Low Power, First Offers, and Reservation Prices: Weak Negotiators are Self‐anchored by Their Own Alternatives

To date, the literature on decision‐making and negotiations has supported a first‐mover advantage in negotiations, which is attributed to the anchoring heuristic (e.g., Galinsky and Mussweiler 2001). However, contemporary studies have revealed a “practitioner‐researcher paradox” (Loschelder et al. 2014, 2016; Maaravi and Levy 2017). This paradox describes the contrast between the findings in the research literature pointing toward a first‐mover advantage, and the tendency of negotiation experts and practitioners to advise making the second move (i.e., the counteroffer) in negotiations (Dell and Boswell 2009), or at least adapting this decision to the specific situation. For example, Loschelder et al. (2014) have shown that when the first offer might reveal one's preferences in a multi‐issue negotiation, deciding whether to move first or second might depend on the counterpart's reputation as pro‐social (better to move first) or pro‐self (better to move second).

In the current research article, we aim to explore this paradox by investigating a paramount factor in most negotiations: power in the form of alternatives to the negotiation (Bazerman and Neale 1993; Thompson 2005; Malhotra and Bazerman 2007). Schaerer and colleagues (2015, 2016) demonstrated that having low power (i.e., inferior alternatives) might reduce the value of one's first offer. It was argued that the psychological mechanism behind this result was that weak alternatives (i.e., having low power) act as detrimental self‐anchors. That is, low‐power negotiators might be influenced (or anchored) by their own inferior alternatives when deciding the amount of their first offers. Inspired by these findings and more recent research (Maaravi and Heller 2021), we aim to shed more light on the practitioner‐researcher paradox by exploring the favorability of making the first versus second offer under low‐power conditions. Specifically, we compared the first offers made by low‐power negotiators to those of both medium‐ and high‐power negotiators in a popular negotiation scenario. Our findings contribute to the current literature by focusing specifically on the favorability of making the first offer rather than only measuring the differences in the first offers of one side of the negotiation across power levels.

First offers in negotiation have become an important area of negotiation research, with extensive research being conducted over the past two decades (e.g., Oesch and Galinsky 2003; Ritov and Moran 2008; Shalvi, Moran, and Ritov 2010; Gunia et al. 2013; Schaerer, Loschelder, and Swaab 2016; Maaravi and Hameiri 2019). The importance of first offers has been attributed to two factors: (1) the universal requirement that one of the parties make the first offer in negotiations; and (2) the effects that first‐offer magnitude have on subsequent counteroffers and the outcomes of negotiations. Indeed, every negotiation must start with one side making the initial offer (Adair and Brett 2005). For example, salary negotiation, a commonplace yet integral organizational activity, usually entails employers either declaring how much they are willing to offer or asking the candidates to state their salary expectations. Moreover, first offers were found to be of great consequence to the process and the outcome of negotiations, affecting counteroffers and settlement prices (Maaravi, Pazy, and Ganzach 2014) via the cognitive process of anchoring and adjustment (Tversky and Kahneman 1974). For example, the propensity to move first and the amount of the first offer were suggested as possible factors explaining salary differences (Babcock and Laschever 2009). This cognitive process is best described as the assimilation of a numeric judgment into a previously considered standard of comparison (Tversky and Kahneman 1974). First offers act as anchors that affect the judgments that negotiation parties make when judging and computing their desired counteroffer (Galinsky and Mussweiler 2001). One study showed that when sellers made the first offers, they made higher ones than buyers, which resulted in higher settlement prices. The same was true, but in the opposite direction, when buyers made the first offers (Galinsky and Mussweiler 2001).

Based on these and similar findings, negotiation researchers have suggested that negotiators make extreme (yet reasonable) first offers if they wish to tilt the negotiation in their favor (e.g., Thompson 2005). Coincidentally and of great consequence, most negotiations involve judgment and decision‐making in uncertain conditions, which are prime circumstances for biases and heuristics (Tversky and Kahneman 1974) such as the anchoring and adjustment heuristic. (For a review of the anchoring heuristic, see Furnham and Boo 2011.)

To quantify the degree to which the negotiation literature recommends moving first as opposed to second, we have relied on a recent review of the literature on first offers in negotiation that was conducted by Lipp, Smolinksi, and Kesting (2022). In their systematic review, the authors searched four popular databases to find articles investigating the “first offer” in negotiation. After analyzing the literature according to strict exclusion/inclusion criteria, they settled on 119 articles published between 1967 and 2021. Of these articles, only 15 (17.85 percent) discussed either a first‐mover disadvantage (i.e., second‐mover advantage) or another contextual limitation that leads first offers to have a non‐positive effect. To sum up, the negotiation literature is replete with studies whose findings both explicitly and implicitly argue for a first‐mover advantage.

