Journal of Asian and African Studies

Journal of Asian and African Studies http://jas.sagepub.com/

Scandals, Media and Good Governance in China and Kenya Jonathan Hassid and Jennifer N Brass Journal of Asian and African Studies published online 13 May 2014 DOI: 10.1177/0021909614528865

The online version of this article can be found at: http://jas.sagepub.com/content/early/2014/05/12/0021909614528865

Published by:

http://www.sagepublications.com

Additional services and information for Journal of Asian and African Studies can be found at:

Email Alerts: http://jas.sagepub.com/cgi/alerts

Subscriptions: http://jas.sagepub.com/subscriptions

Reprints: http://www.sagepub.com/journalsReprints.nav

Permissions: http://www.sagepub.com/journalsPermissions.nav

Citations: http://jas.sagepub.com/content/early/2014/05/12/0021909614528865.refs.html

>> OnlineFirst Version of Record - May 13, 2014

What is This?

Downloaded from jas.sagepub.com by guest on May 14, 2014 JAS0010.1177/0021909614528865Journal of Asian and African StudiesHassid and Brass 528865research-article2014

Article J A A S

Journal of Asian and African Studies 18–1 Scandals, Media and Good © The Author(s) 2014 Reprints and permissions: Governance in China and Kenya sagepub.co.uk/journalsPermissions.nav DOI: 10.1177/0021909614528865 jas.sagepub.com

Jonathan Hassid University of Technology, Sydney, Australia

Jennifer N Brass School of Public and Environmental Affairs, Indiana University, USA

Abstract Conventional wisdom holds that democratic governments listen to their populations, while authoritarian governments do not. This paper questions the extent to which this dynamic applies in cases of government scandals, using the illustrative cases of China and Kenya. We expect democratic countries with free media to be responsive and authoritarian states to ignore public pressure. Counter to this expectation, however, authoritarian China is more responsive to public pressure to clean up scandals than democratic Kenya. Using case studies and quantitative analysis, we argue that while democracy and free media are important for government responsiveness to scandal, they are neither necessary nor sufficient conditions. We assert that political will, state capacity to respond and high public expectations for state action are also necessary.

Keywords Scandal, regime type, governance, public expectations, political will, state capacity, China, Kenya

In 2007, central China’s Henan TV revealed the existence of illegal brick kilns in Shanxi province where kidnapped children were forced to work long hours as industrial slaves. After reports spread across the Chinese media, the kilns were shut down and dozens of colluding government officials were fined, jailed or executed, despite the general reluctance of the Chinese Communist Party (CCP) to hold corrupt party officials accountable at the time (Zhu Zhe, 2007). In contrast, just nine months earlier and half a world away in Kenya, the Goldenberg scandal, festering since 1993, reached its ignominious end with news that the no one would face any charges related to the disappearance of nearly US$600 million, despite years of media pressure and the creation of an anti-corruption watchdog agency. After the verdict, leading newspaper the Daily Nation (2006) proclaimed, ‘the war on graft is, indeed, lost.’

Corresponding author: Jonathan Hassid, China Research Centre, University of Technology, P.O. Box 123, Broadway, NSW 2007, Australia. Email: [email protected]

Downloaded from jas.sagepub.com by guest on May 14, 2014 2 Journal of Asian and African Studies

These contrasting anecdotes pose a challenge to the received wisdom that democracy and free media are necessary for good governance. Political and civic freedoms are core Western values, strongly promoted by international development organizations. Yet in the developing world, countries like China and Kenya appear not to fit the mold. Democratic countries with a free media can sometimes have more unresolved scandals than authoritarian states with strict press control. We demonstrate below that democracy and free media are not required for government responsiveness to scandal. Drawing on case study evidence, we emphasize other necessary factors: citizens’ historically and culturally grounded expectations for their governments, as well as those governments’ willingness and capacity to respond. Where expectations are high and the government has the will and ability to respond effectively, authorities tend to react quickly – regardless of the degree of electoral or media freedom. In contrast, in countries where few expect the state to take scandals seriously, or the state lacks the ability or political will to enforce laws, democratic institutions and free media are often insufficient. Under such conditions, scandals can smolder for years, defying the efforts of activists and the predictions of social scientists. Understanding the role of citizen expectations and state capacity allows us to deepen our comprehension of the conditions under which governments respond to scandals, and it expands our understanding of regime types. Our evidence suggests that some features social scientists normally consider characteristics of democracies can in fact sometimes be observed in autocratic regimes. This observation matters when we attempt to apply theory developed in wealthy Western countries to the developing world. Rather than look at median countries, we deliberately chose outliers on the spectrum of states: China is an unusual autocracy and Kenya a weak democracy. This study of deviant cases using the crucial-case method (Gerring, 2001), allows us to ‘inductively identify variables and hypotheses that have been left out of existing theories’ (Bennett, 2004: 38). In other words, by selecting such different cases, we show the inadequacy of common theories about how democracies and dictatorships function. To arrive at these conclusions, we first outline understandings of scandal and of the relationships among media freedom, democracy and government responsiveness. Next, we examine our cases, China and Kenya. Democratic Kenya does not respond effectively to scandals, despite its open and free media; authoritarian China swiftly responds once scandals are exposed, even under conditions of heavy media censorship. Drawing on a combined two years of field research in these countries, as well as an original survey from Kenya, we argue that while free media and democratic institutions are often correlated with responsive governance, they are not sufficient to produce it. Next, we suggest possible explanations, highlighting the role played by popular expectations, political will and state capacity. We conclude by calling for improved quantitative data collection to assess these claims better.

Understanding scandals, media and regime type Scandals, defined as ‘actions or events involving certain kinds of transgressions which become known to others and are sufficiently serious to elicit a public response’, often provoke government action (Thompson, 2000: 13). We call such state action in the face of public outcry ‘government responsiveness’. Note that we differentiate responsiveness and accountability, the latter concerning the ability of the people to punish state officials for mistakes. In other words, responsiveness refers to how the state reacts to the people, while accountability refers to how the people (can) react to the state. Social scientists and policymakers have long argued that democracy and press freedom encourage better state response to scandals. Besley and Burgess (2001), for example, claim that a developed press and high level of political participation make governments more responsive to popular

Downloaded from jas.sagepub.com by guest on May 14, 2014 Hassid and Brass 3

Table 1. Press Freedom and Elections.

Significant Press Freedoms Yes No Yes United States, Djibouti Kenya Elections No Bhutan (pre-2008) Syria, China (except

Holds Democratic at village level) demands, irrespective of GDP. A free press, they contend, allows information to flow and enables citizens to monitor political behavior. Specifically, media circulation levels and electoral accountability mechanisms are the main factors in determining government responsiveness. Mass media reports, they argue, incentivize official response (Besley and Burgess, 2002: 1415). Likewise, Freedom House (2011), a global good-governance watchdog, asserts that, ‘A free press plays a key role in … contributing to greater accountability, good government, and economic development’. In short, ‘Every dictator dislikes free media’ (Egorov et al., 2009: 645). The World Bank’s Robert Stapenhurst (2000) agrees, citing the crucial role of media in fighting and deterring corruption. World Bank scholars also contend that the media influence government responsiveness by acting as a ‘watchdog’ on public actors, protecting the public from incompetence, corruption and misinformation (Odugbemi and Norris, 2010: 380). According to the British Department for International Development, a free press helps check state abuse of power and allows citizens to make informed decisions when they vote (DFID, 2008). Most scholars and development agencies thus agree that a free press is critical to keeping politicians working for the public (Rose-Ackerman, 1999; Brunetti and Weder, 2003). In democracies, of course, voters have regular opportunities to remove unresponsive leaders from office. As one scholar notes, ‘elections discipline governments to be accountable to the elec- torate’ (Lines, 2009: 8). Markovits and Silverstein (1988) go further, arguing that political scandals are only possible in liberal democracies because only these have public involvement in governing.1 Other scholars focus on the relationship between citizen participation and responsiveness, finding that such participation is critical to the proper functioning of organizations (Hirschman, 1970; Kuklinski and Stanga, 1979). In principle, electoral consequences exist for politicians who refuse to clean house in the face of voter anger. While authoritarian governments can rarely completely ignore public unhappiness, democracies are structurally much more attuned to mass opinion. Drawing from these theories of the relationship between free media and democratic institutions on one hand, and government responsiveness to scandals on the other, we argue that there are four possible configurations of states, as shown in Table 1 below. In the upper left-hand corner are states that have both significant press freedom and democratic institutions; in the bottom right, neither press freedom nor democratic institutions. In the alternate corners are states with either press freedom or democratic institutions, but not both. We expect countries in the top left box to have the highest degree of government responsiveness, with those in the lower right having the least responsive government. Theory predicts that the countries in the alternate diagonals have a mixed response.

