Simona Lištiaková

In DEA, adding a constant to the data is a common practice to avoid negative input or output values. However, the choice of constant value can affect the efficiency scores and the ranking of the companies.

There is no specific maximum value of a constant that can be added in DEA. However, there are some guidelines that can be followed. For instance, the constant value should not be too large, as it may distort the relative efficiency scores or mask differences between the companies. In general, a constant value of 10 to 20 percent of the average value of the inputs or outputs is recommended.

In your case, adding a constant value of 70,000€ to the negative income values may significantly affect the efficiency scores, especially for companies with low positive income values. You may consider using a lower constant value or other methods for dealing with negative values, such as linear programming or inverse DEA. Additionally, you should assess the sensitivity of the results to the constant value by varying it and checking if the conclusions remain robust.

Adding constants to make the data all positive is not a good method. This is because DEA measurements have relative characteristics. You can try to normalize the data and convert it to a positive number. In fact, some DEA models can absorb negative values. MSBM (Modified Slack Based Measure) is an SBM (ERM) proposed by Sharp, Meng, &Liu (2007).