The merit principle asks hiring panels to choose the person most capable of doing the work, yet panels are rarely rewarded for doing so when the most capable person looks unconventional. The safer course, in practice, is to appoint someone who has held the same job title before, because that decision is easy to defend if it goes wrong. The problem, then, is not that panels lack judgement, but that capability-based judgement is harder to defend than precedent.
This article contends that the remedy is to make good judgement defensible, rather than to ask panels to be braver. Four practices would do most of the work: scoring a work sample against the capabilities the role genuinely requires, recording the reasons for the decision, using probation as it was designed, and giving panels explicit cover for documented capability-based decisions. Each moves the defence of a decision away from the candidate’s last title and towards the panel’s reasoning. Hence the aim is a system in which choosing the best candidate is also the safest thing a panel can do.
What the merit principle already asks
The Australian Public Service does not need a new rule to select on capability, because it already has one. Section 10A of the Public Service Act 1999 bases a merit decision on the relative capacity of candidates to achieve outcomes related to the duties, judged against the work-related qualities genuinely required. Nothing in that test refers to the title a candidate last held, and nothing in it prefers precedent to demonstrated capability.
The law, in other words, already asks for the judgement that good selection research recommends. What it cannot do on its own is make that judgement comfortable to exercise. A panel that appoints a candidate whose background differs from the role’s usual pathway must still be confident that its decision will be understood, and supported, if the appointment does not work out.
Thus the gap is not between the law and good practice, which largely agree, but between the law and the incentives surrounding the people who apply it. Closing that gap is a matter of design rather than legislation. It asks what would make a panel confident in choosing the candidate the law intends it to choose.
Why practice drifts towards the job title
If the job title were a strong predictor of performance, relying on it would simply be good practice. The selection literature suggests otherwise. In Schmidt and Hunter’s 1998 meta-analysis, years of job experience correlated with performance at roughly .18, well below work sample tests and structured interviews, and Sackett and colleagues’ 2022 re-analysis left experience among the weaker predictors. A matching title is a cruder signal still, since it records that a candidate held a position, not how well they performed in it. None of this means experience is irrelevant. Relevant experience can build the very capabilities a role requires; the point is that a matching title is a weak proxy for those capabilities, and a panel is better served by assessing them directly.
The persistence of the practice is better explained by how decisions are judged after the fact. When an appointment of a conventional candidate fails, the panel can point to a record that any reasonable person would have accepted, and the failure is attributed to the appointee. When an appointment of an unconventional candidate fails, the same failure tends to be read as an error of judgement by the panel. The success of either appointment, meanwhile, brings the panel little credit. Hence the incentive is not to minimise the risk of a poor hire, but to minimise the risk of being held responsible for one.
Gerd Gigerenzer calls this pattern defensive decision-making: selecting an option that protects the decision-maker, even when another option would better serve the organisation. The pattern is common in the public sector. In a survey of managers in public administration, Artinger, Artinger and Gigerenzer found that 80% reported having made at least one defensive decision, and that on average about one decision in four was defensive. It is not a failing of individual character, and it should not be treated as one. Rather, it is the predictable result of an asymmetry in which credit and blame follow different standards, so the remedy lies in correcting the asymmetry, not in reproaching those who respond to it.
Four practices that make merit defensible
1. Score a work sample against the capabilities the role genuinely requires
A work sample asks candidates to perform a representative task from the role, such as analysing a dataset or drafting a brief. It is among the stronger predictors in the selection literature, and it has a second virtue that is often overlooked: it produces evidence a panel can point to. A candidate who scores highest on the work itself, against criteria written before the applications arrived, has been chosen for a reason that does not depend on their last title.
2. Record the reasons for the decision
The defence of a conventional appointment rests on the candidate’s history, which is already on file. The defence of an unconventional appointment must rest on the panel’s reasoning, which is too often left unwritten. A short record of why the chosen candidate was judged most capable, tied to the capabilities the role requires, gives that reasoning the same standing as a matching title. As such, the record becomes the panel’s protection, and the quality of its reasoning becomes what it is judged on.
