The Role of French Governments in Legislative Agenda Setting

Transcription

1 6/04/05 1 The Role of French Governments in Legislative Agenda Setting Sylvain BROUARD CEVIPOF, Sciences-Po Paper presented at the ECPR Joint Sessions, Granada, April Workshop 22 : The Role of Government in Legislative Agenda Setting. Abstract : Parliament is usually described as «impotent» during the 5th French Republic. The executive dominance in legislative agenda setting is one the main explanation. So, the paper will investigate which instruments of agenda control exist in France and accurately describe their different forms. Some data about the frequencies of their use are presented and preliminary investigated. We will devote particular attention to the time constraints and to the 1995 constitutional amendments, that will be discussed in details. We will conclude on the meaning of «impotence of French parliament» by analyzing the amendment activity in French Parliament. Sylvain Brouard CEVIPOF, Sciences-Po, 98, rue de l université Paris sylvain.brouard@sciences-po.fr

2 6/04/05 2 Parliament is usually described as «impotent» during the Vth French Republic. Only few scholars qualify this opinion or accurately investigate the mechanisms of government control on agenda and their efficiency (Huber 1996, pp ). Since 1958, France is nearly always seen as the model of executive dominance : «The French government is an exception in terms of breadth, depth and variety of institutional weapons at its disposal» (Tsebelis 2003, p. 100). Numerous mechanisms exist by which government determines the agenda of Parliament (see Table 1). In a first part, we present in details the time constraints in French law-making and study the 1995 constitutional reform in which several amendments affect the time constraints in the executivelegislative relationships. Data about used frequency of the various time constraints before and after 1995 are presented to capture an eventual change since Second, we present an exhaustive list of how the legislative agenda is constrained by government. This includes the most studied package vote and confidence vote procedures. But, like Tsebelis, we «focus on weapons of lower range and higher frequency». Then, more generally, we bring evidence about the variable uses of these weapons. Our paper is only a preliminary empirical investigation of the role of French government in legislative agenda setting. A lot of work has to be done to understand better and deeply the empirical variations on this topic. TIME CONSTRAINTS AND THE 1995 CONSTITUTIONAL AMENDMENTS First, the time constraints in French law-making are detailed. Second, we focus on the 1995 constitutional reforms and this effect. Time Constraints In France, the mastering of legislative time belongs to Executive in different ways. First, government dictates how most of legislative time is used. Government determines the legislative agenda. Article 48.1 of the constitution states that government proposals have priority over parliamentary ones and that government fixes the order of discussion. In practice, the Presidents conference set the days and hours of discussion. But government effectively controls the organisation of the legislative agenda in National Assembly because it is allowed to add de jure some days if necessary (article 28.3 & article 50 (NAR)). Moreover, The constitutional court protect the agenda setting power of government : government is allowed to change whenever and freely the agenda set by the Presidents conference and introduce on the floor a new bill (Rousseau 1992).

3 6/04/05 3 Furthermore, according to article 45, only government is allowed to gather the conference committee to solve the disagreement between National Assembly and Senate after two readings in each chamber 1. In France, the principle is the navette between the two chambers until they vote the same bill. Potentially, the navette could last infinitely. Without a declaration of «urgency», government is allowed to call a conference committee after a two-round navette but, in no way, it is constrained to do that (Bernard 2001). So government is able to prevent the adoption of a parliamentary proposal by not calling a conference committee. Since 1969, 839 conference committees were gathered. Another rule favours government in the legislative agenda : Article of National Assembly Rule favours the discussion of amendments from government (but also from committees). Second, legislative time is constrained. Article 28 limits parliamentary meetings. Until 1995, there were two parliamentary meetings by year, the first of 80 days, the second of less than 90 days. Since 04/08/1995, there is only one parliamentary meeting by year limited to 120 days. The same time limits exist for specific laws: Article 47 limits to 70 days the vote of the budget (40 AN, 15 Sénat, 15 resolution) and article 47.1, since 1996, allows 50 days to vote the social security budget (20 AN, 15 Sénat, 15 resolution). The declaration of «urgency», according to Article 45, is also a way for government to fasten the legislative process : there is only one discussion of an «urgent» bill on the floor of each assembly before Prime Minister calls a conference committee. The bill is then discussed on the floor of each assembly and eventually a last time on the floor of National Assembly. «Urgency» procedure is not well-known by scholars whereas it is widely employed by government. For example, in SPPL Data Set (Döring & Hallerberg 2004), government imposed restrictive procedures in 35 of the 42 French bills. For these 35 bills, «urgency» was always invoked. In 4 cases, there was also a confidence vote procedure (and perhaps some others procedures ). In Rationalizing Parliament (1996), Huber also presents similar results : in his data set of 356 bills voted in National Assembly between 1978 and 1989, 46 % are «urgent» bills. More generally, during the Vth Republic, the use of this weapon is frequent : between 1969 and , 672 bills were declared «urgent» (see graph 1). Each year, on average, almost 20 bills (19,8) are urgent. For comparison, during the same period, both package vote and confidence vote procedures concerned less than 280 bills. Clearly, the frequency of its use identifies «urgency» as a key mechanism in the practice of legislative agenda control by French governments. This is particularly true during divided legislature. In these cases, government anticipates an upper level of disagreement with one chamber and use 1 About this topic, see Tsebelis & Money 1997.

