The Paradox of the Veto in Mexico (1917-1997)

by

Eric Magar Departamento de Ciencia Política Instituto Tecnológico Autónomo de México [email protected]

and

Jeffrey A. Weldon Departamento de Ciencia Política Instituto Tecnológico Autónomo de México [email protected]

COMMENTS VERY WELCOME PLEASE CONTACT AUTHORS BEFORE CITING

Prepared for presentation at the 23rd International Congress of the Latin American Studies Association, Marriott Wardman Park Hotel, Washington, D.C., September 6-8, 2001.

2

Abstract “The Paradox of the Veto in Mexico (1917-1997)”, by Eric Magar and Jeffrey A. Weldon. Prepared for presentation at the 23rd International Congress of the Latin American Studies Association, Marriott Wardman Park Hotel, Washington, D.C., September 6-8, 2001. We report work in progress. Existing explanations of veto incidence do not seem to account for the case of classic Mexico. Proceedings of the Chamber of Deputies from 1917 to 1963 report that Mexican presidents returned at least 223 bills to Congress instead of signing them into law. This level of veto incidence is puzzling from numerous theoretical perspectives. For Mexicanists hardly any executive veto should have taken place after 1936, when iron discipline to the mandates of the president was consolidated with the end of the diarquía. For similar reasons, those who explain vetoes on the mistaken evaluations of members of Congress over presidential preferences have little bite on the explanation of Mexico’s vetoes: the president was the leader of the congressional party. Those who explain vetoes as attempts to publicize politicians’ positions in contrast with those of vetoing actors would need to take into account the fact that election politics was never undertaken by campaigning against the incumbent president.

3

The Paradox of the Veto in Mexico (1917-1997)1 by Eric Magar and Jeff Weldon

Empowering certain individuals or groups with a veto over policy lies at the very heart of separation of power (cf. Federalist Papers, 1788). Such veto gates in the decision- making process, it is well established, reduce the chances that government will abuse the rights of the citizenry (or groups of citizens). Presidentialism—the structure of government with separate and independent legislative and executive branches—is one common form of separation of power, a form of government found, with few exceptions, throughout the American continent. The presidential form of separation of power has inspired several strands of literature among comparativists and, especially, Latin Americanists (e.g. Linz and Valenzuela 1994; Mainwaring and Shugart 1997). One strand of literature that has been present for years sustains that Latin American presidentialism is far from operating as intended. Latin American constitutions may contain all sorts of provisions to prevent encroachments of the executive over the legislative branch, but in practice those “paper barriers” (Mecham 1959, p. 262) do not resist the region’s strong- man tradition that undermines the effectiveness of Congressional checks over the president’s broad powers (e.g. O'Donnell 1994). 2 “The fact that the president [in Latin America] rarely has a chance to oppose his veto to a bill from Congress is the best proof [of presidential dominance]” (Lambert and Gandolfi 1987, p. 394). In this paper we begin by showing that assessments of the effectiveness of separation of power by the frequency of veto usage (or lack thereof) are not fruitful as a research strategy. We do so by expositing the principle of anticipated reactions which such strategy fails to account for. This principle has in fact raised another interesting theoretical anomaly: if actors anticipate their reactions, why should there be ever any vetoes? The hundreds, sometimes thousands of executive vetoes that have been recorded in the annals of most presidential systems seem to remain unaccounted for by this common approach. We present two recent theoretical contributions that seek an explanation of veto incidence. One explanation sees vetoes as bargaining ploys, asserting that, even when they try, politicians often overestimate the likeness between their notion of good policy and that of another politician with veto power. The result is a veto to signal a mistaken evaluation, and the sender of the signal may lie in order to attempt and get more concessions from a rival (cf. Cameron 2000). Another explanation sees vetoes as publicity stunts : vetoes, when they take place, show clearly, with observable actions which politician stands for what policy. Politicians often need such type of publicity for the purpose of being elected for office, the same already held or another, and thus engage in veto politics (Groseclose and McCarty 2001; Indridason 2000; Magar 2001). We then apply the theoretical discussion to the case of Mexico from 1917 to 1997. Mexico had the purest breed of presidential authority in the subcontinent. If vetoes should have been absent from any government that was Mexico’s. Representatives from both chambers of Congress were single term- limited; campaigns were run by a national

1

The authors are grateful for the research assistance of Claudia Y. Carmona, Aileen Fernández, María del Carmen Nava Polina y Jorge Yáñez López. 2 See also Dealy (1992), Edelmann (1969), García Hamilton (1990), Meyer and Reyna (1989), Scott (1958), Segovia (1975), Wiarda (1992).

4 party perfectly mirroring the incumbent president’s policy preferences; elections were not exactly democratic; the president was the undisputed head of his party, in Congress and elsewhere. The paper proceeds as follows. Section 1 briefly describes Romer and Rosenthal’s setter model (1978) and their stylization of anticipated reactions. Section 2 reviews two recent modifications of the setter model to explain veto incidence. Section 3 explains the constitutional basis of the veto. Section 4 presents the puzzle of vetoes in Mexico from a theoretical viewpoint: extreme unification of power should have depressed veto incidence to zero throughout the PRI’s reign. Section 5 presents preliminary evidence of veto incidence in Mexico from 1917 to 1997. Section 6 presents conjectures of what could have driven veto incidence. Section 7 concludes. 1 The logic of anticipation

Infrequent use of the executive veto power in the U.S. led many to reach the opposite conclusion of Latin Americanists (for a review, see Pious 1979). U.S. presidents have vetoed, on average, 9 bills every year since World War II; they have signed about 360 on average every year in the same period (Cameron 2000). This is a ratio of 40 signatures per veto. This temp ted many to conclude that U.S. presidents have failed to check the products of the legislature. In this section we review the logic behind the setter model (Romer and Rosenthal 1978) to show the invalidity of assessing the balance of power between the branches of a presidential government by looking at veto incidence. The model stresses the central role that strategic anticipation plays in deflating the occurrence of vetoes. The model shows that a low incidence of executive vetoes is, in fact, synonymous with perfect executive abdication and with perfect executive usurpation of legislative powers. In the parlance of methodologists, when it comes to vetoes abdication and usurpation are observationally equivalent. Romer and Rosenthal assume that a politician with agenda-setting power (hence the name of the model) proposes take-it-or-leave- it policy to another politician. This models a Congressional committee proposing new legislation in its jurisdiction to the floor under a closed rule of amendment; or a president who confronts the bills passed by the assembly with only his or her veto power. If the president dislikes some feature of the bill that he or she is being offered, he or she would have to reject all of it. Turning to the interac tion of such politicians, we can begin by thinking of each as having independently generated notions of where “good” policy should lie in a political spectrum. A president may want, for example, to make existing legislation less ambiguous in defining who is entitled to receive government pensions and who is not, this in order to render free-riding on pensions more difficult. If some free-riders are well represented in Congress this change in policy would in fact look quite “bad” for Congress. Policy deadlo ck ensues because both politicians would rather retain the status quo than let the other “improve” policy. Each uses the veto power to prevent change. 3 Among its logical assumptions the setter model includes the following two strong ones. First, it supposes that politicians only care for policy, and any action of theirs is guided by this maxim. Second, it supposes that politicians know what they want and do not want in policy, but can also map all possible policy outcomes into two exhaustive, mutually-exclusive groups: (1) outcomes they prefer to the status quo; (2) outcomes they disprefer to the status quo.