Although most of the findings in the literature on first offers argue that negotiators should make first offers, a growing body of research has emphasized the disadvantages and limitations of doing so. (An overview of the studies mentioned in this section can be found in Table One below.) One study revealed that negotiators who made first offers had higher anxiety levels and lower satisfaction levels than those who moved second (Rosette, Kopelman, and Abbott 2014). Additional research showed that learning about the anchoring tactic of first offers (i.e., “move first, make a fairly extreme offer”) and then using it in a negotiation resulted in less favorable psychological and economic results over the long term (Maaravi, Pazy, and Ganzach 2014). In one negotiation simulation, the counterparts of first movers were less satisfied with the negotiation than counterparts of second movers and were thus less willing to engage in future negotiations. In a simulated market‐setting negotiation, first movers closed fewer deals than second movers, which reduced their total profits (Maaravi, Pazy, and Ganzach 2014). Finally, as foreshadowed above, making too extreme a first offer may lead to an impasse, defined as “a negotiation in which one or two parties discontinue the interaction, either because one or both parties prefer no agreement, or because they could not reach an agreement despite them benefitting from doing so” (Schweinsberg, Thau, and Pillutla 2022: 50). Schweinsberg et al. (2012) showed that impasses occurred when extreme first offers offended negotiation counterparts.

In light of this growing literature, researchers have begun to cast doubt on the notion that making the first offer is more advantageous in both distributive (Maaravi and Levy 2017) and integrative scenarios (Loschelder et al. 2014). Loschelder et al. (2014) showed that making the first offer backfired in negotiations when the first offer included a compatible issue (thus revealing private information about compatible preferences), which second movers took advantage of to make more beneficial counteroffers. Maaravi and Levy (2017) offered a conceptual framework for analyzing the advantages and disadvantages of making the first offer based on market value information and additional information regarding the counterpart's knowledge of the market value. They found a first‐mover disadvantage where there was information asymmetry between the negotiation parties in a distributive negotiation scenario.¹ The studies by both Loschelder et al. (2014) and Maaravi and Levy (2017) highlight the need to investigate the conditions in which making the first offer is disadvantageous, yet they have examined just one factor—information. To determine the truth behind the practitioner‐researcher paradox, this article studies additional factors. Namely, we aim to shed more light on this paradox by investigating the effects of power on the favorability of making the first offer as a proxy for success in negotiations.

To collect preliminary evidence displaying the practitioner‐researcher paradox, we interviewed expert practitioners who regularly negotiate as part of their jobs. The participant pool included 49 negotiation experts and practitioners from around the world (Israel, Germany, Portugal, Russia, and the U.S.) across various occupational fields (e.g., IT, healthcare, financial services, and government agencies). We asked participants if they preferred to move first or second in a negotiation.² Whereas foundational and more dated negotiation research predicts that most would recommend making the first offer, our results show that only 22.45 percent made such recommendation. Almost thirty‐seven percent (36.73 percent) preferred making the counteroffer, while the remaining 40.82 percent recommended an adaptive approach based on the circumstances of the negotiation (i.e., power, market dynamics, etc.). In other words, it seems that those who practice negotiation do not agree with the academic literature, hence the term “practitioner‐researcher paradox.”

Researchers have investigated various determining factors of first offers, such as market prices (Buelens and Van Poucke 2004), reservation prices (Pruitt and Carnevale 1993), and aspiration prices (White and Neale 1991). More relevant here, additional studies focused on counterparts' or negotiators' alternatives (e.g., Buelens and Van Poucke 2004), specifically their best alternatives to a negotiated agreement (BATNAs) (Schaerer, Swaab, and Galinsky 2015). BATNAs were found to determine negotiators' feelings of power and, consequently, affected first‐offer magnitude and negotiation results (Magee, Galinsky, and Gruenfeld 2007; Schweinsberg et al. 2012). For example, studies have shown that considering one's own or one's counterpart's alternatives before making the first offer significantly impacted the magnitude of first offers, counteroffers, and results (Galinsky and Mussweiler 2001; Maaravi, Ganzach, and Pazy 2011). Magee, Galinsky, and Gruenfeld (2007) have additionally shown that increasing negotiators' power—via priming or better alternatives—increased the likelihood that they would make the first move and, of greater importance to the current article, increased the propitiousness of their first offer. These two behaviors, separately or combined, improved their results due to the anchoring effect discussed above. Thus, it is argued that high‐power negotiators benefit from making the first offer and should strive to do so (Gunia et al. 2013). Past research has suggested two psychological mechanisms to explain these behaviors.

First, the propensity of high‐power negotiators to move first is explained based on the activation of the behavioral approach system (BAS), which is a component of the power‐approach theory (Keltner, Gruenfeld, and Anderson 2003). According to this theory, power may influence the activation of two fundamental behavioral systems: behavioral approach and inhibition systems (Keltner, Gruenfeld, and Anderson 2003). The BAS is hypothesized to motivate high‐power individuals to search for rewards in their surroundings, be more proactive, and have less concern for the social consequences (Keltner, Gruenfeld, and Anderson 2003). Second, the extremity of their offers is explained by the anchoring effect of their strong alternatives (Schaerer, Swaab, and Galinsky 2015). That is, since alternatives are prominent reference points, they influence (i.e., anchor) negotiators when they decide the amount of their first offers (Kahneman 1992).