3. Use probation as it was designed
Probation exists so that an organisation can confirm an appointment against actual performance, yet it is frequently treated as a formality. Used properly, it lowers the cost of a decision that does not work out, because the correction is available early and on clear terms. Thus a panel can afford to back a well-evidenced judgement, knowing that a mistaken one need not be permanent.
4. Give panels explicit cover for documented capability-based decisions
The first three practices create a defensible record, but a panel will rely on it only if senior leaders honour it. Leaders can state, and then demonstrate, that a documented capability-based decision will be judged on the quality of its reasoning rather than on its outcome alone. This applies a principle of credit and blame symmetry: the standard by which a decision would have been praised if it succeeded should be the standard by which it is assessed if it fails. In effect, it removes the extra penalty that currently attaches to choosing the unconventional candidate.
Related work
Much of the ground covered here has been charted before, and that work deserves acknowledgement. The nearest precedent is Peter Cappelli’s Why Good People Can’t Get Jobs (2012), which argues that employers increasingly demand candidates who have already performed the exact job and decline to train anyone who has not. Behind Cappelli lies a longer tradition treating credentials as signals rather than measures, from Michael Spence’s “Job Market Signaling” (1973) to Randall Collins’s The Credential Society (1979).
The behavioural account owes most to Gigerenzer, whose Risk Savvy (2014) popularised the idea of defensive decision-making, and to his later empirical work with Artinger and Artinger on defensive decisions in public administration. The Hidden Workers report by Fuller and colleagues at Harvard Business School and Accenture (2021) added evidence that rigid screening criteria exclude capable applicants for reasons unrelated to the work. Thus the observation that selection filters on proxies, and that the proxies protect those doing the filtering, is well established.
What this article adds lies at the edge of that ground. It locates the problem in an asymmetry between how institutions credit decisions and how they blame them, which suggests a general test for any selection system: does the standard by which a decision would be praised match the standard by which it would be blamed? It also observes that, in the APS, the permissive rule already exists, so the task is to align incentives with a merit principle that is already in force rather than to argue for a new one.
Why it matters
When selection rewards precedent over capability, the cost is rarely visible to those who make the decision. It falls on capable candidates whose experience does not follow a role’s usual pathway, such as career changers, people returning from caring responsibilities or from work overseas, veterans moving into civilian roles, and specialists moving between agencies or sectors. It also falls on organisations that never discover what those candidates might have contributed. Over time it also narrows the workforce, because roles come to be filled from the same pathways regardless of what the work requires.
There is a fairness cost as well. A system that favours those who have already held a role rewards early access to it, and early access is not always distributed by merit. Therefore an institution that intends to select on capability, but in practice selects on precedent, may reproduce advantages it never meant to endorse.
None of this requires panels to take larger risks, and none of it requires new law. It requires a selection process in which good reasoning is recorded and then judged by the same standard whatever the outcome. When choosing the best candidate becomes the safest course for a panel, the safest course will become choosing the best candidate.
References
- Artinger, F. M., Artinger, S., and Gigerenzer, G. (2019). C. Y. A.: Frequency and causes of defensive decisions in public administration. Business Research, 12(1), 9–25. https://doi.org/10.1007/s40685-018-0074-2
- Cappelli, P. (2012). Why Good People Can’t Get Jobs. Wharton Digital Press.
- Collins, R. (1979). The Credential Society. Academic Press.
- Fuller, J. B., Raman, M., Sage-Gavin, E., and Hines, K. (2021). Hidden Workers: Untapped Talent. Harvard Business School and Accenture.
- Gigerenzer, G. (2014). Risk Savvy. Viking.
- Sackett, P. R., Zhang, C., Berry, C. M., and Lievens, F. (2022). Revisiting meta-analytic estimates of validity in personnel selection: Addressing systematic overcorrection for restriction of range. Journal of Applied Psychology, 107(11), 2040–2068. https://doi.org/10.1037/apl0000994
- Schmidt, F. L., and Hunter, J. E. (1998). The validity and utility of selection methods in personnel psychology. Psychological Bulletin, 124(2), 262–274.
- Spence, M. (1973). Job market signaling. Quarterly Journal of Economics, 87(3), 355–374.
- Public Service Act 1999 (Cth), s 10A.