Do Contract Remedies Affect Efficient Renegotiation? An Experiment

2024-06-05

2024-06-30

Share of pairs who do not enter the renegotiation phase;
Share of pairs who are unable to reach an agreement during the renegotiation;
Division of the surplus after the shock.

The experiment comprises two parts. In part 1, we consider a modified version of the “Contract-Breach Game” introduced by Bigoni et al. (2017), where a market shock may occur after the parties agreed to sign a binding contract. After learning about the occurrence of a shock, parties can decide to renegotiate the contractual terms under alternative breach remedies. The key, novel aspect of our design is the comparison between two breach remedies, which allows us to compare a damages remedy (Damage treatment) versus a specific performance remedy (Specific Performance treatment). This modified “Contract-Breach Game” à la Bigoni et al. (2017) allows us to study how frequently an efficient agreement to breach is reached under specific performance, and how frequently an efficient agreement to perform the contract is reached under damages. Furthermore, this game allows us to analyze whether the compensation paid by the promisor to the promisee differ between alternative breach remedies, and between alternative reasons to breach. In part 2, all participants take part in a Dictator Game.

Done in office by a computer

Randomization unit: experimental sessions

Yes

48 6-person matching groups

288 experimental subjects

144 experimental subjects (24 6-person matching groups) per treatment arm

We will test each of our three main hypotheses by means of a two-prong approach. The first approach is more conservative, and it is based on non-parametric two-tailed Wilcoxon Mann-Whitney tests on the difference of the distribution of the outcomes of interest, between treatments. The unit of observation here is the average outcome per matching group, across all rounds, thus we have 48 observations in total. This allows us to detect an effect of 0.85 of a s.d. with α=0.05 and (1-β)=0.8. Based on the data collected by Bigoni et al. (2017), this corresponds to a difference of 3 p.p. in the share of pairs who enter renegotiation when an exogenous shock takes place (HP1), a difference of 5 p.p. in the rate of success of renegotiation (HP2), and a difference of 0.9 units (out of 20) in the surplus allocated to the Seller in the renegotiation stage, when renegotiation is successful The second approach relies on the panel structure of our dataset and allows us to control for the learning dynamics across periods. Based on the simulations obtained via the Stata pc_simulate package (Burlig et al., 2020), this allows us to detect a variation of 4 p.p. in the share of pairs who enter renegotiation when an exogenous shock takes place (HP1), a difference of 6 p.p. in the rate of success of renegotiation (HP2), and a difference of 0.9 units (out of 20) in the surplus allocated to the Seller in the renegotiation stage, when renegotiation is successful, with α=0.05 and (1-β)=0.8.