Party Systems & Coalitions — Explain party systems and coalition formation at a high level - College Political Science

SECTION 1

Historical Context & Motivation

The systematic study of party systems emerged in the mid-twentieth century as scholars of comparative politics grappled with a fundamental puzzle: why do some democracies feature two dominant parties while others sustain five, six, or more? Early democratic theory, rooted in the experience of Britain and the United States, often assumed that a two-party arrangement was the natural or even optimal form of democratic competition. Yet the proliferation of multiparty parliaments across continental Europe, Latin America, and the postcolonial world challenged that assumption and demanded more rigorous frameworks. The concepts of party system classification and coalition formation theory arose to address exactly this diversity, providing tools for comparing political systems across space and time.

1951 • Duverger's Political Parties

Maurice Duverger publishes Les partis politiques, arguing that plurality electoral rules tend to produce two-party systems (Duverger's Law), while proportional representation fosters multiparty systems (Duverger's Hypothesis). This becomes the foundational text linking institutional design to party system structure.

1962 • Riker's Theory of Political Coalitions

William Riker applies game theory to coalition politics in The Theory of Political Coalitions, introducing the minimum winning coalition concept—the idea that rational actors form the smallest coalition sufficient to win, maximizing each partner's payoff.

1976 • Sartori's Typology

Giovanni Sartori refines party system classification in Parties and Party Systems, distinguishing systems not merely by count but by the ideological distance between parties—introducing categories such as moderate pluralism and polarized pluralism.

1984 • Lijphart's Patterns of Democracy

Arend Lijphart begins his comparative project connecting party systems to broader patterns of majoritarian versus consensus democracy, demonstrating that coalition governance is a structural feature—not a deficiency—of consensus democracies.

2000s • Behavioral & Institutional Synthesis

Contemporary scholars integrate electoral rules, social cleavages, and strategic voting into unified models, using the effective number of parties index (Laakso–Taagepera) as a standard metric for cross-national comparison.

The central question driving this literature remains: given that democracies adopt different electoral institutions and harbor different social cleavages, how do we systematically describe the resulting party landscape, and how do parties within that landscape negotiate to form governing coalitions? Answering this question requires both typological precision—classifying what kind of party system exists—and strategic reasoning about why particular coalition configurations emerge rather than others.

SECTION 2

Core Principles & Definitions

Before analyzing specific systems, it is essential to define the core concepts that structure this field. A party system refers not merely to a collection of political parties but to the structured pattern of interaction among them—their number, relative size, ideological positioning, and the ways they compete or cooperate. Understanding a party system requires grasping both its mechanical properties (how many parties, how large) and its dynamic properties (centripetal versus centrifugal competition, willingness to form coalitions).

Party System

The structured pattern of competitive and cooperative interactions among all significant political parties within a polity. It encompasses the number, size, ideological distance, and alliance potential of parties.

Effective Number of Parties (ENP)

A weighted index (Laakso–Taagepera) that accounts for both the number and relative size of parties, providing a single summary statistic for system fragmentation. An ENP of 2.0 indicates a pure two-party system; values above 4.0 suggest high fragmentation.

Coalition

A temporary alliance of two or more parties that agree to govern together, typically by distributing cabinet portfolios and negotiating a joint policy agenda. Coalitions are the dominant governance mechanism in multiparty parliamentary systems.

Minimum Winning Coalition (MWC)

A coalition that commands a legislative majority but would lose that majority if any single member party defected. Riker's size principle predicts that rational actors prefer MWCs because they maximize each partner's share of the spoils.

Cleavage Structure

The set of deep, persistent social divisions (class, religion, ethnicity, region) identified by Lipset and Rokkan that generate distinct voter blocs and, in turn, distinct parties. Cleavage structure is a primary determinant of party system format.

