The paper introduces a new monthly index of financial stress for Russia for the March 2008 – March 2018 period. The index is based on 12 well-established and mostly publicly available standalone metrics of financial instability, including credit-to-GDP gap, debt-service-ratio and real estate price index, provided by the BIS. I seek an optimal method to aggregate the metrics to derive a composite index. Based on the local projections technique [Jordá, 2005, 2009] and Bayesian model averaging, I show that conventional aggregation methods such as principal component analysis (PCA) can be outperformed by the approaches, better capturing the nonlinear and non-Gaussian nature of the standalone indicators of financial instability. Namely, the dynamic factor model with a single factor fares best of all the considered methods.

The composite index based on the dynamic factor model accurately captures the dynamcs of financial instability in the Russian financial sector, with the peaks occurring in the late 2008 and the late 2014 – early 2015. I also show that the financial stress index exerts an adverse effect on industrial production alongside the VIX index, explicitly accounting for oil prices, global and domestic indices of economic policy uncertainty as well as geopolitical risk. This negative effect of financial stress exhibits persistence in the medium run.