If. r .0. V . I “M. g . “Mm...“vqnwfiw. . .. “3.....va ...-. u. -anmfmiu. .4“.- .{no "3“...IPfl -vl». . all}. . .15 memmwmww.nmwpu.nu I ”Mb- l' - 1 . . ‘ t... .u...|.<.:W1.M......l (can “1.)“. . . vxmfifiwumfiw . {h .. .. ... .. 1W. flatulw. REW¢ufli . . Ffifi 4m .. :3... . ”h .V A $ 1‘; .fiwflmfiflhfiwfi. 1 vIWvfimgimwmfl .NHH w Jfimhfi. . :- . . £53594...» rit- ...u....4.. b5 . . .. . “fluivxfi. .mswwunwwmufiwhvc . .. ..._....,...x...$ 3.2% 2.....u5Km. til, J. u I d. MANN. wam.hj.$ l a» .2 $.01. _....nm...usu .iwuumq : 4.......... .7 ....-. . , fiat“... Jammuflnvaéwub 1.. N... , 41ifih4cf1.u.l? fit . , Aliffléuwr .._ JAE. .rfilnkn... . .r...» fluvaw. .nuanIwu‘mwasfinw .0 run” u. . «in... a! (v... .3... w». wfiwenmwfl rah... - ...I.v mun»: -I....\.} a. ... lehnfltv I .I u I) t . Indra. 11‘s.! .. u. t ,. ”aka.- .L. 1. 1h?! 1: .. . a" «#5.... Jun ......,m..s..z “DUO... .vl .. J... 9...). :..x . .. $5. dafif . ....q....z_:;,{!... . . 5'. NY 21.!va .. . . . it“? 1:)..th . . 15"); .v . . ‘3‘- .I" . . - .Aal 5.2.5.5”... @233. a-.. .5. .- . ¢ ’33 2.. . 1v .. Jukmmév hunlchffmnso\ . , ....lvs..n......31-.~t. Huuumum , 1.45. wolci Huewunwuvbmrg \ .. .. .. .3 fi%¢n&flh.kifi.fl 2: m»: i... .. {wilgyfiafi that. .. snunfimnfu... ..:.. . Irv-5311‘: .0 I..- {JIA nlt Ina-Ind. . 21:911. lflw%.~w.mh.§§bflfi .. THY; but f 5.....Snrnuvcir jar .1: u..- ..> .9. . IV... . .. vuulii‘t .Vpll. . .rlollv . ‘11.? .. . Illawnvdflolivflahrfb Jahfidtifisflzvctflfl wags—vi ”FRVWR. 1. UR. . II t! 115.95.; n3“ v! fut-{nu aleg. . ”1..-. I \ “11.1.191... I... o . .LDOuInIJOSJu‘v-I » . DIV “$0... , 93-h. .‘Drnxfl. c .Jolru \ (H.155... . 1A.. 1...».‘Hbs.l-.I»\1lrnuta$lpk. . u.-. a .3... C31: .: .2 . .inqhuilhn 553..."...l..».u!.:u$lo....rn‘o.. . 0' Lu ‘PnAQ‘..ooi-I , .101.v.v . z . . . . ,. , § "1 ”3.5). Iu.n-. 4 . - l: , 2.» PTEMWb-flni ,. .u .«1m..».9...fl.mwrr. thuwurrtu . . ..\.«...|..3nl\xyt 5.. v‘a.lx.oL.w.x. fit. «.2. .. 3! .95.):114 i .....y..¥¢..mnv'llu&:.[v§¥ . . . $33.13....I I ... 6 .3 :1. I 1 . r! ii -eln}5..¢r{w§t£.tu:v 49.43%. s . . .. . «137...... . .li . .l. 4.. .. :hMA..4.r\..l!$fczr. E. . H... 4133...- . . . .. 1.11- 1‘93}... £31.13»... . . .. . . l . . . '1. o. B...» .v...»v.-%1fl..yi.ouv.\!F.....ba: 3. v4.1. . ivy... .. . . 11.1.... . .. 4 r. . -bn...ir4.....s: 33...]; ..... . . . -...»lu....u..\n.. . Ylytl ii.)i|i.. . iKllzivth.0I“I il..- in ,i: P... .F L . . n,‘ ;~‘.. “"1 ul 2". ‘. “Lin ,3 w; 1:0 v y... .0“.- .. 9...»... .vv.‘ v‘pxI-nv ‘ \L 31%.: K “3.5.6.15. 1...}. 7‘... .0. I . Vsioy'. ~ [lo 2... 9-x.ll... . .3. .. £224-. .. u 3.1..“ #1. .dn.....1. 0‘33. .. u.‘ .u'cv. ’12 nulluhhlflfivn. ”hush. LI. 5... lbw _ ol1l.|v v.‘uo.c..v.-r n. n\0.tfl.l..l.o . . via.-. lfl...1;1wvfi.n»flmu..ln . .. .. .. . . .. 1.. alth‘ulllt... Euro...» .ltbnondzul . . £3”... .u . . . . . 4 ‘ 'I..IE.C.(8IIO L. . . . . . ‘ -0 \ v.11- . v. N . . . [Ill I' |l‘ ((((( . . . . . .. . . .. staph.“ LINrI. ,, . ........o.yl.n£.t....l Iii... y ill. I . w...» C-1vrl'|$|fl I‘V‘ . l 1- 1 .IIJPV... ‘ .n 10 I” 9| V .let .v . I .7- I u . s 3. D [vlllvl .l-‘ . I‘ll... . .l .. VI... .1. r u 3. V fir. - . . . I31 E b to 1“!“ Y...’ .lle. 1 .. . ... , . ., . vagleH‘cll.-lelst {Pianlaeifi - 2C! . .i .3'...L.‘1..‘\‘.I‘lu ..Pll.0 fly”. ~01I1|M.'u.l\tu'lounl L... A y... n . .o‘oflvavtuLu -r. i‘uvhflviylll. IL1.....:!..§.€\:.$ _. , . . 1 . . tun: lullaervcnu'u. . . t . . . ,. , . I ut‘ .‘rulx‘ . . 4 . p . :- . , . . 4 . .. , . J . 7 . . .. . . 3...... ‘..L$~rI|u|I . v .1. . u f 7, n . ‘ ..l iuL‘. 1., . . .. i J . u ....u...-y .I|.V4 Ann L . . 1.! .VI. .l .. . 3.-.! .I...l- _ . . ‘Q. 33.19.]; c 37‘ ‘ I I « xx... . . .. .. 3;..ufihfiwfl...€. r. .. 1 xi : .. lam... .rznumn....1n..l.1mm.mrvm..r ENE}?! knuummmwbafnfifin . lvl-‘fili-(lbv\l.l!-!"h.flt~|‘x‘0‘d.4‘ . A v.v‘..| A - I!” . unfit-r .n . u 5.5%}, fi..§.¥u....u.vt.nbfl2 i. : rib... . .. .. . .. . . . j m... . - .ivqoh. . . . . . . .. "magnum; EnanailcruWWwoflwa.:lb. .. . - 1 . .. . , . . . . . _. . . .. . “1.20%“... .1. .11.... H.4f.tsrx1fl.rhnr.w.slwflnhwvv.n. . . . .. . .. .. ., . I» v... 1.4.... .9»! .. . ,. ..l.<. ..' (I‘VIHI . , . , :fl\ ‘1“ 0 . 3". lit. rial-u}: .s b . I yo .. .w . #1.. l ‘ ‘ .113! . ...4l.,l§..r¢$...fi..? .. . , . m... Jul. Tnmkmn¥.n{- 'I. .0: a ‘ .ln'v. vvlflvL "alc‘g I .. .K!‘ .....1..\v$. . I}... Kit! I? i... . 3.0.1-.. 1.5.1:... . 0|. 1 2.1.11. Littllthfikhlwflillll filigvou .. . .o‘ . II..I..4.I~‘1JIIOVI IF. ‘10.»! 441.11 . .. .- -Q9|! \ . ,‘ .I . 1: .. . . t .t .2.» 6m .22.! .. .. |.A .. ‘ 1.. . 3. .uv. . .u... i... I" Ill . . , .l. _ r 1:11;»: .. iv 5 .L... .11. .7 u...n L. w THES!S This is to certify that the dissertation entitled : Remeciau 4dr r94 DEeaS‘es ab fee QNMHM n‘tL F54 (cum 6 wigs m’Tvicamc’ (Md ”((1.2% ML», presented by (dash, BHCLH i) )0 m g kf‘cfi. has been accepted towards fulfillment of the requirements for (PRO degree in '1 PL; b (éi filic’u LL/ céz/Ctficl; 3 LUQL (LA—LkL‘ Major professor te 12/, 2/96“ Da MSU is an Affirmative Action/Equal Opportunity Institution ' 0- 12771 RRRRR LIBRARY M'Ghlgan State Universlty PLACE I! RETURN 30X to romovo this chockout from your rocord. To AVOID FINES rotum on or boron dot. duo. DATE DUE DATE DUE DATE DUE MSU Is An Affirmative Action/Equal Opportunity Intuition W REMEDIES FOR THE DISEASES OF FREE GOVERNMENT: BICAMERALISM, TRICAMERALISM, AND HEALTH POLICY by Brian Paul J aniskee A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Political Science 1996 ABSTRACT REMEDIES FOR THE DISEASES OF FREE GOVERNMENT: BICAMERALISM, TRICAMERALISM, AND HEALTH POLICY by Brian Paul J aniskee The problem that inspired this dissertation is the perception in the United States that our policy making system is in gridlock, a term used by civil engineers to describe an enormous traffic jam. The accused culprit for all of this gridlock was often said to be the separation of powers that includes a bicameral legislature. Academic criticism of the separation of powers imbedded in the United States Constitution is not a recent fad. This criticism is an American tradition. Is the separation of powers to blame for gridlock? Recent scholarship on this question is mixed. Accusations about the existence of gridlock have been used as the basis for arguments in favor of rewriting the Constitution. Before we contemplate such a serious step, the existence of gridlock and its causes must be established empirically. Does our system of separated powers lead to gridlock and, if so, what specific aspects of the system contribute to that state of affairs? The primary method of analysis was time series regression. Two national-level and two state-level models are employed in this dissertation; bicameral and tricameral. The results for the national bicameral model were negative. The results for the national tricameral model were more positive but non-intuitive. Gridlock was found in unified and divided governments. The state bicameral model yielded positive results while the state tricameral model yielded negative results. The tricameral model works at the national level because it takes into account official partisanship that did not exist at the time of the Framing and the broader institutional factors that can guide more complex entities. The bicameral model, on the other hand, could work better in Michigan than at the national level because the Michigan legislature with its 38 senators and 110 house members is more like the Congress at the time of the Framing than the national legislature. Copyright by Brian P. Janiskee 1996 To my grandparents: Corrine, Helen, Robert, and William. Acknowledgments This would not have been possible without the support and guidance of my committee: Carol Weissert(Chair), Jim Granato, Tom Hammond, and Dave Rohde. They have shown remarkable patience and sometimes restraint as they have been forced to read numerous revisions of this manuscript. This aspect of graduate education is unsung but no less vital. I owe them a debt of gratitude. Any errors that remain are strictly my responsibility. I would also like to thank others in the political science department at Michigan State University who have had a positive impact on my education: Cleo Cherryholmes, Scott Gates, Paula Kearns, Arthur Melzer, Ira Polley, and Ken Williams. Whether it was a lively debate in class or an occasional chat in the hallway, they are part of what make MSU such a wonderful place. I would also like to thank my mentors at James Madison College for their fellowship and support: Dean William Allen, Doug Hoekstra, Connie Hunt, Dick Zinman. I would also like to send a very special thanks to the administrative vi staff in the department of political science. I can say without any exaggeration that without them I would not have finished. They are very special people and I will miss them: Karen Battin, Rhonda Burns, Iris Dunn, Elaine Eschtruth. Finally, I would like to thank my lovely wife Jennifer and our beloved daughter Katherine. I have spent many long nights away from them as this project unfolded. They have been very patient with me. I owe them most of all. vii TABLE OF CONTENTS LIST OF APPENDICES ................................................................................. x CHAPTER 1 INTRODUCTION ............................................................................................. 1 Identification of the Problem and Its Importance ................................ 1 Research Questions, Hypotheses, and Design .................................... 14 Overview of Dissertation ..................................................................... 17 CHAPTER 2 NATIONAL BICAMERAL MODEL .............................................................. 20 Importance of Bicameralism and Literature Review ........................ 20 Model .................................................................................................... 27 Summary and Conclusion ................................................................... 43 CHAPTER 3 NATIONAL TRICAMERAL MODEL ............................................................ 46 Importance of Tricameralism and Literature Review ....................... 46 Model .................................................................................................... 52 Summary and Conclusion .................................................................... 66 CHAPTER 4 CASE STUDIES ............................................................................................. 69 Medicare ................................................................................................ 70 Case Study Variables ........................................................................... 80 Health Security Act .............................................................................. 85 Case Study Variables ............................................................................ 97 Comparative Analysis of Factors ......................................................... 101 Conclusion ............................................................................................. 107 CHAPTER 5 STATE BICAMERAL AND TRICAMERAL MODELS ................................. 108 Background .......................................................................................... 108 Bicameral Model .................................................................................. 114 Analysis ................................................................................................ 119 viii Tricameral Model ................................................................................ 12 1 Analysis ............................................................................................... 122 Conclusion ........................................................................................... 123 Summary ............................................................................................. 125 CHAPTER 6 CONCLUSION ............................................................................................... 127 ix LIST OF APPENDICES Appendix 2-1 Banzhaf Voting Power ...................................................... 140 Appendix 2-2 Relative Banzhaf Power .................................................. 140 Appendix 2-3 Alternate Regression Models ........................................... 141 Appendix 2-4 National Relative Change ............................................... 142 Appendix 25 List of Bills ........................................................................ 143 Appendix 2-6 National Occupation ......................................................... 147 Appendix 2-7 National Age ...................................................................... 148 Appendix 2-8 National Retention ............................................................ 149 Appendix 2-9 National Seniority ............................................................. 150 Appendix 2-10 National Bicameral Model .............................................. 151 Appendix 2-11 National Bicameral Diagnostics ..................................... 152 Appendix 3-1 Unified Government .......................................................... 153 Appendix 3-2 National Seat Change ....................................................... 154 Appendix 3-3 Regression Results Using Separate Measures ................. 155 Appendix 3-4 Presidential Popularity ...................................................... 156 Appendix 3-5 Regression Results Using Unrestricted Landslide ........... 157 Appendix 3-6 Interaction of National Party Strength with Unified ....... 158 Appendix 3-7 National Tricameral Model ................................................. 159 Appendix 3-8 National Tricameral Diagnostics ....................................... 160 Appendix 3-9 Modified National Tricameral Model ................................. 161 Appendix 3-10 Modified National Tricameral Diagnostics ...................... 162 Appendix 3-11 Regression Results with GDP .......................................... 163 Appendix 4-1 Four Phases of Medicare ..................................................... 164 Appendix 4-2 Conferees 1965 ..................................................................... 165 Appendix 4-3 Veto Threat .......................................................................... 166 Appendix 4-4 Phases of Health Security Act ............................................. 167 Appendix 4-5 Ideological Distribution ....................................................... 168 Appendix 4-6 Clinton’s Veto Threat ........................................................... 169 Appendix 4-7 Comparison of Variables ...................................................... 170 Appendix 5-1 Michigan Government .......................................................... 171 Appendix 5-2 Michigan Governors ............................................................ 172 Appendix 5-3 Michigan Occupation ............................................................. 173 Appendix 5-4 Michigan Age ......................................................................... 174 X Appendix 5-5 Michigan Retention ............................................................... 175 Appendix 5-6 Michigan Seniority ................................................................ 176 Appendix 5-7 Sample Year ........................................................................... 177 Appendix 5-8 Michigan Bills ........................................................................ 178 Appendix 5-9 Michigan Relative Change ..................................................... 181 Appendix 5-10 State Bicameral Model ......................................................... 182 Appendix 5-11 State Bicameral Diagnostics ................................................ 183 Appendix 5-12 Michigan Seat Change .......................................................... 184 Appendix 5-13 Michigan Gubernatorial Popular Vote ................................. 185 Appendix 5-14 Michigan Party Strength ...................................................... 186 Appendix 5-15 State Tricameral Model ........................................................ 187 Appendix 5-16 State Tricameral Diagnostics ............................................... 188 xi CHAPTER ONE: INTRODUCTION IDENTIFICATION OF THE PROBLEM AND ITS IMPORTANCE Perception of Gridlock The problem that inspired this dissertation is the perception in the United States that our policy making system is in gridlock, a term used by civil engineers to describe an enormous traffic jam. A Nexis databank search of articles mentioning the words "Congress" and "gridlock" from January 1993 to June 1994 found 558 articles. That is a mean figure of almost 35 articles per month (Time, 1994). In a dire prediction, Kevin Phillips warned us that voter anger will be very high in 1996 because the 1994 midterm elections will enhance Congressional gridlock (1994). Accordingly, discussion about the failure of the House, Senate, and president to agree on key issues in 1996 began before the year was even a month old (Dunham 1996). The New York Times proclaimed "gridlock everywhere" (Pear et a1. 1996, 12). The accused for all of this gridlock was often said to be the separation of powers which includes a bicameral legislature. The clash between the 1995 class of Republicans in the Congress and the checks and balances system was described as an "irresistible revolution" meeting an "unmovable constitution," (Rosenbaum 1995, C17). This concern about an "unmovable constitution," however, has not been limited to the popular press. Academic criticism of the separation of powers imbedded in the United States Constitution is not a recent fad. This criticism is an American 2 tradition. The separation of powers became controversial with the coming of progressivism which the scientific management movement later accompanied. The scientific management movement as applied to government became known as "public administration" and it is from this body of scholarship that we find the harshest criticism of the separation of powers and its seemingly concomitant gridlock. As Waldo points out, "The traditional doctrine of the separation of powers became the bete noir of American political science, and the exaltation of the powers of the executive branch its Great White Hope" ( 1948, 36). The discipline of public administration was itself created to undermine the separation of powers system in favor of a more efficient method of government (Marini 1992, Sundquist 1974). Frank Goodnow argued that the "principle of the separation of powers and authorities has proven...to be unworkable as a legal principle" (1900, 15). It was the widely- held view that the societal crisis of the turn of the century made energetic government a necessity and that the "separation of powers made concerted action impossible" (Vile 1969, 266). Woodrow Wilson boldly proclaimed, "It is quite safe to say that were it possible to call together again the members of that wonderful Convention to View the work of their hands in the light of the century that tested it, they would be the first to admit that the only fruit of dividing power had been to make it irresponsible" (1885, 284). Two years after writing these words he went on to write the article that laid the 3 groundwork for the science of public administration in the United States. Herbert Croly described the separation of powers as "a bed of liquid clay...an indiscriminate mass of sticky matter, which merely clogged the movements of every living body entangled in its midst" (1915, 225). More recently, Lloyd Cutler lamented that the separation of powers "almost guarantees stalemate today" (1986,2) and James Sundquist concurs that the proper response "is not to preserve the separation of powers but to overcome it" (1986, 138). Nevertheless, is the separation of powers really that "sticky?” Recent scholarship on this question is mixed. David Mayhew's seminal work reveals that the perception of policy gridlock in the popular press and in academic scholarship is incorrect (1991). Mayhew finds that they pass major policies at the same rate in both divided and unified governments. Alt and Lowry, in a study of state executive budgeting decisions, do find significant evidence that gridlock can occur and be made worse by partisan differences between the separated branches (1994). A third approach holds that gridlock can occur in divided or unified governments (Krehbiel 1993, Rohde 1991, Shepsle 1983). They show that any given equilibrium can yield many policy outcomes. Depending upon the conditions, a divided government could be more productive than a unified government and vice versa. Thus, it appears that a major disagreement over a key issue in American political theory is here. Accusations about the existence of gridlock have been used as the basis for arguments in favor of rewriting the 4 Constitution (Sundquist 1992). Before we contemplate such a serious step, the existence of gridlock and its causes must be established empirically. Does our system of separated powers lead to gridlock and, if so, what specific aspects of the system contribute to that state of affairs? Whether the goal is to attack or defend the separation of powers as framed in the 1787 Convention, one must understand the motives and intentions of the Framers. In the words of Martin Diamond, "Whether one considers that political order to be a blessing or a curse to mankind, it cannot be understood (and therefore, no national self-examination can be truly complete) without a thorough study of the documents of our founding" (Diamond 1981, v). It is this spirit that animates this dissertation and the goal is to provide one small piece in that quest for understanding. Framers The analysis in this dissertation will rely heavily on the Federalist Papers in which Publius gives us an explicit account of how the Framers intended the government to operate.1 It should not be dismissed out of hand 1 Alexander Hamilton, James Madison, and John Jay, when citing quotes from the Federalist Papers, will be referred to by their collective pseudonym "Publius." Besides the convenience of the single moniker, the use of the collective pseudonym expresses a belief in the theoretical proposition that there is a unity in the teaching of the triumvirate (Diamond 1987). The pseudonym used by the triumvirate was in the style of the times. Public officials who wanted to express their opinions in the newspapers would use a pseudonym, usually one from the days of the Roman Republic. The anonymous author would pick a pseudonym whose life history 5 that the desire to create a sound governmental structure guided the Framers though they were partisan combatants in a political fight. Policymakers can have goals of good public policy above the immediate goals of reelection and personal benefit (Fenno 1978). In fact, it might not be out of the question that they were partisans of particular policies because they thought those policies to be reasonable in the light of empirical observation. As McCubbins and Sullivan (1987, 2) put it, "The Federalist Papers, of course, can be read as an elaborate theoretical defense of the Constitutional Convention's structural selections." At the time of the Convention of 1787, every free government that had come into being had disintegrated. Madison studied the histories of previous free governments and noticed a pattern that might account for the tragic state of free government in the world. From the history of Greece, Madison learned that class warfare leads to dictatorship and a lack of union among the poleis made Greece ripe for conquest. This internal instability and corresponded with the authors' position on the current events. Publius Valerius, for example, was instrumental in eliminating the last Roman king and became one of the founders of the Roman Republic. He enacted a series of reforms once he became the chief executive of the Republic. These reforms included: 1) a defendant's right of appeal; 2) lower taxes; 3) the cultivation of a commercial as opposed to an imperial economy; 4) having the lictor's rod in the Assembly lowered toward the people and not the officers so that all could marvel in the "majesty of democracy" (Kessler 1988, 76). Publius eventually became honored with the name , "Publius Publicola" which means, roughly, cherisher of the people." 6 susceptibility to foreign attack allowed Greece to be conquered by Phillip of Macedon, the father of Alexander the Great. The first great European empire, the very antithesis of a republic, came out of the ashes of Europe's first great democracy. The decline of the Roman Republic can be traced to the bloodlust that boiled in its citizens as Rome began to conquer the Mediterranean world. As the Founders were well aware, a constant warlike footing is inimical to a republic. The lesson learned was that the main threat to a republic is instability. This condition would lead to a general feeling of ill-ease which, as in Greece with Pisistratus and Rome with Julius Caesar, a dictator could only solve. Publius desperately wanted to avoid this fate. The remedies for the disease of instability that could affect democracy in America would be Republic and Union (Diamond 1987). Concerning the Union, being united to avoid the petty squabbling and the rampant discord of the patchwork Greek states was important for the states. Shays Rebellion was the first indication that America would not be immune to the diseases that faced all free governments. Concerning the Republic, the system must be one of separated powers to limit an arbitrariness in governmental rule, whether by the masses or an autocrat, which led to the destruction of the Roman Republic and the creation of the military dictatorship known as the Roman Empire. Thus, the centralized power of the federal government would accomplish Union and that centralized power would be made republican 7 through the separation of powers and its key component, bicameralism. This is what Publius meant by "a republican remedy for the diseases most incident to republican government" (Publius [1787] 1961, 84). The strength of conviction in which Publius, and the anti-Federalists for that matter, believed that the separation of powers was indispensable for the protection of liberty may be brought to light through the following quotation: The accumulation of all powers, legislative, executive, and judiciary, in the same hands, whether of one, a few, or many, and whether hereditary, self-appointed, or elective, may justly be pronounced the very definition of tyranny(Publius [1787] 1961, 301) This echoes the thoughts of Jefferson whom Publius quotes in Federalist 48, "The concentration in the same hands is precisely the definition of despotic government" ([1787] 1961, 310-311). The argument about the disease being undemocratic has been made but how truly "republican" is the "remedy?” The argument, made by those public administration scholars and others above, that the Constitution, including the separation of powers, is undemocratic is not borne out by analysis of the Federalist Papers. The intentions of Publius, and for that matter, Jefferson, were to cure the ills of democracy so that democracy could survive. In the words of Harvey Mansfield J r., "The Federalist after all is preoccupied with the diseases of republicanism because it has chosen republicanism"(1989, 254). Furthermore, the separation of powers is not at tension with the ideas of 8 democracy expressed in the Declaration of Independence.2 The separation of powers is the logical result of the theory of human behavior that animates the Declaration. The key theoretical feature of the American regime is the claim that all human beings have an equal nature. The Framers did not claim we are all equally talented in all things, rather than we all have a limited level of rationality and longevity. No one is so talented that he or she could rule over other human beings as if that person was not limited: “The general spread of the light of science has already laid open to every view the palpable truth, that the mass of mankind had not been born with saddles on their backs, nor a favored few booted and spurred, ready to ride them legitimately by the grace of God" (Jefferson 1826 in Peterson 1984, 1517)3 2 This, however, may not have been enough for Woodrow Wilson who, in the following quote, goes beyond the Constitution to attack the Declaration itself, "Liberty rooted in an unalterable law would be no liberty at all" (in Link 1966, 18: 71). Wilson failed to recognize that human freedom must be rooted in an unalterable law. If this is not the case, the argument for human freedom is left to the mercy of the fashion of a particular period in history. The argument then follows that the case for tyranny and the case for freedom are on equal terms. The argument for freedom, if Wilson is right, is not based on anything that amounts to more than a whim or an urge. According to this theory, all urges are created equal and this includes an urge to do good or evil. 3 Jefferson knew very well the extent of his hypocrisy by being the holder of the pen that transcribed the Declaration and the holder of slaves, "What a stupendous, what an incomprehensible machine is man! who can endure toil, famine, stripes, imprisonment, and death itself, in vindication of his own liberty, and, the next moment, he deaf to all those motives whose power supported him through his trial, and inflict on his fellow man a bondage, one hour of which is fraught with more misery than ages of that which he rose in rebellion to oppose" (1786, in J affa 1984, 15) His hypocrisy, however, does Since the institutions of government are a product of reason, a limited human reason, we can only expect government to do limited tasks in a limited fashion. A modern manifestation of this argument is Herbert Simon's idea of "bounded rationality" (1976, 38). Remembering that both Simon and Publius argue that those who govern are bounded is important as well. It would be, so the argument goes, irrational to attempt to solve the problems caused by the limits of human nature by giving a group of human beings, who are themselves limited, unlimited power to solve the problems. Consider the following passages from Federalist 51 : In framing a government that is to be instituted by men over men, the great difficulty lies in this: you must first enable the government to control the governed; and in the next place oblige it to control itself (Publius 1787, 322) But what is government but the greatest of all reflections on human nature? If men were angels, no government would be necessary. If angels were to govern men, neither external nor internal controls on government would be necessary (Publius 1787, 322) Considering the potential cures for the diseases of democracy, most are not democratic. The fact that the Framers attempted the application of the separation of powers doctrine in a way completely untested is an indication of their political and intellectual courage. The claim that democracy was not render the principles of the Declaration false anymore than if a person who had previously denounced murder on one day and then commits murder on the next invalidates the principle that murder is evil. 10 consistent with reason and good government was, then, a very radical claim. The Framers took a leap of faith into uncharted territory. The fact that the reasonableness of free government is accepted without controversy today is a testament to the power of Publius's argument. In fairness to the critics of the separation of powers mentioned above, their motivation is understandable. Strict Construction was the dominant legal theory on the interpretation of the Constitution at the time and the separation of powers was used as an excuse to deny the implementation of child labor laws, monopoly regulation, and other needed reforms. Instead of attacking the interpretation as wrong, the Progressives accepted the Strict Construction approach themselves and attacked the Constitution directly. If the Framers are correct about our rationality being bounded, then they cannot be blamed for constructing the system as such, however imperfect. If, on the other hand, human beings are not as limited as claimed by the Framers then they may have been the worst despots that have ever existed because their system has kept the government from fulfilling its logical role as the chief organizer of society. This is not a matter to be decided out of hand or in the midst of anger over a failed piece of legislation. Despite the final answer their plan is, at the very least, valid. The logic of Publius is as consistent as human logic can be. Whether it is true or not is a matter for another time. The question here, however, is not whether establishing the separation of powers was right but to detect some effects 11 since its establishment. Scope of the Project The attempt to analyze such a complex phenomenon as the separation of powers over an extended period may appear to some as quite brazen. It is because of the enormousness of such a task that the analyst is compelled to maintain the utmost humility. The subject, however, is too important to be avoided simply because it is difficult. Arguments for and against the separation of powers are made as a result of broad perceptions of the system's performance and, therefore, a broad empirical analysis is in order so that we might evaluate these claims on their own terms. Furthermore, a body of literature devoted to the study of the policymaking system on such a broad level has already been developed (Alt and Lowry 1994, Cox and Kernell 1991, Fiorina 1992, Freedman 1995, Krehbiel 1993, Marini 1992, Mayhew 1991, Peterson 1990). Save the examples of Alt and Lowry (1994), Freedman (1995), and Mayhew (1991) who do study the policy effects of the system, these studies are devoted to studying the causes of divided government or whether the president or Congress is more powerful in the legislative process. Neither of these questions will be addressed in this dissertation. The goal of this dissertation is to offer an overall evaluation of the effects of the separation of powers on the production of policies. The separation of powers is more than the product of the checks and 12 balances system between the Executive and Legislative branches, not to mention the Judicial branch. The separation also occurs within the branches themselves. This is especially true for the Legislature which is divided into separate chambers. The separation of powers that occur within the Legislative branch is known as the bicameral system (upper chamber, lower chamber). The separation that occurs between the Legislative and the Executive branch is known as the tricameral system (upper chamber, lower chamber, executive). Within political science the literature devoted to the study of bicameral systems is, for the most part, a discrete entity from the literature devoted to the tricameral system. Thus, the literature for the two systems will be reviewed separately. A summary of each of the remaining chapters can be found at the end of this introductory chapter. Why Health Legislation? Other studies that have analyzed the broad performance of government have focused on areas such as banking, energy, and defense. Analyses of health policy production from this perspective have been less common. Health legislation is the context for this study because it is currently one of the most important areas of domestic policy and, besides its importance now, it has been prominent in the entire period since the Second World War. Health policy spans the jurisdictions of the major committees in Congress. In the House these include: Appropriations, Ways and Means, and Commerce. In the Senate these include: Appropriations, Finance, Labor 13 and Human Resources. In addition, all of the presidents in the postwar period have made the expansion or reduction of public financing and/or provision of health care a major part of their legislative agenda (W eissert and Weissert 1996). Matters of health policy have consistently aroused partisan debate. The question of the proper role of the government in financing and provision of health services goes to the heart of questions concerning what we were intended to be as a nation by those who constructed the system. The different understandings of this system come to the surface very clearly when the issue of health policy is debated the floors of the House and Senate or from the Oval Office (W eissert and Weissert 1996). The reason for the controversial nature of health policy is that the question of health is such a vital component of our everyday lives. Its importance is self-evident. Its inherent complexity as a social phenomenon, however, renders easy answers unobtainable. It is something we must address, knowing that the solution to any problem is bound to create divisions. To borrow the terminology from James Morone (1990), we "yearn" to solve the problems of health care in the United States but "dread" the debate and strife that will surely follow. Thus, health care is a good policy area in which to continue the empirical investigation of the phenomenon of gridlock. 