Cost overruns on multi-period projects are large and frequent in natural environments. Reasons for these overruns include cost understatements in initial project proposals and escalation of commitment to projects when initial actual costs turn out to be higher than expected. Prior literature has suggested, and some firms have implemented, a device that limits escalation and thus potentially reduces cost overruns: changing decision makers (superiors) so that the manager who approves continuation of a project is not the same individual who approved the project initially. We provide theoretical explanations and experimental evidence about how changing versus continuing superiors affect the underestimates of cost in initial project proposals. We find that although changing superiors, as expected, are more likely than are continuing superiors to react skeptically to continuation proposals when first-period cost overruns have occurred, this does not reduce initial cost understatements and overruns. On the contrary, in our setting it leads to greater initial understatements and overruns. Subordinates anticipate that new superiors will be more critical of their projects; hence, they discount later-period payoffs and focus on gaining initial funding by providing understated cost estimates.

Organizations have often been criticized for reliance on a single item of accounting information (e.g., profit) in evaluating performance, because of its incompleteness. We provide theory‐based experimental evidence that under frequently occurring performance‐evaluation conditions (subjective performance evaluation, causal ambiguity, and individual differences in cognitive ability, knowledge, and/or motivation that lead to different interpretations of information), reliance on a single item of accounting information (profit), rather than profit plus additional (e.g., nonfinancial or external) information, can provide two potential benefits which offset the costs of information incompleteness. First, subordinates are more likely to make the management decisions that superiors will evaluate and reward highly—that is, there are fewer coordination failures in management decisions. Second, even after controlling for the presence or absence of coordination failures, subordinates experience less negative surprise about their performance evaluations.