interest rate pass-through - empirical results for the euro area

8/2/2019 Interest Rate Pass-Through - Empirical Results for the Euro Area

1/42

Interest Rate Pass-Through:

Empirical Results for theEuro Area

Gabe J. de BondtEuropean Central Bank

Abstract. This paper empirically examines the interest rate pass-through at theeuro area level. The focus is on the pass-through of official interest rates, approximatedby the overnight interest rate, to longer-term market interest rates, which, in turn, are a

proxy for the marginal costs for banks to attract deposits or grant loans, and thereforepassed through to retail bank interest rates. Empirical results, on the basis of a (vector)error-correction and vector autoregressive model, suggest that the pass-through ofofficial interest to market interest rates is complete for money market interest rates upto three months, but not for market interest rates with longer maturities. Furthermore,the immediate pass-through of changes in market interest rates to bank deposit andlending rates is found to be at most 50%, whereas the final pass-through is typicallyfound to be close to 100%, in particular for lending rates. Empirical results for a sub-sample starting in January 1999 show qualitatively similar findings and aresupportive of a quicker interest rate pass-through since the introduction of the euro.

It is shown that the difference between the adjustment speed of bank deposit andlending rates (typically around one versus three months since the common monetary

policy) can to a large extent significantly be explained by credit risk considerations.

JEL classification: E43; G21.

Keywords: Interest rate pass-through; overnight, market and bank interestrates; euro area.

1. INTRODUCTION

The interest rate pass-through process is from a price and financial stabilityperspective important for monetary policy. Central banks exert a dominantinfluence on money market conditions and thereby steer money marketinterest rates. Changes in money market interest rates, in turn, affect marketinterest rates with longer maturities and retail bank interest rates, albeit tovarying degrees. Bank decisions regarding the yields paid on their assets and

r Verein fur Socialpolitik and Blackwell Publishing Ltd. 2005, 9600 Garsington Road, Oxford OX4 2DQ, UKand 350 Main Street, Malden, MA 02148, USA.

German Economic Review 6(1): 3778


2/42

liabilities have an impact on the expenditure and investment behaviour ofdeposit holders and borrowers and thus real economic activity. In otherwords, a quicker and fuller pass-through of official and market interest rates

to retail bank interest rates strengthens monetary policy transmission andthus may affect price stability. Furthermore, prices set by banks influencetheir margins and therefore bank profitability and consequently thesoundness of the banking system and thus financial stability.

Many factors ultimately influence the adjustment of retail bank interestrates: marginal pricing costs, (expected) bank exposure to interest rate risk,credit and other risk premia, competition and regulation in differentsegments of the deposit and credit market, bankcustomer relations, theadministrative cost of effectively changing interest rates, the degree of passivebehaviour on the part of deposit holders and borrowers, etc. This study

particularly examines the first-mentioned determinant of the interest ratepass-through process at the euro area level. Banks marginal pricing costs areapproximated by interest rate movements in the money and capital markets,since they are the market-based sources for banks to subsequently attractdeposits or grant loans. Banks offer deposits and loans at rates which arecompetitive to those on money market mutual funds and non-bank sourcesof finance.

The main contribution of this paper is that the bank and market interestrates considered have a comparable maturity to avoid distortions from yieldcurve effects and to accurately capture the marginal pricing costs of banks.For the first time both bank deposit and lending rates at the level of the euroarea are analysed. The focus is on the euro area instead of country-specificexperiences in order to analyse the impact of the introduction of a commonmonetary policy in the euro area in January 1999. The shift from nationalcentral banks to the European Central Bank (ECB) may have affected banksbehaviour and therefore the interest rate pass-through process, as theoreti-cally shown by Bagliano et al. (2000). The ECB monetary policy reacts toconditions at the euro area level and not to country-specific developments,and monetary policy may affect banks incentive to collude in the creditmarket through its influence on the cost of raising funds. Therefore, the euroarea experience is examined using three different empirical frameworks for asample starting in January 1996, the first date for which publicly euro arearetail bank interest rate data are available, and a sub-sample starting withStage Three of EMU, i.e. January 1999.

The main lesson of this empirical study is threefold. First, a complete pass-through of official interest rates to money market interest rates up to threemonths is observed, suggesting that monetary policy fully controls moneymarket rates up to three months. The impact of a change in official interestrate to market interest rates declines with the maturity of the marketinstruments. No statistical significant relationship is found between theofficial interest rate and government bond yields, suggesting a crediblemonetary policy at the euro area level. Second, it is shown that euro area

G. J. de Bondt

38 r Verein fur Socialpolitik and Blackwell Publishing Ltd. 2005


3/42

banks adjust their rates on deposits and loans generally closely in line withmarket interest rate developments, but that this pass-through process differsbetween the different segments of the retail bank market in the euro area. The

immediate, i.e. within one month, adjustment of retail bank interest rates toa 100 basis point change in market interest rates is up to 50 basis points,whereas the final or long-term pass-through is close to 100 basis points, atleast for bank lending rates. The interest rate on overnight deposits, depositsredeemable at notice of up to three months and on consumer lending arecomparatively sticky with an immediate pass-through of around 5%. Thethird and final lesson, of a more speculative nature but supported by ourempirical findings, is that the interest rate pass-through process has changedsince the introduction of the euro. Official interest rates are for this sub-sample also immediately completely passed through to money market

interest rates up to three months. This suggests that the monetary policycontrol of the short-end of the yield curve at the euro area level hasstrengthened since January 1999, in line with the introduction of a commonmonetary policy in the euro area. For all retail bank interest rates the speed atwhich they adjust to market interest rate developments is found to havebecome quicker. The mean adjustment speed of retail bank rates to finallyadjust to market interest rate developments is, since January 1999, typicallyaround one month for deposit rates and around three months for lendingrates. The observed difference in the adjustment speed between deposit andlending rates can to a large extent be significantly explained by credit risk.Other likely explanations for the significantly quicker pass-through processsince the single currency are an increase in the degree of competition, asreflected in an increasing role for non-deposit or market-based funding byeuro area banks and a rise in the interest rate elasticity of the demand forretail bank products, and a decrease in asymmetric or switching costs indifferent segments of the euro area retail bank market.