Symmetrically, past research has shown that low‐power negotiators would be less inclined to make the first offer (Magee, Galinsky, and Gruenfeld 2007). Moreover, low‐power negotiators would make lower first offers than medium‐ or high‐power negotiators (Schaerer, Swaab, and Galinsky 2015). First, Magee, Galinsky, and Gruenfeld (2007) have argued that while the BAS explains the behavior of high‐power negotiators, the BIS (behavioral inhibition system) is a complementary system of the same power‐approach theory (Keltner, Gruenfeld, and Anderson 2003) that might account for the behavior of low‐power negotiators. According to this theory, low power activates the BIS, which serves as an alarm system and governs risk assessment and defensive avoidance behaviors. If strong negotiators (activation of BAS) were shown to prefer moving first in negotiations (i.e., searching for rewards in their surroundings, being more proactive), then it can be hypothesized that low‐power negotiators (activation of BIS) will avoid doing so (i.e., avoiding punishments in their surroundings, being more passive). Second, when weak negotiators do make the first offers, their offers should be negatively affected by their inferior alternatives (Schaerer, Swaab, and Galinsky 2015). As explained above, alternatives are such salient reference points (Kahneman 1992) that negotiators are self‐anchored by them when they decide the amount of their first offers. Thus, low‐power negotiators anchored by their weak alternatives are hypothesized to make worse first offers than medium‐ or high‐power negotiators.

As a result, there may be situations in which negotiators' first offer is not only damaged by their power, but even making the first offer may not be as favorable as making the counteroffer. Imagine, for example, an employer negotiating a contract with job candidate A, while having another alternative—candidate B. Not knowing A's expectations, the employer might use B's salary request as a reference point and adjust the first offer accordingly. Assume that according to market information, annual salaries for this job range from $100,000 to $150,000 and that candidate B asked for $160,000. The employer, anchored by this alternative, might offer candidate A $140,000. But if candidate A, having a weak alternative of her own, planned on requesting $120,000, the employer would end up making a disadvantageous first offer and paying more than was necessary. The results of Schaerer, Swaab, and Galinsky (2015) supported this logic, showing that a lack of alternatives was more beneficial to negotiators than having weak alternatives because weak alternatives anchored negotiators to make modest first offers. Yet previous results did not investigate the effect of low power on the favorability of making the first versus the second offer; they only examined its effect on the magnitude of the first offer. Thus, the current article aims to test the following hypothesis:

H1: Low‐power negotiators make first offers that are less beneficial than the offers they would have received if they moved second.

An additional factor known to affect the outcomes of negotiations is reservation price (Raiffa 1982), or the price point where the negotiator is indifferent between accepting the offer or ending the negotiation (i.e., the lowest/highest price a negotiator is willing to accept). Studies have shown that reservation prices impact negotiations' outcomes by creating a “zone of possible agreement” (ZOPA) (White and Neale 1991; White et al. 1994; Kristensen and Gärling 1997a). ZOPA is the area, or overlap, between the parties' reservation prices (Raiffa 1982). For example, if a job candidate's reservation price is an annual salary of $110,000 and the employer's is $140,000, then their ZOPA is the $30,000 between these two numbers and any agreement is expected to fall within this zone. Consequently, the wider the zone of possible agreement, the higher the probability of avoiding an impasse (Schweinsberg et al. 2022) and achieving a settlement (Myerson and Satterthwaite 1983). Thus, reservation prices are crucial in determining the results of any negotiation above and beyond the first offers.

While research has shown that a negotiator's reservation price is determined by market information (Kristensen and Gärling 1997b) and egocentric bias (Babcock et al. 1995), the effects of power on determining reservation prices have not been thoroughly investigated. One exception is an article by White and Neale (1991) clarifying the distinction between BATNAs and reservation prices. Although these concepts are interrelated, the authors argue that they are not conceptually equivalent (as explained above). Being a salient reference point, the BATNA is expected to affect not only first offers but also reservation prices. Essentially, we would expect low‐power individuals to be self‐anchored by their own weak alternatives and consequently harbor less beneficial reservation prices than their counterparts (i.e., lower for sellers and higher for buyers). These reservation prices might be less beneficial by asymmetrically “pulling” the ZOPA in one direction against the low‐power negotiator's interests. Since most negotiations settle on the midpoint of the ZOPA (White et al. 1994; Thompson, Wang, and Gunia 2010), such a shift is expected to result in inferior results. Continuing the example above, if the job candidate has a low BATNA of $80,000 and thus decides that the reservation price would be $90,000, then the ZOPA is between $90,000 and $140,000. Therefore, all other factors being equal, the expected settlement price would be $115,000, which is inferior for this job candidate than $125,000 (the midpoint between the original reservation prices of $110,000 and $140,000).

Taken together, although an inferior reservation price may increase the probability of achieving a deal (by expanding the ZOPA), it may also result in a worse settlement price for the low‐power negotiator. Thus, due to the importance of reservation prices to the outcomes of negotiations, and prior research relating them to power in the form of BATNAs (White et al. 1994), we sought to investigate whether low‐power negotiators might have less beneficial reservation prices as follows:

H2: Low‐power negotiators have less beneficial reservation prices than negotiators with medium or high power.

