Tension in New Zealand's Public Hospitals: Performance Effects of Sharpening Incentives
New Zealand's health sector reforms in the mid-1990s introduced corporate institutions and market disciplines to public hospitals. Yet the reorganisation of New Zealand's public hospitals into Crown Health Enterprises (CHEs) led to severe criticisms. Ultimately the CHEs were replaced with non-profit Hospital and Health Services. This thesis focuses on three major criticisms of the CHEs. We use game theory to provide a formal and novel analysis of interactions that could cause an organisation's performance to differ markedly from the reformers' expectations. The analysis explains how a stylised set of reforms could fail to achieve their objectives. Chapter 2 analyses public hospital throughput data over the reform period. We find that the CHE reforms were independently associated with an increase in hospitals' treatment costs. This chapter motivates the theoretical analyses of the three criticisms of the CHEs. We structure the theoretic analysis using an organisational hierarchy with four actors: a funder, an (hospital) administrator, a (medical) specialist and a (health) consumer. The first criticism was that CHE Boards paid bonuses despite managers failing to achieve performance targets. Chapter 3 examines when a funder may want to revise the budget of an organisation and to pay the administrator a bonus despite failing to meet a target. We introduce three features of the CHE reforms that conventional soft budget constraint models partly or entirely neglect: funder bargaining power, revisable targets and performance bonuses. A flexible budget constraint paired with bonuses can be efficient in the light of uncertainty. The second criticism was that costs escalated despite strong managerial incentives for cost control. Chapter 4 argues that such incentives could disrupt trust in an organisation. We show that sharpening the administrator's incentives for cost control can create a misalignment between the administrator and the specialist and cause costs to escalate. Our result, that incentivising a measurable dimension of performance can worsen performance of that same task, contrasts with the conventional game-theoretic literature. The third criticism was that the reforms let doctors manipulate managers, resulting in inefficiency. The first model of Chapter 5 shows that an administrator might want to encourage a specialist to influence public opinion. We modify the first model to reflect a feature of the reforms: managerial efforts aimed at improving the organisation's operation. The administrator can damage a whistle-blower's credibility, to the detriment of specialists and patients. Both models give original insights into how the reforms could let an administrator take advantage of his role. In this multi-layered model, the administrator may intentionally reduce communication. The CHE reformers expected performance incentives to flow through a corporate structure to improve efficiency. Rather than a cascade of beneficial incentives, incomplete contracts could cause unintentional negative interactions. Tension and perverse incentives could have caused costs to rise, necessitating budget revisions and additional bonus payments, while permitting administrators to silence whistle-blowers. This research shows how complex organisations that rely on soft information can benefit from systems that enhance trust and collaboration, and may be harmed by unhealthy tension.