Abstract
<jats:p>Existing bankruptcy forecasting methods often show contradictory results due to the subjectivity of the choice of factors and fixed weighting factors. The article proposes a hybrid approach combining expert estimates obtained by hierarchy analysis (MAI) and econometric modeling based on binary choice. An integral indicator of financial stability has been formed, the weights of which are determined on the basis of agreed expert judgments. Based on a sample of 45 manufacturing enterprises for the period 2020-2023. Calibration of probit and logit models has been performed, which make it possible to estimate the probability of default depending on the value of the proposed integral criterion. The obtained models have a high predictive ability and can be used to monitor the state of business entities.</jats:p>