Disentangling External flows and policy regulation
1. Disentangling external flows (external shocks) and policy
and regulation effects on the credit activities of banks in
three emerging countries during the Great Recession
Velimir Bole
EIPF, Ljubljana, Slovenia
Milan Lakićević
Faculty of Economics, University of Montenegro, Podgorica, Montenegro
Ana Oblak
Faculty of Economics and Institute for South-East Europe, University of Ljubljana,
Slovenia
Janez Prašnikar
Faculty of Economics and Institute for South-East Europe, University of Ljubljana,
Slovenia, and CEPR
2. Motivation
• The global financial and economic crisis in 2008/2009 has
shown the critical importance of the financial sector for the
amplification of (external and internal) shocks on
macroeconomic activity
• The crisis revisited the theoretical interest in the interplay of
the financial and real sectors in the amplification of shocks
• To the borrower channel and bank’s balance sheet channel a
more elaborated liquidity channel was added
• Two crucial components - funding and market liquidity
• The funding liquidity component is especially important for
studying the financial crisis development in less developed
countries, because of the crucial role that capital inflows
swings had directly or indirectly on the dynamics of the liability
side of the less developed countries’ bank balance sheets
3. Objective
• The main objective of our paper is to show the very mechanism which transmits and amplifies
the effects of external (specifically, foreign capital flows and real demand) shocks impinging on
the domestic economy
• We compare the transmission mechanisms of three Balkan countries (Slovenia, Croatia, and
Montenegro) in the Great Recession.
• To better analyze the details of the transmittion mechanism at play.
• The Great Recession as a natural experiment
• The sizes and intensity of external shocks
• Non-linearities in relations between financial stability and macroeconomic results.
• Balkan countries: no evidence regarding the transmission mechanisms
• Differences: large disparities in structure and sizes of foreign flows, differences in policy freedom
• Similarities: inherited bank-dominated financial systems, used euro or euro- pegged currencies, branches of the same foreign
banks, synhronized GDP.
• We study retail and wholesale funding channels transmitting retail and wholesale fund effects
caused by external shocks, as well as policy and regulation specificities that impact lending to
firms and, separately, to households throughout the boom (2007–2008), bust (2009–2010),
and recovery (2011–2013) periods of the Great Recession in three selected countries:
Slovenia, Croatia, and Montenegro.
4. Objective
• We study retail and wholesale funding channels transmitting
retail and wholesale fund effects caused by external shocks,
as well as policy and regulation specificities that impact
lending to firms and, separately, to households throughout
the boom (2007–2008), bust (2009–2010), and recovery
(2011–2013) periods of the Great Recession.
• Methodology:
• Theoretical model of bank credits
• Operational model of supply and demand factors of credit
amplifications
• Panel data on banks credits to households and firms
• The policy and regulation specifities are derived by using
common factors extracted from a set of variables ecompassing
macroprudential policy and standard macro and structural
policies.
5. The banking environment
• Slovenia
• Pre-crisis: fast and complete freeing of (foreign) financial flows (after
the country entered the EU and accepted the euro)
• Post-crisis: reactions to the financial crisis were slow and procyclical,
partly because of its own mistakes, and partly because of the tough
measures enforced from the EU
• Croatia
• Pre-crisis: several measures were introduced over the 2003–2008
period aimed at reducing capital inflows through debt instruments
• Post-crisis: Croatia managed to preserve relatively stable foreign
inflows to deposit-taking corporations
• Montenegro
• Pre-crisis: the country was characterized by huge (gross and net)
capital inflows, especially of equity capital and FDI
• Post-crisis: the sudden stop of wholesale financing almost caused the
collapse of the banking system, with severe drops of deposits in the
private sector and the low liquidity of banks
6. Figure 2: Capital flows based on financial accounts data
a. Portfolio investment (equity securities) and FDI,
net inflows (percent of GDP), all sectors (percent of
GDP)
b. Portfolio investment (debt securities) and other
investment, net inflows (percent of GDP), all sectors
(percent of GDP)
Portfolio investment (debt securities) and other investment, net incurrence of liabilities (percent of
GDP)
c. Other sectors d. Deposit-taking corporations, except the central
bank
Source: IMF, 2015.
