media regulation

B.J.Pol.S. 40, 75–89 Copyright r Cambridge University Press, 2009

doi:10.1017/S000712340999024X

First published online 17 November 2009

Explaining the De Facto Independence of Public Broadcasters

CHRIS HANRETTY*

Institutions operating beyond direct control of government, such as central banks, constitutional courts and public broadcasters, enjoy guarantees of de jure independence, but de jure independence is no guarantee of de facto independence. This is especially so for public broadcasting, where cultural variables are often assumed to be decisive. In this article, the de jure and de facto independence of thirty-six public service broadcasters world-wide are operationalized, and de jure independence is found to explain a high degree of de facto independence when account is taken of the size of the market for news. Other variables considered in previous literature – such as bureaucratic partisanship and the polarization of the party system – are not found to be significant.

Between 1998 and 2008 laws affecting public service broadcasters (PSBs) were revised in Austria, Portugal, Italy, Spain, Slovenia, the United Kingdom and Finland. Such reforms usually aim to give the PSB concerned greater independence from politics.1 To this end, they propose new legislative provisions concerning appointments, funding and accountability. Unfortunately, we know little about the effects of such legislative provisions on the independ- ence of PSBs. In part, this is because previous work on the politics of public broadcasting has either been country-specific2 or has not upset the null hypothesis.3 Given the dominant market position of most PSBs in Europe,4 this lack of knowledge is regrettable. In part, however, we lack knowledge because the effect of de jure measures designed to

increase independence is uncertain. Scholarly literature on the independence of other bodies operating at arm’s length (outside government control), such as central banks,5

* Department of Social and Political Sciences, European University Institute (email: chris.hanretty@ eui.eu). The author wishes to thank Adrienne Héritier, Michaël Tatham, Alexander Trechsel and the anonymous reviewers of this article, whose comments and criticism greatly improved the argument. He also thanks Niki Yordanova, Elias Dinas, Jeppe Olesen, Lúcio Tomé Féteira, Peeter Vihalemm and Katie De Noël (European Broadcasting Union), all of whom provided data, and Kristen Kufel, who helped with German legal texts.

1 Emilio Lledó Íñigo et al., ‘Informe para la reforma de los medios de comunicación de titularidad del estado’, Report for the Consejo para la reforma de los medios de comunicación de titularidad del estado (February 2005); Paolo Gentiloni, ‘Linee guida per la riforma della Rai’, available online at www. comunicazioni.it (2007).

2 Ellis S. Krauss, Broadcasting Politics in Japan: NHK and Television News (Ithaca, N.Y.: Cornell University Press, 2000); Ken Inglis, This is the ABC: The Australian Broadcasting Commission 1932–1983 (Melbourne: Black, 2006); Jean-Claude Burgelman, ‘Political Parties and Their Impact on Public Service Broadcasting in Belgium’, Media, Culture & Society, 11 (1989), 167–93.

3 Eva Etzioni-Halevy, National Broadcasting under Siege: A Comparative Study of Australia, Britain, Israel, and West Germany (London: Macmillan, 1987).

4 Robert Picard, ‘Assessing Audience Performance of Public Service Broadcasters’, European Journal of Communication, 17 (2002), 227–35.

5 Vittorio Grilli, Donato Masciandaro and Guido Tabellini, ‘Political and Monetary Institutions and Public Financial Policies in the Industrial Countries’, Economic Policy, 6 (1991), 341–92.

constitutional courts6 and regulatory agencies,7 has used indicators of de jure independ- ence and has sidestepped the issue of whether such de jure independence guarantees de facto political independence. Analyses distinguishing between de facto and de jure independence, carried out by economists rather than political scientists, have found that the two items are only weakly correlated.8

Public service broadcasting is an important area in which to distinguish the effects of de jure independence. Most proposed explanatory factors concern features of the wider society that are often assumed to be decisive. If this is so, de jure independence is either superfluous (because de facto independence is guaranteed by features of the wider society), or has no effect (because de facto independence is made impossible by features of the wider society).This article attempts to assess the impact of grants of de jure independ- ence on de facto independence using a sample of thirty-six PSBs drawn from across the world. I have used measures for de jure and de facto independence which I describe and cross-validate with other more direct (but less widely available) measures. My finding – that de jure independence explains a high degree of de facto independence when the size of the market for news is taken into account – shows that the prospects for designing independent PSBs are good.

DE FACTO POLITICAL INDEPENDENCE

By ‘de facto political independence’, I mean the degree to which PSB employees take day- to-day decisions about their output or the output of their subordinates, without receiving and acting on the basis of instructions, threats or other inducement from politicians, or the anticipation thereof; or considering whether the interests of those politicians would be harmed by particular choices about output. I do not consider independence from power- ful economic groups within society, nor do I wish to include political involvement in setting the types of programming and strategic goals of the broadcaster, activities which are more common and less normatively troubling. The definition applies whether inducements or threats come from politicians in govern-

ment or those in opposition. Coppedge and Reinicke use ‘partiality in and greater availability of government-controlled media’ as an indicator of the lack of ‘alternative sources of information’,9 thus impairing voters’ ability to choose between electoral alternatives. Yet interference by elected politicians in general also impairs voters’ ability to choose between alternatives, since it is one of the means by which cartel parties protect

6 Erik Herron and Kirk Randazzo, ‘The Relationship between Independence and Judicial Review in Post-Communist Courts’, Journal of Politics, 65 (2003), 422–38.

7 Robert Elgie and Iain McMenamin, ‘Credible Commitment, Political Uncertainty, or Policy Complexity? Explaining the Discretion Granted to Independent Administrative Authorities in France’, British Journal of Political Science, 35 (2005), 531–48; Fabrizio Gilardi, ‘Policy Credibility and Delegation to Independent Regulatory Agencies: A Comparative Empirical Analysis’, Journal of European Public Policy, 9 (2002), 873–93.

8 Lars Feld and Stefan Voigt, ‘Economic Growth and Judicial Independence: Cross Country Evidence Using a New Set of Indicators’, European Journal of Political Economy, 19 (2003), 505–6; Alex Cukierman and Steven B. Webb, ‘Political Influence on the Central Bank: International Evidence’, World Bank Economic Review, 9 (1995), Table 3.

