Satisficing or maximizing in public–private partnerships? A laboratory experiment on strategic bargaining

ABSTRACT

Cross-sectoral strategic negotiation is a key challenge in PPPs. Based on framing and game theory, we investigate the effect of sectoral agency, affect, and bargaining domain on sectoral agents’ bargaining behaviour in a PPP renegotiation scenario. Results confirm that public agents are more likely to bargain for satisfactory, ‘good enough’ contracts than private agents, who maximize their utility. This difference is stronger in the loss vis-a-vis the gain domain. These experimental findings advance our understanding of psychological mechanisms underlying cross-sectoral negotiations, suggesting that public managers and policy-makers account for partners’ dissimilar bargaining logics to prevent asymmetric loss socialization in PPPs.

KEYWORDS:

• Negotiation
• strategic bargaining behaviour
• bounded rationality
• public-private partnership (PPP)
• laboratory experiment

Disclosure statement

No potential conflict of interest was reported by the authors.

Supplementary material

Supplemental data for this article can be accessed at https://doi.org/10.1080/14719037.2021.2013072

Notes

1. Note that although some authors have pointed out that there are semantic differences between the terms ‘bargaining’ and ‘negotiation’ – cf., Flanders (1968) and Lewicki, Saunders, and Barry (2020), – there is little to no agreement on the significance of those differences. Since the analysis presented in the current study would not suffer from the omission of this distinction, the terms ‘negotiation’ and ‘bargaining’ are used interchangeably.

2. The game mechanics were pretested extensively with focus group laboratory sessions to maximize stimulus realism and minimize response bias. Pretesting revealed that increasing the bargaining phase length (>45 seconds per round) or increasing the number of negotiation rounds did not increase the likelihood of reaching agreement but substantially increased response fatigue, prolonging the whole experimental procedure without additional explanatory value. Following best practice recommendations for designing z-Tree experiments, we opted for the parsimonious setup with two times 15-rounds of 45 seconds each as an optimal balance between creating sufficient bargaining amount variation in a framing treatment design and inhibiting response fatigue due to experiment duration (Katok 2019).

3. The negotiation amounts and experimental mechanics were designed following best practice recommendations regarding the optimal duration and relative size of incentives in z-Tree bargaining experiments (see, e.g. Güth 1995; Crawford 1997; Fischbacher 2007; Katok 2019); see also online appendix A, which explains the theoretical premises, provides an overview of the flow logic (A.1), and the contents of the experiment. The magnitude of rewards and their systematic inter-item variation follows the item magnitude structure of Madden, Petry, and Johnson’s (2009) validated Probability Discounting Questionnaire, an experimental procedure to reveal implicit individuals’ risk preferences by employing rapid economic trade-offs between batteries of probabilistic and secure rewards. This procedure has been confirmed as a highly reliable and structurally robust measure to capture the non-linearity of human risk behaviour (Kahneman and Tversky 1979), and across various numbers of magnitudes, see e.g., Myerson et al. (2003) and Green et al. (2013) for item magnitude structure robustness, and Weißmüller (2021) for a recent example of its application in the context of public management research.

4. The numeric amounts of losses were varied and randomized exactly like the amounts of gains to achieve task balance and to increase internal consistency and measurement validity, see online appendix A.1 for more detail on the game mechanics.

5. The experiment uses the equal split in the loss domain because it was set as the explicit default in the treatment and to increase scenario realism: In cameralistic accounting traditions of public administrative accounting – such as Germany – failure to successfully negotiate the case of unexpected gains (or parts thereof) often results in the expiration of these opportunities (or parts thereof) whereas failure to successfully negotiate the distribution of unexpected losses does not mean that they will disappear.

6. Võ et al.’s (2009) BAWL-r inventory is an extensive list of several thousand common German words which were systematically and empirically tested for their emotional valence, emotional arousal, and imageability on a metric scale in order to be used as psycholinguistic indexes and treatment stimuli in psychological and behavioural experiments. The BAWL-r inventory can also be used to associate open semantic responses with BAWL-r’s values by qualitative coding and quantifying these semantic responses across subjects to create mean implicit affective scores.

7. This scale originally comprises eight items of statistical word problems of varying complexity. The last and most complex item was omitted due to expected questionnaire length-related response fatigue.

Additional information

Funding

Nothing to declare.

Notes on contributors

Kristina S. Weißmüller

Kristina S. Weißmüller is postdoctoral researcher and lecturer at the KPM Center for Public Management, Universität Bern (Switzerland), specialized in behavioural and experimental research methods. Her research focuses on public sector corruption and the psychological effects of ‘publicness’ on behaviour, e.g., regarding strategic choice, motivation, leadership, and negotiation in PPPs.

Robin Bouwman

Robin Bouwman is assistant professor of public management at Erasmus University Rotterdam (The Netherlands). His research focusses on public- and private sector negotiations, individual and group decision making and behaviour.

Rick Vogel

Rick Vogel is full professor of public management at Universität Hamburg (Germany). His current research interests are public sector leadership, cross-sector partnerships, and institutional change in the public sector.