¹ For an interesting calculation of the balance of numbers and potential influence between elected and permanent executives in British government, see Rose, Richard, ‘The Variability of Party Government: a Theoretical and Empirical Critique’, Political Studies, XVII (1969), 413–45.CrossRefGoogle Scholar For a useful discussion of the power of the British civil service, see Gordon, Michael R., ‘Civil Servants, Politicians, and Parties: Shortcomings in the British Policy Process’, Comparative Politics, IV (1971), 29–58.CrossRefGoogle Scholar

² It would be inappropriate in this brief research report to offer an extensive theoretical analysis of ‘responsiveness’ and of the closely related concept of ‘responsibility’. A fully elaborated theory of bureaucratic responsiveness would have to consider explicitly the distinction between social needs and political demands, as well as the issue of differential responsiveness, for responsiveness to one group or constituency may involve systematic neglect of others. I will here leave aside these matters, along with the still knottier problem of developing operational measures of responsiveness. My working assumption is that in some generic sense we can distinguish between more and less responsive bureaucracies. I should perhaps add that responsiveness is not the only standard by which we judge bureaucracies - or governments.

³ Friedrich, Carl J., ‘Public Policy and the Nature of Administrative Responsibility’, in Friedrich, Carl J. and Mason, E. S., eds., Public Policy, 1 (Cambridge, Mass.: Harvard University Press, 1940), p. 2.Google Scholar In the present context Friedrich's ‘responsibility’ is equivalent to my ‘responsiveness’. For a useful discussion of responsibility and responsiveness, see Birch, A. H., Representative and Responsible Government (Toronto: University of Toronto Press, 1964), esp. pp. 17–21 and p. 170.Google Scholar

⁴ See Friedrich, ‘Public Policy’, for a lucid statement of the problem.

⁵ Kingsley, J. Donald, Representative Bureaucracy (Yellow Springs, Ohio: Antioch Press, 1944), p. 274.Google Scholar

⁶ See Friedrich, ‘Public Policy’, and the reply by Finer, , ‘Administrative Responsibility in Democratic Government’, Public Administration Review, I (1941), pp. 335–50CrossRefGoogle Scholar, both conveniently excerpted in Rourke, Francis E., ed., Bureaucratic Power in National Politics (Boston: Little, Brown, 1965), pp. 165–87.Google Scholar

⁷ Cf. C. H. Sisson's view: ‘The main work of government is done by an executive, and where the work is of any complexity that executive must include a large and permanent body of officials. The triumph of democracy in Britain, if it has had a triumph here, is not to have got the work of government done by any popularly elected body but to have informed a large permanent executive with responsiveness to the moods of the elected and of the electorate.’ See The Spirit of British Administration (London: Faber, 1959), p. 151.Google Scholar

⁸ Mannheim, Karl, Ideology and Utopia (New York: Harcourt, Brace, 1946), p. 105.Google Scholar

⁹ See the report in Stern, 9 September 1970.

¹⁰ Chapman, Brian, The Profession of Government (London: Allen and Unwin, 1959), p. 274.Google Scholar

¹¹ Jacob, Herbert, German Administration Since Bismarck (New Haven: Yale University Press, 1963), p. 205.Google Scholar A similar view seems to have prevailed traditionally in Sweden. See Board, Joseph, The Government and Politics of Sweden (Boston: Houghton Mifflin, 1970), p. 164.Google Scholar

¹² Herz, John, ‘Political Views of the West German Civil Service’, in Speier, Hans and Davison, W. Phillips, eds., West German Leadership and Foreign Policy (Evanston, Ill.: Row, Peterson, 1957). pp. 111–13.Google Scholar I would like to avoid here the complex issue of Max Weber's view of the role of the bureaucrat, for while in some ways it is similar to that which I term ‘classical’, in other respects it is closer to the ‘political’ interpretation.

¹³ Related studies have been undertaken as well in Morocco, India, and the British West Indies, though for a variety of reasons, data from these studies are less relevant here. The project has been supported by grants to the University of Michigan from the Ford Foundation and the National Science Foundation.

¹⁴ For a variety of technical reasons, complete data are not available for each respondent. Hence, the ns reported in this paper do not always coincide precisely.

¹⁵ The samples of politicians were simply drawn randomly from the lower houses of the national parliaments, stratified by party. The younger politicians constitute a special sample of parliamentarians aged 40 and under. The lower response rate for Italian politicians reflects the fact that Parliament was recessed for government crisis talks during much of the period of field research in Italy in the spring of 1970.

¹⁶ The one exception to the generally high levels of frankness consisted in a substantial minority - perhaps 25 per cent - of senior Italian officials with whom the interviews were fairly formalized, with the respondents offering a ‘sanitized’ view of politics and administration in Italy. The reticence of this minority can be attributed to the novelty of ‘sociological’ interviewing among these Italian officials and to the climate of mutual suspicion and fear which characterizes Italian public life. The practical impact of this problem is that the Italian results probably understate substantially the alienation of the Italian administrative elite from the world of politics.