Examining the puzzle through case studies We used the crucial-case method to examine this puzzle, meaning that we selected cases in which results are ‘unexpected in light of the causal inference under investigation’ (Gerring, 2001: 220).

Downloaded from jas.sagepub.com by guest on May 14, 2014 4 Journal of Asian and African Studies

These cases emerged from conversations in which the authors, a China specialist and an Africanist, realized that the two countries respond very differently to uncovered scandals, but both with responses that challenge contemporary theory. Specifically, our ‘least-likely’ case (the case in which the theory is least-likely to hold), China, proves positive, while our likely case, Kenya, proves negative. Studies of crucial cases like these are important not only because they can discon- firm existing theory (Gerring, 2001), but also because they allow us to identify variables that are missed in the study of typical cases (George and Bennett, 2005). In other words, each of these two very different cases shows the inadequacy of common theories about how democracies and dictatorships function; combined, they also shed light on alternative explanations. In later sections of the paper, we draw on both cases to hypothesize alternate causal paths to government responsiveness. The fact that such divergent cases point to the same missing variables lends support to these hypotheses. Returning to Table 1, then, we see that Kenya is largely a functioning democracy with free media, whereas China is clearly authoritarian with highly restricted media. The Kenyan government regularly asserts its democratic standing. It lists democracy as a key aspect in achieving its forward-looking Vision 2030 plan (Government of Kenya, 2007). According to Freedom House, however, Kenya receives a score of 3.5, or partly free, compared to China’s 6.5, or not free, on a scale from 1 (most free) to 7 (least free) (Freedom House, 2010). This places China among the 17 least free countries in the world, but puts Kenya in the company of countries such as the Philippines, Bosnia-Herzegovina, and Guatemala.2 The ‘partly free’ ranking corresponds to survey research conducted by one of the authors in 2008, in which 93% of 501 respondents described the political system as a democracy, albeit one that 42.5% described as having ‘major problems’.3 Kenya also has a much more autonomous press, ranking 70th most free of 178 countries, compared to China’s position as 171st, according to watchdog Reporters Without Borders (2010). Both the Kenyan government and members of the press in Kenya publicly mention media freedom in the country (see, for example, Captial Reporter, 2010; The Star Editorial Board, 2012). We recognize the limits associated with selecting these two cases. First, we do not claim that the institutions of democracy function fully in Kenya, only that the country is significantly better than China on both counts – democracy and media freedom – and should therefore also be more responsive to media reports of scandal. Additionally, in spite of the fact that in 2013 a spokesper- son for the Foreign Ministry asserted, ‘China has no media censorship,’ Chinese media are generally not allowed to report on scandals that might prove embarrassing to the CCP.4 In 2008, as China prepared for the Olympics, for example, strict press censorship suppressed a Southern Weekend report on dangerous chemicals added to baby formula (Bandurski, 2008). Only after the Olympics ended – and thousands of babies had been hospitalized – did the story break. Nor is the Kenyan media always ineffective. Intense media coverage on the two murder trials of Tom Cholmondeley, grandson of a founder of British rule in Kenya, is thought to have led to a guilty verdict during his second trial.5 Contrary to accepted wisdom, however, we demonstrate that scandals are actually more likely to result in reported punishments in China than in Kenya. Qualitatively, we draw on some of the most notorious scandals of the past decade in the two countries. While not chosen at random, these stories are representative scandals for both countries. Quantitatively, we employ a computer content analysis of over five hundred newspaper articles pertaining to these and other scandals in both countries. To avoid problems with domestic press censorship, especially in China, we rely on international press coverage. This choice means that only the biggest scandals are likely to receive coverage. We feel this a worthwhile tradeoff since Chinese censorship rules ensure that many scandals, even serious ones, receive little press coverage in local news. Chinese journalists

Downloaded from jas.sagepub.com by guest on May 14, 2014 Hassid and Brass 5 who report scandals frequently face punishment, including dismissal, fines or imprisonment (Brady 2008). Scandals that do receive domestic press scrutiny, moreover, are generally already solved and the offending parties punished. Citizens often hear of government corruption only when it is announced by official news outlets. There are exceptions, but a reliance on the Chinese domestic press would erroneously imply that all corruption that comes to official attention is punished harshly. This is not true – only six percent of provincial disciplinary infractions are turned over to the court system, of which most receive minor sanctions (Wedeman, 2004). International press coverage is not ham- pered by these restrictions and is unlikely to vary systematically across international borders.

Scandals in China: call and response Among the most famous scandals in contemporary China is one involving college-educated migrant worker Sun Zhigang. In April 2003, police detained Sun in Guangzhou for not carrying his temporary residence permit. While in police custody, he was beaten to death by the facility’s staff. After the aggressive Southern Metropolis Daily reported the circumstances of Sun’s death, newspapers throughout China reprinted the article in the face of a national uproar. Pressure on the government was so great that the CCP repealed the twenty-year-old law authorizing the use of internal detention facilities. Public pressure triumphed, forcing the authoritarian regime to implement popular changes (Liebman, 2005; Hand, 2006). In many ways this case is the archetype of how the CCP responds to contemporary scandals (see Table 2 for a list of recent scandals and their outcomes). Cases generally begin when intrepid reporters discover and publicize official abuse of power, despite the dangers inherent in challeng- ing Chinese power holders (Hassid, 2011). Once a scandal has gained national attention, the CCP simultaneously increases censorship and acts decisively to solve the problem, ensuring that public anger is quickly mollified. The melamine milk scandal of 2008 fits this pattern. Although reporters and editors at the Southern Weekend knew in July 2008 that the dairy company Sanlu had for months been selling milk contaminated with melamine, heavy press controls during the Olympics kept it quiet until after the Games ended (Bandurski, 2008). Following an understandable domestic and international uproar, Beijing moved to contain the damage, executing two businesspeople and imprisoning six others, despite their close ties to the state (Zhu Zhe and Cui Xiaohuo, 2009). The head of China’s national Administration of Quality Supervision, Inspection and Quarantine was forced to resign, as was the powerful boss of Shijiazhuang city, the location of Sanlu’s headquarters (Xinhua News Agency, 2008). A similar result is evident in one case of high-level corruption in education, in which one of the highest-ranking officials in the rich city of Shenzhen facilitated the production and mandatory distribution to all Shenzhen middle schools of a film his daughter Niu Niu made. Schools were required to pay a hefty 20 RMB (US$3) per student to view the film. With a population of over 10 million people, this amounted to a substantial revenue stream in a country renowned for extremely inexpensive pirated DVDs (Dimitrov, 2009). After complaints appeared online, however, several of China’s bolder newspapers began investigating. They revealed that in addition to the viewing fees, Niu Niu obtained 7.69 million RMB (US $1.2 million) in investment capital from the government (via her father) (People’s Abstracts, 2004). After newspapers revealed the scale of corruption, the high-ranking official was put under investigation and quickly removed from his powerful position in Shenzhen. While in the past punishments were rarely meted out to very senior officials, the 2011 imprisonment of Railway Minister Liu Zhijun for ‘primary leadership responsibilities for the serious corruption problem within the railway system’ is an encouraging exception (Osnos, 2012). Liu, known

Downloaded from jas.sagepub.com by guest on May 14, 2014 6 Journal of Asian and African Studies

Table 2. Major Scandals and Their Outcomes.