4 6/04/05 4 more frequently urgency to avoid wasting to much time. Data supports this analysis. In fact, the number of conference committees increases significantly from unified legislature (19.15 by year) to divided legislature (30.4 by year). The number of «urgent» bills is significantly correlated (.862) with the number of conference committees. And finally, the number of urgent bills increases from unified parliament (17.3 by year) to divided parliament (22 by year in National Assembly and 25.7 in Senate). By this way, leftwing governments were able to limit the effects of everlasting rightwing majority in the Senate. Furthermore, Article of National Assembly Rule limits the defense of amendments to 5 mn (National Assembly President is allowed to extend or limit this time), except for government that is unconstrained. Last, special parliamentary meetings (see under) called by MPs are limited to 12 days (article 29.2). Third, extension of legislative time is dependent of executive will. On the contrary of precedent rules, government is allowed to extent legislative time. Article 29 allows Prime Minister and majority of MPs to call for a special parliamentary meeting of some days. Prime Minister is the only one allowed to call for special parliamentary meeting before the end of the month after the termination of a special parliamentary meeting (article 29.3). And, as the opposite of parliamentary initiative, special parliamentary meetings called by Prime Minister are not limited to 12 days. Only President decides to open this special parliamentary meeting and determines the legislative agenda. During divided government, President often refuse or threat to refuse to open a special parliamentary meeting and could avoid legislative reading of bills that he dislikes. Every special parliamentary meeting was called by government since the beginning of the Vth Republic. Last, Article 101 of National Assembly Rule permits to call for a second discussion of a bill or a part of a bill just before the final vote. Amendments rejected in this 2 nd reading confirm the 1st reading. During 30 years, government calls 222 second discussions (7,4 times by year) in National Assembly. Majority in Parliament is also allowed to call a second discussion of a bill or a part of a bill and effectively has done it 74 times in the same period. During the legislative reading of the education bill in february 2005, this rule were used. An amendment Tian, so-called after the name of the deputy who has proposed it, is approved by a majority that indirectly would have stopped the affirmative action policy of Sciences Po in its recruitment. Some days later, after newspapers articles and intensive lobbying, government called a second discussion of the amendment before the final vote of the bill in the National Assembly. The amendment was rejected in the second vote. Another example occurred at the end of the legislative reading of the budget Some MPs from UMP with the implicit support of the Finance Minister Sarkozy tried to diminish the financial solidarity between towns discretely in the last hours of the debate. They proposed an amendment that would have changed the

