The Availability of Household Deposits in the Euro Area vs. Regulatory Approach to Funding Stability of Credit Institutions

In addition, Table 1 shows that these three categories of deposits remained correlated to a group of variables describing the economic situation of the countries (GDP pe[r]

(1)

1

The Availability of Household Deposits in the Euro Area vs Regulatory Approach to Funding Stability of Credit Institutions

Keywords: household deposits, banks, credit institutions, monetary financial institutions, liquidity norms,

funding stability, LCR, NSFR

Introduction

The single funding stability regulations of the CRD IV package1 are the results of credit

institutions’2

constraints during the last banking crisis Before 2007, these entities were characterised by clear tendency to finance their dynamically increasing assets with short-term funds from wholesale markets Such activity weakened their ability to deal with liquidity crisis and intensified systemic risk Its ultimate result was the involvement of central banks to stabilize money markets and governments - to rescue individual credit institutions and to strengthen national deposit guarantee schemes The access to long-term and stable funds, insensitive to changing environment, therefore turned out to be a guarantor of entities’ safety during the turmoil

The new liquidity regulations pay special attention to the household deposits3, emphasizing

their stable nature, both under normal and stress conditions This is reflected in announced low “outflow rates” of Liquidity Coverage Ratio (LCR) and projected high “stability weights” of Net Stable Funding Stability Ratio (NSFR) The real sensitivity of these deposits during

destabilization, however, turned out to be geographically diversified4 resulting in bidirectional

changes with differentiated strength of their values Moreover, the recent financial crisis

caused cash transfers between short-term and long-term bank accounts The EBA claims5 that

the lowest volatility happened to overnight deposits, slightly larger – to saving deposits

1 The CRD IV package contains: Directive 2013/36/EU of the European Parliament and of the Council of 26

June 2013 in access to the activity of credit institutions and the prudential supervision of credit institutions and investment firms, amending Directive 2002/87/EC and repealing Directives 2006/48/EC and 2006/49/EC; Regulation (EU) No 575/2013 of the European Parliament and of the Council of 26 June 2013 on prudential requirements on credit institutions and investment firms and amending Regulation (EU) No 648/2012 (http://ec.europa.eu/internal_market/bank/regcapital/legislation_in_force_en.htm) The “package” implements the solutions of Basel II and Basel III

2

The term “credit institution” is defined in Article of the Regulation of the European Parliament and of the Council (EU) No 575/2013, OJ L 176 of 27 June 2013

3See the Articles 421 and 427 of the Regulation of the European Parliament and of the Council (EU) No

575/2013, OJ L 176 of 27 June 2013

4 EBA (2013) Discussion Paper on retail deposits is higher outflows for the purposes of liquidity reporting under the draft Capital Requirement Regulation (CRR) London: EBA / DP / 2013/02, pp 7-9 In accordance

with Article 409 (3) of the Regulation of the European Parliament and of the Council (EU) No 575/2013, this institution is responsible for technical standards of post-crisis liquidity norms

5

(2)

2

(redeemable at notice), and the most significant – to those with agreed maturity This institution also identifies the features responsible for an above-average volatility of some specific types of household deposits

This paper concerns the stability of household deposits placed with monetary financial

institutions6 in the years 2006-2012 in seventeen Eurozone countries7,on the background of

the current regulatory approach This topic appears as new in the literature since there has been no single liquidity ratios in the past Its purpose is to identify the determinants of the developments in deposit levels during the last financial crisis Moreover, it seeks to assess the geographical differentiation of deposits’ sensitivity to economic and financial instability of the Eurozone, thus the availability of the “resistant” ones for local MFIs The research period covers the time span prior to the crisis and the evolving destabilization, which led to the changes of supervisory arrangements

Hypotheses:

 During the recent financial crisis, the household deposits in the Eurozone remained

geographically diverse in terms of their average values and structure Thus, their availability for the MFIs was spatially varied;

 In the years 2006-2012, in individual Euro area countries, there were observed different

mechanisms of the formation of total household deposits per capita However, one can identify subgroups of corresponding countries in this respect;

 In the analyzed period, the stages of the last financial crisis affected the total values of household deposits in the Eurozone member states with varying strength One cannot define a single phase when the greatest changes occurred in all countries;

 In selected years of analytical period, the household deposits per capita in the Eurozone

were shaped by identifiable factors relating to financial market, national economies, and socio-economic characteristics of households

6

Monetary financial institutions (MFIs) are resident credit institutions as defined in European Union (EU) law, and other resident financial institutions whose business is to receive deposits and/or close substitutes for deposits from entities other than MFIs and, for their own account (at least in economic terms), to grant credits and/or make investments in securities See: https://www.ecb.europa.eu/stats/money/mfi/html/index.en.html

7 This working paper presents the results, which are a part of the research project funded by the National Science

(3)

3

The selection of the Eurozone countries8 - Austria, Belgium, Cyprus, Estonia, Finland,

France, Greece, Ireland, Italy, Luxembourg, Malta, the Netherlands, Portugal, Slovakia, Slovenia and Spain – is due to their diversity in a multi-dimensional space of features, ie.: economic development, condition of the financial market, sensitivity to destabilization, public finance distress, social security system and social assistance, living standards and demographic structure These national conditions seem to affect the financial decisions of households, thus the availability of stable deposits for MFIs Furthermore, this group is characterized by discipline in reporting, including transparency and uniformity of the data, which are available in the databases of the European Central Bank and the Eurostat These attributes enable to conduct comparative analysis and inference of regularities occurring in their relatively homogeneous subgroups of countries

The paper is organized as follows: the first part presents regulatory approach to the problem of household deposits’ stability The second one describes the research methods and variables applied in the study The last part comprises of the results of empirical analysis on the diversity of household deposits and its determinants

Regulatory approach to the problem of household deposits’ stability

The package CRD IV / CRR have set the framework for the single supervisory regulations in terms of funding stability of credit institutions The process of the development of detailed

technical standards has not been completed yet The announced solutions relate only to the