Country Scandal Type of Scandal Outcome Kenya Goldenberg (1993 to Corruption Some involved granted immunity or present) exemption from prosecution. Saitoti resigns; reappointed late that year. Case is still pending; plea bargains may be negotiated. No one convicted. Anglo Leasing (2004) Corruption Private firm returned the full down payment it received. Fizzled court cases. No convictions. Grand Regency (2008) Corruption Then-Finance Minister Kimunya is censured and resigns; reappointed within six months as Minister of Trade. No convictions. Tom Chomondeley killings Murders Chomondeley imprisoned for manslaughter. Human Rights Activists Murder No arrests or convictions. Murdered (2009) China Sun Zhigang (2003) Murder Abolition of internal migrant detention facilities, prison terms for reporters and editors who broke the story. Niuniu (2004) Corruption Dismissal of corrupt official (Li Yizhen) from his powerful position, and ultimate removal from government service several years later. Melamine contaminated Corruption Dismissal, imprisonment or execution milk scandal (2008) of responsible officials; supposed improvement in inspection procedures, though contaminated milk continues to be found. Liu Zhijun (2011) Corruption Suspended death sentence for corruption. Zhou Yongkang (2013) Corruption Dismissal of associates, investigation of very senior national politician. as ‘boss rail,’ was for years seen as untouchable because of his institutional power and high-level political connections. He even kept the media in thrall, in part through a system that distributed otherwise unobtainable rail tickets to reporters during peak travel times.6 Eventually, however, media pressure grew and the aftermath of a crash in 2011on China’s vaunted high-speed rail opened the floodgates to negative coverage. After the media uncovered Liu’s ostentatious corruption, including his having 16 cars, more than 350 apartments and 18 mistresses, he was sentenced to a (suspended) death sentence that in practice will result in a prison sentence of at least ten years (Kaiman, 2013). More recently, the new administration led by Xi Jinping has aggressively moved to ‘swat tigers as well as flies’ and to pursue major corruption cases – the so-called ‘tigers’ – that were hitherto ignored. The investigation of former ‘security czar’ Zhou Yongkang and his associates announced in late 2013 is a powerful move in this direction, as ‘it is the first time since the founding of the People’s Republic of China that an official who has held such high office has been the focus of a formal corruption investigation’ (Ansfield and Buckley, 2013). Zhou was previously a member of the nine-person Politburo Standing Committee, the highest political agency in China and one

Downloaded from jas.sagepub.com by guest on May 14, 2014 Hassid and Brass 7 whose members were previously considered sacrosanct. Although in the past the CCP has proved reluctant to investigate graft cases involving those at the very top of government, Party officials seem to have recognized that maintaining credibility increasingly requires that scandals and corruption be tackled at the highest levels. In other word, the CCP has shown clear signs that it is moving toward holding officials accountable at all levels of government, rather than merely con- centrating on local officials.

Kenyan scandals: ignoring the obvious In Kenya, in contrast, media exposure of scandal often leads to public and government outcry, followed by a return to business as usual after only a nominal punishment or a judicial dismissal of charges (see Table 2). Kenya presents an interesting case because it has become significantly more democratic over time, though it is still not a consolidated democracy (Branch and Cheeseman, 2009). Conventional wisdom and international development rhetoric suggest that the government should have become more responsive to scandal as it liberalized; but it has not. In fact, unlike in China, even low-level scandals rarely result in punishment. Kenya’s Goldenberg scandal, the most prominent scandal in the past 20 years, follows this pattern perfectly. Goldenberg involved members of the Kenyan government allegedly subsidiz- ing exports of non-existent or smuggled gold for 35% more than it was worth (Karanja, 2003). The proceeds were paid to Goldenberg International, a company co-owned by former President Moi and with close connections to former Vice President Saitoti and other senior politicians and judges. The scandal cost Kenyan taxpayers around US$600 million (BBC News, 2004), the equivalent of 10% of annual GDP, and is considered to be the principal cause of the near-col- lapse of the Kenyan economy in the 1990s (Transparency International Kenya, 2003). When the scandal first broke, in 1993, a number of implicated senior politicians and civil servants resigned their posts – only to resume work quietly some time later. At that time, Kenya was not a democracy (Freedom House score, 5.5), so this pattern of behavior fits with received wisdom about democracy and responsiveness. When free and fair elections in 2002 brought the first party change since independence, the new government began a much-publicized inquiry into the affair. Despite considerable rhetoric, however, action was minimal. Moi has not been formally accused and Saitoti remained a cabinet minister until his death in 2012. Under a law passed before the elections in 2002 giving amnesty to those who return proceeds of corruption, Kamlesh Pattni, a founder of Goldenberg International, reportedly obtained immunity from prosecution by transferring ownership of the Nairobi Grand Regency hotel to the government of Kenya in 2008 (BBC News, 2008). No one has yet been punished. Despite democratization and increasing media freedoms, the pattern of non-response to scandal remained unchanged. Another major scandal broke in May 2004 (Freedom House score, 3.0). The press reported that the government had contracted a deal with Anglo Leasing and Finance for a computerized passport system for approximately €32 million – even after the firm that would supply the system to Anglo Leasing offered to sell the equipment directly to the Kenyan Government for only €6 million (Vasangar, 2005). The scandal returned to front-page news in 2006 after a report written by Kenya’s ‘anti-corruption tsar,’ John Githongo, was leaked. Several cabinet ministers temporarily resigned, and toothless investigations quickly cleared their names. Other ministers, such as David Mwiraria, simply returned quietly to the Cabinet (BBC News Staff, 2006b) and Kiraitu Murungi escaped conviction and was reinstated as Energy Minister after nine months, despite being audiotaped attempting to bribe Githongo into halting the investigation (BBC News, 2006a).

