5 Presidentialism
\    5.1   DEFINING'PRESIDENTIAL SYSTEM'
i    Just as electoral systems are divided into majoritarian and f     proportional, so are democratic political systems generally k     divided into presidential and parliamentary. However, the ;     latter distinction is more difficult to draw than the former >'.    one. To be sure, presidential and parliamentary systems can be t     defined by mutual exclusion. Sure, a presidential system is f    non-parliamentary and, conversely, a parliamentary system is c     non-presidential. But the distribution of the real world cases into these two classes reveals impermissible bedfellows. The reason for this is, on the one hand, that presidential systems are for the most part inadequately defined; and, on the other hand, t     that parliamentary systems differ so widely among themselves as to render their common name a misnomer for a deceitful togetherness. We shall look into this later. Here, and first, we are required to define presidential systems and thereby to make sure that they are not confused with mock presidential forms or erroneously perceived as mixes, as quasi- or near-parliamentary presidentialisms.
The first defining criterion for a presidential system is the direct or direct-like popular election of the head of state for a fixed time span (that may range from four to eight years). This criterion is doubtless a necessary defining condition; but by no means a sufficient one. Austria, Iceland and Ireland have recourse to a direct popular election of their presidents and yet are only, at most, facade presidentialisms. Regardless of what the constitution says about their power prerogatives,1 the presidents in question are little more than figureheads, and Austria, Iceland and Ireland perform in all respects like parliamentary systems. These countries cannot be classified, then, among the presidential ones, in spite of their popularly elected presidents.2
83
84 Comparative Constitutional Engineering
Presidentialism
85
A second defining criterion is that in presidential systems the government, or the executive, is neither appointed nor dis-missable via a parliamentary vote. Governments are a presidential prerogative: it is the president that discretionally appoints and discharges cabinet members. To be sure, presidents may pick their ministers in a manner that pleases parliament; even so, cabinet members are and remain presidential appointees. Let it be immediately noted that the criterion in question is not violated if a parliament is given the power of censuring individual cabinet ministers, or even by the rare cases in which a parliamentary censure entails that a minister must be removed from office.3 The criterion is not violated because in either case it is still the president that unilaterally retains the nominating power and fills cabinet posts as he or she sees fit.
Do criterion one plus criterion two suffice to identify a presidential system? I would say almost, but not quite. For it must be very clear that a pure presidential system does not allow for any kind of 'dual authority' between the president and his cabinet. On this score Lijphart proposes the criterion 'a one person executive5. But this is perhaps too restrictive; it implies that the head of state must also be the head of government. True, this is generally the case. Still, I prefer a somewhat looser formulation, such as this: that the line of authority is neatly streamlined from the president down. In short, the third criterion is that the president directs the executive.
So, a political system is presidential if, and only if, the head of state (president) (i) results from popular election, (ii) during his or her pre-established tenure cannot be discharged by a parliamentary vote, and (iii) heads or otherwise directs the governments that he or she appoints. When these three conditions are jointly met, then we doubtlessly have a pure presidential system - or so says my definition.4
We still have a loose end to tie up. The first criterion reads, in full, that a president must result from a direct or 'direct-like' election. How open-ended is that? Direct-like accounts for the United States and for countries such as Argentina and, formerly, Chile (up until Allende), whose president is elected by parliament when no candidate receives an absolute majority of
the popular vote. Since the established practice in such cases is to elect the candidate that has obtained the popular relative majority, this kind of indirect election can be assimilated to a direct one. Bolivia practices instead a parliamentary selection among the three front runners (both in 1985 and 1989 it picked the second one) and therefore represents a dubious direct-like case. On the other hand, until 1988 Finland was definitely not direct-like: the elected presidential electoral college was the true president-maker in that its freedom of choice was unrestricted. The cut-off point is, then, whether the intermediary body (electoral college or parliament) is allowed choices of its own. If it performs as a rubber stamp, then the difference between direct and indirect popular election is immaterial; if it can make choices, then criterion one is somewhat violated (but see infra 7.3).
The general point is this: that presidential systems (just like parliamentary systems) are such, and hang together, on account of a systemic logic of their own. Thus, before reassigning presidentialism to another class - whether semi-presidential-ism, near-parliamentarism, and the like - we must check whether a given variance violates that logic or not. What if a president is entitled to dissolve parliament? What if parliament is entitled to impose the dismissal of cabinet members? What if a president can be recalled by a popular vote? Do these and other anomalies bring about a different mechanism that abides by a different logic? If the answer is yes, then let us see what kind of new mechanism we have, and let us reclassify a presidential system accordingly. But let us not rush headlong into discovering or inventing 'new systems' every time that a country borrows a device from another system.
