Do incumbents and opposition mobilize differently?

Rallies as Political Campaigning

Campaign rallies are one of the most visible campaign activities in electoral politics, both in India and elsewhere.

They are widely believed to influence:

• Voter mobilization (Snyder and Yousaf 2020; Sanders and Kaid 1981)
• Media coverage (Brierly and Kramon 2020; Shaw 1999; Lynch 2023)
• Campaign momentum (Feinberg, Branton, and Martinez-Ebers 2022)

These rallies at once function as a Get-out-the-vote (GOTV) measure, a potential tool to energize existing support base, and attract new voters.

We ask three questions:

• Where do political parties place rallies?
• Do incumbents and opposition differ in their placement?
• Do these rallies have an effect?

Summary of Key Findings

Defining Rallies

We define a campaign rally as an episodic, high-intensity collective gathering, characterized by the mass mobilization of the electorate in the pre-electoral period, centered around the strategic presence of elite political actors.

These large scale spectacles very effectively generate support among otherwise dormant supporters by their sheer scale, environment, and grandeur(Paget, Beardsworth, and Lynch 2023)

This grandeur, however, is precisely what constitutes the double-edged nature of rallies as campaign tools.

While early scholarship categorized rallies as primarily ‘labor-intensive’ and non-modern modes of mobilization (Norris 2000) in developed democracies, contemporary research suggests a significant shift. Recent evidence indicates that modern rallies have transitioned into ‘capital-intensive’ operations (Lynch 2023).

Hole in the pocket

Studies argue that in terms of the unit cost of persuasion, rallies occupy an intermediate position: costlier than SMS or automated outreach, yet cheaper than door-to-door canvassing or direct distributive transfers (Enos and Fowler 2018; Brierly and Kramon 2020).

Modern rallies act as hybrid interventions that incorporate elements of localized canvassing, media coverage, and material distribution.

The organization of modern rallies necessitates significant capital outlays.

The bulk of this financial burden is driven by the high overhead of securing and preparing large-scale venues, the logistics of mobilizing and transporting the electorate, and the expenditure required for the transit of political elites. Furthermore, the provision of distributive goods—often essential for turnout—compounds the economic burden, making each deployment a high-stakes investment.

The Time Crunch

Leaders are the campaign’s most valuable—but limited—resource. They face two major bottlenecks:

1. The Day Job vs. The Campaign: Top leaders aren’t just campaigners; they have administrative duties. Every day spent on a stage is a day away from running the government or the party.
2. Too Many Seats, Too Little Time:
   • The Numbers Game: There are hundreds of seats voting, but only a few weeks to visit them.
   • Logistical Lag: A big rally needs days of setup, security checks, and crowd planning.
   • High Demand: Every local candidate wants the “Star” in their backyard.

India: The Case Study

To address these questions, we analyze the world’s largest elections. Indian parliamentary (Lok Sabha) elections are held every five years, with the schedule typically announced four months prior to the end of the term.

Data & Scope

We focus on the two most prominent “Star Campaigners” during the 2019 and 2024 election cycles: Narendra Modi (Bharatiya Janata Party) & Rahul Gandhi (Indian National Congress)

Definitions & Parameters

• Temporal Scope: Excluding “pre-campaign” events. The data covers only the window between the formal announcement of the schedule and the final day of polling.
• Event Formats: We include both high-energy roadshows and large-scale public meetings (Jan Sabhas).

Rally Targeting: 2019 General Election

Rally Targeting: 2024 General Election

The Three-Headed Strategic Logic

Rally deployment is a nested decision across three distinct dimensions:

1. The “When” (Temporal)

Selection by Phase: With elections spread over weeks, parties must decide which phase requires a “star” intervention.

2. The “Where” (Extensive Margin)

Selection by District (PC): Is the seat a “Safe Haven,” a “Battleground,” or a “Lost Cause”?

3. The “Where Within” (Intensive Margin)

Selection by Local Venue (AC): Strategic placement within a PC to influence specific demographic clusters or “swing” assembly segments.