Downloaded from jas.sagepub.com by guest on May 14, 2014 8 Journal of Asian and African Studies

The sale of the same Grand Regency hotel that featured in the Goldenberg Affair was the subject of a separate highly publicized 2008 scandal. After Goldenberg principal Pattni released the hotel to the government, the Central Bank of Kenya sold it to a group of Libyan investors for US$45 million, rather than its recorded value of US$115 million, reportedly in a non-competitive bidding process (BBC News Staff, 2008). The Finance Minister at the time, Amos Kimunya, was accused of corruption and the parliament made a public spectacle by censuring him. His name was quietly cleared several months later following a (secret) official inquiry, and he returned to the Cabinet (Business Daily Staff, 2008). Not all of Kenya’s scandals involve accusations of high-level fraud. In 2009, for example, two prominent human rights activists were assassinated in Nairobi – on the day that the government accused their non-governmental organization (NGO) of being a front for criminal gangs. Many Kenyan civil society NGOs attributed the murders to police acting on government orders and a UN representative who worked with the deceased called for a foreign-led investigation of the murders. The US offered the services of the FBI, but the Kenyan government declined, perhaps because witnesses reported that one of the assailants was wearing a police uniform. Three officers who fired rounds at the University of Nairobi while retrieving the activists’ bodies were arrested, but the government did not arrest or accuse anyone of the murders, despite public outcry (Rice, 2009; Hawaii Free Press, 2009; Amnesty International, 2010). While Kenyan journalists are generally free to publish at their discretion, they occasionally feel the government’s wrath after reporting official involvement in scandals. For example, in September 2009, a media informant was arrested and put on trial for ‘spreading false information’ after news reports cited him as contradicting the government’s official statement on the destination of tanks in a Ukrainian cargo ship hijacked by Somali pirates (Amnesty International, 2009).7

Quantitative results Content analysis of news articles on the two countries paints a similar picture, showing that media reports of scandal in China are strongly correlated with negative repercussions. To arrive at this conclusion, we downloaded 787 English language articles on Chinese and Kenyan scandals from the EBSCO newspaper archive, using pairs of keywords containing the country name (either Kenya or China) and the words ‘corruption,’ ‘scandal,’ or ‘graft’. We choose these keywords because scandals in both countries frequently relate to them. Only articles with available full text were selected, divided roughly equally between China (390) and Kenya (397), all from the period 2000–2010. Of these, 258 articles on China and 248 on Kenya were relevant and on-topic; the remainder were discarded. As discussed above, all articles are from the international press. Next, computer content analysis (CCA) software was used to provide a comparison of the words in the articles with a ‘dictionary’ of pre-defined keywords.8 To determine whether the government responded to scandals, we created categories of words that suggest a judicial response – specifically words related to prison, punishment and the judiciary.9 Analyzing words in these categories allowed us to measure how frequently scandals result in sanctions for those involved. These results are presented in Table 3, along with similar results from a baseline comparison of randomly- selected articles, which ensure a legitimate point of comparison. Two categories of words show statistically significant differences between press coverage of Chinese and Kenyan scandals: words related to prison and to punishment. These results suggest that when the media reveals Kenyan scandals, those involved are less likely to be punished than their Chinese counterparts. Indeed, articles on Chinese scandals are nearly three times as likely as those about Kenya to mention imprisonment and almost twice as likely to mention other punishments, suggesting a substantive difference in outcomes between the two countries. This difference

Downloaded from jas.sagepub.com by guest on May 14, 2014 Hassid and Brass 9

Table 3. Content Analysis Results with Baseline (T-test).

Country Mean Number Mean Number Mean Difference Mean of Words Per of Words Per (China-Kenya, Difference Scandal Article Baseline (non- Scandal Articles) (China-Kenya, scandal) Article Baseline) Judiciary Kenya 1.73 1.45 −0.05 −1.15** China 1.68 0.30 Prison Kenya 0.62 0.60 1.18** −0.06 China 1.80 0.54 Punishment Kenya 0.35 0.21 0.23* −0.10* China 0.58 0.11 Article word count Kenya 978.34 786.61 18.58 −161.24** China 959.76 652.37

**p<.001 *p<.05 is especially pronounced compared to the baseline, non-scandal articles, which discuss the judiciary and punishment more in Kenya than China. Although the absolute difference seems small – around one extra ‘punishment’ word per four articles about Chinese scandals compared to Kenya ones – nearly twice as many newspaper articles mention punishment in China than in Kenya. Content analysis, especially of newspaper articles that average approximately 1000 words, often produces such seemingly small differences (Hassid, 2012). Indeed, because the newspaper corpus reflects dozens of articles on each individual scandal at all stages from discovery to resolution, we should not expect to see wildly divergent reports between the two countries. Here, the results suggest meaningfully higher reports of punishment and imprisonment in Chinese scandals compared to Kenyan ones.10

Why democracy and media freedom fail to improve responsiveness Scandals in China, even those involving prominent national political figures, thus seem to be resolved more quickly than is the case in Kenya. Clearly, government responsiveness in this case cannot be explained by the ‘sunshine’ provided by the media or democratic processes. Why do these mechanisms fail in Kenya and elsewhere? Two features of democracy in less-developed countries help to explain why they are insufficient for government responsiveness. First, the democratic process may itself reduce popular pressure for accountability, because citizens understand that they may eventually remove public officials who are not adequately responsive. In Kenya, for example, the 2007 parliamentary elections saw over 60% of Members of Parliament (MPs) seeking reelection removed from office. Popular opinion at the time, as well as research conducted later, suggested that MPs lost their offices largely for being unresponsive to citizen demands, particularly with regard to mismanagement of Constituency Development Funds (CDF) (Gutierrez-Romero, 2010). These funds, introduced in 2002, disperse 2.5% of government revenue each year through MPs to use for local development projects. In the lead-up to the 2007 elections, reports of CDF misuse were widespread in Kenyan newspapers, but protests regarding this misuse were rare. Indeed, in a 2008 survey of 501 adult Kenyans, we found that only 12% of respondents had ever attended a protest or demonstration. These facts suggest that the Kenyan public is willing to wait for elections to reprimand those involved in scandals rather than

Downloaded from jas.sagepub.com by guest on May 14, 2014 10 Journal of Asian and African Studies demanding immediate relief. However, the violence following the 2007 elections (and the lack thereof in 2013) shows both Kenyan unwillingness to accept rigged elections and their desire for democracy to function (Dercon and Gutierrez-Romero, 2012). In authoritarian states like China, in contrast, citizens may resort to more extreme means of protest if demands are not met quickly. Members of China’s ruling CCP must therefore remain constantly vigilant and responsive to reports of scandal, in the face of a public mood that can quickly turn ugly. Despite spending as much each year on internal policing as on external defense (Tsinghua University, 2010), the number of Chinese protests or ‘mass incidents’ continues to grow, to an astonishing 127,467 in 2008 (China Labor Bulletin, 2009). Many of these ‘incidents’ are mass protests against official corruption, demonstrating the willingness of many Chinese to confront even a coercive state apparatus in a search for clean, responsive government (Chen, 2000). Since the CCP wishes to retain power, it generally responds to protests by punishing those involved in scandal. As Susan Shirk (2011: 17) writes, ‘The elite’s extreme nervousness about potential protests makes them highly responsive when the media report on a problem … Once the media publicize an issue and the issue becomes common knowledge, then the government does not dare ignore it’. The end result is that those exposed often face jail time or worse. In a democracy, in contrast, even leaders of an angry populace rarely have to fear overthrow; instead, regularized succession ensures that popular anger is usually directed toward ‘throwing the bums out,’ rather than remaking the political system anew. Elections can maintain popular anger at scandals at a slow simmer rather than a raging boil. While media freedom and democracy should, in principle, improve governance, other researchers have noted that levels of corruption often tend to be lower in dictatorships than in partial democracies (Montinola and Jackman, 2002). China and Kenya fit this trend. Indeed, Transparency International’s 2010 Corruptions Perception Survey ranks China as the 78th cleanest country in the world, far from good but an enviable position compared to Kenya’s 154th place (Transparency International, 2010).

Possible explanations What then, does explain the difference between these two cases? We speculate below about necessary conditions, drawing on our research in the two countries.

