¹ For reviews of these theories see Taylor, Michael, ‘On the Theory of Government Coalition Formation’, British Journal of Political Science, 11 (1972), 361–73CrossRefGoogle Scholar, and Taylor, Michael and Laver, Michael, ‘Government Coalitions in Western Europe’, European Journal of Political Research, 1 (1973), 205–48.CrossRefGoogle Scholar

² Taylor, and Laver, , ‘Government Coalitions’, p. 207Google Scholar; de Swaan, Abram, Coalition Theories and Cabinet Formation (Amsterdam: Elsevier, 1973), Chaps. 4 and 5.Google Scholar

³ For an empirical testing of these theories see Taylor, and Laver, , ‘Government Coalitions’.Google Scholar

⁴ See Herman, Valentine and Pope, John, ‘Minority Governments in Western Democracies’, British Journal of Political Science, 111 (1973), 191–212CrossRefGoogle Scholar, and Laver, Michael, ‘Dynamic Factors in Government Coalition Formation’, European Journal of Political Research, 11 (1974), 259–70.CrossRefGoogle Scholar

⁵ This point is documented in detail in Budge, Ian and Farlie, Dennis, Voting and Party Competition (London and New York: Wiley, 1977), Chap. 5Google Scholar; also in Budge, Ian, Crewe, Ivor and Farlie, Dennis, eds., Party Identification and Beyond (London and New York: Wiley, 1976), Chap. 20.Google Scholar

⁶ This goes against the standard rational choice tenet that office is the be-all and end-all for politicians. A whole theory of elections has been formulated on this basis; see Downs, Anthony, An Economic Theory of Democracy (New York: Harper and Row, 1957).Google Scholar For a counter-example which demonstrates that the utility of office diminishes rapidly once preferred policies cannot be carried through by the officeholder, see Budge, and Farlie, , Voting and Party Competition, Section 5.1.Google Scholar

⁷ Assumption 3(b) thus takes into account the ideological considerations incorporated into some previous rational choice models, but modifies them through its explicit recognition that their importance will vary over time. Moreover, as we shall specify below (Table 2, Criterion iii), the important ideological difference when ideology is salient is taken as the division between Left and Non-Left parties, not degrees of Rightness or Leftness within the camps. Left-Right differences, when they emerge, are best regarded as a dichotomy rather than the continuum posited in the older theories.

⁸ See Laver, , ‘Dynamic Factors’, pp. 260–4.Google Scholar

⁹ The characteristics of governments of national unity are discussed in Valentine Herman, ‘The Size Principle and “Surplus” Governments’ (unpublished paper, University of Essex).

¹⁰ For a discussion on the ideological placements of parties on a Left-Right continuum see Taylor, and Laver, , ‘Government Coalitions in Western Europe’.Google Scholar Our theory merely requires us to distinguish between, on the one hand, parties of the Left, and those of the Non-Left (or Centre-Right) on the other, thus converting the ideological continuum of older coalition theories into a dichotomy which (with the pro/anti-system dichotomy already mentioned) probably corresponds more closely to the judgements of practising politicians.

¹¹ Establishing an actual percentage cutting-point for a near-majority party is bound to be slightly arbitrary. Before even starting the investigation we felt that 45, 46 or 47 per cent would be equally appropriate – all of them being points at which a party could claim to have missed an absolute majority by bad luck or accident. In the present investigation we use 45 per cent and this is the actual cutting-point which appears in Table 7 below. We have alternatively used 46 per cent, and achieved exactly the same overall success rate, so we do not think that the actual point used within the range makes any significant difference to our success rate.

¹² Establishing percentage cutting points here is even more difficult than for a near-majority party. We have stipulated that being a dominant party in this sense requires that the largest democratic party gets over 37 per cent, and that no other democratic party gets over 25 per cent. Again we suspect that these limits could be varied (up to 40 per cent and down to 20 per cent) without significantly affecting the results of Table 7.