5 6/04/05 5 reference from average national wealth to average regional wealth. The amendment was defeated. The Budget Secretary called immediately but unsuccessfully a second reading and vote of the amendment to try to change the decision. If time scarcity (Döring 1995, 2004) is a major problem in law making, government is a privileged actor in France because government is allowed to determine the content of legislative time, to impose constraints on it and to extend it as it likes. The 1995 constitutional reform Once elected, in 1995, President Chirac initiated two revisions of the constitution. One of them dealt with referendum and parliamentary organization and has affected few articles of the French constitution. President Chirac announced, in a message to MP s (19th may 1995), his will to give to the Parliament «its right place, a central one, allowing the restoration the relationships between citizens and leaders». Nevertheless, in the same time, he has successfully extended the scope of the referendum, which raises doubts about his will to reassert the value of Parliament. Adopted the 31st of July 1995, the constitutional reform has modified the time constraint stipulated in the article 28 : one nine month ordinary parliamentary session replaces two three months ordinary sessions. But, in the same time, the maximum number of working days diminished from potentially 170 days until 120 days. The initial proposition of the President was 130 working days and National Assembly surprisingly has proposed a lesser maximum number (120) One of the explicit motives for these organisational changes is to allow a better control of government by a longer parliamentary session and by a compulsory session of «questions au gouvernement» by week (art 48.2). Article 48.3 introduces also one parliamentary reserved session («niches parlementaires») per month where priority is given to parliamentary proposals. In the end of 1995, a second constitutional reform took place. To cope with the deficit of welfare system, Parliament votes every year the budget. In 1995, a time constraint was introduced in the constitution : the social security budget has to be adopted in a 50 day delay. Ten years after, what are the impact of these constitutional changes? Has this new time constraint affected the parliamentary activity? Has the 1995 constitutional reform stimulated numerous parliamentary bills? Given the scope of the time constraints, described before, at disposal to the government, we hypothesize that the new time constraints should be irrelevant. Everything else equal, it should be no difference. Since 1969, Parliament always votes the budget in the 70 day delay and, since 1996, the social security budget in the 50 day delay.

6 6/04/05 6 Since 1995, parliamentary meeting time decreases to 106 working days on average, under the 120 day constitutional rule (see Graph 2). But from 1969 to 1995, the average number of working days is So the difference is very slight and unsignificant. If we limit the comparison from 1981, the always unsignificant difference is reduced to 0.33 day. An alternative way to test the impact of the 1995 constitutional reform on legislative time is to scrutinize the effective time of legislative reading. This is probably more accurately captured in an indicator of the number of working hours of legislative and budgetary readings. By this mean, we are more able to compare legislative activity before and after 1995 change. In fact, the same number of days could mean some differences in number of working hours. Nevertheless, the two variables are significantly correlated but not perfectly.868 since 1969 and.888 from How has the number of working hours evolved across time? After 1995, the average number of working hours climbs from to and the difference is statistically significant. When we only compare between 1981 and 2002, the difference is slighter and unsignificant. But, the average number of working hours after 1995 counts hours more than before So, if evaluating in hours or days does not give the same results, but the time dedicated to budgetary and legislative readings does not appear significantly changed. So the change of the rule in 1995 does not change the constraints on legislative time. To go further, we specify regressions that control for some variables that could affect the legislative time to better estimate the effect of the 1995 reforms that is one independent variable. Two dependent variables are used, the number of working hours and days by year and two periods are also used : and It is self evident that the number of special parliamentary sessions affects the legislative time. Since 1969, 51 special parliamentary meetings took place (1,54 by year on average) but only 2 since During electoral year, the parliamentary activity is stopped during the campaign. So, we code 1 the year when there is a presidential election and / or legislative elections. Just after an electoral year, majority wants to implement their policy proposals. So, the density of proposed bills is high and special parliamentary meetings are called. To control the post-electoral effect, we use the variable post-electoral year*nb of special meetings. The number of adopted bills should also influence legislative time. More bills should imply more time to cope with them. Last, the impact of intensity of legislative work on legislative time could be captured by the number of proposed amendments. More conflict in Parliament or more desired changes from MPs should spend more legislative time. 2 Sometimes, scholars present mistakenly 149 days as the number of working days before 1995 (Huber, 1996, p. 32). 3 Note that and are missing data in these analysis.

