VoxEU Column Politics and economics Financial Regulation and Banking

“Crime and Punishment”: How Russian banks anticipated and dealt with global financial sanctions

Financial sanctions against Russia’s state-owned and controlled banks were imposed consecutively between 2014 and 2019, allowing banks that would potentially be targeted in the future to adjust their international and domestic exposures. This column explores the informational effects of financial sanctions, showing that compared to similar private banks, ‘not yet sanctioned’ financial institutions immediately reduced their foreign assets while, rather unexpectedly, expanding their foreign liabilities. These informational effects crucially depend on geography, with targeted banks located further from Moscow decreasing their foreign assets by more and raising foreign liabilities by less than those located near the Kremlin. 

Politics affects the banking sector in many ways (e.g. Calomiris and Huber 2014). For example, governments in many countries direct commercial bank lending to specific sectors and/or stimulate lending to small and medium-sized enterprises (e.g. Brown and Dinc 2005). And during the recent COVID-19 pandemic, many governments created emergency loan guarantee schemes that were covering and spurring their banks' lending. In this column, we turn to another recent and striking episode of political impact, namely, the global financial sanctions on Russian banks with close ties to their domestic government that commenced in 2014 and were sequentially imposed on various banks during a five-year period. In general, economic sanctions become increasingly popular from 2010s, being mostly driven by the US to restrain politically unfavourable regimes (Felbermayr et al. 2021). While their effects at the firm level are well studied (Crozet et al. 2021, Ahn and Ludema 2020, Belin and Hanousek 2020),1 the effects of sanctions at the bank level remain unclear. 

The sequential imposition of the Western sanctions against Russia’s largest state banks constitutes a very interesting and policy-relevant laboratory to analyse not only the immediate effects on the already-sanctioned banks, but also the effects on those banks that are not yet sanctioned but that seem to be targeted and may be sanctioned in the near future. The point is that such targeted banks have time to adjust their international operations before the actual sanctions materialise. Henceforth, we refer to the immediate effects of sanctions on the sanctioned banks as direct effects, and we refer to the adjustments of the potentially targeted but not yet sanctioned banks to the anticipated sanctions as informational effects. 

According to the US Office of Foreign Assets Control (OFAC), over the period of 2014–2019, financial sanctions were imposed on 44 banks that were owned or controlled by either the state or major oligarchs in Russia. However, the ownership structure of banks is fuzzy – some banks could be formally private but are in fact influenced by the state through a chain of other state-owned firms and banks. As shown by Karas and Vernikov (2019), who attempted to unfold such chains through a comprehensive analysis of firms’ annual financial reports, there are at least 40 banks that are controlled by the state but left uncovered by the sanctions. This creates an interesting effect of treatment diffusion, since not only the actually sanctioned banks but also the (as yet) uncovered banks could adapt their operations in advance.

In a recent study, we estimate and compare the direct and informational effects of sanctions against the largest Russian banks with respect to their international and domestic operations and address the issue of treatment diffusion due to fuzzy bank ownership structure (Mamonov et al. 2021).

Stylised facts, or what the raw bank data say about the sanctions

From the OFAC database one can infer that there are two major types of sanctions: those affecting debt and those restricting assets. The former represent restrictions mainly on placement of new debt in international markets; the latter impose restrictions on foreign asset holdings of treated banks. Henceforth, we label these two types of sanctions ‘debt' and ‘asset’ sanctions, respectively.2 

Figure 1 plots the evolution of foreign liabilities and foreign asset holdings of selected Russian banks that faced sanctions between 2014 and 2019. The very first sanction arrived in March 2014 and crucially restricted the international operations of the Rossiya bank, owned by the Kovalchuk family (one of richest oligarch families in Russia). The assets sanctions had an immediate negative effect – the bank dramatically decreased its foreign assets (from 25% to 8%) and foreign liabilities (from 5% to 2%) within just one month.3 Other potentially targeted banks follow the Rossiya Bank. 

Figure 1  Evolution of foreign assets and liabilities before and after sanctions for selected largest Russian banks


Note: The figures report foreign liabilities (black line) and foreign assets (grey line), as a percentage of respective total assets, of selected banks that faced sanctions. The red vertical line marks March 2014 — the month in which financial sanctions against Russian banks were imposed for the first time (the Bank Rossiya). The blue vertical line represents the period when individual sanctions were then introduced.