3

The assembly’s veto, in this case, consists of not initiating a new bill.

5 Among its logical assumptions the setter model includes the following two strong ones. First, it supposes that politicians only care for policy, and any action of theirs is guided by this maxim. Second, it supposes that not only politicians know what they want and do not want in policy (which seems rather innocuous to assume) but can also map all possible policy outcomes into two (exhaustive and mutua lly-exclusive) groups: outcomes the other politician prefers to the status quo; and outcomes the other politician disprefers to the status quo (the strong part of the assumption). With these assumptions the analysis of politicians’ interactions becomes simple enough. The agenda setter politician will evidently discard any policy change that is disliked by both him or her and by the veto politician. By the same token, the agenda setter will discard policy that the veto politician disprefers to the status quo, even if the setter prefers it to present policy. The reason is that such proposals will be discarded by the veto politician and policy will revert to the present state of things. Since we have assumed that politicians only care for policy, these attempts would be useless. The agenda setter will, thus, concentrate on the subset of proposals that the veto player prefers to the status quo. Within this subset, the agenda setter will locate the policy that he or she in turn mostly prefers to the present state of things. If such proposal exists the veto player will accept it (because it is within his subset of preference). If such policy does not exist given the preference profile of the two politicians, the agenda setter will refrain from proposing any cha nge, keeping the gates shut. Whether or not such proposal exists, a veto will not take place. Vetoes are avoided by accommodation or by keeping the gates of policy change shut. The important lesson to be drawn from the logic of strategic anticipation is the following. We need not observe actual vetoes to conclude that a veto player is having influence over policy. Agenda setters internalize the preferences of rivals into the proposal in order to avoid their veto. The setter model rests at the base of many studies of the legislative process (e.g. Baron and Ferejohn 1989; Cameron 2000; Cox and Morgenstern 1998; Kiewiet and McCubbins 1988; Magar 2001; McNollgast 1994). 2 Explaining veto incidence

As mentioned in the introduction, the setter model also generates a theoretical anomaly. A central implication of the model is that, in equilibrium, no veto should ever occur. Why does the model predict no vetoes? The answer is that the model’s assumptions take the logic of anticipated reactions too far. More precisely, the assumptions either make politicians too clairvoyant and too single- minded. By the law of anticipation, and in accordance with the model’s assumptions, if (a) the agenda setter knew with certainty that a given range of proposals will be vetoed by the rival with no possibility of override, then (b) because there is nothing to win from such excursions back to the status quo, the agenda setter would refrain from sending a proposal in that range in the first place. By the same token, if (a) the veto player knew with certainty that his or her veto over a given proposal will be overridden by the assembly, then (b) because there is nothing to win from having the assembly override his or her veto, he or she would refrain from vetoing it. Under pure policy outcome orientation politicians are trying to bring policy closer to their ideal. A veto does not influence this distance, only the credible threat of a veto (Schelling 1960). One possible way that a veto can follow in the model is when the (a) clause in either of the sentences above is not true – i.e. if the model is freed from its clairvoyance assumption by introducing uncertainty. This modification is the essence of Cameron’s (2000) model of veto politics with incomplete information. Vetoes, in Cameron’s world, are a combination of mistaken evaluations of other players’ preferences and a strategic

6 reputation-building device in light of incomplete information that is asymmetric. A president sometimes vetoes a bill he does prefer to the status quo in an attempt to extract further concessions form Congress. Vetoes thus become bargaining ploys. Another possible way to account for vetoes is when the (b) clause in either sentence is not true: situations can be conceived where a veto acquires value per se for one or more politicians. Vetoes are a relatively noticeable political event, and as such offer an excellent opportunity to advertise a politician’s or a party’s commitment to a certain cause, and perhaps to embarrass opponents with the public at large (Kernell 1991, p. 102). Such behavior is analogous to a temporary abandonment of the bargaining table in order to rally up support, mobilize followers, and in this fashion strengthen one’s side relative to the adversary’s in the next round of bargaining, typically after an election. This is position-taking, one of three activities that reelection- minded politicians engage in Mayhew’s (1974, p. 61) classic model. Veto occurrence can thus also be explained by freeing the setter model from its single- mindedness assumption – i.e. if position-taking, is included among players’ goals, along with policy (cf. Groseclose and McCarty 2001; Indridason 2000; Magar 2001). By this account players also seek to exploit opportunities to undertake publicity stunts. The agenda setter may send a bill knowing, beforehand, that it will be vetoed – in fact, despite and because of a veto threat – in order to advocate a position dema nded by core constituents. By acting so, he or she can show that the reason why the desired policy is not enacted into law is found in the veto player’s unreasonableness and recalcitrance. The similar logic applies to the veto player who confronts the possibility of an override to his or her veto (see Magar 2001, chapter 2). We now turn to the theoretical inappropriateness of this pair of explanations – vetoes as bargaining ploys and vetoes as publicity stunts – to account for veto incide nce in Mexico after 1936. 3

The veto in Mexico

The executive veto appears in Article 72 of the current (1917) Constitution. After a bill has been approved in identical form by both chambers, the president has 10 days to publish or veto the bill. If he does not veto within 10 days, the bill is “deemed to be approved” by him, though there is nothing in the Constitution that requires him to publish the law, and no other actor has the power to order its publication. If the Congress is not in session, he may return the bill on the first day of the next regular session. Likewise, he may delay publication of the bill to any point between the closure of Congress to the next session. Congress currently has two sessions: September 1 to December 15, and March 15 to Apr il 30. In the years of this study, however, the single regular session was held between September 1 and December 31. If the president has “observations,” he returns the entire bill to the chamber of origin. The Constitution says that he may return the bill if he disagrees totally or partially with the bill; however, if he disagrees with only part of the bill, he cannot exercise a lineitem veto and publish the rest. 4 The partial veto is interpreted as giving the president the right to propose substitute language in the legislation. In practice, if he disagrees with part of the bill, he sends it to the chamber with observations in the form of an amendment.

4

“El proyecto de ley o decreto desechado en todo o en parte por el Ejecutivo, será devuelto, con sus observaciones a la Cámara de su origen…” (Article 72, clause c).