14 RESEARCH QUESTIONS, HYPOTHESES, and DESIGN The research questions being addressed in this dissertation are the following: When will changes in health policy occur and what factors in the bicameral and tricameral system make change more or less likely? The general theory of the separation of powers with respect to policy change is based on the idea that the more different the branches are in terms of characteristics the more stable the system will be. Under a stable policy system, proposals are less likely to become law. To prevent the separated branches of a government from coalescing together, the institutions needed to be structured so that they would have inherent differences. The separated systems of Greece, Rome, and for that matter, England, were all class-based regimes. The separation of powers in a class-based regime is maintained through the official recognition of hereditary rights (Montesquieu [17 48] 1973). The usual pattern was for the top noble to become the executive (monarch), lower nobles would become the members of the upper legislative chamber (Senate, House of Lords), and the propertied commoners would send representatives to the lower legislative chamber (Assemblies, House of Commons). The case is clear that any stability possessed by these past republics was due to the separated system and the official recognition of class was the factor that kept the branches separate (Diamond 1987). The Framers faced a dilemma. The separation of powers was the key to 15 their entire defense of free government. This system, however, seemed to work only if the government officially recognized the idea of hereditary class. To save their new republic they might be forced to accept hereditary nobility which, ironically, was the core principle of the monarchy they had all risked their lives to depose. The answer came in a uniquely American institution, a popular separation of powers. The differences between the branches and between the two chambers of the legislature achieved in the class-based republics of the past was approximated by the Framers through the differences in such things as age requirement and length of terms. These structural differences would manifest themselves in the compositional differences in various characteristics of the House, Senate, and the Chief Executive. These differences would allow the two bodies to be separate enough to fulfill the stability role that in republics past had been accomplished by the antagonism between different classes. Through the invention of a popular separation of powers that relied on the differences of the characteristics of a group of individuals instead of an officially recognized class of people, America became the first wholly representative and wholly popular regime (Diamond 1987, 667). This is not to mention the federal character of the regime. Besides an analysis of the factors in a bicameral or tricameral system that allow for change to occur, another unique feature of this dissertation is the direct comparison of the performance of the separation of powers on the 16 national level with that of the state level. With Fiorina's (1992) study the notable exception, studies that explicitly compare the policymaking systems of the national and state governments are rare. The states provide an opportunity to assess the external validity of constructs normally applied only to the national level. It is the hope that lessons learned in the application of predominantly national theories to state level analysis will help empirical investigations at both levels. The units of analyses in this dissertation are individual pieces of legislation grouped according to an entire two-session Legislature. The temporal component, therefore, is time-series as opposed to a cross-section. The concern with each piece of legislation in the study sample is whether or not it became a law. As Mayhew points out, the production of actual policy is the most important factor in understanding the performance of a policymaking system (1991). The operational definition of the dependent variables, one national and one state, will be discussed in more detail in Chapters Two and Four respectively. The variable will be a ratio level variable making the application of a classical linear regression model the most prudent option. Four empirical models will be used: 1) National Bicameral 2)National Tricameral 3) State Bicameral 4) State Tricameral. The standard diagnostic tests will be applied to help assess the internal validity of each model. The longitudinal examination of the separation of powers will be l7 supplemented by analyses of two case studies: Medicare of 1965 and The National Health Security Act of 1994. Case studies are useful in that they help the analyst to develop an understanding of the subtleties of a phenomenon which are important but not amenable to a broad empirical definition over a large time series. An example of this would be the antagonism between Rep. Wilbur Mills (D-AR) and President Lyndon B. Johnson over Medicare. This factor and factors like it are impossible to capture in a time series regression model yet they were crucial factors in determining the eventual fate of the policy proposal. Thus, it is important, where possible and appropriate, to include both the "micropolitical" and "macropolitical" forces in an analysis of a policymaking system. The comparison of a bicameral and a tricameral model within each level of government, the comparison of the separation of powers between levels of government, and the inclusion of case studies make this study unique in its comprehensive attempt at understanding our nation's policymaking system. OVERVIEW OF DISSERTATION This dissertation is organized in the following manner. In Chapter Two, the separation of powers is analyzed according to a national bicameral model that focuses on the differences between the House and the Senate as explanatory factors concerning policy change. A literature review of the work on bicameralism reveals that the dominant mode of analysis is a comparison of the activities between the two chambers without much attention to the 18 policy outcomes affected by these differences. The existing studies which attempt to analyze the policy outcomes of such a system do so only at the level of formal analysis, thus an initial empirical approach is needed. The variables in this model are derived from the arguments expressed in The Federalist Nos. 52-66 which contain most of the discussions concerning the Congress. In Chapter Three, the separation of powers is analyzed according to a national tricameral model. The literature review in this chapter will reveal that most of the previous tricameral studies attempt to determine which branch, Congress or the President, is the stronger. There are only two major studies which attempt to analyze the effects of the differences between the branches in light of the policy outcomes (Alt and Lowry 1994, Mayhew 1991). This study will highlight certain areas not specifically addressed by these studies. The attributes of each branch which are included in the analysis are culled from the arguments expressed in The Federalist Nos. 4 7-51 and a survey of current research on Congress and the presidency. Chapter Four will consist of an examination of two case studies: Medicare of 1965 and the National Health Security Act of 1994 ( NHSA). One was eventually passed into law (Medicare) and the other was defeated (NHSA). This chapter will serve to give a fuller picture of the tricameralism. The purpose of Chapter Five is to make an explicit national/state comparison in order to see which model works better at different levels of 19 government. Lessons learned from this comparison can be helpful in determining the external validity of national models in the fashion of Fiorina (1992) and also to help us to better anticipate the future performance of federalism. Michigan was selected as the context in which the state study will take place. Michigan is particularly important because, compared to the other states, Michigan's legislature is highly professional (Squire 1992). The legislators are well-paid and work in the legislature full-time and have access to a wide variety of resources including significant staff support and the governor has the veto power as well. The separation of powers in Michigan is, therefore, very similar in character to the separation at the national level. If differences appear, given this similarity, we should expect differences of a greater magnitude to appear in less professional state governments. The final chapter will summarize the findings of the national, subnational, and case study analyses and highlight their significance. CHAPTER TWO: NATIONAL BICAMERALISM IMPORTANCE OF BICAMERALISM AND LITERATURE REVIEW In this chapter, the separation of powers is analyzed according to a national bicameral model that focuses on the differences between the House and Senate as explanatory factors concerning policy change. Most of the research on bicameralism to date has taken the form of describing the differences between the two bodies. Examples of this kind of research include Abramowitz (1978), Baker (1980, 1989), Bernstein and Berkaman, Brown (1922), Carmines and Dodd (1985), Chappell and Suzuki (1993), Clark (1965), Copeland (1989), Fenno (1966, 1982), Michael Foley (1980), Grofman, Griffin, and Glazer (1991), Jacobson (1987, 1990), Jewel and Patterson (1986), Kozak and Macartney (1987), Kuklinski and West (1981), Mansbridge (1983), Mayhew (1991), Oleszek (1984), Ornstein (1981), Pressman (1966), Reedy (1986), Storm and Rundquist (1977), Waterman (1990), Weaver (1972), Woodrow Wilson (1885). What is missing is an analysis of the impact of bicameralism on the legislative process. It is important that this impact be studied given the central role of bicameralism in American government. The institution of legislative bicameralism was the result of Publius's "science of federal government"(Publius [1787] 1961, 129). At the Constitutional Convention at Philadelphia in the Summer of 17 87, the debates over bicameralism were fierce and threatened to undermine the 20 21 adoption of the new Constitution. The antagonists were concerned about the differences between the two chambers because those institutional differences would have consequences on policymaking. The discussion of bicameralism in The Federalist Papers is considered up in numbers 52-66. Fourteen percent of the available space for discussion in the Federalist is devoted to this concern. While there is no sustained empirical treatment of bicameralism as it affects the policy process, there is a body of literature which deals with the formal characteristics of bicameralism as an institution and the effect it has on other institutions in a formal constitutional system. This body of literature takes what has come to be known as a Public Choice perspective. It is the general position of the authors in this body of literature that the structure of the institutions has a significant effect on behavior. Any analysis of policy, therefore, must take into account the political institutions that structure the mode in which that policy is created, debated, ratified, or rejected. One perspective holds that the stark differences in terms of the size of the body should have an effect on the relative power of each branch to affect policy. Steven J. Brams (1989) argues that Publius was correct in assuming in Federalist 52-66 that the House would be the more powerful of the two legislative bodies. Brams applies the Banzhaf Power Index to the bicameral system (Banzhaf 1965). The Banzhaf Index calculates the power of a member 22 of a voting body based on the assumption that the power of a chamber is the number of its critical defections in minimal winning coalitions as a proportion of all the critical defections of the other members combined. According to Brams, an individual Senator is twice as powerful as an individual Representative, but the House as a whole is twice as powerful as the Senate(Appendix 2-1). In addition, Brams's empirical analysis shows that the House is more influential in its legislative activities than the Senate. Brams examined what percentage of bills that start in one body of the legislature get passed by the other (Appendix 2-2). Brams recognizes, however, that the legislation used in the empirical analysis is highly aggregated and that a narrower empirical treatment is necessary due to the fact that all revenue bills are mandated by the Constitution to begin in the House and all appropriations bills begin in the House by tradition. This, according to Brams, might overstate the influence of the House. More specific policy areas need to be analyzed to gain more insight into this question. Mark Petracca (1989) argues that the attempt made by some, especially Brams, to portray Publius as a Public Choice theorist is misleading. Petracca argues that Brams takes the argument made by Publius out of context: With regard to the claims made by Brams, what are the focal points of Madison's and Hamilton's responses? In general, they are reacting to the numerous arguments by anti-Federalists that the Senate, as 23 proposed, was too powerful and undemocratic (Petracca 1989, 159- 160). Petracca is also critical of Brams's application of the Banzhaf Power Index in this context. Petracca argues that this index is reductionist and had little to do with the reality of the complex institutional relationship between Congress, the President, and the Supreme Court. Finally, Petracca agrees with Brams by stating that the results were obtained from highly aggregated legislation and that a more refined analysis must be conducted in order to ascertain a more complete picture of the relative power of each chamber of Congress. Daniel Wirls (1993) offers, to date, the only empirical analysis of bicameralism and its effect on the policymaking process. Wirls is concerned with whether or not the theory of bicameralism "corresponds with political reality" (1993, 2). Wirls examines one particular aspect of Publius's theory of bicameralism, the fostering of greater deliberation which "prevents the passage of hastily considered and poorly conceived legislation" (1993,2). Wirls notes the paucity of analysis regarding the policy effects of bicameralism in the Congressional literature: "the bulk of the material on bicameralism here and abroad is descriptive --topologies of different forms of bicameralism and discussions of House-Senate differences" (1993, 3). Wirls places the elements of American bicameralism into three groups: "the physical, the compositional, and the institutional" (1993, 6). Wirls does point 24 out, however, that since the 17th Amendment passed in 1913, the two houses have begun to converge, "despite staggered election, turnover in both houses is about equal"(1993, 10). Because there is an evolutionary component to bicameralism, Wirls argues that "any evaluation of bicameralism's effects is an analysis of the effect of many factors, some of which have changed over time. Wirls's data consists of important pieces of legislation in the policy areas of national security and civil rights legislation from 1946-1990. The dependent variable was whether or not the bill was passed by the House, Senate, or both chambers. Wirls determines which chamber is the "cooler" by determining which body stopped foolhardy legislation and which chamber was the first to adopt needed reform. Wirls finds no support that any one branch is cooler or more deliberative than the other. Wirls, however, is forced to engage in a normative judgement in order to determine which legislation worthy of passage and which is not. None of the above deals with an important aspect of bicameralism, its effect on stability. Publius was concerned with stability in the making of laws for two reasons; one external and the other internal (Hammond and Miller 1987, Miller and Hammond 1989,1990). Hammond and Miller define instability as a state of constantly changing policy. Instability manifests itself externally by the lack of respect which other nations will hold for a nation that is prone to volatile policy. The internal effects of unstable laws would be to make the 25 laws so complex that they could be used by those well-versed to take advantage of those who cannot keep up with all of their complexities: Law is defined to be a rule of action; but how can that be which, is little known, and less fixed? Another effect of public instability is the unreasonable advantage it gives to the sagacious, the enterprising, and the moneyed few over the industrious and uninformed mass of the people. Every new regulation concerning commerce or revenue, or in any manner affecting the value of different species (types) of property, presents a new harvest to those who detect the change, and can trace its consequences; a harvest, reared not by themselves, but by the toils and cares of the great body of their fellow citizens. This is a state of things in which it may be said with some truth that laws are made for the few, not for the many (Publius [1787] 1961, 381 emphasis in original). The concern which Hammond and Miller address is derived from the problems brought to light by Condorcet (1785). Condorcet showed that for any choice made by a majority of voters there is some other option that is preferred by a different majority. To counteract this volatility, Condorcet and Publius saw the need for moderating institutions like the separation of powers. If there is a set of legislative alternatives that can gain the support of a majority and at the same time are preferable to any other set of potential policies, then there is said to be a core of undominated alternatives. Unicameral legislatures usually lack a core of undominated alternatives. Bicameralism allows for the creation of a core (Hammond and Miller 1987). The literature cited above focuses on a variety of questions. Brams ( 1989) is concerned with determining which chamber is more powerful. Petracca (1989) argues that an analysis of bicameralism can suffer from an 26 approach that is too minute and formal. Wirls (1993) attempts to determine which branch is the more deliberative and responsible. The question I ask is more basic: What effect do the differences between the House and Senate that animate the institution of bicameralism have on the policymaking process? The previous work provides some initial clues but an empirical examination must be attempted to enter the fray of policy relevant research. There are some areas of controversy regarding empirical work in this area. Hammond, Miller, and Kile (1996) point out that statistical models in this area are unlikely to capture the full effect of bicameralism. This is due to the inability to properly control for all possible variables. Nothing in this dissertation is intended to contradict this. The models are exploratory and, therefore, tentative, at this point. The reason for this style of approach is that generalizations about the policy production of American government are made and serious adjustments have been pondered based on broad measures of policy output (Morone 1990, Sundquist 1992, Wilson 1885). Recent studies in this area have focused on such general measures because this is how the system is judged (Fiorina 1992, Freedman 1995, Mayhew 1991). This study is a serious attempt to deal with these arguments on their own terms, however imperfect they may be. 27 MODEL Dependent Variable The time period of both the national bicameral and tricameral studies is 1949-1994. 1949 is the earliest year in which information on all variables is available. The time series will give an almost complete picture of the post war period. This time series is essentially the same as those in Kiewiet and McCubbins (1991) and Mayhew (1991). The dependent variable will be known as "relative change in the status quo." Both the national bicameral and tricameral studies share the same dependent variable. The units of analyses are individual policy proposals within a two-year Congress. There are three aspects of breaking gridlock or changing the status quo: 1) Was a major change proposed? 2) What was the magnitude of that proposed change? 3) How much change occurred relative to what was proposed? With respect to the first item, a reliable and well-tested arbiter of importance was readily available. Was a major change proposed? In every addition of the annual Congressional Quarterly Almanac there is a section entitled "Major Legislation." This section comprises the main body of the almanac and it is divided into various issue areas. Health policy is one of the issue areas. Each year, a select number of health policy proposals are identified by Congressional Quarterly Almanac as most important. Fenno (1973), Kingdon (1989), and Krehbiel (1991) have used the CQ method to determine whether 28 or not a policy proposal was relatively important. Major changes in health policy, however, have recently been a part of omnibus reconciliation bills instead of the more traditional form of a discrete policy proposal. Congressional Quarterly has adjusted to this phenomenon by sorting through the omnibus bills and splitting them into the various individual items enabling the analyst to treat each item as an individual piece of legislation. Once the proposal was identified as important, a determination was made as to whether or not it was a new policy or a continuation of a policy previously enacted into law. The goal was to identify policies that were intended to be a change from, rather than a continuation of, the status quo.“ A proposal was considered to be a continuation if it was merely a reauthorization of an existing policy or program. Only new policy proposals were included in the sample. The major exception is the inclusion of entitlements. Proposed changes in entitlements were included even though by definition they incrementally change only existing proposals. These proposals have been a part of omnibus reconciliation bills and these changes were identified according to the procedure outlined above.5 4 The variables in the bicameral and tricameral models were regressed on a dependent variable that included reauthorizations. The results were almost identical to those obtained with the restricted dependent variable. Please refer to Appendix 2-3. 5 The idea is to get some measure of the initial “shock” to the system in each 29 What was the magnitude of that proposed change? The next step was to determine the magnitude of the policy proposal. This was accomplished by calculating the "real mean annual authorization" for each policy proposal. In the provisions of each proposal there are annual authorization levels, usually, for two or three years. The authorized dollars for each year are added together and then divided by the number of years in the proposal. For example, if a policy proposal was intended to last for three years and the annual authorizations were $35, $40, and $75 million respectively, then the mean annual authorization would be $50 million [(35+45+75)/3].6 This figure is further adjusted to take into account the relative change proposed. A $25 million mean annual authorization in 1989 is not the equivalent of the same authorization in 1949. Each mean annual authorization figure was divided by the Gross Domestic Product for the year policy proposal. This is not intended to be a measure of the actual effect of the policy once the law is passed. 6 There is a potential problem here in focusing strictly on a monetary measure. There is a difference between monetary change and policy change. An alternative measure was constructed which attempted to take this into consideration. This alternative dependent variable was operationalized as the number of lines that CQ Almanac devoted to that bill. One of the problems with this measure is that it was only correlated with the monetary dependent varialbe at the 0.34 level. This was not encouraging. Granted, a monetary measure is not perfect but one would hope that two alternate measures of the same phenomenon would be a bit closer. In the end, the line counting method was considered to be less reliable than the monetary measure. 30 in which it was proposed as a way of determining the scale of the proposal relative the size of the national economy at the time. Each figure was then multiplied by 1,000 in order to avoid dealing with too many digits to the right of the decimal point. This figure, the "real mean annual authorization," standardizes the change from one year to the next so that direct comparisons can be made. Those proposals which did not include an explicit monetary component were excluded from the sample. These issues were purely regulatory in nature and comprise a small number of the proposals. In addition, regulation is primarily a state function and, therefore, the exclusion of these items should not bias the national level sample to any significant degree. Why authorization figures instead of appropriations? The politics of appropriations are the politics of bills already approved and not the politics of whether or not to change the status quo (Kiewiet and McCubbins 1991, Mayhew 1991). An appropriation of money to a particular program occurs only after a particular measure is enacted into law. The decision of how much to allot to a particular program is different than the decision to change or maintain the status quo; "authorizing committees are asking different questions that are appropriators" (Evan 1995, 62). The proposed authorization figures, on the other hand, are part and parcel of a policy proposal. It is the authorization figures that are used in the debates over these bills. In order to evaluate the phenomenon on its own terms, therefore, 31 authorization figures have been used instead of appropriations. If the policy proposal was approved by the House, Senate, and the President, thus becoming a public law, it is counted as "enacted." If it did not become a public law it is counted as "failing to gain approval." Because a policy can be proposed in the first session and enacted in the second session of a Congress, the dependent variable is calculated on a two-year basis. In order to determine the change that occurred relative to the change that did not occur, the "real mean annual authorization" levels of those policies which were enacted were coded as positive and those which were not were coded as negative. The sum of all the proposals constitutes the quantitative measure of relative change in the status quo for that given Congress. A positive number indicates a net positive amount of change relative to what was proposed and a negative number indicates a small amount, if any, change relative to what was proposed. Consider the case of 1965-66 as an example of net positive change. A total of eight policy proposals fulfilled the criteria of being major, non- continuing, monetary items. These items were coded as positive if they became law and negative if they did not. The following is a list of these items with the real mean annual authorization for that item in parentheses immediately after: Medicare (+9.68); Regional Health (+0.168); Mental Health Staffing (+0.06); Medical Professions Training (+0238); State Health Services (+0.10); Hospital Modernization (-1.33); Allied Health Professions 32 (+0.11); International Health (+0.01). The sum of the real mean annual authorization levels for all of the bills is +9.036. This figure is the measure of relative change for 1965-1966. This figure is considered to be representative of a great amount of change relative to what was proposed. Consider the case of 1945-1946 as an example of a small amount of policy change relative to what was proposed. A total of seven policy proposals fulfilled the criteria of being major, non-continuing, monetary items. The following is a list of the proposals: District of Columbia Hospital Center (+0.16); Hill-Burton (+.035); National Mental Health Program (-0.07); Cancer Research (-0.47); National Health Insurance (-7.06); Maternal and Child Welfare (-0.47); Social Security Act Amendments (-1.082). As opposed to the positive number of 1965-1966, the sum of the "real mean annual authorizations" for 1945-1946 is -8.642. A table of the dependent variable is in Appendix 2-4. The most active periods in terms of relative change in the post-war era occurred immediately after the Johnson (1965), Carter (1977), and Reagan (1981) administrations took office. The largest failures, in terms of relative change, took place during the Truman, Eisenhower, Ford, and Reagan presidencies. These negatives were not off-set by other successful policy proposals. A complete list of the policy proposals in the time series can be found in Appendix 2-5. Other Operational Definitions of Similar Concepts. Other studies that attempt to define a dependent variable that measures systemic policy 33 production according to changes in "macro" variables include Freedman (1995) and Mayhew (1991). These "macro" factors are broad measures of governmental change such as whether or not the government is divided or unified. Mayhew's seminal work (1991) provides the starting point from which Freedman (1995) and this study depart. Mayhew's dependent variable is the number of important pieces of legislation that become law from 1947 to 1990. Mayhew defined importance using a two-step process. The first step was to catalog those laws that were identified as very important in the year- end summaries of The Washington