Given the robustness of anchoring effects in general and the hypothesized detrimental self‐anchoring effect of low power on the favorability of first offers and reservation prices (i.e., H1 and H2), we sought to test whether a debiasing technique could be used to help negotiators nullify these effects. One category of strategies has been consistently proven to mitigate judgmental biases: considering the opposite (Lord, Lepper, and Preston 1984). This entails the consideration of alternative information (which is inconsistent with your current knowledge or beliefs and frames problems in a different way), which has been used to attenuate the effects of hindsight bias (Fischhoff 1975), confirmation bias (Snyder and Swann 1978), and belief perseverance (Ross, Lepper, and Hubbard 1975), to name a few.

In the case of the self‐anchoring effect of low power, one of the most relevant “consider the opposite” strategies is to focus on one's own target price, which is the negotiator's ideal or preferred outcome. Galinsky and Mussweiler (2001) demonstrated the effectiveness of using this strategy as an anchor‐debiasing technique. In the debiasing condition, the participant who made the counteroffer was offered the following information: “When preparing your negotiation it is important to think about and focus on your target, the ideal price at which you could sell/buy. A clear understanding of this price will assist you in preparing for the negotiation” (Galinsky and Mussweiler 2001: 664). Their results showed that counteroffers were higher for sellers (and lower for buyers) in the debiasing condition and that the anchoring effect of their counterpart's first offer was entirely eliminated when examining final settlement prices. That is, negotiators in the “focus on target” condition achieved better negotiation outcomes than those in the control condition. Thus, we seek to investigate whether such a strategy could debias the hypothesized detrimental self‐anchoring effects of low power on first‐offer favorability and reservation price. Our hypothesis was:

H3: The detrimental effects of low power are mitigated when low‐power negotiators use an anchor‐debiasing technique (“consider the opposite”).

The importance of the research reported in this article is threefold. First, exploring the determinants of first offers and the factors that make them advantageous or disadvantageous is central for both research and practice. Second, the current research seeks insight into alternative cognitive explanations of judgment and decision‐making in negotiations, as instead of focusing on the anchoring effect of first offers (Galinsky and Mussweiler 2001), it deals with the cues and reference points that negotiators use when deciding the magnitude of their first offer (Schaerer, Swaab, and Galinsky 2015; Maaravi and Levy 2017). Finally, the current research aims to offer novel insights into the specific research domain of low power and first offers. Whereas previous research has investigated the effects of low power on first‐offer magnitude (Schaerer, Swaab, and Galinsky 2015) or on negotiators' tendency to make the first versus second offer (Magee, Galinsky, and Gruenfeld 2007), the current research seeks to explore the effect of low power on the favorability of making the first offer. By examining both parties' first offers and comparing them, we may infer whether the negotiators would have been better off making the counteroffer rather than the first offer. By doing so, the current article contributes to the growing exploration of the practitioner‐researcher paradox of first offers in negotiations by focusing on negotiation power alongside existing research on information.

In Study 1, we used an exhaustive design to test whether negotiators (employers versus job candidates) with low power (in the form of BATNAs) would make less beneficial first offers than the ones they would have received were they to move second (H1). Importantly, we compared their first offer to those of counterparts with varying degrees of power. Additionally, we examined whether low‐power negotiators had less beneficial reservation prices than medium‐ or high‐power negotiators in the same position (H2).

We have chosen a widely used negotiation setting for our research: job candidate versus recruiter. Over the past few decades, this setting has been manipulated and investigated in numerous negotiation studies (e.g., Sinaceur and Tiedens 2006; Magee, Galinsky, and Gruenfeld 2007; Galinsky et al. 2008). Notably, in several of them, the job candidate versus recruiter paradigm was the only setting used (e.g., Swaab, Maddux, and Sinaceur 2011), probably due to its centrality for most negotiators. While other negotiations—such as those involving commercial transactions or divorce—may not occur in one's personal or professional life, most people work to make a living and negotiate as candidates or recruiters (Lipovsky 2006); they easily understand the setting and identify with it. Moreover, such settings are critical in power‐related research. Past research has long established that people interpret information differently depending on the negotiation context and their role (Neale, Huber, and Northcraft 1987). In the candidate versus recruiter setting, participants in a simulated negotiation were particularly sensitive to power issues and their perceptions of power differed; recruiters typically felt more powerful than candidates. In contrast, candidates felt “somewhat at the mercy of a recruiter” (Wolfe and McGinn 2005: 16). These findings are in line with the power‐approach theory presented above (Keltner, Gruenfeld, and Anderson 2003), which is part of the theoretical framework of the current research.

299 participants (121 males, 174 females, and 4 others) with an average age of 27.76 were recruited via the “Prolific” crowdworking platform and were compensated for their participation.

The experiment consisted of a 2 (role: job candidate versus employer) * 3 (power level: weak, mediocre, or strong BATNAs) between‐subjects factorial design. The dependent variables were the first offers made by participants, their stated reservation prices, and their preference to move first versus second.

Participants playing job candidates were asked to imagine they were looking for a job as a computer programmer. They read that the yearly salary for this role ranged between ₤32,000 and ₤72,000, averaging at around ₤56,000 per year. They were told that they already had received a job offer from another company (Company A) and that after negotiating with company A's human resources representative, they were offered ₤34,000 (low power), ₤52,000 (medium power), or ₤70,000 (high power) per year. They then received a description of a different company (Company B) and were told that they were entering negotiations with the company's human resources representative over their yearly salary.