Liquidity Coverage Ratio, defining adequate liquidity of entities during 30-day period,

assuming a scenario combining idiosyncratic and market-wide stress It emphasizes the stability of household deposits by assigning them lower outflow rates than other liabilities of credit institutions The second liquidity standard - Net Stable Funding Ratio - for which the detailed guidance has not been announced yet, imposes the obligation on credit institutions of having adequate and stable funding structure in the long term Its important element are retail deposits which quality is made dependent, as in LCR, on the conditions of their placement

and the relationship between credit institution and depositor9

8 The study does not include Lithuania and Latvia, due to their adoption of a common currency after 2012

(4)

4

According to the decision of the European Commission on the LCR10, the “outflow rate” for

stable household deposits in the EU is established at the level of 5%, with the possibility of its

reduction to 3% from 1st January 2019 However, it refers to the funds covered by deposit

guarantee schemes that meet additionally one of the following:

 They constitute a part of established relationship with clients making withdrawals highly

unlikely i.e.: depositor has a contractual relationship with the credit institution of at least 12 months duration; depositor has a borrowing relationship with the credit institution for residential loans or other long term loans; depositor has at least one active product, other than loan, with the credit institution;

 They are held on transactional accounts

The future application of the underestimated “outflow rate”, is made dependent on the decision of the European Commission and the quality of the national deposit guarantee scheme The latter must be characterized by: the features described in Article 10 of Directive

2014/49/EU11; ready access to additional funding (from public and private sources) in the

event of a large call on its reserves; seven working day repayment period as referred to in Article 8(1) of Directive 2014/49/EU The remaining stable household deposits, including those covered by the guarantee schemes, but not satisfying the additional condition, are attributed with the outflow rate of 10% The premise of its application is abandonment of recognition of the characteristics responsible for significant variation

The scales of outflows for sensitive household deposits are defined in the rates ranging from

10% to 20%12 Despite their higher levels, they stand out against the outflow assigned to other

liabilities of credit institutions According to the EC, the conditions significantly limiting the stability of household deposits are as follows:

 the sum of all client’s deposits in credit institution exceeds EUR 500 000;

 the deposit is an internet account only;

 the deposit offers an interest rate that fulfils any of the following conditions:

o the rate significantly exceeds the average rate for similar retail products,

10

See Article 24 of Commission Delegated Regulation (EU) 2015/61 of 10 October 2014 to supplement Regulation (EU) No 575/2013 of the European Parliament and the Council with regard to liquidity coverage requirement for Credit Institutions, OJ L 11 of 17 January 2015

11 Directive 2014/49/EU of the European Parliament and of the Council of 16 April 2014 on deposit guarantee

schemes, OJ L 173 of 12 June 2014

12 See Article 25 of Commission Delegated Regulation (EU) 2015/61 of 10 October 2014 to supplement

(5)

5

o its return is derived from the return on a market index or set of indices,

o its return is derived from any market variable other than a floating interest rate;

 the deposit was originally placed as fixed-term with an expiry date maturing the 30

calendar day period or the deposit presents a fixed notice period shorter than 30 calendar days, in accordance with contractual arrangements;

 depositor is not a resident of the European Union; currency - other than the euro or the

domestic currencies of the member states

In case of the characteristic indicated in point or two features from points 2-5, the outflow rate for deposit is assumed to range from 10% to 15% However, if the deposit corresponds to point and additionally to at least one of the features from points 2-5, or it would have at least three features from points 1-5, the stated outflow rate is from 15% to 20% The same range is adopted for deposits of non-recognition features The highest outflow rate of 100% is established for cancellable deposits with a residual maturity of less than 30 calendar days and where pay-out has been agreed to another credit institution

Comparing the above rates with those characterizing other debt sources of credit institutions, it can be concluded that the new regulatory environment highlights the household deposits in terms of their positive impact on the safety of the banking systems Thus, an easy access to them shall be crucial for credit institutions’ stability during future crises

Data and metodology

The analysis of the stability of household deposits in the Eurozone member states in the years 2006-2012 was carried out on aggregated variables for each country The set of data contained

information from the ECB’s database13 about household deposits and MFIs’ balance sheet

totals In addition, the study implemented the Eurostat data14 on the condition of the financial

market and national economies, as well as socio-economic characteristics of households These above sets served to carry out following research:

 cross-sectional in selected years, characterizing three periods: pre-crisis (2006), banking

crisis (2008), and sovereign debt crisis accompanied by economic recession (2012) It related to the entire set of countries (the country was a statistical unit) The aim of the study was to identify the common characteristics of the Eurozone member states in terms

13

http://sdw.ecb.europa.eu/browse.do?node=2116074

(6)

6

of the development of particular types of deposits per person and their classification on internally homogeneous subsets;

 dynamic, relating to the formation of household deposits in selected countries

(representatives of specified sub-groups and individual member states considered to be interesting for the aim of this study) in the years 2006-2012 The statistical unit was the unit of time – a quarter

The set of empirical data allowed to form groups of variables for each member state,

characterizing its: household deposits, households, economic situation and financial market Some of them were expressed per person to eliminate the impact of significant differences in population sizes

The variables describing household deposits included: total deposits, total deposits per capita, deposits redeemable at notice up to months, deposits redeemable at notice up to months per capita, deposits redeemable at notice over months, deposits redeemable at notice over months per capita, deposits with agreed maturity up to years, deposits with agreed maturity up to years per capita, deposits with agreed maturity over years, deposits with agreed maturity over years per capita, overnight deposits, overnight deposits per capita

Numerical information on households was provided by the following variables: average size of a household, total household consumption expenditure per capita, household consumption expenditure on durable goods per capita, household consumption expenditure on semi-durable goods per capita, household consumption expenditure on non-durable goods per capita, household consumption expenditure on services per capita, household debts from loans per capita, households at risk of poverty or social exclusion (% of population), average net income, intention of buying a car over the next 12 months, intention of buying or building a house within the next 12 months, intention of renovating a house/flat during the next 12

months, saving rate15, population of the country

15 Savings rate is defined as the value of gross savings divided by gross disposable income, with the latter being

(7)

7

The economic situation of individual countries was characterized by: net saving16 per capita,

unemployment rate, employment rate, rate of inflation (HICP), GDP per capita, general government debt / GDP

Domestic part of the single financial market was described by means of: MFIs’ average

interest rate for ON deposits, MFIs’ assets per capita, share price indices17

, long term government bond yields.