Post and New York Times. Step two involved an identification of those laws which have been determined by scholars and historians to be important. Mayhew did not divide the issues into separate policy areas. Mayhew found that policy proposals became law at essentially the same rate regardless of whether the government was unified or divided. Mayhew found that the policymaking system produces 12.8 major laws per session under unified government and 11.7 laws per session under divided. Freedman (1995) includes the universe of all health care laws that were passed from 1945-1992. She develops a five-point ordinal scale of importance: Trivial, Non-Substantive, Minor, Major, and Historic. Trivial laws are "commemorative laws or other legislation that has no direct policy impact" (Freedman 1995, 7). Non-substantive legislation are bills that are non-trivial but ignored by Congressional Quarterly Almanac. Minor bills are 34 those that are mentioned in CQ Almanac but not in the annual summaries of the New York Times or the Washington Post. Legislation is considered to be major if it is mentioned in the annual summaries of the above papers "but is not identified by these sources as a historic milestone" (Freedman 1995, 7). Finally, historic legislation is that which "satisfies Mayhew's ‘Sweep One' or ‘Sweep Two' criteria" (Freedman 1995, 7). Like Mayhew, Freedman found that there was no essential difference between the number of historic laws passed under unified regimes and that of divided regimes. In fact, to the extent that there is any difference at all it is in favor of the divided regimes as being the most productive! Under unified regimes, Freedman concluded that 0.88 laws per session were passed; under divided regimes that number is 1.27. In addition to this, Freedman offers a startling observation: The data show that Republican presidents operating under the constraints of divided government prove the most productive in terms of historic health legislation -- a finding that is quite counterintuitive to most students of health care policy. Besides averaging a higher number of laws per Congress, the Nixon and Reagan administrations appear to have produced the greatest number of historic health-related laws, each one comparing favorable to the efforts of Kennedy and Johnson (1995, 11). F reedman's linear model of the passage of important health care laws yields only null results. She tested two different theories of policy production. One was a party-based model and the other an ideological model. Neither model had any significant variables at the 0.10 level. Superiority of Relative Change. Before an examination of the 35 differences between the operational definitions of Mayhew, Freedman, and this current study, it is important to highlight some key similarities. This study and Freedman’s both exclude purely regulatory bills at the federal level for the reasons stated above. In addition, like this study, Freedman excludes appropriations bills for similar reasons as well. By constructing a ratio-level weighting system, the dependent variable in this study is not tied to who mentioned the bill in a report or a year-end summary. The importance of the bill has a numerical definition that can account for any number of gradations. Mayhew's scale is nominal: either the bill was important or not important. Freedman's scale is ordinal: trivial, non-substantive, minor, major, historic. The variable in this study is a ratio- level measure that allows the analyst to not only say a bill is more important than another but also by how much. Mayhew and Freedman needed to examine newspapers, Congressional Quarterly Almanac, and various academic sources to construct their sample of bills. In the end they only had ordinal and nominal data respectively. This study only needed to consult one archival resource, Congressional Quarterly Almanac, and my measure yields more information due to its higher level of measurement. In addition, there is no bill of any importance that I did not include which they did. The relative change variable yields more information for less work. Thus, it is more efficient. Beyond the cost of calculation, due to the fact that they each have 36 measures which are nominal or ordinal, conceptual mistakes are made. Freedman claims that Republican presidents under a divided government are the most productive. From the quote above, she claims that Nixon and Reagan compare favorably with Kennedy and Johnson in terms of the production of historic health care laws. This, however, is not the case. Only an ordinal scale like Freedman's which weights the Tax Equity and Fiscal Responsibility Act (TERRA) of 1981 as an equally historic piece of legislation as Medicare of 1965 could make this mistake. According to the scale in this study, TERRA of 1981 is weighted at +1.41 and Medicare is weighted at +9.68. This scale more accurately reflects the relative importance of each piece of legislation. According to the weighted scale in this study, the mean annual production rate of health legislation for the Kennedy-Johnson years was 1.42. The mean annual production rate of health legislation for the Nixon and Reagan years was 0.63. The Nixon and Reagan years were 44% less productive than Kennedy-Johnson! This is a more accurate reflection of the relative importance of the reforms. Another major difference between the dependent variable used in this study and the variables used in the Freedman and Mayhew studies is the inclusion of bills which did not become law. By including only bills which became law Mayhew and Freedman are trying to explain a phenomenon by using only the successful cases. There may be as much to learn, if not more, from what failed (Schick 1995). Take the case of 1993-1994 as an example. 37 Using the weighted values of each proposal and including only those which became law, the first two years of the Clinton administration produced legislation worth a total of +2.03 in value. This compares favorably with the historic 1981-1982 session under Reagan which produced legislation worth +2.64. This is not, however, how people perceive Clinton's first two years. Clinton's first two years are looked upon as a disappointment because so many important proposals did not pass. When the Health Care Reform Act and other proposals which failed are included and coded according to the method mentioned above in a previous section, the value of the first two years of the Clinton administration in terms of health policy change is -0.06 which is a more accurate reflection of what occurred relative to 1981-1982 or 1965-1966 for that matter. It is this feature of the variable in this study which allows for the capture of the phenomenon of gridlock. Frustration is often felt more intensely over the hills which have not passed as opposed to the relative production of different administrations across the decades. In the next section, the independent variables are identified and explicated. The balance of this chapter will be devoted to an examination of the national bicameral model while the next chapter will analyze the same dependent variable but with variables appropriate to a national tricameral model. 38 Independent Variables The differences between the two bodies that are achieved in a class- based society were approximated by Publius through the differences in age requirement and tenure. These structural differences would, according to Publius, manifest themselves in the compositional differences of Occupation, Age, Seniority, and Retention. These differences would allow the two bodies to be separate enough to fulfill the stability role that, in republics past, had been accomplished by the antagonism between different classes. The differences between the two chambers according to the aforementioned characteristics were supposed to provide the requisite stability necessary for a republic to survive. In this section, each of these four characteristics will be defined as a variable and then used as causal variables in a linear time-series regression model based on a theory that the more the chambers become alike, the more policy change (instability) will occur. Occupation. Publius intended for the Senate to be the more stable body. The requirements for eligibility for the Senate would seem to make it more likely that persons of wealth and standing would comprise a greater, proportion of that body than in the House. Both today and in the past lawyers have become the symbol of such wealth and standing. According to the theory as presented in the Federalist Papers, lawyers should make up a greater portion of the chamber in the Senate than in the House. By 39 examining Appendix 2-6 we can see that lawyers have slipped in the House from 56% to 41% from 1953 to 1993 and have remained stable in the Senate at roughly 60%. In terms of occupation the current composition of the two bodies is along the lines predicted by Publius. Age. In order to be a member of the House of Representatives, an individual must be 25 years of age when seated. The Senate's requirement is that the citizen be 30 years of age when seated. This is consistent with Publius' desire to have the Senate be the more rational body: "The qualifications for Senators, as distinguished from those of representatives, consist in a more advanced age and a longer period of citizenship...The propriety of these distinctions is explained by the nature of the senatorial trust, which, requiring greater extent of information and stability of character, requires at the same time that the senator should have reached a period of life most likely to supply these advantages" (Publius [1787-88] 1961, 376). In Appendix 2-7 we see that the mean age in the Senate is consistently greater than that in the House and that the difference between the two is approximately the five year space that is outlined in the Constitution. Retention and Seniority. Terms were intended to be staggered so that some stability in government would be maintained and balanced with the democratic need to have the government remain close to the people: "it was declared to be among the fundamental rights of the people that parliaments ought to be held frequently" (Publius [1787-88] 1961, 328). The entire House, 40 therefore, is up for reelection every two years while only one-third of the Senate is up for reelection in the same time period. This was designed to have a policy impact by allowing change but the change would happen slowly enough to make it more probable that the change would be for the better. The first measurement dealing with terms will be the percent of the membership of the respective chamber that is retained from the previous congress. The second measurement, which is just a long-run variable of the first, is the seniority in each body defined as the mean number of years of service by members in each body. We would expect the Senate to have more retention and, therefore, more seniority. With respect to retention (Appendix 2-8) we see that there is no real discernable difference between the two bodies. Both have varied from approximately 80 to 90 percent since 1953. There have been some "spikes" in the difference between the two bodies. These spikes occurred as a result of the 1964, 1966, 1974, 1982, and 1992 elections. In these cases, it does appear that the Senate retained more of its members than the House due to the fact that only 1/3 of the Senators were forced to face a volatile situation. This difference is in the predicted direction. With respect to seniority, the differences disappear (Appendix 2-9). The mean years of service by members in their respective chambers remained approximately 10 through the entire series with the difference rarely reaching above two years. 41 Data Analysis and Discussion The functional relationship of the national bicameral model can be expressed as: Relative Change = f( Occupation, Age, Retention, Seniority). The theory employed by Publius and operationalized in this study is that the differences between the bodies will introduce stability into the policy making system. The more different the two chambers are in terms of the above characteristics, the less policy change we should expect to see and vice versa. This relationship can be expressed more directly as: Relative Change = f{[Senate Occupation - House Occupation], [Senate Age - House Age], [Senate Retention - House Retention], [Senate Seniority - House Seniority]}. The linear expression of this model, therefore, would be: Relative Change = BO + BI(SOccup - HOccup) + BZ(SAge - HAge) + B3(SRet - HRet) + B4(SSen - HSen). As was explained above, the operational definitions of the independent variables are differences between the two chambers in terms of the macro characteristics outlined by Publius. The larger number for an independent variable, the greater the difference between the two bodies, and the less relative change we should expect to see. The signs for all of the variables, therefore, should be negative. The results of the time-series linear regression can be found in Appendix 2-10. As we can see, none of the variables registers enough predictive power to reject the null hypothesis that there is no relationship. 42 The diagnostic tests of the model that were conducted revealed no violations of any of the classical assumptions of regression analysis (Appendix 2-11)7. The results, however, are not completely discouraging. The results leave some hope that in future studies of different policy areas or government settings, a significant effect may be found. The variable that displayed the best overall performance would be Seniority. This variable has the correct sign and approaches statistical significance. Occupation registers the correct sign. The standard error for this variable, while too large to allow any substantive inference, is smaller than the coefficient. The 7 Ordinary Least Squares Regression has the quality of being “BLUE” among all other linear estimators. BLUE stands for best, linear, unbiased, estimator. The key to interpreting coefficients from a linear regression is the t statistic. The t statistic is obtained by dividing the coefficient by its standard error. There are three main problems with any estimating technique: bias, efficiency, and constancy. An estimator is unbiased if the mean of many iterated estimations equals zero. An estimator is said to be efficient if there is less variance around the mean. An estimator is said to be consistent if the mean of the estimates gets closer to the actual value as the sample size increases. Problems of bias effect the t statistic through the coefficient. Problems of efficiency affect the t statistic through the standard error. Problems of constancy affect the interpretation of estimators obtained from small samples. Violations of the classical assumptions of regression analysis make the estimates obtained less reliable. Under such, conditions, OLS regression is no longer BLUE. This does not meant that if no violation exists the estimators are perfect. It simply means that we can have more confidence in the inferential characteristics of the analysis. The following tests were conducted: The Dufour test for weak exogeneity(consistency), the RESET test for model specification(bias), the J arcque-Berra test for normality(bias), the Durbin-Watson test for autocorrelation(efficiency), the ARCH test for autocorrelation and heteroskedactiticy(efficiency), and the Ferrer-Glauber test for collinearity(efficiency). 43 performance of Age and Retention, on the other hand, are less encouraging. Each of these variables has the wrong sign and the standard errors are larger than the coefficients. SUMMARY AND CONCLUSION The proper construction of the bicameral Congress was the cause of significant debate at the Constitutional Convention for the very reason that the difference would have a significant effect on policy. Most of the research to date on bicameralism, however, has focused only on describing the differences between the two chambers. A spatial analysis of Hammond and Miller (1987) illustrates how bicameralism contributes to stability but an empirical analysis was still left to be constructed. The goal of this study is to show some of the effects of those differences on the enactment of policy. The dependent variable of the model in this chapter is sensitive enough to evaluate three separate dimensions of health policy: 1) Was the change major? 2) What was the magnitude of the change? 3) How much change occurred relative to what was proposed. The strength of this variable with respect to other variables (Freedman 1995, Mayhew 1991) is that it defines the concept at a higher level of measurement and includes proposals that did not pass along with those that did. The dependent variable in this study yields more information at less cost and, therefore, it is the superior variable. The independent variables for the model in this chapter were derived 44 from the arguments of Publius in Federalist Nos. 52-66. The variables were Occupation, Age, Retention, and Seniority. A time series regression analysis of these variables on the dependent variable yielded null results, opening the way for an analysis of the variables according to a more institutional model that would include presidential and partisan characteristics. This analysis will be described in the next chapter. The analysis in this chapter contributes to the body of literature by paving the way for future attempts to show how the differences between the chambers affect policy. The model is unique in the sense that it has a dependent variable that is sensitive to many factors and draws from theoretical arguments both ancient and modern. It would appear then, according to the results of this analysis, that the national bicameral model does not fare well within the context of post-war health legislation. We should keep in mind, however, that the sample is from health policy alone and only at the national level. The results may be different with samples from other policy areas or at the state level. An analysis of different policy areas will have to wait for a subsequent study, but an analysis of the model at the state level will follow in Chapter Five. There is, however, another avenue of approach that must be explored at the national level. The business at hand now is an analysis of the national level using the same dependent variable but different independent variables. It could be the case that a bicameral model does not produce a 45 significant result because the system is really tricameral. It might be an empirical oversight to talk of legislative progress in the Congress without taking into consideration the characteristics of the president. In addition, it may be a mistake to look at the raw demographic characteristics instead of more institutional factors. In Chapter Three, a tricameral model is applied to the same dependent variable. CHAPTER THREE: NATIONAL TRICAMERALISM IMPORTANCE OF TRICAMERALISM AND LITERATURE REVIEW In this chapter the separation of powers is examined from a tricameral perspective. If a two chamber legislature can be thought of as a bicameral system, then the three-part system consisting of the House, Senate, and the president can be thought of as a tricameral system. The argument presented in this chapter is focused on the more institutional factors of the system such as the partisanship and electoral performance of the House, Senate, and the president. There have been several studies on the tricameral system. These include Alt and Lowry (1994), Cox and Kernell (1991), Fiorina (1992), Freedman (1995), Krehbiel (1993), Mayhew (1991), Peterson (1990). These studies represent a variety of perspectives. The separation of powers from the tricameral perspective is one of the key defining features of American government. Most democracies in the world are parliamentary democracies. In a parliamentary democracy, there is no separation between the legislative and executive branches. The cabinet itself is staffed by the top members of the legislature. In addition, there no independent executive elected on a national basis with a veto power. Because of the great potential power of the veto, it was a matter of contention at the Constitutional Convention whether or not to give the veto power to the 46 47 president in our newly created Republic. It is this veto power which animates the institution of the presidency with respect to legislation. One might suggest that this is an overrated argument because, after all, the veto is rarely used. While this is true, it is not the use of the veto that makes it powerful. It is the impending threat of a possible veto that forces members of the House and Senate to take into account when drafting legislation (Cox and Kernell 1991). This factor makes the inclusion of the president a vital tool of analysis in the policymaking system. The following is a summary and evaluation of studies of tricameral analyses. Cox and Kernell (1991) and Peterson (1990) examine the tricameral system in order to determine which branch is more powerful in the policymaking process. The focus for them is a comparative evaluation of the branches. Fiorina (1992) focuses on the causes and condition of divided government at the national and state-level. Krehbiel (1993)analyzes the policy effects of the system from a more formal and theoretical level. The focus of the analysis in Alt and Lowry (1994) is state-level budget and fiscal policy. The goal of the study of Cox and Kernell (1991) is to evaluate the relative strengths and weaknesses of the Congress and the president in enacting their favorite policy proposals. They note that Congress, beginning in the 1970s, began to reassert itself primarily in budget matters. The Budget Impoundment and Control Act of 1974 and the exemption of more 48 than a dozen agencies from OMB review are just a sample of the actions taken. The president, on the other hand, still wields powerful weapons including: 1) concerted public relations, 2) centralized administration, and 3) use of veto threats (1991, 9). The veto, however, according to Cox and Kernell, has its limits. As recent events suggest, the threat and use of a veto can be very helpful to a president in a divided government. This same veto, however, is almost useless in a unified government. It cannot bring people together for it is a negative weapon by nature. It can be used to make a policy less extreme but it cannot be used to move policy out to the cutting-edge. This may be why a veto threat or usage is effective when the Congress attempts to adjust the rate of growth in an already established entitlement system but ineffective in attempting to create a brand new entitlement in terms of national health insurance. It is a weapon that can prevent things from occurring but it cannot make things happen. The key point of the work by Cox and Kernell is that the veto may be the most powerful tool wielded by any single individual in the policy process, but it has its limits. Knowing when and how to use it is the key. Improper usage or threat of the veto will lead to a weakened presidency. The weapon is good but the user may not be: "The veto does not give the president the wherewithal to pry more out of the Congress than it is willing to give" (Cox and Kernell 1991, 103). 49 Peterson (1990) argues that a tricameral, as opposed to a bicameral model or president-centered model, is the proper way to analyze the policymaking system at the macro level. The unit of analysis for Peterson is presidential policy proposals and why they pass or fail. Peterson does not attempt to sort out important from non-important bills. Peterson weighs all policy proposals equally. One of the most compelling findings of Peterson was that policy proposals which have widespread jurisdictions will be less likely to pass because many committee referrals provide many veto points and this favors the status quo. Peterson also addresses the issue of the recent dominance of Congress. The weakness of the presidency at various times, according to Peterson, cannot be ascribed to the institution. The weaknesses instead have been with those who have held the office. The goal of the study of Fiorina (1992) is to describe the prevalence of divided government in recent history at the state and national levels and also to try to discover some of the causes of divided government. Fiorina's state level analysis concurs with J acobson's (1990) argument at the national level that divided governments, until very recently, were occurring with more frequency because of the "decline of Republican legislative strength" (Fiorina 1992, 58). He agrees further with Jacobson that the increased level of professionalism in legislatures hurts Republicans because they are less likely than Democrats to consider lawmaking as a legitimate career choice instead of something one does in one's lifetime as a public service. According to 50 Fiorina, divided governments become more common when there are matters of considerable controversy amongst the American electorate. The more troubled the times, the more likely it will be that divided governments will occur. A large consensus that leads to huge united electoral coalitions can only take place when there is substantial agreement over the direction the nation's policy should take. Such times include the strong Republican eras following the Civil War and at the turn of the century and the strong Democratic eras of the J acksonian and the pre-war FDR years. The object of the work by Krehbiel (1993) is to render an analysis of the tricameral system from a formal perspective. Like Hammond and Miller (1987), Krehbiel uses a spatial analysis to illustrate his ideas. Krehbiel argues that the phenomenon of gridlock is the product of the differences between the president, House, and the Senate in terms of the characteristics of the individuals in those institutions and the institution of the filibuster. The legislative actors in the policymaking system in any given time are the result of many different temporal and geographic majorities. Taking up the issue of temporal majorities first, the current House of Representatives is the representation of the preferences of voters no longer than two years ago. The Senate on the other hand is a combination of the preferences of voters from the last three elections. The president represents the preferences of voters from up to four years ago. Knowing that the moods of the country can change from month to month, the diversity of the different 51 electoral majorities provides a potentially staggering amount of difference in the policy system and this difference can lead to gridlock. This diversity becomes even stronger when we analyze the different geographical majorities. The members of the House of Representatives are a reflection of the preferences of the voting majority in 435 individual districts that are roughly equal in population. The members of the Senate are the representatives of the majorities in each state while the president is the only true national representative. In addition, because of the institution of the filibuster, a bill must, as a practical matter, have the support of at least 60% of the Senate. Krehbiel argues that gridlock can occur in unified governments because of the filibuster. The notion of a unified government is a false one unless the party is united and also holds a supermajority in the Senate. Unless this is the case, divided governments or even weak unified governments may be the best policy producers. The context of the study by Alt and Lowry (1994) is the fiscal policies of American states. More specifically, they are concerned with the extent to which state budget deficits or surpluses deviate from the expected levels. They concluded that Democrats and Republicans behave differently when in control of the government. The parties emphasize different fiscal objectives with Democrats favoring higher levels of spending and taxes. They also concluded that divided government does lead to a form of gridlock which 52 makes it harder for state governments to react to budget crises. This project is different from the above in several respects. Cox and Kernell (1991) and Peterson (1990) are both concerned with the relative strengths of the various branches. Fiorina is concerned with the causes of divided government. Krehbiel offers a formal analysis without any empirical backing. Alt and Lowry focus on the state level alone and use budget figures instead of more discrete policy proposals. Mayhew and Freedman have been analyzed in the previous chapter. All of these studies are important. This study, however, adds to this body of literature by looking at a measure of individual policy proposals, weighing those proposals according to an objective measure of importance, evaluating the production of policy instead of the relative strengths, and focusing not on the causes but on the potential effects of divided government. MODEL The dependent variable is the same as in the national bicameral model. The independent variables, however, are different. The time series will be from 1953 to 1994 because 1953 is the first year in which information is available for all variables. The independent variables that will be used in the national bicameral model were culled fiom the recent scholarship on Congress and the presidency. Unified or Divided Government The success of the president in terms of getting an agenda passed is 53 dependent, for the most part, on the partisan and ideological composition of Congress (Bond and Fleisher 1990, Light 1991).8 Paul Light, through the example of the domestic policy successes of President Lyndon Johnson, demonstrates that "Neither institutional prerogatives nor bargaining skills explain Johnson's dramatic success. Johnson's higher degree of success paralleled the increase in his political resources following the 1964 election" (1990,26). An anonymous OMB official remarked: You ought to think of the presidency as an engine. Each president enters office facing the same model - the horsepower is generally stable and the gears are all there. What differs is the fuel. Different presidents enter with different fuel...Lyndon Johnson entered office with a full tank, while Ford entered on empty (in Light 1990, 14) The individual skills probably have very little to do with whether or not a major health policy initiative can survive the rigors of the tricameral system: Harry Truman could not have gotten national health insurance through the Eightieth Congress had he been Lyndon Johnson, John Kennedy, Franklin Roosevelt, Machiavelli, and Odysseus all rolled into one (King 1983, 254). From the perspective of the member of Congress, it is easier for them to get their legislation into law when they have a fellow partisan in the White House. What contributes to the gridlock of a divided government and what makes the veto such a handy weapon is the fact that both parties are 8 The first divided government was in 1826 when the J acksonian Democrats took control of Congress in a mid-term election. J .Q. Adams, an old-line Jeffersonian Republican, was the president at the time. 54 preparing for the next election in an attempt to unify the government. The president has an incentive to avoid cooperation and the Congress has an incentive to make life difficult for the president. According to Cox and Kernell, the main job of the president should be to prepare for the next election so that the situation in Congress can improve, from the president's point of view (Cox and Kernell 1991). The public commitment to veto is a common tool of presidents in a divided government (Cox and Kernell 1991, 107). Thus, the power of the presidency may be felt the most when it contributes to gridlock. This is because the veto can rarely be used to extract more from a Congress than it is willing to give but it is often used to get less than what the Congress had wanted. The short review of the literature above is representative of the consensus that a unified government variable belongs in a model that attempts to analyze the policy effects of the tricameral system. The operational definition of the variable will take the form of a dichotomy: 1 = unified government and 0 = divided government. During the time series of this study, 1949-1994, 45% of the governments have been unified(Appendix 3-1). Those governments would be Truman from 1949-1952, Eisenhower from 1953-1954, Kennedy/Johnson from 1961-1968, Carter 1977-1980, and Clinton from 1993-1994. All things being equal we would expect more relative change to occur in a unified government. 