The paper proceeds as follows. Section 2 briefly reviews interest rate pass-through studies for individual euro area countries. Section 3 presents atheoretical background and empirical specifications of the interest rate pass-through process. The interest rate pass-through is divided into the pass-through of changes in the official interest rate, approximated by the overnightinterest rate, to longer-term market interest rates, and of market interest rates,as a proxy for the marginal costs for banks to attract deposits or grant loans, toretail bank interest rates. Section 4 describes the euro area data on official andretail bank interest rates and presents a correlation analysis to detect the mostcomparable market interest rates for the bank interest rates considered. Section5 discusses the empirical results based on (i) a vector error-correction model(VECM), (ii) a vector autoregressive (VAR) model, and (iii) a univariate error-correction model (ECM). Section 6 follows with a presentation of the empiricalresults for a sub-sample, starting in January 1999. Section 7 summarizes withconcluding remarks. Appendices A and B present impulse responses and theirconfidence bands based on VAR models for interest rate pairs. Appendix C

Interest Rate Pass-Through: Empirical Results for the Euro Area

r Verein fur Socialpolitik and Blackwell Publishing Ltd. 2005 39


4/42

presents German results for the retail bank interest rate pass-through processto check the hypothesis of Bagliano et al. (2000) that less collusion amongGerman banks is a likely outcome of the shift in monetary policy from the

Bundesbank with a German focus to the ECB with a euro area-wide focus.

2. INTEREST RATE PASS-THROUGH STUDIES FORINDIVIDUAL EURO AREA COUNTRIES

While there is voluminous literature on monetary policy transmission, theretail bank interest rate pass-through process has, at least for several years,been surprisingly underexplored. Table 1 summarizes the main findings ofinterest rate pass-through studies performed for individual euro area

countries. All studies show cross-country differences in the interest ratepass-through without a clear pattern of these cross-country differences.Nevertheless, it seems to be the case that short-term bank lending rates toenterprises in Belgium, Spain and the Netherlands adjust less sluggishly afterthree months compared with the other euro area countries.

As regards bank lending rates, studies from the mid-1990s broadly showthat changes in official and/or money market rates are not fully reflected inshort-term bank lending rates to enterprises after three months, but that thepass-through is higher in the long term (BIS, 1994; Cottarelli and Kourelis,1994; Borio and Fritz, 1995). Recent cross-country studies by Kleimeier andSander (2000, 2002), Donnay and Degryse (2001) and Toolsema et al. (2001)confirm this finding. Mojon (2000) and Heinemann and Schuller (2002) alsofind short-term sluggishness in short-term bank lending rates to enterprises,but assume a priori a complete long-term pass-through. This short-termstickiness is also found for consumer credit rates (Heinemann and Schuller,2002; Kleimeier and Sander, 2002). For long-term bank lending rates allstudies, except BIS (1994) and Heinemann and Schuller (2002) which takelong-term market rates as costs of funds for long-term bank rates, typicallyshow that the pass-through tends to be less complete than for short-term banklending rates to enterprises. This finding may be driven by the fact that themarginal cost prices are approximated by money market interest rates whichmay not be the most appropriate marginal funding costs for long-term loans.

Turning to bank deposit rates, one study examines the adjustment ofdeposit rates to changes in the money market interest rate in individual euroarea countries (Mojon, 2000). The main finding is an incomplete short-termpass-through for deposit rates, notably for savings deposits, and thatderegulation has significantly affected the interest rate pass-through processfor deposits, but not for loans.

Several studies also examine the issue of an asymmetric interest rate pass-through process. The response of bank rates to changes in official rates and/ormoney market rates seems to depend in some cases on whether marketinterest rates are rising or falling (Borio and Fritz, 1995; Mojon, 2000;

G. J. de Bondt



5/42

Table 1 Interest rate pass-through studies for individual euro area countriesa (adju

100 basis points change in money market interest rates in basis points)

Study AT BE DE ES FI FR GR IE

Short-term loans to firmsBIS (1994) ST 85 18 78 15

LT 68 112 106 110 Cottarelli and Kourelis (1994) ST 67 87 78 23 61 107

LT 87 100 94 28 82 107 Borio and Fritz (1995) ST 95 36 100 53

LT 93 98 105 59 1Kleimeier and Sander (2000) ST

LT 110 97 107 195 72 117 101 1Mojon (2000) ST 100 36 55 71

LT 100 100 100 100 1Donnay and Degryse (2001) ST 15 85 66 102 43 36 20

LT 18 92 72 100 75 42 18 Toolsema et al. (2001) ST 76 72 103 53

LT 102 90 114 62 Heinemann and Schuller (2002) ST 44 83 13 75 45 3 1

LT 100 100 100 100 100 100 1Kleimeier and Sander (2002) ST 25 81 10 52 39 27

LT 74 85 63 84 96 64

Long-term loans to firms

Mojon (2000) ST 61 18 42 LT 100 100 100 Donnay and Degryse (2001) ST 21 69 87 23 25 17

LT 10 40 93 50 64 16

r

VereinfurSocialpolitikandBlackwellPublishing

Ltd.2005

41


6/42

Study AT BE DE ES FI FR GR IE

Consumer creditHeinemann and Schuller (2002) ST 60 98 23 46 LT 100 100 100 100

Kleimeier and Sander (2002) ST 15 79 17 69 53 1 LT 85 101 49 65 80 19

MortgagesBIS (1994) ST 48

LT 82 89 27 90 Mojon (2000) ST 5 45 11 41

LT 100 100 100 100 Donnay and Degryse (2001) ST 26 19 20 40 39 16

LT 32 48 44 14 61 6 1Heinemann and Schuller (2002) ST 14 107 99 35 56

LT 100 100 100 100 100 1Kleimeier and Sander (2002) ST 14 32 36 15 30 16 23

LT 72 57 31 66 77 29 57

Savings depositsMojon (2000) ST 27 9 13

LT 100 100 100

Time deposits

Mojon (2000) ST 94 82 15 LT 100 100 100 1

a ST5 short-term pass-through, that is adjustment after 3 months; LT5 long-term pass-through; euro area f

results using January 2001 country weighting structures as applied for euro area retail bank interest rates.Sources: BIS (1994, Table 5, 198493); Cottarelli and Kourelis (1994, Table 1, Model 2); Borio and Fritz (19Sander (2000, Table 5, 199498); Mojon (2000, Table 2a, 199298); Donnay and Degryse (2001, Table 3, 199