7 The veto, along with the observations, is treated like any other legislation, according to Article 136 of the Reglamento of the Congress. 5 Therefore, the committee that originally reported the bill must prepare a report on the veto and the observations. If the committee agrees with the amendments proposed by the president, they issue a report incorporating those modifications. The chamber then votes on the report. If they approve the changes—by simple majority—the bill is sent to the other chamber for its approval. Presumably, the president still has an opportunity to veto again (though it is unlikely, because the chamber accepted his observations). Since the veto is subject to regular parliamentary procedures, the committee or the floor could modify the observations. If approved by both chambers, again the president may have the right to veto the legislation a second time. In case of a total veto, if the committee in the chamber of origin decides that the president was correct in vetoing, it issues a favorable report. If this is accepted by a simple majority of the chamber, the veto is accepted, and the legislative process ends. If the committee issues a negative report on a total or partial veto, a two-thirds vote is required on the floor to override the veto, and the bill is sent to the other chamber. If the second chamber also overrides the veto with a two-thirds vote, then the bill is sent to the president for publication. 6 However, there is nothing in the Constitution that requires the president to publish the bill, so there is a possibility that the president can exercise a de facto absolute veto. 7 Constitutional scholars argue that the president cannot veto appropriations bills (Burgoa 1994, p. 692; Carpizo 1978, pp. 86-87; Tena Ramírez 1985, pp. 263-267). This is because, according to Article 74 of the Constitution, appropriations are approved only by the Chamber of Deputies. The veto is authorized in Article 72, which deals only with bills seen by both chambers. There is further confusion in that clause j of Article 72 explicitly rejects the veto for certain resolutions for which the approval of only one chamber is required (such as calling special elections, or impeachment resolutions); since appropriations are not listed in clause j, then by absence, it can be argued that the vetoes of appropriations are permitted. Nonetheless, though the authors of the 1917 Constitution intentionally denied to the Senate the power to approve appropriations, it is not at all clear that they meant to deny a veto over the same to the president. Regardless of the arguments of the constitutional scholars, there have been 45 vetoes of appropriations legislation, although none has taken place again since 1933. When these occurred, the Chamber of Deputies did not argue that the vetoes were unconstitutional, but rather ruled on the merits of the veto and sent their resolutio n to the Executive, (and in addition not to the Senate, as Article 72 would call for). 4 The politics of Mexican presidencialismo

In Mexican political science, presidentialism, as a constitutional system, is contrasted with presidencialismo, which is characterized by the predominance of the federal executive. Presidents have been able to dominate the other federal branches,

5

The Reglamento para el Gobierno Interior del Congreso General de los Estados Unidos Mexicanos is the permanent rules of both chambers. 6 Only the two-thirds rule for override is specified in the Constitution. The other procedures, such as accepting the total veto, or modifying the bill with the president’s substitute amendment, are derived from the Reglamento and have been incorporated in practice over the years. 7 Carpizo (1978, pp. 90-91) describes a case of a second veto on a legislative pension bill by Lázaro Cárdenas in 1935 that was protested by Congress. See also the discussion in Carpizo (1978, pp. 93-94) and Tena Ramírez (1985, pp. 461 -63).

8 legislating with minimal interference from the legislative and judicial branches. They have been able to remove governors through the constitutional procedures involving the Senate, or by merely requesting their resignations. They have been able to appoint their successors. These have been called the metaconstitutional powers of the president, the powers that he has beyond the Constitution (Carpizo 1978: 190-99). Many have looked to the roots of strong presidencialismo in Mexico in the authoritarian political culture of the country. 8 They have a hard time explaining recent change in Mexico. Others emphasize the supposedly strongly presidentialist Constitution of 1917.9 In fact, later comparative research proved that constitutionally, the Mexican president is one of the weakest in Latin America (Shugart and Haggard 2001, 80-81, 9899). Rather, very strong presidentialism is an equilibrium that requires a presidentialist constitutional system plus three additional conditions that are based on the relationship between the president and his party. The three “metaconstitutional” conditions are 1) unified government, 2) high party discipline in Congress, and 3) the president must be head of his party. 10 Not all of these conditions were in full force during the 20th Century. There was divided government before the official party—the PNR (the precursor of the PRI)—was founded in 1929. Carranza (1917-20) and Obregón (1920-24), for example, faced opposition majorities in at least the lower chamber (Weldon 1997c). During most of the 1920s, there was a highly fragmented party system in Congress (Dulles 1961, Garrido 1982). Consequently, presidents lacked most of their metaconstitutional powers, and were legislatively weak. For example, in 1920-22, less than 20 percent of Obregón’s bills were approved by the Chamber of Deputies (Weldon 1997c). Through 1926, a majority of all legislation passed by the lower chamber had been originally introduced by deputies, not the Executive. Divided government ended with the formation of the official party in 1929, and the PRI or its precursors maintained a majority in both chambers until 1997. In the 57th Legislature (1997-2000), the PRI lost its majority in the Chamber of Deputies, but maintained a majority in the Senate. In the 58th Legislature (2000-2003), no party has a majority in either chamber. The second metaconstitutional condition is high discipline in the majority party. Although there is no evidence that discipline had ever been particularly low, several institutional reforms during the 1930s guaranteed that it would remain high for the rest of the century. The main source of division within the party was not interbranch, but rather more federal in nature. The political system was originally based on a coalition of local political bosses, and the federal Congressmen were agents of their local machines rather than to the national leadership of the party (Weldon 2001). There were four main institutional reforms that increased discipline during the period. First, immediate reelection of deputies and senators was prohibited in 1933 (see Nacif 1995). This removed the reelection incentive to members of Congress to represent local interests, and they began to take cues instead from the national party. Second, the local parties affiliated with the PNR were abolished, beginning in 1933 (though it would take the rest of the decade to complete the process). Third, nomination procedures for members of Congress were closed. The PNR’s open primaries were replaced by closed conventions controlled by party leadership (Garrido 1982, pp. 220-21; Goodspeed 1947, pp. 295-96). Closed

8

See, for example, Hansen (1971), Loaeza (1989), Segovia (1975), and Villa Aguilera (1987). See especially Carpizo (1978) and Cosío Villegas (1973). 10 For detail on the metaconstitutional conditions, see Weldon (1997b and n.d.). The authors thank Molinar for the formula, which was adapted from Cosío Villegas (1973, p. 29). 9

