Division Sensitivity analysis

In the previous empirical analysis, the MDD contract was defined as the contract in the dataset, for a given year, that presented the highest price difference in absolute terms (RON/MWh) between the observed price and its corresponding estimated price.

In order to check the robustness of the results, a sensitivity analysis was performed. The same analysis was now performed, but defining the MDD contract as the contract with the largest difference between the estimated price and its corresponding observed price, expressed as a percentage of the estimated price. In the sensitivity analysis, the MDD contract was now defined as the contract with the largest difference between the estimated price and its corresponding observed price, expressed as a percentage of the estimated price. While this analysis should not affect significantly the selection of the MDD contract, it could affect the contracts that are selected. The results are shown in Tables 6-9 below. They confirm the validity of the results from the main analysis.

The prices of electricity in contracts between Hidroelectrica and Luxten-Lighting, Electrocarbon and Elsid are still found not to be in line with the benchmark price. The percentage price differences are considerably higher for the contracts with Luxten-Lighting for the years 2008 and 2009 and with Electrocarbon and Elsid for the period 2007-2010. In addition, two other contracts appear, namely with Alro in 2008 and Energy Holding in 2009. The percentage price difference in both cases is very close to the percentage price difference of the MDD contract. The observed price for the 2008 contract with Alro is […] % lower than its estimated price, whereas the percentage price difference in relation to the MDD contract was […] % in 2008. The observed price for the 2009 contract with Energy Holding is […] % lower than its estimated price, whereas the percentage price difference in relation to the MDD contract was […] % in 2009. Overall, these two contracts should be considered to be in line with the benchmark price.

In addition to the main sensitivity analysis performed by defining the MDD as a percentage difference in relation to the benchmark price, which confirms the validity of the results, further sensitivity tests of the results obtained from the preferred definition of MDD in absolute values were carried out, as follows:

• an arbitrary 10 % mark-up was deducted from the absolute value of the MDD: the modified MDD does not affect the conclusion for the contracts under investigation from 2008, 2009 and 2010. For 2007, a difference of 2,5 RON/MWh below the applied price of […] RON/MWh would appear for EFT (i.e. the benchmark-compliant price would have been […] RON/MWh); this difference, not observed for EFT prices in 2008, 2009 and 2010, is deemed to be negligible,

• likewise, use of the second downward deviation (2nd DD) as reference would capture only two more contracts (Luxten-Lighting and EFT) for 2007, but this does not mean that any other of the contracts under investigation in the period analysed would appear as being below the benchmark prices or that the contract with EFT would appear as being below the benchmark price for 2008, 2009 and 2010,

• regression data from 2011, not initially tested, was introduced: even if the values shown in Tables 2 to 5 change slightly, the qualitative results remain unchanged,

• investigated contract prices deemed to be in line with the benchmark price were reintroduced in the regression as dataset (in-sample): in this case too, the conclusion for the contracts under investigation does not change.

It follows that these additional sensitivity tests relating to the main approach confirm the robustness of the results of the econometric analysis.