= Party Systems, Democratic Positions, and Regime Changes: Introducing the Party-System Democracy Index =

'''Party Systems, Democratic Positions, and Regime Changes: Introducing the Party-System Democracy Index''' (DOI: https://doi.org/10.1017/S0007123424000887) was written by Fabio Angiolillo, Felix Wiebrecht, and Staffan I. Lindberg in 2025. It was published in the ''British Journal of Political Science'' (vol. 55).

The authors investigate the role of parties in the [[PoliticalScience/Democratization#Waves|third wave of autocratization]]. They construct a novel index of democratic positions in party systems, i.e., preference for democracy or preference for autocracy. At the far ends of the index, the parties in a system are homogenously in favor of democracy (autocracy).
 * Use [[VarietiesOfDemocracy/VarietyOfPartyIdentityAndOrganization|V-Party]], particularly the anti-pluralist index (API) (v2xpa_antiplural) which measures commitment to core elements of democracy.
 * Each party's API is weighted by seat share.

The authors hypothesize that autocratization is made more likely when the opposition does not defend democracy. And bearing in mind that autocracies generally do not allow for oppositions, they hypothesize that the emergence of an opposition makes democratization more likely. Therefore moving toward the interior of the index is a marker for upcoming regime change.

The test this with [[Statistics/ProbitModel|probit regressions]] on the novel index (and the square of the index, anticipating nonlinear effects). Autocratization of democratic regimes and democratization of autocratic regimes are modeled separately.
 * Dependent variables drawn from [[VarietiesOfDemocracy/EpisodesOfRegimeTransformation|ERT]].
 * They control for the cubic transformation of regime age, as recommended by Maerz et al. (2023)
 * They also "use a set of economic variables from the [[MaddisonProject|Maddison Project Database]] (MPD) in the form of one-year per cent GDP growth, GDP per capita (logged), and one-year per cent GDP per capita change".
 * They take several 1-year lagged covariates from [[VarietiesOfDemocracy|V-Dem]]: "average democratic level of the political regions", "whether the incumbent accepted electoral outcomes, levels of legislative constraints, and the extent of civil society participation".
 * They control for Marquardt's (2021) measure of 'identity-based exclusion'. It aggregates data from V-Dem and from the Ethnic Power Relations Project.

The authors also fit models with lagged levels (by 2 or 3 electoral cycles) of the novel index, as to test if there is a near- or long-term causal effect instead of an inmediate one.

The authors find a small but statistically significant effect on autocratization. This is true in both the immediate and near-term formulations. Effects on democratization are not statistically different from zero.

The authors apply [[MeasurementValidity|Adcock and Collier's framework]] to validate their index.

For content validity, they test the index against V-Dem’s Electoral Democracy Index (EDI), Polity5's Polity Score, and [[FreedomInTheWorld|Freedom in the World]]. The expectation is that highly democratic nations should have a party system that strongly prefers democracy. They do note that correlation of low democracy and autocratic preferences is tautological, as autocratic systems feature a "''de jure'' or ''de facto'' lack of opposition parties".

They also test it by examining how the index changes over time in nations and years which we know autocratization (democratization) occurred.

For convergent validity, they test the correlations with [[DemocraciesAndDictatorshipsInLatinAmerica|Mainwaring and Pérez-Liñán's measure]] and then with the aforementioned measure from Marquardt (2021)

For construct validity, they test the index against left-right positions. They use two versions of party labels: V-Party and Chapel Hill expert surveys in Europe (CHES-Europe). They transform the party-level data into an index using the same (seat share weighted) method as for their novel index itself. This test is performed with the expectation that "democratic regime preferences can be
undermined when party systems’ centrifugal directions are predominant".

They also test against Kim's (2023) party-system institutionalization index, expecting a correlation between democracy and institutionalism. Note this is only defined for democracies.

They also test against constitutional structure: presidential, semi-presidential, and parliamentary. Parliamentary systems are generally accepted as being more democratic.

Lastly, they replicate the aforementioned work by Mainwaring and Pérez-Liñán while replacing their measure for the author's novel index.



----
CategoryRicottone CategoryReadingNotes