Efficiency of Market Discipline in the Interbank Market

Efficiency of Market Discipline in the Interbank Market: The Case of Russia Irina Andrievskaya (Università degli Studi di Verona) Maria Semenova (NRU HSE, Moscow) April, 12 th 2013

Outline Introduction and motivation Literature Model Data Results 2

Introduction and motivation The interbank loan market plays an important role in the efficient functioning of the whole financial system Distribution of liquidity among banks Increased linkages among banks Contagion effect Loans provided in this market are uninsured Financial institutions should have incentives to monitor those they are lending to and punish them for excessive risk-taking Market discipline should reduce threat of contagious closure of credit limits 3

Introduction and motivation Market discipline: ‣ A phenomenon when “financial markets provide signals that lead borrowers to behave in a manner consistent with their solvency” (Lane, 1993) ‣ Its significance is recognized in several policy initiatives (e. g. Basel II, Pillar 3) and is necessary for supporting macroprudential supervision ‣ Rochet and Tirole (1996) stress that the efficiency of peer monitoring (market discipline) substantially decreases if there is some government intervention such as state insurance of interbank claims or a too-big-to fail policy 4

Introduction and motivation Russian market for interbank loans is… …an important source of funds Average share of bank loans in total liabilities 9 -10% May amount to 70% …rather fragile and is “frozen” at crises times 1995, 1998, 2004, 2008 Including credibility gap (“crisis of trust”) …a network and may be exposed to contagion effects 5

Introduction and motivation 6

Introduction and motivation 7

Introduction and motivation 8

Introduction and motivation According to the strategy of the banking sector development until 2015, the regulatory standards Basel II and Basel III are going to be fully adopted starting from 2019 Research questions: Is there market discipline in the Russian market for interbank loans? Is market discipline efficient in constraining the excessive risk-taking behavior of banks? 9

Literature Existence of market discipline: Market of retail and corporate deposits: price discipline (Hannan, Hanweck, 1988), (Ellis, Flannery, 1992); quantity discipline (Jordan, 2000), (Goldberg, Hudgins, 1996); maturity shifts mechanism (Murata, Hori, 2006), (Semenova, 2007) Securities market: stock prices (Brewer, Lee, 1986), (Distinguin et al. , 2006); debt prices (Ashcraft, 2008), (Goyal, 2005) Interbank market: (Furfine, 2001), (King, 2008), (Cocco et al. , 2009), (Angelini et al. , 2009) 10

Literature Efficiency of market discipline (extent to which market discipline affects the risk -taking behaviour of banks) Market of retail and corporate deposits: efficiency is lower the more the government intervention is (Demirgüç-Kunt, Huizinga, 1999), (Billett et al. , 1998), (Demirgüç-Kunt, Huizinga, 2004) Interbank market: Still limited research, controversial conclusions Major contributors: (Nier, Baumann, 2006) for 729 banks from 32 countries, (Dinger, von Hagen, 2009) for Central and Eastern Europe, (Liedorp et al. , 2010) for Netherlands 11

Model 1: Existence of market discipline MDi, t (ld_mbkb) - market discipline variable - growth rate of loans received by a bank i from other banks in quarter t Heckman estimation procedure ‣ Not all banks participate in the market ‣ The growth rate of interbank borrowings is observed only for those that take part in this market (when borrowings are greater than zero) ‣ Zero values of the growth rate for banks, which do not enter the market, do not represent their corner solution ‣ The data are not censored at zero 12

Model 1: Existence of market discipline Sample selection problem and we have to account for the non-random nature of the sample. In this case OLS regression will produce inconsistent estimates of the coefficients Heckman sample selection model Main equation + participation equation Participation equation: Dependent variable: indicator of bank’s involvement in the interbank market (it is equal to 0 when a bank has no interbank borrowings, and to 1 when interbank borrowings are greater than 0) Independent variables: N 1, the ratio of total loans over total assets, logarithm of total assets and interbank borrowings over total assets 13