Under this proviso I trust that we now have a definition that neatly sorts out what 'presidentialism' includes and, conversely, excludes. On its basis we come up, today, with some 20 countries, mostly concentrated in Latin America.5 The reason that Europe has no pure presidential systems, whilst we find them from the Canadian border down through all of the two Americas, is historical and does, not attest to any deliberate choice. When the European states started practicing constitutional government, all (except France, which became a repub-
86
Comparative Constitutional Engineering
lie in 1870) were monarchies; and monarchies already have a (hereditary) head of state. But while in Europe there was no room (at least until 1919) for elected presidents, in the New World almost all the new states became independent as republics (Brazil and, in a way, Mexico, being the temporary exceptions), and therefore had to have elected heads of state, i.e., presidents. The division between presidential versus parliamentary systems did not result, then, from any theory that debated whether one form was superior to the other. But the time has come for this debate to occur and, thus, for a comparative assessment of how the two systems perform.
By and large, presidentialism has performed poorly. With the sole exception of the United States, all other presidential systems have been fragile - they have regularly succumbed to coups and breakdowns.6 However, the exception of the United States, lonely as it is, is an important one. Furthermore, the United States provides the original from which all the other presidential systems are derived. So from here we start.
5.2   THE AMERICAN PROTOTYPE
The Washington model is characterized, more than by any other single feature, by the division-separation of power between President and Congress. This separation is not easily spelled out. For Neustadt (1960, p. 33) the Founding Fathers did not create a government of'separated powers' but, instead, 'a government of separated institutions sharing power'. But Jones (1990, p. 3) corrects: we now 'have a government of separated institutions competing for shared power'; and he ultimately perceives the American polity as a 'truncated system'. These and other important nuances notwithstanding, the essence of the matter is this: separation consists of'separating' the executive from parliamentary support, whereas power sharing means that the. executive stands on, and falls without, the support of parliament.7 And on this criterion what the United States has is indeed power separation.
With power separation, then, a parliament cannot interfere in the internal affairs, in the interna corporis, of the executive
Presidentialism
87
realm, and especially cannot dismiss (impeachment aside) a president. By virtue of the same logic, or of the same principle, power separation implies that a president cannot dissolve a parliament. And this is indeed the case in the United States and in most presidential systems. But what if a president is given such power? Does this constitute a violation of the separation principle? Yes, even though my inclination is to consider the power of dissolution of parliament an anomaly that does not transform - if all my three defining criteria apply - a presidential system into another kind of system. True, the power of dissolving parliament enhances the presidential /power; but the efficacy of this deterrent is generally overestimated and, in any event, on this ground alone I do not see sufficient grounds for reclassifying a presidential structure.
The American system is also characterized by most authors as a system of checks and balances. Right; but this is hardly a distinctive qualification, for all truly constitutional systems are systems of checks and balances. The point is that we can have checks and balances without power separation, and that the uniqueness of American-type presidentialism is precisely that it checks and balances power by dividing it. Therefore (I insist) the defining and central feature of the Washington model is an executive power that subsists in separateness — on its own right as an autonomous body.
This does not imply in the least that the American president is indifferent as to whether he may, or may not, enjoy Congressional support. Indeed, the more we have a divided power structure, the more we need - it would seem - a 'united government', i.e., a same majority in control of the executive and of Congress. This has been both the theory and the practice of American governance for a century and a half. However, the prevailing pattern has now become one of 'divided government'. Eisenhower was, in 1954 and again in 1957, the first president in seventy-two years to be confronted by a Congress controlled by the opposition party. Since then 'from 1955 through 1992 the government was divided for twenty-six of thirty-eight years'; and 'from 1969 through 1992 divided government prevailed in twenty of twenty-four years' (Sundquist, 1992, p. 93). The Clinton presidency re-
88
Comparative Constitutional Engineering
Presidentialism
89
established an undivided majority which lasted, however, only two years (1993-4). So the trend has unequivocally been, in the last forty years, a trend of minority presidents, of presidents whose party did not have a majority in the Houses. While Republicans have held the White House for all but four years (under Carter) between 1968 and 1992, Democrats have constandy controlled (until 1994) Congress and - but for six years - both Houses since 1955.