9 primaries were reintroduced by the PRI in 1946, but these were prohibited by modifications to the federal electoral law in 1951 (Medina 1978, pp. 20-25). The third metaconstitutional condition is that the president must also be head of the party. This was most emphatically not the case during the Maximato (1928-1936), when former president Calles was known as Jefe Máximo, the chief boss of the PNR. 11 Calles was the person who benefited from the disciplined deputies, not the presidents. He was particularly powerful between about 1928 and 1934, when three substitute presidents completed the term of Obregón, who had been assassinated in July 1928, just weeks after his election. A few years into his sexenio, Cárdenas expelled Calles from the country. He then reorganized the party along lines that privileged sectors over territorial organizations, with his allies at the top of each of the sectors. Seats in Congress were apportioned among the sectors of the new PRM: peasant, labor, and popular (the latter including mostly bureaucrats, but also the military and some survivors of the Revolution). The sectors then assigned candidates. Once Cárdenas was head of the party, the other metaconstitutional powers, such as influence over legislation, increased tremendously (Weldon 1997a). For the most part, the president was the Jefe Máximo of his party at least from the day of the election, and often from the day that he was named as candidate of the official party—by the incumbent president. Therefore, between 1917 and 1929, there was divided government and relatively weak presidents. Between 1928 and 1936, the president was not head of the party. After about 1937, we have strong metaconstitutional presidentialismo most of the time. Table 0 Summary of Weldon’s conditions between 1917 and 1997 Government Parties unified? disciplined? Period – “era name” 1917-1929 – “divided government” 1929-1936 – “Maximato ” 1937-1997 – “strong presidencialismo ” 1997-2000 – “” 2000-2003 – “divided government”

CONDITION 1

CONDITION 2

No Yes Yes

? Yes Yes

President leader of his party? CONDITION 3 No No Yes

No No

Yes Yes

Yes No

The changing partisan conditions that followed the Revolution are summarized in Table 0. It We should stress that executive- legislative relations under the first two postrevolutionary eras – divided government and the Maximato – bear important similarities with those taking place under a democracy. The logic of policy bargaining (à la Cameron) and that of position-taking (à la Groseclose and McCarty or Magar) remain plausible in Mexican politics; the only difference is that the ‘third actor’ is different. To see this consider that the constituents of congresspersons in the 1920s and 1930s were not ordinary citizens, as in a representative democracy. Instead, groups of legislators represented local bosses and their private armies in Congress. 12 The president – a boss himself – was supported by or coalition of local bosses and their guns in arrangements that were far from stable: the end of single-termed presidencies brought

11

See Córdova 1995, Dulles 1961, Medín 1982, Meyer 1978, and Meyer, Segovia & Lajous 1978 This is what Roeder (199*) calls the “selectorate” in the game of policy delegation of the former Soviet Union. 12

10 new episodes of civil war with each. Yet an equilibrium was reached after 1929 in which all incumbents would belong in a single party – an oversized coalition – inside which power and influence would be distributed. By cooptation and by splits (all were violently dismantled), the party gradually gained the allegiance of all local bosses. Ballots, even if tainted with irregularities of all sorts, gradually replaced bullets (see Molinar 1991). In the divided government era each branch was controlled by a different coalition of local bosses; branches would bargain the products of policy, much in the way elected branches representing voters do nowadays. It is conceivable that the branches would engage in bargaining over the spoils of government, struggling to get the better share of (mutually beneficial) deals (cf. Schelling 1960). A president could plausibly have attempted to exploit a congressional majority’s incomplete information about his true preference to get extra concessions, just as Cameronian politicians would in the U.S. today. The same happens under the Maximato , when Congress was under the firm control of the PNR while the National Executive Committee of the Party was firmly controlled by Calles until Cárdenas had him deported in 1936. In both eras the use of the veto for position-taking is plausible for the same reason. Congress could pass popular legislation knowingly disliked by the president in order to provoke his veto. The idea behind such maneuvers was to weaken the president’s coalition by showing where he stood with respect to policy and its spoils. For similar reasons it is conceivable that presidents would veto bills knowing beforehand that Congress would override them, in order to rally up his supporters. This seems particularly likely in 1924-28, when one “revolutionary” faction (the obregonistas, for example) seek support from peasant and rural organizations, while the other main faction (the callistas) depended on urban labor support. During the Maximato, despite unified government, 13 there were widely publicized policy differences between President Ortiz Rubio (1930-32) and Jefe Máximo Calles. The president could attempt to woo support among bosses against Calles’ party by engaging in veto politics of bills that hold the coalition together. However, o nce all three metaconstitutional conditions are in place, we should not expect vetoes. The mode of bargaining described so far makes little sense when the president is the head of his party in Congress. The president proposed, and the Congress disposed in name only. Vetoes became unnecessary as a bargaining ploy. Under strong presidencialismo the president managed to get between 95 and 100 percent of his bills approved by Congress (Weldon 1997a); an idea of the share of policy that this represented is given by the fact that almost 9 out of 10 public bills approved by the lower chamber originated in the executive branch. Politicians under strong presidencialismo also became unwilling to engage in vetoes as publicity stunts to woo supporters. There was little competition in the electoral arena from other parties until the 1980s. Since the foundation of the PNR, elections were decided more by fraud than by policy, so there was no reason for position-taking by either the president, Congress, or the party. The president and party leadership controlled the nomination procedures sufficiently to prevent factionalism in the party, especially

13

All members of each chamber were from the PNR. The PNR began as a party of incumbents, regardless of the ideology or factions that the incumbents belonged to.

11

Table 1 Vetoes by President 1917-1963

President

Years

Carranza 1917-20 De la Huerta 1920 Obregón 1920-24 Calles 1924-28 Portes Gil 1928-30 Ortiz Rubio 1930-32 Rodríguez 1932-34 Cárdenas 1934-40 Ávila Camacho 1940-46 Alemán 1946-52 Ruiz Cortines 1952-58 López Mateos 1958-64* Díaz Ordaz 1964-70 Echeverría 1970-76 López Portillo 1976-82 De la Madrid 1982-88 Salinas 1988-94 Zedillo 1994-2000 Fox 2000-02* 1917-2002 Totals *Data through September 1964.

Type of Legislation Budget Pensions Non Budget 7 0 8 7 6 14 3 0 0 0 0 0 0 0 0 0 0 0 0 45

0 0 0 29 18 29 19 16 1 20 3 4 15 0 0 0 0 0 0 154

8 0 3 7 3 2 4 10 4 0 2 2 1 0 0 0 0 0 2 48

Total

Concur

15 0 11 43 27 45 26 26 5 20 5 6 16 0 0 0 0 0 2 247

3 0 5 39 11 30 3 16 4 8 3 0

122

Resolution of Veto Not Override Disputed 6 6 0 0 3 3 3 1 2 12 4 11 7 11 8 2 1 0 2 6 2 0 0 0

38

52

Withdrawn 0 0 0 0 2 0 5 0 0 4 0 0

11

12 Table 2 Vetoes by Year 1917-1963 Type of Legislation Year 1917 1918 1919 1920 1921 1922 1923 1924 1925 1926 1927 1928 1929 1930 1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955 1956 1957 1958-63 Total

Resolution of Veto

Budget Pensions Non Budget 0 1 6 0 2 6 0 0 0 0 0 7 6 8 1 6 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 45

0 0 0 0 0 0 0 0 0 6 20 3 17 7 21 11 9 1 6 0 5 0 3 2 1 0 0 0 0 2 0 3 10 3 2 0 0 0 0 2 1 0 135