Primary effects of financial sanctions: International operations

To test the sanction effects, we first match sanctioned banks with never sanctioned banks using observable characteristics (1:4 nearest neighbourhood matching). We then run a difference-in-differences regression analysis on the matched sample of banks showing how the not yet sanctioned banks adjusted their international operations vis-à-vis matched banks in a specific time window around March 2014 (see Figure 2).

The estimation results clearly indicate that, first, not yet debt-sanctioned banks raised, rather than decreased, their international borrowings after March 2014 (by 3.8% of their total assets at peak). This implies the banks were treating foreign financial markets as an important source of (possibly cheaper than domestic) funds. 

Second, not yet asset-sanctioned banks exhibited different reactions. After March 2014, they turned to decreasing both international borrowings (by 2.5% of their total assets at peak) and international asset holdings (by 2.2%). These figures suggest the banks decided to avoid gambling for Western funds.

Figure 2 Evolution of the informational effects of financial sanctions on foreign assets and liabilities (percentage points changes in terms of total assets)



Note: The figures report the difference-in-differences estimates on expanding windows [-k,k] with k=1,2... 36 months after the sanction imposition on the bank, Rossiya (March 2014). Sanctioned and non-sanctioned bank groups are matched within 1 year prior to March 2014.

Our further analysis shows that geography matters a lot in explaining these informational effects of the sanctions. First, not yet debt-sanctioned banks were less likely to expand foreign liabilities if located further from Moscow. Second, those not yet debt-sanctioned banks whose headquarters were located further from Moscow were more likely to reduce their international assets. Therefore, these banks could reveal a fear of asset freezes while being less sure on which of the two types of sanctions will be introduced. Third, not yet asset-sanctioned banks behave differently – specifically, they were less likely to reduce their international borrowings in the months after March 2014 if they were located outside Moscow. Geography may proxy for a differential exposure of these banks to the information on upcoming sanctions.

Secondary effects of financial sanctions: Domestic operations

Regarding domestic borrowed funds, we find that neither private nor corporate depositors organised withdrawals on not yet sanctioned banks. However, when the sanctions arrived, the sanction-based withdrawals amounted to -2.2% and -10% of the debt- and asset-sanctioned banks' total assets, respectively, despite the fact that the deposits insurance system was working perfectly well. The government worked fast. It stepped in and – either directly or indirectly (through inter-bank market) – supported the banks, thus preventing their disorderly failure.

Second, we reveal a ‘credit reshuffling’ effect. Both not yet debt- and asset-sanctioned banks turned to reducing loans to non-financial firms (Figures 3a and 3b) and raising loans to individuals (Figures 3c and 3d). The estimated size of this reshuffling is 4% of Russian GDP (average across 2014-2019). We interpret this result as the banks' forward-looking willingness to insure the profitability of their loan portfolios from a rising risk of sanctions against Russian firms per se.4

Figure 3 How banks adjusted their domestic lending after sanctions? (by sanction type, percentage points change in terms of total assets)



Note: The figures report the difference-in-differences estimates of the informational and direct effects of the financial sanctions on the domestic loans to individuals and firms by Russian targeted banks. The estimates are obtained by running DiD on expanding windows [-k,k] with k=1,2... 36 months after either bank-specific sanction date (direct effects, black lines) or the date of sanctions against the bank Rossiya (informational effects, pale red lines).

Hidden control by the state and treatment diffusion

The nearly 40 banks that were indirectly controlled by the government but were left uncovered by the sanctions (call them diffused banks) were in between the asset and debt-sanctioned banks in terms of size and had similar structure of their international operations. It is clear that the already sanctioned banks, if necessary, could transfer a part of their prohibited international operations to their unsanctioned subsidiaries, thus dampening the overall effects of sanctions.

We argue that the subjectively perceived probability of being sanctioned in the future crucially depends on the share of government-connected persons5 on the board of directors of either diffused or not yet sanctioned banks – the greater the share, the easier the recognition by Western countries, and the higher the subjective probability of being sanctioned. To create the government share variable, we manually collect the data on each and every member of the board of directors for each and every state-controlled bank that had or had not eventually been sanctioned. We extract this information from several sources, starting from the banks' annual financial reports, the persons' CVs, and Google Search.

Our results suggest that a one standard deviation increase in the share of government-connected persons on the board of directors raises the probability of being debt-sanctioned by between 1% and 4%, depending on the month, whereas the effects are near zero for asset sanctions.6

We find that those banks with government-connected persons on the board of directors were likely to behave very similarly to those banks that were eventually sanctioned. First, those who could anticipate debt sanctions were raising international borrowings, especially if located in Moscow, and decreasing their foreign assets, especially if located farther from Moscow. Second, those who could anticipate asset sanctions were reducing international borrowings and selling foreign assets in advance.