Note: All data are yearly and given in
percentages of GDP; in Figure 2b,
data for Croatia does not include IMF
funds; other sectors include
households, non-financial corporations
and other financial corporations.
7. The banking environment (2)
a. Increment of loans to households (percent of total
assets of banking sector)
b. Increment of loans to non-financial corporations
(percent of total assets of banking sector)
c. Increment of deposits of non-financial sector
(percent of total assets of banking sector)
d. Increment of financial sector funding (percent of
total assets of banking sector)
Sources: HBN, 2015; CBCG, 2016; BS, 2016.
Note: Data are presented as increments (value in year t less value in year t-1) in balance sheet percentages (in year t);
deposits of the non-financial sector include deposits of households, firms (non-financial corporations) and government;
wholesale funding is defined as total liabilities less deposits of the non-financial sector (households, non-financial
corporations, and government) and capital.
8. Stylized facts
• In the whole cycle of the recent financial crisis economic activity
was synchronized between the analyzed countries, although the
amplitudes of swings differed
• For the banking amplification of foreign shocks, swings in the bank
wholesale and retail funding were crucial
• Gross inflows through debt and other investments had
incomparably larger amplitude than inflows through FDI and equity
instruments
• Differences among countries in credit trajectories through the
period of the Great Recession resulted from differences in demand
or economic activity, differences in the described exogenous
shocks (direct and indirect gross foreign financial flows) to the
banking sector and the real economy, as well as differences in
policy interventions or the corresponding market institutions
buildup
9. Macroprudential interventions
Categories
Slovenia Croatia Montenegro
boom bust recovery boom bust recovery boom bust recovery
capital buffers 0 -2 0 -1 2 0 0 0 0
lending standards
restrictions -1 0 0 0 0 0 -1 0 0
limits on credit growth and
volume 0 0 1 -1 3 1 -1 1 0
limits on large exposures
and concentration -1 1 0 0 -2 2 0 0 0
liquidity requirements and
limits on currency and
maturity mismatch 0 0 0 1 0 2 0 0 0
loan-loss provisioning 1 -1 -2 0 0 0 -4 3 0
minimum capital
requirements -1 0 0 0 -1 0 -1 1 -1
other measures -1 0 -2 0 0 0 0 -2 0
risk weights 0 0 0 -1 1 0 0 0 0
Source: Macroprudential Policies Evaluation Database, Budnik and Kleibl (2018; own collection)
Note: Macro prudential interventions; cumulative number of loosening less number of tightening in the indicated period; 28 subcategories
included ; boom (2007-2008); bust (2009-2010); recovery (2011-2013)
10. Standard macro policy measures;
changes in indicators
Slovenia Croatia Montenegro
boom bust recovery boom bust recovery boom bust recovery
fiscal deficit -0.45 2.75 2.55 -0.25 1.90 2.45 0.35 2.45 2.15
sales of state
firms 0.05 0.10 0.05 0.00 0.00 0.00 -1.25 0.30 -0.20
government
borrowing -0.50 6.15 1.55 1.70 2.60 2.15 -0.95 -0.50 2.15
CB credits to
banks 1.23 2.74 -0.68 1.09 -0.06 -1.21 -2.62 1.99 1.56
Source: Eurostat; Central banks; own collection
Note: Indicators of standard policy measures; average changes in the indicated period; in percentages of GDP; boom (2007-2008); bust (2009-
2010); recovery (2011-2013).
11. Structural policy measures
Changes in
legislation of
Slovenia Croatia Montenegro
boom bust recovery boom bust recovery boom bust recovery
wages in public
sector 1 -1 -3 -1 -2 -3 2 -1 -2
privatisation 0 0 2 0 2 -1 1 1 1
labour market 0 1 -1 1 -1 -1 0 -1 -1
capital flows 1 0 1 0 2 2 2 1 -1
Source: Yearly IMF Reports (2006-2014); own collection.