This strikes observers as a momentous turnaround which confronts the American system with deadlock and gridlock. However, according to David Mayhew there is no reason to worry, for in his findings 'unified as opposed to divided control has not made an important difference in . . . the enactment of a standard kind of important legislation: . . . important laws have materialized at a rate largely unrelated to party control' (1991, p. 4). But Mayhew is, I believe, quite wrong. The difference between unified and divided government cannot be belittled, and the reason why this difference does not show up in his findings makes the case for the American system more worrisome than ever.8 I come back to this shortly^.First I wish to state the full case as I see it.
The basic assumption about presidential systems is that they are conducive to strong and effective government — both per se and in comparison to parliamentary systems. But this is an assumption that stands on thin ice. The fact that the American system has long coped with its problems does not detract from the fact that a divided power structure engineers paralysis and stalemates better than any other. And, indeed, does the American system still work? Looking back, we see that the division of power has been compensated not only by consonant majorities, by the coinciding of the president's majority with the parliamentary one, but also, customarily, by con-sociational practices, especially bi-partisan concurrence in foreign politics.
However - and barring an unlikely return to enduring cycles of undivided government - the pattern that has emerged from the Fifties onward confronts us with a split, antagonistically divided polity whose two major component elements perceive their respective electoral interest to be, across the board, in the
■ failure of the other institution. For a Democratic controlled :' Congress to go along with a Republican president is to help bring about another Republican presidency. Conversely, a minority president (in Congress) seeking to restore undivided government is prompted to run against Congress playing, as it were, the blame game.
Ironically, then, the belief that presidential systems are strong systems draws on the worst possible structural arrangement - a divided power defenseless against divided government - and fails to realize that the American system works, or has worked, in spite of its constitution, hardly thanks to its consdtution. To the extent to which it is still able to perform, ; it requires, in order to unblock itself, three factors: ideological unprincipledness, weak and undisciplined parties, and locality centered politics. On these counts a president can win over the congressional votes that he may need by 'horse trading' constituency favors. We thus end up with the institutionalization of pork-barrel politics - nothing much to be admired. And what we have, structurally, is in fact a weak state.
We may now revert to Mayhew's findings that the divided party control of the presidency and of Congress does not appear to affect and worsen in any significant way the ' legislative output of Congress. Let us assume, for the sake of argument, that this finding does hit upon something. If so, however, it does not speak to the point intended by Mayhew but, rather, to the increasing atomization of the American party. That is to say that the factor or variable at play is the crossing of party lines in Congressional voting. Whether a president has, or does not have, 'his majority' in Congress matters and makes a difference on the assumption that the notion of majority is a meaningful one, that there is something there that hangs together and acts cohesively. If, however, a majority exists only on paper, if it has to be assembled occasion by occasion, then it stands to reason that the difference between united and divided government may make little difference. The point thus becomes that even when the majority is undivided on paper, the reality of the matter is that today an American president never has a true and reliable majority in Congress.
90
Comparative Constitutional Engineering
Presidentialism
91
To be sure, the interest of a Democratic congress is to have a 1 Democratic president who succeeds. It is also evident that :| there must be, for the consumption of public opinion, some | showing that united government does provide 'unity' of sup- | port, that is, greater support than otherwise. Yet the overriding 11 consideration has become, for each member of Congress, how | well his or her voting record fares, vote by vote, in their i| districts. It is widely admitted that American parties are little | more than electoral parties, and such only in the feeble sense that | they provide labels for two candidates to run their campaigns | one against the other in single-member districts. But are they 1 any more than that as parliamentary parties? I would say that the ■ -1 increasing dissolution of the electoral party must somehow | reflect itself also at the Congressional level.9 We are thus f witnessing an increased and increasingly localistic, constitu- | ency driven fragmentation of the American congressional- % legislative party (though a lesser one, for obvious reasons, in § the Senate). And a parliament in which politics becomes retail politics {supra 4.1), in which more and more members perform ;| as constituency lobbyists, as constituency errand boys, is a parliament in which majorities easily become a fleeting, va- % porous entity. An astute observer, Nelson Polsby (1993, p. 33), writes:
In the practical politics of today, legislation frequently requires a complicated sort of agreement: a coalition must be built that crosses . . . party. This coalition is the product of a series of negotiations not only on the substance of various measures but also on the apportionment of credit for the benefits those measures may generate and blame for the pain they may cause . . . Passing legislation that causes pain is thus a risky thing for elected officials to do . . . Amazingly, sometimes they do it, if a deal can be worked out to share the blame. :
Polsby knows his characters and portrays them well. But can the process that he describes be called coalition building? I realize that 'coalition' is used sloppily, in the American politics field, as a blanket term for almost everything. Still, coalition is here a misleading term. Coalitions, proper, did exist in Con-
gress in the past, for the Southern Democrats did perform as a coalition-like bloc. Perhaps coalitions, proper, still exist today. But what Polsby describes is definitely not a coalitional state of affairs. Coalitions are agreements: but not single-issue, day by day agreements. The notion of coalition assumes - in order to be meaningful - a modicum of solidity, that is to say, some kind of enduring understanding across a congruent range of issues. Overnight coalitions are not coalitions; and the deals described by Polsby are not coalition-like; they are much more patchwork-like. The point is thus reiterated that with daily, piecemeal collages we do not obtain real majorities. One of the major stakes of the 'reunited government' of President Clinton's initial administration surely is the economic and deficit-reducing package submitted to Congress in August 1993; a package that was approved (after much wheeling and dealing) with a two-vote majority in the House, and a bare one-vote margin (the vote of the vice-president that presides) in the Senate. Does that attest to a majority supported president? Hardly, I would say.
All things considered, the bottom line is, that while common partisan control of executive and legislatures does not assure affective and energetic government, division of party control hampers it. And the basic point is that common partisan control by no means assures a common partisan majority. Regardless, then, of whether the American government happens to be divided or undivided, in either case much of its decision-making requires issue-by-issue localistic side-payments leading to ill-made mosaics rather than to sound compromises. The American system works (in its particular and unique manner) because Americans are determined to make it work. It is as simple as that and, by the same token, that difficult. For Americans do have a constitutional machine made for gridlock, a defect that shows in all its might when their presidentialism is exported to other countries.
5.3  THE LATIN AMERICAN EXPERIENCE
As we leave the United States we naturally land in Latin America, It is here that the bulk of presidential systems reside.
92 Comparative Constitutional Engineering '<-
And it is equally here that presidential systems display a worrisome record of fragility and instability.
In terms of longevity, Costa Rica is the best performer to \t date, for it has remained 'unbroken' since 1949, followed by -Venezuela, continuous since 1958, Colombia (since 1974), and 'f Peru (which returned to civilian government in 1979).   Most -Latin American countries (notably Argentina, Uruguay, Brazil, Chile) have reestablished their presidential democracy only ; in the 1980s.11 And while the last South American bastion of J" old-style dictatorship - Paraguay - fell in 1989, a considerable number of countries in the area remain uncertain democracies ■ and/or polities 'highly vulnerable to breakdown overthrow' > (Diamond et al. 1989, p. xviii): for example, Ecuador, Bolivia, : Honduras, Guatemala and the Dominican Republic;    let ; alone Nicaragua, which reverted in 1990 to democracy under : Sandinista guardianship. In other areas, the Philippines is again, since 1986, presidentially democratic but has yet to prove how well it can perform.13 All in all, then, the record of , the presidentially governed countries range - but for one ■ exception - from poor to dismal and prompts ^us to wonder . whether their political problem might not be presidentialism : itself.
It is always speculative to separate politics from its economic, social and cultural substratum. And in the case of Latin America it must be especially acknowledged that the diflicul-ties of presidential government relate to, and are mightily intensified by, economic stagnation, glaring inequalities, and : socio-cultural legacies. Yet the one handle that we have for ; confronting and solving problems is politics. Granted that , politics can make and often does make matters worse, still it is from 'good politics' that we expect the good things that we seek.
What is wrong, then, with Latin American presidentialisms? The previous analysis of the model, of the US prototype, facilitates the answer. That the presidential systems are strong systems of energetic government has never been quite true for . the US, and is clearly an illusion leading to delusion in the countries that seek inspiration from the Washington model. ' One problem is, of course, that many Latin American pre-
Presidentialism 93
sidentialisms sit upon 'wrong' party systems.14 But the major line of argument, and the one to be pursued first, is that Latin American presidents are by no means as all-powerful as they may appear. Quite to the contrary, 'most Latin American Presidents have had trouble accomplishing their agendas. They have held most of the power for initiating policy but have found it hard to get support for implementing policy' (Mainwaring, 1990, p. 162).