3, 19802000); Heinemann and Schuller (2002, Table 6, 199599); Kleimeier and Sander (2002, Table A4,

Table 1 Continued42

r

VereinfurSocialpolitikandBlackw

ellPublishingLtd.2005


7/42

Heinemann and Schuller, 2002) or whether bank interest rates are below orabove equilibrium levels as determined by cointegration relations (Kleimeierand Sander, 2000, 2002).1

Industrial organization-based literature examines the pricing behaviour ofbanks using bank data. The focus of this strand of the literature is typically onthe link between bank interest rate margins and the market structure of thebanking system (Hannan and Berger, 1991; Neumark and Sharpe, 1992;Angbazo, 1997; Hannan, 1997; Wong, 1997; Corvoisier and Gropp, 2002).The main lesson of these banking structure studies is that the pricingbehaviour of banks may depend on the degree of competition and contest-ability in the different segments of the retail bank market. For instance,Corvoisier and Gropp (2002) conclude that for demand deposits and loansincreasing bank concentration in individual euro area countries during the

years 199399 may have resulted in less competitive pricing by banks,whereas for savings and time deposits the opposite seems to be the case.

3. INTEREST RATE PASS-THROUGH MODEL

3.1. Theoretical background

Central banks exert a dominant influence on money market conditions andthereby steer the short-term money market interest rates, which, in turn, may

affect market interest rates with a longer maturity (first stage of the interestrate pass-through process). Changes in market interest rates, in turn, affectretail bank interest rates, albeit to varying degrees (second stage of the interestrate pass-through process). As regards this second stage, in the textbook worldof perfect competition with complete information prices equals marginalcosts and the derivative of prices with respect to marginal costs equals one.This derivative typically becomes less than one when the perfect competitionand information assumption are relaxed. Applying this idea to the pricesetting of banks results in the following marginal cost pricing model equation(Rousseas, 1985):2

br g0 g1mr 1

where bris the price set by banks, that is the bank interest rate, g0 is a constantmarkup and mr is the marginal cost price approximated by a comparablemarket interest rate. The underlying idea is that market interest rates are the

1. See Scholnick (1996) for an analysis of an asymmetric interest rate pass-through process inMalaysia and Singapore.

2. Another approach to model the interest rate pass-through, not followed in this paper, is along

the lines of the Klein (1971)Monti (1971) model or Tobin (1982) model and extensions ofthese models (Dermine, 1986). These studies particularly focus on the impact of capitalrequirements on bank pricing policy.




8/42

most appropriate marginal cost prices, because of their accurate reflection ofthe marginal funding costs faced by banks.

The coefficient g1 depends on the demand elasticity of deposits and loans

with respect to the retail bank interest rate. If the demand for deposits andloans is not fully elastic, parameter g1 is expected to be less than one. Depositdemand is expected to be relatively elastic with respect to the bank depositrate when close substitutes for deposits exist, for instance, money marketfunds. The loan demand elasticity depends, among other factors, on whetherborrowers have access to alternative sources of finance.

Parameter g1 will also be less than one if banks have some degree ofmarketpower. Retail bank interest rates in less competitive or oligopolistic segmentsof the retail bank market adjust incompletely and only with a delay, whilebank interest rates set in a fully competitive environment respond quickly

and completely (Laudadio, 1987). A wide range of factors influence marketpower. For instance, entry into the banking sector is restricted by regulatoryagencies, creating one of the preconditions for a degree of monopoly powerand administrated pricing (Niggle, 1987). Market power and an inelasticdemand for retail bank products may also result from the existence ofswitching costs and asymmetric information costs.

Switching costs may arise when bank customers consider switching fromone bank to another, for example when a household intends to transfer itssavings deposits from bank A to bank B. Switching costs, such as costs ofacquiring information and search and administrative costs, are potentiallyimportant in markets where significant information or transaction costs exist.They are also expected to be high in markets with long-term relationships andrepeated transactions (Sharpe, 1997; Kim et al., 2003).3 However, even in thepresence of small switching costs, the theory predicts that the smaller theproportion of customers that are new to the market, the less competitiveprices will be. Klemperer (1987) shows that generally the existence ofswitching costs results in market segmentation and reduces the demandelasticity. Even with non-cooperative behaviour, switching costs result in aretail bank interest rate adjustment of less than one to a change in the marketinterest rate (Lowe and Rohling, 1992).

As regards the setting of lending rates by banks, asymmetric information costsintroduce problems of adverse selection and moral hazard (Stiglitz and Weiss,1981). If banks increase their lending rates they may attract riskier borrowers(adverse selection) or the increase of lending rates will give adverse incentivesfor borrowers to choose riskier projects (moral hazard). In other words, banksexpected receipts may actually fall when they increase their lending rateseven if funding costs increase, if the probability of default rises sufficiently.

3. Kim et al. (2003) empirically show that for the Norwegian banking industry around a quarter

of the bank customers added value is due to the lock-in phenomenon generated by switchingcosts. Their model estimates an average duration of the bankcustomer relationship of 1312years.

G. J. de Bondt



9/42

Consequently, banks will set lending rates below the market clearing ratesand ration the amount of credit supply accordingly. In this case of creditrationing, bank lending rates exhibit upward stickiness.

However, this result of lending rate stickiness does not necessarily hold upif credit is not rationed. Consider a world in which there are two broad classesof borrowers to which banks can lend. For the first class of loans, such as fullysecured lending, the probability of default is zero, while for the second classof borrowers, the probability of default is positive and increasing in thelending interest rate through adverse selection and moral hazard. Assumethat banks can distinguish between the two borrower types, but not betweencustomers within each class and that banks are risk neutral. Given perfectcompetition, banks must earn the same expected return on both classes ofloans, as formulated in equation (2):

br1 1 Pbr2br2 g0 mr 2

where br1 is the rate charged by banks on the riskless loan, P( ) is theprobability of default on the second class of loan and br2 is the bank lendingrate on these loans.

