The Spanish Competition Authority (CNMC) plans to support judges by developing guidelines on quantifying damages caused by competition law infringements. The CNMC held a public consultation on the draft version of these guidelines. The contributions received, including CDC’s, are published on the website of the CNMC. While the issues we picked up were partially addressed by the draft guidelines, we believe they merit further elaboration. In response to the public consultation, we arranged our remarks under six headings.
- Datasets: it is not just volume and veracity but also variety that matters.
We suggest that the CNMC do not identify the size of the dataset as an overly strict condition to be considered by judges.[1] In particular, a smaller dataset that contains information on purchases from multiple cartel infringers may be preferable to a larger dataset of a single infringer. Greater variety in data provides a clearer distinction of the infringement effects from factors that may affect firms and transactions differently.
- Availability of data: the absence of documentary evidence is not evidence of the lack of cartel effects.
The draft guidelines mention that data must be reliable, transparent, complete, and verifiable.[2] But the guidelines also mention that ‘most datasets are incomplete or otherwise imperfect.’[3] Finding a balance between desire and reality is difficult. The ideal situation of data completeness formulated by the draft guidelines should be read in the light of the EU law principle of effectiveness, which stipulates that the requirements for obtaining full compensation may not make exercising this right practically impossible or excessively difficult.[4] Therefore, we suggest that the CNMC moderate the statements about data completeness and discuss the use of non-transaction data and evidence when transaction-level data is unavailable.
- Sources of cost data: issues with firm-level cost data suggest that public data on costs is more reliable.
The draft guidelines indicate that cost is one of the most used explanatory variables of prices.[5] Even so, one should be aware that the costs of the infringer(s) has limited explanatory power for prices during the infringement. The draft guidelines correctly point out the issues with (i) X-inefficiency and (ii) the use of the infringer’s accounting figures to explain economic costs.[6] Therefore, we suggest that the CNMC clarify – in the final version of the guidelines – that publicly available cost proxies are more suitable for analysis. And if used, the cost data obtained from the infringers should be scrutinized.
- Interpreting regressions: statistical measures are no substitute for common sense.
Annex 3 of the draft guidelines discusses general statistical measures of econometric models, like statistical significance and goodness of fit. We find that there is too much emphasis on the reliability of these measures in verifying a model. For example, cartels should not escape paying compensation just because the cartel effect is statistically insignificant due to things like having few data points. As for the goodness of fit measures, they often are misleading, as with time-series models. Further issues with general descriptive statistics are outlined in Section 2.4. of the CDC submission. In general, we argue that statistical measures should not replace “common sense” about how the market works.
- Cartel dating: infringement period may differ from the period affected by the cartel.
The draft guidelines recognize that the effects of the cartel may occur at different times rather than strictly within the official start/end date of the cartel (as defined in the infringement decision of a competition authority).[7] Because determining the beginning/end of the cartel effects can seriously affect the estimated overcharge, we suggest that the CNMC add a separate section on cartel dating. This will highlight the possibility of higher than competitive prices even after the official end date of the infringement, the impact of long-term contracts signed during the cartel period, the time needed to rebuild capacity in the market that has been reduced due to the cartel activity, etc. A return to free competition will be more difficult and protracted the longer the cartel effects continue.[8]
- Real vs. nominal prices, inflation, and interest: models do not have to work with inflation-corrected prices, and interest is an integral part of full compensation.
We provide some observations to the questions on interest, inflation, and discounting because parts of the draft guidelines could be understood as stating that models should always work with inflation corrected (“real”) prices.[9] If a model uses inflation corrected prices and other monetary explanatory variables, close attention must be paid to the choice of the inflation/deflation rate(s) or discount rate(s) used (the draft guide identifies this issue[10]). In addition, transforming all monetary variables to inflation corrected values can introduce additional variation that mask collusive patterns. Furthermore, models that generate nominal overcharge estimates allow a clearer interpretation of the results.
Finally, we suggest adding a chapter on the calculation of legal interest on (nominal) damages resulting from infringements of competition law. A study coordinated by CDC and the European University Institute on the interest in damages resulting from competition law infringements also contains a country report for Spain.
[1] Draft guidelines, p. 18, 69.
[2] Idem, p. 18.
[3] Idem, p. 6., fn. 3 (CDC translation of the Spanish original).
[4] European Commission (2013), Quantifying Harm in Actions for Damages Based on Breaches of Article 101 or 102 of the Treaty on the Functioning of the European Union, Commission Staff Working Document, 11.6.2013, SWD (2013) 205, para. 17.
[5] Draft guidelines, p. 17.
[6] Idem, p. 29.
[7] Idem, Section II.2.a).
[8] See CJEU, judgment of 8 December 2011, Case C-389/10 P, KME Germany & Ors. v Commission, para. 75.
[9] Draft guidelines, p. 24., 27., 31., 36., 40.
[10] Idem, pp. 24-25, 28.