7 6/04/05 7 The results of this simple regression analysis (see Table 2) show that the 1995 reform has a negative and unsignificant effect on the number of working days in National Assembly in both periods. Using the number of working hours as dependent variables, the sign associated to 1995 reform variable changes and is, in both cases, unsignificant. In the four regression analysis, number of amendments, post-electoral and electoral years have the correct expecting sign and the coefficients are often statistically significant. They seem to have a significant impact on the legislative time contrary to the 1995 reform. So, 1995 constitutional reform allows less working days for parliamentary readings. Nevertheless, because of the scope of mastering of time by government, there is de facto no change in time dedicated to budgetary and legislative readings in National Assembly. A posteriori, French scholars had good reasons to avoid studying the impact of the new time constraints introduced in The new parliamentary reserved session should improve the number of bills initiated by parliament. But their number should also be marginal compared to the number of bills initiated by the government and could in no way change substantially the current equilibrium. The oppportunity to parliamentary initiatives seems better since 1995 (see Graph 3). On average, there are adopted bills from Parliament by year. Between 1969 and 1995, the average number of non-government bills reaches 13.7 by year (12, 8 between 1981 and 1995) 4 whereas since 1997, it reaches more than by year. This increase is undoubtedly associated with the new parliamentary reserved session each month, used jointly with the traditional «complementary agenda». The «complementary agenda» permits Parliament to discuss a bill when the «ordinary agenda» is finished, if there is time left. What is striking is that, at the same time, the number of adopted bills significantly decreased since On average, there are 83.8 bills adopted by year. But, since 1997, the mean number of adopted bills in France dropped to 71 versus 85 between 1981 and 1995 and 86,6 between 1981 and Consequently, the average proportion of parliamentary bills in the total number of adopted bills has increased significantly from 15.83% ( ) or ( ) to 24.61% ( ). So, contrary to our prior expectations, the government control on parliamentary agenda seems empirically looser or less monopolistic since Nevertheless, Vth French Republic gives numerous and efficient weapons to governmental control. In my mind, time constraints are probably the key point to explain the role of government in legislative agenda setting, and consequently, executive dominance. French government has a large 4 According to Mattson (1995), between 1978 and 1982, the average number of non-government bills reaches 7 by year. 5 Nevertheless, the relative importance of these bills could be discussed as the real origin of the bills.

8 6/04/05 8 repertoire of means to control the legislative time : 1. government determines how most of legislative time is used ; 2. legislative time is constrained and government is able to sharpen it ; 3. Executive is allowed to extent legislative time according to its needs. LEGISLATIVE AGENDA SETTING POWERS OF FRENCH GOVERNMENT After time constraints, we complete the 7 left forms of agenda control established by Rasch and Tsebelis to cope with the role of French governments in legislative agenda setting. Closed or Restrictive Rule If amendment is a constitutional right in France (article 44), its exercise is closely constrained (Baufumé 1993). Two main restrictions to the scope of amendments exist. Constitution of the French Vth Republic in Article 40 introduces a financial inadmissibility of amendments that increase spending or reduce revenues of the State. budget. Financial inadmissibility is clearly the most used restrictive rules to exclude amendments : since 1969, nearly amendments were excluded because of article 40. The average number of amendments excluded on this ground by year is 368. Between 1981 and 2003, the average has climbed to 464 rejected amendments by year which represents 5,12 % of the number of non-governmental amendments introduced during this period. And, with no doubt, this rule prevents the introduction of numerous other amendments. Article 41.3 establishes a legislative inadmissibility of amendments, that excludes amendments that violates the distinction between the «domain of law» (article 34) devoted to Parliament and the «domain of regulation» (article 37) reserved for government. The current President of the National Assembly advocates new constitutional amendments to reinforce legislative inadmissibility and the strict separation between the «domain of law» (article 34) and the «domain of regulation» (article 37). Another restrictive rule, less famous in France, is the Germaness rule. Two articles of National Assembly Rule (Art , Art. 127) and, at least, one article of social security Organic Law (Art ) require that amendments have to be inside the frame of the bill. Like legislative inadmissibility, Germaness rules are quite never applied to avoid amendements : since 1969, both rules only eliminate 40 amendments. But, if germaness rules are seldom a visible constraint for parliamentary amendments, they limit government behaviour because constitutional court vetoes amendments from government that do not fit the germaness rule. In the «Décision n DC du 5 août 2004 Service public de l'électricité et du gaz et entreprises électriques et gazières», the change of the age limit for manager of public firms has been invalidated because the amendment from government in Senate on this topic did not respect germaness requirements. The package vote (Article 44.3, «vote bloquée») and the confidence vote procedure (article 49.3) are well-known mechanisms of rationalized