But if this is so, why not seek a remedy by enhancing the power of presidents? This path has in fact been pursued, for most Latin American presidents do have more extensive powers than an American president. In many cases they are given the line-item veto on bills that was denied to the White House until 1995; they are largely permitted, albeit to different extents, to govern by decree; and they are often allowed extensive emergency powers.15 And steps are still being taken to reinforce, presidential power.16 However, the prevalent current mood is, rather, to clip the president's wings,17 for the past, repeated dictatorial seizures of power are perceived — at times rightly, though often wrongly - to result from presidential all-powerfulness. Be that as it may, the problem ultimately resides in the separation of power principle; a separation that keeps the Latin American presidentialisms in perennial, unsteady oscillation between power abuse and power deficiency.
We seemingly have, here, a Gordian knot; a knot for which Juan Linz and, in his wake, a number of important scholars see no solution other than the proverbial one of cutting it. Linz and others have thus reached the conclusion that the remedy is not - in Latin America - in improving presidentialism, but in dismissing it altogether and in adopting in its stead a parliamentary form of government. The Linzian argument was formulated in 1985, and its gist was, and remains, that presidentialism is less likely than parliamentarism to sustain stable democratic regimes (see Linz, 1990, 1994). His central point is not, however, that a presidential structure is conducive to structural stalemate but, rather and more generally, that presidential systems are 'rigid' while parliamentary systems are 'flexible', and that flexibility is to be preferred to rigidity
94 Comparative Constitutional Engineering
Presidenlialism
95
especially because flexibility is risk-minimizing. Thus the \§ Linzian case essentially rests on the point that a flexible,
parliamentary type of polity is far less risk-exposed — on i
account of its self-correcting mechanisms - than a rigid one. %
As Valenzuela felicitously encapsulates it, 'The crises of par- |
liamentary systems are crises of government, not regime' I
(1987, p. 34). The point is well taken. No doubt, presidential |
systems cannot handle major crises.58 Nonetheless it seems to |
me that the argument cannot be left at that. In the first place, |
one of the possible alternatives to presidenlialism is semi- f
presidentialism; and I shall argue that semi-presidentialism |
largely takes care of the rigidity problem, that it affords the |
flexibility that presidentialism lacks [infra, ch. 7). In the second I
place, and above all, the Linzian proposal fails to explain |
convincingly why and how a switch to parliamentarism would | solve the governability problems that presidentialism either
generates or leaves unresolved.19 |
5.4   IS PARLIAMENTARISM A REMEDY?^
Whether parliamentary systems are superior to presidential ? ones is an issue that must await the analysis of parliamentarism [infra ch. 6). At the moment I can address only the transition from the American-like to a European-like system on the basis of this simple admonishment: that parliamentary democracy cannot perform - in any of its many varieties - unless it is served by parliamentary fit parties, that is to say, parties that have been socialized (by failure, duration, and appropriate incentives) into being relatively cohesive and/or disciplined bodies. Mind you, the above is not always true for presidential systems, for which the argument is - we know - that under conditions of divided government stalemate is avoided precisely by party indiscipline. Instead, disciplined parties are a necessary condition for the 'working3 of parliamentary systems. Conversely put, with undisciplined parties parliamentary systems become non-working assembly systems.
The question thus is: if Latin American countries adopted a parliamentary system, would their parliamentary performance .
be any better than the assemblear one of much of continental Europe all the way up to the Twenties and Thirties? I very much doubt it, for it is clearly the case that the bulk of Latin America does not have, and is still far from acquiring, parliamentary fit parties. Brazil eminently speaks to the point; and since Brazil did go as far as to submit to a referendum, in 1993, the option of parliamentarism, let me pause on the Brazilian case.
Probably no country in the world currently is as anti-party, both in theory and in practice, as is Brazil, Politicians relate to their party as a partido de aluguel, as a rental. They frequently change party, vote against the party line, and refuse any kind of party discipline on the ground that their freedom of representing their constituency cannot be interfered with (see Mainwaring 1991). Thus parties are truly volatile entities; and the Brazilian president is therefore left to float over a vacuum, over an unruly and eminently atomized parliament. Can we expect on this soil that a switch to a parliamentary system would bring about party solidification, and this because in the new system parties would be required to sustain a parliamentary-derived government? This was indeed the argument of the propounders of the referendum (which was defeated). But there is no comparative nor historical evidence that goes to support that expectation.