For the purposes of the dynamic analysis, the following dummy variables identified

sub-periods: P0: IQ 2006-IV Q 2006 (a period of relative development of the financial market); P1:

IQ 2007-IIIQ 2007 (pre-crisis period); P2: IVQ 2007 – IIIQ 2008 (the crisis in the USA, until

the collapse of Lehman Brothers Holdings Inc.); P3: IVQ 2008 - IVQ 2010 (the crisis after the

collapse of Lehman Brothers Holdings Inc., until the implementation of new limit of deposits

covered by guarantee schemes - EUR 100 000), P4: IQ 2011 - IVQ 2012 (the sovereign debt

crisis and growing economic distress)

The Ward's method was used for the identification of similarities and differences among the countries in the values and structure of household deposits per capita in selected years, while the application of regression models allowed to indicate the variables responsible for the variability of household deposits in individual member states and the disparities in their values across the Eurozone

Regression models were applied to analyze the impact of the phases of financial crisis on the overall values of household deposits in selected member states The attempt to determine differences in the development of the total household deposits under financial turmoil was

carried out according to the following dynamic models18, implementing separate sub-periods:

P1, P2, P3, P4 as explanatory variables:

 linear:

(1)

16Net saving measures the portion of gross national disposable income that is not used for final consumption

expenditure, but takes into account the consumption of fixed capital

17This index reflects changes in the market prices of the shares that it represents

18G S Maddala (2009) Introduction to Econometrics West Sussex: Wiley&Sons Ltd.; J Pociecha, B Podolec,

(8)

8

 exponential:

(2) which, after logarithmic transformation gave:

(3)

where:

– dummy variables expressing belonging of the i-th quarter to the j-th sub-period (j = 1,2,3,4) Pij =1 when i-th quarter belongs to j-th sub-period, otherwise Pij =0 The basis of

comparison is sub-period P0;

– values of total HH deposits observed in the i-th quarter (i = 1,2, , 24); – residual

Regression models were also used in the analysis identifying specific variables from groups characterizing: households, economic situation and domestic conditions of the financial market, which statistically significantly used to affect the values of household deposits per capita in the Eurozone, in the years: 2006, 2008 and 2012 There were tested linear and exponential models

The best results in statistical sense gave linear models in two variants:

 variant A – simple regression model for particular type of household deposits per capita:

(4)

where:

yi –value of analyzed type of household deposits per capita in i-th member state,

xi - value of selected explanatory variable in i-th country, ɛi – residual;

 variant B – multiple regression model for specific type of deposit per capita (stepwise

regression determines the input of explanatory variables):

(5)

where:

xij - value of j-th explanatory variable in i-th country (i=1,2,…,17; j=1,2,…,k).

(9)

9

The last banking crisis revealed the lack of single sensitivity for analyzed types of household deposits in the group od countries Awareness of this phenomenon accompanied by new liquidity standards can contribute to the interest of credit institution in their particular categories As indicated in the introduction, the EBA has identified ON deposits as the most stable retail product type during stress period, which were followed by saving deposits (redeemable at notice) and finally - term deposits (with agreed maturity) The availability of these types remained geographically diverse in the group of surveyed countries Acting on the single financial market – free from barriers to cross-border activities - credit institutions may, therefore, define their preferences for targeted areas of deposit rising In order to assess the reliability of such a scenario, the study paid attention to the formation and structure of household deposits in the Eurozone countries during the recent financial crisis It should be noted that the information on household deposits in the ECB database was aggregated, so the opportunity to identify above-average volatile deposit categories (of high values, easy to withdrawn within a 30-day period, with unlimited access, and specific interest rates, country of origin of depositor or currency) was limited

 Average values and structure of household deposits in individual countries

In the years 2006-2012, the Eurozone member states remained geographically diverse both in terms of average values and structure of total household deposits placed with MFIs, as well as transfers of funds between short-term and long-term accounts (Annex 1) The cases of the

dominant share of ON deposits19 in total deposits located in MFIs were recorded for the

following countries: Finland, Italy and Luxembourg (since the second half of 2009) Their significant fractions were also noted for: Ireland and Estonia The largest share of deposits with agreed maturity characterized the sectors of MIFs in: Austria, Spain (since the second half of 2007), Greece (since 2008), Luxembourg (until the first half of 2009), Portugal, Slovenia, Slovakia, Cyprus and Malta In the Netherlands and Belgium, deposits redeemable up to months proved to be a dominant category In the majority of countries (Austria, Estonia, Finland, Spain, Greece, Ireland, Luxembourg, Portugal, Slovenia, Slovakia, Cyprus and Malta), ON deposits coexisted with deposits with agreed maturity, representing together almost all household sums located in MFIs In the cases of Italy, Belgium and the Netherlands, deposits redeemable at notice proved to be complementary to ON deposits The

(10)

10

greatest diversification in deposit types was observed in France and Germany, where three main categories - ON, with agreed maturity and redeemable at notice – remained complementary

On the basis of simultaneous and multidirectional changes in the sums located by households in monetary financial institutions, there was inferred funds’ transfer between accounts, and substitutability of some deposit categories (Annex 1) During the evolving instability in 2008, this phenomenon was evident between ON deposits and deposits with agreed maturity located in Spain, Finland, Greece, Ireland and Luxembourg In the Netherlands and Belgium this replacement was observed for deposits with agreed maturity and redeemable at notice In Germany, this fact was due to funds’ transfer from ON deposits and deposits redeemable at notice to deposits with agreed maturity

The group of 17 countries have not remained uniform in terms of the direction of changes in the total values of deposits The attention was drawn to the downward trend of analyzed variable for Greece (2010-2012), Luxembourg (2008-2010), Ireland (2010-2012), Spain (2011-2012) and Malta (2008-2009) In other countries, the development of the crisis, however, was accompanied by an increase in the aggregate value of household deposits