PC Targeting Model

PC Targeting — Decentralized (State × Phase FE)

Alternative Specification: Centralized Targeting

Relaxes state-level coordination; assumes a highly centralized campaign strategy.

\[ rally_{pc} = \mathbf{X}_{pc}'\beta + \alpha_{phase} + \gamma_{year} \]

• \(\alpha_{phase}\): captures phase-specific timing effects common across all states
• \(\gamma_{year}\): controls for election-wide differences

Interpretation: Identification comes from cross-constituency variation within phases, pooling across states.

PC Targeting — Centralized (Phase FE Only)

Discussion: PC Targeting Logic

• Narendra Modi:
  • Victory Margin is the most consistent predictor (\(p < 0.001\)). Confirms a “Swing Seat” logic.
  • Resource Optimization: Driven by political variables (incumbency/margins) rather than demographics.
• Rahul Gandhi:
  • Victory Margin is statistically insignificant.
  • Institutional Presence: Higher likelihood of rallying for INC Incumbents (\(0.177^*\)).
• The “Alliance” Effect: Both prioritize their own party over alliance partners.

Stage 3: The Within-AC Selection Model

Once a Parliamentary Constituency (\(pc\)) is targeted:

\[ rally_{ac,t} = \beta_1 \log(electors)_{ac,t} + \beta_2 touch\_count_{ac} + \beta_3 \Delta electors_{ac,t} + \mathbf{Z}_{ac}'\delta + \alpha_{pc,t} + \epsilon_{ac,t} \]

Where: * \(\alpha_{pc,t}\): Fixed effects for the Parent PC \(\times\) Election Year. * \(touch\_count_{ac}\): Number of neighboring PCs (measures Centrality). * \(\mathbf{Z}_{ac}'\): Controls for Vegetation (VCF) and Terrain (TRI).

AC-Level Targeting Results

Discussion: The Logistics of Centrality

• Avoidance of Borders: PC_touch_count is negative and significant (\(p < 0.001\)). Campaigns prefer the geographic center to maximize internal pull.
• Scale: Strong preference for larger ACs (log(ac_electors)), anchoring to urban/semi-urban hubs.
• Infrastructure: For Rahul Gandhi, vcf (Vegetation) is negative and significant (\(-0.007^*\)), suggesting placement in cleared/developed areas.

Summary: Political Divergence, Logistical Convergence

The Mobilizing Effect of Rallies on Turnout

Discussion: Turnout & Mobilization

• The Turnout Bump: A visit from either leader is associated with a ~0.82 pp increase in turnout (\(p < 0.01\)).
• Sticky Participation: Previous turnout remains the strongest predictor.
• Uniformity: Despite divergent strategy on where to go, the impact of the spectacle is uniform.

Spatial Spillovers: Do Rallies Boost Adjacent Turnout?

Analysis: The Spillover Divergence

• Direct Mobilization: Both see a primary turnout bump of ~1.25 pp.
• The Rahul Spillover: Rahul’s rallies result in significant “Secondary Mobilization” in neighboring PCs (+0.69 pp, \(p < 0.01\)).
• Model Fit: High proportion of variance explained (Adj. \(R^2 \approx 0.87\)).

Direct Impact: Rally Effect on Vote Share

Spatial Spillovers: Vote Share Expansion

Analysis: The Regional Persuasion Effect

• The Rahul “Multiplier”: INC rallies act as powerful “Regional Converters.” They boost vote share in the host PC (+4.17 pp) and create a significant “halo effect” in neighboring constituencies (+1.69 pp, \(p < 0.1\)).
• The BJP Baseline: BJP vote share is highly structural and less sensitive to marginal rally interventions.
• “Saturation Effect”: where the base vote share is already high, the marginal utility of a single rally for vote conversion is diminished.
• Persuasion vs. Mobilization: While the BJP uses rallies to successfully maintain turnout, the INC relies on them as essential tools to actively expand and consolidate their vote share from a lower baseline.