Model 1: Existence of market discipline BFi, t-1 - vector of bank fundamentals with one quarter lag CAMEL C (h 1) – regulatory bank capital over risk-weighted assets ratio - “N 1” A (bl_loan) – NPL over total loans M (pe_profit) - personnel expenses over total revenues E (roa) – ROA L (h 3) - short-term (liquid) assets over short-term liabilities ratio “N 3” Size (lnas) – logarithm of total assets Ii, t-1 (mbkb_as)- share of loans received from other credit institutions in total liabilities of a bank i with one quarter lag Tt – dummy variables for each quarter 14

Model 2: Efficiency of market discipline Panel data model A choice among pooled OLS, fixed effect and random effect models is done based on a set of appropriate tests (the Hausman test, the Breusch-Pagan test and the test for differing group intercepts) yi, t includes indicators of bank’s capital level, credit, liquidity and overall bank risk in quarter t Capital level - capital adequacy ratio N 1 (h 1) Credit risk - ratio of NPL over total loans (bl_loan) and ratio of reserves over total assets (res_as) Liquidity risk - ratio of instant liquidity N 2 (h 2), ratio of current liquidity N 3 (h 3), ratio of long-term liquidity N 4 (h 4) Overall risk – ratio of risk-weighted assets over total assets (rwa_as) 15

Model 2: Efficiency of market discipline xit-1 – interbank variables with one quarter lag Ratio of total interbank borrowings over total assets (mbkb_as) Ratio of interbank foreign borrowings over total assets (fmbkb_as) zit-1 - bank-level control variables with one quarter lag Share of interest income in total revenues (in_rev) Share of total loans in total assets (tl_as) ROA (roa) Size (lnas) Growth rate of total assets (ld_lnas) 16

Data Mobile database (“Banks and Finance” Analytical System) Banks’ financial statements (CBR website) 665 banks (86% of total assets of the system) 1 Q 2004 -2 Q 2011 Panel is unbalanced Estimations are done for the period before the crisis (1 Q 2004 -1 Q 2008), for the period afterwards (2 Q 2008 -2 Q 2011) as well as for 1 Q 20102 Q 2011 ‣ Is there any effect of the government intervention on market discipline? 17

Results: Existence of market discipline 18

Results: Existence of market discipline Whole period: market discipline exists ‣ ‣ Bank fundamentals are jointly significant at 1% confidence level Higher capital levels correspond to quicker growth of interbank borrowings Share of interbank borrowings in total assets is significant and has a negative sign Size of a bank - significant at 1% confidence level – positively influences the interbank borrowings’ growth rate ‣ too-big-to-fail policy? 19

Results: Existence of market discipline ‣ Before 1 Q 2008 ‣ All bank fundamentals are significant at 1% confidence level ‣ The level of short-term liquidity and the size of a bank have a statistically significant positive effect on the growth rate of interbank borrowings ‣ The level of personnel expenses and the level of interbank borrowings have a statistically significant negative effect on the growth rate of interbank borrowings ‣ After 1 Q 2008 ‣ The level of banks’ bad loans and the level of interbank borrowings have a statistically significant negative sign ‣ The level of short-term liquidity and the size of a bank have a statistically significant positive effect on the growth rate of interbank borrowings ‣ 1 Q 2010 -2 Q 2011 ‣ No evidence of the existence of market discipline 20

Results: Efficiency of market discipline 21

Results: Efficiency of market discipline Most efficient discipliners - foreign lenders ‣ Foreign interbank borrowings have a statistically significant effect on the level of banks’ capital (N 1) ‣ Efficiency of market discipline decreases after 1 Q 2008 ‣ Market discipline is inefficient during 1 Q 2010 -2 Q 2011 No effect on liquidity levels, credit and overall bank risk Results are robust if we consider top-30 banks and banks with positive interbank borrowings 22

Thank you for your attention!