A salary range in a job advert is usually justified on grounds of fairness. Behavioural economics offers a more sober argument: without a published range, every applicant measures an offer against their own expectation, and it is precisely these differing expectations that give rise to pay gaps that no qualification can later explain.
The same offer triggers two different reactions
Suppose a mid-sized company is looking for someone to join its controlling team, and in the end two candidates remain who barely differ in education, years of experience or the impression they make. After the second interview, both receive the same offer of €62,000 a year. Anna had expected around €58,000, so she is pleased, thanks them and accepts. Jonas, on the other hand, had imagined €66,000, so he is disappointed, asks for time to think and comes back with a counter-demand, which is eventually settled at €64,000. The example is invented, but the mechanism behind it is well documented, and anyone who carries out a pay analysis two years later will find a gap between the two that no qualification explains.
Nobody in this story did anything wrong. Anna accepted an offer that was above her expectation, while Jonas renegotiated one that was below his, and the company gave in to a counter-demand that stayed within reason. The gap arose nonetheless, not in the negotiation itself but in the minds of the two candidates, where no personnel file records it.
We measure every offer against an internal reference point
In 1979, Daniel Kahneman and Amos Tversky used prospect theory to describe why people make decisions under uncertainty differently from what classical economics assumes [1]. The core of the theory is that we do not judge an outcome by its absolute value but by whether it lies above or below a reference point, that is, above or below the amount we perceive as neither a gain nor a loss. Anything above it feels like a gain, and with gains we are cautious, because we would rather secure what we have than put it at stake for a little more. Anything below it feels like a loss, and to avert a loss we accept considerably more risk, including the risk that a conversation turns sour.
Applied to salary negotiations, this means that the decision to accept an offer or to renegotiate depends less on the offer itself than on where the person has drawn their own zero line. This is exactly where it becomes relevant to the question of equal pay.
Men set their zero line higher than women
In 2024, a team at the University of Queensland led by Jonathan Levy, Olivia Ru and Agnieszka Tymula examined whether women and men set their reference points at different levels [2]. In an experiment with 579 people who are representative of the US population on several characteristics, they measured it in various ways, namely through people's own salary expectations, through the expected pay for a specific task and through choices between lotteries. Across all measures, men set their reference point higher than women, and people with a higher reference point were also more willing to take risks [2].
This shifts a debate that has been going on for years. That women on average make more risk-averse decisions has been described in many studies, mostly as though it were a fixed trait. But if willingness to take risks is linked to the reference point, then part of this caution is simply the consequence of the same offer landing more often in the gain zone for women and more often in the loss zone for men. Anna does not behave more cautiously than Jonas because she is more cautious, but because she believes herself to be in a different position.
The authors also find indications that shared experience narrows the gap between reference points, and they themselves name transparent pay structures and regular pay equity audits as possible starting points [2]. The study is a working paper and has not yet been peer-reviewed, but it clearly suggests that expectations can be aligned when everyone has the same information.
A single sentence in the advert changes who negotiates
This fits with a field experiment by Andreas Leibbrandt and John A. List, in which almost 2,500 job seekers were randomly assigned to job adverts that differed in one detail [3]. Some of the adverts stated explicitly that the salary was negotiable, while the others left this open. Where nothing was said, men negotiated more often than women, whereas where it was stated, the difference disappeared and even tended to reverse. In addition, men were more likely to apply for positions where the rules of pay setting remained unclear, and that is precisely where the gap was largest [3].
From the perspective of prospect theory, this is hardly surprising. Renegotiating without being invited to is a risk of unknown size, and anyone who experiences this risk as a potential loss, because it might cost goodwill or the offer itself, is more likely to refrain. A single sentence in the advert turns the question of whether one may negotiate into the question of how much one asks for.
The employer's first offer sets the anchor
That leaves the other side of the table. Marcus Dittrich, Andreas Knabe and Kristina Leipold had 122 students at TU Dresden conduct face-to-face wage negotiations, with one person representing the employer and the other the worker in each case [4]. Female employees achieved significantly less on average, but not because they negotiated less well over the course of the negotiation, where the authors found no significant difference. Rather, the gap was already there in the first round, because men in the employer role offered women significantly less from the outset than they offered men [4].
In this logic, the first offer is more than a number, because it sets the reference point against which every subsequent move is measured. Whoever starts lower shifts the entire negotiation downwards, and even if both sides negotiate equally skilfully afterwards, the gap from the start remains.
A published range gives everyone the same reference point
Put these three findings side by side and a fairly simple argument for the salary range in the job advert emerges, one that has little to do with fairness as an attitude and a great deal to do with how people perceive numbers.
A published range gives all applicants the same external reference point before they set their internal one. Anna would have known that €58,000 to €68,000 had been earmarked for the position, and €62,000 would no longer have felt like a gain to her but like an offer from the lower half. Jonas might have expected less than he had assumed without the information. Both would have measured the same offer against the same figure.
If the advert also states explicitly that negotiation takes place within this range, renegotiating loses part of its risk as well, and the ambiguity in which, according to Leibbrandt and List, the gap is largest disappears [3]. A range without this note leaves the most important question open.
Beyond that, the range also binds the employer itself. An employer that has published a range must place its first offer within it and should be able to explain why it sits at the lower or upper end. This turns precisely the moment at which the gap arose in the Dresden experiment into a decision based on criteria rather than impressions.
A range alone does not close the gap
A few caveats belong here. None of the three studies directly tested what a salary range in a job advert achieves, so the argument connects their findings but does not replace a study of its own. A range also only helps if it is realistic. €50,000 to €80,000 for a clerical position does not provide a shared reference point; it leaves applicants guessing, just within a wider frame. It is also conceivable that some applicants orient themselves towards the lower end of the range because it feels safe, which could reopen the gap elsewhere. Employers that publish ranges should therefore track where women and men actually end up within them.
In closing
For employers that have to carry out their pay analysis anyway, there is one more practical thought here. An adjusted pay gap that has built up over years is difficult to correct, whereas one that never arises at the point of hiring will never need to be explained later.
Frequently asked questions
1. Does the Pay Transparency Directive require the salary range to appear in the job advert?
No. Article 5 requires that applicants learn the initial pay or its range before the job interview; the job advert is one possible way of doing this, but not an obligation [5]. Employers may not ask about previous pay.
2. What is a reference point in a salary negotiation?
The amount a person perceives as neither a gain nor a loss. According to prospect theory, where an offer lies relative to this point helps determine whether someone accepts it or renegotiates [1] [2].
3. Is a published range enough to prevent gaps in starting salaries?
Not on its own. The range should be realistically set and include a note that negotiation takes place within it [3]. It is also worth tracking where women and men actually end up within the range.
Sources
[1] Kahneman, D. & Tversky, A. (1979): Prospect Theory: An Analysis of Decision under Risk. Econometrica 47(2), pp. 263–291.
[2] Levy, J., Ru, O. & Tymula, A. (2024): The gender reference point gap: Evidence from a representative sample. Life Course Centre Working Paper 2024-17, The University of Queensland. DOI: 10.14264/787065c.
[3] Leibbrandt, A. & List, J. A.: Do Women Avoid Salary Negotiations? Evidence from a Large-Scale Natural Field Experiment. NBER Working Paper 18511.
[4] Dittrich, M., Knabe, A. & Leipold, K. (2012): Gender Differences in Experimental Wage Negotiations. CESifo Working Paper No. 3862.
[5] Directive (EU) 2023/970 of 10 May 2023 (Pay Transparency Directive), Art. 5.
