French Fourth Republic Roll Call Data

 

Howard Rosenthal and Erik Voeten

 

 

Citations

Description of Data

Datasets

Variable Definitions

Vote choices

Party Codes

Status Codes

Data Download

 

Citations

 

If you use this data, please cite the following two works:

 

Howard Rosenthal and Erik Voeten. 2004. “Analyzing Roll Calls with Perfect Spatial Voting: France 1946-1958.” American Journal of Political Science 48(3): 620-32. (for a more detailed analysis, you may also want to look at our paper: Government Instability with Perfect Spatial Voting: France 1946-1958 , which is available as Russell Sage working paper #197.)

 

Duncan MacRae Jr. 1967. Parliament, Parties, and Society in France 1946-1958. New York: St. Martin’s Press.

 

Description of Data

 

The data are based on Duncan MacRae’s French Representation Study 1946-58 (ICPSR Study No. 52). The original data separates out the votes of legislators and their party and or cabinet status, something that changed frequently for many legislators even within a legislative session. Codebooks are available from the ICPSR.

 

We have read the data from card format into a format suitable for modern statistical software. We have also scanned information about the roll-calls in the codebooks such that it is now integrated into the databases. We have also reorganized the data such that it is more suitable for analysis.

 

The data are from the first (1946-51), second (1951-56) and final (1956-58) legislatures of the Fourth Republic. The sample of roll calls consists of two parts (See MacRae, Appendix A, for more detail). First, it contains all 739 roll calls mentioned in L’Année Politique. Sampling in this way is roughly similar to constructing a sample for Congress using only “key” votes reported in Congressional Quarterly Weekly.  L’Année Politique includes important votes, such as investitures, confidence votes, important budget votes, and other policy votes.  Second, it includes a random sample of 50 votes per legislature from the remainder of all roll call votes.

 

Among the legislators in the dataset are well-known French politicians such as Charles de Gaulle, Valerie Giscard d’Estaing, Francois Mitterand, Pierre Mendes-France, and Jean-Marie Le Pen. The sample includes votes on topics such as NATO, the European Defense Community, Algeria, Indo-China, Italy treaty, Paris accords, military budget, atomic energy development, labor strike regulations, taxes, public health, Parisian transportation, colliery schools, National Assembly reform, agriculture, land reform, fiscal reform, inflation, salary scales, and amnesty.  It includes a total of 33 votes to invest a new prime minister, and 136 votes of confidence. The data allow the analyst to identify what vote is an investiture/confidence vote and whether a legislator is in the cabinet at the time of the vote. It is thus very suitable for analyses of cabinet formation in a parliamentary system.

 

Datasets

 

  1. Pooled. Uses legislators as the unit of analysis with their roll-call vote choices in the columns. The file is pooled over all three legislatures. Assigns a party code for each legislator based on the first party he/she belongs to in each legislature (party1, party2, party3).

 

  1. Partyswitcher. This dataset creates a new unit of analysis when a legislator changes party affiliation (see Rosenthal and Voeten 2004). MacRae assigned party codes for each individual roll call. We are therefore able to identify exactly when deputies decided to change their party affiliation. We include only changes within the major party categories given in Rosenthal and Voeten (2004).  Thus, we do not treat as switches changes between party groups we identify as “Right” (see also below).  We only count alterations if these endured for at least 25 votes.  We do treat deputies who become unaffiliated with any party as switching. 

 

  1. Legislator Cabinet and Party Status. Provides information for each legislator on each vote what party he/she is in at that time and whether the legislator is currently a member of the cabinet. This file is useful for more detailed analyses of cabinet formation in a parliamentary system.

 

  1. Roll Call Information. The unit of analysis is the roll call. The file includes an identification variable that matches the variable names in the first three files (itemid). It further includes a verbal description of the issue, a categorization variable (categ) that identifies whether the vote is an investiture vote (1), a confidence vote (2), part of the random sample (3), or another type of vote (0). It further contains information that refers to L’Année Politique (roll-call number (scrutin), page of description) and the roll-call marginals (pour and contre). All information is based on a scan of original codebooks. A few roll calls are missing due to some missing information in the codebook. None of the missing votes are investiture or confidence votes.

 

  1. Legislator Ideal Points and Scales. The unit of analysis is the legislator. This file includes two-dimensional W-NOMINATE and optimal Classification ideal point estimates for legislators in individual legislatures and pooled over all three legislatures. It also includes a set of scales that MacRae developed for his book and were the basis for many of his analyses on intra-party conflict. For more on these, see MacRae’s book cited above (we transferred the legislator scores on scales from the tables to the database).

Variable Definitions

 

Identification variables: Mid is the unique legislator identifier that should be used when merging files. Id1, id2, id3 are ids that can be used when analyzing separate legislatures.

