispersion, and voter fragmentation become intrinsic features of the democratic process.

This study adopts a systemic approach to the political market of Rome, examining how the entropy of the electoral offer evolves and mays interacts with voter participation over time. The analysis covers a thirty-year period (1993–2023)¹, encompassing 29 electoral cycles across four distinct levels of competition: local (municipal), regional, national, and European elections. This temporal and institutional breadth allows for a comparative exploration of how political fragmentation and voter behavior vary across scales of governance and over time.

The empirical focus on Rome—a metropolitan area with approximately 2.3 million eligible voters—offers a unique opportunity to observe electoral dynamics within a dense and politically diversified urban system. Rome’s electoral arena combines characteristics of national politics (party competition, coalition structures) with the local specificities of urban governance, making it a particularly rich case for analyzing political entropy.

Entropy is computed following the Shannon formulation, normalized to yield relative entropy values ranging from 0 to 1.

Where p i represents the proportion of votes obtained by each party list, and H max denotes the theoretical maximum entropy given the number of parties. This relative measure allows for comparisons across different electoral contexts with varying numbers of competitors, as suggested by Ferreira and Dionisio (2008). This normalization ensures that the index does not mechanically increase with the mere addition of parties: a higher number of lists only raises relative entropy if electoral support becomes more evenly dispersed among them.

In order to focus on changes in the dynamics of the political market offer, the analysis is restricted to valid votes only. This choice allows entropy to capture the effective distribution of electoral support, excluding invalid or blank ballots that do not convey information about political competition. Furthermore, to isolate the dynamics of political fragmentation rather than representation outcomes, only party lists obtaining at least 1% of the total vote are included. This threshold ensures that entropy reflects the effective structure of the political offer, rather than the long tail of marginal actors whose negligible support could artificially inflate disorder without producing meaningful systemic effects².

By treating elections as successive states of a complex adaptive system (Byeon, 2005; Marmani et al., 2020), the study interprets entropy as an indicator of systemic uncertainty. Fluctuations in entropy levels thus capture not only the degree of competition but also the broader informational and emotional coherence of the political field. The methodological design allows for both cross-level and temporal analyses, facilitating the identification of cyclical or self-correcting dynamics in the evolution of Rome’s political market.

Drawing on previous studies that conceptualize entropy as both a determinant and a reflection of political stability and participation (Ferreira and Dionisio, 2008; Byeon, 2005), the research tests three main hypotheses:

Together, these hypotheses reflect an integrated understanding of electoral entropy as a dynamic equilibrium between fragmentation and consolidation, uncertainty and adaptation. In the Roman context, this cyclical balance may reveal how local and national actors adjust their strategies within a shared political ecosystem, offering new insight into the systemic rhythms of democratic life.

The scatter plot (Figure 1) illustrates the evolution of relative entropy (HREL) in the Roman political market from 1993 to 2023, across four levels of electoral competition—Local (L), Regional (R), National (N), and European (E). Each observation represents the degree of systemic uncertainty and dispersion of the political offer in a given electoral cycle, while the dashed regression line summarizes the overall temporal trajectory.

The data reveal a slight upward drift in entropy over the thirty-year period, though this pattern remains modest and should be interpreted with caution rather than as a strong monotonic trend. In this sense, the Roman political system appears to have become somewhat more fragmented and less predictable, with indications of greater diversity and volatility in the political offer, even as levels of entropy continue to oscillate across electoral cycles and arenas. From an informational standpoint, this points to a gradually more complex environment for voters, who are confronted with an expanding and intermittently reconfigured set of competing signals and heterogeneous alternatives.

A closer inspection of the election types reveals some interesting nuances. Local elections (L) exhibit the highest entropy levels, confirming the more volatile and plural nature of urban political competition, where new civic lists and personal movements frequently enter the arena. Regional (R) and European (E) elections show more irregular patterns, reflecting their varying degrees of politicization and voter engagement over time. National elections (N) tend to display intermediate values, consistent with the stabilizing influence of stronger party brands and structured coalitions.

Overall, the evidence suggests that the entropy of the Roman political market has gradually increased, pointing toward a systemic transition from a more centralized to a more dispersed structure of competition. This evolution can be interpreted through the dual lens proposed by Ferreira and Dionisio (2008) and Marmani et al. (2020): higher entropy corresponds to a more fragmented and uncertain political offer, which may simultaneously stimulate voter choice diversification and foster disaffection or abstention when uncertainty exceeds a tolerable threshold.