9 Michael Coppedge and Wolfgang H. Reinicke, ‘Measuring Polyarchy’, in Alex Inkeles, ed., On Measuring Democracy (New Brunswick, N.J.: Transaction, 1991), p. 50, emphasis added; Robert A. Dahl, A Preface to Democratic Theory (Chicago: University of Chicago Press, 2006), p. 70; Robert A. Dahl, Polyarchy: Participation and Opposition (New Haven, Conn.: Yale University Press, 1971), pp. 2–3.

76 HANRETTY

themselves against political insurgents not currently represented in the legislature.10 Thus, political independence in general is desirable. It is also desired: the proposed 2007 reform of Rai (the Italian PSB) sought specifically to grant it autonomy ‘from the parties’, not just from the government.11

Ideal measurement of this concept would rely on surveys of PSB journalists and the considerations influencing their work, yet such surveys are rare. Therefore, I use a proxy measure of independence based on executive turnover. In doing so, I follow Cukierman and Cukierman and Webb, who developed two proxies for de facto independence: the rate of turnover of chief executives (TOR), and the political vulnerability index (VUL).12 TOR is equal to the reciprocal of the average tenure of the chief executive in years. VUL is the percentage of government changes followed within six months by a change in the chief executive. The logic behind the use of TOR is as follows. Where chief executives are in office for a

short period of time, they lack knowledge of the broadcaster, and consequently lack the capacity to defend it. By contrast, where the chief executive has been in office longer than one legislative term – and perhaps longer than many legislators – then she will know more about the broadcaster, and be better able to defend it. The logic behind the use of VUL is as follows. If, following a new government, there is a

change in the chief executive, then either the chief executive reached the end of her term, or left early. If she reached the end of her term, it may be that the terms of chief executives are designed so as to coincide with changes in government (the case in Spain before 2006). If this is the case, then one may assume that the chief executive is, in some sense, the expression of a government choice. If the terms do not coincide by design, then the fact that they did so may create this impression in any case. If, by contrast, the chief executive left early, she was either constrained to resign, or did so of her own accord. If she was constrained to resign, this may represent the introduction of some new constraint connected to the government. If she left of her own accord, this may reflect a belief that the government should have a ‘clean slate’ to influence the forthcoming selection of a chief executive. Since the use of multiple imperfect proxies can reduce the error present in each single

proxy, I operationalize independence, I, as the average between TOR and VUL. Since higher values of TOR and VUL indicate lower independence, the formula for independence is:

I ¼ 1" ðTORþ VULÞ

2

where higher values of I indicate greater de facto independence. Table 1 shows values of I for the thirty-six PSBs included in my sample.13 These values need to be treated sensitively. The

10 Carina S. Bischoff, ‘Political Competition and Contestability: A Study of the Barriers to Entry in 21 Democracies’ (doctoral dissertation, European University Institute, 2006), pp. 112, 117.

11 Gentiloni, ‘Linee guida per la riforma della Rai’, y2.6. 12 Alex Cukierman, Central Bank Strategy, Credibility, and Independence: Theory and Evidence

(Cambridge, Mass.: MIT Press, 1992); Cukierman and Webb, ‘Political Influence on the Central Bank’. 13 Information on executive turnover was obtained from broadcasters directly and from LexisNexis

news coverage. The most recent chief executive of each PSB has been omitted where including their tenure would reduce the value of TOR. Data on government changes was taken from Ian Budge, Jaan Woldendorp and Hans Keman, ‘Party Government in 20 Democracies: An Update (1990–1995)’, European Journal of Political Research, 33 (1998), 125–64; Ferdinand Müller-Rommel, Katja Fettelschoss and Philipp Harfst, ‘Party Government in Central Eastern European Democracies: A Data Collection (1990–2003)’, European Journal of Political Research, 43 (2003), 869–94; and subsequent issues of the European Journal of Political Research, except for Chile (Dieter Nohlen, ‘Chile’, in Dieter Nohlen, ed.,

Explaining the De Facto Independence of Public Broadcasters 77

range of values is limited, and the discrimination at higher levels of I seems limited. The Swiss broadcaster SRG-SSR has a higher value of I than the British Broadcasting Corporation (BBC), but this reflects the fact that BBC directors-general who have retired have done so after seven or ten years whilst SRG-SSR directors have done so after twenty years. This difference, though it affects TOR, does not reflect differences in de facto independence. Additionally, there seems to be no readily identifiable cut-off point above and below which one can talk of ‘independent’ and ‘not independent’ broadcasters respectively.

TABLE 1 Ranking of Broadcasters according to Perceived Independence (I)

Public Service Broadcaster Abbreviation Period I

SRG SSR idée suisse SRG-SSR 1950–2006 0.96 Norrikskringskasting NRK 1948–2006 0.93 Oy Yleisradio YLE 1945–2005 0.92 Zweites Deutsches Fernsehen ZDF 1962–2007 0.91 Australian Broadcasting Corporation ABC 1965–2006 0.91 British Broadcasting Corporation BBC 1944–2006 0.89 Danmarks Radio DR 1961–2005 0.89 Televisión Nacional TN 1995–2007 0.89 Radio Telefı́s Éireann RTÉ 1960–2007 0.88 Nippon Housou Kyoukai NHK 1946–2007 0.87 Vlaamse Radio- en Televisieomroep VRT 1996–2007 0.87 Sveriges Television Ab SVT 1978–2007 0.86 Canadian Broadcasting Corporation CBC 1945–2007 0.86 Österreischischer Rundfunk ORF 1960–2007 0.85 Eesti Televisioon EETV 1994–2007 0.82 Radiotelevisione Italiana (1) RAI 1975–1993 0.81 Radio-Télévision Belge de la Communauté française RTBF 1994–2007 0.79 Israel Broadcasting Authority IBA 1989–2006 0.79 Television New Zealand TVNZ 1986–2006 0.78 Corporation for Public Broadcasting CPB 1969–2007 0.75 Telewizja Polska TVP 1994–2007 0.75 Latvijas Valsts Televizija LT 1995–2007 0.74 France Télévisions (1) FT82 1982–2000 0.72 France Télévisions (2) FT73 1973–1982 0.72 Radiotelevizija Slovenija RTVSLO 1991–2007 0.68 Societatea Româna de Televiziune RO/TVR 1993–2007 0.65 Česká televize CTV 1992–2007 0.63 South Africa Broadcasting Corporation SABC 1994–2007 0.59 Television Española SA TVE 1981–2006 0.56 Radiotelevisão Portuguesa SA RTP 1977–2007 0.56 Bâlgarsko Nationalno Radio BNR 1997–2007 0.55 Lietuvos Radijas ir Televizija LRT 1996–2007 0.51 Radiotelevisione Italiana (2) RAI 1993–2005 0.42 Slovenská televı́zia SK/STV 1994–2007 0.39 Magyar Televizió MTV 1990–2007 0.35 Bâlgarska Nationalna Televizija BNT 1991–2007 0.26