In the group of member states, one may indicate years of the highest growth in the total household deposits, however, in individual countries they were recognized at various stages of financial crisis Around 2008, such tendency was observed for the MFIs in: Austria, Germany, Finland, Portugal, Slovakia and Spain In Italy, this phenomenon occurred earlier, in 2007, while in Greece and Ireland later - in 2009 These differences encouraged to analyze the significance of the various phases of the financial crisis for the total levels of household deposits located in individual countries The results are presented further in this paper

 Populations' ability to possess deposits

(11)

11

represented total deposits per capita (Figure 1) The differences demonstrated at the levels of deposits per capita showed a lack of single capacity of the populations to provide stable funding to domestic credit institutions The highest, significantly deviating from the rest of the values, characterized Luxembourg In the group of countries, such as: Austria, Belgium, Cyprus, the Netherlands and Germany, these values stood close to EUR 20 000 The lowest levels of deposits per capita were recorded for post-communist countries: Slovakia, Slovenia and Estonia

The structure of deposits per capita in selected years, showed the diversity of household preferences, in terms of the placement rules in MFIs Recalling the EBA’s statement that during the banking crisis ON deposits and deposits redeemable at notice were more stable than those with agreed maturity, it can be concluded that credit institutions in Belgium, the Netherlands, France or Italy were provided with easier access to funds less vulnerable to shocks than those in Austria, Cyprus, Spain, Greece, Malta, Portugal or Slovenia The comfort of financing based on household deposits was thus in the Eurozone unequal and determined locally

Figure The values (in EUR) and structure of total household deposits per capita in individual Eurozone countries, in the years 2006, 2008 and 2012

(12)

12

The above findings have raised the question: Can anyone indicate any similarities, in terms of

the values and structure of total deposits per capita, in such an individualized group of countries? Trying to answer, I grouped the countries according to the Ward’s method

Comparing the tree diagrams for the years: 2006, 2008 and 2012 (Figures 2-4), despite the changes in the economic situation and the financial market, there were clearly seen similarities between the following member states: Belgium and the Netherlands; Slovakia and Slovenia Furthermore, Germany and France remained in one subgroup in all analyzed years In 2006, they formed one subset with Belgium and the Netherlands, but in 2008 they were accompanied by Italy, and in 2012 by Finland It could therefore be concluded that, on the one hand, there were visible targeted changes in household deposits per capita in some countries, on the other hand instability of composition of the subsets proved progressive structural transformation of the Eurozone in terms of the formation of analyzed variable as well as continuing dissimilarity of Luxembourg It seemed particularly important to identify countries changing affiliation to subsets They were characterized by MFIs with instable access to household deposits This could effectively hinder entities’ refinancing in time of evolving problems on the financial market and in economies

Figure Tree diagram for the EA member states, due to the level of total household deposits per capita located in MFIs in 2006

Source: Own calculations based on ECB’s data

(13)

13 Source: Own calculations based on ECB’s data

Figure Tree diagram for the EA member states, due to the level of total household deposits per capita located in MFIs in 2012

Source: Own calculations based on ECB’s data

 The impact of the stages of financial crisis on HH deposits

(14)

14

Due to the instability of the financial market in the years 2006-2012, I made an attempt to assess the impact of the various stages of the crisis on the levels of household deposits with MFIs For the safety of credit institutions it is important to recognize the sub-periods with the greatest modifications of the analysed variable Moreover, I raised the following question:

Was the significance of individual stages observed at national level, or was it common throughout the Eurozone? On the basis of the results obtained from the Ward’s method I

indicated the countries, which were the representatives of selected subgroups (Germany, the

Netherlands and Slovakia), as well as those particularly difficult(Spain, Portugal and Greece)

and keeping distinct position (Luxembourg) A study was carried out for the above member states, using regression models (1-3) The results are presented in Annex

For Germany, Spain and Slovakia, the impact of evolving instability on dependent variable could be traced In the case of Germany the best fit (in statistical sense) of the model was obtained for its exponential form It explained 94% of the variation in the total value of

household deposits in the years 2006-2012 All sub-periods (P1-P4) proved to be statistically

significant, but the principal difference in the dependent variable was recorded in the sub-period P4 It was higher on average by 25.9%20 than the value of deposits in a period of relative stability, which was the basis of comparison Period P3 with its events had slighter impact on the level of deposits It began with the collapse of Lehmann Brothers Inc and ended with the official deadline for revising the conditions of deposit guarantees under national schemes21 The order of the impact of time variables led to the conclusion that the evolving instability in economic and financial background accounted to an important incentive for households in Germany to accumulate surplus funds in the MFIs In the case of Spain, all time variables proved to be statistically significant with an effect on the dependent variable During the sovereign debt crisis, the total value of deposits in the MFIs was higher on average by EUR 212 198.1 million from the level in the period of relative growth This model allowed to explain 95% of the variability of the analyzed phenomenon Examining the situation in Slovakia, the linear model explained 97% of the variability in total deposits and also pointed the significance of all time variables

20

21 Laid down in: Directive 2009/14/EC of the European Parliament and of the Council of 11 March 2009

(15)

15

In Luxembourg, the MFIs reported the biggest changes in total deposits during the sub-period

P2, in relation to the basis of comparison Then the dependent variable reached the value

higher on average by over EUR billion than in the time preceding destabilization The last two sub-periods affected to a similar extent the value of the variable, which was higher on

average by more than EUR billion then in sub-period P0 The equation explained 73% of the

changes in analyzed value

In the cases of the Netherlands and Portugal, the importance and the order of sub-periods P2

