Government Oversight of Organizations Engaged in Multiple Activities: Does Centralized Governance Encourage Quantity or Quality?

Abstract

In situations where organizations perform multiple tasks, is it better for government oversight to be centralized or decentralized? In the private sector, strong incentives are used to induce effort: sometimes in the form of bonuses, stock options, piece-rate schemes, and commission sales. Strong incentives, however, are frequently unable to achieve efficient outcomes. If one is operating in an environment whereby multiple tasks are performed but are not all easily observed, a strong incentive scheme will fail to produce an efficient outcome. Under this type of scenario, previous research has suggested that slack incentives should be used. Although much of the literature on multitasking has focused on private sector incentives, similar problems exist in the public sector. In the public sector, different tools are available to address this problem. A public sector organization generally has limited resources and is often not able to use performance based rewards. In the public sector, it has become popular to develop a set of performance measures to assess a governmental unit’s productivity. We analyze how centralization of oversight and funding affect research activities within a public institution for a set of public universities, focusing on quantity and quality measures. Our results suggest that quality is higher at institutions within a centralized framework. Moreover, these results hold only for those institutions that are not considered to be the state's flagship university. To explain why centralization may be better in this instance, we explored several alternative hypotheses. We find the evidence supports the hypothesis that centralization improves performance insofar as it encourages potential externalities to be internalized when several public organizations are forced to compete and cooperate with each other.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 4 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.