Participants in the employer role were told they were recruiters for a big tech company looking to recruit a computer programmer. They received the same market information (salary range and average) as those in the candidate role. They were then told that they already had met a promising candidate (Candidate X) and that after negotiations, the candidate requested a salary of ₤70,000 (low power), ₤52,000 (medium power), or ₤34,000 (high power) per year. They then received a description of an additional candidate (Candidate Y) and were told that they were about to enter negotiations with the latter over the yearly salary.

After reading the scenario, participants stated whether they preferred to make the first offer or let their counterpart make it. Following this question, they were asked to imagine that they were the first movers and were asked to make the first offer. Finally, participants declared their reservation price. That is, candidates stated the lowest salary they would be willing to accept and employers the highest salary they were ready to offer.

A chi‐squared test of independence was performed to examine whether candidates and employers differed in their preference to make the first move (across all power conditions). The results showed a significant relationship between role and preference to make the first offer. Whereas only 22.2 percent of the candidates preferred to make the first offer, 40.4 percent of employers showed this preference. These results echo the power‐approach theory (Keltner, Gruenfeld, and Anderson 2003) and its behavioral approach system (BAS). They suggest that employers might be more proactive and action‐oriented than job candidates due to their innate power (Magee, Galinsky, and Gruenfeld 2007).

Another chi‐squared test of independence was performed to examine whether participants in different power conditions (both candidates and employers) differed in their preference to make the first move. The results showed a non‐significant relationship between power and preference to make the first offer. Interestingly, when controlling for the role, a chi‐squared test of independence showed a marginally significant association between power and preference to move first among candidates, such that the higher the power, the greater the proportion of participants preferring to move first. Removing the medium‐power group resulted in a significant relationship, where only 13.2 percent of low‐power candidates preferred to make the first move compared to 31.3 percent of high‐power candidates. This finding would be rational if having low power was found to produce a first‐mover disadvantage, which is a question we deal with in the following analyses.

The same analysis found a non‐significant relationship between power and initiation preference among employers. This difference between roles might be explained by the initial perception of power among recruiters (feeling more powerful) versus candidates (who are said to feel more passive and, as noted above, “at the mercy” of the recruiters) (Wolfe and McGinn 2005: 16).

Next, we ran a general linear model with the first offer as the dependent variable and both role and power as the independent variables. Results showed a significant effect of role on the first offer, such that candidates' first offers (in thousands of pounds sterling per year) were significantly higher than those of employers (M_candidate = 59.49, M_employer = 46.28). There was no significant effect of power on first offers, yet when controlling for the role, power significantly affected both the candidate's and the employer's conditions. Whereas in the candidate role, higher power resulted in higher first offers, in the employer role, higher power resulted in lower first offers, as can be seen in Table Two. Tukey's HSD post‐hoc test for both role conditions showed significant differences in first offers between all three power groups.

When looking at low‐power negotiators, we sought to compare their first offer with the one they would have been given were they to move second. When both candidates and employers were low in power, their first offers were significantly different, such that employers offered a significantly higher salary (M = 56.34) than the candidates asked for (M = 50.26). In other words, when both negotiators have low power (weak BATNAs) they should have preferred to move second because the first offer made by their counterpart was more beneficial than their own first offer (H1).

When comparing low‐power candidates (M = 50.26) with medium‐power employers (M = 48.53), there was no significant difference between their first offers. We argue that here too, candidates were better off letting employers move first since it would have allowed candidates to bargain and increase the salary from that point instead of making the same offer themselves and allowing employers to bargain and decrease the salary. Only when low‐power candidates were compared to high‐power employers (M = 35.08) was there a significant difference, such that low‐power candidates would benefit from moving first.

When shifting the focus to employers and comparing low‐power employers (M = 56.34) with medium‐power candidates (M = 58.59), there was no significant difference between their first offers. Echoing the findings regarding low‐power candidates versus medium‐power employers, low‐power negotiators (employers) were better off not making the first offer. That is, letting candidates move first would allow employers to bargain and decrease the salary from that point instead of making the same offer themselves and allowing candidates to bargain and increase the salary. Only when low‐power employers were compared to high‐power candidates (M = 70.65) was there a significant difference, such that low‐power employers would benefit from moving first and setting the beneficial anchor. In other words, when low in power, both candidates and employers should avoid making the first offer in most cases (except when their counterpart has high power). Not making the first offer, they would receive more favorable (or at least similar) offers than they would have made themselves (H1).

Finally, we ran a similar model with reservation price as the dependent variable and both role and power conditions as the independent variables. Results showed a significant effect of role on reservation price, such that candidates' reservation price (in thousands of pounds sterling per year) was significantly lower than that of employers' (M_candidate = 52.88, M_employer = 54.33). This pattern would be expected given that, as previously stated, reservation price is the lowest (highest) salary that candidates (employers) would be willing to accept (offer). There was no significant effect of power on reservation price, yet when controlling for the role, power significantly affected both the candidate and the employer. Whereas in the candidate condition, higher power resulted in higher reservation prices, in the employer condition, higher power resulted in lower reservation prices, as can be seen in Figure One. Tukey's HSD post‐hoc test for both role conditions showed a significant difference in reservation price between all three power groups. In other words, it seems that the effect of low power is not limited only to first offers; rather, it also negatively affects negotiators' reservation prices and thus has another channel for determining the final settlement price (H2).

