Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Electoral Polls and Economic Uncertainty: an Analysis of the Last Two U.S. Presidential Elections

Giampiero M. Gallo, Demetrio Lacava, Edoardo Otranto

TL;DR

Using daily Trump polling data and multiple uncertainty proxies, the paper investigates how voter sentiment co-moves with economic and financial uncertainty during the 2020 and 2024 U.S. elections. It applies Dynamic Conditional Correlation and extensions NLARC to model time-varying correlations, with a two-step estimation involving univariate GARCH volatilities and de-GARCHed residuals. The results show a strong, event-driven link between polls and uncertainty in 2020, but correlations near zero and stable in 2024, suggesting polarization and non-economic shocks muted the economic channel. The study contributes methodologically by applying DCC and NLARC to political data and provides policy-relevant insights into when economic fundamentals influence electoral dynamics.

Abstract

This paper examines the dynamic relationship between electoral polls and indicators of economic and financial uncertainty during the last two U.S. presidential elections (2020 and 2024). Using daily polling data on Donald Trump and measures such as the Aruoba-Diebold-Scotti Business Conditions Index, the 5-year Breakeven Inflation Rate, the Trade Policy Uncertainty index, and the VIX, we estimate conditional correlation models to capture time-varying interactions. The analysis reveals that in 2020, correlations between polls and uncertainty measures were highly dynamic and event-driven, reflecting the influence of exogenous shocks (COVID-19, oil price collapse) and political milestones (primaries, debates). In contrast, during the 2024 campaign, correlations remained close to zero, stable, and largely unresponsive to shocks, suggesting that entrenched polarization and non-economic events (e.g., assassination attempt, candidate changes) muted the economic channel. The study highlights how the interplay between voter sentiment, financial markets, and uncertainty varies across electoral contexts, offering a methodological contribution through the application of Dynamic Conditional Correlation models to political data and policy-relevant insights on the conditions under which economic fundamentals influence electoral dynamics.

Electoral Polls and Economic Uncertainty: an Analysis of the Last Two U.S. Presidential Elections

TL;DR

Using daily Trump polling data and multiple uncertainty proxies, the paper investigates how voter sentiment co-moves with economic and financial uncertainty during the 2020 and 2024 U.S. elections. It applies Dynamic Conditional Correlation and extensions NLARC to model time-varying correlations, with a two-step estimation involving univariate GARCH volatilities and de-GARCHed residuals. The results show a strong, event-driven link between polls and uncertainty in 2020, but correlations near zero and stable in 2024, suggesting polarization and non-economic shocks muted the economic channel. The study contributes methodologically by applying DCC and NLARC to political data and provides policy-relevant insights into when economic fundamentals influence electoral dynamics.

Abstract

This paper examines the dynamic relationship between electoral polls and indicators of economic and financial uncertainty during the last two U.S. presidential elections (2020 and 2024). Using daily polling data on Donald Trump and measures such as the Aruoba-Diebold-Scotti Business Conditions Index, the 5-year Breakeven Inflation Rate, the Trade Policy Uncertainty index, and the VIX, we estimate conditional correlation models to capture time-varying interactions. The analysis reveals that in 2020, correlations between polls and uncertainty measures were highly dynamic and event-driven, reflecting the influence of exogenous shocks (COVID-19, oil price collapse) and political milestones (primaries, debates). In contrast, during the 2024 campaign, correlations remained close to zero, stable, and largely unresponsive to shocks, suggesting that entrenched polarization and non-economic events (e.g., assassination attempt, candidate changes) muted the economic channel. The study highlights how the interplay between voter sentiment, financial markets, and uncertainty varies across electoral contexts, offering a methodological contribution through the application of Dynamic Conditional Correlation models to political data and policy-relevant insights on the conditions under which economic fundamentals influence electoral dynamics.
Paper Structure (5 sections, 8 equations, 5 figures, 2 tables)

This paper contains 5 sections, 8 equations, 5 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Dynamic Correlation Models
  3. Preliminary Analysis and First-step Estimation
  4. Second-step Estimation: Dynamic and Constant Correlations
  5. Final remarks

Figures (5)

  • Figure 1: Time-series of the polls in favor of Trump during the 2020 electoral campaign. The vertical lines represent some events related to inflation (blue vertical line), business condition (gray), VIX (red), and trade uncertainty (green). Sample period: March 5, 2020 - November 3, 2020.
  • Figure 2: 5-day rolling correlations (black line) end relevant electoral events (gray vertical lines, see text). Sample period: March 5, 2020 - November 3, 2020.
  • Figure 3: Estimated conditional correlations (black line) from the DCC model and relevant electoral events (gray vertical lines, see text). Sample period: March 5, 2020 - November 3, 2020.
  • Figure 4: Estimated conditional correlations (black line) from the DCC model. Sample period: March 8, 2024 - November 5, 2024.
  • Figure 5: Time-series of the polls in favor of Trump during the 2024 electoral campaign. The vertical lines represent relevant electoral events (see text). Sample period: March 8, 2024 - November 5, 2024