1 6 1 0 1 0 0 2 0 0 4 3 2 2 0 2 0 3 0 0 0 1 0 9 1 2 1 0 0 0 0 0 0 0 0 0 0 2 0 0 0 0 43

Concur Total 1 7 7 0 3 6 0 2 0 6 24 13 25 17 22 19 11 4 6 0 5 1 3 11 2 2 1 0 0 2 0 3 10 3 2 0 0 2 0 2 1 0 223

1 1 2 0 0 5 0 0 0 6 21 12 10 15 9 9 1 0 2 0 2 1 1 10 2 2 0 0 0 1 0 1 5 0 0 0 0 1 0 1 1 0 122

Not Override Withdrawn Disputed 0 0 0 3 2 1 1 4 0 0 0 0 2 1 0 0 1 0 0 0 0 1 1 0 0 0 0 0 0 0 2 1 0 1 0 0 1 12 2 1 1 0 4 9 0 0 9 1 5 2 3 2 1 1 2 2 0 0 0 0 3 0 0 0 0 0 2 0 0 1 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 2 4 1 0 0 3 0 0 2 0 0 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 0 0 0 38 52 11

13 after the mid-1950s, so there was little reason for him to stake an ideological position different from the party, or expect the party to take a position far from his ideal point. 5 The incidence of vetoes in Mexico, 1917-1997

Vetoes were common in post-Revolutionary Mexican politics. Between 1917 and 1957, there were at least 223 vetoes, with a mean of 5.4 vetoes per year. 14 Carpizo (1978, p. 91) identifies an additional 23 vetoes after 1963, which is the last year covered by the database. The last veto recorded by Carpizo was in 1969, and is generally accepted as the last veto before Fox vetoed the Ley de Desarrollo Rural in March 2001. Table 1 presents the vetoes by president, while Table 2 shows vetoes by year. 15 The data only includes vetoes that entered the Chamber of Deputies. It is possible that some vetoes were sent to the Senate if the legislation originated in that chamber; if the Senate did not vote on the veto, it never arrived at the Chamber of Deputies and would be missing data. The tables classify the bills in three categories according to type of legislation. Budget vetoes are vetoes of appropriations bills (which according to constitutional scholars cannot be vetoed because they are approved by only one chamber). Sometimes, the president vetoed the entire budget, or specific parts of the budget (these are not line- item vetoes, but rather cases —before 1929 or so—when the budget was submitted in several parts, much like U.S. appropriations). Often, the president also vetoed specific spending projects submitted by deputies, usually bills that spent money outside of the regular budget in a specific state or municipality—pork barrel expenditures. Pension vetoes are vetoes of private bills that authorized or amended pensions to specific persons, usually veterans of the Revolution, or families of Revolutionary veterans or politicians. Some pensions are treated as appropriation bills (and thus considered only by the Chamber of Deputies), while others are treated as regular authorizing legislation. All pensions, regardless of the legislative procedures used, are classified in this category. The third category is non budget legislation, which includes all regular public bills (including taxes) that do not involve appropriations. Of the 223 vetoes, over 60 percent (135) were of pensions. The reasons for these vetoes is unclear, though they might respond to changes in the composition of the socalled “Revolutionary family” (Brandenburg 1964). After the De la Huerta rebellion of 1923, it is likely that family members of the rebels were purged from the pension roles by presidential vetoes. Later, when the PNR is formed, manipulating pensions might have been a way to define who was a genuine member of the party (ie., a callista), and who had just been lucky to be in power in 1929 and required exclusion later on. These vetoes begin in 1926 and continue to the end of the Maximato. There were 45 vetoes of budget bills, beginning in 1918 and ending in 1933. One out of five vetoes fell on the budget. There were no vetoes between 1923 and 1927 for the simple reason that the president issued these budgets by extraordinary decree; no bill was sent to Congress, so there was nothing to veto (Weldon n.d.). The vetoes end in 1933 for two reasons. First, the budget law was reformed so that the president would send a unified budget instead of sending budgets for specific programs or min istries. Since there is only a package veto in Mexico, it is more costly to veto the whole budget than just one area. Second, reelection of deputies and senators was prohibited in 1933, and the number

14

The data is from a database covering all legislation in the Chamber of Deputies between 1917 and 1964, compiled by Jeffrey A. Weldon, María del Carmen Nava Polina, and Jorge Yáñez López, for Molinar and Weldon (n.d.). 15 Much of the data on vet oes between 1917 and 1946 was originally published in Weldon (1997a and 1997c).

14 of pork barrel bills fell suddenly to nearly zero once they no longer had to please local constituents (Weldon n.d.). 16 The last category is regular legislation. These vetoes are more or less evenly distributed until about 1943. These vetoes are among the most interesting, because they are vetoes of regular public bills, over which we have the expectation that the president always got what he wanted from the Congress. Between 1917 and 1964, only two presidents did not veto legislation: De la Huerta, who ruled for six months in 1920, between the assassination of Carranza and the inauguration of Obregón; and López Mateos, who governed during years of strong presidentialismo. Note that some of the other vetoes identified by Carpizo would correspond to López Mateos’ term. The two presidents who vetoed the most bills were Calles (1924-28) and Ortiz Rubio (1930-32), with 43 and 45 vetoes respectively. It is not too surprising to find these presidents at the forefront, because both had conflicts with the Jefe Máximo of the time. Although the official party had not bee n founded yet, Obregón was the leader of the Revolutionary family during the Calles presidency. Ortiz Rubio was the president who was most in conflict with Calles when the latter had become Jefe Máximo, to the point that the former had to resign. Bargaining and posturing should have been present. There were 45 vetoes after 1937: 20% of all vetoes took place under the presidencialismo era. The president was supposed to be dominant in this period, and there is no apparent reason to engage in policy bargain ing with him. And given that elections were largely won by fraud, there appears to be no third party watching politicians’ positions, hence no reason to seek publicity. Data are interesting on the side of Congress’ response to executive vetoes. Of the 223 executive vetoes, 11 were withdrawn by the president at some point afterward. The presidents cancelled vetoes more or less randomly throughout the period under study, though 5 were during the Maximato . One explanation of these peculiar withdrawn vetoes could be a change in players’ updated preferences over policy. The Chamber of Deputies actively concurred in about 55 percent of the vetoes (122 in all). In these cases, the lower chamber, voted either to agree that the bill should have been discarded, or the deputies voted to accept the modifications to the bill suggested by the president. A rather high incidence of post-veto agreement. If the deputies agree that the bill should have been vetoed, then why did they pass it in the first place? The Chamber of Deputies voted to override 52 of the vetoes, over 23 percent of the time. Veto overrides are examples of greatest conflict, and in a sense are the most difficult to explain in any situation. Why would the president veto if he knew the chamber would override? Again, sometimes a bargaining explanation makes some sense, especially in the periods of divided government or divided dyarchy (president/Jefe Máximo). The publicity argument could also apply in these occasions when the president takes a stand against the opposition in a divided government situation, or within the party in divided diarchy. All 12 overrides of vetoes over budget or regular public bills happened during periods of divided government or during the Maximato. However, the Chamber of Deputies overrode 6 vetoes—all pension bills—in the late 1940s and early 1950s, when the metaconstitutional presidency was strong. Why would a president risk his image as undisputed leader of the party by vetoing a bill that is later overridden? One possible explanation: these years correspond to the short period in which the PRI 16

By local constituents we mean political bosses, not voters, due to the disfranchisement from electoral fraud.