Note: I is a proxy measure of independence.

(F’note continued)

Elections in the Americas, Volume 2 (Oxford: Oxford University Press, 2005)), and for the French language community and the Flemish community in Belgium.

78 HANRETTY

The proxy nevertheless agrees with more direct measurements. Polls conducted in Denmark, Britain, Canada, France and Italy show that broadcasters with higher scores were more likely to be perceived as independent by the public: 42 per cent and 38 per cent of respondents believed that the BBC and Danmarks Radio respectively were independent of the government;14 a slightly lower percentage (35 per cent) believed CBC was independ- ent of the Canadian government;15 22 per cent of respondents thought that the French media in general was independent of political interests;16 in a subsequent question, France Télévisions was neither substantially more trusted, nor substantially less trusted, suggesting that if a question had been asked about its independence, France Télévisions would not score substantially better. Finally, older internal polling for Rai showed that only 4.1 per cent of respondents believed Rai to be ‘outside of politics’, which I take to be equivalent to ‘politically independent’.17 These responses correspond with the broad- casters’ ranking and relative distance shown in Table 1.

FOUR POTENTIAL EXPLANATORY FACTORS

The potential explanatory factors identified below are, with the exception of de jure independ- ence, all derived from previous work on the politics of the media. Not all, however, are mutually compatible, or enjoy the same logical standing. Foreshadowing the descriptions below, explanation in terms of the partisanship of the bureaucracy works by altering politi- cians’ perceptions of the appropriateness of interfering; it therefore falls within a ‘logic of appropriateness’ model.18 Explanations in terms of the polarization of the party system and the size of the market for news function by altering PSB journalists’ attitude to their work and, a fortiori, politicians’ perceptions of journalistic output at the PSB. When politicians’ perceptions are positive (or at least not negative), their motives for interfering are reduced (or, more dryly, the net utility derived from interference relative to the status quo ante decreases). Finally, explanations, in terms of de jure independence, work by partially constraining politicians’ attempts to sanction the broadcaster, given adverse views of PSB output. Where attempts to interfere are severely constrained, the politician cannot interfere, no matter how pressing the motive. These three explanations fall within a (constrained) consequentialist logic. Since de jure independence acts as a constraint after other factors have provided politicians

with motives to interfere, one might think that its influence should always be felt. However, legal provisions are never entirely clear or entirely self-enforcing. Where other factors, such as the size of the market for news and party system polarization, furnish politicians with motives to interfere, their interpretation of de jure provisions may be determined more by their motives than by disinterested interpretation of the relevant text. That non-institutional background factors in society have usually been held to be determining in this way is strongly suggested by the way in which typologies of media systems and descriptions of public broadcasters have tended to use geographic and social labels, rather than institutionalist ones, to describe different media systems. This tendency would be more lamentable if further

14 MORI, ‘Quantitative Research to Inform the Preparation of the BBC Charter Review 2004’, report of a research study conducted on behalf of the Department for Culture, Media and Sport, 2004.

15 COMPAS, ‘Attitudes Toward Broadcast Issues, Canadian Content and the CBC’, survey carried out for the Friends of Canadian Broadcasting, 1999, p. 7.

16 CSA/Marianne, ‘L’image des journalistes et l’objectivité des médias’ (27 February 2003). 17 Istituto Eurisko and Maria Pia Montesi, ‘Immagine della RAI’ (internal company document, 1988). 18 Johan P. Olsen and James G. March, ‘The Logic of Appropriateness’, Arena Working Paper 04/09

(Centre for European Studies, University of Oslo, 2004).

Explaining the De Facto Independence of Public Broadcasters 79

comparative work had perpetuated it; unfortunately, much ostensibly comparative work is primarily based on the experiences of either Britain or the United States.

BUREAUCRATIC PARTISANSHIP

The one factor which has been cited in a comparative work on PSB is the partisanship of the bureaucracy. ‘Since a public national broadcasting corporation exists in the same normative framework and in the same political arena as the government bureaucracy, it is likely to have some features in common with it.’ Etzioni-Halevy hypothesizes: ‘countries that have party- politicised bureaucracies are also more likely to have party-politicised public broadcasting corporations as compared with other countries where the bureaucracies have become largely non-partisan.’19 Suppose that, subsequent to the establishment of a public broadcaster, politicians engage in a search for appropriate models to guide their behaviour towards the broadcaster, using an availability heuristic.20 Given that PSBs in certain respects resemble government departments more than they do corporations, politicians may use the same model that they use to govern their relationships with bureaucrats in their subsequent relationship with broadcasters. Where the model of politician–bureaucrat relations involves party-politicization, politicians will be used to giving partisan orders to bureaucrats, and will subsequently give such orders to broadcasters (and will expect them to be carried out). Where, by contrast, the model of politician–bureaucrat relations involves ‘professionalized’ bureaucracy, politicians will be used to having bureaucrats disallow certain orders as incompatible with basic norms of professional conduct, and will thus not make partisan requests of broadcasters, and/or will accept broadcasters rebuffing such requests. To operationalize the partisanship of the bureaucracy, I use data from the International

Institute of Management Development, which includes in its survey of business executives a question on whether ‘public service y is [or is not] independent from political inter- ference’.21 Scores range from 1 to 6; higher scores indicate greater independence.