Note: Cummulative number of changes (number of loosening less number of tightening) in the indicated period; boom (2007-2008); bust (2009-
2010); recovery (2011-2013).
12. Operational model of bank credit
dynamics to households and firms
dloans_to_househ_bil = 0b_h+ 2dbank_fin_bil + 3ddeposits_bil +
ɣ ɣ ɣ
4cost_impar_bil
ɣ (-1)+ 5g_ngdp + 6fac_prudent1
ɣ ɣ (-1) + 7fac_prudent2
ɣ (-
1) + 8fac_pol_struc1
ɣ (-1) + 14fmo + 15size+const+
ɣ ɣ ε
(1a)
dloans_to_firms_bil = δ0b_h+ δ1b_n + δ2dbank_fin_bil + δ3ddeposits_bil +
δ4cost_impar_bil(-1)+ δ5g_ngdp + δ6fac_prudent1(-1) +
δ7fac_prudent2(-1)+ δ8fac_pol_struc(-1) + δ14fmo + δ15size+ const+ε
(1b)
where (dloans_to_househ_bil) is the yearly change in bank loans to households (per
unit of the total balance sheet), (dloans_to_firms_bil) is the yearly change in bank
loans to firms (per unit of the total balance sheet), (b_h) and (b_n) are correction
factors, (dbank_fin_bil) is the wholesale (bank) funding channel (change in loans to
banks per unit of the total balance sheet), and (ddeposits_bil) denotes change of the
total deposits (per unit of the total balance sheet); (cost_impar_bil) denotes yearly
costs of impairment (per unit of the total balance sheet); (g_ngdp) is the growth of
GDP. Variables (fac_prudent1_1 and fac_prudent2_1) indicate the lagged values of
the first two factors extracted from the set of indictors encompassing macroprudential
interventions, while (fac_pol_struc_1) stands for the lagged value of the first factor
extracted from the standard macro policy variables and indicators of policy structural
interventions. Variable (fmo) is the dummy for a foreign-owned bank. Variable (size)
is the dummy for the size of a bank. Finally, (const) is the intercept and ε the error
term.
13. Hypotheses
H1: The funding channel was a sizable driver of the credit trajectory
throughout the Great Recession episode in the three Balkan countries.
H2: The wholesale funding of banks was more important for credit activity to
firms than for credit activity to households.
H3: Erratic (unsystematic) use of macroprudential policy interventions and
regulation didn’t prevent destabilization of credit activity and so
contributed to financial instability in observed three countries during the
Great Recession.
H4: Other policy interventions and regulations (standard macroeconomic
policies, structural policies) didn’t mitigate destabilization of credit growth
and so increased the procyclicality of the credit trajectory throughout the
Great Recession .
H5: Other bank characteristics (ownership and size) had significant effects on
credit trajectory in the observed Balkan countries through the Great
Recession.
14. Data
• 55 banks from Croatia, Montenegro and Slovenia
• (2007-2013)
• Unbalanced panel
• in year 2010, it encompasses 30 out of 33 in Croatia, 8
out of 11 in Montenegro, and 17 out of 22 in Slovenia
• The main source of data was Bankscope, which was
augmented with hand-collected data from the banks' annual
reports
• 3 policy factors are exctracted from the corresponding sets
of policy indicators by factor analysis (poliychoric correlation
matrix is caculated: Ekstrom (2010); Kolenikov (2016)).
• 2 factors of macroprudential interventions
• 1 factor of standard policy measures and structural
interventions
15. Methodology
• Instrumental variables are from different sources
• Data on number of employees, number of branches,
number of ATMs were collected from banks’ annual
reports
• Real estate prices and data on FDI inflows are taken
from official statistics (IMF, 2015; HBN, 2015; CBCG,
2016; BS, 2015)
• The 2GSLS estimation method was used
• We instrumented costs of impairment, retail and
wholesale bank funding with the number of employees,
number of branches, number of ATMs, prices on the real
estate market, FDI inflows and the interactions among
the mentioned variables
• We used panel estimation for the entire period.