 

Itemid identifies the roll-calls in the roll call information file and corresponds to the variable numbers in the other three data-files (so v1001 corresponds to the case where itemid has the value 1001). V1001- are votes taken in the first legislature, v2001- in the second legislature, v3001- in the third.

 

Vote choices

 

1 = “Yes”

6 = “No”

7 = “Abstain”

8 = “Not Present”

9 = “Not Seated”

 

Party Codes

 

Two types of party codes are used. First, MacRae’s original party-codes. These codes vary by legislature!!! In the pooled data:

 

Party1: 0 'Not seated' 1 'Communiste' 2 'Progressiste' 3 'Rass. Dem. Afr.' 4 'Socialiste' 5 'Musul, Ind.' 6 'TLD Algerie' 7 'Gauche Ind.' 8 'Radical-Soc.' 9 'MRP' 10 'UDSR' 11 'Ind. Outremer' 12 'PRL' 13 'Act. Pays Soc.' 14 'Un. Dem. Soc.' 15 'Rep. Dem. Soc.' 16 'Act. Dem. Soc.' 17 'Rep. Pop. Ind.' 18 'Non Inscrit'.

 

Party2: 0 'Not seated' 1 'Communiste' 2 'Progressiste' 3 'Rass. Dem. Afr.' 4 'Socialiste' 5 'Musul, Ind.' 6 'TLD Algerie' 7 'Gauche Ind.' 8 'Radical-Soc.' 9 'MRP' 10 'UDSR' 11 'Ind. Outremer' 12 'Ind. Paysan' 13 'Paysan' 14 'Cent. Dem. Soc.' 15 'Rep. Indep.' 16 'RPF, URAS, RS'  17 'ARS' 18 'Non Inscrit'.

 

Party 3: 0 'Not seated' 1 'Communiste' 2 'Progressiste' 3 'Rass. Dem. Afr.' 4 'Socialiste' 5 'Mouv. Soc. Afr' 6 'Rad. (Morice)' 7 'Rad. (Faure)' 8 'Radical-Soc.' 9 'MRP' 10 'UDSR' 11 'Ind. Outremer' 12 '' 13 'Paysan' 14 'IPAS-Rurale' 15 'IPAS' 16 'Rep. Soc.'  17 'Uff (Poujade)' 18 'Non Inscrit'.

 

The second coding scheme groups the individual parties into party families (see Rosenthal and Voeten 2004). This coding is used in the pooled data for party1new, party2new, party3new and is the basis for the partyswitcher data:

 

0 'Not seated'  1 'Communists' 2 'Socialists' 3 'Radicals' 4 'Rural parties' 5 'Christian Democrats' 6 'Gaullists' 7 'Poujadists' 10 'Other' 11 'Not affiliated'.

 

The Right includes Parti Républicain de la Liberté, Républicains Indépendants, Indépendants, Paysans, Indépendants d’Outremer, Indépendants Paysans, Paysan, IPAS. The Gaullists include RPF, URAS, and ARS.

 

Below are the SPSS recode-statements to go from the MacRae-coding to the common party-grouping coding:

 

RECODE

party1

  (0=0)  (1=1)  (4=2)  (2=10)  (5=10)  (8=3)  (9=5)  (10=3)  (12=4)  (13=4)

  (15=4)  (18=11)  (SYSMIS=0) (else=10)  into party1new .

 

 

RECODE

  party2

  (0=0)  (1=1)  (3=10) (4=2)  (2=10)  (7=10)  (8=3)  (9=5)  (10=3) (11=4) (12=4)  (13=4)

  (15=4) (16=6) (17=6) (18=11) (SYSMIS=0) (else=10)  into party2new .

 

 

RECODE

 party3

  (0=0)  (1=1)   (4=2)  (2=10)  (7=3)  (8=3)  (9=5)  (10=3) (11=4)   (13=4)

  (15=4) (16=6) (17=7) (18=11)   (SYSMIS=0) (else=10)  into party3new.

 

Status Codes

 

The codes in the legislator status data file correspond to the original MacRae partycodes (see above for how to transform these, remember v1001- is first legislature, v2001- second legislature, v3001- third legislature). If a legislator is in the cabinet at the moment of the vote, 100 is added to the code. So a value of 101 for variable 2001 refers to a communist who is in the cabinet at the time vote id 2001 is taken. A value 1 on the same vote corresponds to a communist not in the cabinet.

 

Scales and Ideal Points

 

X12 is the 1st dimension NOMINATE coordinate in the 1st legislature (from a 2-dimensional solution), y12  the 2nd dimension coordinate. X22 is the 1st dimension coordinate in the second legislature, etcetera. _o indicates optimal classification coordinates. Variable labels are descriptive. File also includes legislator specific fit characteristics (GMP for NOMINATE, classification proportion for OC).

 

Scales are all named after the tables in MacRae’s book. So, sca731 is the first scale in table 7.3. Variable labels are descriptive of scale content.

Data Download