Political will The experiences of Kenya and China demonstrate that political will may be a necessary component for government responsiveness to scandals, particularly combined with adequate state capacity. In less-developed countries, democracy is often unconsolidated, suffering from weak political parties and corruption. Because of this, conditions of ‘partial democracy’ override some of the government responsiveness we expect of mature democracies. In the case of Kenya, parties are anemic, lacking distinguishable ideology or platform. Parties die and are formed with each election cycle, each remaining little more than amalgams of ever-shifting elite politicians (Murunga and Nasong’o, 2006). As a result, corrupt actors from previous and current regimes interweave, making it difficult for voters to respond systematically to scandals of prior administrations (Wolf, 2006). Thus we encounter the ‘politics of collusion’ among the political elite in Kenya, who lack the will to respond to scandal. In China, on the other hand, the Party has the necessary political will. Despite the importance of informal politics and factions inside the CCP, China’s sustained economic growth has been impres- sive enough that Party legitimacy is increasingly tied to meeting ever-stricter economic performance standards, not ideological ones (Downs and Saunders, 1998). With economic growth

Downloaded from jas.sagepub.com by guest on May 14, 2014 Hassid and Brass 11 slowing somewhat, however, the government cannot rely on this alone for legitimacy. Chinese officials thus appear increasingly to feel they must act quickly to assuage public anger when scandals appear, especially those involving corruption or abuse of power. Anything less risks a wholesale challenge to the fundamentally performance-based legitimacy of Communist rule. While in Kenya the unconsolidated democratic process itself gives the incumbent regime a measure of insu- larity from legitimacy crises, Chinese officials are constantly worried that if perceived performance drops too low they may lose their jobs. This system, then, creates huge political will to act quickly on most scandals.

Low expectations Is the presence of political will sufficient to explain the finding that authoritarian China is more responsive to scandals than democratic Kenya? Clearly, enforcement mechanisms matter, but both countries’ judiciaries are so feeble and corrupt – especially when dealing with government officials – that we can effectively discount the independent influence of these bodies (Wedeman 2004; Ringera 2007), though this may be changing in Kenya. We argue that another factor, public expectations, may be the critical missing link needed for government responsiveness. Looking at the relative expectations of the Kenyan and Chinese publics, we find that in China, expectations for state action and responsiveness to scandal are quite high; in Kenya, they are extremely low. This difference is based on, and reinforced by, both countries’ political institutions. In China, heightened expectations exist for at least three reasons. First, the CCP is reluctant to loosen its monopoly on addressing social problems and it retains deep suspicions about even innocuous extra-party organizations. Long wary of civil society, the CCP generally attempts to co- opt social organizations and to suppress those it cannot co-opt. The end result is that far from chal- lenging or monitoring state action, NGOs and other civil society groups often actively curry favor with the powerful (Gallagher, 2004). This ensures that citizens turn directly to state authorities when problems arise. Put simply, the Party/state is the central actor in Chinese public life and remains the natural focus of citizens’ demands. Second, the Chinese central government itself has deliberately raised expectations of its good intentions and effectiveness. At the national level, the government in Beijing takes care to ensure that it deflects anger from central leaders to local ones, maintains social stability and appears to be acting in the best interest of all citizens. The central government tends to involve itself in high profile or particularly serious cases and once it does concessions to an angry public are more likely. At the local level, state officials are most responsive to public outcry over scandal where they are ‘embedded’ into local communities in China. Lily Tsai (2007) found that Chinese officials enmeshed in ‘solidary groups,’ often local temples and heritage associations, are more likely to provide reliable and honest public service provision than those without strong ties to local communities. Ultimately, these local ties provide a strong motivation for many public officials to act honestly; indeed, given that there are no electoral consequences to graft, media investigations, cultural norms and public outcry seem to provide the only incentives for clean local government. Finally, the Chinese public itself demands that the government responds quickly to scandals. For example, the 2005 World Values Survey found that nearly 80% of Chinese rate democracy as ‘absolutely important’ or very important, and 55% believe that one of the ‘essential’ criteria of democracy is that ‘criminals are severely punished’ (compared to only 17% in the United States, of valid answers only). We interpret this to mean that as long as the CCP remains responsive to public outrage over individual scandals and punishes wrongdoers, pressure to reform the political system and institute national electoral democracy may remain low.

Downloaded from jas.sagepub.com by guest on May 14, 2014 12 Journal of Asian and African Studies

In Kenya, in contrast, society appears to be more indulgent of scandal, though there are signs that this is changing. Relative tolerance in Kenya can be attributed to the way that government and citizenship were created and institutionalized, and to learned powerlessness. The combination of these factors has produced a situation where it is arguably rational for individuals to maintain low expectations of government responsiveness to scandal. In one of the most widely-cited essays on African institutions, Peter Ekeh (1975) explains that the manner in which African governments have developed has created a situation in which use of public office for personal or group gain is largely, if begrudgingly, accepted. Colonialists built Africa’s government institutions, transferring them to Africans upon decolonization. Because the institutions were considered foreign instruments of domination, Ekeh (1975: 108) argues, ‘it [became] legitimate to rob the civic public’ – especially if the benefits are (mostly) distributed to one’s tribal and voluntary associations. State-based patronage became the norm in most of Africa, holding deeply divided, multi-ethnic societies together through strategic distribution of resources (Bates, 1981). Corruption, a major precursor to scandal, is often seen as amoral if the corrupt are fleecing the state for the benefit of their ethnic or religious associations. Moreover, citizens rely on their community patron retaining ties to the state. As long as voters expect to benefit from the patronage system, systematic pressure to solve corruption scandals withers. In these relatively new states, moreover, high-level officials are also given discretionary leeway by the nature of the relationship between rulers and ruled. According to Michael Schatzberg (2001), for example, state leaders act – and are treated as – father figures to their citizens. Schatzberg details how power in Africa is a fine balance between an acceptable appetite on the part of the ‘father’ and political corruption. Individual Africans often act more as subject populations than as citizens will full rights and responsibilities (Mamdani, 1996; Schatzberg, 2001). In such situations, individuals may not feel empowered to act, even when scandal is brazen. Some non-state elites recognize this phenomenon, though it baffles them. In an op-ed in Kenya’s leading newspaper, for example, Willy Rotich (2010) writes about Kenyan ‘indifference to blatant corruption, at times an inexplicable attraction to people indicted in graft, impunity or whatever evil’. Rotich describes the efficient, modest-living public servant as the object of ridicule and high- level crooks and bandits as objects of reverence. This is not to say that Kenyans are satisfied with low state responsiveness to scandal. Indeed, politicians, the mass media, civil society organizations and public opinion polls provide a steady drumbeat of complaint against the status quo (Mutua, 2011). But while demands for change are high, expectations for change are low. Scandals produce outcry, but outcry that is tempered by (exasperated) acceptance. Countering Rotich’s cynicism, it may be that many Kenyans have, in essence, ‘exited’ the system (Hirschman, 1970). Rationally, if perhaps not consciously, these citizens have recognized that their government will not punish official malefactors. Getting involved in attempts to change the status quo often leads to anger and disillusionment, as anti-corruption NGOs in Nairobi experience.11 As one scholar argues, ‘Repeated breaches of trust can generate an attitude of deep mistrust among some sectors of the population; individuals may be inclined to distrust, not just particular leaders … but politicians per se’ (Thompson, 2000: 255). In interviews Brass conducted with NGO leaders and civil servants between 2005 and 2008, many respondents mentioned that the Kenyan people don’t expect much from government and are even empathetic to the government’s inability to provide [citation redacted]. The end result is a ‘culture of impunity,’ as identified by former US Ambassador Michael Ranneberger (Hepple and Smith, 2010). In interviews and surveys Brass conducted, for example, she found that popular support for both Kenyan political leaders and the civil service is quite high – indeed surprisingly so.12 In a 2008 survey of 500 secondary school students replicating work from the immediate post-independence era (a time of great optimism), for example, we find that – contrary to popular scholarly opinion

