Quarterly financial accounts and sectoral balance sheets’ statistics in the context of financial instruments and sectors of the economy, formed on the basis of microdata, is a reliable information basis for a comprehensive and deep macroeconomic analysis. Most organizations in the Financial Corporations sector (S12) report on an annual and quarterly basis, but for some of them, including organizations of the Other Financial Institutions subsector (S125), which perform non-licensed activities, data is only available on an annual basis. On a quarterly basis, only a small part of these organizations’ reporting is available. Therefore, to ensure the completeness of the range of companies in the formation of statistics of financial accounts and sector balance sheets, it is necessary to restore gaps in the quarterly indicators of financial statements of organizations. In this article the results of restoring omissions in the quarterly indicators of financial statements for the Other Financial Institutions subsector (S125) in the Russian Federation, which perform nonlicensed activities, are presented. In particular, the results of the traditional methods (regression analysis, individual growth rates, cluster analysis) and Machine learning-based methods, that can be applicable to recover data, such as random forest and generative neural network.