Russian bank data: Reasons of bank closure

For every closed Russian bank I collect its closure reasons. I divide these reasons into categories and report them in the form of tags. The data come from online sources. Combined with Russian banks' financial statements, freely available on the Central Bank of Russia's website, this dataset provides ample possibilities for empirical research: on bank failure prediction models that take into account the after-the-fact revealed reasons of bank closure; on the extent, determinants, and effects of financial misreporting by Russian banks; on the relative performance of banks pursuing particular business models, such as captive banks.

For every closed Russian bank I collect its closure reasons. I divide these reasons into categories and report them in the form of tags. The data come from online sources. Combined with Russian banks' financial statements, freely available on the Central Bank of Russia's website, this dataset provides ample possibilities for empirical research: on bank failure prediction models that take into account the after-the-fact revealed reasons of bank closure; on the extent, determinants, and effects of financial misreporting by Russian banks; on the relative performance of banks pursuing particular business models, such as captive banks.

Data
The dataset [1] represents a table in Excel. Variables are in columns; variable names are in the first row. Rows correspond to different observational units, that is, different Russian credit institutions. Each institution is identified by its unique official registration number reported in the first column. This registration number should be used as key when merging the dataset with other data on Russian banks [2]. The list of institutions is limited to those that have (been) closed down. The closure date is reported in the second column. In what follows I use the words 'bank' and 'credit institution' interchangeably.
For each bank the third column reports a list of 5-character tags separated by spaces. Each tag corresponds to a specific reason of bank closure. Missing values represent banks for which closure reasons are not available. The next section provides further details. E-mail address: a.karas@ucr.nl.

Experimental design, materials, and methods
When a bank closes down the Central Bank of Russia (CBR) issues an official communication in which it explains the reasons of closure. Excerpts from these official communications are reported on a number of online portals. In what follows, I draw on the 2700þ excerpts reported in the Memory Book of Banki.ru [3]. I read the excerpts. As I read, I encounter reasons of bank closure. For each newly encountered reason I generate a tag, which I then place next to every bank whose excerpt contains that reason. For example, a tag S1mac is placed next to every bank that closed down because of a merger or acquisition. Banks with multiple reasons get multiple tags.
The correspondence between tags and closure reasons is summarized in Table 1. The first column reports the tags. The second column reports the reasons of closure. The third column reports the most common phrases the Central bank uses in its official communications that correspond to those reasons. All tags are divided into 8 categories. I discuss these categories below.
A-tags concern asset risk. A1llr tags banks that create insufficient provisions for possible losses. The CBR reports these banks to underestimate risk, overestimate asset quality, and create insufficient loss reserves. A2rsk tags banks that take too much risk. Their excerpts mention high risk credit policy, excessive risk taking, low asset quality, large amount of distressed assets, and similar. The two A-tags often co-occur, that is, banks that take excessive risk tend to underprovision for it.
C-tags concern capitalization and solvency. C1fin tags banks whose financial health the CBR considers generally unsatisfactory. This dissatisfaction is revealed in statements like 'unsatisfactory financial condition', 'hides real financial condition', 'fails to improve the situation', 'experiences problems', and other pejorative statements. The reasons for CBR's dissatisfaction are often unspecified. C2sol tags banks that are (on the verge of being) insolvent. The CBR reports these banks to be insolvent, bankrupt, in need of measures to avoid bankruptcy, or having lost all of their capital. C3c2p tags banks with too low capital adequacy. Their excerpts mention capital adequacy being critically low or below two percent. C4low tags banks whose capital falls short of established requirements, most frequently, short of established minimum value of statutory capital. The CBR often identifies these banks as failing to bring capital in line with requirements.
E1neg tags loss-making banks, that is, banks with negative net income. L-tags concern liquidity. L1deb tags banks whose excerpts mention 'debit balance on correspondent account (with the CBR)'. This 'debit balance' represents a form of borrowing from the Central bank. It

Value of the Data
These data report closure reasons of Russian banks.
The data benefits researchers who study Russian banks or compare banking systems across countries. The data can be used to empirically develop failure prediction models that take into account the after-the-fact revealed reasons of bank closure. The data can be used to study financial misreporting and money laundering by Russian banks.
Specifications Table   Subject Finance Specific subject area Empirical Banking in Russia Type of data was prohibited in 1995. L2lbl tags banks that have (partly) defaulted on their obligations. The CBR identifies these banks as being unable to honor, or effectively not honoring, (some of) their obligations. M-tags concern specific aspects of the bank's business model. M1cap tags captive banks. The CBR reports these banks to primarily serve the business interests of (i.e. lend to) a narrow group of customers connected to the bank's owners (or managers). M2shd tags banks engaged in dubious activities. Their excerpts mention dubious, shadow, doubtful, fictitious, transit (and other synonyms of shady) operations/activities or examples thereof. M3tun tags banks whose excerpts mention asset tunneling, for example, in the form of replacing liquid assets with illiquid. M4frd tags banks whose closure relates to some sort of fraud. Their excerpts mention 'fraud', 'financial pyramid', and 'fake'. M5ter tags banks financial statements submitted late misses deadlines for submitting financial statements; delayed submissions whose excerpts mention violations of the anti-money laundering law (i.e. the federal law "On Counteracting the Legalization of criminal proceeds and the Financing of Terrorism") or related regulations. R-tags concern violations of rules and regulations (with the exception of anti-money laundering law violations). R1fed tags banks that violate federal banking laws. R2cbr tags banks that violate CBR regulations. R3nrm tags banks that violate prudential standards for statutory ratios (norms). R4req tags banks that violate rules on reserve requirements. R5act tags banks that got a license but did not start operations. Their excerpts mention 'no activity', 'no transactions', and 'delay in starting operations'. R6lic tags banks that engaged in certain activities without having the appropriate license. R7stc tags banks that did not follow the right procedure when paying their initial capital. The CBR reports these banks to have violated rules for statutory capital formation.
T-tags concern transparency. T1mis tags banks engaged in financial misreporting. The CBR reports these banks to submit unreliable financial statements, to violate principles of accounting, to misreport specific balance sheet items, and similar. Note that insufficient provisions for possible losses are also a form of misreporting. For that reason, all banks tagged with A1llr should be considered as also tagged with T1mis. T2non tags banks that did not submit their financial statements to the CBR. T3aud tags banks that did not undergo a yearly audit. T4ddl tags banks that did not submit their financial statements on time.
S1mac tags banks that discontinue operations because of a merger, acquisition or transformation into a subsidiary. These banks come in two varieties. Some remain healthy until merger or acquisition. Others experience problems and are subject to Central Bank intervention. For these latter banks their acquisition is often facilitated by the Central Bank. Unfortunately, my data do not allow to distinguish these two varieties. Finally, S2liq tags banks that discontinue operations because of a voluntary liquidation.