Downloaded from jas.sagepub.com by guest on May 14, 2014 Hassid and Brass 13

– legitimacy levels in Kenya have not fallen. In addition, in a second survey of adults in 2008, Brass found that 65% of respondents agreed or strongly agreed that, ‘the Government of Kenya shares my values and does the right thing’. Moreover 74% of respondents reported ‘a great deal’ or ‘quite a lot’ of confidence in the Government of Kenya generally. We find that even in middle class neighborhoods of Nairobi, where access to news is universal, few challenge the legitimacy of the central state. When asked about confidence in the President and the Government generally, 71% and 60% respectively reported high confidence ratings. This phenomenon has been dubbed ‘expectation-disconfirmation theory’, where citizens rate government performance higher when they have very low expectations than when they expect more (Van Ryzin, 2004). These survey data suggest that most of Kenya’s officials need not be overly concerned with state legitimacy, and they have few worries that even festering scandals will result in a loss of confidence in the government. CCP officials, in contrast, remain concerned about their perceived legitimacy, a worry sharp- ened by events like the awarding of the 2010 Nobel Peace Prize to imprisoned Chinese activist Liu Xiaobo. Many analysts regarded the Chinese response – weeks of denunciations followed by the creation of a ‘competitor’ Confucius Prize – as an extreme reaction conditioned by fears that the regime lacked political legitimacy. When a government feels secure, as it does in Nairobi, it has a less urgent need to respond to every social scandal or uncovered crime. Insecure governments, in contrast, must fear that every unresolved scandal can conceal the germ of revolution. Individual Kenyans, then, are rational to have low expectations for government responsiveness to scandal. After decades of living with the pattern ‘media exposure of scandal; government outcry; nominal (often temporary) punishment; return to business as usual’, most Kenyans have become resigned to the status quo. In a situation where not a single high-level official has been convicted for a scandalous crime – in a country where one scandal cost 10% of GDP – disillusioned ‘exit’ from participation makes sense. Unfortunately, a vicious cycle of expectation and response can develop, where the lack of government responsiveness to scandal lowers public expectations of a response.

Conclusions In China, a one-party dictatorship with a muzzled press, authorities react quickly to punish officials involved in scandal, satisfying public expectations by rapidly quelling public outrage. In Kenya, in contrast, politicians act with impunity and even egregious scandals fester unsolved for years, despite democratic elections and a free press. This suggests that our perceived wisdom about the role of democracy and free media in government responsiveness to scandal may not be entirely correct. Some authoritarian regimes are more responsive than their democratic counterparts. We hypothesize that crucial factors explaining this include political will and the expectation among ordinary citizens that the state authorities will act. Neither the Kenyan nor Chinese public is happy about scandals, and both publics in surveys and interviews demand that these problems are resolved quickly. While ordinary Kenyans demand responsive government, however, they certainly do not expect it. The weak consolidation of democracy, political structures dominated by ethnic identity and strong patron-client relationships, and decades of official negligence have all gradually reduced public expectations for clean government. Kenya is hardly unique. Major Western democracies like Italy maintain a similar level of unac- countable corruption and unresolved scandals for decades, despite pluralistic competition and a free media. As Colazingari and Rose-Ackerman (1998: 469) write, ‘A shift from authoritarian to democratic rule does not necessarily reduce payoffs’. In other words, democratization may not be the right approach to increasing responsiveness to scandal. Italy’s example also provides some

Downloaded from jas.sagepub.com by guest on May 14, 2014 14 Journal of Asian and African Studies positive news, however, suggesting that when public expectations change, even entrenched corruption can be (mostly) swept away. Manion (1996: 191) explains, ‘major changes in structures may be ineffective in eliminating corruption or very significantly reducing its incidence’, but in contrast, ‘altering expectations may be effective’ provided that the changes in expectations are large enough. Sunshine, in short, may not always be the best disinfectant. The implications of our finding are several. Most importantly, understanding the role of citizens’ expectations and state capacity not only allows us to deepen our comprehension of the conditions under which governments respond to the public pressure but also expands our understanding of regime types. Our evidence suggests that political and societal features social scientists normally consider as characteristics of democracies can in fact sometimes be observed in autocratic regimes. We therefore draw attention to the need to refine state theory and to unpack the usual simple regime type continuum – to consider a multitude of important variables in our discussions of ‘democracy’ and ‘dictatorship’. As such, this work fits broadly into the efforts by Collier and Levitski (1997), Sartori (1970), and others to encourage scholars to carefully define and parse such broad terms. But there is a crucial difference: we are not arguing about how to characterize countries like Kenya and China, but rather are suggesting that government responsiveness to mass pressure should not be associated exclusively with democracy. Under some conditions, autocracies can be more responsive to their citizens than democracies. A democratic state, in other words, is not necessarily responsive to popular pressure day-to-day. Nor is an authoritarian regime necessarily deaf to mass demands. In short, we aim to decouple broad terms like ‘democracy’ from the assump- tions that many scholars have implicitly held about regime performance. Finally, our findings have implications for policymakers, particularly those aiming their efforts at less-developed countries. In the short run, efforts to reduce scandal rates – particularly corruption scandal – might best be targeted at enforcement mechanisms and incentives, rather than at wholesale change of the electoral structure. While democracy-promotion should be lauded, programs cannot stop at the level of partial consolidation. Indeed, poorly consolidated democracies tend to be more corrupt on average than even the most hardened authoritarian regimes. In the end, programs to manage citizens’ expectations and to encourage them to demand government accountability might prove more effective than simply promoting democracy in a one-size-fits-all approach.

Acknowledgements Thanks to anonymous reviewers, Bart Watson, and those who gave comments at the 2011 annual meetings of the Association for Asian Studies and the American Political Science Association.

Funding The order of names was determined by coin toss. Support for the research reported in this article came from a Fulbright-Hays Doctoral Dissertation Research Award Grant, a National Science Foundation Graduate Research Fellowship, the University of Technology, Sydney and Indiana University.

Notes 1. Emphasis in original. 2. While Kenyan elections have been associated with violence, particularly following the 2007 election, the country also holds parliamentary, county and municipal elections at regular intervals without problems. Additionally, the country has had two recent democratic changes in the party in power (2002 and 2013), suggesting democratic consolidation. 3. Brass conducted this survey in Machakos and Mbeere districts in working-class neighborhoods of the capital, Nairobi. While the responses are biased toward ethnicities living in Eastern Province, responses did not vary significantly by ethnicity. For example, only 4% of Luo in the sample did not think Kenya

Downloaded from jas.sagepub.com by guest on May 14, 2014 Hassid and Brass 15

was a democracy in 2008, and 39% thought it was a democracy with ‘major problems’ – a smaller per- centage than the sample as a whole. 4. China has extensive media censorship, of course, but the official pronouncement is that it does not exist. See http://hk.aboluowang.com/2013/0105/276760.html. 5. When Attorney General Wako dismissed the first charges, in 2005, the media excoriated the decision. When Cholmondeley was accused a year later of a second murder, media reports kept the scandal in the spotlight. 6. Author interview with senior Chinese journalist, Beijing, 2011. 7. The Kenyan government claimed that it ordered the tanks for its own use, but others – including an official in South Sudan – say that Kenya planned to re-export them to that country. Kenya has featured prominently in the peace accords between the Sudanese governments in Khartoum and in Juba. 8. The software, Yoshikoder, is available for free online, and has been used in similar ways by other scholars. 9. The prison category counted the following words and stems: ‘hard labor,’ ‘imprison*,’ ‘incarcer- ate*,’ ‘jail*,’ ‘prison*,’ ‘re-educat*,’ ‘reeducat*,’ and ‘sentenced’. Punishment measured: ‘demoted,’ ‘execute,’ ‘executed,’ ‘fined,’ ‘fired,’ ‘punish*,’ and ‘stripped’. Judiciary measured: ‘appeal,’ ‘appel- late,’ ‘attorney(s),’ ‘barrister(s),’ ‘charged,’ ‘court(s),’ ‘defendant(s),’ ‘indict*,’ ‘judge(s),’ ‘lawyer(s),’ ‘magistrate(s),’ ‘plaintiff(s),’ ‘procurator*,’ ‘prosecut*,’ ‘solicitor(s),’ ‘trial(s),’ and ‘tribunal(s)’. 10. Expanding on these results, we replicated this method across all countries that have a 2010 Freedom House ranking (n = 162), using an original data set of 17,160 articles compiled in the same way described above. Confirming the results we present for China and Kenya, no combination of Freedom House’s Civil Liberties or Political Rights measures (or similar measures from the World Bank and Reporters Without Borders) was statistically significant in combinations controlling for GDP per capita, population and other variables in OLS regression using the same dependent variables. In other words, we find that variation in international state response to scandals cannot be adequately explained by traditional measures. 11. Some of this anger and frustration is documented in a leaked diplomatic communication (US Embassy Nairobi, 2009). 12. Other survey researchers have found similar, surprisingly optimistic results (Wolf et al., 2004), crediting them to post-electoral ‘euphoria’ after the democratic transition in 2002.