We believe our results may have important policy implications for both the Russian government and Western countries. For the former, our estimates imply that, if the imposition of sanctions were not phased-in, the negative effect could have been larger, which is economically inefficient for a country with long-lasting recessions. For the latter, our results indicate that, despite the phasing-in, the sanctions still had a significant effect.


Ahn, D, and R D Ludema (2020), “The Sword and the Shield: The Economics of Targeted Sanctions”, European Economic Review 130: 103587.

Belin, M, and J Hanousek (2020) “Which Sanctions Matter? Analysis of the EU/Russian Sanctions of 2014”, Journal of Comparative Economics 49: 244–257.

Brown, C, and I S Dinc (2005), “The Politics of Bank Failures: Evidence from Emerging Markets”, Quarterly Journal of Economics 120(4): 1413–1444.

Calomiris, C, and S Haber (2014), Fragile by Design: The Political Origins of Banking Crises and Scarce Credit,  Princeton University Press.

Crozet, M, J Hinz, A Stammann and J Wanner (2021), “Firms’ exporting behaviour to countries under sanctions”,, 5 March.

Davydov, D, J Sihvonen, and L Solanko (2021), “Who Cares about Sanctions? Observations from Annual Reports of European Firms”, BOFIT Discussion Papers 5/2021.

Felbermayr, G, A Kirilakha, C Syropoulos, E Yalcin and Y V Yotov (2021), “The ‘Global Sanctions Data Base’: Mapping International Sanction Policies from 1950-2019”,, 18 May.

Karas, A, and A Vernikov (2019) “Russian Bank Data: Birth and Death, Location, Acquisitions, Deposit Insurance Participation, State and Foreign Ownership”, Data in Brief 27: 104560.

Mamonov, M, A Pestova, and S Ongena (2021), “‘Crime and Punishment?’ How Russian Banks Anticipated and Dealt with Global Financial Sanctions”, CEPR Discussion Paper no. 16075.


1 Several studies complement our bank-level analysis by exploring the effects of sanction at the firm level. Belin and Hanousek (2020), for example, focus on Russian non-financial firms and study the effects of sanctions on their international trade flows vis-à-vis their US and EU trade partners. Ahn and Ludema (2020) also investigate the effects of sanctions against Russian firms, showing that the targeted approach to sanctions (i.e., smart sanctions), was new but efficient since they negatively affected the firms' activities while causing minimal ‘collateral’ damage. Davydov et al. (2021) analyse how European firms perceive Russia-related sanctions. Crozet et al. (2021) study the impact of sanctions against Russia (and other countries like Iran, Cuba, and Myanmar) on the probability of serving a market at the firm-level using monthly custom data on French firms.

2 According to the US Department of the Treasury, debt sanctions are called ‘sectoral’ while assets sanctions are titled ‘entity’.

3 Before the sanctions, the bank Rossiya had intensive international operations borrowing funds from financial markets and granting loans to foreign banks and foreign non-financial firms. All these became minor after the sanctions in the long run. Another implication of sanctions is that Visa and Mastercard had blocked all operations of the bank's credit cards. The bank had lost its ability to carry out transactions in foreign currency. However, the Russian government had fully, and even over-, compensated these restrictions to the bank by increasing its deposits and by replacing “Alfa-bank” (the largest private bank in Russia, inside top-10 banks in terms of assets, never facing sanctions) with the bank Rossiya as an operator of the wholesale energy-market in the country (with annual turnover equalling 1.5% of GDP).

4 The firms themselves could face sanctions and stop repaying their debts while individuals (at least, those not in the OFAC’s Specially Designated Nationals list) were free of such ‘sudden’ constraints. Our conclusion on reductions of loans to firms is consistent with the findings in Ahn and Ludema (2020), who document that the sanctions indeed had a negative effect on Russian firms.

5 For instance, federal or municipal ministers, senators, city mayors, or regional governors from the ruling political party Edinaya Rossiya (literally, “United Russia”), oligarch families with close ties to the Kremlin, governors of other recognised state-controlled entities, and so on.

6 However, in respective logit regressions, we control for many other observable characteristics such as international operations, the structure of domestic assets and liabilities, quality of loans, profitability, and so on, so that we still obtain enough variation in the predicted probabilities of asset sanctions.

4,199 Reads