-P4 were duplicated In the first one, the fluctuation of the total value of household deposits corresponded to the magnitude of difficulties on the financial market Phase preceding the crisis in the Eurozone and characterized by symptoms of impending imbalances proved to be statistically insignificant The equation explained 93% of the variation in total deposits In Portugal, the impact of the situation on the financial market and the economy also revealed

during P2 and intensified in the subsequent sub-periods The total value of household deposits

during P2 was higher on average by approximately EUR 12 billion than in the period of

comparison, in P3 by nearly EUR 24 billion, while in P4 by about EUR 36 billion Similarly to

the equation for the Netherlands, the model explained 95% of the variability in household deposits

In Greece, since 2009 the financial and economic crises have led to the decline of the total value of household deposits in MFIs, and thus to the ongoing reduction of deposit base in domestic credit institutions Disclosure of the critical situation of general government sector, applying for international financial aid, inability to stabilize locally the financial market and restricted household access to sums located on bank accounts resulted in both: a loss of confidence in MFIs, as well as the necessity of consumption of previously accumulated savings The linear regression model allowed to explain 84% of the variability of deposits

with two statistically significant explanatory variables - P3 and P2 During the banking crisis

(16)

16

The above observations did not allow to confirm a single impact of the phases of the financial crisis on the values of household deposits in analyzed countries, thus across the Eurozone Their sensitivity to the environmental changes therefore has developed individually

4 The determinants of the deposit heterogeneity in the Eurozone

Since the sensitivity of household deposits in the Eurozone countries proved to be heterogeneous, I made an attempt to explain its causes The aim of this part of the study was to identify economic, financial and socio-economic factors, which significantly influenced the level of analyzed variable The study was conducted for selected years: 2006, 2008 and 2012, characterizing different conditions of the financial market and national economies The variables from the groups of: household deposits, households, economic situation of the country and local financial market, expressed average values of features in analyzed countries Some of them were converted into “per capita” to eliminate the impact of different population sizes

Some of the variables were found to be statistically correlated with the values of household deposits per capita (Table 1) For the aim of this research, it was important to identify linkages between selected types of deposits per capita: with agreed maturity up to years, and ON The relationship between ON deposits per capita and total deposits per capita in all years remained positive and close to one It should be noted that for the last year, the correlation coefficient for deposits with agreed maturity up to years per capita and total deposits per capita was significantly lower (0.63) than for previous years In addition, Table shows that these three categories of deposits remained correlated to a group of variables describing the economic situation of the countries (GDP per capita, net saving per capita), the living conditions (average net income, saving rate, average household size) and financial market (MFIs’ average interest rates for ON deposits, share price indices and MFIs’ assets per capita)

Table Pearson’s correlation coefficients for household deposits per capita (by type) and selected variables, in the years 2006, 2008 and 2012

HH deposits per capita Tot

dep p.c

Dep red up to

3M p.c

Dep red over 3M p.c

Dep with agr mat up to

Dep with agr mat over

(17)

17

2 Y p.c

2 Y p.c Total deposits per capita

(Tot dep p.c.)

2006 2008 2012 1.000 1.000 1.000 - - - - - - 0.947 0.967 0.632 - - - 0.944 0.959 0.921 Deposits redeemable

up to 3M per capita (Dep red up to 3M p.c.)

2006 2008 2012 - - - 1.000 1.000 1.000 - - - - - - - - - - - -Deposits redeemable

over 3M per capita (Dep red over 3M p.c.)

2006 2008 2012 - - - - - - 1.000 1.000 1.000 - - - - - - - - -Deposits with agreed maturity up to Y

per capita

(Dep with agr mat up to Y p.c.)

2006 2008 2012 0.947 0.967 0.632 - - - - - - 1.000 1.000 1.000 - - - 0.921 0.935 -Deposits with agreed maturity over Y

per capita

(Dep with agr mat over Y p.c.)

2006 2008 2012 - - - - - - - - - - - - 1.000 1.000 1.000 - - -ON deposits per capita

(ON p.c.) 2006 2008 2012 0.944 0.959 0.921 - - - - - - 0.921 0.935 - - - - 1.000 1.000 1.000 GDP per capita

2006 2008 2012 0.932 0.940 0.928 - - - - - - 0.817 0.852 - - - - 0.920 0.929 0.929 Average net income

2006 2008 2012 0.790 0.748 0.799 - - - - - - 0.625 0.606 - - - - 0.732 0.705 0.691 MFIs’ average interest rates for ON deposits

2006 2008 2012 0.685 0.631 - - - 0.537 - 0.616 0.677 0.614 0.741 - - - 0.572 0.599 -Share price indices

2006 2008 2012 0.558 - - - - - - - - 0.560 - 0.612 - - - 0.546 - -MIFs’ assets per capita

2006 2008 2012 0.959 0.957 0.927 - - - - - - 0.925 0.929 - - - - 0.976 0.976 0.984 Saving rate 2006 2008 2012 - - 0.529 - - - - - - - - - - - - - - -Net saving per capita

2006 2008 2012 0.877 0.887 - - - - - - 0.735 0.775 - - - 0.866 0.877 0.862 Average size of household

2006 2008 2012 - - - - 0.497 0.504 - - - - - - - - - - - -(-) statistically insignificant values

Source: Own calculations based on ECB’s and Eurostat’s data

The above results encouraged me to use econometric model, which explained the geographical differentiation of total household deposits per capita in the Eurozone, in the years: 2006, 2008 and 2012 Some independent variables were correlated with each other, causing their reduction and the implementation of different model variants - A and B (4-5) The results from the single regression model are presented in Annex 3, in the order of the

(18)

18

explanatory variable on the total values of household deposits per capita was confirmed (p-level) In the case of multiple regression model (Annex 4), the explanatory variables were ordered according to the strength of their impact on the dependent variable

Total deposits per capita:

Simple linear regression model (4) indicated the significant influence of the following variables:

• MFIs’ assets per capita,

• GDP per capita,

• net saving per capita,

• ON deposits per capita,

• deposits with agreed maturity up to years per capita,

• average net income,

on the levels of total deposits per capita in the Eurozone, in the years: 2006, 2008 and 2012 (Annex 3) Individual regression equations explained from 40% to 93% of the total geographical diversification of the dependent variable Strong impacts of ON deposits per capita and deposits with agreed maturity up to years per capita were due to their high share in the total deposits per capita in all three years The influence of this second category was highlighted, in particular, in equations for the years: 2006 and 2008 For 2012, it became reduced due to the dynamic growth of total deposits, caused by increased interest of the Eurozone residents in shorter-term deposits The results also confirmed the importance of the condition of financial market (measured by MFIs’ assets per capita and MFIs’ average interest rate for ON deposits - 2006), and the economic condition of the countries (defined as GDP per capita, net saving per capita) for the formation of the dependent variable In addition, household characteristics, such as average net income, significantly affected the geographical diversification of deposits per capita in the Eurozone

(19)

19

variables: deposits with agreed maturity up to years per capita and GDP per capita In total, it explained 98% of the geographical differentiation in total household deposits per capita The parameter b* (for standardized variables) indicated the dominant role of deposits The regression equation for 2012 recognized the importance of the same variables as the previous one, but with reverse influence (GDP per capita proved to be decisive) The equation interpreted 95.7% of the diversity of total deposits per capita Statistical criteria (high significance of the estimated structural parameters and the parameters of stochastic structures) indicate that it thoroughly explained the formation of the dependent variable across the Eurozone

Concluding, the models allowed to identify a group of variables, positively affecting the values of total deposits per capita in the Eurozone Higher levels of the regressand characterized those member states that distinguished in the studied periods with relatively higher level of economic and financial development, but also living standards This, therefore, denies a popular theorem that in developed countries, households focus on other forms of savings and marginalize their deposits in banks Strong interaction of deposits with agreed maturity up to years per capita and ON deposits per capita resulted from their dominant position in the structure of total deposits per capita The variant A of the regression model showed that before the financial crisis, the dependent variable was influenced by MFIs’ average interest rate for ON deposits The appearance and intensification of financial and economic destabilization, contributed to the loss of importance of returns on investments and drew households’ attention to other feature - safety

ON deposits per capita:

In the years: 2006, 2008 and 2012, the value of overnight deposits per capita in the Eurozone remained under the influence of the following variables (variant A):

 MFIs’ assets per capita,

 total deposits per capita,

 GDP per capita,

 net saving per capita,

 average net income

(20)

20

explanatory variables significantly affected the level of ON deposits per capita, explaining between 36% and 97% of its differentiation in the analyzed group of countries

In the years 2006 and 2008, geographically different levels of deposits remained under the influence of the values of deposits with agreed maturity up to years per capita In 2012, the changes in economic condition of the countries and investment preferences of the populations, however, resulted in a lack of correlation between these variables During the banking crisis in

2008, the ON deposits per capita also remained under the influence ofMFIs’ average interest

rates, but the intercept of the regression equation turned out to be statistically insignificant This is the only year from the three analyzed, in which households considered the profitability of investments as a decisive factor for the type of deposits held, despite their short-term nature In the equations for the years: 2006 and 2008, in which the explanatory variables were: total deposits, deposits with agreed maturity up to years and average net income, intercepts remaind statistically insignificant, thus limited their informational value In the equation for 2012, the impact of deposits with agreed maturity up to years on the formation of ON deposits per capita was not proved In other equations for the last year all parameters were found to be statistically significant The above-mentioned explanatory variables positively influenced ON deposits per capita in the Eurozone, in the years 2006, 2008 and 2012 This means that, among others, the maturity of the MFI sectors, measured by the values of their assets fostered the growth of these short-term deposits The economic situation of the countries also constituted the factor positively affecting the willingness of individual customers to keep a part of their incomes in the form of liquid deposits

The variant B of the linear model allowed to explain more than 97% of geographic diversity of ON deposits per capita during the banking crisis (Annex 4) Dominant influence was found in MFIs’ assets per capita, with additional impact of MFIs’ average interest rates for ON deposits The trials to construct a multiple regression model for ON deposits in the years: 2006 and 2012 proved a failure

Deposits with agreed maturity up to years per capita:

In 2006 and 2008, the average value of deposits with agreed maturity up to years per capita in the Eurozone remained under the strong influence of:

 total deposits per capita,

(21)

21

 MFIs’ assets per capita,

 ON deposits per capita,

 net saving per capita

The regression equations explained from 54% to 93% of the differentiation of deposit values across the Eurozone (Annex 3) They pointed out a significant impact of domestic economic and financial conditions, in addition to the impact of selected categories of deposits on the dependent variable The results for 2012 confirmed the importance of the MFIs’ ON interest rates and the situation on the major stock exchanges These equations explained respectively 55% and 44% of the analyzed geographical differentiation of deposits’ level It should be noted that the value of the dependent variable was under a negative influence of the prevailing situation on the capital market in 2012 Visible signs of recovery in a part of the countries favored redirecting there sums from bank accounts The construction of regression model in variant B turned out to be unsuccessful in statistical sense

Other categories of household deposits:

The study was repeated for other, less important categories of household deposits in the Eurozone It revealed only a weak interaction between the dependent variable and the explanatory variables listed in Annex Noteworthy were models describing the formation of deposits redeemable up to months per capita and deposits redeemable over months per capita These two categories were assessed by the EBA as moderately stable under stress In 2006 and 2012, the latter was affected by MFIs’ ON interest rates, but the coefficients of determination were only 29% and 38% In 2008 and 2012, the deposits redeemable up to months per capita were under the influence of an average size of the household in individual member states In both cases the impact of the explanatory variable proved to be negative and relatively poor (R-squared = 25%)

(22)

22

Conclusions

The aim of this paper was to identify the determinants of household deposit stability in the Eurozone during the last financial crisis, on the background of new regulatory environment The results of empirical analysis proved all hypothesis statements specified in this paper (see: Introduction)