For the first type of loans @br1/@mr51; that is, changes in the marginalcost of funds get transmitted one-for-one into changes in the lending rate onthe riskless loans. However, when banks are lending to the second borrowertype, @br2/@mr41, since @P/@br240. For these loans banks must increase

their lending rate by an amount greater than the increase in the marketinterest rate to compensate for the decrease in the probability of repayment.At some interest rate banks will not be able to increase the interest ratesufficiently to compensate for this risk and all lending will be made to thefirst type of borrower, also known as the flight-to-quality phenomenon.However, until this happens, the bank rate on these loans should be verysensitive to changes in the market interest rate. In other words, a more thanone-to-one adjustment of bank lending rates to changes in market interestrates, as shown in some cases in Table 1, suggests that bank credit was onaverage not rationed and consisted of relatively risky loans during the period

under review.

3.2. Empirical model specifications

Three different empirical frameworks are used in order to check therobustness of the empirical results. First, a VECM is applied to simultaneouslymodel both stages of the interest rate pass-through process. The mainadvantage, besides the one-step or simultaneous approach, is that adistinction is made between the short-term dynamics and the long-term orequilibrium relationships. For instance, in the presence of fixed adjustmentcosts retail bank interest rates will adjust to changes in market interest ratesonly if those adjustment costs are lower than the costs of maintaining a




10/42

non-equilibrium bank rate (Hannan and Berger, 1991). Furthermore, demandcurves are likely to be more inelastic in the short than in the long term. Thegreater the elasticity of demand for deposits and loans, the higher the cost of

keeping retail bank interest rates out of equilibrium. Second, the two stages ofthe interest rate pass-through process are separately examined by performingan impulse response analysis based on a VAR model of interest rate pairs inlevels. This framework maximizes the long-term information in the datasetand delivers super-consistent coefficient estimates. It avoids the possibility ofimposing inappropriate cointegration relations, which can lead to biasedestimates and hence may bias the impulse responses. The third and finalempirical framework is a univariate ECM, which has the advantage of aneconomically very appealing interpretation of the model parameters. Thisframework explicitly examines the immediate or within one-month pass-

through, the final pass-through, and the mean speed at which the final pass-through is reached. At the same time, however, any interaction betweenofficial, market and bank interest rates is fully ignored.

3.2.1. Vector error-correction modelA three-variable VECM is specified to simultaneously model the interest ratepass-through process:

Yt cXn1

i1

GiD

Yti Y

Ytn e

t 3

with:

Yt official ratemarket rate

bank rate

24

35

t

The rank of the matrixQ

determines the number of cointegration vectors. Auniform rank of one, as generally indicated by the trace and maximum

eigenvalue tests ( Johansen and Juselius, 1990; Johansen, 1991), has beenassumed to allow differences in the pass-through process across instrumentsto be compared. The cointegration relation has been normalized on the bankrate. In addition, in all cases the long-term coefficient of the official interestrate is set to zero in the cointegration relation. This binding restriction on theofficial interest rate is tested by the likelihood ratio test. The lag order isdetermined by a lag structure test and is typically one or two months and insome cases up to four months.

3.2.2. Vector autoregressive modelFor a VAR model in levels a uniform lag order is applied for all interest ratepairs to allow differences in the pass-through process across instruments to be

G. J. de Bondt



11/42

compared. Moreover, overparametrization is considered to be a largerproblem than underestimation of the lag order given the short sample. Thelag order is therefore set at two months, as low as possible given the wide

range of optimal lag orders derived from the Akaike, HannanQuin andSchwartz criteria and the residual properties. This leads to the followinguniform VAR model specification:

Yt cX2i1

AiYti et 4

with:

Yt official rate

market rate

t

4a

or

Yt market rate

bank rate

t

4b

In computing the impulse-response functions the missing identification ofthe contemporaneous relationships between each pair of interest rates issolved by using the traditional Cholesky decomposition of the residualvariancecovariance matrix (Hamilton, 1994). This decomposition relies on

the idea that a shock on the last ordered variable in the system does notcontemporaneously affect the previous one. For the interest rate pass-throughanalysis, the Cholesky decomposition method matches well with economicintuition. With the order of equation (4), a shock in the market/retail bankinterest rate will have no contemporaneous effect on the official/marketinterest rate, while shocks in the official/market interest rate may have animmediate impact on market interest rates with a longer maturity/retail bankinterest rate.

3.2.3. Univariate error-correction modelA univariate error-correction model, which also examines separately the twostages of the interest rate pass-through process, is specified as follows:

Dmrt a1 a2Dort b1mrt1 b2ort1 et 5a

Dbrt a1 a2Dmrt b1brt1 b2mrt1 et 5b

The coefficient a2 reflects the immediate or short-term pass-through,parameter b2 the final or long-term pass-through and (1 a2)/b1 equals themean adjustment lag at which the official/market interest rate is fully passedthrough to market/bank interest rates (Hendry, 1995). The existence of co-integration between the interest rate pairs can be directly tested by examining




12/42

the significance of the coefficient of the error-correction term, b1, using thecritical values as proposed by Kremers et al. (1992) and Boswijk (1994).

4. DATA

4.1. Official interest rate

Instead of using the ECB interest rate on the main refinancing operations, aninterest rate which changes only infrequently,4 the overnight interest ratehas been used in the empirical analysis, since 1 January 1999 (the euroovernight index average, EONIA). The overnight interest rate is the interest

rate central banks try to control by using instruments such as reserverequirements, standing facilities and open market operations (Perez Quirosand Rodrguez Mendizabal, 2001). Control normally means an attempt tokeep the daily overnight rate around an official rate which in some countries,for example the United States, is a target rate and in others, for example theeuro area, is the rate of the open market operations.