In this sense, the Roman case exemplifies the dynamics of a complex adaptive democratic system, where entropy acts as a sensitive indicator of the balance between pluralism and coherence, competition and stability. The slow but visible upward trend over time hints at a long-term transformation in the structure of political competition—one that mirrors broader processes of party system de-institutionalization and voter volatility observed in many European democracies.

The first hypothesis, grounded in the existing literature, conceptualizes entropy as a potential explanatory factor of voter abstention. However, the empirical evidence from the Roman case does not provide support for this relationship. In the observed electoral cycles, the level of relative entropy shows only a negligible correlation with voter turnout (Pearson’s r = −0.07). This result remains unchanged when considering lagged entropy values (t–1), for which the correlation is similarly weak (r = −0.03). Given the limited size of the sample and the focus on a single political market, these findings do not warrant further investigation within the present study. Assessing the relationship between entropy and abstention would require a broader comparative design, involving multiple cases and a more extensive temporal or territorial scope.

To test the second hypothesis, a regression analysis (Table 1) examines the determinants of relative entropy (HREL) in the Roman political system between 1993 and 2023, using a set of explanatory variables derived from the theoretical framework. Relative entropy, interpreted as an indicator of systemic uncertainty, is regressed on variables capturing both temporal dependence and key features of the electoral context.

The first variable, relative entropy (t–1), captures the inertia of the system—whether high entropy in one election tends to persist in the next, as suggested by Coleman (1995). The negative coefficient (−0.338) indicates a self-correcting dynamic, where periods of high fragmentation are often followed by lower entropy in subsequent elections. However, the result is not statistically significant (p = 0.115), suggesting that while the adaptive tendency exists, it is not strong enough to confirm a stable autoregressive pattern in the Roman case.

The variable total number of lists is the only significant positive determinant (coeff. = 0.0026, p = 0.007). This aligns with both the theoretical expectations and empirical literature (Väyrynen, 1972; Ferreira and Dionisio, 2008; Marmani et al., 2020). A larger number of competing lists increases the fragmentation of the political offer, thereby raising systemic entropy. In practical terms, the proliferation of electoral options amplifies uncertainty in voter decision-making, disperses informational cues, and reduces predictability in the political market. This finding strongly supports the idea that the structure of political supply is a key driver of systemic uncertainty.

The variable full duration (indicating whether an election was held to term or called early) shows a small and insignificant negative coefficient (p = 0.775). This suggests that early elections do not systematically affect the level of entropy. While intuitively one might expect premature elections to signal political instability and thus raise entropy, the data indicate that their occurrence in Rome has not translated into meaningful shifts in systemic fragmentation.

Finally, election day concurrency, representing the simultaneity of multiple elections, displays a negative and marginally significant coefficient (−0.071, p = 0.006). This is coherent with Marmani et al. (2020), who found that simultaneous elections tend to reduce entropy by concentrating voter attention and reinforcing partisan alignment. When multiple contests occur on the same day, citizens often rely on the same cognitive and informational shortcuts across ballots, leading to a more coherent distribution of votes.

Taken together, the model explains about 46% of the variation in entropy (R² = 0.465), suggesting a moderate degree of explanatory power given the inherently noisy nature of electoral data. F-statistic value around 5 and a significance at p = 0.01 allow us to consider not all the parameters equal to zero. The results confirm that entropy is mainly driven by the fragmentation of the political offer rather than by past dynamics, turnout, or institutional timing. The Roman political market thus appears highly sensitive to the structural complexity of electoral competition, with the number of competing lists functioning as the most reliable predictor of systemic uncertainty.

To better capture the systemic dynamics underlying the regression results, it is necessary to look beyond aggregate measures and examine how relative entropy (HREL) evolves within each electoral level—local, regional, national, and European. Figure 2 provides this disaggregated view, illustrating the temporal evolution of entropy in the Roman political market between 1993 and 2023. By isolating each arena, we can observe how the system’s behavior varies across scales of governance and electoral engagement.

Across all four levels, the patterns reveal a distinct wave-like motion—a cyclical alternation between phases of high and low entropy. This empirical regularity confirms the systemic interpretation of Byeon (2005), who argued that political systems tend to oscillate between moments of disorder and periods of re-stabilization. In Rome’s case, periods of high entropy correspond to fragmented, uncertain, and competitive electoral environments, while subsequent declines in entropy signal adaptive rebalancing, as voters and political actors respond to excess fragmentation by reconsolidating support. This oscillatory movement suggests that the Roman political market behaves as a self-correcting adaptive system, continually adjusting its internal equilibrium in response to fluctuations in political supply and voter sentiment.