SIZE OF THE MARKET FOR NEWS

The size of the market for news differs widely between European countries. In Sweden and Switzerland, almost 400 newspapers are printed for every 1,000 people. In Spain and Italy, the figure is closer to 100.22 These relative differences are stable across time and correlated with past growth in literacy.23 They also have important consequences for the perception of journalism in a country. A smaller market for news means that newspaper

19 Etzioni-Halevy, National Broadcasting under Siege, pp. 8–9. 20 Daniel Kahneman and Amos Tversky, ‘Availability: A Heuristic for Judging Frequency and

Probability’, in Daniel Kahneman, Paul Slovic and Amos Tversky, eds, Judgement under Uncertainty: Heuristics and Biases (Cambridge: Cambridge University Press, 1982).

21 International Institute for Management Development, IMD World Competitiveness Yearbook 2006 (Lausanne: International Institute for Management Development, 2006). In so far as PSBs are public services, respondents’ answers to this question may be affected by the independence of the PSB, causing endogeneity problems. However, the context of the question, which is preceded and followed by questions on state administration and bureaucracy, suggests that respondents’ primary focus when answering will be on state bureaucracy.

22 UNESCO, ‘Daily newspapers: Total average circulation per 1,000 inhabitants’, http://stats.uis. unesco.org/ (2007).

23 Daniel C. Hallin and Paolo Mancini, Comparing Media Systems: Three Models of Media and Politics (Cambridge: Cambridge University Press, 2004), p. 64.

80 HANRETTY

publishers facing fixed costs will struggle to break even or profit from sales and adverti- sing revenue alone.24 Consequently, publishers who remain in business will either not have to care about profit or take subventions from other interested groups.25 Examples of the former case include newspapers owned by political parties or large industrial groups (the so-called editori impuri in the Italian case). These publishers or groups will seek favourable coverage in exchange for their subventions, either by pressuring existing journalists or by recruiting on the basis of ideological affiliation. As Giovanni Bechelloni writes of the Italian case, ‘[p]olitical–ideological militancy became, from the twenties and thirties onward, a winning card for entry into the profession. This conditiony has meant that the body of Italian journalists has been, in large part, made up of successive cohorts of individuals selected on the basis of their political militancy’.26 Such conditions make it difficult for the idea of objective or impartial journalism to flourish.27 Journalists are more likely to adopt an idea of journalism as a politically engaged activity.28 Patterson and Donsbach found that 74 per cent of Italian journalists rated ‘championing values and ideas’ as very or quite important in their work, compared to 45 per cent in higher-circulation Britain and 36 per cent in Sweden with still higher circulation.29

This has consequences for the type of journalism practised by the PSB. The more partial the journalists for hire – partial in the sense that they champion certain values and ideas which are not widely held across society but instead held by particular identifiable groups within society – the more partial the resulting stock of journalists within the PSB, other things being equal. The more partial the stock of journalists within the PSB, the more likely politicians are to see the output of the PSB as a biased activity and, as such, potentially damaging to them and worthy of sanctioning. Conversely, the larger the market for news and the better established the idea of objective or impartial journalism, the more likely it is that PSB journalists will be committed to such ideals, and the more likely politicians are to see the output of the PSB as an attempt to report the facts. Facing negative coverage from today’s facts, they may believe that tomorrow’s facts will lead to better coverage; or, they may believe that sanctioning in response to such negative coverage will be ineffective if journalists are strongly committed to the idea of reporting the facts. These developments are, of course, mediated by the strategies and the agency of the

executives of the broadcaster. Broadcasters may be better at hiring impartial journalists; or may constrain journalists by adopting certain procedural rules governing their journal- ism. In its most extreme form, they may require journalists to read out news supplied by other companies. Thus, the BBC, Sweden’s Radiotjänst, and Denmark’s Statsradiofonien all began by broadcasting the news they were supplied with by the press agencies. Agency influence persisted: as late as 1936 the Ullswater Committee on Broadcasting was able to state: ‘The BBC y is dependent for the bulk of its news on four commercial

24 James Hamilton, All the News That’s Fit to Sell: How the Market Transforms Information into News (Princeton, N.J.: Princeton University Press, 2004).

25 Maria Petrova, ‘Newspapers and Parties: How Advertising Revenues Created an Independent Press’. Unpublished working paper (2007), available from http://fir.nes.ru/en/people/professors/Documents/ newspapers_parties_12-12-08.pdf.

26 In Alessandro Mazzanti, L’obiettivita giornalistica: un ideale maltratto (Naples: Liguori, 1991), p. 14. 27 Mazzanti, L’obiettivita giornalistica, pp. 49, 91, 189. 28 Daniel Hallin and Styliano Papathanassopoulos, ‘Political Clientelism and the Media: Southern

Europe and Latin America in Comparative Perspective’, Media, Culture & Society, 24 (2002), 175–95. 29 Thomas Patterson and Wolfgang Donsbach, ‘News Decisions: Journalists as Partisan Actors’,

Political Communication, 13 (1996), 455–68.

Explaining the De Facto Independence of Public Broadcasters 81

agencies’;30 and in Denmark, Radioavisen’s news was supplied by the press agencies until the 1960s. Consequently, the mould for public broadcast journalism was set early by press agencies specializing in a terse, factual style. Yet these news agencies are themselves a product of a large market for news: whilst the BBC in 1936 could count on four news agencies to supply them with material, EIAR, Rai’s predecessor in Italy, could only count on one agency, Agenzia Stefani, which supplied only ‘insufficient and stale’ dispatches.31

To operationalize the size of the market for news, I have used the natural log of daily newspaper circulation per 1,000 population for each country in 1975. That year is used since reliable data are available for it, and later data would in certain cases be later than the early years of the broadcaster (thus implying reverse causation), and because there is some comparable data for countries which did not exist in 1975.32 Newspaper circulation is used because newspapers are the medium which concentrates most on news instead of entertainment generally, because markets in radio and television have for much of the period been state monopolies or otherwise restricted markets, and because newspapers have had a historical role in setting norms for journalism, both in general and through operating as recruitment pools for companies in other media. Because newspaper circula- tion in Communist countries was artificially inflated, I also use a dichotomous control variable to account for this inflation.33

PARTY SYSTEM POLARIZATION

Party system polarization has been cited both as a correlate of certain media systems and as a specific explanation of low levels of independence in PSBs. As Mancini and Hallin write, ‘the notion of politically neutral journalism is less plausible where a wide range of competing world views contend’.34 Their view is echoed by PSB journalists and managers. Ruggero Orlando, Rai’s London correspondent in the1950s, wrote:

Whereas in Britain, the existing agreement of ‘90% of the people on 90% of the issues’ (Sir Winston Churchill’s figures) leaves ample scope to ‘non-controversial’ political broadcasts, the

30 Ullswater Committee on Broadcasting, Report of the Broadcasting Committee 1935, Cmd. 5091 (London: Her Majesty’s Stationery Office, 1936).