✦

KEY TAKEAWAY

Think of a party system like a marketplace: just as the number of competing firms, their market shares, and the nature of the products they sell collectively define the structure of an industry, the number, size, and ideological character of parties define the structure of political competition. A monopoly (one-party dominance) differs fundamentally from perfect competition (many small parties), and in each case the dynamics of coalition-building—like mergers and alliances in business—follow predictable strategic logic.

SECTION 3

Visual Explanation — Party System Typology

The following diagram illustrates the major types of party systems arranged along a continuum of fragmentation, from one-party dominance through two-party systems to moderate and polarized multiparty systems. Each type is annotated with real-world examples and its typical effective number of parties (ENP) range, showing how the abstract classification maps onto concrete democratic experience.

The diagram arranges four party system types along a fragmentation continuum. Note that the effective number of parties (ENP) increases from left to right, and the bottom panel identifies three key drivers—electoral rules, social cleavages, and strategic behavior—that interact to determine where a given democracy falls on this spectrum.

Sartori's framework, visualized above, adds an important qualitative dimension to what might otherwise be a purely numerical exercise. A system with five parties arranged along a narrow ideological spectrum (moderate pluralism) operates very differently from a system with five parties spanning extreme ideological poles (polarized pluralism). In the former, coalition bargaining is relatively fluid because any combination of centrist parties can produce a workable government. In the latter, the presence of anti-system parties at the extremes constrains coalition options, often trapping centrist parties in unstable alliances and driving what Sartori calls centrifugal competition—a dynamic in which parties move toward the poles rather than the center.

SECTION 4

Analytical Framework — Measuring and Modeling Party Systems

While party systems are ultimately political phenomena shaped by history, institutions, and culture, comparative politics has developed several quantitative tools that allow for precise cross-national measurement. The most widely used is the effective number of parties (ENP) index proposed by Markku Laakso and Rein Taagepera in 1979. This metric weights each party by its size, ensuring that a system with two large parties and several negligible ones is not treated the same as a system with five equally sized parties.

EFFECTIVE NUMBER OF PARTIES (ENP)

N = 1 / Σ pᵢ²

Where N = effective number of parties, pᵢ = the proportion of seats (or votes) held by the ith party, and Σ denotes summation over all parties. The index equals 1.0 when one party holds all seats and approaches the raw count n when all parties are equally sized.

The ENP formula is mathematically equivalent to the inverse of the Herfindahl–Hirschman Index (HHI) used in industrial economics to measure market concentration—a striking parallel that reinforces the marketplace analogy for party competition. Just as a high HHI indicates monopolistic concentration, a low ENP indicates that political power is concentrated in very few parties.

DUVERGER'S LAW (HEURISTIC FORM)

FPTP districts → strategic voting + strategic entry → N ≈ 2

Under first-past-the-post (FPTP) rules, voters abandon third-party candidates they perceive as non-viable (the mechanical effect of wasted votes), and elites refrain from funding hopeless candidacies (the psychological effect). The result is convergence toward two serious competitors per district.

RIKER'S SIZE PRINCIPLE

Preferred Coalition Size = 50% + 1 seat (minimum winning coalition)

Riker's size principle predicts that office-seeking parties form coalitions just large enough to secure a legislative majority, because adding unnecessary partners dilutes the payoff (cabinet portfolios, policy influence) for each member. In practice, information asymmetries, policy preferences, and institutional norms often produce oversized or minority coalitions, but the MWC remains the baseline prediction.

ℹ

Beyond Office-Seeking

Coalition formation theories have evolved beyond Riker's purely office-seeking model. Policy-seeking models (De Swaan, Axelrod) predict that parties prefer ideologically adjacent partners, forming minimum connected winning coalitions—alliances that span the smallest possible ideological range. Vote-seeking models add electoral considerations: parties may prefer partners that do not threaten their voter base in the next election.

SECTION 5

Coalition Types & Formation Logic

Coalition governments are the norm rather than the exception across parliamentary democracies worldwide. Kaare Strøm's data show that roughly two-thirds of all governments formed in Western European democracies between 1945 and 2000 were coalition governments. However, not all coalitions are alike. Scholars distinguish among several coalition types based on their size, ideological composition, and relationship to the legislative majority threshold. The following diagram maps these coalition types in terms of the strategic logic that produces them.