Taken together, these results point toward a first‐mover disadvantage when negotiators have low power for both roles and in one third of all combinations (see Table Two). Additionally, low power affects the preference to move first versus second, such that higher‐power negotiators prefer to move first to a greater degree than low‐power negotiators. Our findings suggest that this preference pattern is entirely rational, although negotiation courses, textbooks, articles, and seminars based on existing research may advise otherwise (e.g., Thompson 2005).

After confirming the existence of the adverse effect of low power on our dependent variables, we designed Study 2 to test the robustness and strength of this effect by testing whether it persists even in the presence of a classic anchor‐debiasing technique.

In Study 1, we established that low‐power negotiators make less beneficial first offers than the ones they would have received had they moved second. In Study 2, we aimed to investigate whether such negotiators can restore the classic anchoring advantage of making the first offer by using the bias‐attenuating technique of “considering the opposite” (Lord, Lepper, and Preston 1984; Mussweiler, Strack, and Pfeiffer 2000). Specifically, a “focus on your target” strategy should lead negotiators to consider its inconsistencies with their weak BATNA, reducing any BATNA‐related anchoring effect.

199 participants (73 males, 124 females, and 2 others) with an average age of 27.77 were randomly recruited via the “Prolific” crowdworking platform and were compensated for their participation.

The experiment consisted of a 2 (role: job candidate versus employer) * 2 (use of debiasing technique: no versus yes) between‐subjects factorial design. All participants were randomly assigned to either the low‐power candidate or low‐power employer role. They were then assigned to either a counterfactual thinking or control intervention condition. The dependent variables were the first offers made by participants, their stated reservation prices, and their preference of moving first versus second.

We used the same scenario as in Study 1 (job candidate versus employer), except that both candidates and employers were only given weak BATNAs. Additionally, and before filling out the dependent variable measures, participants in the debiasing technique condition were given the following text (adapted from Galinsky and Mussweiler 2001): “Important Preparation Information: Research has shown that when negotiators have a weak alternative to their current negotiation, they tend to make a suboptimal first offer. To improve your results, we advise you to ignore your current alternative and instead focus on your target—the ideal salary you would like to get. A clear understanding of this target salary will assist you in preparing for the negotiation.” (In the employer condition, the text stated: “…the ideal salary you would like to offer rather than the ideal salary you would like to get.”) They were then asked to think about their target price and write it down in the provided text box.

After reading the scenario and the above text (for those in the target intervention condition), participants were required to state whether they preferred to make the first offer or let their counterpart do it. Following this question, they were asked to imagine that they were the first movers and to state their first offers. Finally, participants declared their reservation price—the lowest (highest) salary they would be willing to accept (offer) as candidates (employers) in this negotiation.

A chi‐squared test of independence was performed to examine whether candidates and employers differed in their preference to make the first move. As in Study 1, the results showed a significant relationship between role and preference to make the first offer. Whereas only 25 percent of the job candidates preferred to make the first offer, 46.5 percent of employers showed this preference. Consistent with the power‐approach theory (Keltner, Gruenfeld, and Anderson 2003) and its behavioral approach system (BAS), this pattern may suggest that employers tend to be more proactive and action oriented, given the innate power of their role (Magee, Galinsky, and Gruenfeld 2007). Being in the target intervention condition had no significant effect on this disparity, and there was no significant relationship between this condition and the preference to move first. In other words, these results echo those of Study 1, which showed a clear difference in initiation preferences between candidates and employers. Yet, focusing on one's target price did not change this preference, a finding that would be rational only if target‐focus would not change the adverse effects of low power on first offers.

Next, we ran a general linear model with first offer as the dependent variable and both role and debiasing‐technique condition as the independent variables. Results showed a significant effect of role on the first offer, such that candidates' first offers (in thousands of pounds sterling per year) were significantly lower than those of low‐power employers' (M_candidate = 51.45, M_employer = 54.74). This result is in line with the similar analysis made in Study 1 and shows how low‐power negotiators should usually strive to move second (H1).

There was no significant effect of the intervention on first offers, yet this is precisely what we would expect to find given that the debiasing technique, target‐focus, should affect first offers in a different direction depending on the role (thus, eliminating the effect). When controlling for the role, the debiasing technique had a significant effect in the candidate condition but not in the employer condition. Whereas in the candidate role target‐focus resulted in higher first offers (M_target = 55.13, M_control = 47.45), in the employer role, target‐focus did not result in lower first offers, as can be seen in Figure Two. In other words, it seems that target‐focus is a good debiasing technique (i.e., leads to more beneficial first offers) against being self‐anchored by one's own weak BATNA for low‐power candidates but not necessarily for low‐power employers. This pattern may stem from differences in the activation of the behavioral approach system versus behavioral inhibition system, which is also evident in the preference to move first (Magee, Galinsky, and Gruenfeld 2007). That is, it is possible that due to the innate power of employers, this intervention was not strong enough to change their behavior.

