Jenny Säve-Söderbergh, SOFI, Stockholms universitet'

2010-2016

The aim of the project and changes during the course of the project
The aim of the project was to study the role of wage requests at the time of employment, a research area within which the empirical literature is limited. Our aim was thus to illuminate the role of wage requests using different empirical methods: field data over wage requests, experimental data and through data from job ads in Platsbanken. The project further aimed to study how wage requests depend on risk preferences and how wage requests relate to the gender pay gap.

Apart from the below described change, the project has been executed as planned and we have gathered unique data on wage requests and risk preferences using two of the proposed methods: field data and experimental data. The field data contains new and unique survey data on wage requests for over 35 000 respondents between 1999 and 2012 that are based on yearly surveys executed by Jusek. Of great importance is that beyond wage requests the data contain information on why the respondent did not negotiate and attitudes towards the negotiation. We have also executed a comprehensive research experiment in several sessions with a large number of participants. In the experiment we had participants elicit their wage requests under different contexts, their risk preferences using three independent measures and survey questions on their willingness to negotiate, their willingness to compete, economic background and reservation wages.

Regarding the third method, the work has to some extent changed during the course of the project. Our initial ambition was to test and explain employers' use of wage requests with data from job ads in Platsbanken against a theoretical model. A smaller dataset was also gathered but the material did not provide the opportunity to do a meaningful analysis. The experiment was also more time consuming than anticipated, partly as a consequences of attaining an adequate number of observations, partly as a consequences of testing alterations in the design.

Due to the above reasons, the project has been extended to include two related research questions on risk preferences and social context. The first research question regards if, for example, wages or wage requests, can be explained by contextual factors like in which social environment within which the decision is taken. To answer this question a second field data was gathered which is based on 10-11 year old children and adults' behavior in the game-shows Jeopardy and Junior Jeopardy. From this we can study how the willingness to take risk and performance changes depending upon a random assignment into competing against men or women, and for children, against boys and girls. The second research question that the project has been extended with is whether we find an expected positive association between female wages and working directly for a female as opposed to male manager. The positive association is based on the argument that female managers will be particularly beneficial for female employees through, e.g., role modeling, mentoring or providing other incentives to enhance female productivity. Our study does no find this. Instead we find a negative association yet only among female employees working for a lower-level manager. The study discusses possible explanations like differences in bargaining power or resource division between manager genders in different positions in an organization.

The three most important results
The research has contributed with several new and important findings to the research field. The results also provide important insights to the design of policy measures. Below, I summarize the three most important results.
1. Based on the field data, the research shows that wage requests are not uncommon as 40 percent of the applicants state wage requests to their prospective employer. The analysis also shows that there is a selection into the group that state wage requests. Those who did not state a wage request are younger, more often in some educational groups, apply for jobs somewhat more in the public sector and are found in jobs with lower starting salaries. Moreover, the analysis shows that the reason for why the applicant have not stated a wage request is largely because the applicant believed the wage was fixed, or because the applicant did not care about the wage, alternatively was happy with the wage offered. Very few indicate psychological reasons like not daring as to why they did not state a wage request.
2. Based on the field data, the research shows that there are important and not previously shown gender differences in the propensity to state a wage request, the level of wage requests and in the negotiation outcome conditional on the wage request. Firstly, controlling for observables between those stating wage requests and those who do not, women are more prone to state a wage request compared to men. Previous literature has shown that women are less or equally likely to negotiate. There are also no large differences in the reason to why they did not state wage requests, not in attitudes or knowledge, commonly suggested as important explanations to a gender gap. Secondly, mean wage requests are lower among women compared to men and the difference increases throughout the distribution. Third, even though women request lower salaries are women more likely to attain a starting salary below their requested salary. The results are robust to controls for many typical differences by gender like in the choice of field of major, experience, private or public sector work, but also for differences in how men and women found the jobs, working hours, choice of industry and the share of unemployed within the field of major.
3. A third new and important result is that, despite the gender composition of opponents being random and should not matter for the behavior, both women and girls' behavior is susceptible to the social context, in this case the gender composition, within which the decision is taken. While women take lower risk, girls perform worse and employ inferior strategies when they are randomly assigned to compete in a male-dominated context. These results can neither be explained by any strategic reason nor by any game-specific behavior. This is noteworthy in the light of the large economic incentives in the game. The research further shows that the gender difference commonly found among adults - women taking less risk than males do - is not found among the children despite the groups having the same circumstances for their decision-making. The results are robust to controls for the contestants' performance, relative performance and several game-specific aspects. Thus, one explanation to a gender wage gap, and to women's lower wage requests, could be that women already as girls change their behavior depending upon the gender composition of the environment in which they act.

The most important publications, dissemination of results and publication strategy
The project has generated a number of studies. The first two consider specifically the field data on wage requests "Gender differences in salary negotiations - evidence from the field" and "Attitudes, ambiguity and social costs in salary negotiations: an evaluation using field data". These studies are unique contributions to the field and will within short be published as CESIfo working papers, and then be submitted to high ranking top field journals. The third study "Children do not behave like adults: gender gaps in performance and risk-taking in the high-stakes game shows Jeopardy and Junior Jeopardy" is already a CESIfo working paper. This study is also revise and resubmit in Economic Journal. The fourth study "Immediate manager gender and wages of female employees: the importance of manager position" is a SSRN working paper and is under review in Labour Economics. The publication strategy for all studies is to publish in internationally well-known journals to ensure scientific evaluation, quality and open access. The studies have also been presented at conferences e.g. European Economic Association, 2011, European Association of Labor Economics, 2011, 2013, European Society for Population Economics 2013, at seminars and within two interdisciplinary networks.

Has the project generated new research questions?
The project has generated many new research questions. Importantly, the comprehensive material gathered through the research experiment can be used to answer many interesting and related research questions. The advantage with this experiment is that we have been able to isolate the negotiation behavior from other confounding factors while also allowing for controlling for many otherwise unobserved differences between men and women that could affect the negotiation decisions. Moreover, against the background of this project's third result we also aim to study how the wage requests varies in the experiment depending upon whether the participant competes against men or women. The survey material can also be used to answer a number of questions related to differences in negotiation behavior.

Publications

Publikationer:
Säve-Söderbergh, J. (2014) “Children do not behave like adults:  Gender gaps in performance and risk taking within a  random social context in the high-stakes game shows Jeopardy and Junior Jeopardy”, joint with
Sjögren-Lindquist, G., CESIFO Working paper 4595, Revise & Resubmit in  Economic Journal.

Säve-Söderbergh, J. (2014) ”Female Wages and the Gender of the Immediate Manager”, joint with  Halldén, K. och Å. Rosén, SSRN Working Paper, under review in  Labour Economics.

Säve-Söderbergh, J. (2015) “Gender Differences in Salary Negotiations: Evidence from the Field”, Working Paper.  

Säve-Söderbergh, J. (2015) ”Gender gaps in Attitudes, Ambiguity and Social Costs in Salary Negotations: An Evaluation Using Field Data”, Working Paper.