This flowchart traces the decision logic from election results to coalition type. When no party holds a majority, bargaining begins—and the resulting coalition type depends on whether parties prioritize office payoffs (MWC) or policy coherence (minimum connected winning coalition). Oversized coalitions, while seemingly irrational under the size principle, emerge when parties value stability, constitutional supermajorities, or crisis management.

Coalition Type	Definition	Real-World Example
Minimum Winning Coalition (MWC)	Holds a majority but would lose it if any partner withdrew. Maximizes portfolio share per partner.	Germany's CDU/CSU–FDP coalitions (e.g. 2009–2013): exactly two blocs combining for a slim majority.
Minimum Connected Winning Coalition	An MWC in which all partners are ideologically adjacent on the left-right spectrum, minimizing policy compromise.	Scandinavian red-green coalitions: Social Democrats + Greens + Left, all on the center-left.
Oversized (Surplus) Coalition	Includes more parties than strictly necessary for a majority, providing a legislative buffer.	Switzerland's 'magic formula' Federal Council: four parties sharing executive power since 1959.
Grand Coalition	The two largest parties (typically one center-left, one center-right) govern together, usually as a last resort.	Germany's CDU/CSU–SPD grand coalitions (2005–2009, 2013–2021): formed when smaller partners were insufficient.
Minority Government	A single party or small coalition governs without a formal majority, relying on ad hoc legislative support.	Denmark and Norway frequently produce minority governments sustained by confidence-and-supply agreements.

Major coalition types in parliamentary democracies

SECTION 6

Worked Example — Computing ENP & Identifying Coalition Options

Consider a hypothetical parliament of 200 seats distributed among five parties after an election under proportional representation. We will compute the effective number of parties, classify the system, and identify the set of possible minimum winning coalitions.

Analyzing a Five-Party Parliament

1
Step 1 — Identify Seat Shares

The election results yield: Party A = 70 seats (35%), Party B = 50 seats (25%), Party C = 40 seats (20%), Party D = 25 seats (12.5%), and Party E = 15 seats (7.5%). A governing majority requires at least 101 seats (50% + 1).

Seat proportions: pₐ = 0.35, $p_B$ = 0.25, $p_C$ = 0.20, $p_D$ = 0.125, $p_E$ = 0.075
2
Step 2 — Compute Squared Proportions

We square each proportion: pₐ² = 0.35² = 0.1225, $p_B$ ² = 0.25² = 0.0625, $p_C$ ² = 0.20² = 0.0400, $p_D$ ² = 0.125² = 0.015625, $p_E$ ² = 0.075² = 0.005625.

Σ pᵢ² = 0.1225 + 0.0625 + 0.0400 + 0.015625 + 0.005625 = 0.24625
3
Step 3 — Calculate ENP

Applying the Laakso–Taagepera formula: N = 1 / Σ pᵢ² = 1 / 0.24625 ≈ 4.06. Although there are five parties in the legislature, the effective number is approximately 4, reflecting the fact that the smallest party (E) carries relatively little weight. Under Sartori's typology, an ENP near 4.0 places this system in the moderate pluralism category.

ENP ≈ 4.06 → Moderate Pluralism
4
Step 4 — Enumerate Minimum Winning Coalitions

A minimum winning coalition (MWC) must command ≥ 101 seats but would fall below 101 if any single partner left. We systematically check combinations. A + B = 120 seats (removing A → 50, removing B → 70; both below 101) → MWC. A + C = 110 → MWC. A + D + E = 110 → removing D yields 85, removing E yields 95; both below 101 → MWC. B + C + D = 115 → removing D yields 90 (below 101) → MWC. B + C + E = 105 → removing E yields 90 (below 101) → MWC. A + D = 95 (below 101) → not viable. Several other combinations are either non-winning or oversized.