55 Seat Change Another factor which the literature identifies as an important factor in the tricameral system is seat change. The number of seats that switch from one party's control to the other's should have an impact on whether or not major policy proposals are enacted into law. A large change in favor of the party of the president should help to promote change. A large change away from the president should limit the possibilities of the tricameral system converging on an agreement to change the status quo. Members of Congress "are notorious for watching electoral trends" (Light 1991, 30). Mayhew states that: Nothing is more important in Capitol Hill than the shared conviction that election returns have proven a point. Thus the 1950 returns were read not only as a rejection of health insurance but as a ratification of McCarthyism (197 4, 71). The new legislators may be more likely to toe the party line. In fact, they may be the party line. They are important because they are the passionate representatives of what Rohde refers to as "natural partisanship," the partisanship that is the result of electoral forces (Rohde 1991, 171). Brady, Cooper and Hurley argue that: Perhaps the most important external variable, at least in terms of time devoted to it by congressmen, is elections. While research directly linking voting behavior to elections has not revealed any clear pattern of relationships, there is evidence that freshman congressmen vote in accordance with the party position more often than their senior counterparts (1987, 240). 56 The operational definition of this variable is the result of a three stage process. First, the net seat change in each chamber was coded as positive if it was in favor of the president's party and negative if it was not. Second, the net seat change in each chamber was divided by the number of seats in the respective chamber and multiplied by one hundred. The use of a percent figure provides a control for the size of the chamber. Third, the mean of the two percentage numbers for each chamber was calculated. From Appendix 3- 2 we see that the highest values were reflected in the years of 1949-1950, 1953-1954, 1965-1966, and 1981-1983. All things being held equal, we should expect more change to occur as this variable increases in value. We should also expect this effect to get stronger under conditions of unified government. An interaction term between Seat Change and Unified Government will be created and placed in the model. This variable is different from the dichotomous variable of unified government because there is variation in the Seat Change numbers under unified governments while there is no variation within all unified governments. The value is always one.9 9 There are two possible ways to model this variable. One way is to measure the Congress as a whole which is the method used in this model. Another way is to calculate the changes for each chamber separately and enter them in the model as their own variables. An estimation of such a model can be found in the Appendix ( 3-3). The results indicate that a comprehensive seat change variable is superior in this instance. 57 Presidential Popularity It is a matter of contention whether or not public opinion polls are a significant source of presidential influence. There seems to be a consensus in the literature that, at the very least, the electoral performance of the president in the previous election is a source of strength or weakness depending on the numbers. Popularity is the single-most important resource that a president has and even then, "the effects of presidential popularity are marginal at best" (Bond and Fleisher 199, 25). There is even evidence to suggest that a popular president, by being popular, may actually do as much damage as good for the president's favored legislation: "popular presidents tend to receive more support from members of their party but less support from members of the opposition" (Bond and Fleisher 1980, 75). Bond and Fleisher demonstrate that "liberal Democrats, for example, did not become solid supporters of President Reagan even at the zenith of his popularity" (1990, 29). Whatever popularity that does count with Congress, however, does not appear to be public approval (Collier and Sullivan 1995, 197). It is electoral performance, the proven ability to win at the polls, which will influence those on Capitol Hill. How the president's party fares on election night will have more to do with the president's success than any other variable (Bond and Fleisher 1990). The percent of the popular vote received in the last election is the indicator that members of Congress take seriously (Light 1991). 58 The operational definition of the presidential popularity variable is the percent of the popular vote received by the president in the previous election (Appendix 3-4). The highest figures occurred as a result of the 1964,1972, and 1980 elections which elected Presidents Johnson, Nixon, and Reagan respectively. The lowest figures were the result of the 1968 and 1992 elections which elected Presidents Nixon and Clinton respectively. We should also expect this variable to be a stronger contributor to relative policy change under two special conditions. The first condition is unified government. Another presidential popularity variable will be created through an interaction with unified government. The second condition is a landslide election. There are certain elections that are extraordinary. In the aftermath of these rare elections, policy proposals that would normally fail can pass. This variable will be an interaction between unified government and all elections in which the president received over 60% of the popular vote.10 Party Strength The existence of strong or weak parties helps to determine the 10 In order to test if the operational definition of landslide was too strict, a test was conducted on the full model with landslide defined as any president receiving over fifty-five percent of the popular vote. The results for this model indicate that the less restricted version is not a good predictor of the behavior of the dependent variable(Appendix 3-5 ). 59 character of a government, be it divided or unified. The parties in Congress will be weak when there are policy divisions within the party. In a divided government, weak parties may allow for change to occur and strong parties may lead to gridlock: All of them [the postwar presidents] were forced to varying degrees to seek bipartisan support to get their programs enacted. In the House, the Democratic Party always included 60 or more southern conservatives who were opposed to much of their party's economic and social programs. And a score of moderate to liberal Republicans frequently were at odds with the party's conservative majority (Korns et al. 1982, 156). It could also be the case, however, that a strong Congress can get legislation enacted into law if the president is perceived to be weak or disinterested in domestic affairs. Rohde demonstrates that this may have been the case in the latter stages of the Reagan administration: Wright devised a list of priorities...despite the opposition of the president and House Republicans in almost all instances. In the case of these priorities, partisanship was not muted by divided government, although compromises were eventually reached in some instances. Nor was stalemate the result, and certainly not inaction. Clearly the president did not set this agenda; he opposed it. Yet not only did every one of the ten items on Wright's list of priorities pass the House; every one of them eventually became law in one form or another (1991, 175). In a unified government, on the other hand, weak parties may inhibit the convergence of policies to the median preferences of those in the tricameral system and strong parties may allow the party in control to pass an impressive number of policies like the Johnson administration in 1965-1966. unified may depend on the extent to which members are cross-pressured 60 (Keefe and Ogul 1987). Members like to go along with the party leadership when they can. Whether they can or not will depend upon whether the member is under any pressure from constituents to go against the party leadership in that body (Brady, Cooper, and Hurley 1987, 248). The lessening of cross-pressures will lead to an increase in party control and this will help enable the party to behave as a unit in favor of enacting change or maintaining the status quo (Keefe and Ogul 1987). Rohde demonstrates that congressional partisanship has recently been on the rise. Rohde isolates four main causes of this phenomenon: 1) The elections of the late 1950s and 19603 brought in more liberal Democrats who took power away from committee chairmen and gave it to the majority caucus, 2) In the 1970s (post-reform) electoral forces had an impact through national and district-level changes that resulted in electoral coalitions of representatives that were more similar within parties and more different between them, 3) This increased homogeneity provided the basis for more aggressive use of reform-granted powers by the Democratic Leadership,11 4) Partisanship was enhanced by the impact of individual personalities. President Reagan and Speaker Wright are good examples (1991, 162-163). 11 The Legislation Reorganization Act of 1970 forced all committee roll-call votes to be recorded and created the recorded teller vote on the floor and in the Committee of the Whole. 61 Rohde is careful to point out, however, that partisanship is conditional. It will only be evident on issues of significant import, where there is controversy among the electorate and between parties instead of within them. The operational definition for party strength is the mean percent of party unity votes in the two chambers in a given Congress when interacting with unified government. This interaction is necessary in order to maintain parameter constancy. Under a condition of divided government, we would expect this variable to be negatively related to relative policy change but under unified to be positively related. This interaction is necessary in order to be able to interpret the coefficient( Appendix 3-6). Data Analysis and Discussion The functional relationship of the model can be expressed as: Relative Change = f( Unified, Seat Change, Seat Change*Unified, Presidential Popularity, Presidential Popularity* Unified, Party Strength*Unified, Landslide). The linear expression of this model, therefore, would be: Relative Change = B0 + B1(Unified) + Bz(Seat Change) + Ba(Seat Change * Unified) + B4(Presidential Popularity) + B5(Presidential Popularity * Unified) + BS(Party Strength * Unified) +B7(Landslide), The results of the time series regression analysis can be found in Appendix 3-7. The diagnostic tests of the model that were conducted 62 revealed that there is a substantial collinearity problem with the model (Appendix 3-8). This renders and inferential interpretaion suspect. The problem variables are unified government and the interactive term of party strength with unified government. These two varaibles were removed from the model and a second estimation was conducted which yeilded the results presented in Appendix 3-9.12 The diagnostic tests can be found in appendix 3-10. No violations of the classical assumptions of regression analysis were detected”. The overall quality of the model is strong. The R2 is 0.61 which is quite strong. Some might suggest that this is relatively weak for a time series, but one must keep in mind that, unlike many time series, this does not include a lagged dependent variable. For a time series without a lagged endogenous variable, these are strong overall numbers. The F statistic which is a judge of the 12 GLS is not used because it is essentially atheoretical. It puports to fix a problem without an identification of the cause. This is like an engineer painting over a crack in a bridge (Granato 1992). Diagnostic problems are best viewed as modelling problems and best solved by adjustments to the model, the theory, or both. 13 In addition to the tests conducted on the model in Appendix 3-7, a recursive least squares analysis was conducted on the model in Table 3-9 to test for weak exogenetity. The Chow tests that were conducted revealed no violations of the assumption of parameter constancy in the 1 step, increasing, and decreasing break point tests. In addition, reduction techniques applied using the method in Granato (1992) reveal that dropping the interactive term of popular vote with unified government from the model increases the performance of the model. 63 explanatory power of the model as a whole is significant at the .01 level.14 The strongest performing coefficient is Landslide. From the intercept of a relative change figure of -6.55, a Landslide election will contribute 12.60 points of relative change to the policy system. This is in the expected direction and it is also the strongest variable. This makes sense given that it is the rarest and most extreme variable and it is here that we would expect the most change. It is also significant at the 99% level of probability in a two- tail test. The next best performing variable is Seat Change. It is significant at the 95% level of probability for a two-tail test. The sign of the coefficient is in the expected direction. A one-unit increase in mean bicameral change will contribute 0.31 positive units in relative change. The performance of Presidential Popularity, while not significant, is encouraging. The sign of the coefficient is in the proper direction and the t- value is moderately strong indicating that the popular vote received in the previous election may still be of some use in empirical analysis. Presidential Popularity when interacting with Unified government is statistically insignificant The performance of Seat Change with Unified Government requires some explanation. 14 A model which included an economic factor, the GDP deflator, was estimated. The results indicate that this variable adds nothing to the model(Appendix 3- 11 ). 64 The interactive term of Seat Change with Unified Government is statistically significant at the 99% level of probability in a two-tail test. The coefficient, however, has the wrong sign. The coefficient indicates that, in unified governments only, a one unit increase in mean bicameral seat change, will cause there to be a 0.77 unit decrease in relative change. This is non-intuitive because we expected the positive effect of positive mean bicameral seat change to become even stronger in a unified government. This is especially disconcerting since the actual seat change variable was significant. Krehbiel (1993) in a formal analysis mentioned above, suggests that the filibuster may be responsible for causing what Krehbiel refers to as "unified gridlock" a condition in which the status quo is maintained even though a single party has a numerical majority in the Congress and controls the presidency. The filibuster, which requires a vote of at least 60 members to end, has turned, for all intents and purposes, the Senate into a supermajoritarian body. The majority status of a party is deceiving because a committed block of 41 Senators could, if unified against a certain policy issue, halt the progress of the bill even if it is supported by the president, a majority in the House, and a majority in the Senate. It may indeed be the filibuster which gives us such strange results as less relative change when the president's party gets stronger in Congress under a unified government. What is the cause of all of this? 65 It may be the case that a minority Senator may be more likely to compromise in a unified government where the majority party in the Congress is not very strong. In a situation where the government is unified and the majority party is strong, the minority Senators may look upon themselves as the last outpost of their beliefs and be more willing to tenaciously deny policy victories to the majority. This situation, as was the case in 1965, may be overcome by a massive landslide that renders a filibuster inoperative and the model does indeed display this characteristic. Short of a landslide, divided governments or weak unified governments may give us the best chance at policy change. With a filibuster, the least likely scenario for policy change may be a unified government that has done reasonable well in the elections but falls short of a landslide (Krehbiel 1993)! These results, even though they were obtained from a substantially different method of operationalization and methodological approach, confirm the findings of Mayhew (1991) and Freedman (1995) that it is possible that divided governments can do as well as, if not outperform, unified governments in terms of policy production. The lesson that Mayhew draws is that gridlock is not as serious a problem as we may have been led to believe. On the other hand, Mayhew (1991) and Freedman (1995) don't offer much of an explanation as to why this might happen. An attempt to offer an explanation may lead to another answer which is different than the one given by Mayhew and Freedman. 66 The possible lesson that can be drawn from the results in this analysis is that gridlock may exist in both unified and divided governments. The next step would be to test this hypothesis in other policy areas and levels of government to test the external validity of these propositions. While an analysis of this phenomenon in another policy context will have to wait for the time being, an analysis of this phenomenon at the state level will be explained in the next chapter. In comparison with the national bicameral model, the national tricameral model is superior. What can be learned from this? The national bicameral model is constructed with demographic characteristics of the two chambers while the national tricameral model is constructed with partisan institutional variables. A tentative conclusion which can be drawn fi‘om this study is that in an analysis of the policymaking system along macro lines, such as this study, Mayhew (1991) and Freedman (1995), one should focus on the more institutional factors such as the presidency and partisanship and less on the demographic factors. A tentative result of this comparison of the two models against the same dependent variable is that institutions matter and partisanship is important, even though the results are counter to how we expect partisanship to influence the process. SUMMARY AND CONCLUSION In this chapter the separation of powers was analyzed according to a tricameral model. If the two chamber legislature can be thought of as a 67 bicameral system, then the three-part system consisting of the House, Senate, and the president can be thought of as a tricameral system. The separation of powers from the tricameral perspective is one of the key defining features of American government. Most democracies in the world are parliamentary democracies. The argument presented in this study adds to the literature on bicameralism by examining a measure of individual policy proposals, weighing those proposals according to an objective measure of importance, evaluating the production of policy instead of the relative strengths, focusing not on the causes but on the potential effects of divided government. The dependent variable was the same as in the national bicameral model, however, the independent variables were different. There were seven independent variables; five after adjusting for multicollinearity: Unified Government, Seat change, Seat Change interacting with Unified Government, Presidential Popularity, Presidential Popularity interacting with unified government, Party Strength interacting with unified government, and Landslide. The overall quality of the model was strong. The strongest performing coefficient was Landslide. The effect was in the expected direction and it was also the strongest variable. The larger the landslide, the more likely health policy change. The next best performing variable was Seat Change. The sign of the coefficient was in the expected direction. The performance of the interactive 68 term of Seat Change with Unified Government required some explanation. It had a negative sign when a positive sign was expected. This could possibly be the result of the availability of the filibuster. These results, even though they were obtained from a substantially different method of operationalization and methodological approach, confirmed the findings of Mayhew (1991) which have been controversial. In addition to this, the results of the data analysis suggest that institutional factors such as the branch of government and partisanship can be factored into a statistical model and yield insights concerning the very real world of policy production, which is the core of our governmental system. The larger the seat change, the larger the change in health policy with the exception of unified governments. CHAPTER FOUR: CASE STUDIES The analysis in this chapter is of two bills, on their own terms, with the spotlight on real human beings and their undetermined behavior. The goal guiding this analysis is to discover both the unique and generalizable circumstances of each case in order to better understand tricameralism. This chapter is not a supplement of the tricameral model. Case studies are important tools that can be used to examine variables in a level of detail that might not be possible in a more aggregated study. The methodology that will be employed is inspired by the theory of design presented by King, Keohane, and Verba (1994). There is a brief review of the origins of each proposal. This is the more familiar "story" component which has become associated with case studies. The analysis, however, continues with a systematic review of a number of factors with a comparison at the conclusion of the chapter. The cases which will be examined are Medicare and the Clinton Administration's Health Security Act. They were selected because they offered a rich source of descriptive data from which to analyze the key factors. It is possible that some aspects could have been squeezed into a quantitative mode. The emphasis in this chapter, however, is on the analysis of factors as they apply to disaggregated cases. This is especially important in this dissertation where the quantitative 69 70 dependent variable is an aggregate summary taken over a two year periods stretching from 1953-1993. The variables which will be examined are the following: electoral context, presidential support, conference committee, filibuster, veto, domestic context, international context, interest groups, multiple referrals, early or late introduction into Congress. MEDICARE Brief Review of Medicare The history of Medicare can be divided into four phases (Appendix 4-1). The first phase was 1935 to 1949. The Wagner-Murray-Dingell proposals for comprehensive national insurance were in the spotlight during this period. The second phase began in 1957 when Aime Forand sponsored a more limited form of national health insurance. The Forand plan was to be financed through the social security system and would only be offered to seniors. The Forand bill became the basis for the Kennedy proposal of 1960 which formed part of the final Medicare bill. The third and penultimate phase was what Marmor called "the era of possibility" (1973, 39). The election of Kennedy left the White House controlled for the first time in eight years by a proponent of national health insurance. As this era came to a close in 1964, the Senate finally passed a Medicare proposal. The only obstacle the remained was in the House Ways and Means committee and the election of 1964 made the removal of that obstacle inevitable. The final phase consists of the 71 maneuvers that led to the final passage of Medicare. 1935-1949 The genesis of Medicare was the Social Security bill of 1935. The original bill contained just one line that hinted at the possibility of financing health care through the social security system. That single line caused so much controversy that it was pulled from the bill. It was feared that the entire social security proposal would be defeated if it were included (Feingold 1966). The idea was so controversial that another plan for national health insurance was not introduced until 1939. Instead of basing the system's finances on the social security model, the bill called for the provision of grants to the states so that they could set up their own programs (Skidmore 1970). Pressure from the American Medical Association, however, prevented the bill from being reported out of committee. The remainder of the first phase was taken up with the consideration of the Wagner-Murray-Dingell (WMD) proposals of 1943 to 1949. They are so named because of the bill's sponsors; Senator Robert Wagner (D-NY), Senator James Murray (D-MT), and Representative John Dingell Sr. (D-MI). The WMD bill would have provided a system of national health insurance financed through payroll deductions (Skidmore 1970). The first conservative alternative to the WMD plan appeared in 1947. This bill was sponsored by Senator Robert Taft (R-OH), a leading Republican 72 candidate for the presidential nomination. The difference between the WMD and Taft bills was a microcosm of the difference of opinion between liberals and conservatives. Neither the Democrats nor the Republicans disagreed over the basic notion that the government should help people in need. The controversy was over the method of providing that help. The essential rift can be exposed by identifying the difference between the notion of social insurance and charity. The WMD proposals were based on the social insurance model of assistance and the Taft proposals were based on the charity model. The social insurance model was formulated in America after the publication of I.M. Rubinow's 1913 classic, Social Insurance. Those who favor the social insurance model have the view that goods like health care and conditions like economic security are basic rights. They are on the same level as the rights recognized in the first ten amendments to the Constitution. Because it is an accepted right, one does not need to prove a need for a service. The service, in this case health care, should be available on equal terms regardless of income level. The preferred method of finance in a social insurance system is a payroll tax. The practice of social insurance dates back to the efforts of Bismarck in Germany in 1883. The charity model is based on the notion that the market can best provide for the health needs of citizens. If the market fails, the government should step in and make it possible for people to obtain services. In order to 73 receive benefits an individual would have to pass what is known as a "means test." The test would determine if the individual had the means to provide the needed services. The means test can be traced back to the English Poor Laws under Queen Elizabeth I. It is the liberal belief that health care is a right. A means test, therefore, is both demeaning and unnecessary. From the liberal point of view, proving a need for health care would be tantamount to proving a need for freedom of speech or freedom of religion. From the conservative viewpoint, the provision of services via government finance to those who could otherwise afford it is an impingement on the individual's liberty and an inefficient economic practice. The Medicare package that gained final approval in 1965 included all perspectives. Part A of Title 18 of the Social Security Act was derived form the social insurance principle. Title 19 of the Social Security Act, Medicaid, was derived from the charity notion. Part B of Title 18 of the Social Security Act is a compromise between the two approaches. 1957-1960 The period from 1957 to 1960 can be defined as a period of narrowed focus and issue formation, but little legislative action. A group of former Roosevelt administration officials and labor leaders decided after the last WMD bill failed in 1949 that 1957 was the time to act once more. Private insurance continued to grow throughout the 19503. In an attempt to keep 74 premiums competitive, private companiew were shutting seniors out of the health market (David 1985, Harris 1966). This group thought that a national health insurance plan, limited to those over 65 and financed through the social security system, would have a better chance of passing than the comprehensive WMD plans. Forand, the fourth ranking majority member of the House Ways and Means committee, was the highest ranking member of Congress willing to sponsor the bill. He was unenthusiastic about its chances. The Chair of Ways and Means, Wilbur Mills, refused to even hold hearings on the bill. Liberals and conservatives alike were barely aware of its existence and those who were, were not hopeful of passage. Slowly but surely, however, grass- roots enthusiasm began to build for the proposal. The Forand bill never made it out of the Ways and Means Committee but it became the basis for the debate that would take place in the following years (Feingold 1966). From 1958 until the passage of Medicare in 1965, annual hearings were held on some form of the Forand bill. It was the presidential campaign of 1960 that brought the issue to the top of the policy agenda. There were three main alternatives in 1960. The first proposal was the one offered by the liberal Democrats. The Democratic presidential nominee, Senator John Kennedy, was squarely behind a Forand-type proposal. While his running-mate, Senate Majority Leader Lyndon Johnson had been an 75 opponent of the Forand bill, he changed his mind when he joined the ticket. The second proposal was supported by moderate Republicans. The Republican nominee, Vice-President Richard Nixon, did not want to yield the ground of aggressive policy solutions to the Democrats. Nixon, through the sponsorship of Senator Jacob J avits (R-NY), supported a plan that would subsidize private insurance companies. They would then provide insurance on a charity basis to those who could not otherwise afford it. A third plan had the support of the more conservative Democrats and Republicans. It was sponsored by Senator Robert Kerr (D-OK) and Representative Wilbur Mills (D-AR). This proposal was based on the charity model. The Kerr-Mills plan would be financed by the federal government out of general revenues but administered by the states. Only the Kerr-Mills plan emerged from Congress. This plan, a precursor to Medicaid, was intended to provide the means for states to finance health care for the poor through federal block grants. It was signed by President Eisenhower. With the election of Kennedy to the White House, however, the hopes were very high among liberals that F orand-type legislation would be enacted in 1961. 1961-1964 This period was one of rising expectations for Medicare. The chances of enactment increased with each passing year. The politics of this period centered upon the liberal claim that the Kerr-Mills plan was inadequate and 76 the conservative counter-claim that it was. Kerr-Mills was supported by the American Medical Association. It was the position of the AMA that Kerr-Mills was the proper policy and that any attempt at compulsory health insurance for the aged would be the first step to socialized medicine (Harris 1966, David 1985). Kennedy realized the rhetorical power of such charges and sought to narrow the coverage under the original Forand proposal in order to increase the probability of passage. This move was necessary because Congress was dominated by a conservative coalition: "That strategy softpedalled the innovative character of the program in an attempt to widen agreement on the legitimacy of government involvement in health insurance" (Marmor 1973, 40). The sponsors in the Congress were Sen. Clinton Anderson (D-NM) and Rep. Cecil King (D-CA). The King-Anderson bill was the Forand bill with minor modifications. Neither was considered to be a particularly influential legislator, and in Anderson's case not particularly liberal, but they were the highest ranking legislators who were willing to sponsor the measure. The conservatives based their rhetoric on the claim that Kerr-Mills was adequate. The co-sponsor of the measure, Wilbur Mills, was the Chair of the House Ways and Means Committee and was not interested in backing the administration’s proposal. As it became clear that the chances of a Forand- type bill getting passed in the House would be slim, the administration was forced to turn to the Senate for support. 77 The King-Anderson bill was not first introduced into the Senate because it was essentially a revenue bill. This made it necessary for any original proposal to pass through the gauntlet of Mills's Ways and Means Committee. The administration came up with the tactic of attaching Medicare to Social Security bills that had already passed the House. They then hoped to work out a compromise in conference. The most prominent of these "riders" was the Anderson-Javits bill of 1962. The bill, however, could only muster the support of five Republicans. The reason for this lack of bipartisan support was the massive lobbying campaign waged by the American Medical Association (Harris 1966). In the aftermath of the Kennedy assassination and the impending victory of Johnson, the chances for Medicare improved. Sensing this, Wilbur Mills attempted to outflank Medicare supporters by proposing a broad liberalization of the Kerr-Mills program. This proposal was attached to the social security bill that cleared the House in 1964. The administration sought to add Medicare as a rider to the Senate bill and hope for a compromise in conference. This rider passed on a close 49 to 44 vote. For the first time in the history of the health debate, a chamber of the United States Congress went on record in its support for Medicare. The conferees, however, were unable to reach a compromise. The issue would have to wait for the aftermath of the elections. The elections of 1964 brought a large number of Northern Liberal 78 Democrats into Congress. In addition, Democratic losses were concentrated in the South. Liberal Democrats simultaneously increased their ranks while the ranks of their conservative Democratic opponents dwindled (Feingold 1966, Rohde 1991). Shortly after the election, reading the writing on the wall, Mills indicated his willingness to support a Medicare bill in the upcoming 89th Congress. The battle for Medicare had been won. The details were left for the debate of 1965. 