Preventing innovative cooperations: the legal exemptions unintended side effect

Abstract

In 2004, European competition law had been considerable changed by the introduction of the new Council Regulation No. 1/2003. One of the major renewals was the replacement of the centralized notification system for inter-company cooperations in favor of a so-called legal exemption system. We analyze the implications of this reform and its arising uncertainty on the agreements firms implement, especially on innovative agreements like vertical R&D agreements. By means of a decision theoretic approach, we show that the law’s intention to reduce the incentive to establish illegal cartels will be reached but innovating cooperations might be prevented. To avoid this unintended side effect, fines but not the monitoring activities should be increased.

Appendices

Appendix 1

See Table 1.

Table 1 Comparison of application of Article 81—old (Reg. No. 17/62) and new (Reg. No. 1/2003)

Appendix 2

The expectation value for the innovative path is given by:

$$ E(I) = \pi_{II} + s(\pi_{Ic} - \pi_{II} ) - c[p_{Ic} s + (1 - s)p_{II} ] - p_{II} f(R_{II} )q(1 - s) - p_{Ic} f(R_{Ic} )(1 - q)s. $$

(23)

The second term states the defending costs, which depend on the probability of being detected and the probability that the company received wrong information ex-ante. This term will always reduce the maximum possible expectation value, which is determined by the first two terms, as $c[p_{Ic} s + (1 - s)p_{II} ]$cannot become negative.

As c is per definition a positive value, the latter term can only become negative if:

$$ p_{Ic} s < - (1 - s)p_{II} $$

(24)

$$ p_{Ic} s < ( - 1 + s)p_{II} . $$

(25)

Since $(1 - s)p_{II} > 0$it implies that

$$ - (1 - s)p_{II} < 0. $$

Hence, $p_{Ic} s < 0$.

Because 0 ≤ q ≤ 1 and 0 ≤ s ≤ 1, the term (1 − s)p _II on the right hand side cannot become negative (the lowest value for the right hand side is given by zero), such that the total value of the right hand side can never reach a positive value. Consequently, if the left hand side has to approach a lower value than the right hand side, this implies that p _Ic s < 0, which is precluded by definition.

Appendix 3

Again, for a cartel to be preferable, it must hold that E(C) > E(I). Assuming that R _cc  > R _Ic  > R _II and π _cc  > π _Ic  > π _II , this yields to:

$$ \begin{gathered} \pi_{cc} - p_{cc} [c + f(R_{cc} )(1 - q)] > (1 - s)[\pi_{II} - p_{II} [c + f(R_{II} )q] \hfill \\ +\,s[\pi_{Ic} - p_{Ic} [c + f(R_{Ic} )(1 - q)]] \hfill \\ \end{gathered} $$

(26)

Assuming further that the probability of detection, as well as the fine depend linear on excess profits (π).³² Define: $\Updelta_{1} \pi_{cc} = \pi_{Ic}$ and Δ₂ π _cc  = π _II , with Δ₁, Δ₂ < 1 which yields also $\Updelta_{1} R_{cc} = R_{Ic}$ and $\Updelta_{2} R_{cc} = R_{II}$, and, consequently, implies that$\Updelta_{1} p_{cc} = p_{Ic}$, $\Updelta_{2} p_{cc} = p_{II}$. With respect to these conditions, inequality (26) can be rewritten as:

$$ \begin{gathered} \pi - pc + pf - pfq > \Updelta_{2} \pi - \Updelta_{2} pc - \Updelta_{2}^{2} pfq + \pi s(\Updelta_{1} - \Updelta_{2} ) + spc(\Updelta_{2} - \Updelta_{1} ) \hfill \\ +\,pfqs(\Updelta_{2}^{2} + \Updelta_{1}^{2} ) - pfs\Updelta_{1}^{2} \hfill \\ \end{gathered} $$

(27)

Solving for the probability q that AA makes a mistake, the following solutions can be distinguished:

$$ \begin{gathered} {\text{Solution}}\;1: \hfill \\ q < {\frac{{(\pi - pc + pf - \Updelta_{2} \pi + \Updelta_{2} pc - \pi s\Updelta_{1} + \pi s\Updelta_{2} - spc\Updelta_{2} + spc\Updelta_{1} + pfs\Updelta_{1}^{2} )}}{{pf(1 - \Updelta_{2}^{2} + s\Updelta_{2}^{2} + s + \Updelta_{1}^{2} )}}}, \hfill \\ {\text{if}} \hfill \\ 0 < pf - \Updelta_{2}^{2} pf + pfs(\Updelta_{2}^{2} + \Updelta_{1}^{2} ) \hfill \\ {\text{Solution}}\;2: \hfill \\ q > {\frac{{(\pi - pc + pf - \Updelta_{2} \pi + \Updelta_{2} pc - \pi s\Updelta_{1} + \pi s\Updelta_{2} - spc\Updelta_{2} + spc\Updelta_{1} + pfs\Updelta_{1}^{2} )}}{{pf(1 - \Updelta_{2}^{2} + s\Updelta_{2}^{2} + s + \Updelta_{1}^{2} )}}}, \hfill \\ {\text{if}} \hfill \\ 0 > pf - \Updelta_{2}^{2} pf + pfs(\Updelta_{2}^{2} + \Updelta_{1}^{2} ). \hfill \\ \end{gathered} $$

Solution 1 states that if the probability that AA slips up (q) is smaller than the ratio on the right hand side and if the additional constraint, which is given by $0 < pf - \Updelta_{2}^{2} pf + pfs(\Updelta_{2}^{2} + \Updelta_{1}^{2} )$holds, a cartel is preferred. Furthermore, a cartel overrules an innovative agreement if q exceeds the right hand side and if $0 > pf - \Updelta_{2}^{2} pf + pfs(\Updelta_{2}^{2} + \Updelta_{1}^{2} )$.

Therefore, an additional statement about the parameters in the constraint has to be added.

Considering $0 < pf - \Updelta_{2}^{2} pf + pfs(\Updelta_{2}^{2} + \Updelta_{1}^{2} )$ and solving for the probability of a firm’s error (s)—which highlights the probability that the firm assumes ex-ante that its agreement is innovative but which emphasizes ex-post to be a cartel—this has to fulfill:${\frac{{\Updelta_{2}^{2} - 1}}{{(\Updelta_{2}^{2} + \Updelta_{1}^{2} )}}} < s.$

As $\Updelta_{2}^{2}$<1, the ratio on the left hand side becomes negative. Hence, solution 1 is valid for all probabilities s.

From solution 2, it follows:

$$ {\frac{{\Updelta_{2}^{2} - 1}}{{(\Updelta_{2}^{2} + \Updelta_{1}^{2} )}}} > s, $$

which can never hold as s determines a probability and, thus, has to lie in the range of 0 ≤ s ≤ 1. Therefore, a positive value for the left hand side can only be reached if Δ ²₂  = 1, which is excluded by assumption.³³

Concluding, solution one gives the necessary parameter constellation such that the expectation value for the cartel exceeds the innovative agreement. The same relation holds for exponential dependence of fees and detection probabilities on excess profits.