31 Valerio Castronovo, Luciana Giacheri Fossati and Nicola Tranfaglia, La stampa italiana nell’eta liberale (Bari: Laterza, 1979), p. 64.

32 Data for the Baltic countries were taken from Svennik Hoyer, Epp Lauk and Peeter Vihalemm, eds, Towards a Civic Society: The Baltic Media’s Long Road to Freedom (Tartu: Baltic Association for Media Research/Nota Baltica, 1993); data for the Czech Republic, Slovakia, Flanders and the French language community in Belgium were imputed from parent countries. It is likely that the model fit would have improved had separate 1975 data for these latter four areas been available; the Belgian French-language broadcaster has lower levels of independence than predicted, and the French language community reads fewer newspapers than Flanders; the same is true for Slovakia compared with the Czech Republic.

33 It might be thought that this control variable is in fact be an independent variable in its own right, reflecting either the legacy of a Leninist philosophy of the press (cf. Andrew Milton, ‘Bound but Not Gagged: Media Reform in Democratic Transitions’, Comparative Political Studies, 34 (2001), 493–526) or turbulence related to processes of democratization. First, for all that they were under Soviet influence, it would be a mistake, however, to think that a Leninist philosophy of the press applied equally well in Poland, Estonia and Russia; or to think that stated commitments to such a philosophy were incompatible with the development of professional norms (see Jane Lefwich Curry, Poland’s Journalists: Professionalism and Politics (Cambridge: Cambridge University Press, 1990)). Secondly, it is not the case that democratization-related turbulence led to inflated values of TOR or VUL: of the shortest-serving executives in each of the nine post-Communist countries included in the analysis, six started as chief executive after 1999; the remaining three started between 1991 and 1994.

34 Hallin and Mancini, Comparing Media Systems, pp. 59, 61.

82 HANRETTY

disagreement between government and pro-Communist opposition runs deep to the founda- tions of the national constitution which makes it very difficult, and often impossible, to plead absolute impartiality.35

The sentiment is not restricted to Italy. Oliver Whitley, chief assistant to the BBC Director-General, claimed that ‘the nation divided always has the BBC on the rack’.36

The intuitive idea is that politicians react badly to claims repeated on radio or television that are far from their own preferred positions; they assume that anyone who repeats what they see as such a mistaken claim – a claim which involves what they see as fundamentally incorrect views on the constitutional settlement or nature of democracy – cannot merely be presenting views impartially, but must thereby be endorsing them. Journalism which repeats such claims can defend itself when such claims are rare, as they are in party systems with limited polarization; but journalists cannot so defend themselves when they report claims made by politicians in polarized party systems, where claims made are by definition far apart. Consequently, politicians react by interfering in broadcasting to avoid such objectionable claims being made in the future. This may happen even when the broadcaster repeats the views of extreme positions in a scrupu- lously fair fashion. At one point, the PSB may be attacked from the left; at another, from the right. Politicians react in this way – runs the argument – not because they are con- cerned about medium- or long-run balance, but because they believe that some in the broadcasting staff hold deeply mistaken views which make a mockery of impartiality. Outside of Britain and Italy, this idea seems less applicable. Mancini and Hallin claim

that political journalism in Spain is similar to that in Italy because ‘Spain and Portugal y had a form of polarized pluralism y during brief periods of democracy early in the twentieth century’,37 but it is unclear why polarized views ninety years ago should affect independence now. Adopting the polarization of the party system as an explanation also means grouping Italy with other countries frequently described as polarized, such as Finland, where, despite a tendency towards self-censorship,38 the broadcaster is relatively independent, according to the measures presented in Table 1. To operationalize polarization, I have used the unweighted range of party left–right

scores for each country for parties reported by Huber and Inglehart.39 Huber and Inglehart’s measures are used in preference to other data, since they track party position, rather than policy stands: to the extent that polarization is thought to affect independ- ence, it is through the broadcasting of extreme politicians’ statements, rather than politicians’ extreme policy proposals.

DE JURE INDEPENDENCE

Perhaps because it has sometimes been used as a proxy for de facto independence, the idea of de jure independence has often been unclear. De jure independence must refer to legal

35 Ruggero Orlando, ‘Letter to Harman Grisewood, Director, BBC Spoken Word’, BBC Written Archives Centre File E1/1008/3 (11 July 1954).

36 Quoted in Asa Briggs, Governing the BBC (London: BBC, 1979), chap. 1, p. 20. 37 Hallin and Mancini, Comparing Media Systems, p. 61. 38 Ari Heinonen, ‘The Finnish Journalist: Watchdog with a Conscience’, in David Weaver, ed., The

Global Journalist (Cresskills, N.J.: Hampton Press, 1998). 39 John Huber and Ronald Inglehart, ‘Expert Interpretations of Party Space and Party Locations in 42

Societies’, Party Politics, 1 (1995), 73–111.

Explaining the De Facto Independence of Public Broadcasters 83

provisions, but which ones? It may seem tautological to suggest that one should include those legal provisions which are thought to have a bearing on actual, de facto independ- ence, but this should draw attention to the fact that every empirical link between de jure and de facto independence itself depends on a theory of which legal provisions are thought to influence independence.40

Potential links between legal provisions and de facto independence are numerous. Provisions might involve moral suasion: the presence of a commitment to independence might shame politicians and/or journalists into living up to such a provision. Altern- atively, provisions might create (dis)incentives for independent behaviour – say, if licence fee settlements were decided by viewers’ panels. The most common links, however, are the extent of sanctioning methods and the possibilities for selecting executives.

