Performance Measurement in Multi-Period Agencies

Abstract

this paper, we examine a multiperiod moral hazard problem. In contrast to most of the earlier literature on multiperiod agency relationships, we allow for the possibility that the agent's unobservable effort in any given period affects not only the cash flow for the current period, but also future cash flows. For contracting purposes, the firm can rely on realized cash flows (current and past) and on expected cash flows to be received in future periods. We consider alternative arrangements for combining these informational variables so as to create optimal incentive provisions. We find that the information reported by a standard accrual accounting system is sufficient for creating optimal incentives. In our model, the accrual accounting system capitalizes future receivables in an asset base. The system also measures income by combining current cash flows and changes in the asset base (the accruals). Residual income is usually defined as income minus an interest charge on the assets held. We demonstrate that residual income identifies precisely the information the principal desires for performance evaluation purposes. In particular, it is optimal for the principal to reward the agent in each period on the basis of current residual income. In addition to the familiar moral hazard problem, we allow for "financing activities" by which the manager can manipulate the intertemporal distribution of cash flows. Given any stream of cash flows resulting from his effort choices, the manager can induce a variation of the basic cash flows provided the overall present value is unchanged. Our interpretation here is that an outside observer cannot distinguish whether cash displacements were part of the firm's ordinary transactions or whether they were induced deliberately by the manag...

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 — 11 citations exist upstream.

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