9 6/04/05 9 parliamentarism (Huber 1992, 1996). The package vote permits government to select some articles and amendments and to call a take-it or leave-it vote from Parliament on this package. Article 49.3 is a tool in government hands to avoid a vote on a bill which is adopted if government is not defeated by a motion of censure. Thus, package vote and confidence vote are rare, on average respectively 6 times and 2 times by year since Since 1981, their use is more frequent 2.87 after 1981 vs before, for confidence vote, and 7.38 vs for package vote. One important empirical question is how many amendments are reserved by these ways? On his data base 6, J. Huber evaluates the average number of reserved amendments by package vote to 44.2 by bill, by confidence vote to 1.5 and by both to 58.3 by bill. So, when employed, package vote procedure succeeds in eliminating numerous amendments whereas confidence vote procedure seems less efficient for this task. We do not go deeper on these two rules. These procedures are now well known. Their logics, in France, are explained and empirically validated : package vote protects agreement between majority parties and confidence vote procedure works along an electoral logic (Huber 1996). Last, according to article 45, after an agreement during the conference committee, the proposed bill is voted on the floor of each assembly under closed rule, unless government accepts amendements (see below). At the first glance, restrictive rules allowed French governments to sharply restrain amendments activity in Parliament. Expansive Rules Financial inadmissibility (Article 40) prohibits financially unbalanced amendments from Parliament. But this rule does not apply to government. So, de facto, government is allowed to make amendments that are prohibited to others in a key domain : decrease tax or increase spendings. Sequencing Rules French government has the possibility by different means to present the last offer. The package vote and the confidence vote procedure are presented before. But two other mechanisms permit government to make amendments at times when other amendments are not permitted. First, Article 44.2 establishes an inadmissibility of amendment for no committee reading. Government can exclude amendments by this way. In fact, amendments should be drafted 4 days after the official report of the committee about the law in the first reading, 2 days after a new setting in the legislative agenda, and always before the beginning of the discussion in case of change in the agenda (Article 99 of National Assembly Rule). This delay is not applied for amendments from government or from committee. Nevertheless, when government introduces new amendments or new articles in a bill, amendments are allowed but committee, its 6 Package vote and confidence vote procedures concern 37 of the 356 bills of his data base.

10 6/04/05 10 President and its reporter act as gatekeepers and could accept or refuse amendments. Empirically, its efficiency seems limited in time and in number. Inadmissibility for no committee reading was used, since 1969, only in 1986, 1990, 1993 and and prohibits 2657 amendments. Second, in case of failure of the navette between National Assembly and Senate, an conference committee is called by Prime Minister to reach a deal. The proposed bill is then voted on the floor of each assembly under closed rule, unless government accepts amendments (article 45). In this case, amendments are clearly reserved to government. So, in the last reading of a controversial law, government benefits from an expansive rule that allowed it to make the last offer. This happens, on average, for 24 bills each year. Between 1985 and 1991, 42 bills were amended after an agreement in the conference committee and 11 between and When there is a desagreement in the conference committee or when both chambers do not approve the same text, National Assembly is allowed to decide. At this stage, amendments adopted in Senate may be introduced : between and , 8 bills are concerned. Nevertheless, Constitutional Council recently limits governmental amendment right to the part of the bill discussed during the conference committee. So, now, government is not allowed to amend the part of the bill about which both chambers agree before the conference committee (Camby 2000 ; Binczak 2001). So, government can intervene whenever during legislative readings whereas MPs cannot. Nevertheless, Committees seem to be important player in law making. This is confirmed by a close look at the origins of amendments (see below). Voting Order Rules Article 42 states that governmental bill is the basis of discussion on the floor, not the committee proposition. On the contrary, amendments are voted first according to a successive procedure 7. Moreover, amendments that favour suppression of a part of a bill are voted first. It is not self evident that the governmental bill gains a privileged position in the order of voting in France. Vote-Counting In France, only majority of valid votes (simple majority 8 ) is usually necessary to adopt a bill or an amendment in Parliament whoever is the author. So, votes in favour or against government proposals are not counted differently than other proposals. Nevertheless, there is an 7 About voting order and vote counting rules, see Rash Bjørn Eric In successive procedure, alternatives are voted one-by-one in a specified order whereas, in elimination procedure, there are pairwise comparison of alternatives. 8 Simple majority : «an alternative needs to receive more than1/2m of the votes to beat its challengers, where m is the number of aye and nay votes ( )» (Rash Bjørn Eric 1995).