For the Eurozone credit institutions, the access to stable funding such as household deposits, determines the ability to meet the regulatory criteria for both short- and long-term liquidity As presented in this paper, during the period of instability, these funds remained varied across the Euro area in terms of their values and structure Thus their availability for MFIs was not homogenous in the region In most countries, the crisis was a kind of incentive to accumulate deposits However, there were cases where, despite the strengthening of guarantee schemes and the lack of alternative investment opportunities, households have begun to reduce their receivables from MFIs (eg Luxembourg - up to 2009, Greece, Ireland, Spain and Malta) The paper also indicates the countries, which were characterized by the domination of ON deposits - of the highest stability under the stress (Finland, Italy, Luxembourg – from 2009), and those in which MFI sectors were dominated by more volatile ones - deposits with agreed maturity (Austria, Spain, Cyprus, Greece, Luxembourg – until 2009, Portugal, Malta, Slovakia, Slovenia)

(23)

23

The phases of the crisis affected with varying degrees the total value of household deposits in individual countries – representatives of selected subgroups and those particularly interesting because of their financial and economic problems In some of them (eg Germany, Spain and Slovakia) evolving destabilization motivated people to locate increasing sums in the MFIs, while in others the response to the changing environment was late (in the Netherlands, Portugal), or disclosed only in selected sub-periods (in Greece) Luxembourg was an example of the state where the strongest change in the value of household deposits appeared not in the last, but in earlier period (identifying the financial crisis in the US up to the collapse of Lehman Brothers Holdings Inc.) All above led to the conclusion that surveyed countries were characterized by individual sensitivity to the changes on financial market and in national economies, making the conclusion about single tendency in the Eurozone impossible, in terms of the availability of household deposits to MFIs in the years 2006-2012

The disparities in economic development of countries and their MFI sectors, as well as populations’ wealth and propensity to save proved to be responsible for the geographical diversification of household deposits per capita They influenced the formation of two the most important categories of deposits per capita in the Eurozone (ON and with maturity up to years), in terms of their average values, and thus average values of total household deposits in the years: 2006, 2008 and 2012 For deposits redeemable at notice per capita, one of important factors was referring to the socio-economic characteristics of households – an average size of a household

Concluding, the described heterogeneity of deposit nature may become a significant problem for credit institutions in a part of the Eurozone, caused by the implementation of the single funding stability regulations According to the new standards, household deposits will positively influence the safety of entities as well as their further development The study did not lead to the results confirming equal access of the Eurozone’s MIFs to this sources of funding in the period of instability, nor its equal quality in the whole group of countries The value and structure of deposits developed in a clearly individual way This in turn may contribute in the future to the entities’ geographically targeted preferences within the regions providing household deposits

(24)

24

The presented results are a part of the research project funded by the National Science Centre in Poland within the grant "Sonata 6", No UMO-2013/11 / D / HS4 / 04056

Bibliography:

Commission Delegated Regulation (EU) 2015/61 of 10 October 2014 to supplement Regulation (EU) No 575/2013 of the European Parliament and the Council with regard to liquidity coverage requirement for Credit Institutions, OJ L 11 of 17 January 2015;

Directive 2009/14/EC of the European Parliament and of the Council of 11 March 2009 amending Directive 94/19/EC on deposit-guarantee schemes as regards the coverage level and the payout delay, OJ L68 of 13 March 2009;

Directive 2013/36/EU of the European Parliament and of the Council of 26 June 2013 in access to the activity of credit institutions and the prudential supervision of credit institutions and investment firms, amending Directive 2002/87/EC and repealing Directives 2006/48/EC and 2006/49/EC, OJ L 176 of 27 June.2013;

Directive 2014/49/EU of the European Parliament and of the Council of 16 April 2014 on deposit guarantee schemes, OJ L 173 of 12 June 2014;

EBA (2013) Discussion Paper on retail deposits is higher outflows for the purposes of

liquidity reporting under the draft Capital Requirement Regulation (CRR), London:

EBA/DP/2013/02;

Maddala G.S (2009) Introduction to Econometrics West Sussex: Wiley&Sons Ltd.;

Pociecha J., Podolec B., Sokolowski A., Zajac K.(1988) Taxonomic methods in

socio-economic studies Warszawa: PWN;

Regulation (EU) No 575/2013 of the European Parliament and of the Council of 26 June 2013 on prudential requirements on credit institutions and investment firms and amending Regulation (EU) No 648/2012, OJ L 176 of 27 June 2013;

(25)

25

Annex Total HH deposits and their division by type (EUR million) in the Eurozone

member states, in the years 2006-2012

(26)

26 2000 4000 6000 8000 10000 01 Jan u ar y 006 01 J an ua ry 2007 01 J an ua ry 2008 01 Jan u ar y 009 01 J an ua ry 2010 01 J an ua ry 2011 01 J an ua ry 2012 Malta 100000 200000 300000 400000 500000 01 J an ua ry 2006 01 Jan u ar y 007 01 J an ua ry 2008 01 J an ua ry 2009 01 J an ua ry 2010 01 J an ua ry 2011 01 J an ua ry 2012 Netherlands 50000 100000 150000 01 J an ua ry 2006 01 Jan u ar y 007 01 J an ua ry 2008 01 J an ua ry 2009 01 J an ua ry 2010 01 J an ua ry 2011 01 J an ua ry 2012 Portugal 5000 10000 15000 20000 25000 30000 01 J an u ar y 20 06 01 J an u ar y 20 07 01 J an u ar y 20 08 01 J an u ar y 20 09 01 J an u ar y 20 10 01 J an u ar y 20 11 01 J an u ar y 20 12 Slovakia

(27)

27

Annex The estimation of the parameters in exponential and linear regression models for total household deposits in selected Eurozone member states, in the period: IQ 2006 – IVQ

2012

Germany (exponential model)

Beta22 Std error B23 Std error t-Statistic p-value

Constant 14.14 0.01 1282.37 0.000000

P4 1.20 0.07 0.23 0.01 17.20 0.000000

P3 0.80 0.07 0.15 0.01 11.31 0.000000

P2 0.26 0.06 0.06 0.01 3.99 0.000573

P1 0.06 0.06 0.02 0.02 1.04 0.307990

F-statistic 104.85; R-squared 0.94; sɛ 0.02

Spain (linear model)