References Amnesty International (2009) Kenya – Amnesty International Report 2009. London: Amnesty International. Amnesty International (2010) Kenya – Amnesty International Report 2010. London: Amnesty International. Ansfield J and Buckley C (2013) China Focusing Graft Enquiry on Ex-Official. The New York Times, 15 December, p. A1. Bandurski D (2008) Press Controls Feed China’s Food Problem. The Wall Street Journal Asia, 7 October, p.16. Bates RH (1981) Markets and States in Tropical Africa: The Political Basis of Agricultural Policies. Berkeley, CA: University of California Press. BBC News Staff (2004) Moi ‘Ordered’ Goldenberg Payment. BBC News, 17 February. BBC News Staff (2006a) ‘Taped Evidence’ in Kenya Scandal. BBC News, 9 February. BBC News Staff (2006b) Travel Ban in Kenya Scam Inquiry. BBC News, 14 February BBC News Staff (2008) Probe into Kenya Hotel Sell-Off. BBC News, 30 June. Bennett A (2004) Case Study Methods: Design, Use, and Comparative Advantages. In: Sprinz DF and Wolinsky-Nahmias Y (eds) Models, Numbers, and Cases: Methods for Studying International Relations. Ann Arbor, MI: University of Michigan Press. Besley T and Burgess R (2001) Political Agency, Government Responsiveness and the Role of the Media. European Economic Review 45(4–6): 629–640. Besley T and Burgess R (2002) The Political Economy of Government Responsiveness: Theory and Evidence from India. The Quarterly Journal of Economics 117(4): 1415–1451. Brady A-M (2008) Marketing Dictatorship: Propaganda and Thought Work in Contemporary China. Lanham: Rowman & Littlefield.

Downloaded from jas.sagepub.com by guest on May 14, 2014 16 Journal of Asian and African Studies

Branch D and Cheeseman N (2009) Democratization, Sequencing, and State Failure in Africa: Lessons from Kenya. African Affairs 108: 1–26. Brunetti A and Weder B (2003) A Free Press Is Bad News for Corruption. Journal of Public Economics 87(7–8): 1801–1824. Business Daily Staff (2008) MPs Censure Kimunya over Sale of Grand Regency Hotel. Business Daily (Kenya), 2 July. Captial Reporter (2010) Kibaki Vows Kenya Media Freedom. Capital News, 18 March. Chen F (2000) Subsistence Crises, Managerial Corruption and Labour Protests in China. China Journal, 4–63. China Labor Bulletin (2009) Going It Alone: The Workers’ Movement in China 2007–2008. Research Reports. Hong Kong: China Labor Bulletin, pp.1–57. Colazingari S and Rose-Ackerman S (1998) Corruption in a Paternalistic Democracy: Lessons from Italy for Latin America. Political Science Quarterly 113(3): 447–470. Collier D and Levitsky S (1997) Democracy with Adjectives: Conceptual Innovation in Comparative Research. World Politics 49(3): 430–451. Daily Nation. (2006) The Nation (Kenya) - AAGM: War On Graft Is, Indeed, Lost. The Daily Nation. Nairobi, Kenya. Available at: http://www.nation.co.ke/oped/-/1192/136750/-/r3mxt2z/-/index.html (accessed 15 April 2014). Dercon S and Gutierrez-Romero R (2012) Triggers and Characteristics of the 2007 Kenyan Electoral Violence. World Development 40(4): 731–744. DFID (2008) Briefing Note: Media and Good Governance. London: BBC World Service Trust and DFID Politics and the State Team. Dimitrov MK (2009) Piracy and the State: The Politics of Intellectual Property Rights in China, Cambridge and New York: Cambridge University Press. Downs ES and Saunders PC (1998) Legitimacy and the Limits of Nationalism: China and the Diaoyu Islands. International Security 23(3): 114–146. Egorov G, Guriev S and Sonin K (2009) Why Resource-poor Dictators Allow Freer Media: A Theory and Evidence from Panel Data. American Political Science Review 103(4): 645–668. Ekeh PP (1975) Colonialism and the Two Publics in Africa: A Theoretical Statement. Comparative Studies in Society and History 17(1): 91–112. Freedom House (2010) Combined Average Ratings – Independent Countries. Washington DC: Freedom House. Freedom House (2011) Freedom of the Press. Washington DC: Freedom House. Gallagher ME (2004) China: The Limits of Civil Society in a Late Leninist State. In: Alagappa M (ed.) Civil Society and Political Change in Asia: Expanding and Contracting Democratic Space. Stanford, CA: Stanford University Press, pp. 419–452. George AL and Bennett A (2005) Case Studies and Theory Development in the Social Sciences. Cambridge, MA: MIT Press. Gerring J (2001) Social Science Methodology: A Criterial Framework. Cambridge: Cambridge University Press. Government of Kenya (2007) Vision 2030: The Popular Version. Nairobi: Government Press. Gutierrez-Romero R (2010) Decentralization, Accountability and the 2007 MPs Elections in Kenya. Center for the Study of African Economies, University of Oxford, Working Paper Series. Hand KJ (2006) Using Law for a Righteous Purpose: The Sun Zhigang Incident and Evolving Forms of Citizen Action in the People’s Republic of China. Columbia Journal of Transnational Law 45(1): 114– 195. Hassid J (2011) Four Models of the Fourth Estate: A Typology of Contemporary Chinese Journalists. China Quarterly 208: 813–832. Hassid J (2012) Safety Valve or Pressure Cooker? Blogs in Chinese Political Life. Journal of Communication 62(2): 212–230. Hawaii Free Press Staff (2009) Wikileaks Writers Killed in Kenya. Hawaii Free Press, 9 March. Available at: http://www.hawaiifreepress.com/ArticlesMain/tabid/56/ID/502/Wikileaks-writers-killed-in-Kenya. aspx (accessed 15 April 2014).

