Mergers, Common Ownership, and Anti-Competitive Behaviour: Experimental Evidence
A recent literature on the effects of common ownership of product market rivals on market outcomes has triggered far-reaching policy proposals to reform the basic workings of the asset management industry. However, given limitations of theory and measurement challenges, questions remain about the extent to which the existing results should be interpreted as causal effects, which corporate mechanisms are at play, and which policy interventions are likely to be effective.
We study these questions in a controlled laboratory environment, in which the experimenter can vary ownership and thus infer causal effects with greater confidence. Despite the inherent limitations of experiments, policymakers may wish to consult these results to assess the likely effectiveness and possible side effects of planned large-scale interventions. In our experiment, there are two firms, managed by two managers and owned by two shareholders. Managers choose production quantities. The price in the market depends on the total production of both firms. Subjects play the experiment for multiple rounds.
Shareholders initially own 100% of one of the firms each (divided ownership). During the experiment, they have the option to exchange half of their shares so that they each own 50% of both firms. The shareholders' incentives change with the ownership. If they each own shares in only one firm, they are best off when their manager produces the Cournot duopoly quantity. If they each own shares in both firms, they are best off when their manager produces half of the monopoly quantity. Overall shareholder profits are higher if both managers produce half of the monopoly quantity than if both produce the Cournot duopoly quantity.
Shareholders also have the opportunity to incentivise managers. The incentive options differ between the two treatments we run. In treatment 1, shareholders can choose incentive contracts before the managers make their decision. Shareholders have the option to choose flat incentives, which should have no impact on manager behaviour Moreover, they can commit to paying managers a bonus, if the profit they make is at least as high as that of the competing firm (relative profit incentives). This incentivizes managers to compete aggressively, Finally, they can commit to paying managers a bonus if their firm's margin is at least as high as that of the competing firm, which incentivizes managers to compete less aggressively. In treatment 2, shareholders can voluntarily pay a bonus after observing the manager's choice, which gives them the option to reward desirable behaviour by the managers which should induce those to keep showing this behaviour. Subjects play the experiment for 15 rounds.
A significant share of managers, albeit less than half, produces the Cournot duopoly quantity (26) and hence behaves anti-competitively. Moreover, around 40% of the shareholders vote to exchange shares. In our first analysis, we study if managers respond to incentives set by the shareholders. A regression shows that relative profit incentives do not affect the managers' decisions. This could be explained by the fact that managers have a preference for producing the Cournot duopoly quantity in the absence of incentives so that further incentives to do so have no effect. Relative margin incentives on the other hand make it significantly more likely that the manager produces half of the monopoly quantity. A larger bonus makes it significantly less likely that the manager changes their production.
Next, we study if managers set the correct incentives given their diversification status. We find a negative link between the incentive dummy and the diversification dummy, which indicates that shareholders set more anticompetitive incentives if they are diversified. We also find a strong positive link between the bonus and the interaction term, which indicates that shareholders pay a larger bonus if the manager produced half of the monopoly quantity when the shareholder was diversified.
Finally, we study if shareholders choose to diversify in order to be able to set anticompetitive incentives and profit from reduced competition. A regression analysis shows that shareholders who voted to diversify in the last round are more likely to do so again, suggesting a somewhat stable preference for diversification. However, if an ownership exchange actually happened (i.e. if both shareholders agreed to an ownership exchange in a given round) this makes shareholders less likely to vote to exchange ownership again. Hence, we do not see that shareholders learn about the advantages of diversification, once they have experienced it. This suggests that the markets will not converge towards diversification in the long run. Our study also includes a comprehension check, which subjects can pass either on their first attempt. We use this information as a proxy for how well subjects understand the experiment and we do not find that subjects with a better understanding of the experiment are significantly more likely to diversify.
In summary, we find that managers respond to incentives and that shareholders correctly use incentives to influence managers to produce the quantity that is in their best interest given their diversification status. However, we do not find that shareholders anticipate this and chose to strategically diversify. Our markets do not converge towards diversification. In future research, we plan to study if this is driven by the default, which was no diversification in both treatments, by running sessions in which shareholders start out diversified and can choose to become concentrated owners.