15 experimented again with closed primaries instead of closed conventions or direct nominations by sectors. Could these vetoes be signals to certain factions within the official party? Table 3 Resolution of Vetoes, by Type of Legislation 1917-63

Type of Legislation Budget Pensions Non Budget Totals

Concur 33 64 25 122

Resolution of Veto Not Disputed Override 5 24 9 38

5 40 7 52

Withdrawn

Total

2 7 2 11

45 135 43 223

Table 4 Presentation of Vetoes, by Month 1917-63 Month January February March April May June July August September October November December

Vetoes 18 18 43 4 7 8 0 4 15 2 5 11 135

% 13.3% 13.3% 31.9% 3.0% 5.2% 5.9% 0.0% 3.0% 11.1% 1.5% 3.7% 8.1% 100.0%

Congress in session

Table 3 disaggregates the vetoes by type of legislation and resolution of the veto. We find that over 75 percent of the vetoes (excluding those withdrawn) of appropriations legislation was agreed to by the lower chamber. The chamber overrode about 11 percent of these vetoes. The Chamber of Deputies concurred in over 60 percent of the vetoes of regular public bills, while 17 percent were overriden. Pension vetoes, on the other hand, received a different treatment. The deputies concurred in the veto in half of the cases, but they overrode more than 30 percent of the vetoes of pensions. This would seem to indicate differences between the branches on membership in the Revolutionary family.

16

Table 5 Vetoes, by Period 1917-1963

Period Divided Government Maximato Strong Presidentialism Totals

Period Divided Government Maximato Strong Presidentialism Totals * Withdrawn vetoes are excluded.

Type of Legislation Years Budget % Pensions % Non Budget % Total 1917-28 22 31.9% 29 42.0% 18 26.1% 69 1929-37 23 21.1% 77 70.6% 9 8.3% 109 1938-63 0 0.0% 29 64.4% 16 35.6% 45 45 135 43 223 Resolution of Veto* Years Concur Not Override Disputed 1917-28 48 70.6% 10 14.7% 10 1929-37 48 47.1% 18 17.6% 36 1938-63 26 61.9% 10 23.8% 6 122 38 52

Total 14.7% 35.3% 14.3%

68 102 42 212

Table 6 Resolution of Vetoes, by Period and Type of Legislation* 1917-1963

Period Divided Government Maximato Strong Presidentialism Totals * Withdrawn vetoes are excluded.

Years

Concur

1917-28 1929-37 1938-63

14 19 0

Budget Not Disputed 3 2 0

33

5

Override

Concur

4 1 0

27 24 13

5

64

Pensions Not Override Disputed 2 0 15 34 7 6 24

40

Concur 7 5 13 25

Non Budget Not Override Disputed 5 6 1 1 3 0 9

7

Total 68 102 42 212

17

Tables 5 and 6 break down the vetoes by historical period, type of legislation, and resolution. We divide post-Revolutionary Mexican political history into three periods, each closely correlated with the presence or absence of the conditions for metaconstitutio nal presidentialism. The first period is “Divided Government,” which extends from 1917 to 1928, the year before the foundation of the PNR. There were 69 vetoes in this period, averaging 5.8 per year. In these years, 32 percent of the vetoes were of appropriations legislation, the highest proportion of any of the three periods. More than a quarter of the vetoes were on regular public bills. These types of vetoes are to be expected in periods of divided government under any model. Over 14 percent of the vetoes were overriden in this period, 4 appropriations bills and 6 other public bills. The second period is the “Maximato,” covering the years 1929-37. The latter year was selected in part because the exile of Calles was in 1936, and in part because the 37th Legislature (1937-40) was more completely cardenista than the previous one (Enríquez Perea 1988). These are years of unified government but divided dyarchy. They were also the years with the most vetoes, 109 total, averaging over 12 per year. Vetoes over appropriations were just as common as in the years of divided government, though there were relatively fewer vetoes of other public bills. On the other hand, this was the period when the vetoes of the private pensions were most common, 57 percent of all pension vetoes occurred in this period, and pensions make up over 70 percent of all vetoes during the Maximato. This was the period with the most veto overrides; 35 percent of the vetoes were overturned, at least by the Chamber of Deputies. Among the overrides, 34 were of pension bills, and only one was a budget bill and another was a regular public bill. During the third period, “Strong Presidentialism” (1938-63), all three of the metaconstitutional conditions were in effect. It is very difficult to explain vetoes with orthodox models in this period. However, there were still about 1.7 vetoes per year on average in this period. As discussed before, there were no longer vetoes of appropriations legislation, but more than a third of all vetoes were of public bills. Furthermore, more than 14 percent of the vetoes were overriden—all pensions. 6 Conjectures on Mexican vetoes

The frequency of vetoes and veto overrides in Mexico would be surprising to most mexicanólogos. We are accustomed to the old myths of Goodspeed (1947, p. 444), who claimed, “Congress does not pass legislation that would be inimical to the wishes of the executive branch of government,” or “No piece of legislation desired by Obregón and effectively pushed by him ever failed of passage. No legisla tion unfavorable to the president was ever passed. No bill that did not have presidential approval was ever successful. Better than 98 per cent of the legislation passed by Congress was initiated by the president. The use of a presidential veto was unknown” (p. 149). These myths became the stylized facts. We now know that in his first term, less than 20 percent of Obregón’s bills were passed, most of the legislation approved in his whole term had been introduced by deputies, and Obregón himself vetoed 11 bills. Worse, three of these were overriden. These vetoes, however, make sense with various models. So do the vetoes of the Maximato, though the explanation is more nuanced (the two veto players being the president and Jefe Máximo rather than two branches). Nonetheless, how do we explain