For the euro area this control is not perfect, but the EONIA closelyfluctuates around the ECB main refinancing rate. In the first eighteen monthssince the start of Stage Three of EMU the mean spread between the EONIAand the rate on the main refinancing operation of the eurosystem was 7 basis

points (Bindseil and Seitz, 2001). This average spread may, however,overestimate the actual spread because the sample is dominated by a cycleof expectations of interest rate increases. In addition, the volatility of theEONIA, which is mainly driven by liquidity concerns, may be high, inparticular at the end of the maintenance period (Gaspar et al., 2001). Sharpadjustments at the end of the maintenance are, however, usually taken backat the beginning of the new maintenance period. It is assumed that thevolatility dynamics are not dominant, since monthly averages of daily EONIArates are used in the empirical analysis. In other words, the overnight interestrates considered are more related to changes in the expectation of official

interest rates and less to liquidity issues. Empirical studies support theassumption that the EONIA reflects relatively well official interest ratedecisions. Gaspar et al. (2001) show, using daily EONIA rates, that markets areable to predict the ECBs interest rate decisions quite accurately within areserve maintenance period. Hartmann et al. (2001) show with thick data thatrate expectations by the market have been relatively precise during the periodunder review, with the absolute expectation error in the EONIA beforemonetary policy decisions averaging only about 5 basis points. The lattercompares to the usual 25 and 50 basis point step size in monetary policydecisions.

4. For instance, in the first three years of a single monetary policy in the euro area, 12 ECBinterest rate changes occurred.

G. J. de Bondt



13/42

4.2. Retail bank interest rates

The euro area retail bank interest rates on the deposits and loans considered

are average monthly data, including synthetic euro area data before 1 January1999. From 1 January 2001 onwards, euro area retail interest rate data includeGreece. The bank rates on five deposit categories (overnight deposits, savingdeposits up to and over three months notice and time deposits up to andover two years) and four loan categories (loans to enterprises up to and overone year and loans to households for consumer lending and house purchase)are examined. The retail bank interest rates considered reflect almost 100%new business, although the exact definition of new business may differ acrosseuro area countries.5 Differences in country weights across instruments arelarge, but fairly stable over time. For instance, the interest rate on deposits

redeemable at notice of over three months for the euro area is actually onlybased on German data. The retail bank interest rate statistics and furtherexplanations are made available on the Euro area statistics download pageof the ECBs website (http://www.ecb.int).

Figure 1 plots the average monthly interest rates charged by monetaryfinancial institutions (MFIs) in the euro area on the deposits and loansconsidered, respectively, for the sample period January 1996 to May 2001.6

Figure 1 illustrates that the level of bank deposit rates depends on maturity:typically the lower the maturity of the deposits the lower the deposit rate. Theinterest rate on overnight deposits is also low compared to other bank deposit

rates, partially due to the fact that in some countries this deposit rate isadministrated to be zero during the period under review. Turning to banklending rates, prima facie the maturity of the instruments also plays a role indetermining the level of bank lending rates: the lower the maturity the higherthe level of the bank lending rate. This is noteworthy since the yield curve isgenerally positively sloped. However, the differences in the level of banklending rates can be explained by differences in secured and unsecuredlending practices. Short-term borrowing by enterprises, such as overdrafts, isunsecured, while long-term lending to enterprises is mostly secured oncorporate assets. Mortgages are collateralized, while this is typically not the

case for consumer credit. Another explanation is that the information andmonitoring costs of banks are higher for short-term loans, because borrowers

5. The exceptions are the interest rate on overnight deposits, deposits redeemable at notice ofup to three months and on loans to enterprises up to one year. For the two deposit rates thedistinction between amounts outstanding and new business is rather artificial because of the

short-term nature of these deposits. As regards the lending rate, it is only for one country thatthe lending rate does not reflect close to 100% new business. Furthermore, the euro areaaggregate data are based on non-harmonized country data, with the potential risk that the

empirical results for the euro area aggregate retail bank interest rates are to some extent

distorted by an aggregation bias.6. See ECB (2001) for charts of the spread between retail bank interest rates and comparable

market interest rates.




14/42

with severe information asymmetries will tend to borrow more heavily atshort maturities.7

Turning to the relative importance of the different retail bank interest rates,the most important deposits in terms of the total deposits outstanding wereovernight deposits (demand deposits), closely followed by deposits redeem-able at notice of up to three months (savings deposits). They accounted eachto 25% to 30% of total deposits at the end of 2000. Deposits with an agreedmaturity of over two years and up to two years (time deposits) each amountedat around 20% of total deposits. Deposits redeemable at notice over threemonths were very small compared to the other deposit categories considered.As regards loans, loans to households for house purchase and loans toenterprises over one year were the most important in terms of amountsoutstanding (each around 30% of total loans). Loans to households forconsumer lending and loans to enterprises up to one year were of almost equalimportance and amounted to around 18% of total loans at the end of 2000.

4.3. Correlation analysis to detect comparable market interest rate

A correlation analysis between retail bank rates and market interest rates isperformed to detect which market interest rates are most closely related to thebank interest rate analysed. The idea is to capture adequately the marginalcost price. A distinction is made between the correlations of the variables inlevels and in first differences (the change in the interest rate). The marketinterest rates analysed, with Reuters as data source, are the overnight marketinterest rate, money market rates at 1, 3, 6 and 12 months maturities andgovernment bond yields at 2, 3, 5 and 10 years maturities. Correlations arecomputed both across maturities and for different lags of the market interest

rates. Table 2 presents the results of the correlation analysis for the sample

Bank deposit rates Bank lending rates

0

1

2

3

4

5

6

0

1

2

3

4

5

6

4

6

8

10

12

14

Jan-01Jan-00Jan-99Jan-98Jan-97Jan-96Jan-01Jan-00Jan-99Jan-98Jan-97Jan-964

6

8

10

12

14

Figure 1 Retail bank interest rates (percentages per annum; monthly averages)

Source: ECB.

