The paper explains the use of composite indicators as objective functions in econometric decision models. We consider an example of minimizing the harm from two kinds of pollution at the workplace, which can be reduced by two types of anti-pollution measures restricted to a given budget. At first, we define the pollution harm as a sum of two factors and find the optimal budget distribution to minimize it. Next, the harm is regarded as a probabilistic risk, and the cumulative effects from the interaction of two factors is taken into account. The corresponding indicator of the occupational hazard becomes a quadratic function, resulting in a different budget distribution. We also discuss the methodology of using composite indicators and refer to advanced methods of their construction.