When looking at target‐focused negotiators, we sought to compare their first offer with the one they would have been given were they to move second. Importantly, we compared each target‐focused role to the entirety of their counterpart's role (both target and control conditions) because when entering a negotiation, negotiators do not know whether their counterparts are using debiasing techniques or not. Thus, these comparisons should yield more accurate and representative results. A comparison of target‐focused candidates' first offers to employers' (both target and control) first offers resulted in no significant difference. In this case, and like Study 1, letting employers move first would allow candidates to bargain and increase the salary from that point, instead of making the same offer themselves and allowing employers to bargain and decrease the salary. The opposite comparison—target‐focused employers versus candidates—showed similar results. In this case, letting candidates move first would allow employers to bargain and decrease the salary from that point instead of making the same offer themselves and allowing candidates to bargain and increase the salary. In other words, it seems that the first‐mover disadvantage of low‐power negotiators is strong and robust enough to resist the target‐focus technique, meaning our hypothesis (H3) was not affirmed.

Finally, we ran a similar model with reservation price as the dependent variable and both role and debiasing‐technique conditions as the independent variables. Results showed a significant effect of role on reservation price, such that candidates' reservation price (in thousands of pounds sterling per year) was significantly lower than that of employers' (M_candidate = 40.12, M_employer = 65.24). This pattern would be expected given that, as previously stated, reservation price is the lowest (highest) salary that job candidates (employers) would be willing to accept (offer). There was no significant effect of the debiasing technique on reservation price. Yet, when controlling for the role, the intervention significantly affected the candidate's but not the employer's condition. Whereas in the candidate role, target‐focus resulted in higher reservation prices (M_target = 41.82, M_control = 38.27), in the employer role, target‐focus did not result in lower reservation prices. In other words, target‐focus might affect the negotiation outcome for candidates by leading them to hold more favorable reservation prices (H3). Taken together, these results generally replicate the findings of Study 1 and point to the significant influence of power (specifically, low power) on first offers and consequently on solving the practitioner‐researcher paradox. While using the powerful debiasing technique of considering the opposite (one's target price) did make low‐power negotiators' (only candidates') first offer more favorable (in terms of magnitude), it did not reverse their first‐mover disadvantage.

Additionally, target‐focus did not affect the preference to move first versus second (most preferred to move second), which seems rational given that it does not eliminate the first‐mover disadvantage that low power brings. In summary, when low in power, negotiators should strive to make a counteroffer. If this option is unavailable to them, then using debiasing techniques should still result in a more favorable negotiation (compared to making the first offer without debiasing), yet not as good as when moving second.

While scholars regularly advise negotiators to take advantage of the anchoring and adjustment heuristic (Galinsky and Mussweiler 2001) and make the first offer, the current study echoes recent findings showing that this may not always be the best course of action. In this article, we examined whether negotiation power (in the form of BATNA) would affect the favorability of making the first offer and whether this effect could be mitigated using a popular anchor‐debiasing technique. Specifically, we hypothesized that: (1) low‐power negotiators would have received a more beneficial offer were they to make the counteroffer; (2) low‐power negotiators would have less beneficial reservation prices compared to those with high power; and (3) these detrimental effects of power would be attenuated if negotiators were to use an anchor‐debiasing technique.

Our hypotheses were largely confirmed across two studies using the same scenario—a salary negotiation between employer and candidate. In Study 1, candidates and employers were assigned to either low‐, medium‐, or high‐power conditions (i.e., in the form of best alternatives to the current negotiation). As preliminary support of our hypotheses, results showed that, within roles, lower power resulted in less favorable reservation prices (H2) and first offers (i.e., lower first offers for candidates and higher for employers). More importantly, when facing low‐power counterparts, low‐power negotiators made first offers that were less favorable than the offers they would have received were they to move second (H1). Although there was no significant difference in first‐offer magnitude when comparing the low‐power negotiators with their medium‐power counterparts, we argue that making the first offer would still have been less favorable. As shown in past research (Maaravi and Heller 2021), when both sides' first offers are equal, moving second allows negotiators to bargain and move the negotiation in their favor. In fact, only when facing strong counterparts—who were shown to make extreme first offers in their own favor (Schaerer et al. 2020)—should low‐power negotiators follow the advice set forth in most research and move first. See Figure Three for a flowchart summarizing recommended first‐offer decision as a consequence of power.

Study 2 examined whether this detrimental effect of low power on first‐offer favorability could be nullified using a popular anchor‐debiasing technique. While we used the same scenario as in Study 1, participants were assigned to only low‐power conditions since that comparison was where the detrimental effects were shown to be strongest. Half of the participants in each role used the anchor‐debiasing technique. Again, low‐power negotiators made first offers that were less favorable than the offers they would have received from their (low‐power) counterparts (H1). Additionally, our results showed that the anchor‐debiasing technique made candidates' (but not employers') first offers and reservation prices more self‐serving (in terms of their magnitude). More importantly, we found that when comparing low‐power negotiators using the debiasing technique with their low‐power counterparts (both debiasing technique and control conditions), there was no significant difference in their first offers. Following the rationale of previous research exploring first‐ versus second‐mover advantage (Maaravi and Heller 2021), we claim that when both sides make equal first offers, it is, in fact, better to move second. Moving second may allow negotiators to bargain and push the settlement price in their direction. Thus, even using the powerful anchor‐debiasing technique of “consider the opposite” does not change the findings of this research (and thus rejects H3): weak negotiators are better off moving second.

