This paper presents possible approach how different sources of data at farm level, national statistics and analytical models could be merged in simulation process to analyse income risk at the sector level. Baseline is production structure resumed out of annual subsidy applications as key information per each agricultural holding within the sector. Presented approach utilises potential of random number generator and random distributions of Monte Carlo to roughly reconstruct different sources of risks in different states of nature that may occur with diverse probabilities at the particular farm. In such a manner income situation at sector level is analysed. The developed approach is tested on the 21 farm types further divided into 13 economic classes. Obtained preliminary results suggest that this could be useful approach for rough estimation of income risk and points on some limitations and drawbacks that should be further improved.