1965 Marmor calls the beginning of 1965 "the politics of legislative certainty" (197 3, 59). So that nobody would miss the point, the respective Medicare bills in the House and Senate were numbered HR 1 and S 1. With the certainty of something being passed before the end of the year each side staked out its position hoping for a compromise. The liberal and conservative positions had not changed since 1960. The liberals rallied around the Medicare bill (Forand/King-Anderson) and the conservatives, backed by the AMA, rallied around the extension of the Kerr-Mills program and dubbed it "Eldercare". The only new proposal, "Bettercare," was put forth by mainstream Republicans. The Republican leadership in the House came up with a plan which would provide a wider set of benefits than either Medicare or Eldercare. It was to be financed from general revenues and a progressive payroll tax instead of the flat tax of the Medicare plan. The bill was sponsored by 79 Wisconsin Representative John Byrnes. This bill caught both the liberals and conservatives by surprise. Mills, and other Democrats, were concerned that the moderate Republicans were going to steal the thunder of the Democrats. Not to be outdone, however, Mills had a surprise of his own. Mills fashioned a bill humorously dubbed "Medi-Elder-Better Care." This bill can be envisioned as a three-layer cake. The Forand/King- Anderson-type plan was the first layer, the Byrnes plan was the second, and Kerr-Mills was the third. The first layer, King-Anderson, would be financed by an increase in the social security tax. This second layer, the Byrnes bill, would provide an insurance plan to pay physician’s fees. The third layer was an expansion of the Kerr-Mills plan. Mills's plan to include all three versions in one bill stunned both supporters and opponents of Medicare alike. The Republican tactic of trying to outflank the Democrats had backfired. Wilbur Cohen offers this vivid account of the Ways and Means hearing in which Mills first announced his plan: Like everyone else in the room, I was stunned by Mills's strategy," Cohen said afterward [Wilbur Cohen was Assistant Secretary of HEW and the intellectual leader of the Medicare movement]. "It was the most brilliant legislative move I'd seen in thirty years. The doctors couldn't complain, because they had been carping about Medicare's shortcomings and about its being compulsory. And the Republicans couldn't complain because it was their own idea. In effect, Mills had taken the A.M.A.'s ammunition, put it in the Republicans gun, and blown both of them off the map." Byrnes too was stunned. "He just stood there with his mouth open," a member of the committee said later (Harris 1966, 187). Johnson was pleased with the compromise and although there was some last 80 minute tactics by Senate conservatives to kill the bill, it passed the Senate, made it through the conference, and was signed in Independence, Missouri, in the presence of former President Harry Truman. Case Study Variables Electoral Context Lyndon Johnson recorded a historic victory over Barry Goldwater in the 1964 elections. Johnson received a stunning 61.1% of the popular vote and 486 electoral votes while Goldwater received 38.5% and 52 votes respectively. Turning to the legislature, the composition of the 89th Congress was the biggest Democratic majority since 1937. The Democrats had a 68 to 32 advantage in the Senate and a 295 to 140 advantage in the House. More importantly, the Democrats who came into the House and Senate tended to be more liberal than in the past. So, unlike the partisan majority of 1937, this was an ideological majority. Presidential Support Both Presidents Kennedy and Johnson were supporters of Medicare HEW co while they were in office. Kennedy, however, never made it a top priority. Kennedy was more concerned about getting his tax cut through the Congress than getting Medicare passed (David 1985). He did not want to press Mills too hard for fear of alienating him on the tax issue. In addition, the resources of uld not be focused on Medicare with any strength because 81 the agency was fighting a loosing battle over federal aid to education (Feingold 1966). It was not until 1965, when other important issues such as civil rights had finally been addressed, that Medicare could become the top priority. Conference Committee The final version of the bill was more like the House plan than the Senate's and this was to be expected. The Senate was the more extreme of the two parties from the status quo. They would still be willing to accept a version closer to what the House was willing to offer. Let Appendix 4-2 illustrate the situation. Let q represent the status quo of no Medicare program. Let 3 represent the position represented by the Senate conferees and let h represent the position represented by the House conferees. Positions considered to be liberal would be to the left and conservative to the right. The farther to the right, the more conservative the actor. The farther to the left the more liberal the actor. Any point to the left of the status quo would be more preferable to the Senate than the status quo so the House could dictate the final terms of the proposal and this is, indeed, What happened (Cox and Kernell 1991). The critical assumptions being made are a unidimensional policy space and a one-shot game.15 15 A great deal of this depends upon who is setting the agenda. The assumption in this model is that the House delegation, led by Mills, is setting the agenda. 82 The Senate's version called for $7.6 billion and the House version called for $6.0 billion. The final version was estimated at $6.5 billion. The conference rejected the Senate's plan to lower the age of beneficiaries for Social Security cash benefits from 62 to 60. They also rejected the Senate's proposal to include payments to radiologists, anesthesiologists, physical therapists, and pathologists (RAPP) in Part A of the Plan (King-Anderson). One of the concessions to the Senate, however, was the acceptance of the higher taxable income base. Filibusters Oddly enough, the only filibuster that affected Medicare came from the ideological left in 1964. Minority Leader Everett Dirksen (R-IL) had attached a rider to a foreign aid bill that would have halted the court-ordered reapportionment plans for state legislatures following Baker v. Carr and Reynolds v. Sims. These reapportionment plans were expected to cost Republicans in terms of seats. Some liberals filibustered the Medicare bill to stop the Senate from completing action on it and then proceeding to consider the Dirksen rider. Behind the scenes negotiation resolved the filibuster and Medicare was eventually free to be considered (Feingold 1966). The Veto President Eisenhower certainly would have vetoed the Forand bill had it been presented to him during his administration. Veto threats from Presidents Kennedy and Johnson were not forthcoming. They were 83 supporters of Medicare. Vetoes are negative weapons, and it is very difficult to make a negative instrument perform a positive task. Vetoes cannot pry more out of a Congress than they are willing to give. An illustration of this phenomenon can be found in Appendix 4-3. The notation is the same as Appendix 4-2 save the fact that a p is added to represent the president. When the president wants to change the status quo, the veto is not a credible threat. The president, in this scenario will be willing to accept anything that is to the left of the status quo (Cox and Kernell 1991). Again, the assumptions of this model and the veto model in the Health Security Act case are a unidimensional policy space and a one-shot game. Domestic Context The Johnson administration did not face the level of domestic distraction that it had in the past. The successful resolution of legislative issues concerning civil rights helped alleviate major sources of controversy from previous years. The nation was experiencing the quiet before the storm of the Viet Nam years and the civil rights struggles which needed to be won on the state and local levels. International Context The war in Viet Nam escalated in 1965 and it might have come closer to affecting the passage of Medicare than anyone would have predicted in 1964. In the first half of 1965 there was general chaos in the Republic of Viet Nam. On January 2 ARVN [Army of the Republic of Viet Nam (South)] forces 84 were defeated by the Viet Cong at Binh Gia. On February 7, the Viet Cong staged a surprise attack on the US. base at Pleiku. The situation turned very serious as ARVN forces retaliated by taking the war directly into the North. It was becoming clear that the United States, in order to keep its publicly made promises would need to increase it commitment of forces on the ground. The war, at this point, overshadowed and began to threaten domestic legislation like Medicare (CQ Almanac 1965). On the day immediately following the final passage of Medicare in the House, President Johnson announced that the United States' presence in Viet Nam would be increased from 75,000 to 125,000. By the end of the year, American forces in Viet Nam numbered 180,000. If the opponents of Medicare could have delayed the bill for a few more weeks, this may have forced Johnson to put off the announcement of the troop build-up. This might have increased the chances of a sudden military disaster in Viet Nam. If this had occurred, Medicare may have been delayed well into 1966 and, due to the impending election results and the nature of the eventual 90th Congress, it is doubtful that Medicare would have passed. The international context was probably the greatest threat to the passage of Medicare. Interest groups The AMA was Medicare's chief opponent and the AFL-CIO was its chief proponent. The lobbying effort reached its peak in 1964-1965. The AMA reported spending $952,000 in the first quarter of 1965 alone (CQ 85 Almanac 1.965). The AMA's total 1965 spending of $1.152 million was the third highest amount ever recorded for a lobbying effort. In 1961 the AMA had also created a political action committee to support candidates favorable to their policy positions. American Medical Association Political Aciton Committee (AMPAC) funds were concentrated on Ways and Means members who supported the AMA's position. The counterpart of AMPAC on the AFL-CIO side was COPE (Committee on Political Education). COPE worked especially hard to defeat members of Ways and Means who were opposed to Medicare. Both groups acted in both the primaries and the general elections. HEALTH SECURITY ACT Brief Review Not long after the passage of Medicare and Medicaid, costs began to escalate and cost control became the dominant policy concern. This period can be divided into three phases (Appendix 4-4). The first phase lasted from 1971-1974.16 The major proposals were advanced by Sen. Kennedy (D-MA) and the Nixon administration. Each had a plan for universal comprehensive health insurance. The Kennedy plan was based on the Canadian model in 16 While President Nixon did indicate that there was a health crisis as early as 1969, the Kennedy plan was the first serious attempt to deal with the issue of costs and that plan did not rise near the top of the agenda until 1971. 86 which the federal government would control the finances of the health care system. The Nixon plan would subsidize the purchase of private insurance and impose employer mandates. The second period lasted from 1978-197 9. The proposals in this period were more modest in comparison with the first phase. President Carter, under the pressures of budget austerity, proposed a bill that would only provide catastrophic coverage and would sharply curtail hospital spending. The third phase was 1991-1994. True to his "New Democrat" image, President Clinton's proposal resembled the Republican plans of the early seventies more than the traditional Democratic single- payer proposals. 1971-1975 From 1960 to 1965 medical costs as a percentage of the national economy rose from 5.2% to 5.9%. By 197 0, however, the percentage had increased to 7.1% (Congress and the Nation 1969-197 2, 522). The Nixon administration, congressional leaders, and the American Medical Association proposed their own solutions to this problem. The only major proposal to pass any chamber of Congress in this period was the Nixon administration's plan. The Nixon proposal had three components. The first was a requirement that every worker be provided health insurance by his or her employer. Each plan would be forced to provide a mandatory set of benefits. The second component was the replacement of Medicaid with subsidies to 87 low-income families so they could purchase private insurance. The third component was the development of Health Maintenance Organizations (HMOs). Unlike the traditional fee-for-service arrangement, an HMO provides care to an individual or a family for a set fee which is paid in advance. The hope is that primary care will be utilized more often preventing the development of more expensive serious ailments. The Nixon plan was introduced by Senator Wallace Bennett (R-UT) and the unwitting author of Medicare Part B, John Byrnes (R-WIS). The first two components never made it out of committee. The HMO plan passed the House but was not acted upon by the Senate. The liberal alternative, sponsored by Sen. Kennedy and Rep. Martha Griffiths (D- MI), was a Canadian-style single-payer plan in which the federal government would take control of medical care. Physicians would receive compensation in the form of an annual salary or nationally established fees. Hospitals and other facilities would operate on an annual budget. The new system would provide comprehensive medical care to all citizens and would be financed through a 3.5% increase in the payroll tax and from general revenues. The most conservative of the three approaches was the AMA's Medicredit proposal. The AMA feared the employer mandate and HMO components of the Nixon plan. Medicredit would have provided income subsidies on a progressive scale to assist the purchase of private insurance. 88 The bill, sponsored by Senator Clifford Hanson (R-WO) and Representative Richard Fulton (D-TN), was never reported out of any committee in either chamber. While the Nixon and AMA proposals faded away after 1975, the Kennedy-Griffiths plan remained a viable reform option. It was a more direct approach than the other two, and therefore, easier to understand. Understanding the concept of government controlled finance is more lucid than insurance subsidies, risk pools, employer mandates, or HMOs. It also had the strong support of labor groups. A certain veto by President Ford kept the Congress fiom seriously considering the proposal even though it was quite popular. Meanwhile, health costs continued to soar. 197 8-1979 From 1965 to 1978, medical costs had increased 350% and were reaching an annual rate of increase of 15% (Congress and the Nation 1977- 1979, 601). The increase was the result of three factors. The first is the fact that the market for medical care is not a traditional market. An overabundance of suppliers in a traditional market would drive prices down, but because of the nature of medical services, the increasing supply of physicians, aided by a system of third-party payment, tends to increase demand for medical services. The second factor at work was the rapid advance of medical technology which could treat more diseases faster and better but at a higher cost. The third factor was the rising frequency and cost 89 of medical malpractice suits. In addition to the cost problem, there was great concern over the ever-growing number of elderly in the population and the large number of uninsured. The issue once again reached political saliency when candidate Jimmy Carter promised to have a comprehensive health insurance plan introduced into the Congress. Carter backed off of this promise because he thought that costs must be controlled first. Things changed, however, in 197 8. After intense lobbying from labor groups and political pressure from Senator Kennedy, President Carter announced that HEW Secretary Joseph Califano would draft a bill that would be ready in 1979. Carter's plan was modest in comparison with the earlier Kennedy proposals. Carter argued that a more comprehensive plan was not a fiscal possibility. This Carter plan was a version of the Nixon employer mandate, focusing on the aspects of medical care that were contributing most to the cost increase. The Carter plan required all employers to provide a basic package of catastrophic coverage or face severe tax penalties. The American Medical Association was ambivalent about the Carter plan. It certainly preferred it to the Kennedy plan but believed Carter's cap on hospital costs to be particularly onerous. Even this modest proposal, however, was never reported out of committee. There was a rising mood of conservatism in the nation and the Democratic party failed to unify behind a single alternative. 90 1991-1994 In one form or another, a plan for comprehensive national health insurance had been a major policy option from 1935 until 1980. This string was broken with the election of Ronald Reagan to two terms and the election of George Bush in 1988. The focus of these administrations was to control medical costs through the Medicare program. All of this changed on the first Tuesday in November of 1991 (Mann and Ornstein 1995). A special election took place to fill the seat of the late Senator John Heinz (R-PA) who was killed in a helicopter crash. The Republican nominee was former Pennsylvania Governor and United States Attorney General Richard Thornburgh. With President Bush riding high in the polls and the Democratic challenger a relative unknown, the election was expected to be a victory for Thronburgh and a harbinger of 1992. With the aid of James Carville it turned out to be a harbinger indeed but not the one predicted. Wofford rode a message of universal health coverage to victory and a seat in the United States Senate. The Democratic nominee for president, Governor Bill Clinton of Arkansas, then made universal health coverage a central feature of his campaign. By the time of his election, it was becoming clear that his political fortune and that of his party might be decided by the outcome of the health policy debate that would take place in the first two years of the Clinton administration. Unlike Presidents Kennedy and Johnson, Clinton made 91 health care reform the clear policy priority of his administration (Rovner 1995). By 1993, health costs as a percentage of the entire economy had approached 14% (CQ Almanac 1993, 336). It was also widely reported that 38 million people in the United States had no health insurance. There appeared to be wide-spread public support for a change in the health policy system. Because of budget restrictions, the Clinton administration could not simply offer to extend Medicare and Medicaid benefits until everyone was covered. This was a financial impossibility without massive tax increases. Clinton needed a different plan. The only major source of untapped money was in private insurance which comprised half of all health care spending in the United States. The federal government and the states contributed thirty and twenty percent respectively. The Clinton plan proposed to utilize the private insurance dollars through what were originally known as "health cooperatives." The name was changed to "health alliances" because it sounded less "socialistic" (Rovner 1995). Consumers of medical services would be pooled together into large groups called "alliances". These alliances would be able to negotiate a fairer price from health care providers because of their large buying power. This buying power would lower the cost of premiums while maintaining an acceptable level of service. Costs would be controlled by a global budget which would place a maximum, set by a national health council, on the 92 number of dollars that could be spent on health care in the nation in a given year. Access would be ensured by an employer mandate requiring all workers to be provided a basic plan of coverage approved by the government. This plan was formulated by the Task Force on National Health Care Reform headed by First Lady Hillary Rodham Clinton. The Task Force produced the plan in the Summer of 1993. The President introduced the plan to the nation in a dramatic nationally televised speech on September 22, 1993. The President produced a prototype of a health security card representing health care that would be universal and could never be taken away. Once the plan was formally introduced into Congress, a turf war began in both the House and the Senate. Turf wars are not just jurisdictional disputes. They are wars because people with different views on policy desire to control the turf (Evans 1995). The fight in the House was primarily between two committees, Ways and Means and Energy and Commerce. Because of the employer mandate and the tax increases, Ways and Means claimed jurisdiction. Energy and Commerce, however, had jurisdiction over many health areas including Medicaid. When asked which part of the bill his committee would control, Energy and Commerce Chair John Dingell (D-MI) replied "We have health" (CQ Weekly Report, October 9, 2734). Speaker of the House Thomas Foley eventually split the bill between the two committees and also allowed for a substantial role for the House Committee on Education 93 and Labor. The multiple referral would prove a problem for the prospects of the bill. A multiple referral provides more veto points for any proposal (Evans 1995). Senate rules do not allow for multiple referrals of bills. The two contending committees were Finance, chaired by Daniel Patrick Moynihan (D-NY) and Labor and Human Resources, chaired by Kennedy. Neither Kennedy or Moynihan was interested in a compromise. The animosity between the two chairs was out on the surface (Rovner 1995). Moynihan had made it clear that health care reform was not a top priority with him and his staunch desire to control it could be viewed as an attempt to kill it (Schick 1995). Senate Majority Leader George Mitchell (D-ME) had no choice but to refuse to refer the Administration's bill and let each committee draft its own proposal. Because of the open rules of the Senate concerning amendments, Mitchell hoped to work out a compromise on the floor. The administration's plan was not, however, the only proposal. Rather, there were a spate of proposals, seemingly one for every gradation of ideological conviction in the Congress (Appendix 4-5).17 The proposal most to 17 The author is well aware of the difficulty of precisely placing actors on an ideological spectrum. The scale should not be viewed as an ordinal or interval scale but as a nominal scale. Each ideological group is a rough description. The more control over finance that was ceded to the federal government, the more liberal the proposal was considered to be. The more the plan relied on subsidies or savings accounts to purchase private insurance, the more conservative. 94 the ideological left was championed by Rep. Jim McDermott (D-Wa) and Sen. Paul Wellstone (D-MN). This proposal was almost identical to the previous Kennedy proposals of 197 1-197 5. The federal government would replace private insurance as the clearinghouse for funds in the health care system. The "left" and "center-left" bills were versions of the Clinton proposal and differed in the level of benefits awarded and the extent of the employer mandate. The "center-right" plans of Chafee, Cooper, and Grandy focused on a reshuffling of the insurance market rather than more substantive change such as employer mandates or health alliances. The "right" and "ultra-right" proposals were built on the idea of subsidizing private insurance. The Gramm and Armey plans emphasized Medical Savings Accounts (MSAs) which could be used to pay for day-to-day expenses and would be tax free. As it turned out, the most serious alternative to the Clinton proposal was the Cooper-Grandy plan. It was widely accepted that the ultra left plans were proposed to stake out a bargaining position (CQ Almanac 1994, Rovner 1995). January 25, 1994, was the height of the health care drama. In his State of the Union Address, brandishing the presidential pen, President Clinton, made the fateful promise of his first two years: I want to make this very clear. I am open, as I have said repeatedly, to the best ideas of concerned members of both parties. I have no special brief for any specific approach, even in our own bill, except this: If you send me legislation that does not guarantee every American private health insurance that can never be taken away, you will force me to take this pen, veto the legislation, and we will come right back here and start all over again (Clinton 1994, 6-D, CQ Almanac) 95 While this threat made for good theater, it made for bad politics because it provided a clear opportunity to his opponents to delay the enactment of the bill. All his opponents had to do was to get a compromise for something less than universal coverage and the president had promised to start the whole process again. In effect, the president created an opportunity for every significant group in either the House or Senate to block the bill. This realization led the Administration, later in the year, to claim that 95% coverage was "universal" (CQ Almanac 1994, 351). On the other hand, the threat may have been a daring attempt to resurrect the idea of reform which had been losing ground in public approval since October of 1993 (CQ Almanac 1993, 321). It did energize the administration's most ardent supporters in Congress and helped to move the bill along in the committees (CQ Almanac 1994). It also energized the opposition. Only one Republican in a major committee, Senator Jim Jeffords (R-VT), supported the administration's proposal. The most severe blow to the plan came fiom a Congressional Budget Office report in February of 1994 which claimed that the Clinton plan was not, as it claimed, budget neutral. Rather, CBO estimated that the Clinton plan would contribute $60 billion to the deficit. This gave opponents an opportunity to oppose the plan on grounds other than a simple policy dispute. Opponents could claim the moral high ground of protecting the nation's finances. What was forgotten in the report was that the CBO agreed that the 96 president's plan would provide universal coverage as it had promised (Mann and Ornstein 1995, Rovner 1995, Schick 1995). This episode continued the recent trend of disputes over dollars instead of substance (Mann and Ornstein 1995). In the pall cast by the CBO report, Kennedy's Senate Committee on Labor and Human Resources, on June 9, became the first to report a comprehensive bill. Unlike the Kennedy proposals of the past, this bill was not a Canadian-style system. Kennedy wanted to produce legislation that had the best chance of passing. The Kennedy plan was modeled after the Clinton plan. The next Committee to announce its intentions was the House Committee on Energy and Commerce and the news was not good for supporters of reform. When the chair, John Dingell (D-MI), announced that his committee would be unable to reach a compromise, this was the beginning of the end for reform. Energy and Commerce was viewed as a representative committee and thus a bellwether for what would happen on the floor (CQ Almanac 1994, 321). In an almost pro forma fashion, the House Ways and Means and Senate Finance Committees reported their bills. Although the Congressional leadership attempted to produce passable bills in August of 1995, it was to no avail. On September 26, Senate Majority Leader Mitchell announced that health care reform would have to wait until the next Congress. Even though 97 it had failed, it had come closer to passing than any other comprehensive health reform proposal in United States history (Rovner 1995). Case Study Variables Electoral Context Bill Clinton received a modest plurality of the popular vote at 43% and a comfortable majority of electoral votes, 370-168. The results from the 1992 congressional elections were even less impressive. The Democrats maintained their majority in the House but they lost 9 seats leaving the balance at 258-176. They also maintained their majority in the Senate but lost a seat, leaving the balance at 56-44. This is relevant information because President Clinton was planning a reform of the health care system on a large scale with a very slim mandate. Executive Support There were many domestic policy distractions which took attention away from health care reform. The main policy disputes were the budget, NAFTA, and gun control. All things considered, however, President Clinton made it clear that health care reform was his administration's top priority (CQ Almanac 1993). Conference Committee The conference committee was not a factor because the plan did not emerge from either chamber. 98 Filibuster The shadow of an eventual filibuster hung over and changed the nature of the process. It was widely expected that the Republicans, led by Dole and Gramm, would filibuster any bill that made it to the Senate floor. This certainty affected action in the House as well. It tempered the ability to fashion a bill liberal enough to keep the Democrats together in order to get 218 votes. Veto This is one of the more curious matters in the entire case. This may have been the single biggest mistake made by any individual in the process. It is difficult for the veto to be successfully employed to extract more from a Congress than it is willing to provide. Consider Appendix 4-6. Let q represent the status quo of the health care system. Let 3 represent the position represented by the majority in the Senate,18 let h represent the position represented by the House, and let p represent the position of the president. Positions considered to be liberal would be to the left and conservative to the right. This figure represents the situation in 1993-1994. A veto threat by the 18 I am aware of the argument made in Krehbiel (1993) that the filibuster pivot is the key to understanding the Senate in a spatial context. I rely on this argument in other portions of this dissertation. Adding this to the spatial model here, however, would not add anything to the explanation while complicating the picture unnecessarily. 99 President is not credible because any movement to the left would be more preferred by the president than the status quo. In fact, when told he would have to accept something less than him most preferred policy, he agreed. Domestic Context The Clinton administration was significantly distracted by other pressing domestic issues. The most significant distraction was a close vote on the Budget in 1993. The measure passed the House by one vote and Vice President Gore needed to use his tie-breaking prerogative in the Senate. NAFTA and gun control also proved to be controversial measures that sapped political strength that could otherwise have been devoted to health reform. International Context Somalia, Bosnia, Haiti, and North Korea were significant foreign distractions. On October 3, 1993, 18 American soldiers were killed in a battle with the forces of Mohammed Farah Aidid (CQ Almanac 1994, 486). In Bosnia, America moved to the precipice and back several times. In Haiti, the difficulties eventually led to American intervention. Throughout 1993, North Korea played hard ball with the Clinton administration and dropped out of an international organization designed to limit the spread of nuclear weapons. Three of these crises weakened the president's position; Somalia, Bosnia, and North Korea. Haiti, on the other hand, turned out to be one of President Clinton's foreign policy successes. 