Jack Bailey
Research Interests
Make your next move.
Prepare your application
Review drafts before you send them.
Explore related searches
Articles (19)
The Gallagher index is the most popular measure of disproportionality in political science. While many assume that it can range from 0 (perfect proportionality) to 1 (perfect disproportionality), I prove that its maximum value is constrained by the size of the party system. For any election which features N_V effective vote-winning parties, the Gallagher index cannot exceed √(N_V + 1)/(2N_V). In some cases, this is a problem, since it makes it difficult to compare disproportionality across time and space, especially given rising fragmentation. As such, I propose a new normalised index that runs from 0 to 1, no matter the size of the party system.
Year:
2026
Brexit accelerated the long-term realignment of how demographic groups vote in Britain. However, Europe is no longer salient to voters, support for the pro-Brexit Conservatives has collapsed, and the 2024 UK General Election was the most volatile on record. So, did the Brexit realignment persist? We address this question by examining the relationship between Brexit preferences and demographic groups with vote choice. We find that the demographic alignments that underpinned Brexit support continue to shape political behaviour through support for party blocs. Thus, while Brexit itself no longer drives party competition directly, the trends that it intensified are still crucial for understanding British politics.
Year:
2025
Social democratic parties across Europe have increasingly adopted restrictive immigration rhetoric in response to radical right challengers' electoral successes. The consequences of accommodation remain contested. This paper leverages a pre-registered quasi-experimental design exploiting within-individual variation from as-good-as-random exposure to UK Labour leader Keir Starmer’s “Island of Strangers” speech -- a major rhetorical turn toward anti-immigration positions. We find that exposure significantly altered perceptions of Labour, making it appear more anti-immigration and right-leaning. Crucially, these shifts carried electoral costs: support for Labour declined within individuals following the speech. While we detect no corresponding gains for the radical right Reform UK party, there is also no evidence that Labour’s adoption of nativist rhetoric diminished Reform’s appeal. Our findings highlight the risks of strategic convergence, showing that accommodation of exclusionary rhetoric by social democratic parties is electorally costly for them rather than for their radical right competitors.
Year:
2025
The Gallagher index of disproportionality is perhaps the most popular measure of electoral distortion in all of political science. While many assume that the index ranges from 0 (perfect proportionality) to 1 (perfect disproportionality), I prove that in any free and fair election its maximum possible value is bounded above by sqrt((N-1)/(2N)), reaching a maximum of 1/sqrt(2) where the number of parties, N, tends to infinity. This bound arises in the limiting case where all parties receive equal vote shares but only one secures representation. Building on this result, I then propose a new normalised version of the index that rescales disproportionality to lie between 0 and 1 under democratic conditions.
Year:
2025
The Gallagher index of disproportionality is perhaps the most popular measure of electoral distortion in all of political science. While many assume that the index ranges from 0 (perfect proportionality) to 1 (perfect disproportionality), I prove that in any free and fair election its maximum possible value is bounded above by sqrt((N-1)/(2N)), reaching a maximum of 1/sqrt(2) where the number of parties, N, tends to infinity. This bound arises in the limiting case where all parties receive equal vote shares but only one secures representation. Building on this result, I then propose a new normalised version of the index that rescales disproportionality to lie between 0 and 1 under democratic conditions.
Year:
2025
I present a new model of the effective district-wide threshold: the vote share at which any given party has a 50/50 chance of winning its first seat in a district. To derive my model, I rely only on simple combinatorics. The resulting expression suggests both that, all else being equal, we should expect parties to be more likely than not to win at least one seat in a district and that the effective district-wide threshold depends only on the district’s magnitude and the number of vote-winning parties. Using district-level data from 375 elections in 110 countries, I then estimate a non-linear Bayesian model that strongly corroborates my theoretical logic. As such, the model that I present here serves both to advance the broader Taageperan research agenda of interlocking electoral models and allows for new quasi-experimental research designs that exploit as-good-as-random variation around electoral thresholds.
Year:
2025
I present a new model of the effective district-wide threshold: the vote share at which any given party has a 50/50 chance of winning its first seat in a district. To derive my model, I rely only on simple combinatorics. The resulting expression suggests both that, all else being equal, we should expect parties to be more likely than not to win at least one seat in a district and that the effective district-wide threshold depends only on the district’s magnitude and the number of vote-winning parties. Using district-level data from 375 elections in 110 countries, I then estimate a non-linear Bayesian model that strongly corroborates my theoretical logic. As such, the model that I present here serves both to advance the broader Taageperan research agenda of interlocking electoral models and allows for new quasi-experimental research designs that exploit as-good-as-random variation around electoral thresholds.
Year:
2025
I present a new model of the effective district-wide threshold: the vote share at which any given party has a 50/50 chance of winning its first seat in a district. To derive my model, I rely only on simple combinatorics. The resulting expression suggests both that, all else being equal, we should expect parties to be more likely than not to win at least one seat in a district and that the effective district-wide threshold depends only on the district’s magnitude and the number of vote-winning parties. Using district-level data from 375 elections in 110 countries, I then estimate a non-linear Bayesian model that strongly corroborates my theoretical logic. As such, the model that I present here serves both to advance the broader Taageperan research agenda of interlocking electoral models and allows for new quasi-experimental research designs that exploit as-good-as-random variation around electoral thresholds.
Year:
2025
Information theory now informs fields as diverse as quantum physics, neuroscience, and economics. In this short note, I show that it informs political science too. In particular, I show that the effective number of parties—perhaps the most widely-used measure in all of comparative politics—is the antilogarithm of a common measure of information entropy: the Rényi entropy. This equivalence bridges the two fields, allowing advances in information theory to inform our understanding of electoral systems.
Year:
2025
Information theory now informs fields as diverse as quantum physics, neuroscience, and economics. In this short note, I show that it informs political science too. In particular, I show that the effective number of parties—perhaps the most widely-used measure in all of comparative politics—is the antilogarithm of a common measure of information entropy: the Rényi entropy. This equivalence bridges the two fields, thereby allowing us to use advances in information theory to improve our understanding of electoral systems and their consequences.
Year:
2025
Year:
2025
Year:
2024
Year:
2024
Year:
2024
Year:
2023
Year:
2023
Year:
2021
Year:
2021
Year:
2020
Collaborators (5)
Jonathan Mellon
Assistant Professor
United States Military Academy
Robert Johns
Professor of Politics
University of Southampton
Stuart James Turnbull-Dugarte
Associate Professor in Quantitative Political Science
University of Southampton
Christopher Prosser
Associate Professor
Royal Holloway, University of London
Will Jennings
University of Southampton
Make your next move.
Prepare your application
Review drafts before you send them.