The information about tax changes’ effects on aggregate output is highly important for economic policy, especially in times of economic contractions. Russian economy underwent the series of tax changes during 2003–2020. For better tax policy design, it is necessary to understand and to evaluate the effects of this changes on aggregate output, which is the purpose of this study. To solve the problem of endogeneity we use two methods – “narrative approach” and “classical” approach. The first one uses data on exogenous, not driven by economic conditions, tax changes from official documents and forecasts. The second one uses cyclical component of the aggregate tax receipts as tax shocks indicator. Using both methods we estimated a VAR model of Russian economy for period 2003–2020. The implementation of “narrative approach” did not provide any significant effect possibly due to vulnerability towards the measurement error. Based on the classic approach we found that tax changes affect output with a 1-year lag and a 1 percentage point shock of aggregate tax receipts to GDP ratio lowers output growth by 0.7–0.88 percentage points. This result is robust to inclusion of additional factors in the model. The results are mostly consistent with existing research. Implementation of
“narrative approach” proved to be restricted in Russia. “Classical” approach allows to conclude that tax changes could serve as an appropriate tool of countercyclical policy in Russia. On the other hand, increasing tax burden in times of downturn could be highly harmful for recovery. These results should be interpreted taken into consideration the limitations of the VAR method used.