¹⁷ For a preliminary report of the Swedish data, see Anton, Thomas J., Linde, Claes, and Mellbourn, Anders, ‘Bureaucrats in Politics: a Profile of the Swedish Administrative Elite’, delivered to the Sixty-second Annual Meeting of the Society for the Advancement of Scandinavian Study, New York, May 1972.Google Scholar

¹⁸ Data from Sweden suggest that as many as a third of the younger bureaucrats there are from working-class homes, reflecting the impact of more than three decades of Social Democratic rule. See Anton, et al. , ‘Profile of the Swedish Administrative Elite’, pp. 12–15.Google Scholar

¹⁹ Some caution is appropriate in interpreting responses to these questionnaire items, for they are potentially subject to ‘acquiescence response set’, that is, the tendency for some people to agree with such items regardless of content. This problem is much less important when analyzing responses to interview questions.

²⁰ Across our set of three national samples the mean intercorrelation among the six items is r = 35. The Spearman-Brown internal reliability coefficient is, thus, 77. See Selltiz, Claire, et al. , Research Methods in Social Relations (New York: Holt, Rinehart, and Winston, 1959), pp. 183–4.Google Scholar

²¹ Data from the Swedish sample suggest relatively high tolerance for politics there, too; the mean Swedish score on the ITP is 58.

²² Note that the standard of comparison here is essentially a relative one: respondents are compared with the mean score (50) for the three-nation sample. However, given the scoring system for the ITP, this mean score happens to be the dividing line between those who, on balance, tended to agree with the six items, and those who, on balance, tended to disagree. Thus, it is possible to interpret this mean in an absolute sense, as well.

²³ The standard deviation for the ITP and for nearly every one of its component items is markedly higher for the German sample than for the other two.

²⁴ Across our total sample of bureaucrats from the three countries, the correlation between the questionnaire-based ITP and the measure of Enthusiasm for Politics is r = 25. Statistical purists may be offended by the use of interval-level statistics on ordinal-level data. My purpose here, however, is merely to indicate roughly the order of magnitude of the statistical relations, and in this context Pearson's r is not a misleading measure. For a recent discussion of this issue, see Labovitz, S., ‘Some Observations on Measurements and Statistics’, Social Forces, LVI (1967), pp. 151–60.CrossRefGoogle Scholar

²⁵ Across our total sample of bureaucrats from the three countries, the mean intercorrelation among the three items is r = ·35. The mean national scores are as follows: Britain, 49; Germany, 46; Italy, 32; Sweden, 61. By this measure Swedish bureaucrats are considerably more committed to programmatic policy-making than their counterparts in the other three countries.

²⁶ The correlation between the ITP and the IPC across our total sample of bureaucrats is r = ·50.

²⁷ The correlation between the ITP and the ‘legalism’ questionnaire item is r = .56; that between the measure of Enthusiasm for Politics and this item is r = 19. The correlation between the two measures of tolerance and Enthusiasm for Politics and a measure of the ‘activism’ of the bureaucratic role described by the respondent ranges from r = .17 to r = .34.

²⁸ The correlation between the measure of Enthusiasm for Politics and approval of some relaxation of the anonymity rule in Britain is r = .20. The correlation between the measure of Enthusiasm for Politics and approval of ‘politicization’ of the civil service in Germany is r = .60. The comparable correlation for approval of the gabinetto ministeriale in Italy is not yet available but is certain to lie within this range.

²⁹ The correlation between the ITP and the ‘misanthropy’ item is r = .45; that between the ITP and the ‘pro-lobbying’ item is r = .41; that between the measure of Enthusiasm for Politics and the item on social conflict is r = .20.

³⁰ The correlation between the ITP and the item on political propaganda is r = –·47.

³¹ The correlation between the Elitism Index and the ITP is r = –·47.

³² The correlation between age and the measure of Enthusiasm for Politics in Italy is r = –.22; age is correlated positively with the Elitism Index, r =.20, and negatively with the Index of Programmatic Commitment, r = –·2 5. The correlation between age and the item expounding a ‘legalist’ interpretation of the role of a civil servant is r =.26. Note that even the ‘young’ Italian bureaucrats are not so young; the median age of the Italian high-fliers is 44, as contrasted with 36 and 39 for Britain and Germany respectively. Differences across age cohorts in data gathered at a single point in time are notoriously ambiguous. Broadly speaking, they may reflect life cycle changes, generational differences, or ‘natural selection’. (For a detailed discussion, see my ‘Studying Elite Political Culture: the Case of “Ideology”, American Political Science Review, LXV (1971), p. 673).Google Scholar Strictly speaking, we cannot yet exclude the life-cycle explanation for these Italian data, but the generational explanation is inherently plausible, given the upheavals in Italian history in the last half-century. More rigorous evidence is available for the German case, discussed below.