11 6/04/05 11 important exception. According to Article 49.3, during confidence vote procedure, when opposition wants to reverse government by a motion of censure, only the vote against government are counted and majority of elected MP s (absolute majority 9 ) is necessary to win. This rule protects deliberately French government and de facto favours its position in lawmaking. Gatekeeping Rules It is possible for the government to hold an issue away from consideration by the parliament by two tools detailed before and by referendum : - Article 49.3 establishes the confidence vote procedure that induces no vote would occur on the law when confidence vote procedure is called. - Article 41 enables government on the grounds of legislative inadmissibility to avoid that a matter is discussed in Parliament given that «domain of regulation» is reserved for the government. - Article 11, since 1995, extend the scope of the referendum and President is allowed to organize a referendum (after a proposition of Parliament or government) on every bill relative to organization of the State or social or economic reforms. For 10 years, the new content of the article 11 was never used. Exclusive Government Jurisdiction Executive decrees are offered to the government by the constitution : - Article 38 allows executive decrees («Ordonnances») for a fixed time provided an explicit delegation from Parliament. These executive decrees have to be adopted afterwards by Parliament in a specific bill. - Article 34 and Article 37 define the limits of law and stipulates that the unwritten matters and domains belong to regulation that is the monopoly of government activity. Since the birth of the 5th Republic, French executive dominance is traditionally based on the supposed extensive scope of the exclusive government jurisdiction. But some scholars advocate that, in this domain, the revolution didn t occur (Rivero 1978; Rousseau 1992) and numerous bills contain a lot of regulation provision. Conclusion Undoubtedly, French government has numerous means at its disposal to control the legislative agenda. Nevertheless, the frequencies of uses for each of them vary widely. We have focused on time constraints given their scope and their formal changes. One of the main effects of 1995 constitutional amendments is to significantly increase the number of 9 Absolute majority : «to be accepted an option needs positive support from more than 1/2n of the assembly (n refers to total membership)» (Rash Bjørn Eric 1995).

12 6/04/05 12 adopted bills from Parliament. The equilibrium between government and parliament in law-making was, surprisingly but significantly, changed too. Like dissuasion by nuclear weapon, the effects of restrictive rules is rather difficult to estimate empirically because of anticipation behaviours. The fact that a rule is not explicitly employed to erase some amendments does not mean that this rule is inefficient (for example germaness rule). This rule could efficiently prevent some behaviours because the cost of the behaviour writing amendments - become too expensive compared to its benefits. Nevertheless, some forbidden amendments have signalling functions that preserve the interest to write it : by their content for specific voters or by their number in parliamentary guerrilla to show the will of the opposition to abort a bill. To conclude, it is self evident that the limits of parliamentary amendment rights in French Parliament are numerous. But are they so efficient that we could conclude to legislative impotence? One the most efficient restrictive rules is financial inadmissibility that establishes an expansive rule on financial issues. Nevertheless, the numerous government amendments on budgetary bills reflect compromises with legislative majority on financial issue and thus take into account parliamentary preferences. Undoubtedly, on this ground, the last offer power belongs to government. Consequently, government is able to extract its best obtainable policy result. It is clearly not a detail. But there is a paradox. The intensive use, by government, of the financial expansive rule, during budgetary readings, is probably a paradoxal sign of Parliamentarian influence. Moreover, the analysis of amendment activity shows an unexpected legislative activity from Parliament (See Table 3). First, only a minority of bills is adopted without any amendment : 680 of 2227 adopted bills (30,5 %) since 1969 had not been amended, whereas bills with between 1 and 50 adopted amendments (more than 50 adopted amendments) represent 53,1% (16,4 %) of adopted bills. Second, as expected, government amendments are in minority in the total number of introduced amendments. Since 1969, 6,9 % of the amendments were introduced by government. More surprisingly, amendments from government were also not the more numerous in the total number of adopted amendments. Only 19.5 % of succeeded amendments came from governments. And last but not least, the rate success of amendments is almost equal between government and committees. Thus, 79.8 % of amendments from committees were adopted, whereas the success rate for governmental amendments is 83.5 % 10. But, this rate did not apply on the same amount. Committees introduced, since 1969, amendments which represents more than 3 times the number of amendments introduced by government (16 928). So, in no way, government is dominant in amending bills: 61,5 % of the total number of 10 Since 1981, the success rates of amendments from government and committee are closer, respectively 84,7 % and 83,3 %.