Sanctioning

Assume that PSB executives’ utility is a function of their own income, autonomy and reputation, and the success (commercial or otherwise) of the broadcaster. Politicians can act so as to reduce (increase) these goods contingent on the broadcasters’ actions, thereby sanctioning (rewarding) PSB executives, who will come to anticipate sanctioning and avoid the kind of programming that leads to sanctions, thereby reducing independence. Where sanctioning mechanisms are heavily constrained by law, this effect will be negligible, and the broadcaster will consequently enjoy a high degree of de facto independence.

Appointments

Whilst sanctions operate through changing executives’ actions despite their preferences, appointments may reduce independence by changing the preferences of those at the top of the broadcaster. Where few (political) actors are involved in appointments, the closer the match between the ideal point of the appointing individual and that of the appointed person; conversely, where multiple (political) actors need to agree on appointments, the match between any appointing actor and the appointed person will be looser. Where appointments last for a long time, the greater is the potential that the appointed person’s preferences will diverge from those of the appointing agent over time, thus increasing the possibility for independence. Conversely, the more frequent and more easily revoked these appointments are, the stronger this effect.41

Table 2 presents thirteen indicators used to construct an index of de jure independence using these two links. These indicators are heavily reliant on prior work done by Gilardi and by Elgie and McMenamin.42 Items in the first column refer to sanctions; items in the second column refer to appointments. Each is based on legal provisions commonly found in the laws establishing PSBs. The first two indicators – reporting to government and reporting to parliament – are

already used in the literature. Where an agency must report to a political body frequently, the psychological costs of appearing before intense committee pressure and/or ques- tioning and media intrusion can operate as sanctions where members of the committee

40 Therefore, it is entirely possible that the low correlation observed in previous works between de jure and de facto independence results from a mis-specification of those legal features which affect de facto independence.

41 Powers of dismissal may, of course, also be interpreted as a method of sanctioning. 42 Gilardi, ‘Policy Credibility and Delegation to Independent Regulatory Agencies’; Elgie and

McMenamin, ‘Credible Commitment, Political Uncertainty, or Policy Complexity?’

84 HANRETTY

wish to sanction. Therefore, the more often legislation allows for such reports, the greater the opportunity for politicians to sanction the broadcaster. ‘Borrowing’ is another term used in Gilardi. The idea is this: due to the capital-intensive

nature of broadcasting, broadcasters occasionally require large capital expenditure which cannot be met by current revenue streams, but must be borrowed. Failure to invest will result in competitive disadvantage. Where borrowing requires ministerial permission, the minister can threaten to withhold permission either to punish the broadcaster for prior behaviour, or to blackmail the broadcaster and so influence future behaviour. In either event, where legislation requires the broadcaster to seek ministerial permission to borrow, the greater the opportunity for the relevant minister(s) to sanction the broadcaster.

TABLE 2 Thirteen Indicators of De Jure Independence

Sanctions Appointments

1. Reporting to government 1. Term of office of first executive group No reporting requirement: 1 More than six years: 1 Annual written reporting to executive: 0.66 Six years: 0.8 Annual in-person report to executive: 0.33 Five years: 0.6 Greater than annual in-person reporting: 0 Four years: 0.4

2. Reporting to parliament Less than four years: 0.2 No reporting requirement: 1 No fixed term: 0 Annual written reporting to parliament: 0.66 2. Appointing body for first executive group Annual in-person report to parliament: 0.33 Management board members: 1 Greater than annual in-person reporting: 0

3. Borrowing Complex mix of executive and legislature: 0.75

Borrowing unrestricted: 1 The legislature: 0.5 Borrowing requires ministerial permission: 0 The executive collectively: 0.25

4. New operations, sub-contracting One or two ministers: 0 No requirement for ministerial approval: 1 3. Dismissal of first executive body Require ministerial approval: 0 Dismissal not possible: 1

Dismissal for non-policy reasons: 0.55. Term of service contracts Dismissal at appointing body’s convenience: 0

Greater than six years: 1

4. Term of office of second executive group Six years: 0.8

More than six years: 1 Five years: 0.6

Six years: 0.8 Four years: 0.4

Five years: 0.6 Three years: 0.2

Four years: 0.4 Less than three years: 0

Less than four years: 0.2 6. Mechanisms for altering funding

No fixed term: 0 Automatically uprated licence fee: 1

5. Appointing body for second executive group Discretionally uprated licence fee: 0.75

Executive director or other independent organisation: 1

Advertising: 0.5

Complex mix of executive and legislature: 0.75

Pluriannual grant from parliament: 0.25

The legislature: 0.5

Annual grant from parliament: 0

The executive collectively: 0.25

7. State participation

One or two ministers: 0

Independent foundation: 1

6. Dismissal of second executive body

Non-majority state participation: 0.5

Dismissal not possible: 1

Total or majority state participation: 0

Dismissal for non-policy reasons: 0.5 Dismissal at appointing body’s convenience: 0

Explaining the De Facto Independence of Public Broadcasters 85

‘New Operations’ is not, to my knowledge, found in other indexes. Due to the high- technology nature of broadcasting, new operational techniques may become necessary, or old operational techniques obsolete; broadcasters may therefore need to adopt new techniques or jettison old ones in order to avoid competitive disadvantage. As with borrowing, where new operations require ministerial permission, the minister can threaten to withhold permission either to punish or blackmail the broadcaster for prior behaviour, thereby influencing its future behaviour. ‘Contract term’ is specific to public service broadcasting. Some public broadcasters are

established in perpetuity, as is their qualification as a public service broadcaster. Other public broadcasters are established as companies for a limited time; or their public service concession is time-limited; or, increasingly, agreements regulating the public service concession are set for a limited number of years.43 Where such contract terms are limited, the broadcaster is more likely to face periodic intrusion into its operations, and the threat of sanctioning. Funding is obviously of key importance. Where the funding mechanism guarantees a

secure revenue stream, the de jure independence of the broadcaster will be high, as politicians cannot intervene arbitrarily to reduce company revenue. Where the funding mechanism guarantees a secure revenue stream but future increments are subject to political discretion, the broadcaster will be less secure; where the funding mechanism generates a short-term income stream only, and where continuation of that income stream is subject to political discretion, the opportunities to sanction will be numerous. Finally, state participation in the ownership of a company may generate opportunities for sanctioning, where, if the broadcaster is structured as a normal public company with share ownership, the requirement that certain resolutions be approved at annual general meetings may give the share-holding minister power over the broadcaster. (In recognition of this possibility, some laws – i.e., the Portuguese law on radio and television restructuring – explicitly forbid the minister this power.) The second group of indicators, ‘Appointments’, is largely drawn from Elgie and