Beta Std error B Std error t-Statistic p- value

Constant 518253.2 9853.52 52.60 0.000000

P4 1.24 0.07 212198.1 12068.04 17.58 0.000000

P3 1.13 0.07 187285.6 11842.45 15.81 0.000000

P2 0.55 0.07 111114.6 13219.88 8.41 0.000000

P1 0.18 0.06 54966.8 17066.79 3.22 0.003787

Slovakia (linear model)

Beta Std error Std error t-Statistic p- value

Constant 14885.00 378.19 39.36 0.000000

P4 1.21 0.05 10432.00 463.19 22.52 0.000000

P3 0.95 0.05 7900.33 454.53 17.38 0.000000

P2 0.33 0.05 3309.00 507.40 6.52 0.000001

P1 0.09 0.04 1429.00 655.05 2.18 0.039621

Luxembourg (linear model)

Constant 39844.00 721.91 55.19 0.000000

P2 1.11 0.15 7295.40 968.55 7.53 0.000000

P4 0.96 0.16 5385.50 884.16 6.09 0.000003

P3 0.96 0.16 5186.00 867.63 5.98 0.000004

P1 0.37 0.13 3627.00 1250.39 2.90 0.008059

the Netherlands (linear model)

Constant 282141.20 4895.33 57.63 0.000000

P4 1.23 0.08 90463.40 5995.52 15.09 0.000000

P3 0.85 0.08 60909.30 5883.40 10.35 0.000000

P2 0.32 0.08 28044.40 6567.77 4.27 0.000287

Portugal (linear model)

Constant 93398.67 1335.44 69.94 0.000000

P4 1.19 0.06 36456.21 1766.62 20.64 0.000000

P3 0.83 0.06 24562.33 1724.04 14.25 0.000000

P2 0.33 0.05 11859.33 1980.78 5.99 0.000004

Greece (linear model)

Constant 134674.00 5011.98 26.87 0.000000

P3 1.10 0.13 52950.60 6023.65 8.79 0.000000

22

Standardized coefficients inform which of the independent variables have a greater effect on the dependent variable in a multiple regression analysis

23

(28)

28

P2 0.49 0.11 28587.00 6724.27 4.25 0.000301

F-statistic 29.23; R-squared 0.84; sɛ 10024.00 Source: Own calculations based on ECB’s data

Annex The estimation of the parameters in linear regression model (variant A) for selected

types of household deposits in the Eurozone, in the years: 2006, 2008 and 2012

B Std error t-Statistic p-value

Total HH deposits per capita - 2006

Constant

MFIs’ assets per capita

11116.31 0.03 1472.29 0.00 7.550 12.680 0.0000 0.0000 R-squared 0.92; S.E of regression 5462.3

Constant

ON HH deposits per capita

Constant

HH deposits with agreed mat up to Y per capita

Constant GDP per capita

-24366.70 371.90 4749.04 38.67 -5.131 9.617 0.0002 0.0000 R-squared 0.87; S.E of regression 6998.1

Constant

Net saving per capita

Constant

Average net income

Constant

MFIs’ average interest rate for ON deposits

Constant

Average net income

-17258.90 2.10 9292.21 0.49 -1.857 4.360 0.0830 0.0006 R-squared 0,56; S.E of regression 14916,0

Constant

(29)

29

ON HH deposits per capita 1.19 0.13 9.166 0.0000

R-squared 0.85; S.E of regression 7623.5 Constant

Constant

Average net income

Constant

ON HH deposits per capita - 2006

Constant

Total HH deposits per capita

Constant

Average net income

Constant

MIFs’ assets per capita

Constant

Average net income

Constant

(30)

30

Constant

Average net income

HH deposits with agreed maturity up to Y per capita - 2006

Constant

-1340.37 20848.20 2161.01 4875.02 -0.620 4.277 0.5443 0.0007 R-squared 0.55; S.E of regression 49071

Constant

Share price indices

7836.40 -2,4531.70 1396.69 7178.82 5.611 -3.417 0.0000 0.0038 R-squared 0.44; S.E of regression 5481.5

HH deposits redeemable up to M per capita - 2008

Constant

Average size of HH

21588.61 -7347.04 8399.20 3309.60 2.570 -2.220 0.0213 0.0423 R-squared 0.25; S.E of regression 3671.4

HH deposits redeemable up to M per capita - 2012

Constant

Average size of HH

31533.20 -11326.90 12352.54 5008.89 2.553 -2.261 0.0221 0.0390 R-squared 0.25; S.E of regression 5259.0

HH deposits redeemable over M per capita - 2006

Constant

(31)

31

R-squared 0.29; S.E of regression 306.60

HH deposits redeemable over M per capita - 2012

Constant

-141.43 777.44

113.90 256.94

-1.242 3.026

0.2339 0.0085 R-squared 0.38; S.E of regression 258.63

(32)

32

Annex The estimation of the parameters in linear regression model (variant B) for selected

types of household deposits in the Eurozone, in the years: 2006, 2008 and 2012

Beta Std error B Std error t-Statistic p-value Constant

HH deposits with agreed mat up to Y per capita Average net income

0,34 0,47 0,25

0,14 0,12 0,07

-914,35 0,01 0,86 0,68

2888,31 0,01 0,22 0,18

0,317 2,525 3,887 3,762

0,757011 0,026688 0,002161 0,002711 R-squared 0,98; S.E of regression =3319,1

Beta Std error B Std error t-Statistic p- value Constant

HH deposits with agreed mat up to Y per capita GDP per capita

0,61 0,42

0,07 0,07

-12344,40 1,00 209,90

2964,69 0,11 32,66

-4,165 9,260 6,427

0,0010 0,0000 0,0000 R-squared 0,98; S.E of regression =2986,2

GDP per capita

H deposits with agreed mat up to Y per capita

0,80 0,33

0,06 0,06

-20424,20 327,20 0,90

2675,15 24,23 0,16

-7,635 13,504 5,609

0,90 0,16

0,05 0,05

1602,59 0,01 2224,47

720,51 0,00 700,76

2,224 17,653 3,174

Định dạng
Số trang	32
Dung lượng	1,4 MB