7. See Section 1.6 on debt maturity theories in de Bondt (2000).

G. J. de Bondt



15/42

Table 2 Correlation analysis between bank and market interest rate

Bank rate

Marketrate Correlation

Lag inmonths

Marketrate Correlation

Lag inmonths

Marketrate Cor

1996.12001.5

a

1996.11998.12

a

1999.1

Level

Deposit rate

Overnight Overnight 0.89 8 Overnight 0.92 8 Overnight Up to 3 months notice 1 month 0.89 7 1 month 0.94 1 1 month Over 3 months noticeb 1 year 0.79 1 1 year 0.30 1 6 months

Maturity up to 2 years 2 years 0.96 5 3 months 0.96 1 3 months Maturity over 2 years 3 years 0.97 1 5 years 0.98 1 2 years Lending rate

Up to 1 year to firms 6 months 0.90 7 6 months 0.91 0 3 months Over 1 year to firms 5 years 0.94 4 10 years 0.98 3 1 year

Consumer lending 10 years 0.98 7 10 years 0.98 4 2 years House purchase 5 years 0.97 2 10 years 0.99 0 2 years

First difference

Deposit rateOvernight 1 month 0.54 1 1 month 0.22 2 1 month Up to 3 months notice 1 month 0.53 2 1 month 0.57 2 1 month

Over 3 months noticeb 1 year 0.77 1 1 year 0.78 1 6 months Maturity up to 2 years 6 months 0.73 1 3 months 0.67 1 3 months

Maturity over 2 years 2 years 0.70 0 10 years 0.64 0 2 years Lending rateUp to 1 year to firms 3 months 0.60 1 3 months 0.38 1 6 months

Over 1 year to firms 1 year 0.70 0 10 years 0.46 0 1 year Consumer lending 1 year 0.52 3 1 year 0.36 3 1 year House purchase 10 years 0.71 0 5 years 0.72 1 2 years

a Data for the interest rate on loans to enterprises with a maturity of over one year is available since Noveb Results for the samples starting in January 1996 are biased downward, since the interest rate on depositmonths is a 100% German interest rate and the money market rates considered are euro area money mark

Sources: ECB, Reuters and authors calculations.

r


Ltd.2005

51


16/42

starting in January 1996 as well as two sub-samples: 1996.11998.12 and1999.12001.5. The last column of Table 2 shows the comparable market ratesselected for the empirical analyses of Sections 5 and 6.

Looking at the total sample results (see columns 24 in Table 2), thecorrelation coefficients in level terms broadly vary between 0.89 and 0.98,implying that bank and market rates move closely together. The lag with thehighest correlation varies between 1 and 8 months, suggesting that retailbank rates do not collectively react by the same speed to market interest ratechanges. In first differences, the correlation coefficients are lower with arange of 0.52 and 0.77. For all bank interest rates, the same or a loweroptimal lag applies than for the levels.

As regards the two sub-samples, the same qualitative picture emerges.However, two striking differences emerge when comparing the two sub-

samples. First, the correlation coefficients for the sample starting in January1999 are higher than for the years 199698. Second, the lags with the highestcorrelation are typically lower for the sample starting in January 1999 thanfor the three-year sample starting in January 1996. These findings suggest thatthe co-movement between retail bank rates and market interest rates hasbecome closer since the introduction of the euro, suggesting a quickerinterest rate pass-through process since the start of Stage Three of EMU.Section 6 addresses this issue in more detail.

5. EMPIRICAL RESULTS

This section examines the interest rate pass-through process in the euro areaon the basis of three different empirical methods: VECM (Section 5.1), VAR(Section 5.2) and a univariate ECM (Section 5.3). Broadly speaking, all threeframeworks show that the interest rate pass-through process clearly differsbetween the short- and long-end of the yield curve and across retail bankinterest rates in the euro area. The impact of a change in the official interestrate declines with the maturity of market interest rates. Monetary policy fullycontrols money market rates up to three months. At the same time,

government bond yields are not significantly affected by changes in theofficial interest rate. Retail bank interest rates are found to be sticky in theshort term, whereas in the long term a close to complete pass-through ofmarket interest rate developments to retail bank interest rates, most notablylending rates, is observed.

5.1. Vector error-correction model

The VECM results show that the long-term pass-through is typicallyincomplete for bank deposit rates and complete for bank lending rates (seeTable 3). In particular, the bank rate on overnight deposits and depositsredeemable at notice of up to three months do not fully adjust to changes in

G. J. de Bondt



17/42

the market interest rate developments. Explanations for this finding might bethat these segments of the retail bank market are not fully competitive, have arather inelastic demand and/or that the switching costs of demand andsavings deposits are comparatively high. In contrast, the rate on loans toenterprises of up to one year adjusts more than one-to-one to market interestrate developments. This overshooting may, among other factors, be explainedby asymmetric information costs without credit rationing, as described inSection 3. If banks increase their lending rates exactly one-for-one withmarket interest rates they will attract a riskier class of borrowers. Conse-quently, banks have to increase the lending rate premium charged. Theresults of the likelihood ratio test show that the imposed restriction of nolong-term impact of the official interest rate in the determination of the retailbank interest rate can typically not be rejected. The only exceptions are therates on time deposits up to two years and short-term corporate loans.

5.2. Vector autoregressive model

Figure 2 summarizes the impulse response results for the official interest ratepass-through. The figure plots how a temporary shock to the overnight rate, asa proxy for the official interest rate, is passed through to market interest rateswith a longer maturity. The immediate or one-month pass-through is around50% at the short-end of the yield curve (up to three months) and well below50% for the market interest rates with longer maturities. The pass-throughafter 12 months is close to complete for the one- and three-month money

Table 3 Interest rate pass-through based on a vector error-correction model

Retail bank rate

Long-term

pass-through

Adjustment

coefficient

Complete

pass-through?

LR

testa

Deposit rateOvernight 0.32** 0.04** No** b

Up to 3 months notice 0.25** 0.27** No** 3.43Over 3 months notice 0.50** 0.08* No** 0.17Maturity up to 2 years 1.08** 0.03 Yes** 25.1**

Maturity over 2 years 0.71** 0.31** No** 1.17

Lending rateUp to 1 year to firms 1.37** 0.07* No** 19.7**Over 1 year to firms 0.88** 0.11* Yes** 1.02

Consumer lending 0.90** 0.12** Yes** 0.56House purchase 0.95** 0.21** Yes** 0.03

Notes: Results based on equation (3) using a sample 1996.12001.5. ** and * denote significanceof the F-statistic at the 1% and 5% levels, respectively.aw2 test of long-term restrictions: market rate coefficient51 (normalization) and overnight rate50.b No binding long-term restriction, since market rate is overnight rate.Sources: ECB, Reuters and authors estimations.




18/42

market interest rate, around 40% for the six-month money rate, 0% for the12-month money market and negative for government bond yields. The lattersuggests that at the long-end of the yield curve an increase in the officialinterest rate has typically lowered long-term inflation expectations during theperiod under review, thereby bringing government bond yields down.