Downloaded from jas.sagepub.com by guest on May 14, 2014 Hassid and Brass 17

Hepple G and Smith D (2010) Wikileaks Cables: Rampant Corruption ‘Could Push Kenya Back into Violence’. The Guardian (UK), 8 December. Available at: http://www.theguardian.com/world/2010/ dec/08/wikileaks-cables-kenya-violence-china (accessed 15 April 2014). Hirschman AO (1970) Exit, Voice, and Loyalty: Responses to Decline in Firms, Organizations, and States. Cambridge, MA: Harvard University Press. Kaiman J (2013) Liu Zhijun, China’s Ex-railway Minister, Sentenced to Death for Corruption, The Guardian (UK), 8 July. Available at: http://www.theguardian.com/world/2013/jul/08/liu-zhijun-sentenced-death- corruption (accessed 15 April 2014). Karanja W (2003) Kenya: Corruption Scandal. World Press Review 50: 10. Kuklinski JH and Stanga JE (1979) Political Participation and Government Responsiveness: The Behavior of California Superior Courts. American Political Science Review 73(4): 1090–1099. Liebman BL (2005) Watchdog or Demagogue? The Media in the Chinese Legal System. Columbia Law Review 105(1): 1–157. Lines K (2009) Governance and the Media: A Survey of Public Opinion. London: BBC World Service Trust. Mamdani M (1996) Citizen and Subject: Contemporary Africa and the Legacy of Late Colonialism. Princeton, NJ: Princeton University Press. Manion M (1996) Corruption by Design: Bribery in Chinese Enterprise Licensing. Journal of Law, Economics, and Organization 12(1): 167–195. Markovits AS and Silverstein M (1988) Introduction: Power and Process in Liberal Democracies. In: Markovits AS and Silverstein M (eds) The Politics of Scandal: Power and Process in Liberal Democracies. New York, NY: Holmes and Meier. Montinola GR and Jackman RW (2002) Sources of Corruption: A Cross-Country Study. British Journal of Political Science 32(1): 147–170. Murunga GR and Nasong’o SW (2006) Bent on Self-destruction: The Kibaki Regime in Kenya, Journal of Contemporary African Studies 24(1): 1–28. Mutua M (2011) Political Class Testing Kenyans’ Patience. Daily Nation (Kenya), 22 January. Available at: http:// www.nation.co.ke/oped/Opinion/Political+class+testing+Kenyans+patience+/-/440808/1094010/-/ 10yyaylz/-/index.html (accessed 15 April 2014). Odugbemi S and Norris P (2010) Assessing the Extent to Which the News Media Act as Watchdogs, Agenda Setters, and Gatekeepers. In: Norris P (ed.) Public Sentinel: News Media and Governance Reform. Washington, DC: World Bank Publications: 379–394. Osnos E (2012) Boss Rail. The New Yorker, 22 October, 44–53. People’s Abstracts (人民文摘) (2004) Gaining Perspective on the ‘Seven Hour Time Zone’ Movie Incident (透视电影《时差七小时》事件). People’s Abstracts (人民文摘) 12. Reporters Without Borders (2010) Worldwide Press Freedom Index [Online]. Available at: http://en.rsf.org/ press-freedom-index-2010,1034.html (accessed 15 April 2014). Rice X (2009) Two Kenyan Rights Activists Shot Dead. The Guardian (UK), 6 March. Available at: http:// www.theguardian.com/world/2009/mar/06/kenyan-activists-shot-dead (accessed 15 April 2014). Ringera AG (2007) Corruption in the Judiciary. Presentation of the Honourable Mr Justice (Rtd) Aaron G Ringera, Director and Chief Executive of the Kenya Anti-Corruption Commission. Washington, DC: The World Bank. Rose-Ackerman S (1999) Corruption and Government: Causes, Consequences, and Reform. London and New York: Cambridge University Press. Rotich WK (2010) Kenyans Suffer from an Ambivalence That May Soon Stymie the New Order. Saturday Nation (Kenya), 13 December. Available at: http://www.nation.co.ke/oped/Opinion/-/440808/1071820/-/ nu9tvhz/-/index.html (accessed 15 April 2014). Sartori G (1970) Concept Misinformation in Comparative Politics. American Political Science Review 64(4): 1033–1053. Schatzberg MG (2001) Political Legitimacy in Middle Africa: Father, Family, Food, Bloomington, IN: Indiana University Press. Shirk SL (2011) Changing Media, Changing China. In: Shirk SL (ed.) Changing Media, Changing China. New York, NY: Oxford University Press: 1–37.

Downloaded from jas.sagepub.com by guest on May 14, 2014 18 Journal of Asian and African Studies

Stapenhurst R (2000) The Media’s Role in Curbing Corruption. Washington, DC: World Bank Publications. The Star Editorial Board (2012) Kenya: Kibaki Outstanding on Press Freedom. The Star, Kenya, 13 December. Available at: http://allafrica.com/stories/201212170484.html (accessed 15 April 2014). Thompson JB (2000) Political Scandal: Power and Visibility in the Media Age. Cambridge: Polity Press. Transparency International Kenya (2003) Fools Gold: The Abuse of Export Compensation Schemes in Kenya and Argentina. Kenya, Nairobi: Transparency International. Transparency International (2010) Corruption Perceptions Index. Available at: http://www.transparency.org/ policy_research/surveys_indices/cpi/2010/results (accessed 15 April 2014). Tsai LL (2007) Accountability without Democracy : Solidary Groups and Public Goods Provision in Rural China. New York, NY: Cambridge University Press. Tsinghua University (2010) A New Way of Thinking About Maintaining Stability (维稳新思流). Southern Weekend (南方周末), 14 April. US Embassy Nairobi (2009) Visas Donkey: Corruption 212(F) Visa Denial. Secret cable to US Secretary of State released via Wikileaks. Van Ryzin GG (2004) Expectations, Performance, and Citizen Satisfaction with Urban Services. Journal of Policy Analysis and Management 23(3): 433–448. Vasagar J (2005) Kenyan President Faces Rebellion on Sleaze. The Guardian (UK), 24 February. Available at: http://www.theguardian.com/world/2005/feb/24/kenya.jeevanvasagar (accessed 15 April 2014). Wedeman A (2004) The Intensification of Corruption in China. China Quarterly 180: 895–921. World Values Survey (2008) Official Aggregate 1981–2005, Madrid: World Values Survey Association, Aggregate File Producer ASEP/JDS. Wolf TP (2006) Immunity or Accountability? Daniel Toroitich arap Moi: Kenya’s First Retired President. In: Southall R and Melber HL (eds) Legacies of Power. Cape Town: HSRC Press. Wolf TP, Logan C and Owiti J (2004) A New Dawn? Popular Optimism in Kenya after the Transition. Afrobarometer Working Paper Afrobarometer: Cape Town, South Africa. Working paper no. 33. Xinhua News Agency (2008) Crisis Management Helps China’s Dairy Industry Recover. XinhuaNet English, 25 September. Available at: http://news.xinhuanet.com/english/2008-09/25/content_10112354.htm (accessed 15 April 2014). Zhu Zhe (2007) More Than 460 Rescued from Brick Kiln Slavery. China Daily. 15 June. Available at: http:// www.chinadaily.com.cn/china/2007-06/15/content_894802.htm (accessed 1 February 2011). Zhu Zhe and Cui Xiaohuo (2009) Two Get Death in Tainted Milk Case. China Daily. 23 January. Available at: http://www.chinadaily.com.cn/china/2009-01/23/content_7422983.htm (accessed 15 April 2014).

Author biographies Jonathan Hassid is a Postdoctoral Research Fellow at the University of Technology, Sydney’s China Research Centre. He has published on Chinese politics and the Chinese news media in Comparative Political Studies, The China Quarterly, the Journal of Communication, Asian Studies and elsewhere. He received his PhD in political science from the University of California, Berkeley. Jennifer N Brass is Assistant Professor at the School of Public & Environmental Affairs at Indiana University. Her research focuses on service provision, governance and state development in sub-Saharan Africa. She holds a PhD and MA from the University of California, Berkeley and BS from Georgetown University. Her articles appear in World Development, Governance, Development and Change, and elsewhere.

Downloaded from jas.sagepub.com by guest on May 14, 2014