13 6/04/05 13 adopted amendments were introduced by committees. In French Parliament, committees play a significant role during legislative readings. The amendments rules give only small opportunities to MPs to change bills but, by the way of committees, a parliamentary majority preferred position on a bill has a high probability to influence the feature of the law. In our mind, government is seldom a dictator but more frequently an agenda setter that just can propose its best obtainable policy position to a parliamentary majority whose preferences often diverge from governmental ones. Bibliography : Baufumé Bruno Le droit d amendement et la Constitution de la Vème République. Paris : LGDJ. Bernard Sébastien «La commission mixte paritaire.» Revue Française de Droit Constitutionnel 47 : Binczak Pascal «Le Conseil Constitutionnel et le droit d amendement : entre «errements» et malentendus.» Revue Française de Droit Constitutionnel 47 : Camby Jean-Pierre «Droit d amendement et commission mixte paritaire.» Revue de Droit Pulic (6) : Döring Herbert «Time as scarce resource : government control of the agenda.» In Döring Herbert (ed.), Parliaments and Majority Rule in Western Europe. Frankfurt/New York : Campus/St Martin s (pp ). Döring Herbert «Controversy, time constraint, and Restrictive Rules.» In Döring Herbert & Hallerberg Mark, Patterns of Parliamentary Behavior. Aldershot : Ashgate (pp ). Döring Herbert & Hallerberg Mark Patterns of Parliamentary Behavior. Aldershot : Ashgate. Huber, John D «Restrictive Legislative Procedures in France and the United States.» American Political Science Review 86 (3): Huber John D Rationalizing Parliament: Legislative Institutions and Party Politics in France. Cambridge/New York: Cambridge University Press. Mattson Ingvar «Private Members Initiatives and Amendments.» In Döring Herbert (ed.), Parliaments and Majority Rule in Western Europe. Frankfurt/New York : Campus/St Martin s (pp ). Rash Bjørn Eric «Parliamentary Voting Procedures.» In Döring Herbert (ed.), Parliaments and Majority Rule in Western Europe. Frankfurt/New York : Campus/St Martin s (pp ). Rivero Jean «Le domaine de la loi et du réglement.» Actes du colloque d Aix-en-Provence. Marseille : P.U.A.M. Rousseau Dominique Droit du contentieux constitutionnel. Paris : Montchrestien. Tsebelis George Veto Players: How Political Institutions Work. New York/Princeton, NJ/Woodstock: Princeton University Press.

14 6/04/05 14 Tsebelis George & Money Jeannette Bicameralism. Cambridge: Cambridge University Press Graph 1 : Number of urgent bills by year in National Assembly 11 Nb of working hours Nb of working days Graph 2 : Working time in National Assembly by year After 1995, when we write 1996, for example, it means year

15 6/04/05 15 Nb of adopted bills (NA) Nb of Parliamentary bills (NA) Graph 3 : Number of bills by year (National Assembly) After 1995, when we write 1996, for example, it means year After 1995, when we write 1996, for example, it means year

