Analyzing the dynamics in multilateral negotiations

Abstract

The purpose of this article is to examine phase-differences in the patterns of actors involved in decision making. Two phases are distinguished, the negotiating phase, and the decision-making phase. During the first phase the agent’s primary goal is to influence the policy position of other agents, whereas during the second phase the agent’s goal is to find a solution (especially if there is a credible threat of a worse outcome). Convincing others is easier in a bilateral setting than in a multilateral one, and thus the decision-making phase is expected to result in an increase in multilateral interactions. To reveal the participation pattern of agents, we discuss methods for quantitative analyses of complex negotiations. A dynamic analysis of participation in multilateral negotiations gives insight into the continuity and change of participation during a negotiating process. The process of determining prices for specialized medical care in The Netherlands is used as a case study. The quantitative methods seem to be useful for analyzing such a complex case. Following our hypothesis, in the case study multilateral contacts are shown to be more frequent over time as compared to bilateral negotiations.

Introduction

An important feature of complex decision-making processes in modern administration is that, very often, a great variety of agents participate in them. Negotiations between agents take place over extended periods of time and are conducted both multilaterally and bilaterally. Two phases are commonly distinguished: the negotiating phase where actors try to influence each other’s preferred outcomes, and the decision-making phase where the actual decision is taken (Stokman and Van den Bos, 1992). During the negotiating phase, the agent’s active participation depends on the saliency of the issue at stake, the difference between the expected and preferred outcome, and the ability to influence the decision making. According to Achterkamp (1999), it is not only the agent’s potential power (e.g. voting rights) but primarily his effort in affecting the collective outcome that determines how much he influences the final decision. Some agents participate because of formal procedures; others are involved because of informal procedures or because they deliberately seek to influence the outcome of the negotiating process irrespective of formal procedures.

The participation of a great variety of agents in complex negotiations does not imply that they all participate during the entire negotiating process. While some agents participate all the time, others may only participate temporarily or intermittently. Different combinations of agents may also successively participate in the negotiating process. Participation patterns may therefore change over time.

In network analysis interaction may be a source of influence because of the flow of information between actors (Marsden and Friedkin, 1994). It is easier to convince others of your point of view in a dyadic interaction situation than in a multilateral setting with possibly many conflicting interests. Coalition building does not solve this problem because collective decisions are often indivisible making the exchange of sets of positions problematic (Knoke et al., 1996). Without a multi-issue agenda agents are more likely to behave as competitors than as partners forming coalitions. The long-term perspective of a perfectly rational agent would suggest a stable negotiating strategy over time (Macy and Flache, 1995). However, we will argue a discontinuity in the agents’ behavior as a result of a shift in the saliency of goals.

In the negotiating phase, the agent’s primary goal is to influence the policy position of the other agents. The focus on influencing other agents’ policy positions during the negotiating phase relegates the salience of other goals to the background (see Lindenberg and Frey, 1993, Wittek, 1999). It structures the negotiation process by the agent’s overriding goal of maximizing personal profits (i.e. competitive gain frame) as opposed to distributing profits according to the principle of equity (i.e. cooperative frame). There is some evidence that in an experimental setting individuals behave very differently depending on whether they are instructed to be competitive or cooperative (see Carnevale and Lawler, 1986). In the decision-making phase the agent’s goal shifts to finding a solution, especially if there is a credible threat of a worse outcome. An example of such a threat is an intervention by an external broker. The agent may expect the outcome to be even farther removed from the outcome he prefers. If the decision-making power is controlled by an external agent, entering the decision-making phase is likely to switch the agent’s frame from a gain frame (i.e. pushing one’s policy position) to a cooperative frame (i.e. striving for an equitable solution).

There is yet another reason why we might expect a frame switch in the decision-making phase of the negotiations. The goal of gain seeking may be tempered by relational concerns regarding the future (Axelrod, 1984). Anticipated future interactions in other decision-making processes stress the importance of keeping the relationship going as a goal in itself. As a result of the attempts to influence other agents’ policy positions we hypothesize that agents are more likely to engage in bilateral interactions in the negotiating phase, whereas in the decision-making phase they are more likely to interact multilaterally to find an equitable solution.

Complex negotiations have often been analyzed by means of qualitative methods, however quantitative methods can help better understand the course and the outcomes of negotiations (see for example Payne, 1999, Rojer, 1999). As discussed by Willer and Willer (2000) most network studies neglect the dynamics of networks. Moreover, studies on negotiations do not take framing effects into account. We are particularly interested in a dynamic analysis of participation in multilateral negotiations. Our methods should give insight into the continuity and change of participation during a negotiating process. By contrast, a static analysis does not yield such an insight because static analysis does not take “time” into consideration. It only provides information on the varieties of agents and participation patterns in the negotiating process.

The aim of this article is two-fold. First, to study phase-differences in the patterns of actors involved in the decision-making process, and second to discuss some methods for quantitative analyses of complex negotiations. To illustrate phase-differences in participation patterns we use a case from Lieverdink (1999) who studied the negotiating process on tariffs (prices) of medical services by medical specialists in The Netherlands (also see Lieverdink and Maarse, 1995). Section 2 offers a brief description of the negotiating process under study. 3 Quantitative modeling, 4 Dynamic analysis give an overview of the participation patterns by our quantitative methods to model complex negotiations. In Section 5 we discuss limitations and further applications of our findings.