Figure 3 summarizes the retail bank interest rate pass-through according toan impulse response analysis. The figure plots how a temporary shock to the

-40

-20

0

20

40

60

80

100

121110987654321

One-month money market rate

Three-month money market rateSix-month money market rate

Twelve-month money market rateTwo-year government bond yield

Five-year government bond yieldTen-year government bondyield

Figure 2 Official interest rate pass-through based on VAR framework (percentageshares of cumulative official interest rate shock passed through to market interest rate)

Notes: Results based on equation (4a) using a sample 1996.12001.5. See Appendix A for

underlying impulse responses including standard errors.Sources: ECB, Reuters and authors estimations.


Figure 3 Retail bank rate pass-through based on VAR framework (percentage sharesof cumulative market interest rate shock passed through to bank rate)

Notes: Results based on equation (4b) using a sample 1996.12001.5. See Appendix A forunderlying impulse responses including standard errors.

Sources: ECB, Reuters and authors estimations.

G. J. de Bondt



19/42

market interest rate is passed through to bank deposit and lending rates,respectively. Three main conclusions emerge from Figure 3.

The first conclusion is that shocks in the market interest rates are not

immediately reflected in retail bank interest rates. In other words, euro areabank interest rates are sticky in the short term, in line with previous countrystudies. The immediate response of deposit rates to a 1 percentage pointshock to the market interest rate varies between around 5 basis points forovernight deposits and deposits redeemable at notice of up to three monthsand around 35 basis points for deposits with an agreed maturity of over twoyears. The immediate pass-through of a 1 percentage point shock in themarket interest rate to a bank lending rate varies between around 15 basispoints for consumer lending and 50 basis points for loans to enterprises ofover one year.

Second, the pass-through of a market interest rate shock is higher in thelonger term. After 12 months the pass-through for deposits rates variesbetween 15% (overnight deposits) and 68% (time deposits), and for banklending rates between 44% (consumer lending) and 76% (loans to enterprisesup to one year and loans to households for house purchase). After 36 monthsthe pass-through is far from complete for the overnight deposits and depositsredeemable at notice of up to three months, but almost fully complete(around 90%) for time deposits and loans. The exception is loans toenterprises up to one year that shows, in line with the VECM outcome, apass-through of around 145%. This overshooting of the bank lending ratesuggests a move towards riskier borrowers during the period under review.

The third conclusion, closely related to the other two conclusions, is thatthe interest rate pass-through process clearly differs across instruments. Theinterest rates on overnight deposits and deposits redeemable at notice of upto three months are sticky compared to other bank deposit rates. The interestrate on consumer lending is sticky compared to the other bank lending rates.

5.3. Univariate error-correction model

Four conclusions emerge from the estimation results of equations (5a) and(5b), as summarized in Tables 4 and 5, respectively. Overall, the resultsconfirm the findings based on the VECM and VAR frameworks.

First, changes in the official interest rates are completely passed through tomarket interest rates at the short-end of the yield curve in the euro area, i.e.up to three months. No significant statistical relationship exists between theofficial interest rate and government bond yields, suggesting a crediblemonetary policy. The declining impact of the change in the official interestrate to market interest rates as the maturity rises is in line with Perez-Quirosand Sicilia (2002). They find that the impact of monetary policy shocks onbond yields declines with the maturity of the bonds.

Second, the immediate pass-through of market interest rates to retail bankinterest rates is found to be incomplete. The proportion of a given market




20/42

interest rate change that is passed through within one month is found to betypically around 30% during the sample period. The highest immediate pass-through is found to be 54% in the case of the interest rate on lending to

Table 4 Official interest rate pass-through based on a univariate error-

correction model

Retail bank rate

Marketrate

Finalpass-through

Adjustmentspeed

in months

Immediatepass-through

complete?

Finalpass-through

complete?a2 b2 (1 a2)/b1 a251 b251

Money market rates1-month deposits 0.63** 0.90** 0.62** No** Yes**

3-month deposits 0.45** 0.68** 3.12** No** Yes**6-month deposits 0.35** No** No**12-month deposits 0.28** No** No**

Government bond yields2 years 0.17 No** No**

5 years 0.01 No** No**

10 years 0.05 No** No**

Notes: Results based on equation (5a) using a sample 1996.12001.5. ** and * denote significance

of the F-statistic at the 1% and 5% levels, respectively.Sources: ECB, Reuters and authors estimations.

Table 5 Retail bank interest rate pass-through based on a univariate error-

correction model

Retail bank rate

Immediatepass-

through

Finalpass-

through

Adjustmentspeed

in months

Completepass-

through?Cointegration

relation?a2 b2 (1 a2)/b1 b251 b150

Deposit rateOvernight 0.02 0.41** 15.3** No** Yes**

Up to 3 months notice 0.05 0.35** 8.7** No** Yes**

Over 3 months notice 0.35** 0.87** 23.8 Yes** No**

Maturity up to 2 years 0.32** 0.98** 4.8** Yes** Yes**Maturity over 2 years 0.35** 0.76** 3.0** No** Yes**

Lending rateUp to 1 year to firms 0.24** 1.53** 8.7** No** Yes**Over 1 year to firms 0.54** 0.92** 3.9** Yes** Yes**

Consumer lending 0.13** 0.93** 10.2** Yes** Yes**

House purchase 0.26** 0.94** 2.8** Yes** Yes**

Notes: Results based on equation (5b) using a sample 1996.12001.5. ** and * denote significance

of the F-statistic at the 1% and 5% levels, respectively.Sources: ECB, Reuters and authors estimations.

G. J. de Bondt



21/42

enterprises over one year. This may be explained by a relatively elastic loandemand, since firms have access to alternative sources of funds, such as tradecredit (Kohler et al., 2000) and the corporate bond market (de Bondt, 2002).

Third, the final pass-through of market interest rates to retail bank interestrates is typically complete or even well above 100%, as in the case of loans toenterprises of up to one year. The final pass-through is clearly incomplete forthe interest rates on overnight deposits and deposits redeemable at notice ofup to three months.

