Bayes’ application to HRM is limited to event probability but is a topic that is mentioned in passing in a paper on shirking and presenteeism by S. Brown (2004) recommended by Dr. Andrew Bryce (Sheffield), which was written over 20 years ago this year.
Brown (2004) reads as follows:
“Such ‘shirking’ is potentially costly to firms and may incite them to undertake monitoring. BST** envisage a monitoring technology in which there is some probability, α < 1, of each absentee’s true state of health being revealed to the firm.”
After deciding to initiate a brief scoping review for the B812 literature topic of choice (‘Wellbeing’). I thought I would check in with the blog and provide some justification and background for this choice of theme.
This spurious love affair with Bayes’ theorem has loomed over my educational learnings but only in its form as sub-game perfect in non-cooperative game theory. Big thanks to Melvyn Coles, Pierre Regibeau, and Franco Squintani for their lectures and classes from our days in Colchester on Economics.
I started the HRM course in Nov 2023 and while I am still somewhat aware of some concepts surrounding Bayes, things have changed. Nowadays, Bayes’ theorem (10+ years on) is being used in combination with what we call supervised learning and algorithmic techniques such as neural networks.
So, how do we proceed? Perhaps, it is wise to proceed with caution. A brief scoping review will get me up to speed and updated with new research as much as is feasibly possible.
References
Brown, S. and Sessions, John (2004) “Absenteeism, Presenteeism and Shirking”, Economic Issues, 9(1), pp. 15-22 – Available at: https://econpapers.repec.org/article/eisarticl/104brown.htm (Accessed on 13 March 2024)
**Barmby, T. A., Sessions, J. G. and Treble, J. G. (1994) “Absenteeism, Efficiency Wages and Shirking”, Scandinavian Journal of Economics, 94(4), pp. 561-566 – Available at: https://www.jstor.org/stable/3440797 (Accessed on 13 March 2024)
-------
This post was written by Alfred Anate Mayaki, a student on the MSc in HRM, and was inspired by the author's previous learnings and experiences.
Comments
Monitoring as a Feedback Mechanism: Boundary Problems in the Shirking Model
This post opens a genuinely systemic question beneath its game-theoretic framing. Bayesian SPNE treats the firm's monitoring probability as a parameter, but it functions as a feedback mechanism. The worker's behaviour responds to the monitoring system, which in turn responds to observed outcomes. Structure determines behaviour. The shirking model also illustrates a boundary problem. The firm defines the system boundary around observable attendance. The worker's actual health state sits outside that boundary. When the monitoring system cannot distinguish between shirking and genuine illness, it produces an unintended consequence: presenteeism. The control mechanism generates the behaviour it was designed to prevent. The shift from Bayesian updating to supervised learning does not resolve this. It changes the boundary again. The worker becomes a data point in a model they did not design and cannot contest. That is a systems question, not a statistical one. Regards,
Russell Larke
BA (Hons) Business Management | MSc Candidate (Systems Thinking)
Trading Beyond Charts