This thesis studies a 2002 attempt to fundamentally reform the Pakistan Police Service (PSP) through the promulgation of Police Order 2002. This reform was aimed at converting the PSP from an instrument of coercive government to a force responsive to and protective of the citizenry. It constituted, as its introduction explicitly stated, a significant move toward democratic consolidation in Pakistan. PO 2002 was introduced nationwide on 14 August 2002 with wide support from Pakistani elites and under the direction of the autocratic but reform-oriented government of General Pervez Musharraf. Yet PO 2002 failed. It was amended in 2004 and annulled in 2010. Understanding why it failed is important for understanding the challenges confronting democratising developing states, like Pakistan. Existing accounts attribute PO 2002 failure either to ‘loss of political will’ or to ‘bureaucratic politics’. The present research inclined toward the latter explanation until evidence gathered in the field pointed in quite another direction. The thesis employed a combination of process-tracing of the history of PO 2002 and the analytical framework of advocacy coalition developed by Sabatier and Jenkins-Smith. The use of Advocacy Coalition framework (ACF) directed research toward the most important factors before the actors involved. Semi-structured interviews with these revealed their beliefs and various aims regarding the initiation, progress and fate of PO 2002. ACF also alerted the research to the impact on the policy domain of external perturbations, internal shocks and changes in socio-political conditions. Interviews with key personnel of the Musharraf regime and relevant political and civil parties revealed considerable unanimity of opinion: the demand for police reform originated in the 1990s when simultaneous strategies of democratic transition and neoliberal transformation conflicted, aggravating chaos and distributional conflicts in multi-ethnic urban centres of Pakistan. Governments used police in an attempt to control the situation but brutal, highly politicised policing failed, prompting widespread acceptance of the need for police reform. ACF analysis found that the policy design phase of the resulting PO 2002 was monopolised by a material coalition of PSP that identified the problem as control of police by central executive power. It recommended transfer of control to local communities. Absent an epistemic community capable of analysing this proposal’s merits, and in an atmosphere of general public distrust of elected politicians, it was approved by a military government pursuing community empowerment and by liberal elites who saw it as the pathway to democratic policing. But research revealed a contradiction in PO 2002’s stated objectives – first, to improve police performance and, second, to make police autonomous. Increased autonomy worked against performance by serving the motives of PSP officers more interested in removing bottlenecks from their careers than in effective policing. Moreover, the transfer of control to divided, often mutually hostile communities, in the times of market liberalisation merely fragmented politicisation and led to intercommunal violence. Exacerbating the situation was the fact that, even before PO 2002 was implemented, the aftermath of 9/11 caused unprecedented anti-American sentiment in Pakistan. Given popular perceptions of the US as anti-Islam, President Musharraf’s assistance to George W. Bush in his ‘war on terror’ cost him public support. Meanwhile, thanks to PO 2002, his government could not employ police to fight terrorists hiding in local communities. Belated recognition of the need to align police with the policies of the central executive led to the 2004 amendment of PO 2002. Control of provincial heads of police was acquired by the state via the intermediary home secretary, while control of ancillary police agencies was left with local actors and communities. But problems once again re-emerged when the Pakistan People’s Party (PPP) came to power after Musharraf’s resignation in 2008. Strong evidence against the lower ranks of police force victimising the accused of blasphemy came forward, creating a scandalous national crisis. PPP government and state institutions were blamed for neglecting their responsibility and not protecting the rights of the accused. Further inquiry identified both the internal and external structures of PO 2002 as problem parameters. The additional internal structures introduced by PO 2002 had introduced procedural delays which significantly increased the risks for the blasphemy accused. These procedural delays further increased the probability of exploitation of police force by the local political actors, business groups and the extremist factions in some communities. Lower level police constables inspired by the extremist ideology or overtaken by their own sentiments even killed the accused of blasphemy, especially in Punjab. One, such incident also led to the murder of the governor of Punjab in 2010. The strong evidence before the PPP government left little choice but to repeal PO 2002 in 2011, and revert back to PO 1861, but it was not without incremental changes. PO 2002 failed because of its own internal weaknesses. Removing central executive control created serious security issues and singular focus of the PSP material coalitions on removing career bottle neck in their careers introduced procedural delays between reporting of the crime and initiation of its investigation. This delay increased the risk of victimisation of already vulnerable blasphemy accused. Devolving policing responsibility to local communities may seem democratic, but in a multi-ethnic country it is a recipe for conflict. The study concludes by exploring models aimed at reconciling control of police with democratic imperatives but argues that the lack of political trust in the executive, which began with Pakistan’s transformation to a neoliberal regime, remains the biggest challenge for democratic consolidation in Pakistan and perhaps in other developing countries.