100 Interest Groups The strongest support of the Clinton Administration's bill came from the American Association of Retired Persons (AARP). The AARP sent out massive mailings and commissioned numerous public policy studies (CQ Almanac 1994, 91). The American Medical Association did not take the lead in the opposition to health care reform as it had in the past. President Clinton, in a wise move, was very careful not to challenge the AMA directly. He realized he was working with a different generation of physicians and softened his tactics (CQ Almanac 1994). While the AMA did eventually oppose the plan, it did not do so with the same intensity that is opposed Medicare. In fact, a schism had developed with the national elite supporting the Clinton plan and the state and local societies opposing it. The strongest opponent of the Clinton plan was the National Federation of Independent Businesses. The NFIB was animated by the opposition of their members to the employer mandate. The NFIB skillfully used the tactic of cross-lobbying to deny Clinton the centrist allies he needed. Cross-lobbying is the act of pressuring other lobby groups to support or not support a position. One of the favorite targets of the NFIB was the AMA. The NFIB was credited with keeping Jim Cooper fi'om compromising with Dingell. Another prominent interest group was the Health Insurance Association of America. The group was headed by former Republican 101 member of Ways and Means, Bill Gradison. Gradison was widely recognized as one of the foremost experts on health care reform. The HIAA developed the famous "Harry and Louise" ads which depicted a couple discussing over the kitchen table all of the supposed terrible things that would happen if the Clinton plan were to pass. Ironically, the HIAA supported much of the Clinton plan. They were, however, against the health alliances. Comparative Analysis of Factors An aggregated statistical analysis can reveal patterns but a disaggregated case analysis affords the opportunity to examine factors that would slip through the statistical cracks. This does not mean, however, that the analysis is free from the need to be systematic (King, Keohane, and Verba 1994). In this section the factors from each case will be compared to the other (Appendix 4-7). Electoral Context. The electoral context is important because it is a visible sign of the president's "political capital" for each case was quite different (W eissert and Weissert 1996). Lyndon Johnson had engineered a massive electoral victory at the presidential and congressional levels. He could rightly claim that he had received a mandate to forge the enactment of Medicare with 155 more Democrats than Republicans in the House and 36 in the Senate. President Clinton, on the other hand, came to office in concert with a congressional election where over 40% of the new members were 102 Republican . To make matters worse for the incoming president, the Democratic party actually lost seats in both chambers of Congress. Even with the unimpressive electoral context, passage of some kind of reform was not out of the question with the case of the HSA. The idea of reform, the reality notwithstanding, was very popular with the public. A more modest step towards the ultimate goal of universal coverage might have passed. It is important to remember that even the strongest of conservatives, like Phil Gramm, were hesitant at first to attack the idea of reform. There was a window of opportunity to get something passed. Alas, for the supporters of health care reform, the Clinton administration attempted to pass a large package through the eye of the needle. Presidential Support and Early or Late Introduction. Both Medicare and HSA were top presidential priorities. Each proposal was the cornerstone of the preceding presidential campaign. President Johnson made it clear that Medicare was his top priority by making sure that the proposal was the first bill introduced into each chamber of the new Congress. President Clinton made it clear that the HSA was his top priority by designating the First Lady, Hilary Rodham Clinton, as the leader of the task force that would construct the bill. The difference in the two approaches, however, is the time of year in which the bill was introduced. Johnson's was introduced at the very beginning of the Congress. Medicare was also extremely well known. The Clinton proposal, on the other hand, was unlike any of the proposals that 103 had been introduced by Democrats in years prior to 1993. There was no bill ready to go. In fact, the Clinton proposal was more like the Nixon proposal of 1971 than any previous Democrat bill. The best people to help draft the bill early enough to give it a reasonable chance of passing would have been Republican members of Congress who were involved with the National Health Insurance debate of the early seventies. Ironically, the best possible person to work with would have been Bob Dole. In order for a bill to have been introduced at the beginning of the 103rd Congress, it would have to have been a Canadian-style, single-payer plan. One form of this bill or another has been introduced in Congress since 1939. Once again, the old AMA warning about "socialized medicine" killed another attempt at reform. This time the effect was indirect. A single-payer plan was not chosen in order to avoid the appearance of a government takeover of health insurance. The health alliance scheme was constructed to give the impression that the system was still "private." It was this complexity that led to other weaknesses in the political maneuvering. To avoid the claim of socialism, the Administration constructed a bill that was probably too complex to pass. Ironically, in trying to avoid the claim that a single-payer system was socialistic, the Clinton Administration produced an actual scheme that was reminiscent of the corporatist takeover of the Italian economy in the 19208. Conference Committee and Filibuster. The conference committee was 104 only a factor in the case of Medicare because the HSA did not emerge from either chamber. With respect to the filibuster, direct comparisons can be made. In 1965, the filibuster was not commonly used and there was not a committed group of Senators who either threatened or actually carried out a filibuster in the case of Medicare. With the HSA, on the other hand, the threat to filibuster was real and would most likely have been organized by the Minority leadership. This dynamic became important because the two early front-runners for the Republican nomination were Senators Dole and Gramm for whom the drama of a filibuster might have provided the perfect opportunity to define the differences between themselves and the Clinton Administration. If a filibuster would have begun it would have been difficult to end because neither Dole or Gramm would have wanted to be the first to compromise and thus lose the support of the core primary constituency in the Republican party. This filibuster threat even overshadowed the drafting of the bills in the House (CQ Almanac 1994, Rovner 1995). The threat forced Democrats to move any package as much to the center as possible. The hope was to gain the support of enough Republicans to convince them to help end a filibuster. This in turn alienated the staunch supporters of health care reform within the Democratic party. Therefore, no base of support was developed from which to compromise. Veto. The threat of a presidential veto was not a factor in the case of Medicare but it certainly played a dominant role in the HSA struggle. 105 Clinton curiously attempted to use the negative weapon of the veto to perform the positive task of extracting more concessions than Congress may have been willing to give. The threat to veto was damaging because it gave the opponents of reform one more vehicle to try to delay the passage of the bill. If, on the other hand, the president chose to back-down and compromise he would lose credibility with the legislature and even with the public since the threat was made on national television. Clinton did eventually back down from the promise but it was too late to do much good. In the end, he got the worst of both worlds. The plan did not pass and he lost face for backing down on his pledge. Domestic Context. The domestic context was more distracting for the HSA than it was for Medicare. While it is difficult to make the case that the Clinton administration faced active tumult and chaos, a case can be made that items such as NAFTA and gun control were, at the least, mild irritants. International Context. Johnson and Clinton faced different international contexts. Johnson, however, faced the most difficult and his bill still passed. Medicare was more threatened by Viet Nam than the HSA was by Somalia, Bosnia, Haiti, and North Korea (CQ Almanac 1965). In Viet Nam in 1965 we already had a substantial commitment of American troops and material in a hot war. This was not the case with any of the trouble spots in the Clinton administration. The international context was a factor but not a determining one. Otherwise, the Johnson bill would have failed 106 and the Clinton bill would have had a better chance of passing. Interest Groups. The interest groups were of a different configuration in each case and this was driven by the nature of each proposal. In the case of Medicare, the two main interest groups were the AMA and the AFL-CIO. They had been battling each other for decades over this and similar issues. Each was the undisputed leader of the opposition and advocacy respectively. A great deal of institutional memory, therefore, was developed by each side. The case of the HSA was much different. Their numbers were much larger than in 1965 and they were scattered and did not operate under an umbrella of two large groups like the case of Medicare (W eissert and Weissert 1996). The National Federation of Independent Businesses was opposed to the employer mandate but not necessarily the reorganization of the insurance market into health alliances. The Health Insurance Association of America was, conversely, supportive of the employer mandate but not in favor of health alliances. The AMA was not as vocal in its opposition to the Administration's plan as it had been about other plans in the past. In addition, the AARP was not nearly as supportive of the HSA as the AFL-CIO had been of Medicare. The plan had a little bit of everything for everybody. Accordingly, it also had a little pain for everybody as well giving many groups a reason to oppose one aspect or another. Multiple Referrals. Because of the focused nature of the Medicare proposal, there was no doubt about which committees would have 107 jurisdiction. The Medicare bill was designed to be an amendment to the Social Security Act of 1935 and thus clearly within the purview of the House Ways and Means and Senate Finance Committees. The HSA on the other hand was not a focused proposal. It dealt with such wide-ranging issues as workman's compensation, pension benefits, and mental health. The complexity of the bill made it difficult to develop a base of support. The turf fights and the subsequent multiple referrals did not kill the HSA, but they did create more veto points that eventually helped kill the proposal. CONCLUSION The variables that were associated with the fate of the two bills in the expected "direction" were: electoral context, filibuster, veto, domestic context, multiple referrals, and early or late introduction. The variables that did not perform in the expected was the international context. Variables that displayed inconclusive results were interest groups and conference committees. The variable which was not a factor at all in this analysis was the level of presidential support. Medicare became the law of the land and was superior in the following categories: better electoral context, no no filibuster threat, a less distracting domestic context, the significant committees in each chamber had sole jurisdiction over the bill, and was introduced early in the term. In the case of international context, the bill that faced the most danger passed shedding some doubt, in this analysis at least, on the linkage between domestic and foreign policies. CHAPTER FIVE: CASE STUDIES Background The United States is a federal system. Because of this, an analyst has up to fifty other contexts in which to test theories first constructed at the national level. The experience of Michigan, as a laboratory of democracy, may yield insights for the study of bicameral and tricameral systems. The diversity of Michigan mirrors national diversity in many ways. Michigan has two competing political cultures. While some recent research has discounted the effect of political culture on mass attitudes, the link with elite attitudes is still vital and reflected in policy outcomes (Baker 1990). The first is a moralistic political culture developed by settlers from New England, Scandinavia, and the Netherlands. The second political culture is an individualistic one developed by immigrants fiom the Southern United States, Southern Europe, and Eastern Europe (Elazar 1966). The individualistic culture tends to dominate in the Detroit area while the moralistic culture is more prevalent in out-state areas. The diversity of Michigan is not limited to political culture. Michigan, contrary to its image as being dominated by the auto industry in and around Detroit, is economically diverse. Grand Rapids is a center for the retail service and furniture industries. Battle Creek is the home of several worldwide food processing firms. Bay 108 109 City and Muskegon have been ship-building and lumber centers. Midland is home to a large chemical producing and research center and the Kalamazoo area is the home to one of the world's largest pharmaceutical producers. Besides manufacturing, tourism and the agriculture industry are critical to the state's economy. The lakeshores of the state are home to many tourist locations which form the crux of the state's second largest industry. In terms of agriculture, Michigan ranks third in the nation in the diversity of its products. All of this diversity has led to the creation of many interest groups. These groups tend to be very active and well-organized. There is, therefore, a great deal of competition for favorable policy enactments (Browne and VerBurg 1995). Michigan is a competitive two-party state (Dye 1984). From 1951 to 1994, 79% of the governments in the state of Michigan have been divided (Appendix 5-1).19 Remarkably, 21% of the legislatures have had one of the chambers in a virtual tie! During this time period there have been three Democratic governors (Williams, Swainson, and Blanchard) and three Republican governors (Romney, Milliken, and Engler). The Republicans had the governorship for 24 years and the Democrats for 22. During this time 19 The time series in this example is 1951-1996 and is intended to be the basis for a general description. The time series for the empirical analysis is limited to 1962-1994 because of the constraints of the data. They are intended to serve two different yet complimentary purposes. 110 period the Democrats controlled the house for 26 years, the Republicans 14 years, and the house was equally divided for 6 years. The senate was controlled by the Republicans for 30 years, the Democrats 12 years, and the senate was equally divided for 4 years. Michigan has a professional legislature and a professional executive branch. A strong governor and legislative leaders counteract many of the effects of the intense lobbying in a state like Michigan (Wiggins, Hamm, and Bell 1992). The executive branch includes a governor with full veto powers and a bureaucracy that operates according to a merit system which ranks in the 89th percentile of all state public administrations (Sigelman 1976). It is true that there has been a nationwide trend towards more powerful governors, Michigan has been far out in front of that trend. This is significant because governors in the American states have traditionally been very weak. This condition was a reaction to the harsh treatment experience by the colonists under British governors (Sabato 1983). Under the previous constitutions of 1835, 1850, and 1908, the governor did not have a four-year term like the national chief executive. It was the 1963 Constitution, which was crafted for the most part under the leadership of eventual Governor George Romney, that gave the governor a four-year term. Romney was responsible for the institution of a strong governorship in Michigan (Sabato 1983). Like the president, the governor of Michigan proposes the annual budget and while the president will have a line-item 111 veto starting in 1997, the governor of Michigan already exercises such power (Browne and VerBurg 1995). Unlike the president, however, who is the only nationally elected executive, the governor of Michigan must compete with other state-wide elected executives such as the Attorney General and the Secretary of State. These executives have independent bases of power and can use this power to "critique or embarrass the governor" (Browne and VerBurg 1995, 77). The governor does have the power to appoint the top members of the bureaucracy but this power is overrated (Haas 1989). The more important power, the power to fire, is more restricted. This is especially true in the case of Michigan. Many states have the rules for hiring and firing in statutes. Michigan, with its moralistic political culture, has these rules embedded in the Constitution. This reduces the flexiblity of the governor to act promptly and in the self-interest of the governorship (Beyle 1989). The governors during this time series of the study in this chapter, 1962-1994, can be found in Appendix 5-2. John B. Swainson was a strong- minded governor who tried to carry on the tradition of liberal social reform started by the previous governor, G. Mennen "Soapy" Williams. George Romney was a reformer of the process of governing and aimed at building consensus instead of creating a polarized situation. He inherited the tradition of the progressive reformers of the turn of the century. William G. Milliken followed in the footsteps of Romney in promoting an active state government and a neutral competent state bureaucracy. James G. Blanchard 112 focused on how the government could promote Michigan businesses and John Engler is a promoter of leaner government and lower taxes. The governor plays an important role in setting the tone for Michigan policy because the legislature habitually defers much of the policy leadership to the governor, and until 1992 governors in Michigan had no term limit and since the early 19608 have served multiple terms (Browne and Verburg 1995). This is in line with the national trend of governors acting as chief legislators (Beyle and Dalton 1983). The Michigan senate and the house of representatives have 38 and 110 members respectively. Each senator represents approximately 245,000 residents while each member of the house represents about 85,000. The Speaker of the Michigan House is the most powerful legislator in the Capitol. The majority leader of the senate does not wield as much power, "Because the senate has fewer members and senators serve longer terms, the majority leader is more subject to challenge" (Browne and VerBurg 1995, 113). Browne and Verburg also cite an anonymous former speaker who sums it up in the following manner, "The senate has 38 prima donnas, and the house, 109 sheep" (1995, 113). These similarities and differences extend into the context of health policy as well. While bicameralism has been an important factor in the legislative battles at the state level, there is little research documenting its effects. Michigan is a particularly important state to study because, compared to the 113 other states, Michigan's legislature is highly professional. A recent study places the Michigan Legislature as second only to New York in professionalization (Squire 1992).20 The legislators are well-paid, work in the legislature full-time, and have access to a wide variety of resources including significant staff support. Unlike many other states, Michigan, in having these characteristics, is similar to the national government. Therefore, for the bicameral model, we would expect it to act in ways that were expected of national government. The anticipations are different with respect to tricameralism in Michigan. The explanation for the behavior of the national tricameral model was the Senate filibuster. Michigan, like all other states, does not have the institution of a filibuster. We would expect, therefore, that any variables interacting with a unified government to be positively associated with relative change instead of the negative association found at the national level. 20 Squire (1992) also found that professionalization is related to a smaller than average percentage of women in the legislature. Rule (1990) claculated the national average for women as a percentage of state legislatures as 5.7 in 1974 and 12.1 in 1984. Michigan had figures of 4.0 and 9.0 for the respective years. The highest was New Hampshire with 20.0 and 27.0. The lowest was Alabama with 1.0 and 8.0. Rule's data show that New England Republican states have the highest percentage of women legislators while Southern Democrat states have the lowest. 114 Bicameral Model As was the case with the national bicameral model, the state bicameral model will consist of the four explanatory variables (occupation, age, seniority, and retention) and the dependent variable, relative change. According to the bicameral theory as used by Publius, the differences and similarities between the two bodies animate the institution. The more the chambers become alike, the more policy change (instability) will occur. The more the chambers become unlike each other, the less policy change will occur. There is a difference in the scope of the time series between the national and state models. Information on the dependent variable could only be collected from 1962 to the present. The time series for the Michigan models, both bicameral and tricameral, is from 1962 to 1994. Independent Variables Occupation. According to the theory presented in The Federalist Papers, lawyers should make up a greater portion of the chamber in the senate than in the house.21 By examining Appendix 5-3, we can see that lawyers have consistently been a greater percentage of the body in the senate than they have been in the house. At the end of the current time-series, however, the trend appears to be changing. More data points need to be 21 The figures for Michigan may appear lower than in other states because of the influence of the moralistic political culture in Michigan. Attorneys tend to dominate in states with a more traditional political culture (Baker 1990). 115 added to this trend in order to determine if the trend will be sustained. In the overall picture, the composition of the bodies is along the lines predicted by Publius. Age. In Appendix 5-4, we see that the mean age in the Michigan Senate and in the Michigan House is non-distinguishable. The mean age for both chambers has rarely been above 50 years and has never been below 40. This explanation for this lies in the fact that, unlike the Congress, there is no difference in the age requirements for Michigan senators and representatives. In order to be eligible to serve in either chamber of the Michigan legislature a person must be at least 21 years of age. Turnover and Seniority. Differences in the terms is another way to make sure that differences in compositional characteristics will occur between an upper and lower chamber. In the Congress this is accomplished by the Senate having only one-third of its membership up for reelection every two years while the entire House is up for reelection during the same period. In Michigan this difference is that the entire house is up for reelection every two years, but the entire senate is up for reelection every four years. The year in which the entire senate, along with the house, is up for reelection coincides with the gubernatorial election. The first measurement dealing with terms will be percent of the membership of the respective chamber that is retained from the previous legislature. As we can see from examining Appendix 5-5 that the expected 116 pattern develops. In years in which the Senate is up for reelection along with the House, the House retains a higher percentage of its members. In years in which the Senate is not up for reelection, tautologically, the House retains less. The second measurement, which is just a long-run variable of the first, is the seniority in each body defined as the mean number of years of service by members in each body. From Appendix 5-6, we can see that there is no clear pattern of dominance by either chamber in terms of the mean number of years served. There is a slight overall trend upward for both chambers, but, in terms of the difference, no trend emerges. The collection of these variables was time-consuming. These figures are not published in a summary fashion in any almanac. To collect these variables I examined the biographies of all members for each legislature in the Michigan Manual from 1962 to 1994. Dependent Variable Relative Change. For the national level variable, the source was the Congressional Quarterly Almanac. There is no state level version of the almanac and I was forced to seek another journalistic source for the construction of the variable. The nearest thing to CQA in Michigan is The Michigan Report by the Gongwer News Service. The Michigan Report is a daily newsletter that covers issues of import to state legislators, staff, and those in the executive branch. The first few pages of each daily newsletter 117 are taken up with headline stories. If a bill dealing with a health issue was the subject of a headline, it was counted as a major bill and included in the sample. Like the independent variables, the construction of this variable was tedious. The Michigan Report is a daily newsletter and there is no subject index until 1993. This means that every single page of the Michigan Report from 1962 to 1992 was examined and the index was used to locate health stories in the 1993 and 1994 editions. I examined approximately 85,100 pages of the Michigan Report, one page at a time. The pool of bills that were considered major by The Michigan Report was further classified as either "negative" or "positive." If the policy proposal was approved by the house, senate, and the Governor, thus becoming a public law, it was counted as "positive." If it failed to become a public law it was coded as "negative." For each two-year legislature, the total bills were pooled together and the sum of the bills was taken as the relative change figure for that legislature. An example is the legislature for the years 1973-1974 (Appendix 5-7). Each bill that was included in the sample for those years was coded as "+ 1" if it became a public law and "-1" if they did not. The sum for the sample year is "+9." This is the change that occurred relative to what was proposed in the manner of the national variable. A complete list of those policy proposals included in this series can be found in Appendix 5-8 and a table of 118 the relative values in Appendix 5-9.22 One difference between the state level and national level variables is the inclusion of non-fiscal proposals at the state level. Regulatory policy in health has traditionally been more important at the state level than at the national level (Levine 1984). In fact, regulatory policy is the main duty of the state in this policy area. Regulation is what states "do." States have had the more important role in the development and administration of health policy (Lipson 1991). Altman claims that, "the extent of state regulation is enormous" (Altman 1994, 97). This role includes insurance, rate structures, licensing, legal liability, experiments, hospitals, and employment (Biles 1980, Esposito 1982, Holoweiko 1988, Levine 1984, Merrill 1987, Thomas 1986). Since regulation sets the rules for every other aspect of health policy it can be argued that the state role is primary (Levine 1984). The national level regulatory concerns, on the other hand, are limited to product control under the jurisdiction of the Food and Drug Administration and financing. Even though the impact of federal finance is undeniable, it 22 The years 1962, 1963, and 1964 are, for the purposes of this analysis, counted as their own legislatures. The reason for this is that, prior to 1965, a bill had to become a public law in the same year it was introduced or it died and had to be reintroduced the next year even if it was in the same legislature. After 1965, a bill needed to become law within the same two year legislature or it died. Thus, 1965-1966 and all subsequent two year groupings after that are counted as one data point while 1962, 1963, and 1964 each have their own data point. 119 can still be argued that the state regulatory role is primary due to the state's police power (Brecher 1990, Levine 1984). This is why the non-fiscal policy proposals are not only included in the Michigan sample, they clearly dominate. Given the non-fiscal dimension of these proposals, each is given equal weight. Analysis The functional relationship of this stability model can be expressed as: Relative Change = f(Occupation, Age, Retention, Seniority). This relationship can be expressed more directly as: Relative Change = f(Senate Occupation - House Occupation, Senate Age - House Age, Senate Retention - House Retention , Senate Seniority - House Seniority). The linear expression of this model, therefore, would be: Relative Change = B0 + Bl(S Occup - H Occup) + Bg(S Age - H Age) + B3(S Ret - H Ret) + B4(S Sen - H Sen). The results of the time series analysis are in Appendix 5-10. As we can see the results for the state bicameral model are much stronger than the national bicameral model. The overall statistics are solid. The R2 is a respectable 0.57 and the F-statistic is significant at the 98% level of probability. In addition, three of the four variables are statistically significant and the one that is not approaches significance. The diagnostic tests that were conducted on the model revealed a problem with Autoregressive Conditional Heteroskedacticity(Appendix 5-11). This 120 violation renders the results somewhat suspect. Since I will be adding other states and policy areas to this model in the future, this problem, if it persists after then, will force a serious reevaluation of this theory. The strongest variable by far is age. The sign is in the predicted direction. For every unit of difference between the two bodies in terms of mean age, the relative change reacts by registering a loss of nearly three units. This is particularly striking because the difference in age between the two bodies is so small. The dependent variable, therefore, is extremely sensitive to any difference in age.23 The other two significant variables, occupation and retention, had signs which were the opposite of what was predicted. One reason why this variable may not have performed well is that 100 percent of the senate is retaine every two years. A possible consolation is that the substantive effects of these variablee are quite small. The seniority variable has the correct sign and a realtively large coefficient but fell short of accepted levels of statistical significance. It is encouraging that the two largect coefficients each had the correct sign. Perhaps if the the time series could have been extended, a better result may have been obtained. The next step was to analyze relative change according to a tricameral 23 It could be the peculiar configuration of the data points in the 32 year time series but a statistical test is designed to detect effects which are merely random. 