16 6/04/05 16 Table 1 : AGENDA SETTING POWERS OF FRENCH GOVERNMENT Mechanisms of government agenda setting : Time Constraints : Government proposals may have priority over parliamentary ones; in addition it may be possible for governments to restrict time for discussion to a level that hinders or prohibits the introduction of amendments Closed or Restrictive Rule : No amendments or few amendments to government proposals are accepted (for example, introduction of germaneness requirements, or restriction to amendments that increase spending or reduce revenues) Expansive Rules : Permitting the government to make Rules : - Art.48.1 Priority for government in legislative agenda - Art. 28.3, allows government to add de jure some days (also, article 50 (NAR)). - Art since 04/08/1995, birth of parliamantary reserved session («niches parlementaires») - Art. 45, Prime minister has the power to gather the conference committe - Art (NAR) give priority to the discussion of amendments from government and committe - Art 28. limit parliamentary meeting to 120 days (art , Prime Minister can add more days) - Art. 47 limit of 70 days to vote the budget (40 AN, 15 Sénat, 15 resolution) - Art limit of 50 days to vote the social security budget (20 AN, 15 Sénat, 15 resolution) - Art. 45 Prime Minister can proclaim «the urgency» for a law that limit to one discussion on the floor of each assembly before an inter-chamber committe. - Art (NAR), advocacy of amendments limited to 5 mn (NA President is allowed to extend or limit this time), excepted for government. - Art. 29.2, special parliamentary meetings called by a majority of MPs are limited to 12 days. - Art. 29 favours Prime Minister to call for a special parliamentary meeting (length and delay). - Art. 101 (NAR), government is allowed to ask for a new discussion of a law or a part of it just before the final vote. Amendments rejected in this 2 nd discussion confirm the 1st discussion - Art. 40, financial inadmissibility - Art.41.3, legislative inadmissibility - Germaness rule (Art (NAR), Art. 127 (NAR), Art social security Organic Law) - Art. 44.3, package vote («vote bloquée») - Art. 49.3, confidence vote procedure - Art. 45, after a conference committe, the bill is voted on the floor of each assembly under closed rule, unless government accepts amendements. - Art. 40, financial inadmissibility

17 6/04/05 17 government to make amendments that are prohibited to other parties Sequencing Rules : Permitting the government to make amendments at times when other amendments are not permitted Voting Order Rules : The governmental bill (rather than the proposal from the committee) gains a privileged position in the order of voting Vote-Counting : Votes in favour or against government proposals are counted differently than other proposals. Gatekeeping Rules : It is possible for the government to hold an issue away from consideration by the parliament Exclusive Government Jurisdiction : Executive decrees can be offered to the government by the constitution, a general act of Parliament, or by specific Parliamentary votes - Art. 45, after a conference committe, the bill is voted on the floor of each assembly under closed rule, unless government accepts amendements. - Art. 44, inadmissibility for no committee reading - Art. 42, government s bill is the basis of discussion on the floor, not the committee proposition, but amendements are voted first. - Art. 49.3, during confidence vote procedure, when opposition want to reverse government («Motion de censure»), vote against government are only counted and majority of elected MP s is necessary to win. Only majority of valid votes is usually necessary otherwise. - Art 49.3, confidence vote procedure : no vote occurs on the law. - Art.41, legislative inadmissibility. - Art. 11, referendum. - Art. 38. Executive decrees («Ordonnances») by delegation of Parliament. - Art 34 & Art 37, law vs regulation

18 6/04/05 18 Table 2 : Determinants of legislative time (unstandardized estimates from OLS Regression; significance levels in parentheses) Working days Independent variables : by year constitutional (.234) reform Electoral year (.04) Post-electoral year*nb of (.103) special meetings Number of.0014 amendments (.008) Number of.044 adopted bills (.738) Number of special (.235) parliamentary meetings Dependent variable : Working days by year (.127) (.251) 4,445 (.159) (.129).198 (.385) -1,633 (.588) Working hours by year (.426) (.109) (.042).0124 (.011).170 (.892) (.114) Working hours by year (.544) (.152) (.028).0067 (.159).985 (.598) (.339) Ajusted r

19 6/04/05 19 Table 3: Amendement activity in National Assembly ( ) Year Nb of Gvt amendements Nb of Adopted Gvt amendements Nb of Committe amendements Nb of adopted Committe amendements Nb of Total amendments Nb of Total amendments adopted Total Total