McMenamin, comprising three indicators (‘Tenure’, ‘Appointing Body’, and ‘Dismissal’) for the first and second executive bodies. An example of a first, or upper executive body, is the former Board of Governors of the BBC or the Fernsehrat of ZDF; an example of a second executive body is the post of Director-General, Intendant, or, more rarely, a multiple-member executive board.44

Broadcasters were first assigned scores for each indicator on the basis of the appropriate legislative text.45 The two groups of indicators – ‘Appointments’ and ‘Sanctions’ – were then

43 Tomas Coppens and Frieda Saeys, ‘Enforcing Performance: New Approaches to Govern Public Service Broadcasting’, Media, Culture & Society, 28 (2006), 261–84.

44 In cases where the number of members was large and the methods used to appoint them divergent, I have assigned different scores for some part of the board, and averaged these scores. For example: in Italy following the passage of the 1975 reform of Rai, six members of the sixteen member administrative council were nominated by the majority shareholder (the state, coded here as the executive), whilst the remaining ten members were nominated by a parliamentary committee. The score is therefore equal to [(63 0.25)1 (103 0.75)]/16.

45 Information on legislative standing was taken from: Australia: The Australian Broadcasting Corporation Act 1983, as amended; Austria: 1966 Broadcasting Act; Belgium (Flemish-language community): ‘Décret relatif à la transformation de la BRTN en une société anonyme de droit public’ (29th April 1997) (French-language translation); Belgium (French language community): ‘Décret du 14 juillet 1997 portant statut de la Radio-Télévision belge de la Communauté française (RTBF)’; Bulgaria: The

86 HANRETTY

averaged, and the average of the two groups taken as the overall value for the broadcaster’s de jure independence. Alternative specifications of the index of de jure independence do not change the results appreciably. Gran and Patterson argue that the de facto independence of agencies depends not on the mean value of legislative guarantees of independence, but on the minimum, since politicians will attack the ‘weakest link’.46 Yet countries which score highly on Appointments also score highly on Sanctions; the correlation between the two terms is 0.91. The specification also matches expert attempts to insulate PSBs from political pressure: the score for one model law on public broadcasting would be the second-highest in the sample, beaten only by Switzerland.47

METHODOLOGY

The universe of public service broadcasters operating at national level in consolidated democracies consists of approximately fifty cases. From these, I have drawn an availability sample of the thirty-six PSBs listed in Table 1, periodicized according to the duration of the legislative text. The sample is an availability sample in that broadcasters were excluded if they lacked information on legislative standing and/or turnover. Values for the partisanship of the bureaucracy were imputed in five cases using the mean.48

This sampling is likely to underestimate the number of low independence broadcasters, since a number of the excluded PSBs have low values of I: numerous government changes in Turkey have been followed by a change in the director-general of TRT, and the average

(F’note continued)

Radio and Television Law of 1998; Canada: the 1991 Broadcasting Act; Chile: law 19/132 of the 30th March 1992; Czech Republic: 483/1991 Coll., Act of the Czech National Council of 7 November 1991 on Czech Television and Metykova (2005, ch. 4); Denmark: Law no. 215 of the 11th June 1959, law no. 421 of the 15th June 1973, and law no. 374 of the 10th June 1987; Estonia: The Broadcasting Act of 1994, as amended; Finland: Act on Yleisradio Oy of 1993, as amended; France: Loi ordinaire 74-696 du 07/08/ 1974 relative à la radiodiffusion et télévision, loi n8 82-652 du 29 juillet 1982 sur la communication audiovisuelle; Germany: Statute of ZDF (April 1962), as amended; Hungary: Act on Radio and Tele- vision Broadcasting, act I of 19th May 1996; Israel: Israel Broadcasting Authority Law, no. 5725 – 1965; Japan: The Broadcast Law, no. 132 of 2nd May 1950, as amended; Latvia: Radio and Television Law (consolidated); Lithuania: Law on the National Radio and Television, No I-1571 of October 1996, as amended; New Zealand: The Television New Zealand Act 2003; Norway: The Broadcasting Act of 4 Dec. 1992 no. 127, as amended; Poland: The Broadcasting Act of the 29th December 1992; Portugal: the Television Broadcasting Act, Law no. 32/2003; Republic of Ireland: The Broadcasting Act, 2001; the Broadcasting Act, 1976; the Broadcasting Authority Act, 1960; Romania: Law no. 41 of June 17 1994, ‘On the Organization and Operation of the Romanian Broadcasting Corporation and of the Romania Television Corporation’; Slovakia: The Act on Slovak Television and Monika Metykova, ‘Regulating Public Service Broadcasting: The Cases of the Czech Republic, Slovakia and Ireland’ (doctoral dis- sertation, Masaryk University, Brno, 2005), chap. 4; Slovenia: The Law on Radio and Television of 1994; South Africa: the Broadcasting Act 1999; Sweden: Information made available by SVT; Switzerland: ‘Statuto della Società svizzera di radiotelevisione (SRG SSR idée suisse)’, of 22 November 1991, as amended; United Kingdom: Successive Royal Charters and Agreements with the BBC; United States of America: The 1967 Public Broadcasting Act, as amended.