Compared to the Brazilian parties, the German ones during the Weimar period were 'model parties'; yet their fragmentation was never overcome and their parliamentary performance between 1919-33 neither improved nor provided governability. Nothing changed in the behavior or the nature of parties during the Third and Fourth French Republics. The average duration of governments over the forty-year period of the Third Republic (1875-1914) was nine months. And the same would apply to pre-fascist Italy. The point is that party solidification and discipline (in parliamentary voting) has never been a feedback of parliamentary government. If a system is assembly-based, atomized, unruly, magmatic, on its own inertia it will remain as it is. I cannot think of any party system that has evolved into a veritable 'system' made of strong, organization-based mass parties on the basis of internal parliamentary learning. The metamorphosis from an unstruc-
96
Comparative Constitutional Engineering
Presidentialism
97
tured to a structured party system has always been triggered ■ by exogenous assault and contagion. The earlier parties of ' notables and of opinion either perished or changed their ways ? in response to the challenge of externally created (and largely anti-system) mass parties characterized by strong ideological ties and fervor. Now, all the foregoing elements are notably absent in Brazil. Furthermore, the anti-party creed and rheto- ; ric (let alone a typically anti-party electoral legislation) thai r permeates the country make any kind of parliamentary-fit : parties not only unlikely, but altogether inconceivable. The : point is, then, that the current Brazilian political culture and tradition nurtures parliamentary unfit parties. That under such circumstances a parliamentary experience would lead Brazil \ out of chaos into some kind of efficient parliamentary government is, in my opinion, against all odds.
On the other hand, there are, in Latin America, three major countries that might conceivably afford - in terms of their party system - a switch to parliamentarism, namely, Chile and > two-party Argentina and Venezuela. Chile has the most ; European-like multiparty arrangement of the continent. However, Chile also has a past of'polarized pluralism', of strong ; polarization coupled with high party fragmentation. On this background, would it be wise for Chile to adopt a parliamentary system? I doubt it. If Chileans were to decide to abandon their presidential system, they would be well advised, in my ' opinion, to seek a semi-presidential, not a parliamentary solution.
Argentina is instead a two-party presidential system that comes close to enjoying undivided majorities.20 As a purely conjectural question, would Argentina benefit from a parlia-mentary transformation? Again, I doubt it. Argentina's parties are not 'solid' parties. What keeps them together and brings about their coalescence is the presidential system, that is, the overriding importance of winning a non-divisible prize: the presidency.25 I would thus expect a different arrangement to bring about a party fragmentation that Argentina does not : need. All told, then, Venezuela appears (appeared) to be the one South American country that can afford - on the basis of its two strong and disciplined parties - to run the risk of a
parliamentary experiment. But in 1993 Venezuela's two-part-ism was deeply shattered.22 I am thus prompted to conclude without further ado that the variety of parliamentarism that would most probably emerge across most of Latin America would be its assemblear form at its worst.23
That presidential systems perform poorly - as Linz maintains (1990, p. 52) - in countries with deep cleavages and with a fragmented party system, is very true. But could they perform better - these conditions remaining equal - under parliamentary forms? Ceteris paribus, I think not.
NOTES
1. The Irish president truly has very little power, while Austria's and especially Iceland's presidents do have significant power prerogatives which remain, however, dead letters. The reasons for this will be given infra 7.2. Here the point is only that while a direct popular election doubtless establishes an independent legitimacy, this factor alone is not, per se, of consequence.
2. Duverger classifies them as semi-presidentiat. Duverger's is, however, a purely legalistic categorization, as he himself admits: 'the concept of a semi-presidential form of government ... is defined only by the content of the constitution' (1980, p. 166). In due course I shall argue that even this classification is misleading and unacceptable (infra 7.2 and 7.3).
3. Thus to speak on the aforesaid grounds of Chile (between 1891 and 1925) and of Peru as parliamentary-like systems is a misclassification leading to serious misperceptions. A similar point is currently raised with regard to Uruguay, but again without merit (or so I contend below, note 11).
4. For other definitions and a discussion, see Shugart and Carey, 1992, pp. 18-22. My definition omits their fourth criterion, namely, that 'the president has some constitutionally granted lawmaking authority*. For me this is too vague to constitute a criterion, and is already incorporated in my criteria. If a president heads/appoints the executive, it goes without saying that he or she has and must have some lawmaking authority. But no harm follows if, ad abundantiam. this criterion is added to the three that I propose.