Five MWCs identified: {A,B}, {A,C}, {A,D,E}, {B,C,D}, {B,C,E}
5
Step 5 — Apply Policy-Seeking Logic

If we assume a left-right ordering A (center-right) – B (center) – C (center-left) – D (left) – E (far-left), then the minimum connected winning coalitions are those whose members are adjacent on this spectrum. {A,B} (center-right + center) and {B,C,D} (center + center-left + left) satisfy this criterion. {A,C} would skip over B, making it non-connected. A policy-seeking model would therefore predict either an {A,B} center-right government or a {B,C,D} center-left government as the most likely outcomes.

Policy-seeking prediction: {A,B} or {B,C,D} as minimum connected winning coalitions

SECTION 7

Strengths & Limitations of Party System Theories

The theoretical frameworks described in this lesson—Duverger's Law, Sartori's typology, and Riker's coalition theory—have proven remarkably productive, generating testable hypotheses and structuring decades of comparative research. However, each framework carries significant limitations that scholars have identified and debated. The following table summarizes the key strengths and weaknesses.

Framework	Strengths	Limitations
Duverger's Law / Hypothesis	Parsimonious and broadly predictive; correctly links FPTP to two-party tendencies and PR to multipartism; widely tested across dozens of democracies.	Fails in cases of strong regional/ethnic parties (e.g. India, Canada); treats electoral systems as static when they evolve; the 'law' is really a tendency with many exceptions.
Sartori's Typology	Adds qualitative ideological dimension to party system classification; distinguishes meaningfully between moderate and polarized multipartism; accounts for anti-system parties.	Category boundaries are somewhat subjective; harder to operationalize than ENP; developed primarily from European cases and may not translate perfectly to non-Western contexts.
Riker's Size Principle	Elegant formal prediction derived from game theory; provides a clear baseline against which to measure empirical deviations; focuses attention on bargaining dynamics.	Assumes perfect information and purely office-seeking actors; cannot explain oversized coalitions, minority governments, or the role of policy/ideology; real-world coalitions routinely violate the prediction.
ENP Index (Laakso–Taagepera)	Easy to compute; allows precise cross-national and cross-temporal comparison; widely adopted as a standard metric in the discipline.	A single number cannot capture ideological distance, coalition potential, or anti-system character; two systems with the same ENP can function very differently.

Comparative assessment of major party system and coalition frameworks

✦

KEY TAKEAWAY

No single theoretical framework captures the full complexity of party systems and coalition politics. The ENP provides quantitative precision but is blind to ideology; Sartori adds ideological depth but is harder to measure; Riker's coalition models offer elegant formal logic but struggle with empirical messiness. The most productive comparative analyses use these tools in combination, triangulating quantitative metrics with qualitative case knowledge—much as a physician combines lab results, imaging, and clinical judgment rather than relying on any single diagnostic tool.

SECTION 8

Connections to Advanced Theory

The foundational concepts covered in this lesson serve as the entry point to several advanced research programs in comparative politics. Understanding where the basic frameworks connect to more sophisticated models is essential for students moving into upper-division coursework and graduate study. The table below maps each introductory concept to its advanced theoretical extensions.

Introductory Concept	Advanced Extension
Duverger's Law	Cox's (1997) strategic coordination theory formalizes Duverger's mechanisms using game-theoretic models of voter and elite coordination, explaining when and why the 'law' fails (e.g. strong local parties, presidential systems).
Sartori's Typology	Mainwaring's work extends party system analysis to presidential systems and new democracies, showing how presidentialism interacts with multipartism to create governance challenges (the 'difficult combination').
Riker's Size Principle	Laver and Shepsle's (1996) portfolio allocation model adds institutional structure: coalition bargaining is shaped by which party controls which ministry, introducing spatial modeling and credible commitment problems.
Cleavage Theory (Lipset–Rokkan)	Post-materialist value change (Inglehart), dealignment, and the emergence of new parties (Greens, populist radical right) challenge the 'frozen cleavages' thesis and require dynamic models of party system change.
ENP Index	Electoral volatility indices (Pedersen index), party system institutionalization measures (Mainwaring & Scully), and decomposition techniques disaggregate fragmentation into its mechanical and behavioral components.

