We include elections and cabinets in established democracies. More specifically, we include democratic national lower house elections and EP elections for all EU and most OECD members. For the latter, we exclude presidential systems. We record all elections and succeeding cabinets after 1945 or after full democratization according to Boix ea. (2013).
We have started to include elections and cabinets before 1945. We record information after 1900 or after the last democratic transition (Boix ea. 2013). As of today, these observations are experimental and may be incomplete.
We include parties winning more than 1.0% vote share in elections we cover
We avoid including minor parties or candidates that win a seat only in one election due to low threshold requirements.
New parties are recorded in 'party_change' and 'party_name_change'.
Mergers and party splits are only added as a new party if the (largest) predecessor party won less than 75% of the combined vote of all preceding parties in the last election. Otherwise the largest party is just renamed.
Party names are sentence case if no national (language) convention requires title case
Party names and delimiters · recoding to be finished
Year added to smaller party name if identical names exist in a country (eg. SZ-92; Green Party -- 1992; Strana zelených – 1992)
We provide aggregated party positions in four major dimensions. These positions are time-invariant unweighted mean values of information from party expert surveys on a 0 to 10 scale. All expert surveys are linked with ids from the 'party' table and original values are rescaled before calculating the mean
Missing party positions for each dimension are imputed by mean values for the respective party family. We distinguish mean and imputed values by the number of decimal places. Mean values based on external datasets have five decimal places and imputed values have one decimal place only.
Variables and sources
We classify parties into families by their position in an economic (state/market) and a cultural (liberty/authority) left/right dimension.
The classification leads to eight party family categories: Communist/Socialist, Green/Ecologist, Social democracy, Liberal, Christian democracy, Agrarian, Conservative, Right-wing.
We add further information about party families in a separate table (see party_family).
We include election results for
Earlier versions of ParlGov included only parties with seats in parliament and updating all countries to the 1.0% vote share rule was completed in December 2014.
Sorted by preference for coding selection
We record a new cabinet for these events (cf. Budge/Keman 1993: 10)
All parties with ministers in cabinet are included (Indridason/Bowler 2014: 396)
Damgaard (1994: 194-95) and Müller/Strom (2000: 12) provide a more comprehensive discussion of cabinet definitions.
Cabinets with a limited legislative mandate (cf. McDonnell/Valbruzzi 2014)
Government status (minority, minimum winning or surplus majority) is determined only by the seat share of government parties in parliament and not coded manually. Currently, the script does not take into account any changes in the composition of parliament during the legislative term.
If there is an electoral alliance with separate seat shares but one of the parties is not a cabinet member, the government will be treated as a minority government (eg. UK 1951). Similarly, if any of the governing parties can be removed and the other governing parties still hold a majority in parliament, the cabinet is considered to be a surplus majority cabinet.
Experimental version only — coding has yet to be completed
Coding rule: The same events that define a new government are used to define the termination date of a government. Cabinets may remain in office for a short period after the initial terminal event. The 'description' field should give a short description of events other than an election that led to the fall of a government and these events should be coded in the 'data_json' field.
We include all party-affiliated heads of state. Short-term acting presidents are not included. See table president.
For new cabinets and revisions of observations, we derive information about cabinet termination from news sources, preferably from the news agencies Reuters, AFP or the main national news agency
For West European countries we double checked our initial observations with the data in Müller/Strom (2000). For Central- and Eastern European countries we compared our information to Müller-Rommel ea. (2008).