18 the relatively frequent vetoes between 1937 and 1969? Furthermore, how do we reconcile these vetoes with the complete absence of vetoes between 1970 and 1997? There are several possible answers. The first is that vetoes are just simply mistakes, trembling hands on the part of either Congress or the president. Or perhaps they are merely corrections of legislation, a natural part of the review process. No one was necessarily bargaining, taking a position, or claiming credit. The presidents were just fixing aberrant legislation by somewhat “incompetent” legislators (cf. policy valence in Londregan 2000). The six overrides in the years of strong metaconstitutional presidentialism, however, challenge this explanation. Another possibility lies in the unusual nature of the timing of the vetoes in Mexican politics. Most legislation in this period was approved at the end of the regular session in December, very often on the very last day (Weldon N.d.2). Congress then closed its doors, and the president still had time left on his ten-day limit to consider the legislation. There is no pocket veto in Mexico, so the president could return the bill at any time in the eight- month period between January 1 and September 1, when the next regular session began. Nor does he have to publish the bill at any particular point in the interim. Since there is often no rush to publish bills, it gives the president ample time to think about the legislation and perhaps even change his mind. So it is possible that the frequency of vetoes in Mexico is due in part to changes in preferences over time on the part of the president. Table 4 presents the month in which the veto message was sent, if available. We find that vetoes were most common in March, 32 percent of the total, at least two months after the bill was passed. There were also many vetoes presented in September, when the Congress opened again (11 percent of the total), allowing 8 months for the president to change his mind. Furthermore, there is no time limit on voting on vetoes in Congress. Bills that are not voted on never die, including vetoes. Some vetoes were resolved in a matter of days, while others took years—up to ten years—to come to a vote. 17 This may explain some of the overrides. For example, a president vetoes a bill, Congress does not dispute the veto right away, but some future Congress, perhaps under a different administration, has a change of heart and overrides the veto. In these cases, the overrides could even have the approval of the current president, so executive dominance is not really challenged. A third explanation for vetoes in the years following 1937 can be derived from the economic historian Raymond Vernon (1963). He recognizes that the president of Mexico is extraordinarily powerful, and does not dispute that his strength comes from his being the head of the party. However, he emphasizes that the party itself—even in the post World War II years —is exceptionally diverse and heterogeneous. There are competing demands from agrarian, labor, bureaucratic, and business interests. It is almost impossible for a president to please everyone in the party at the same time. This leads to vacillating executives, who move from one side of the spectrum to the other in order to keep the peace within the party. Or it leads to presidents muddling through the middle, going practically nowhere.

17

Although there is no evidence of this in the record, a Congress could wait until intervening elections led to majorities sufficiently large to override the veto…and there is nothing that the president could do about it!

19 Although Vernon does not mention vetoes, it is possible that the president uses the veto power when he needs to return the party to the positio n that he most prefers — perhaps something of a signaling game. This may explain the occasional vetoes in the final period. It does not, however, explain why the vetoes disappeared in the 1970s. Perhaps the increasing electoral competition in the 1980s and 1990s forced greater cohesion on the PRI? 7 Conclusion

1. Vetoes have occurred frequently in post-revolutionary Mexico. Many took place when government was divided between factions of the Revolutionary Family (1917-1929) and under the Maximato (1929-1936), when Calles controlled Congress by means of the party. Yet one out of five executive vetoes took place under strong presidencialismo. 2. One out of five executive vetoes was overridden by the lower chamber of Congress. Most overrides involved vetoes on pensions bills distributing the spoils among members of the party. Yet 6 overrides took place under strong presidencialismo . 3. Presidencialismo presents empirical anomalies. Contradicts conventional wisdom. 4. Data seem to confirm that Mexican authoritarian regime was not a monolith; rather Swiss-cheese with holes (Knight 19xx). The agency that the president performed in favor of party members did not function smoothly all the time. In this case bargaining and posturing would still seem to play a role in strong presidencialismo . 5. Fox’s veto: this looks very much like a search for publicity right before and right after the 2000 election in Mexico. Bill = Pork for PRI’s organized peasants. Passed by PAN+PRD and delayed by PRI in Senate; passed then by PRI+PRD in the Senate after the elections and vetoed by Fox (see Magar 2001, chapter 6).

20

References Baron, David P., and John Ferejohn. 1989. Bargaining in Legislatures. American Political Science Review 83 (4):1181-1206. Brandenburg, Frank. 1964. The Making of Modern Mexico. Englewood Cliffs, N.J.: PrenticeHall. Burgoa, Ignacio. 1994. Derecho constitucional mexicano. 9th ed. Mexico City: Porrúa. Cameron, Charles M. 2000. Veto Bargaining: Presidents and the Politics of Negative Power. New York: Cambridge University Press. Carpizo, Jorge. 1978. El presidencialismo mexicano. Mexico City: Siglo XXI. Córdova, Arnaldo. 1995. La Revolución en crisis: la aventura del maximato. Mexico City: Cal y Arena. Cosío Villegas, Daniel. 1973. El sistema político mexicano. Mexico City: Joaquín Mortiz. Cox, Gary W., and Scott Morgenstern. 1998. Reactive Assemblies and Proactive Presidents : A Typology of Latin American Presidents and Legislatures. Paper read at the XXI International Congress of the Latin Americ an Studies Association, at Pittsburgh. Dealy, Glen C. 1992. Pipe Dreams: The Pluralistic Latins. In Politics and Social Change in Latin America: Still a Distinct Tradition?, edited by H. J. Wiarda. Boulder: Westview. Dulles, John W. F. 1961. Yesterday in M exico: a Chronicle of the Revolution, 1919-1936. Austin: University of Texas Press. Edelmann, Alexander T. 1969. Latin American Government and Politics: The Dynamics of a Revolutionary Society. Homewood: Dorsey Press. Enríquez Perea, Alberto. Los sectores populares a la Cámara de Diputados: XXXVII Legislatura del Congreso de la Unión, 1937-1940. Revista Mexicana de Ciencias Políticas y Sociales 34:134 (Oct.-Dec.), pp. 161-173. Federalist Papers 1788 [1961], by Madison, Hamilton, and Jay, edited by C. Rossiter. New York: Penguin. García Hamilton, José Ignacio. 1990. Los orígenes de nuestra cultura autoritaria (e improductiva). Buenos Aires: Calbino. Garrido, Luis Javier. 1982. El Partido de la Revolución Institucionalizada. Mexico City: Siglo XXI. Garrido, Luis Javier. 1989. The Crisis of Presidencialismo. In Mexico's Alternative Political Futures, ed. Wayne A. Cornelius, Judith Gentleman, and Peter H. Smith. La Jolla, Calif.: Center for U.S.-Mexican Studies, University of California, San Diego, 417-34. Goodspeed, Stephen Spencer. 1947. “The Role of the Chief Executive in Mexico: Politics, Powers and Administration.” Ph.D. dissertation, University of California. Groseclose, Tim, and Nolan McCarty. 2001. The Politics of Blame: Bargaining Before an Audience. American Journal of Political Science 1:100-119. Hansen, Roger D. 1971. The Politics of Mexican Development. Baltimore: Johns Hopkins. Indridason, Indridi H. 2000. Process, Politics and Institutions: Three Essays in Positive Political Theory. Ph.D. dissertation, University of Rochester. Kernell, Samuel. 1991. Facing an Opposition Congress: The President's Strategic Circumstance. In The Politics of Divided Government, edited by G. W. Cox and S. Kernell. Boulder: Westview. Kiewiet, D. Roderick, and Mathew D. McCubbins. 1988. Presidential Influence on Congressional Appropriations. American Journal of Political Science 32:713-36. Lambert, Jacques, and Alain Gandolfi. 1987. Le système politique de l'Amérique latine. Paris: Presses Universitaires de France. Linz, Juan J., and Arturo Valenzuela, eds. 1994. The Failure of Presidential Democracy. Edited by J. J. Linz and A. Valenzuela. 2 vols. Baltimore: Johns Hopkins University Press.