From introductory frameworks to advanced research programs

One particularly vibrant area of current research concerns the rise of populist and anti-establishment parties across established democracies. These parties challenge existing coalition norms—mainstream parties often refuse to include them in governing coalitions (the cordon sanitaire strategy), which constrains the coalition space and can lead to grand coalitions, minority governments, or novel cross-bloc arrangements. This phenomenon illustrates how the formal models of coalition theory must be supplemented by attention to norms, democratic values, and the politics of legitimacy.

SECTION 9

Practice Problems

Question 1: CONCEPTUAL

Explain why Sartori argued that simply counting the number of parties in a legislature is insufficient for classifying a party system. What additional dimension does his typology incorporate, and why does it matter for understanding political competition?

Show answer

Sartori argued that the raw number of parties tells us little about the dynamics of competition without also considering the ideological distance between them. His typology incorporates the dimension of polarization—specifically, the degree of ideological spread among relevant parties and the presence or absence of anti-system parties at the extremes. This matters because a five-party system with parties clustered near the center (moderate pluralism) produces centripetal competition and relatively easy coalition-building, whereas a five-party system with extreme poles (polarized pluralism) produces centrifugal competition, irresponsible opposition, and governance instability. The qualitative character of competition, not just the quantity of competitors, determines democratic performance.

Question 2: BASIC CALCULATION

A legislature has three parties with the following seat shares: Party X = 45%, Party Y = 35%, Party Z = 20%. Calculate the effective number of parties (ENP) using the Laakso–Taagepera formula and classify the resulting party system.

Show answer

First, convert percentages to proportions:

p_X

= 0.45,

p_Y

= 0.35,

p_Z

= 0.20. Square each: 0.45² = 0.2025, 0.35² = 0.1225, 0.20² = 0.0400. Sum the squares: 0.2025 + 0.1225 + 0.0400 = 0.3650. Apply the formula: N = 1 / 0.3650 ≈ 2.74. Although three parties exist, the ENP of approximately 2.74 indicates that the system functions closer to a 'two-and-a-half' party system—a configuration common in countries like Germany and Austria before the proliferation of smaller parties. Under Sartori's framework, this falls at the boundary between a two-party system and moderate pluralism.

Question 3: INTERMEDIATE

In a 150-seat parliament, five parties hold the following seats: A = 55, B = 40, C = 30, D = 15, E = 10. A majority requires 76 seats. List all minimum winning coalitions (MWCs) and identify which among them qualify as minimum connected winning coalitions if the ideological ordering from left to right is E – D – C – B – A.

Show answer

We need coalitions with ≥ 76 seats where removing any single party drops the total below 76. Checking systematically: {A,B} = 95 (remove A → 40, remove B → 55; both < 76) → MWC. {A,C} = 85 (remove A → 30, remove C → 55; both < 76) → MWC. {A,D} = 70 → not winning. {B,C,D} = 85 (remove D → 70 < 76) → MWC. {B,C,E} = 80 (remove E → 70 < 76) → MWC. {A,D,E} = 80 (remove D → 65, remove E → 70; both < 76) → MWC. {B,D,E} = 65 → not winning. Now checking connectivity on E–D–C–B–A: {A,B} spans only A and B (adjacent) → connected. {A,C} skips B → not connected. {B,C,D} spans B through D (adjacent chain) → connected. {B,C,E} skips D → not connected. {A,D,E} skips B and C → not connected. Therefore, the minimum connected winning coalitions are {A,B} and {B,C,D}.

