Monitoring for worker quality: a theory of task assignment, job ladders, wages and mobility in internal labor markets

拟议研究的广泛目标是了解信息的性质和获取信息的技术如何影响工资，晋升途径和企业内的任务分配（内部劳动力市场）。狭义的目标是解释非管理人员内部劳动力市场结构的变化。与大多数文献相反，PI并不关注旨在激励努力工作或阻止道德风险的政策，这可能是一个直观上没有吸引力的起点，并可能导致与数据不一致的结论。相反，这个项目的重点是监测，以评估工人。基线模型是一个有两种类型的工人，好的和坏的，两种类型的任务，高和低。在这种模式下，大多数工作都是例行公事（至少对受过适当培训的工人来说是这样），但工人偶尔会遇到“危机”，好工人可以解决，但坏工人不能。失败很容易被观察到，但成功只有在公司监控工人的情况下才能被观察到。常规工作在“高”任务中更有效率，但失败的代价更高。由于监控成本高昂，公司可以选择三种策略之一。首先，它可以将工人置于高任务中并希望最好。第二，公司可以将工人分配到低级别的任务，并对工人进行监控。当危机来临时，它会观察成功和失败，知道工人是好是坏。解决危机的工人将被转移到高级任务。最后，它可以将工人分配给低任务，而不是监视和等待失败。随着时间的推移没有失败，工人成功解决危机的可能性越来越大。当足够长的时间过去后，公司将把工人分配到高任务。该项目的第一个任务是完成一个最优分配模型，该模型扩展了部分监控、假阳性和假阴性以及考虑复杂性不断增加的多个任务。第二个任务是将该模型嵌入到更广泛的劳动力市场和工资决定模型中。这种分析有望使PI能够解释非管理劳动力市场的重要特征。

The broad goal of the proposed research is to understand how the nature of information and the technology for acquiring it affect wages, promotion paths and task-assignment within firms (the internal labor market). The narrower goal is to explain some of the regularities and variation in the structure of internal labor markets for non-managerial workers. In contrast with much of the literature, the PIs do not focus on policies intended to incentivize hard work or discourage moral hazard, which may be an intuitively unappealing starting point and might lead to conclusions that are inconsistent with the data. Instead, this project focuses on monitoring in order to assess workers.The baseline model is one with two types of workers, good and bad, and two types of task, high and low. In this model most work is routine (at least for suitably trained workers), but workers face occasional "crises" which good workers can resolve but bad workers cannot. Failures are readily observed, but successes are only observed if the firm monitors the worker. Routine work is more productive in the "high" task but failures are more costly. Since monitoring is costly, a firm can choose one of three strategies. First, it can place a worker in the high task and hope for the best. Second, the firm can assign the worker to the low task and monitor the worker. When a crisis arrives, it will observe success as well as failure and know whether the worker is good or bad. Workers who resolve the crisis will be moved to the high task. Finally, it can assign the worker to the low task, not monitor and wait for a failure. As time passes without a failure, it becomes increasingly likely that the worker has successfully resolved a crisis. When sufficient time has passed, the firm will assign the worker to the high task.The first task of this project is the completion of an optimal assignment model with extensions such as partial monitoring, false positive and false negatives as well as the consideration of multiple tasks of increasing complexity. The second task is to embed the model in a broader model of the labor market and wage determination.This type of analysis is expected to enable the PIs to explain important features of non-managerial labor markets.