¹³ The success of the hypothesis based on the two typologies also constitutes strong evidence that the typologies themselves are valid and relevant – if they were not, we should have expected the hypotheses to fail.

¹⁴ In the case of Germany, Italy and Japan, we include the first parliaments of the post-war years, before the peace treaties and official termination of occupation. Since the early parliaments were elected on the same procedures as later ones, and were substantially autonomous, there seems no reason to exclude them.

¹⁵ In North-west Europe the threat was obviously less serious than further south, which perhaps accounts for the emergence of single party governments rather than grand coalitions in Norway after the war.

¹⁶ This is operationally attested by the fact that the success rate for Criteria i (Table 4) is one of the lowest for any of the criteria: had there been a concealed tautology it should have been highest.

¹⁷ Austria and the Federal Republic of Germany during the 1960s are exceptions to this general trend, since the rigid Left-Right divisions prevailing in the 'fifties were modified by a pragmatic socialist leadership in the 'sixties. Japan has been unusual in maintaining relatively extreme Left-Right distinctions over the whole period.

¹⁸ Our data are from Keesing's Contemporary Archives; Cook, Chris and Paxton, John, European Political Facts, 1918–1973 (London: Macmillan, 1975)Google Scholar; Mackie, Thomas T. and Rose, Richard, The International Almanac of Electoral History (London: Macmillan, 1974)Google Scholar; and von Beyme, Klaus, Die parlamentarischen Regierungssysteme in Europa (München: R. Piper, 1970).Google Scholar

¹⁹ See Hurwitz, Leon, ‘An Index of Democratic Political Stability: A Methodological Note’, Comparative Political Studies, IV (1971), 41–68CrossRefGoogle Scholar, and Sanders, David and Herman, Valentine, ‘The Stability and Survival of Governments in Western Democracies’, Acta Politica, XII (1977), 346–77.Google Scholar

²⁰ See de Swaan, Abram, ‘An Empirical Model of Coalition Formation as an N-Person Game of Policy Maximization’, in Groennings, Sven, Kelley, E. W. and Leiserson, Michael, eds., The Study of Coalition Behaviour (New York: Holt, Rinehart and Winston, 1970), pp. 424–44.Google Scholar

²¹ Domination in the sense that the majority has taken into government small ‘normal parties of government’. Thus in Italy the combination of Christian Democrats with Republicans would be a success whereas a combination with a larger party would be regarded as voluntarily ceding more policy objectives than they needed, and hence as a failure of the Criterion.

²² One or two countries, however, particularly France under the Fourth Republic, do have their governments well characterized by this criterion since the normal parties – the Socialists, Radicals and MRP – were sizeable enough either to form governments on their own or to dominate them. Criterion V(c), incidentally, also provides a full specification for the permanent four-party coalition in Switzerland, which is omitted from consideration in tables and text because it is not a parliamentary democracy in the same sense as the other countries.

²³ A study of governments which formed in twelve western democracies between 1945 and 1971 revealed that 35·8 per cent and 30·2 per cent of governments which formed were minority and ‘surplus’ administrations, respectively. The functions of these governments were neither predicted nor postdicted by existing coalition theories which would thus explain only a maximum of 34 per cent of the governments that formed. See Herman, and Pope, , ‘Minority Governments in Western Democracies’, pp. 192–4.Google Scholar

²⁴ See Browne, Eric C. and Franklin, Mark N., ‘Aspects of Coalition Payoffs in European Parliamentary Democracies’, American Political Science Review, LXVII (1973), 453–69.CrossRefGoogle Scholar Unpublished research by David Robertson of the Department of Government, University of Essex, has shown that British governments of differing party composition show quite clear contrasts in broad areas of expenditure. We expect this will be true of other countries.

²⁵ The theory can be cast in a form analogous to certain of the Newtonian laws of motion and similar to the predictive theory of election outcomes discussed in Budge, Ian and Farlie, Dennis, ‘Newtonian Mechanics and Predictive Election Theory’, British Journal of Political Science, VII (1977), 413–18.Google Scholar