21 Loaeza, Soledad. 1989. El llamado de las urnas. Mexico City: Cal y Arena. Londregan. 2000. Legislative Institutions and Ideology in Chile. CUP. Magar, Eric. 2001. Bully Pulpits: Posturing, Bargaining, and Polarization in the Legislative Process of the Americas. Ph.D. dissertation, University of California, San Diego. Mainwaring, Scott, and Matthew Soberg Shugart, eds. 1997. Presidentialism and Democracy in Latin America. New York: Cambridge University Press. Mayhew, David R. 1974. Congress: The Electoral Connection. New Haven: Yale University Press. McNollgast. 1994. Legislative Intent: The Use of Positive Political Theory in Statutory Interpretation. Law and Contemporary Problems 57 (1):3-37. Mecham, J. Lloyd. 1959. Latin American Constitutions: Nominal and Real. Journal of Politics 21:258-75. Medín, Tzvi. 1982. El minimato presidencial: historia política del maximato. Mexico City: Era. Medina, Luis. 1978. Evolución electoral en el México contemporáneo. Mexico City: Comisión Federal Electoral. Meyer, Lorenzo. 1978. El conflicto social y los gobiernos del maximato, vol. XIII of Historia de la Revolución Mexicana. Mexico City: El Colegio de México. Meyer, Lorenzo, and José Luis Reyna. 1989. Introducción. In Los sistemas políticos de América Latina, edited by L. Meyer and J. L. Reyna. Mexico City: Siglo XXI. Meyer, Lorenzo, Rafael Segovia, and Alejandra Lajous. 1978. Los inicios de la institucionalización: la política del maximato, vol. XII of Historia de la Revolución Mexicana. Mexico City: El Colegio de México. Molinar, Juan. 1991. El tiempo de la legitimidad: elecciones, autoritarismo y democracia en México . Mexico City: Cal y arena. Molinar Horcasitas, Juan, and Jeffrey A. Weldon. N.d. Procedimientos legislativos de la Cámara de Diputados, 1917-1964. Serie I, vol. I, tomo 2 of Enciclopedia Parlamentaria de México. Mexico City, Instituto de Investigaciones Legislativas, Cámara de Diputados/Porrúa. Nacif, Benito. 1995. “The Mexican Chamber of Deputies: the political significance of nonconsecutive reelection.” Unpublished Ph.D. dissertation, University of Oxford. O'Donnell, Guillermo. 1994. Delegative Democracy. Journal of Democracy 5 (1):55-69. Pious, Richard M. 1979. The American Presidency. New York: Basic Books. Romer, Thomas, and Howard Rosenthal. 1978. Political Resource Allocation, Controlled Agendas, and the Status Quo. Public Choice 33:27-44. Schelling, Thomas C. 1960. The Strategy of Conflict. Cambridge: Harvard University Press. Scott, Robert E. 1958. Legislatures and Legislation. In Government and Politics in Latin America, edited by H. E. Davis. New York: Ronald Press. Segovia, Rafael. 1975. La politización del niño mexicano. Mexico City: El Colegio de México. Shugart, Matthew Soberg, and Stephan Haggard. 2001. Institutions and Public Policy in Presidential Systems. In Presidents, Parliaments, and Policy, ed. Stephan Haggard and Mathew D. McCubbins. Cambridge: Cambridge University Press, pp. 64-102. Tena Ramírez, Felipe. 1985. Derecho constitucional mexicano. 21st ed. Mexico City: Porrúa. Vernon, Raymond. 1963. The Dilemma of Mexico's Development: the Roles of the Private and Public Sectors. Cambridge: Harvard University Press. Villa Aguilera, Manuel. 1987. La institución presidencial: el poder de las instituciones y los espacios de la democracia. Mexico City: Coordinación de Humanidades, Universidad Nacional Autónoma de México. Weldon, Jeffrey A. 1997a. El crecimiento de los poderes metaconstitucionales de Cárdenas y Avila Camacho: su desempeño legislativo, 1934-1946. Diálogo y Debate 1:1 (April-June 1997), 11-28.

22 Weldon, Jeffrey A. 1997b. The Political Sources of Presidencialismo in Mexico. In Presidentialism and Democracy in Latin America, ed. Scott Mainwaring and Matthew Soberg Shugart. Cambridge: Cambridge University Press, 1997, 225-58. Weldon, Jeffrey A. 1997c. El presidente como legislador, 1917-1934. In El Poder Legislativo en las décadas revolucionarias, 1908 -1934, ed. Pablo Atilio Piccato Rodríguez, serie I, vol. I, tomo 3 of Enciclopedia Parlamentaria de México. Mexico City, Instituto de Investigaciones Legislativas, Cámara de Diputados, LVI Legislatura, 1997, 117-45. Weldon, Jeffrey A. 2001. El Congreso, las maquinarias políticas y el Maximato: las reformas antireleccionistas de 1933. In El debate sobre la reelección legislativa en México: una perspectiva histórica e institucional, ed. Fernando F. Dworak. Weldon, Jeffrey A. N.d. The Legal and Partisan Framework of the Legislative Delegation of the Budget in Mexico. In Legislative Politics in Latin America, ed. Scott Morgenstern and Benito Nacif. Cambridge: Cambridge University Press, forthcoming (2001). Weldon, Jeffrey A. N.d.2. Las estrategias del presidente con gobierno dividido. In Episodios republicanos: experiencias de mayoría dividida en México a partir de 1867, ed. María Amparo Casar and Ignacio Marván Laborde. Mexico City: Océano, in press. Wiarda, Howard J. 1992. Social Change, Political Development, and Tradition. In Politics and Social Change in Latin America: Still a Distinct Tradition?, edited by H. J. Wiarda. Boulder: Westview.

23 Table 7 Probability of Overriding Veto 1917-63 Probit Estimation with Robust Standard Errors in parentheses Variable

Coefficient

Budget

-0.836* (0.271) -0.39 (0.245) -0.904* (0.265) -0.055* (0.02) 1.542 (0.7)

Non Budget Divided Government Trend (year) Constant * Significant at the .01 level.

212 obs. Log likelihood = -106.5. Chi2 = 18.97, significant at 0.0008.