Comparative Constitutional Engineering
Presidentialism
99
The countries in question are listed in my table at p. 174. My listing neglects very small countries and the as yet 'fluid polities' (for reasons explained in Sartori, 1976, ch. 8). It is also too soon to include in the analysis the Eastern European and the formerly Soviet countries. And here I leave aside Mexico, a unique case that I pick up later {infra chs. 9 and 11). On Mainwaring's counting (1993, pp. 204-7), since at least 1967 there have been, in the world, 31 continuous democracies; of these, only four {Colombia, Costa Rica, the US, and Venezuela) have presidential systems. On Riggs's counting (1993, p. 219-20), 'Among 76 open polities, 33 were presiden-tialist. . . . Their failure rate was far greater than that of the parliamentary system: 91 percent (30 cases) for the former, by contrast with 31 percent (13 cases) for the latter'. To be sure, these statistics cannot prove much without the backing of a rationale.
This is more precise than 'mutual independence', for the presidency and congress are not reciprocally independent in all respects.
I leave aside how valid Mayhew's findings are. It is notoriously very difficult to appraise qualitatively a legislative output; and Mayhew is also bound to miss the omissions, that is, how many projects are not ever put on the agenda when one knows beforehand that they do not have a shred of a chance. Furthermore, under divided government deals are bound to be struck in one way, and under unified government in another way. In all these respects, and still others, I consider Mayhew's evidence suspect and highly unconvincing. But if Mayhew were right, my argument would be reinforced. The divide is, here, whether a president can negotiate deals with congressional party leaders that can thereafter deliver their votes, versus the extent to which deals have to be negotiated with members of Congress on a one-by-one basis. However, in 1992 and 1993 Venezuela (which stands out, not only in duration but also in terms of economic affluence, as one of the most 'solid' Latin American democracies) witnessed -much to everybody's dismay - two military coup attempts; and Peru has stumbled into some 18 months of unconstitutional rule. To wit, President Fujimori staged in April 1992 an autogolpe, a self-coup, and has governed with emergency powers and military backing until he narrowly won, on October 31, 1993, a referendum for a new constitution (which allows him, inter
alia, to run for re-election) drafted by a Constituent Congress somewhat dubiously elected in November 1992. The end of the story is, however, that Fujimori was triumphantly re-elected in 1995.
11. Uruguay has displayed uncharacteristic constitutional oscillations between 'quasi-presidentialism' (1830, 1934, 1942, 1966) and not (1918, 1952), but I would definitely consider its present-day system (following the 1973 coup and the 1973-84 interruption) a presidential one. That the legislature may censure ministers and that the president is empowered to dissolve the legislature represent a deviation from the United States model but does not contradict my defining criteria of presidentialism (supra 5.1) and hardly affects the substance.
12. Ecuador is in near permanent crisis; Bolivia has undergone, between 1952 and 1982, some 17 military interventions; Hond uras and Guatemala perform messily and largely under de facto military control; and the record of the Dominican Republic is very poor.
13. Mrs. Aquino came to power in 1986 only because 'people power' in the streets of Manila (assisted by US persuasion) forced Marcos to flee. Thus the election of President Ramos in 1992 was the first free, regular transmission of power in 26 years. And while Mrs. Aquino was a symbolic president 'by acclamation', Ramos is an outsider (a former- general) with somewhat uncertain parliamentary support,
14. This matter will be analyzed and discussed infra 11.2, for I have still to establish how presidentialism relates to multiparty situations and/or to party atomization.
15. All these points are covered in detail infra ch. 10.
16. For instance, Chile's constitution of 1989 empowers the president to dissolve the Chamber of Deputies (article 32, sect. 5). Since I have already discussed (supra 5.2) how this 'anomaly' affects the division of power principle, here the point only is that reinforcing the presidential prerogatives is a matter of concern.
17. This mood shows, e.g., in the proposal put forward in Venezuela to submit presidents to popular recall.
18. In this respect the Chilean case with Allende is emblematic. As Mainwaring (1993, p. 208) sums it up, 'in Chile in 1973 opponents of the Popular Unity government feared that by allowing Allende to complete his six-year term (1970-6) they would open the door to authoritarian socialism. Allende had