³³ Weber, , ‘Bureaucracy’, in Rourke, , ed., Bureaucratic Power in National Politics, pp. 8–10.Google Scholar That the Italian higher bureaucracy in the past has shown a similar ability to resist the impact of political revolutions is suggested by the fact that as late as 1937 every single one of the top thirty civil servants in the all-important Ministry of Corporations had entered the civil service before 1916. See Cole, Taylor, ‘Italy's Fascist Bureaucracy’, American Political Science Review,XXXII (1938), 1143–57.CrossRefGoogle Scholar

³⁴ Hanf, Kenneth, ‘Administrative Trends in West and East Germany ’, paper prepared for the Conference, Group on German Politics, Western Regional Meeting, Sacramento, California, 04 1970, pp. 11–12Google ScholarDahrendorf, Ralf, Society and Democracy in Germany (Garden City, N.Y.: Doubleday, 1967), pp. 240–1.Google Scholar

³³ These data are borrowed from a preliminary and unpublished research report. I am grateful to Uwe Schleth for allowing me to see these results.

³⁶ The correlation between membership in or sympathy for the SPD and the ITP is r = ·25; for the IPC, r = ·23; for the Elitism Index, r = –·24; for the interview-based measure of Enthusiasm for Politics, r = ·26; for the questionnaire item on political propaganda, r = ·22.

³⁷ The correlation in Germany between age and the ITP is r = -.30; that for the IPC is r = -.30; that for the Elitism Index, r = -.21; that for the interview-based measure of Enthusiasm for Politics, r = -.30. These findings are broadly consistent with those reported in Ellwein, Thomas and Zoll, Ralf, Berufsbeamtentum – Anspruch and Wirklichkeit (Düsseldorf: Bertelsmann Universitätsverlag, 1973).Google Scholar

³⁸ An alternative and perhaps complementary explanation is offered by C. H. Sisson. ‘The British administration enjoys an extraordinary degree of freedom from political intrusion, and there is no doubt that this freedom greatly facilitates the responsiveness of the administration to the purposes of cabinet government.’ (Sisson, The Spirit of British Administration, p. 112.) In the subsequent analysis of our data we will want to look closely at the interaction between the values and behavior of bureaucrats and the values and behavior of politicians.

³⁹ The correlation between ‘specialist’ status and the ITP is r = -.26. The correlation for the Elitism Index is r = .24; for the item tapping propensity to control political propaganda, r = .25; and for answers to my question about the ‘grey area’ between politics and administration, r = .24.

⁴⁰ The correlation in Britain between age and the ITP is r = -.30. That for the IPC is r = -.37; for the Elitism Index, r = .27; for the interview-based measure of Enthusiasm for Politics, r =-.19. The point, of course, is not that the older British civil servants are so opposed to programs, politics, pluralism, and equality, but that the younger ones are so firmly in favor of these things. In assessing the extent of the generational differences in Britain as contrasted to the continental countries, one must recall that the Italian and German samples of high-fliers were nominated by our senior incumbents, whereas the British sample was selected by more objective criteria. Hence, the continental samples may systematically understate the real generational differences, whereas this is unlikely to be true of the British sample.

⁴¹ See Anton et al., ‘Profile of the Swedish Administrative Elite’. A similar set of findings for Sweden is reported in Ulf Christoffersson, Molin, Björn, Månsson, Lennart and Strõmberg, Lars, Byråkrati och politik (Stockholm: Bonniers, 1972).Google Scholar

⁴² Wheare, Kenneth, ‘Seven Characters in Committee Work’, as reprinted in Rose, Richard ed., Policy-Making in Britain (New York: Free Press, 1969), pp. 107–8.Google Scholar

⁴³ This statement is ascribed to the late MrMorrell, Derek by Sampson, Anthony in ‘The Men Who Run Britain’, Observer Review, 21 02 1971, p. 19.Google Scholar

⁴⁴ Simon, Herbert, Administrative Behavior (New York: Free Press, 1957), p. xii.Google Scholar

⁴⁵ Simon, , Administrative Behavior, pp. 57–8.Google Scholar Cf. David Marquand's view that ‘… there is nothing odd or distressing in politicians basing their decisions on a value-judgment. One of the reasons why it is better to be governed by politicians than by civil servants is that civil servants are apt to disguise their value judgments as judgments of fact. Politicians are slightly more honest about bringing theirs into the open.’ 'A War of Ideologies’, New Society, XVII, no. 438 (18 02 1971), p. 279.Google Scholar