Question 4: APPLIED

After Germany's 2021 federal election, the Bundestag included the following approximate seat shares: SPD (25.7%), CDU/CSU (24.1%), Greens (15.7%), FDP (11.5%), AfD (10.3%), Left (4.9%), SSW (0.1%), and others (7.7%). Compute the effective number of parliamentary parties and discuss why the resulting coalition (SPD–Greens–FDP, the 'traffic light' coalition) is surprising from a policy-seeking perspective.

Show answer

Using seat proportions (approximated): Σ pᵢ² ≈ 0.257² + 0.241² + 0.157² + 0.115² + 0.103² + 0.049² + 0.001² + 0.077² ≈ 0.066 + 0.058 + 0.025 + 0.013 + 0.011 + 0.002 + 0.000 + 0.006 ≈ 0.181. ENP ≈ 1/0.181 ≈ 5.52, placing Germany firmly in the moderate-to-high pluralism range. The 'traffic light' coalition (SPD + Greens + FDP) is surprising from a policy-seeking perspective because the FDP (classical liberal, economically right) is not ideologically adjacent to the Greens (left-libertarian, pro-regulation). A minimum connected winning coalition would have predicted either SPD–Greens–Left (left bloc) or CDU/CSU–FDP (center-right), both of which are more ideologically coherent. The traffic light coalition was driven by political realities—the Left party was too weak and internally divided, and the CDU/CSU was weakened by leadership struggles—illustrating how real-world coalition formation departs from pure theory.

Question 5: CRITICAL THINKING

Riker's minimum winning coalition theory and Sartori's polarized pluralism model make seemingly contradictory predictions about coalition size in highly fragmented, polarized party systems. Riker predicts minimal coalitions; Sartori's framework suggests that polarization constrains the viable coalition space, often producing oversized or grand coalitions. Develop an argument that reconciles these two perspectives, using at least one empirical example to support your reasoning.

Show answer

The apparent contradiction dissolves when we recognize that Riker's model assumes all parties are potential coalition partners, whereas Sartori's concept of polarized pluralism introduces anti-system parties that mainstream parties refuse to include. In a polarized system, the effective coalition space shrinks because parties at the extremes are excluded by a cordon sanitaire or by their own unwillingness to compromise. Within this constrained space, the set of possible MWCs may be empty—no combination of acceptable parties produces a bare majority. This forces mainstream parties into oversized coalitions or grand coalitions simply to achieve a governing majority without relying on anti-system partners. Italy's First Republic exemplifies this dynamic: the presence of the PCI (Communist) on the far left and the MSI (neo-fascist) on the far right excluded roughly 40% of parliamentary seats from coalition consideration, compelling the DC (Christian Democrats) to build oversized coalitions with multiple small centrist parties. Riker's logic still operates within the constrained space—parties do not add unnecessary partners beyond what the exclusion of extremist parties requires—but the structural constraint of polarization overrides the pure size principle prediction. Thus, the two frameworks are not contradictory but operate at different analytical levels: Riker describes the bargaining logic within a given coalition space; Sartori defines the boundaries of that space.

SUMMARY

Summary — Party Systems & Coalitions

Summary

A party system is the structured pattern of competitive and cooperative interaction among political parties, shaped by electoral rules (Duverger's Law links FPTP to two-party systems and PR to multipartism), social cleavages (the Lipset–Rokkan framework), and strategic behavior by voters and elites. The effective number of parties (ENP) index (N = 1 / Σ pᵢ²) provides a standard quantitative metric for system fragmentation, while Sartori's typology adds qualitative depth by distinguishing moderate from polarized pluralism based on ideological distance and the presence of anti-system parties.

In multiparty parliamentary systems, governance typically requires coalition formation. Riker's size principle predicts that office-seeking parties form minimum winning coalitions, but policy-seeking models predict minimum connected winning coalitions among ideologically adjacent partners. Real-world outcomes—including oversized coalitions, grand coalitions, and minority governments—reflect the interplay of office motivations, policy preferences, institutional constraints, and political norms. Mastering these frameworks equips students to analyze and compare democratic governance across a wide range of national contexts.