121 model. After the presentation of these results, there will be a general discussion of the two models and the conclusions that can be drawn from comparing them withthe models at the national level. Tricameral Model Independent Variables Seat Change. The operational definition of this variable, as was the case with the national level variable, is the result of a three-stage process (see national Seat Change Variable) . Remember that a positive change is in favor of the governor's party and a negative change is in favor of the party not holding the governorship. The variable is displayed in Appendix 5-12. As we can see, the largest shifts occurred following the 1964, 1966, and 1974 elections. Compared to the national level variable, the Michigan seat change variable is quite stable. All things being held equal, we should expect more change to occur as this variable increases in value. In addition, because Michigan does not have the institution of the filibuster, we would expect this trend to be even stronger in unified governments. Gubernatorial Popularity. The operational definition of this variable is the percent of the popular vote received in the previous election (Appendix 5- 13). Ceteris paribus we would expect more change to occur as the variable increases. Party Strength. In a divided government, weak parties may allow for 122 change to occur and strong parties may lead to gridlock. In a unified government, on the other hand, weak parties may inhibit change and strong parties encourage it. At the national level the operational definition of this variable was the mean of the percent of party unity votes in the two chambers in a given Congress under a condition of unified government. This data, however, does not exist at the state level and the cost of collecting it from original archival resources at this time is prohibitive. A proxy measure for party strength at the state level will be the percentage of seats the majority party controls in the legislature. As we can see from Appendix 5-14, the main periods of strength for this variable were the early 19608 and the 19708. Analysis The functional relationship of this model can be expressed as: Relative Change = f(Seat Change, Seat Change * Unified Gov 't, Gubernatorial Popularity, Party Strength, Party Strength * Unified Government). The linear expression of this model, therefore, would be: Relative Change = B0 + Bl(Seat Change) + BZ(Seat Change * Unified) + B4(Gubernatorial Popularity) + the 85(Party Strength) + B6 (Party Strength * Unified). The results of the regression analysis are in Appendix 5-15. As we can see, none of the variables registered a significant effect. In addition, the overall measures of the statistical model are not very strong. The R2 is a discouraging 0.25 and the F-statistic was 0.78 which is statistically insignificant given the number 123 of variables and the number of cases. No violations of any of the classical assumptions of regression analysis were found in the model. The results of the diagnostic tests can in Appendix 5-16. Three out of the five variables have the predicted sign. Gubernatorial Popularity, Party, and the interactive term of Party with Unified Government. The strongest of the three in terms of the coeficient is Party Strength. The other two have coefficients that were quite weak. The remaining two variables had coefficients with signs that were the opposite of what was predicted. The interactive term of Seat Change with Unified Government was by far the strongest of the two in terms of strength even though both had t-statistics that were insignificant. The Seat Change variable has a coefficient with a negative 0.29 value. Conclusion At the state level the bicameral model outperforms the tricameral model with the understanding that there was an ARCH violation which renders these results somewhat suspect. At the national level, on the other hand, the tricameral model is the best performer. In both cases, the best model is the only model with significant coefficients. In terms of the external validity of each theory regarding the policymaking system, there are many problems with these models using Michigan as a testing context. What could be a possible explanation for such strange results? A possible explanation is 124 that the bicameral model outlined by Publius was intended to work in a Congress that had 67 members of the House and 26 members of the Senate not the 435 and 100, respectively, that we have today. Publius was open about the expectation that other states would be added which would increase the size of Congress. But Publius was concerned that a point would be reached that would change the way they thought the chambers would operate. There was special concern for the House. Publius was worried that the House would reach a point where they would be so numerous they would overwhelm the Senate's ability to deal with legislation. It is in Federalist 58 that Publius proclaims what came to be known as the iron law of oligarchy, "It is that in all legislative assemblies the greater the number composing them may be, the fewer will be the men who will in fact direct their proceedings" (Publius [1787-17 88] 1961, 360). The tricameral model works at the national level because it takes into account official partisanship which did not exist at the time of the Framing and the broader institutional factors which can guide larger and more complex entities. The bicameral model, on the other band, could work better in Michigan than at the national level because the Michigan legislature with its 38 senators and 110 house members is more like the Congress at the time of the Framing than the national legislature. These explanations are initial and are by no means definitive. More tests in different contexts and states will need to be conducted in order to come to a firmer conclusion. It is, 125 however, a start and an intriguing one at that. A complete list of the size of the lower and upper legislative chambers in the states is provided in Table 5- 6. Perhaps the most compelling reason for the difference, however, is the nature of the dependent variable. The national level variable has a fiscal focus and the state level variable has a regulatory focus. There is no easy way to solve this dilemma. The national and state governments have different roles. Perhaps comparison is out of the question and we need to analyze them each on their own terms. Summary Michigan is a suitable context for the external test of the national models because of the great diversity of Michigan. It is a competitive two- party state. Michigan has a professional legislature and a professional executive branch. The influence of the moralistic political culture made Michigan a leading innovator in the Progressive Era. Accordingly, the executive branch includes a governor with full veto powers and a bureaucracy that operates according to a merit system. Michigan does, however, differ from the national system becuase of the lack of a filibuster in the upper chamber. It was expected, therefore, that the results in Michigan would be different. Much of the data for this chapter, like much of the data at the national level, was original archival data collected specifically for this study. This originality allowed for the exploration of previously unanswered questions. 126 In fact, the "answers" raised more questions than had existed before. The results for the state bicameral model were much stronger than the national bicameral model. The strongest variable by far was age. None of the variables in the state tricameral model registered a significant effect. The reasons for this may have to do with the relative sizes of each legislature. The Michigan legislature is more like the Congress of 1789 in terms of size than the modern Congress. In all fairness, however, there are other states with legislatures closer in size than Michigan's. This might be a fruitful area of future comparative state analysis. CHAPTER SD(: CONCLUSION The perception that our policymaking system is in a state of "gridlock" has been a popular one among citizens, the popular press, and political scientists for quite some time. The blame for this perceived condition has been squarely placed on the separation of powers instituted by the framers of the Constitution. This separation makes it possible for different political parties to control different branches of the government at the same time. Recent scholarship, however, has begun to challenge this perception (Mayhew 1991). Mayhew, in a longitudinal analysis of legislative production in the postwar period, found that major legislation was produced at the same rate regardless of whether the government was divided or unified. Other studies, on the other hand, have found conditions in which divided governments have been more productive than unified ones and vice versa (Alt and Lowry 1994, Krehbiel 1993, Rohde 1991, Shepsle 1983). Since there is such serious disagreement between scholars concerning the prevalence and cause of gridlock a fresh look at this problem fi'om a different perspective could make a valuable contribution to the discussion. This is especially true since the existence of policy gridlock has been the source for serious suggestions that the Constitution be revised to eliminate the potential for divided government (Morone 1990, Sundquist 1992, Wilson 1885). Whether the goal is to attack or defend the separation of powers as 127 128 framed in the 17 87 Convention, one must understand the motives and intentions of the Framers. It is this spirit that animated this dissertation. At the time of the Convention of 1787, every free government that had come into being had disintegrated. Madison studied the histories of previous free governments and noticed a pattern that might account for the tragic state of free government in the world. Madison determined that instability was the main cause of the demise of previous democracies. The remedies for the disease of instability that could affect democracy in America would be Republic and Union (Diamond 1987). Concerning the Union, it was important for the states to be united to avoid the petty squabbling and the rampant discord of the patchwork Greek states. Concerning the Republic, the system must be one of separated powers in order to limit an arbitrariness in governmental rule, whether by the masses or an autocrat. The Union would be accomplished by the centralized power of the federal government and that centralized power would be made republican through the separation of powers and its key component, bicameralism. Thus, Madison devised the remedy for the diseases of free government. There are many potential ways to cure the diseases of democracy. Most of them do not involve anything approaching democratic methods. The fact that the Framers attempted the application of the separation of powers doctrine in a way completely untested is an indication of their political and intellectual courage. The claim that democracy was 129 consistent with reason and good government was, at that time, a very radical claim. The logic of Publius is as consistent as human logic can be. Whether it is true or not is a matter for another time. The question at hand, however, is not whether it was right to establish the separation of powers but to determine some of the effects since its establishment. While the attempt to analyze a phenomenon of such scope may appear quite brazen, arguments for and against the separation of powers are made as a result of broad perceptions of the system's performance and, therefore, a broad empirical analysis is in order so that we might evaluate these claims on their own terms. Health legislation is the context for this study because it is currently one of the most important areas of domestic policy and, in addition to its importance now, it has been prominent in the entire period since the Second World War. The question of the proper role of the government in financing and provision of health services goes right to the heart of questions concerning what we were intended to be as a nation by those who constructed the system. The research questions being addressed in this dissertation are the following: When will changes in health policy occur and what factors in the bicameral and tricameral system make change more or less likely? The In addition to an analysis of the factors in a bicameral or tricameral system that allow for change to occur, another unique feature of this dissertation is the direct comparison of the performance of the separation of powers on the 130 national level with that of the state level. National Bicameral Model Two national-level models are employed in this dissertation; bicameral and tricameral. The dependent variable is the same as in the national bicameral model. The independent variables, however, are different. Most of the research on bicameralism to date has taken the form of describing the differences between the two bodies. What is missing is an analysis of the impact of bicameralism on the legislative process. It is important that this impact be studied given the central role of bicameralism in American government. While there is no sustained empirical treatment of bicameralism as it affects the policy process, there is a body of literature which deals with the formal characteristics of bicameralism as an institution and the effect it has on other institutions in a formal constitutional system. This body of literature takes what has come to be known as a Public Choice perspective. It is the general position of the authors in this body of literature that the structure of the institutions has a significant effect on behavior (Brams 1989, Hammond and Miller 1987, Petracca 1989, Wirls 1993) . Brams (1989) is concerned with determining which chamber is more powerful. Hammond and Miller (1987) focus on the formal properties that make stability possible. Petracca (1989) argues that an analysis of bicameralism can suffer from an approach that is too minute and formal. Wirls (1993) attempts to determine which branch is the more deliberative and responsible. 131 The previous work provides some initial clues but an empirical examination must be attempted to enter the fray of policy relevant research. Other studies that attempt to define a dependent variable that measures systemic policy production according to changes in "macro" variables include Freedman (1995) and Mayhew (1991). Mayhew found that policy proposals became law at essentially the same rate regardless of whether the government was unified or divided. Like Mayhew, Freedman found that there was no essential difference between the number of historic laws passed under unified regimes and that of divided regimes. The scale for Mayhew's dependent variable is nominal; either the bill was important or not important. Freedman's scale is ordinal; trivial, non-substantive, minor, major, historic. The variable in this study is a ratio-level measure that allows the analyst to not only say a bill is more important than another but also by how much. The relative change variable yields more information for less work. Thus, it is more efficient. Another major difference between the dependent variable used in this study and the variables used in the Freedman and Mayhew studies is the inclusion of bills which did not become law. By including only hills which became law Mayhew and Freedman are trying to explain a phenomenon by using only the successful cases. There may be as much to learn, if not more, from what failed (Schick 1995). The independent variables in the bicameral model are Occupation, Age, Seniority, and Retention. The differences between the two chambers 132 according to the aforementioned characteristics were supposed to provide the requisite stability necessary for a republic to survive. Publius intended for the Senate to be the more stable body. The requirements for eligibility for the Senate would seem to make it more likely that persons with occupations of wealth and standing would comprise a greater proportion of that body than in the House. In order to be a member of the House of Representatives, an individual must be 25 years of age when seated. The Senate's requirement is that the citizen be 30 years of age when seated. This is consistent with Publius' desire to have the Senate be the more rational body. Terms were intended to be staggered so that some stability in government would be maintained and balanced with the democratic need to have the government remain close to the people. This was intended to leave the two chambers with different levels of seniority and retention. This would have a policy impact by allowing change but the change would happen slowly enough to make it more probable that the change would be for the better. The results of the time-series linear regression for the bicameral model, however, found that none of the variables registers enough predictive power to reject the null hypothesis that there is no relationship. It happened that the bicameral model was not strong because of the failure to take into consideration the characteristics of the president and partisan control of the branches of government. 133 National Tricameral Model There have been several studies as of late on the tricameral system. These include Alt and Lowry (1994), Cox and Kernell( 1991),Fiorina (1992), Freedman (1995), Krehbiel (1993), Mayhew (1991), Peterson (1990). Cox and Kernell (1991) and Peterson (1990) examine the tricameral system in order to determine which branch is more powerful in the policymaking process. Fiorina (1992) focuses on the causes and condition of divided government at the national and state level. Krehbiel (1993) analyzes the policy effects of the system fiom a more formal and theoretical level. The focus of the analysis in Alt and Lowry (1994) is state level budget and fiscal policy. All of these studies are important. This study, however, adds to this body of literature by looking at a measure of individual policy proposals, weighing those proposals according to an objective measure of importance, evaluating the production of policy instead of the relative strengths, and focusing not on the causes but on the potential effects of divided government. The independent variables in the tricameral model are a dichotomous Unified / Divided Government variable, Seat Change, an interactive variable of Seat Change with Unified / Divided Government, Presidential Popularity, an interactive variable of Presidential Popularity with Unified/ Divided Government, and an interactive variable of Party Strength with Unified / Divided Government, and Landslide. The overall quality of the model is strong. The strongest performing coefficient is Landslide. The next 134 best performing variable is Seat Change. The interactive term of Seat Change with Unified Government is statistically significant at the 90% level of probability in a two-tail test. The coefficient, however, has the wrong sign. The coefficient indicates that, in unified governments only, a one unit increase in mean bicameral seat change, will cause there to be a 0.78 unit decrease in relative change. This is non-intuitive because we expected the positive effect of positive mean bicameral seat change to become even stronger in a unified government. Case Studies An aggregated statistical analysis can reveal patterns but a disaggregated case analysis affords the opportunity to examine factors that would slip through the statistical cracks. This does not mean, however, that the analysis is free from the need to be systematic (King, Keohane, and Verba 1994). The case study offered confirmation of the observation that gridlock can occur in unified governments as well. The cases are Medicare and the Clinton Administration's Health Security Act. The variables included in the case study are the following: electoral context, presidential support, conference committee, filibuster, veto, domestic context, international context, interest groups, multiple referrals, early or late introduction into Congress. The variables that were associated with the fate of the two bills in the expected "direction" were: electoral context, filibuster, veto, domestic context, multiple referrals, and early or late introduction. The variables that did not 135 perform in the expected was the international context. Variables that displayed inconclusive results were interest groups and conference committees. The variable which was not a factor at all in this analysis was the level of presidential support. Medicare became the law of the land and was superior in the following categories: better electoral context, no filibuster threat, a less distracting domestic context, the significant committees in each chamber had sole jurisdiction over the bill, and was introduced early in the term. In the case of international context, the bill that faced the most danger passed shedding some doubt, in this analysis at least, on the linkage between domestic and foreign policies. State Models The experience of Michigan, as a laboratory of democracy, may yield insights for the study of bicameral and tricameral systems. The diversity of Michigan mirrors national diversity in many ways. Michigan has two competing political cultures. Michigan is a competitive two-party state. Michigan has a professional legislature and a professional executive branch. As was the case with the national bicameral model, the state bicameral model consists of the four explanatory variables (occupation, age, seniority, and retention) and the dependent variable, relative change. The results of the time series analysis for the state bicameral model are much stronger than the national bicameral model. The strongest variable by far is age. The other two significant variables, occupation and 136 retention, had signs that were the opposite of what was predicted. One reason why this variable may not have performed well is that 100 percent of the senate is retained every two years. A possible consolation is that the substantive effects of these variables are quite small. The seniority variable has the correct sign and a relatively large coefficient but fell short of accepted levels of statistical significance. The next step was to analyze relative change according to a tricameral model. The dependent variable was the same but the independent variables were different: Seat Change, an interactive variable of Seat Change with Unified Gov't, Gubernatorial Popularity, Party Strength, an interactive variable of Party Strength with Unified Government. None of the variables registered a significant effect. The tricameral model works at the national level because it takes into account official partisanship which did not exist at the time of the Framing and the broader institutional factors which can guide larger and more complex entities. The bicameral model, on the other hand, could work better in Michigan than at the national level because the Michigan legislature with its 38 senators and 110 house members is more like the Congress at the time of the Framing than the national legislature. Contributions to the Field The construction of a unique measure of policy change, an explicit empirical test of the Framer's structural model, a direct comparison of demographic versus institutional models, a direct comparison of national and 137 state behavior, and a demonstration of the relevance of case studies as a compliment to statistical analysis rank as the contributions to the field. The dependent variable in this study is the first of its kind to measure policy output on a ratio scale while accounting for passed and failed legislation. This is an originally constructed variable. While the sources for the national level variable have been used in the past, the application is certainly novel. At the state level, any similarity to other measures disappears. The depth of such a complex dependent variable allows for a much richer analysis than ordinal or nominal level measures and allows for the avoidance of important conceptual errors. A substantial portion of this dissertation also involves an explicit test of the structural theories of the Framers. What came to the surface as a result of the analysis in this study was a fuller appreciation of how the advent of parties and the modern executive have made the Framer's bicameral model irrelevant at the national level. Even more intriguing, however, were the sub-national results with respect to this question. At the state level, the bicameral model of the Framer's is stronger than the tricameral model. This is the case even though Michigan has had a history of strong governors and a competitive party system. An initial explanation of this fact is the differing scope of policy in the two systems. There may be an interactive effect of the size of the legislature at the national level and the scope of policy considered with political parties and the executive branch. 138 Another contribution to the field comes from the inclusion of case analysis with statistical analysis. Both specific case studies and aggregate statistical analysis have weaknesses in terms of properly describing a phenomenon. This study was an attempt to try to simultaneously provide a view of the "forest" while maintaining the perspective of the "trees." Future Research The potential fruitful areas of future research include: the addition of more policy areas, the addition of more states, and examination of other "Framing" documents such as Jefferson's Notes on the State of Virginia, and a closer examination of the institution of the filibuster. It may indeed be the case that the patterns that developed in this study are peculiar to health policy. An interesting and particularly arduous task, when one considers the inclusion of states, would be a study that would include all policy areas. There could be a great deal to learn from such an omni-policy study as was attempted by Mayhew (1991). Given the greater degree of sensitivity of the dependent variable in this study, however, it will be intriguing to see the results. In addition, The inclusion of more states would shed more light on the interesting results obtained in the Michigan models. This would allow the analyst to control for the variety of factors that may indeed be unique to the Great Lakes State. How would the inclusion of less professional states with larger or smaller legislatures affect the results? It might also be the fact the this study has focused too much attention 139 on the Federalist Papers. A strong case can be made that even though it was a political tract it contained wisdom of a theoretical nature. There are, however, other Framing documents that could analyzed in a similar fashion in order to better gauge the intentions and foresight of the leaders at the time. Jefferson's Notes and Hamilton's Report on Manufactures are just two examples. There is much to be learned here. It was a unique period in history when the slate was as clean as a slate can get in political life and those who held the chalk were simultaneously scholars and political leaders. Finally, there is much yet to be learned about the effects of the filibuster. The results of the analysis in this study give some credence to the suspicion that the supermajoritarian aspects of the Senate are severely distorting the republican qualities of the union constructed over 200 years ago. The filibuster was not an explicit or implicit creation of the Framers. Referring to the metaphor that is the title of this study, the right institutions can provide the vaccine that will make it more likely that the better angels of our nature will win in the end. But some vaccines can be fatal. The institution of the filibuster is an inoculation that is potentially too strong, thus, allowing the disease to take over. APPENDICES 140 Appendix 2-1 Banzhaf Voting Power Actor Individual Collectivity Individual Collectivity __ ordinary ordinanc amendment amendmenL House 0.00146 0.63316 0.00185 0.80471 Senate 0.00329 0.32881 0.00195 0.19529 * This table is an adaptation of Tables 8.1 and 8.2 in (Brams 1989, 134-136) Appendix 2-2 Relative Banzhaf Power ACTOR WE 2W WENT House 76.4 66.2 83.0 Senate 60.7 45.0 27.0 * This table is an adaptation of Tables 8.1 and 8.2 in (Brams 1989, 134-136) * %PASS OTHER = The percent of legislation that originates in the chamber that is subsequently passed by the other chamber * %VETO OVERRIDE = The percent of veto overrides that originate in this chamber that are also overriden in the other chamber. * %AMENDMENT = The percent of Constitutional amendments originated in this chamber that are subsequently passed in the other chamber Appendix 2-3 Alternate Regression Models Bic rlM l 141 VARIABLE COEFFICIENT STD ERROR t-VALUE Constant 1.7930633 occupati -.2740070 age .4679250 seniorit -2.2312260 retentio . 1673642 R2 = ,3222945 0 = 3.8335753 F( 4, 16) = TricameraLModel 6.46572 .27732 .21939 -1.24895 .94345 .49597 1.18819 -1.87783 .23774 .70397 1.90 [ .1592] DW = 2.592 VARIABLE COEFFICIENT STD ERROR t-VALUE CONSTANT -5. 1527068 unified 3.8420187 seatchg .3294477 seat*uni -.7648123 pres .0573663 pres*uni -.0496219 landslid 13.3314865 R2 = .6629994 0 2.9651952 F( 6, 16) = 2.93005 -1.75857 11.74341 .32716 .16573 1.98786 .24647 -3.10310 .05566 1.03067 .22971 -.21602 4.11353 3.24089 5.25 [.0037] DW = 1.828 Appendix 2-4 National Relative Change Year 1945 1947 1949 1951 1953 1955 1957 1959 1961 1963 1965 1967 1969 1971 1973 1975 1977 1979 1981 1983 1985 1987 1989 1991 1993 Value 4164 4185 4156 4106 4106 4185 (100 4L96 +4107 4L13 44104 44118 4122 4124 4L82 4154 +J”16 {L96 44164 4110 £156 1169 4185 «455 4106 142 143 Appendix 2-5 List of Bills 4 - h on 8 DC Hospital Center +0.16, Hill-Burton +0.35, National Mental Health Program -.07, Cancer Research -0.47, National Health Insurance -7.06, Maternal and Child Welfare -0.47, Social Security Act Amendments -1.082 -8.642 1 4 - 4 National Health Insurance -7.06, Dental Institute -0.012, Cancer Research +0.065, School Health +0.07, Water Pollution Control +0.09, Hill-Burton +0.0 -6.847 4 - 1 n 8 National Health Insurance -7.06, Voluntary Prepayment -2.18, Child Health Plan -.133, Medical Education -0.19 -9.563 1951-1952 82nd Congress Public Health Centers -0.06, Total -0.06 1 - 4 r n e Health Reinsurance -0.06 Total -0.06 - 4 n e Polio Vaccine +0.09, Polio Vaccine II +0.03, National Health Insurance -7.06, Research Facilities +0.07, Alaska Mental Health +0.02 -6.85 W Total +0.00 WWW Federal Health Program +0.23, Health for Peace +0.017, Medicare -2.21 -1.963 W Medical Training -0.24, Freedman's Hospital +0.02, Medicare -2.03, Water Pollution +0.232 +0.07 1 -1 4 11 Medical Training +0.12, Mental Health +0.21, Medicare -1.60, Nurse Training +0.09, Veteran's Nursing Care +0.05 -1.13 W 144 Appendix 2-5 cont. Medicare +9.68, Regional Health +0.168, Mental Health Staffing +0.06, Medical Professions Training +0.238, State Health Services +0.10, Hospital Modernization -1.33, Allied Health Professions +0.11, International Health +0.01 +9.036 -1 88 Partnership for Health +0.25, Rat Control -0.03, Alcoholism -0.03, Mental Retardation +0.06, Hospital Modernization -0.07, Eye Institute +0.0 +0.18 1969-19le 91st Congress Drug Abuse +0.13, Alcoholism +0.10, Communicable Diseases +0.08, Family Practice +0.08, Family Planning +0.13, National Health Insurance -2.23, Physicians and Pharmeceuticals -0.54, Migrant Workers' Health +0.03 -2.22 W Black Lung Disease 078, Cancer Research +0.50, Dental Health 005, Drug Abuse -0.32, Drug Listing -0.00, Veteran's Drug Treatment -0.02, National Health Insurance -2.23, Sickle Cell Anemia +0.03, Veteran's Medical Care +0.06, Safe Drinking Water +0.04, Heart and Lung Disease+0.4, HMOs+1.5, Emergency Medical Care -0.07, Cooley's Anemia +0.00, Mental Health -0.30 -1.24 1 -1 74 r n HMOs +0.06, National Health Insurance -2.23, Child Abuse +0.02, Crib Death +0.00 Drinking Water +0.04, Diabetes +0.01, Biomedical Reserach +0.15, Nurses' Training -0.16, Indian Health Care +0.228, Arthritis +0.01 -1.82 W Health Services Training +0.47, Drug Abuse Prevention -0.12, Disease Control -0.10, Health Care for the Unemployed -0.56, National Health Insurance -4.07 4, Health Manpower +0.44, Disease Control +0.06, Preliminary National Health Insurance +0.03, Health Block Grants -5.9, Indian Health Care +0.21 -9.54 mm Clinical Laboratory Standards -0.01, Hospital Cost Control +0.034, Cancer Research +0.63, Family Planning +0.13, Health Services, Centers +0.45, Medicare, Medicaid -0.07 +1.164 WWW 145 Appendix 2-5 cont. Asbestos +0.03, Child Health Assurance -0.82, Hospital Cost Control +1.78, Vietnam Vets Counseling +0.02, Military Doctor Bonuses +0.02, Medicare, Medicaid -0.17, National Health Insurance -2.47, Mental Health +0.09, National Health Insurance Start-Ups -0.44 -l.96 WW Health Manpower +0.10, Medicaid +0.36, Agent Orange +0.02, Health Program Spending Cuts +0.75, TEFRA +1.41, Orphan Drugs +0.00 +2.64 W DRGs +0.17, Drug, Alcohol Abuse +.03, Health Insurance for the Unemployed -0.40, Health Emergency Fund +0.01, Child Health Plan (CHAP) -0.06, Abortion - 0.23, Health Institutes and Professions -0.67, Organ Transplants +0.00, DEFRA +0.87 Baby Doe +0.18 -0.1 W Medicare Payment Freeze -0.93, COBRA, OBRA +0.26, Medical Abuse -0.87, AIDS -0.01, Bubonic Plague -0.01 -l.56 WW Catastrphic Costs +0.15, Indian Health Care +0.02, Infant Health -0.01 Long-Term Care 915, Vaccine Compensation +0.02, OBRA +0.62 Omnibus Health +0.668, Rights of Mentally Ill +0.00, Regional Trauma Centers -0.01 -7.692 W Catastrophic Coverage +0.14, Physician Payments +0.72, Physician Referrals +2.16 Medicaid +0.36, Minority Health +0.05, Trauma Care +0.01, AZT -0.01, Medicare +1.61, Medicaid Expansion +.04, FDA -0.02, AIDS +0.16, Disabilities Prevention +0.00, Anti-Smoking -0.02, Organ and Bone Marrow +0.00, NIH +0.01, Cancer +0.01 Access, Long-Term Care -12.07, -6.85 W Health Care Reform -5.29, Health Professionals +0.18, Disease Prevention +0.02, Disability Prevention +0.00, Alcohol, Drug Abuse, Mental Health 146 Appendix 2-5 cont. +0.56, Lead Poisoning -0.07, Cancer Registries +0.01, Medical Training +0.04 —4.55 W Health Care Reform -2.01, Health Care Budget +2.01, Disabilities Assistance +0.02 Disability Technology Grants -0.01, Organ Transplants -.07 -0.06 Appendix 2-6 National Occupation Year 1953-1954 1955-1956 1957-1958 1959-1960 1961-1962 1963-1964 1965-1966 1967-1968 1969-1970 1971-1972 1973-1974 1975-1976 1977-1978 1979-1980 1981-1982 1983-1984 1985-1986 1987-1988 1989-1990 1991-1992 1993-1994 56.78 56.32 53.79 55.63 56.09 57.47 56.78 56.55 55.63 54.25 50.80 50.80 51.03 47.13 44.60 45.98 43.68 43.30 42.30 42.30 41.61 147 61.46 62.50 61.46 63.54 63.00 66.00 67.00 68.00 68.00 65.00 68.00 67.00 68.00 65.00 59.00 61.00 61.00 62.00 63.00 61.00 57.00 Hausa Senate Difference. 