18. Credits to firms – policy effects
-.02
-.01
0
.01
.02
credits
to
firm
s
-
policy
effects
2006 2008 2010 2012 2014
yearabs1
Slovenia Croatia
Montenegro
Source: Model simulations
Note: Credits to firms; effetcs of macroprudential as well as standard macro and structural
policy interventions; effects on credits increment per unit of balance sheet
19. Credits to households – policy effects
-.02
-.01
0
.01
.02
.03
credits
to
households
-
policy
effects
2006 2008 2010 2012 2014
year
Slovenia Croatia
Montenegro
Source: Model simulations
Note: Credits to households; effetcs of macroprudential as well as standard macro and
structural policy interventions; effects on credits increment per unit of balance sheet
20. The estimated effects of
wholesale and retail
(deposit) funding, and the
effects of policy and
regulations, on credits to
households
Funding effects Policy effects
Actual credit
dynamics Wholesale Retail Prudential
Macro and
structural
Boom
Croatia 0.0329 0.0001 0.0011 -0.0030 -0.0007
Montenegro 0.1171 0.0106 0.0014 0.0021 0.0026
Slovenia 0.0270 0.0125 0.0009 -0.0044 -0.0010
Bust
Croatia -0.0004 0.0000 0.0006 0.0046 0.0056
Montenegro -0.0219 0.0003 0.0008 0.0004 0.0051
Slovenia 0.0107 -0.0018 0.0008 0.0004 0.0094
Recovery
Croatia 0.0017 0.0001 0.0006 0.0028 0.0022
Montenegro 0.0081 -0.0019 0.0009 -0.0013 -0.0041
Slovenia 0.0005 -0.0079 0.0001 0.0045 -0.0010
Results -funding and policy effects on
credits to households
21. The estimated effects of
wholesale and retail (deposit)
funding, and the effects of
policy and regulations, on
credits to firms
Funding effects Policy effects
Actual credit
dynamics Wholesale Retail Prudential
Macro and
structural
Boom
Croatia 0.0412 0.0004 0.0217 -0.0129 -0.0003
Montenegro 0.1353 0.0527 0.0287 0.0077 0.0010
Slovenia 0.0980 0.0620 0.0188 -0.0039 -0.0004
Bust
Croatia 0.0294 0.0000 0.0114 0.0053 0.0021
Montenegro 0.0299 0.0017 0.0160 0.0009 0.0019
Slovenia -0.0013 -0.0089 0.0156 0.0010 0.0035
Recovery
Croatia 0.0281 0.0006 0.0120 0.0027 0.0008
Montenegro -0.0101 -0.0095 0.0175 -0.0037 -0.0015
Slovenia -0.0219 -0.0393 0.0027 0.0056 -0.0004
Results – funding and policy effects on
credits to firms
22. Concluding observations
• Our study is a good presentation of the simultaneous
workings of external (capital surge) and internal factors
(procyclical policy intervention) of financial crisis
amplification in the environment of capital scarcity in
developing countries
• We have shown that foreign financial flows influenced bank
retail and wholesale funding channels, and through them
also credits to households and credits to firms.
• Wholesale elasticities much greater for both, firm and
houdeholds credits
• Wholesale funding far more demaging regarding the stability of
banks
• Procyclical volatility much more pronounced in credits to firms
for all three countries
23. Concluding observations
• Small effects of policy interventions + large effects of external
flows → capital control
• Erratic policy orientatiton in all three countries:
• Only Croatia in the boom countercyclical
• Only in the bust countercyclical (still weak) in all three countries
• Procyclical policy in recovery in all three countries.
• Other policy goals are important factors of erratic policy
orientation.
• On average macroprudentil interventions much more effective for
credits to firms., while standard macro and structural for credits to
households.
• Both domestic and foreign banks behave similarly, as well as
bigger and smaller banks.
• The study sheds some light on similar events that occurred in the
past and envisages possible future developments