46 Brian Gran and Robin Patterson, ‘Law and Weak Links of Independence: A Fuzzy-Sets Analysis of Children’s Ombudspersons’ (unpublished working paper), available from http://www.northwestern.edu/ rc19/Gran.pdf; Garry Goertz, Social Science Concepts: A User’s Guide (Princeton, N.J.: Princeton University Press, 2006), pp. 129–39.

47 Werner Rumphorst, ‘Model Public Service Broadcasting law’ (Geneva: UNESCO, 1999). 48 Results from the regression without imputation did not differ significantly; they are available on

request.

Explaining the De Facto Independence of Public Broadcasters 87

tenure of directors-general of ERT in Greece during the 1980s was eight months.49

Nevertheless, such omitted cases seem to be in line with theoretical expectations: both Turkey and Greece have extremely low newspaper circulation per population. Periodicization of the broadcasters according to legislative standing may seem artificial,

given the historical continuity present in the development of the market for news. Given the theoretical interest in identifying the effect of de jure independence, however, period- icization does seem necessary; as for the remaining variables which vary less over time, one can only note that:

one faces continuous problems of ‘aggregation’ of the temporal units y [if] for one aspect of the research the temporal unit of reference is the legislative period, those properties/variables which can be evaluated only at the temporal level of the regime will have to be considered as constant for each legislative period internal to the regime.50

The shortest period under consideration is the nine years of France Télévisions under Loi no. 696/1974; the longest is the BBC (1944–2006). I have used multivariate ordinary least squares (OLS) linear regression to estimate the

relative effects of the four potential explanatory factors listed above. The use of a dependent variable bounded between zero and one means that checks for hetero- scedasticity are necessary. Therefore, I have reported statistics for the Breusch–Pagan test alongside each model.

RESULTS

Table 3 shows the results of the regression analysis. The first column of data shows the full model, with all four main explanatory variables and one interaction term included. With an n of 36, this model has too high a ratio of variables to observations. It also faces problems of heteroscedasticity. It is, however, clear from the model that the partisanship of the bureaucracy and the degree of polarization add little to the model, and can safely be removed. The second column shows the reduced model, with all three terms significant at the 0.05 level. All coefficients are in the predicted direction; the value of Cook’s distance is less than 0.25 for all cases, and problems of heteroscedasticity faced by the previous model disappear. The resulting model explains over half of the variance in PSB independence.51

If the model is broadly correct – and if the unexplained variance is not the result of some other systematic factor – then the prospects for designing independent public service broadcasters are relatively good. Assuming average press circulation, an increase from no legal protection to full legal protection would result in an increase from negligible to almost total de facto independence. Of course, these values are never observed in real life, as every PSB law grants the broadcaster some protection. Within the range of observed values, however, the difference between the minimum de jure score of 0.15 and the maximum of 0.82, is, nevertheless, still large. We can understand the substantive

49 Georgia Chondroleou, ‘Policy Networks in Comparative Perspective: Media Policy Networks in Britain and Greece’ (paper presented at the ECPR General Conference, 2001).

50 Stefano Bartolini, ‘On Time and Comparative Research’, Journal of Theoretical Politics, 5 (1993), 131–67, at p. 148.

51 An additional model including audience share – on the basis that politicians may avoid interfering in PSBs which dominate the market lest they be seen as engaged in a power grab – was also tested, but the effect of ‘share’ was not significant. I thank an anonymous reviewer for suggesting this possibility. Results are available on request.

88 HANRETTY

significance of these coefficients better if we apply them to a real-world example of PSB reform. In Spain, the Zapatero government, following a report by a committee of sages, drafted a new law on RTVE establishing a single-tier board of twelve members (nominated by the parliament for a non-renewable six-year term), one of whom would be the president of the new corporation. The reformed legal position scores much higher than the previous law of 1980 (0.8 compared with 0.5). The likely effect of the law on RTVE’s independence will be positive: from an independence score of 0.49, the mean predicted independence score would rise to 0.78 (standard deviation of 0.068), fulfilling the intent behind the legislation of liberating RTVE from excessive governmental control.52

CONCLUSION

Given the magnitude of the effects of de jure independence noted above, the potential for designing independent PSBs seems great. If, as I claimed at the beginning, PSB independence is both desired and desirable, then this conclusion should be heartening: within the limits of what can be explained by existing theories, an important determinant of independence – legislative provisions – is easily tractable. Questions of institutional design, however, are rarely simple. Designers often have a multiplicity of goals in mind; there is no guarantee that PSB independence is compatible with other desirable characteristics of the broadcaster. Additionally, asking about the capacity of institutional design to resist politicians’ current desires to interfere in the PSB begs the question of whether these institutions are, as Riker memorably put it, ‘congealed preferences’,53 and, if so, just how congealed? If institutional designs intended to preserve independence could easily be rewritten, their worth is greatly diminished. Fortunately, the de jure independence granted by PSB reforms seems only to increase: reforms of the French, Italian and Spanish laws on PSB have all granted greater de jure independence to the broadcaster. Thus, de jure PSB independence tends to congeal upwards, a positive finding if PSB independence is indeed valued as I have suggested.

TABLE 3 Regression Analysis with Full and Reduced Models

Full model Reduced model

Coeff. p-value Coeff. p-value

Constant 20.119 0.665 20.125 0.598 De jure independence 0.322 0.087 0.411 0.020* log(Circulation) 0.109 0.051 0.125 0.007** Bureaucratic partisanship 0.031 0.142 Polarization 0.001 0.951 Log(Circulation) * PostCommunist 20.037 0.000*** 20.038 0.000***

N 36 36 Adj. R-squared 0.580 0.537 Breusch-Pagan 12.150 (5 df) 0.033* 3.5 (3 df) 0.321

*, **, *** indicates significance at the 95%, 99%, and 99.9% level, respectively.

52 Bootstrapped estimates after 1,000 iterations obtained using Zelig, by Kosuke Imai, Gary King and Olivia Lau, ‘Zelig: Everyone’s Statistical Software’, http://gking.harvard.edu/zelig (2006).

53 William Riker, ‘Implications from the Disequilibrium of Majority Rule for the Study of Institutions’, American Political Science Review, 74 (1980), 432–46, at p. 445.

Explaining the De Facto Independence of Public Broadcasters 89