4.68 6.18 7.67 7.91 6.91 8.53 10.22 11.45 12.73 10.75 17.20 16.20 16.97 17.87 14.40 15.02 17.32 19.00 20.70 18.70 15.39 148 Appendix 2-7 National Age Year ngse Senate Riflerence 1953-1954 52.00 56.60 4.60 1955-1956 51.40 57.20 5.80 1957-1958 52.96 57.90 4.94 1959-1960 51.70 57.10 5.40 1961-1962 51.00 59.90 8.90 1963-1964 51.70 56.80 5.10 1965-1966 50.50 57.70 7.20 1967-1968 50.80 57.70 6.90 1969-1970 52.20 56.60 4.40 1971-1972 52.24 57.10 4.56 1973-1974 51.10 55.30 4.20 1975-1976 49.80 55.50 5.70 1977-1978 49.30 54.70 5.70 1979-1980 49.01 53.00 3.99 1981-1982 48.20 51.80 3.60 1983-1984 48.57 53.72 5.15 1985-1986 49.44 54.90 5.50 1987-1988 50.25 54.30 4.05 1989-1990 51.14 55.50 4.36 1991-1992 52.16 56.20 4.04 1993-1994 50.77 56.86 6.09 149 Appendix 2-8 NatmnaLRetentien Year Hense Senate Difl‘enenm 1953-1954 82.30 85.00 2.70 1955-1956 88.05 86.00 2.05 1957-1958 90.11 90.00 0.11 1959-1960 83.22 84.00 0.78 1961-1962 87.13 94.00 6.87 1963-1964 86.67 90.00 3.33 1965-1966 80.23 93.00 12.77 1967-1968 83.68 93.00 9.32 1969-1970 91.72 86.00 5.72 1971-1972 88.28 89.00 0.72 1973-1974 85.06 87.00 1.94 1975-1976 79.08 89.00 9.92 1977-1978 85.52 83.00 2.52 1979-1980 83.22 80.00 3.22 1981-1982 83.68 82.00 1.68 1983-1984 81.84 95.00 13.16 1985-1986 90.57 93.0 2.43 1987-1988 89.43 87.00 2.43 1989-1990 93.10 90.00 3.10 1991-1992 89.89 98.00 8.10 1993-1994 74.71 90.00 15.29 Appendix 2-9 National Seniority Year 1953-1954 1955-1956 1957-1958 1959-1960 1961-1962 1963-1964 1965-1966 1967-1968 1969-1970 1971-1972 1973-1974 1975-1976 1977-1978 1979-1980 1981-1982 1983-1984 1985-1986 1987-1988 1989-1990 1991-1992 1993-1994 Renae 9.80 10.40 11.00 11.20 11.60 11.40 11.00 11.20 11.40 12.00 11.40 10.80 9.20 10.00 9.80 9.20 9.40 11.20 11.60 10.62 9.21 150 Senate 8.50 8.40 9.60 9.40 9.70 9.90 11.10 11.60 11.20 11.50 11.20 11.50 10.60 9.60 8.50 9.60 10.10 9.60 9.80 12.57 11.86 Difl’erence 1.30 2.00 1.40 1.80 1.90 1.50 0.10 0.60 0.20 0.50 0.20 0.70 1.40 0.40 1.30 0.40 0.70 1.60 1.80 1.95 2.65 151 Appendix 2-10 National Bicameral Model VARIABLE COEFF S. E t constant .93 6.23 .15 occupation -.28 .21 -1.30 age .42 .91 .47 seniority -1.77 1.17 -1.52 retention 17 .23 .75 R2 = .29 F( 4, 16) = 1.63 [.2142] 152 Appendix 2-11 National Bicameral Diagnostics Testing for Serical Correlation from Lags 1 to 2 Chi-Squared(2) = 1.619 F-Form (2, 14) = .58 [0.5702] Error Autocorrelation Coefficients: -0.3331 -0.0156 ARCH Test Residuals Scaled by 0.3772D +01 Constant Lag Coeff. 0.8578 -0.0746 S.E. 0.4198 0.2646 RSS = 0.37590D+02 delta = 1.63859 Chi-squared = 0.113 F(1,14) = 0.08 [0.7821] Chi-Squared Test for Normality 1.079 RESET F-Test for adding Yhat-squared F(1, 15) = 1.250 [0.2812] Ferrer-Glauber for Multicollinearity: no violations 153 Appendix 3-1 Unified Government he; n'fi r ivi ed? 1949- 1950 Unified 1951-1952 Unified 1953-1954 Unified 1955-1956 Divided 1957-1958 Divided 1959-1960 Divided 1961-1962 Unified 1963-1964 Unified 1965-1966 Unified 1967-1968 Unified 1969-1970 Divided 1971-197 2 Divided 1973-1974 Divided 1975-1976 Divided 1977-1978 Unified 197 9-1980 Unified 1981-1982 Divided 1983-1984 Divided 1985-1986 Divided 1987-1988 Divided 1989-1990 Divided 1991- 1992 Divided 1993-1994 Unified 154 Appendix 3—2 National Seat Change Year ic 1949-1950 13.10 1951-1952 -5.85 1953-1954 3.05 1955-1956 -2.55 1957-1958 -3.50 1959-1960 -11.90 1961-1962 -1.80 1963-1964 0.20 1965-1966 4.75 1967-1968 -7.50 1969-1970 4.05 1971-1972 -0.90 1973-1974 0.40 1975-1976 -8.00 1977-1978 0.60 1979-1980 -3.35 1981-1982 10.00 1983-1984 -2.60 1985-1986 1.45 1987-1988 -4.55 1989-1990 -0.35 1991-1992 -1.30 1993-1994 -0.55 155 Appendix 3-3 Regression results using separate measures VARIABLE COEFF. S.E. t CONSTANT -8.55 3.92 -2.17 s chg .30 .21 1.41 s chg*un -.25 .59 -.43 h chg -.06 .28 -.22 h chg*un -.25 .39 -.63 popvote .11 .07 1.53 pop*uni .02 .02 .96 landslid 11.91 4.02 2.96 R2 = .67 F = 4.35 [.0081] Appendix 3-4 Presidential Popularity Year 1949-1950 1951-1952 1953-1954 1955-1956 1957-1958 1959-1960 1961-1962 1963-1964 1965-1966 1967-1968 1969-1970 1971-1972 197 3-1974 197 5-1976 1977-1978 197 9-1980 1981-1982 1983-1984 1985-1986 1987-1988 1989-1990 1991-1992 1993-1994 156 Pereent Pepular Vete 49.6 49.6 55.1 55.1 57.4 57.4 49.7 49.7 61.1 61.1 43.4 43.4 60.7 0* (Ford was not elected) 50.1 50.1 50.7 50.7 58.8 58.8 53.4 53.4 43.0 157 Appendix 3-5 Regression results using less restrictive landslide VARIABLE COEFF. S.E. t CONSTANT -6.60 3.97 -1.66 unified -36.34 26.89 -1.35 party*un .21 .20 1.04 seatchg .27 .23 1.16 popvote .07 .08 .90 pop*uni .54 .36 1.49 schg*uni -.51 .32 -1.59 land 2 -.09 2.56 -.03 R2 = .46 F( 7, 15) = 1.85 [ .1502] 158 Appendix 3—6 Interaction of National Party Strength with Unified Year M 1949-1950 51.00 1951-1952 60.05 1953-1954 42.05 1955-1956 Divided 1957-1958 Divided 1959-1960 Divided 1961-1962 49.50 1963-1964 46.50 1965-1966 45.00 1967-1968 34.00 1969-1970 Divided 1971-1972 Divided 197 3-197 4 Divided 197 5-197 6 Divided 1977-1978 40.50 197 9-1980 44.50 1981-1982 Divided 1983-1984 Divided 1985-1986 Divided 1987-1988 Divided 1989-1990 Divided 1991-1992 Divided 1993-1994 63.15 159 Appendix 3-7 National Tricameral Model VARLABLE COEFF. S.E. CONSTANT -6.55 2.86 unified 26.49 28.28 seatchg .28 .16 schg*uni -.75 .24 popvote .08 .05 pop*uni -.36 .39 party*un -. 14 .20 landslid 15.18 4.96 R2 = .67 F( 7, 15) = 4.35 [.0081] t -2.29 .94 1.75 -3.08 1.41 -.93 -.70 3.06 160 Appendix 3—8 National Tricameral Diagnostics Testing for Serial Correlation from Lags 1 to 2 CHI2(2) = 2.662 F-Form(2, 12) = .79 [ .4781] Error Autocorrelation Coefficients: .2002 -.3810 ARCH Test Residuals Scaled by .2923D+01 CNST 1 LAG COEFF. .5121 .1650 S.E.'s .2716 .2896 RSS = .10515D+02 o = .93609 CHI'~’( 1): .580 F( 1, 12) = .32 [ .5793] Chi-Squared Test for Normality: CHI2(2) = .487 RESET F-Test for adding Yhat2 F( 1, 13) = .45Table 1.1 Regression Results: National Bicameral Ferrer-Glauber for multicollinearity: Unified Government is collinear with the interactive term of popular vote with unified government, the interactive term of party with unified, and landslide. The interactive term of popular vote with unified is collinear with the interactive term of party with unified government. 161 Appendix 3-9 Modified National Tricameral Model VARIABLE COEFF. S.E. Constant -5.67 2.87 landslid 12.60 3.42 seatchg .31 .15 schg*uni -.77 .23 popvote .06 .05 pop*uni .01 .03 R2 = .61 F( 5, 19) = 5.88 [ .0019] -l.97 3.69 2.15 -3.29 1.10 .40 162 Appendix 3-10 Modified National Tricameral Diagnostics Testing for Serial Correlation from Lags 1 to 2 CHI'*’( 2) = .375 and F-Form(2, 17):.13 [.8794] Error Autocorrelation Coefficients: .1107 -.0851 ARCH TEST Residuals Scaled by .3039D+01 CNST 1 LAG COEFF. .6330 .0019 S.E.'s .2061 .1742 RSS = .10054D+02 o = .76903 CHI”( 1): .000 with F( 1, 17) = .00 [ .9914] CHI-SQUARED Test for NORMALITY :CHI’(2) = .803 RESET F-TEST for adding Yhat’ F(1, 18) = .399 [.5355] Ferrer-Glauber Test for Multicollinearity: no violations 163 Appendix 3-11 Regression results with GDP VARIABLE COEFF. S. E. t CONSTANT -5.27 3.28 -1.60 seatchg .32 .15 2.07 schg*uni -.78 .24 -3.15 popvote .05 .05 1.05 pop*uni .01 .02 .25 landslid 12.60 3.50 3.59 gdp def -.53 1.99 -.27 R2 = .60 F( 6, 18) = 4.67 [.0050] 164 Appendix 4-1 Four Phases of Medicare Phase Years F tur r s l I 1935-1949 Wagner-Murray-Dingell bill II 1957-1960 Forand bill III 1961-1964 King-Anderson bill IV 1965 Mills's "Three-Layer Cake" 165 Appendix 4-2 Conferees 1965 s h 1Q 166 Appendix 4-3 Veto Threat p s h r ID Appendix 4-4 Phases of Health Security Act Bhaae I II III 167 YearsEeatnredJZreneeals 1971-1975 1978-1979 1993-1994 Kennedy-Griffiths (Canadian) Nixon (employer mandate) Kennedy (Canadian) Carter (catastrophic coverage) Clinton (alliances, employer mandate) McDermott-Wellstone (Canadian) Gramm-Armey (medical savings accounts) 168 Appendix 4-5 Ideological Distribution Ultra Center Center Ultra Canadian Universal 95% Managed Subsid. MSAs Coverage Coverage Compet. Insurance Wellstone Kennedy Mitchell Chafee Dole Gramm McDermott Dingell Moynihan Cooper Armey Waxman Rostenkow. Grandy Stark Gephardt 169 Appendix 4-6 Clinton’s Veto Threat p h s a Appendix 4—7 Comparison of Variables W Electoral Context Presidential Support Conference Committee Filibuster Veto Domestic Context International Context Interest Groups Multiple Referrals Introduced in Congress 170 19.6.5 MEDICARE Strong Yes Difficult No No No Distraction Strong Distraction One major proponent One major opponent No Early Weak Yes N/A Yes Yes Mild Distraction Mild Distraction Variety against different aspects Yes Late Appendix 5-1 Michigan Government EAR 1951-1952 1953-1954 1955-1956 1957-1958 1959-1960 1961-1962 1963-1964 1965-1966 1967-1968 1969-1970 1971-1972 1973-1974 1975- 1976 1977- 1978 1979-1980 1981-1982 1983-1984 1985-1986 1987-1988 1989-1990 1991-1992 1993-1994 GQXERNQR Democrat Democrat Democrat Democrat Democrat Democrat Republican Republican Republican Republican Republican Republican Republican Republican Republican Democrat Democrat Democrat Democrat Democrat Republican Republican 171 HQLLSE Republican Republican Republican Republican Tie Republican Republican Democrat Tie Democrat Democrat Democrat Democrat Democrat Democrat Democrat Democrat Democrat Democrat Democrat Democrat Tie SENATE Republican Republican Republican Republican Republican Republican Republican Democrat Republican Republican Tie Tie Democrat Democrat Democrat Democrat Democrat Republican Republican Republican Republican Republican 172 Appendix 5-2 Michigan Governors 1961-1962 1963-1969 1969-1982 1983-1990 1991- John B. Swainson George W. Romney William G. Milliken James J. Blanchard John M. Engler 173 Appendix 5-3 Michigan Occupation ha; ' r n ' cu 1962 16.33 1963 10.33 1964 10.33 1965-1966 12.67 1967-1968 2.00 1969-1970 1.67 1971-1972 7.33 1973-1974 12.33 1975-1976 12.67 1977-1978 8.67 1979-1980 1.00 1981-1982 1.33 1983-1984 8.33 1985-1986 13.33 1987-1988 7.33 1989-1990 4.67 1991-1992 1.33 1993-1994 6.00 174 Appendix 5—4 Michigan Age Year Differenceinnge 1962 2.67 1963 2.67 1964 2.67 1965-1966 2.67 1967-1968 0.67 1969-1970 2.00 1971-1972 0.67 1973-1974 3.33 1975-1976 0.67 1977-1978 1.33 1979-1980 2.00 1981-1982 0.67 1983-1984 1.33 1985-1986 2.67 1987-1988 1.33 1989-1990 1.33 1991-1992 2.00 1993-1994 4.00 175 Appendix 5-5 Michigan Retention Yeags Di er n ' n ion 1962 7.78 1963 3.33 1964 3.33 1965-1966 7.78 1967-1968 11.11 1969-1970 8.89 1971-1972 13.33 1973-1974 25.56 1975-1976 7.78 1977-1978 2.22 1979-1980 12.22 1981-1982 21.11 1983-1984 8.89 1985-1986 1.11 1987-1988 1.11 1989-1990 5.56 1991-1992 5.56 1993-1994 20.00 Appendix 545 Michigan Seniority Year 1962 1963 1964 1965-1966 1967-1968 1969-1970 1971-1972 1973-1974 1975-1976 1977-1978 197 9-1980 1981-1982 1983-1984 1985-1986 1987-1988 1989-1990 1991-1992 1993-1994 Dif r 2.00 0.44 0.44 0.89 0.89 0.11 0.11 1.78 0.44 0.89 1.11 0.67 0.33 1.22 1.00 0.67 2.11 1.22 i 176 i rit 177 Appendix 5-7 Sample Year 1973-197 4 Alcoholism Treatment +1 Abortion -1 Medical Referral Service -1 Generic Drugs +1 Hospital Bond Authority +1 Abortion II +1 HMOs +1 Mental Health Code +1 Alcoholism Treatment +1 Nursing Home Inspection +1 Outpatient Facilities +1 Paramedic Regs. +1 Physician Fraud +1 Sum = +9 178 Appendix 5-8 Michigan Bills 1962 taxes to Support Mentally Ill -1, Health Facilities -1, Community Health Centers -1, Pharmacists and Packaged Liquor -1, Old Age Assistance -1, Hospital Council -1, -6 1.9.6.3 Dental Care Corps. +1, Community Mental Health -1, 0 1.9.65 Paregoric +1, Blue Cross Regs. -1 0 rename Directors Actions Review -1, Podiatrists +1, Family Planning +1, Osteopathic School at MSU -1, Community Mental Health +1, Chiropractors -1, Blue Cross/Chiropractors -1, Medical Examiners -1, Medicare +1, Medicare II +1 0 1361-13518 Teaching Hospitals -1, Hospital Licensing +1, Medicaid +1, Board of Physicians -1, Ionia State Hospital +1, Basic Science Act -1, Abortion -1, Fluoridation +1, Sterilization -1, Kidney Research -1 -2 13624319 Abortion -1, Physician Licensing -1, Medical Evaluation -1, Organ Donors +1, Pharmacy Regs. +1, Hospital Finance Authority +1, Critical Health Problems +1, Blue Cross Regs. -1 0 -1 72 Alcohol Treatment -1, Abortion Law -1, Blue Cross -1, Drug Control +1, Certificate of Need +1, Drug Abuse Treatment +1 0 W Alcoholism Treatment +1, Abortion -1, Medical Referral Service -1, Generic Drugs +1, Hospital Bond Authority +1, Abortion II +1, HMOs +1, Mental Health Code +1, Alcoholism Treatment +1, Nursing Home Inspection +1, Outpatient Facilities +1, Paramedic Regs. +1, Physician Fraud +1 +9 1975-1976 Health Manpower +1, Definition of Death +1, Health Manpower II +1, 179 Appendix 5-8 cont. Medical Malpractice -1, Workmen's Comp. +1, Health Manpower III +1, Smoking Prohibition +1, Prescription Drugs +1, Riverside Center +1 Audiologist/Radiologist Regs. -1 +6 1911-1918 Blue Cross Regs. -1, Labeling Food Additives -1, Revise Public Health Code +1, PPB +1, Dentist Reg. +1, Health Manpower +1, Sex Education +1, Health Manpower +1, Single Business Tax and Hospitals +1, Mental Health -1, Nursing Home Regs. +1, Pharmacy Regs. +1, Abortion -1, Medical Fraud +1, Certificate of Need -1, Medical Malpractice +1, Hospital Bonds -1, Laetrile -1, Hospital Bonds +1, Sales Tax Exemptions +1 +6 1919:1989 Medical Treatment -1, Prescriptions -1, Health Manpower -1, Insurance Reform -1, Health Care Corps. +1, Hospice +1, University Hospitals +1, Medical Services +1, Certificate of Need +1, Prohibit Medicaid Abortions +1, Marijuana for Cancer +1, Smoking Regs. -1, Smoking Regs. II -1 +1 1991;1982 Sewage -1, Hospital Bond Regs. +1, Health Facilities +1, Abortion +1, Health Facilities II +1, Health Manpower -1, Mental Health +1, Health Manpower II +1, Substance Abuse\lnsurance Coverage +1, Medicare +1, Medicaid Fraud +1, Insurance Regs. +1, Group Insurance -1 +7 1353-] 354 Weight Loss Clinic Regs. -1, Cancer +1, Power of Attorney -1, Drug Abuse +1, Mental Health +1, Mental Health Youth Care +1, Health Services for Minors +1, PPOs +1, Medicaid Regulation -1, Diagnostic Drugs +1, Hospice Care +1, Medical Service Corps. +1, Medicaid Fraud +1, Smoking Prohibition +1, HMOs +1, HMOs 11 +1, Prohibit Medicaid Abortion -1, Imitation Drugs +1, Medical Malpractice -1, Coordination of Benefits +1, Health Corporation Regs. -1, Health Insurance Regs. +1 Health and Safety -1 +9 1W Prohibit Medical Abortion -1, Informed Abortion -1, Health Manpower +1, Health Manpower II +1, Malpractice Liability +1, AIDS -1, AIDS II +1, Health Manpower III +1, Insurance Regs. +1, Abortion II -1, Smoking +1, Smoking II +1, Health Manpower IV +1 +3 180 Appendix 5-8 cont. 1997;1999 AIDS +1, AIDS 11 +1, AIDS III +1, LTC +1, Drugs +1, Health Manpower +1, Chiropractors -1, LTC II -1, Health Manpower II +1, AIDS IV +1, AIDS V -1, AIDS VI +1, LTC III -1, Help for Working Poor -1, HIV Tests -1, Prescription Drugs -1, Licensing -1, Health Manpower III -1M, Prescriptions II +1 AIDS VII +1, Drug Abuse +1, AIDS VIII +1, Drug Abuse +1, Smoking -1, Smoking II +1, Pollution from Health Care Facilities -1, Health Manpower IV +1 5 1999:1999 Mental Health +1, Prohibit Abortion -1, Injury Research -1, Waste Disposal +1, Mental Health II -1, Abortion +1, Abortion II -1, Right to Die -1, Medicare -1, Insurance -1, Environmental Protection +1, Mental Health III -1, Diseases +1, AIDS +1, AIDS II -1, Smoking -1, Birth Control -1, Hospitals +1, Health Care Misc. -1, Assisted Suicides -1 -6 1991:1992 Drugs -1, Nursing Homes +1, Smoking -1, Prescriptions -1, Optometrists -1, Speech Pathologists -1, Assisted Suicide +1, Children's Health Insurance +1, AIDS +1, Mental Health -1, Smoking II -1, Respiratory Therapists -1, Tobacco +1, AIDS II -1, AIDS III -1, Health Care for the Poor -1, Facilities -1, Smoking III +1, AIDS IV -1, Malpractice -1, Assisted Suicides -1, Malpractice II -1, Health Code +1, Abortion -1, Abortion +1, EMS Liability -1, Physician's Assistants +1, Chiropractors -1 Insurance Regs. -1, Physician's/Torts -1, Smoking IV +1, Abortion -1 -12 1993:1994 Health Occupation Regulation +1, AIDS +1, HMOs -l, Forensic Labs +1, Medical Waste -1, Abortion -1, Health Insurance Regs. -1, Community Mental Health -1, Optometrists -1, Patient Advocate -1, Health Technologies -1, Nursing Homes -1, Smoking +1, HMOs +1, Health Care -1, Dental Hygiene -1, Smoking +1, Access to Health Insurance -1, Immunizations -1, Medical Savings Accounts +1, Health Facilities Safety -1, Midwives -1, Wheelchair Lemon Law +1, Smoking -1, Abortion -1, Tobacco -1, Medicaid -1, Malpractice +1, Assisted Suicide +1, Health Professions Regs. +1, Abortion +1, Nursing Homes -1, Certificate of Need +1, Smoking -1, Smoking +1, AIDS -1, Medical Malpractice +1, Psychiatric Hospitals +1, Clinical Care Mgmt. +1, Tobacco -1, Right-to-Die -1, Assisted Suicide -1 -8 Appendix 5-9 181 Michigan Relative Change Year 1962 1963 1964 1965-1966 1967-1968 1969-1970 1971-1972 1973-1974 1975-1976 1977-1978 1979-1980 1981-1982 1983-1984 1985-1986 1987-1988 1989-1990 1991-1992 1993-1994 Belatixefihanee 00:95” I N.) 182 Appendix 5-10 State Bicameral Model Coeff. S.E. t Constant -1.29 2.78 -0.46 Occup. 0.79 0.26 3.06 Age -2.98 0.97 -3.06 Seniority -3.19 2.21 -1.44 Retention 0.39 0.17 2.31 n=18, R-squared = 0.57, F = 4.35 [0.02] 183 Appendix 5-11 State Bicameral Diagnostics Testing for Serial Correlation from Lags 1 to 2 Ch12(2)= 1.322 F-Form(2,12)= .48 [.6328] Error Autocorrelation Coefficients = .0322 .3113 ARCH Test Residuals scaled by .435d+01 CNST LAG COEF F. .6845 .0182 S.E.s .3356 .2751 RSS = .1235D+02 CHIZ(1) = 6.784 with F (1,2) = 7.97[.0154] Chi-Squared Test for Normality CHI2 Test for Normality: CH12(2) = .942 RESET F-Test for adding Y-hat2 F(1,13) = 1.354 [.2655] 184 Appendix 5—12 Michigan Seat Change Yea: Seattlhange 1962 0 1963 +2.29 1964 +2.29 1965-1966 -19.67 1967-1968 +1828 1969-1970 -1.64 1971-1972 -2.73 197 3-1974 -1.64 1975-1976 -10.38 1977-1978 +1.71 1979-1980 -3.28 1981-1982 -3.83 1983-1984 -6.56 1985-1986 -5.46 1987-1988 +2.86 1989-1990 -2.19 1991-1992 -0.55 1993-1994 +6.28 Appendix 5—13 185 Michigan Gubernatorial Popular Vote Year 1962 1963 1964 1965-1966 1967-1968 1969-1970 1971-1972 1973-1974 1975-1976 1977-1978 1979-1980 1981-1982 1983-1984 1985-1986 1987-1988 1989-1990 1991-1992 1993-1994 Per n 51.56 52.44 52.44 57.78 61.33 61.33 51.56 51.56 52.44 52.44 57.78 57.78 53.33 53.33 69.33 69.33 51.55 51.55 Vte 186 Appendix 5-14 Michigan Party Strength Year P r r 1962 54.89 1963 57.11 1964 57.11 1965-1966 63.33 1967-1968 52.22 1969-1970 52.67 1971-1972 52.22 1973-1974 52.67 1975-1976 62.00 1977-1978 63.33 1979-1980 63.78 1981-1982 61.11 1983-1984 54.44 1985-1986 51.78 1987-1988 56.22 1989-1990 54.44 1991-1992 54.44 1993-1994 51.78 187 Appendix 5-15 State Tricameral Model Coeff. S.E. t Constant -25.07 25.98 -0.96 Seat -0.29 0.26 -0.10 Seat*Unified -1.61 1.28 -1.26 Governor Pop. 0.003 0.27 0.14 Party 0.45 0.41 1.08 Party*unified 0.05 0.07 0.66 n=18, R2=0.25, F5,12=0.78, prob.F.=0.5832 188 Appendix 5-16 State Tricameral Diagnostics Testing for Serial Correlation from Lags 1 to 2 CHI"'(2) = 4.288 F-Form( 2, 10) = 1.56 [ .2565] Error Autocorrelation Coefficients: .6399 -.0595 ARCH Test Residuals Scaled by.6197D+01 CNST 1 LAG COEFF. .5980 .0975 S.E.'s .3616 .3153 RSS: .15027D+02 0: 1.22585 CHI‘~’( 1): .161 with F( 1, 10) = .10 [.7635] CHI-Squared Test for Normality: CH12(2) = .130 RESET F-Test for adding Yhat2 F( 1, 11) = .020 [.8909] 189 References Abramowitz, Alan I. 1980. "A Comparison of Voting for US. Senator and Representative in 1978." American Political Science Review 74:633- 640. Alt, James E. and Robert C. Lowry. 1994. "Divided Government, Fiscal Institutions, and Budget Deficits: Evidence From the State. American Political Science Review 88: 811-828. Altman, Stuart H., Marc Rodwin. 1994. Competitive Markets and Ineffective Regulation. Health Policy and Nursing, Edited by Chatlene Harrington and Carroll Estes. Boston: Jones and Bartlett. Baker, John R. 1990. "Exploring the Missing Link: Political Culture as an Explanation of the Occupational Status and Diversity of State Legislation in Thirty States. Western Political Quarterly 43(3): 597 - 612. Baker, Ross K. 1980. Friend and Foe in the US. Senate. New York: Free Press. Baker, Ross K. 1989. House and Senate. New York: W.W. Norton and Company. Banzhaf, John F. 1965. "Weighted Voting Doesn't Work: A Mathematical Analysis." Rutgers Law Review 19:317-343. Bernstein, Robert S. and Michael B. Berkman. 1988. "Do Senators Moderate Strategically?" American Political Science Review 82:237-245. Beyle, Thad L. and Robert Dalton. 1983. "The Governor and the State Legislature" In Being Governor: The View from the Office, Thad Beyle and Lyn R. Muchmore Eds. Durham: Duke University Press. Beyle, Thad L., Scot Mowur. 1989. Governors: The Power of Removal. Policy Studies Journal. 17(3): 495-513. Bond, Jon R. and Richard Fleisher. 1980. "The Limits of Presidential Popularity as a Source of Influence in the U. S. House." Legislative Studies Quarterly 5: 69-78. 190 Biles, Brian, Carl Schramm, and J. Graham Atkinson. 1980. Hospital Cost InflationUnder State Rate Setting. New England Journal of Medicine 303: 664-667. Brady, David, Joseph Cooper, and Patricia A. Hurley. 1987. "The Decline in Party in the US. House of Representative, 1887-1968." In McCubbins, Mathew D. and Terry Sullivan. 1987. Congress: Structure and Policy. Cambridge, UK: Cambridge University Press. Brecher, Charles. 1990. The Government's Role in Health Care. In Health Care Delivery in the United States, Edited by Anthony Kovner. New York: Springer. Brown, George Rothwell. 1922. The Leadership of Congress. Indianapolis: Bobbs-Merrill. Browne, William P., Kenneth VerBurg. 1995. Michigan Politics and Government. Lincoln: University of Nebraska Press. Carmines, Edward G. and Lawrence C. Dodd. 1985. "Bicameralism in Congress: The Changing Partnership." In Congress Reconsidered, ed. Lawrence C. Dodd and Bruce I. Oppenheimer. Washington DC: CQ Press. Chappell, Henry W. and Motoshi Suzuki. 1993. "Aggregate Vote Functions for the US. Presidency, Senate, and House." The Journal of Politics 55:207-217. Clark, Joseph S. 1965. Congress: The Sapless Branch. New York: Harper and Row. Clinton, William Jefferson. 1994. State of the Union Address. In Congressional Quarterly Almanac 1994, 6-D. Collier, Kenneth and Terry Sullivan. 1995. "New Evidence Undercuttin the Linkage of Approval with Presidential Support." Journal of Politics 57: 197-209. Condorcet, Marquis de. 1785. Essai sur l'application de l'analyse a la probabilite des decisions rendues a la pluraliste des voix. Paris: De l'Imprimerie royale. 191 Copeland, Gary W. 1989. "Choosing to Run: Why House Members Seek Election To the Senate." Legislative Studies Quarterly 24:549-565. Congress and the Nation. Washington D.C.: CQ Press. Congressional Quarterly Almanac. Washington DC: CQ Press. Congressional Quarterly Weekly Report, October, 9, 1993, p.27 34. Cox, Gary W. and Samuel Kernell. 1991. The Politics of Divided Government. Boulder, CO: Westview Press. Croly, Herbert. 1915. Progressive Democracy. New York: Macmillian and Co. ' Cutler, Lloyd. 1986. "To Form a Government." Separation of Powers--Does it Still Work?, eds. Robert A. Goldwin and Art Kaufman. Washington D.C.: American Enterprise Institute. David, Sheri I. 1985. With Dignity: The Search for Medicare and Medicaid. Westport, CT: Greenwood. Diamond, Martin. 1981. The Founding of the Democratic Republic. Itasca, Illinois: F.E. Peacock. Diamond Martin. 1987. The Federalist. History of Political Philosophy, ed. Leo Strauss and Joseph Cropsey. Chicago: University of Chicago Press. Dunham, Richard S. 1996. Here Comes Gridlock '96. Businessweek, 8 January. Dunn, Charles W. 197 5. The Future of the American Presidency. Morristown, NJ: General Learning Press. Dye, Thomas R. 1984. "Party and Policy in the States." Journal of Politics 46: 1097-1116. Elazar, Daniel. 1966. American Federalism: A View From the States, 4th Ed. New York: Thomas Y Crowell Erler, Edward J. 1992. The American Polity. New York: Crane Russak. 192 Evans, C. Lawrence. 1995. Committees and Health Jurisdiction in Congress. In Mann, Thomas E, Norman J. Ornstein. Intensive Care: How Congress Shapes Health Policy. Washington DC: American Enterprise Institute. Esposito, Alfonso, Michael Hupfer, Cynthia Mason, and Diane Rogler. 1982. Abstracts of State Legislated Hospital Cost Containment Programs. Health Care Financing Review. 4(2): 129-158. Feingold, Eugene. 1966. Medicare: Politics and Policy. San Francisco: Chandler. Fenno, Richard F. Jr. 1966. The Power of the Purse. Boston: Little, Brown, and Co. Fenno, Richard F. Jr. 1973. 1973. Congressmen in Committees. Boston: Little, Brown, and Co. Fenno, Richard F. Jr. 1978. Homestyle. Glenview, Illinois: Scot, Foresman, and Company. Fenno, Richard F. Jr. 1982. The United States Congress: A Bicameral Perspective. Fiorina, Morris. 1992. Divided Government. New York: Macmillan and Company. Foley, Michael. 1980. The New Senate. New Haven: Yale University Press. Freedman, Grace Roegner. 1995. Toward a Macro Theory of Health Care Policymaking: Lessons from the Enactment of Health Care Legislation, 1945-1992. Manuscript Delivered at the Annual meeting of the American Political Science Association, September, in Chicago, Illinois. Goodnow, F. J. 1900. "Politics and Administration." Annals 16: 177-188. Granato, James. 1992. “An Agenda for Econometric Model Building.” Political Analysis. 123-156. Grofman, Bernard, Robert Griffin, and Amihai Glazer. 1991. "Is the Senate More Liberal Than the House? Another Look." Legislative Studies 193 Quarterly 26:281 - 295. Haas, Peter J ., Deil S. Wrightsus. 1989. Public Policy and Administrative Turnover in State Government: The Role of the Governor. Policy Studies Journal 17(4): 788-803. Hammond, Thomas H and Gary J. Miller. 1987. "The Core of the Constitution." American Political Science Review 81:1155-1174. Harris, Richard. 1966. A Sacred Trust. New York: New American Library. Heidegger, Martin. [1955] 1977. The Question Concerning Technology. In The Question Concerning Technology and Other Essays. Translated and Edited by William Lovitt. New York: Harper Torchbooks Holoweiko, Mark. 1988. "What Happens When Profit Motive Runs Wild." Medical Economics 18: 150-168. Jacobson, Gary C. 1987. The Politics of Congressional Elections. Glenview, Illinois: Scott Foresman and Company. Jacobson, Gary C. 1990. The Electoral Origins of Divided Government: Competition in United States House Elections, 1946-1988. Boulder: Westview Press. Jewell, Malcolm E. and Samuel C. Patterson. 1986. The Legislative Pracess in the United States. New York: Random House. Keefe, William J. and Morris S. Ogul. 1989. The American Legislative Process: Congress and the States. Englewood Cliffs, NJ: Prentice Hall. Kessler, Charles. 1988. The Founders and the Classics. The American Founding: Essays on the Formation of the Constitution, ed. J. Jackson Barlow, Leonard W. Levy, and Ken Masugi. New York; Greenwood Press. Kiewiet, Roderick and Mathew McCubbins. 1991. The Logic of Delegation. Chicago: University of Chicago Press. King, Anthony. 1983. "A Mile and a half is a Long Way." In Both Ends of the Avenue ed. Anthony King, 246-273. Washington DC: American Enterprice Institute. 194 King, Gary, Robert Keohane, and Sidney Verba. 1994. Designing Social Inquiry: Scientific Inference in Qulaitative Research. Princeton: Princeton University Press. Kingdon, John. 1989. Congressmen '3 Voting Decisions. Ann Arbor: University of Michigan Press. Korns, William, Mary Cohn, Irwin B. Arieff, Michael D. Wromser, and Lynda McNeil. 1982. Origins and Development of Congress. Washington D. 0.: Congressional Quarterly Press. Kozak, David C. and John D. McCartney. 1987. "Congress and Public Policy." PS 19:15-22. Krehbiel, Keith. 1991. Information and Legislative Organization. Ann Arbor: University of Michigan Press. Krehbiel, Keith. 1993. "A Theory of Divided and Unified Government." Unpublished Manuscript. Graduate School of Business, Stanford University. Kuklinski, James H. and Darrell M. West. 1981. "Economic Expectations and voting Behavior in United States House Representatives and Senate Elections." American Political Science Review 7 5:436-447 Lacayo, Richard. 1994. "Down for the Count?", Time, August 22, p. 34. Light, Paul C. 1991. The President '8 Agenda. Baltimore: Johns Hopkins University Press. Levine, Peter B. 1984. An Overview of the State Role in the United States Health Scene. Health Care Politics and Policy. Albany: Delmar. Link, Arthur S. 1966. The Papers of Woodrow Wilson. Princeton: Princeton University Press. Lipson, Deborah. 1991. Overview of State Role in Health Care Policy. Health Politics and Policy Edited by Theodor Litman and Leonard Robins. Albany: Delmar. Mann, Thomas E, Norman J. Ornstein. 1995. Intensive Care: How Congress Shapes Health Policy. Washington DC: American Enterprise 195 Institute. Mansbridge, Jane J. 1983. Beyond Adversary Democracy. Chicago: University of Chicago Press. Mansfield, Harvey 0., Jr. 1989. Taming the Prince: the ambivalence of modern executive power. New York: The Free Press. Marini, John. 1992. The Politics of Budget Control. New York: Crane and Russak. Marmor, Theodore R. 197 3. The Politics of Medicare. New York: Aldine Mayhew, David R. 1974. The Electoral Connection. New Haven, CN: Yale University Press. Mayhew, David R. 1991. Divided We Govern. New Haven: Yale University Press. McCubbins, Mathew D. and Terry Sullivan. 1987. Congress: Structure and Policy. Cambridge, UK: Cambridge University Press. Merrill, Teri. 1987. "Regulation vs. Competition: Three States Chose." Hospitals 61 (13): 34-35. Miller, Gary J. and Thomas H. Hammond. 1989. "Stability and Efficiency in a Separation-of-Powers Constitutional System." In The Federalist Papers and the New Institutionalism, ed. Bernard Grofman and Donald Wittman. New York: Agathon. Miller, Gary J. and- Thomas H. Hammond. 1990. "Committees and the Core of the Constitution." Public Choice 66:201-227. Oleszek, Walter J. 1984. Congressional Procedures and the Policy Process. Washington DC: CQ Press. Ornstein, Norman J. 1981. "The New House and the New Senate." In The New Congress, ed.Thoma8 E. Mann and Norman J. Ornstein. Washington DC: American Enterprise Institute. Peterson, Mark A. 1990. Legislating Together. Cambridge, MA: Harvard University Press. 196 Peterson, Merrill D. 1984. ed. Jefferson: Writings. New York: Library of America. Petracca, Mark P. 1989. "The Distribution of Power in the FederalGovernment from The Federalist Papers -- A Critique." In The Federalist Papers and the New Institutionalism, ed. Bernard Grofman and Donald Wittman. New York: Agathon. Pressman, Jeffrey L. 1966. House vs. Senate. New Haven: Yale University Press. Publius. [1787-1788] 1961. The Federalist Papers, ed. by Clinton Rossiter. New York: Mentor. Reedy, George E. 1986. The US. Senate. New York: Crown. Rohde, David W. 1991. Parties and Leaders in the Postreform House. Chicago: University of Chicago Press. Rosenbaum, David E, "Gidlock," New York Times, 20 December, 1995, C17. Rosenbloom, David H. 1983. Public Administration Theory and the Separation of Powers. Public Administration Review 43: 219-227. Rovner, Julie. 1995. Congress and Health Care Reform. In Mann, Thomas E, Norman J. Ornstein. Intensive Care: How Congress Shapes Health Policy. Washington DC: American Enterprise Institute. Sabato, Larry. 1983. "Changing Contours of State Government" In Goodbye to Good-time Charlie: The American Governorship Transformed. Washington DC: CQ Press. Schick, Allen. How a Bill Did Not Become a Law. In Mann, Thomas E, Norman J. Ornstein. Intensive Care: How Congress Shapes Health Policy. Washington D.C.: American Enterprise Institute. Schmitt, Carl. 1976. The Concept of the Political. Rutgers: Rutgers University Press. Shepsle, Kenneth A. "Institutional Equilibrium and Equilibrium Institutions." Paper presented at the 1983 APSA meeting in Chicago) 197 Sigelman, Lee. 197 6. "The Quality of Administration." Administration and Society 8: 107-144. Simon, Herbert. 1976. Administrative Behavior, Third Edition. New York: The Free Press. Sinclair, Barbara. 1989. The Transformation of the United States Senate. Baltimore: Johns Hopkins University Press. Skidmore, Max J. 197 0. Medicare and the American Rhetoric of Reconciliation. University, AL: University of Alabama Press. Squire, Peverill. 1992. "Legislative Professionalization and Membership Diversity in State Legislatures. Leigslative Studies Quarterly 17: 69- 79. Strom, Gerald S. and Barry S. Rundquist. "A Revised Theory of Winning in House- Senate Conferences." American Political Science Review 71:448-453. Sundquist, James. 197 4. Reflections on Watergate: Lessons for Public Administration. Public Administration Review 34. Sundquist, James. 1986. Quoted in Ann Stuart Anderson. 1986. A 1787 Perspective on Separation of Powers. Separation of Powers--Does it Still Work ?, eds. Robert A Goldwin and Art Kaufman. Washington D.C.: American Enterprise Institute. Sundquist, James. 1992. Constitutional Reform and Effective Government. Washington DC: Brookings Institution. Time, "Gridlock?," July 4, 1994. de Tocqueville, Alexis. [1835] 1966. Democracy in America, Edited by J .P. Meyer and Max Lerner, Translation by George Lawrence. New York: Harper and Row Vile, M. J. C. 1967. Constitutionalism and the Separation of Powers. Oxford: Clarendon Press. 198 Waldo, Dwight. 1948. The Administrative State. New York: Ronald Press. Waterman, Richard W. 1990. "Comparing Senate and House Electoral Outcomes: The Exposure Thesis." Legislative Studies Quarterly 25:99- 114. Weaver, Warren. 1972. Both Your Houses. New York: Praeger. Weissert, Carol S. and William G. Weissert. 1996. Governing Health: The Politics of Health Policy. Baltimore: Johns Hopkins University Press. Wiggins, Charles W., Keith Hamm, and Charles Bell. "Interest Group and Party Influence in the Legislative Process." Journal of Politics 54(1): 82-100 Wilson, Woodrow. [1885] 1966. "Congressional Government." In The Papers of Woodrow Wilson, ed.n Arthur S. Link. Princeton: Princeton University Press. Wilson, Woodrow. [1887] 1955. The Study of Administration. Washington D.C.: Public Affairs Press. Wirls, Daniel. 1993. "To Cool the Coffee: The Effect of Bicameralism on Congressional Action." Presented at the annual meeting of the American Political Science Association, Washington. DC. S 53 MICHIGAN TATE UNIV. LIBRARIES lllllllllllll lllllllllllllllllllllllllllllllllll 312 3015795937