When the analysis is disaggregated, each electoral level displays its own rhythm.

At the local level, entropy shows the most pronounced oscillations. The volatility of municipal politics—driven by the frequent emergence of civic lists, local movements, and short-lived alliances—produces sharp increases in entropy, often followed by stabilization when the electorate reorients toward stronger or more institutionalized parties. Regional elections display a similar cyclical pattern but with moderate amplitude. Here, the interplay between local identities and national party structures generates temporary spikes in fragmentation, yet the system tends to re-stabilize as governing coalitions consolidate.

At the national level, the trend is smoother, suggesting the partial stabilizing effect of party brands, leadership continuity, and ideological anchoring. Nonetheless, even national contests reveal fluctuations linked to systemic shocks—such as party realignments or leadership crises—that temporarily raise entropy before the system recentres.

Finally, European elections show irregular but still cyclical trajectories. Their entropy levels oscillate according to the varying salience of EU politics in domestic discourse and the alternating presence of populist or issue-based formations that attract disaffected voters.

Taken together, these trajectories reveal that entropy in the Roman political system is not random variation but a manifestation of cyclical adaptation. Each level of election contributes differently to this balance: when one arena becomes fragmented, others tend to stabilize, allowing the system as a whole to avoid sustained disorder. This pattern supports the idea of an “oscillatory equilibrium,” in which the system absorbs shocks and reorganizes itself through the interactive behavior of voters and political actors.

From a theoretical perspective, these findings refine the picture emerging from the regression results. While structural variables such as the number of competing lists remain the strongest predictors of systemic uncertainty, the temporal and multi-level analysis shows that entropy also reflects adaptive feedback mechanisms inherent to democratic systems. Rather than signaling progressive destabilization, the fluctuating entropy of the Roman political market illustrates its capacity for renewal and re-equilibration—a hallmark of resilient, though increasingly complex, democratic dynamics.

The findings of this study contribute to the growing body of literature that interprets entropy as a systemic indicator of political complexity, uncertainty, and adaptive behavior. In line with existing research on the informational interpretation of political competition, the results show that a broader and more diversified political offer—measured by the number of lists competing in each election—tends to fragment the vote and reduce participation. When citizens are faced with an increasingly complex electoral environment, their ability to decode political messages and identify meaningful distinctions among parties becomes strained. This informational overload often translates into disaffection or disengagement, as suggested by Ferreira and Dionisio (2008) and supported here by the significant relationship between entropy and the structural fragmentation of the political offer.

At the same time, the analysis of temporal dynamics across electoral levels reveals that entropy does not increase indefinitely. Instead, it follows a wave-like pattern, confirming the oscillatory behavior theorized by Byeon (2005). After moments of heightened uncertainty and dispersion, the system appears to self-correct, returning to more moderate levels of entropy. This dynamic equilibrium reflects a form of collective adaptation, in which both voters and political actors respond to excessive fragmentation by re-stabilizing preferences, alliances, and electoral strategies. In this sense, entropy can be understood not only as a measure of disorder but also as an indicator of the system’s resilience and self-regulation.

These oscillations, observed consistently across local, regional, national, and European elections, suggest that the Roman political market functions as a complex adaptive system—one that continuously negotiates between pluralism and coherence. Periods of high entropy stimulate political innovation and diversification, while phases of lower entropy favour consolidation and predictability. The alternation between these two states embodies the cyclical rhythm of democratic competition in a multi-level context.

Some clear limitations of this study should be acknowledged, in line with its exploratory scope as outlined in the introduction. Focusing on a single territorial political market does not allow for general conclusions about the broader dynamics of political complexity, and future research should extend the analysis to multiple cases and contexts. Moreover, while the present approach captures the structural dispersion of the political offer, it does not explicitly account for power asymmetries among political actors or for qualitative differences between party lists. Incorporating these dimensions would allow for a more refined application of the complexity-based perspective and a more nuanced interpretation of electoral entropy (see Bischof, 2015).

Finally, the capacity to monitor and interpret these entropic dynamics opens up intriguing possibilities for future research and even for strategic forecasting. If entropy levels indeed capture the systemic uncertainty of the political market, they could serve as early indicators for shifts in electoral behavior or as tools to anticipate phases of fragmentation and consolidation. Understanding the patterns of entropy thus offers not only analytical insight into the structure of democratic competition, but also a potential framework for interpreting and managing the information dynamics that shape the stability and evolution of complex political systems like Rome.