These findings and their implications add to the emerging literature regarding the circumstances in which making the first offer may be less beneficial (Maaravi, Pazy, and Ganzach 2014; Rosette, Kopelman, and Abbott 2014). Specifically, our studies show the effects of low power on negotiators' first offer and how counterparts' power affects the favorability of one's first offer. Our results suggest that low‐power negotiators with no knowledge of their counterparts' power should strive to make a counteroffer. In most cases, they will receive an equal or better offer if their counterparts make the first offer. Additionally, our results suggest that weak negotiators should adopt anchor‐debiasing techniques such as considering the opposite (i.e., focusing on one's target price) (Galinsky and Mussweiler 2001) because they mitigate the detrimental effects of low power on first offers. Nevertheless, this mitigation is not significant enough to render making the first offer more favorable; thus, low‐power negotiators applying anchor‐debiasing techniques should still strive to make the counteroffer.

Our studies are not without limitations, and future research is needed to further investigate the dynamics of low power and the first‐ versus second‐mover advantage. First, all negotiation scenarios in both studies culminated in the first offer made by participants and did not require them to undertake a complete negotiation with a final settlement price. The procedure was designed this way due to the exploratory nature of our research questions and hypotheses. Yet, it limits our ability to fully understand the effects of low power and anchor‐debiasing techniques on the negotiation's process and outcome. Relatedly, past research has investigated the impasse phenomenon in negotiation and its relationship to low power and first offers. One line of research found that extreme first offers offended low‐power negotiators and caused them to walk away (Schweinsberg et al. 2012). Another study suggests that if low‐power negotiators make unfavorable first offers (as our results indicate), they might be immediately accepted, resulting in feelings of regret and an eventual impasse (Galinsky et al. 2002). Investigating impasses requires a complete negotiation process, which is missing in our current research. However, since impasses are not the focus of this investigation, and in light of the ample literature detailing the causal relationship between first offers and negotiation outcomes, we predict that future studies examining complete negotiations will replicate our findings.

Second, our experimental procedure lacked clear incentives for the participants to achieve better outcomes in the negotiation. Furthermore, we used the same scenario for both our studies—the salary negotiation. Both factors may contribute to a lack of generalizability of our findings. Nevertheless, experimental paradigms lacking incentives are standard practice in negotiation research (e.g., Gunia et al. 2013) and the salary negotiation scenario is not only popular in the literature (e.g., Lipovsky 2006; Marks and Harold 2011), but is a common negotiation event that most individuals go through in their lifetime. Moreover, the job candidate versus recruiter research paradigm is highly relevant to the focus of the current studies, specifically to their theoretical framework: the power‐approach theory (Keltner, Gruenfeld, and Anderson 2003). Still, scholars have found that people interpret information differently depending on the negotiation context and their roles (e.g., Neale, Huber, and Northcraft 1987). For example, job negotiations are heavily scripted with normative expectations of how people should behave and thus may influence negotiation dynamics in the lab. Future studies should implement incentives for successful negotiations and, more importantly, use varied scenarios to test the generalizability of our findings.

Third, settlement prices and negotiation outcomes depend not only on negotiators' alternatives but also on the supply–demand balance of alternatives in the market (i.e., the number of candidates or positions available). In line with past research, we manipulated power by controlling the quality of the alternatives, not their quantity (Magee, Galinsky, and Gruenfeld 2007). Therefore, we propose that future research explore the effects of market conditions (e.g., supply–demand) on power and their impact on first‐offer favorability.

Finally, we used only one of the multiple anchor‐debiasing techniques—focusing on one's target price—to investigate their efficacy at mitigating the detrimental effects of low power on first offers. While this technique is representative of the “consider the opposite” class of debiasing strategies (Galinsky and Mussweiler 2001), future studies should test more techniques from within the same category or others. Some examples of other anchor‐debiasing techniques include considering why anchors are inappropriate, focusing on counterparts' alternatives, framing the negotiation differently (Lord, Lepper, and Preston 1984), or debiasing via psychoeducation, affective debiasing, mindfulness, and other methods (Croskerry, Singhal, and Mamede 2013).

To conclude, our studies' purpose was to explore one of the factors affecting the favorability of making the first offer in negotiations while possibly shedding light on the subtleties of the researcher‐practitioner paradox. We found that low‐power negotiators should strive to make counteroffers (except when faced with high‐power counterparts) even if they use anchor‐debiasing techniques. These findings echo the emerging literature on first‐mover disadvantage. They offer practical guidelines to negotiation practitioners and make theoretical contributions to research.

The authors declare no potential conflicts of interest with respect to the research, authorship, or publication of this article.

This study was supported by the Israel Science Foundation grant 1916/19.

These studies were approved by the Institutional Review Board of the Adelson School of Entrepreneurship, Reichman University (IDC), Herzliya, Israel.

The data that support the findings of this study are openly available in the Negotiation Data Repository (NDR) at https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/IFOHGP.