The fourth and final conclusion is that the average speed for retail bankinterest rates to fully adjust to market interest rate changes is typicallybetween 3 and 10 months. Exceptions to this finding are the slow speed ofadjustment between 1 and 2 years for the interest rate on overnight depositsand deposits redeemable at notice of over 3 months. The latter finding is

biased because this deposit rate is actually a German interest rate, while a euroarea market interest rate is considered. This also explains why for this depositrate no cointegration relation with the market interest rate is found. In allother cases a long-term equilibrium relationship exists between the retailbank interest rate and the comparable market interest rate. Regarding lendingrates, the speed of adjustment of around 9 months as found for the interestrates on lending to enterprises with a maturity up to 1 year and on consumerlending is much slower than for the other lending rates which show a meanadjustment speed of around 3 months. This difference might be explained bythe fact that the asymmetric information costs are higher for unsecuredlending than for secured lending.

6. EMPIRICAL RESULTS BASED ON STAGE THREE OF EMU

This section examines the hypothesis of a change in the interest rate pass-through since the introduction of a common monetary policy by applyingthe three empirical methods to a sub-sample starting in January 1999 (EMUsample). The tentative conclusions, albeit robust across the three empiricalframeworks, are that the monetary policy control at the short-end of the yieldcurve has become immediately complete at the euro area level since theintroduction of the euro. In addition, a quicker retail bank interest rate pass-through has been observed since the common monetary policy in the euroarea. The quicker pass-through can, among other factors, be explained byless collusion among banks, which, in turn, may be affected by the fact thatthe ECB monetary policy reacts to euro area conditions and have con-tributed to less variability in money market interest rates (Bagliano et al.,2000). It is also shown that asymmetric information costs, as reflected incredit risk concerns, may significantly explain the slower speed at which banklending rates adjust to market interest rate developments compared withdeposit rates.




22/42

6.1. Vector error-correction model

Table 6 summarizes the interest rate pass-through process on the basis of a

VECM using a sample starting in January 1999. The most striking differencebetween the results of Table 6 and Table 3 of the full sample is that theadjustment coefficients have generally increased since the introduction of acommon monetary policy, suggesting a quicker interest rate pass-through.The long-term pass-through is found to be around 20% for the rate onovernight deposits and deposits redeemable at notice of up to three months,and around 80% for the other deposit and lending rates. The exception is therate on loans to households for house purchase for which a complete long-term pass-through is found. The likelihood ratio tests show that the bindingrestriction that the official interest rate is not part of the cointegration

relation between the bank and market rate cannot be rejected in all cases atthe 1% confidence level.

6.2. Vector autoregressive model

Figures 4 and 5 provide an overview of the interest rate pass-through process sinceJanuary 1999 based on the VAR approach. Broadly speaking, the picture emergingfrom Figures 4 and 5 is similar to that of Figures 2 and 3: monetary policy controlsthe short-end of the yield curve, retail bank interest rates are sticky in the shortterm, the pass-through is close to complete in the long term, and striking

Table 6 Interest rate pass-through process based on vector error-correction

modela

Retail bank rateLong-term

pass-throughAdjustmentcoefficient

Completepass-through?

LRtestb

Deposit rateOvernight 0.18** 0.48** No** c

Up to 3 months notice 0.24** 0.26** No** 6.38*

Over 3 months notice 0.74** 0.58* No** 1.71Maturity up to 2 years 0.77** 0.16 No** 2.90Maturity over 2 years 0.85** 0.78 No** 3.63

Lending rateUp to 1 year to firms 0.85** 0.37** No** 2.16Over 1 year to firms 0.84** 0.34 No** 2.18Consumer lending 0.62** 0.22** No** 0.09House purchase 1.18** 0.22** Yes** 0.28

Notes: Results based on equation (3) using a sample 1999.12001.5. ** and * denote significance

of the F-statistic at the 1% and 5% levels, respectively.a

w2

test of long-term restrictions: market rate coefficient5

1 (normalization) and overnight rate5

0.b No binding long-term restriction, since market rate is overnight rate.Sources: ECB, Reuters and authors estimations.

G. J. de Bondt



23/42

differences in the pass-through process exist across retail bank instruments. Thereare, however, two striking differences between both sets of charts.

First, the immediate pass-through of changes in the official interest rate tomoney market interest rates up to three months has increased from around60% to 80%, suggesting a strengthening of the monetary policy control at theshort-end of the yield curve.

-60

-40

-20

0

20

40

60

80

100

120

1 2 3 4 5 6 7 8 9 10 11 12

One-month money market rate Three-month money market rate

Six-month money market rate Twelve-month money market rateTwo-year government bond yield Five-year government bond yieldTen-year government bond yield

Figure 4 Official interest rate pass-through based on VAR framework since January1999 (percentage shares of cumulative official interest rate shock passed through to

market interest rate)

Notes: Results based on equation (4a) using a sample 1999.12001.5. See Appendix B forunderlying impulse responses including standard errors.



0

20

40

60

80

100

1 4 7 10 13 16 19 22 25 28 31 34 1 4 7 10 13 16 19 22 25 28 31 34

0

20

40

60

80

100

0

20

40

60

80

100

120

140

0

20

40

60

80

100

120

140

Figure 5 Retail bank rate pass-through based on VAR framework since 1 January1999 (percentage shares of cumulative market interest rate shock passed through to

bank rate)

Notes: Results based on equation (4b) using a sample 1996.12001.5. See Appendix B forunderlying impulse responses including standard errors.





24/42

The second difference relates to the extent by which market interest rateshocks are immediately passed through to retail bank interest rates. The shareof the immediate pass-through of a market interest shock to the interest rate

on deposits redeemable at notice of over three months, time deposits and onloans to households for house purchase has increased since the introductionof the euro by around 15 percentage points. The main difference with respectto the long-term pass-through is that the interest rate on lending to enterprisesup to one year no longer overshoots, in line with the VECM outcome.

6.3. Univariate error-correction model

Table 7 summarizes the official interest rate pass-through process since January 1999. For the EMU sample official interest rates are immediatelycompletely passed through to money market interest rates up to threemonths. This suggests that the monetary policy control of the short-end ofthe yield curve at the euro area level has increased since January 1999. This isnot a surprising finding given the fact that in this period a commonmonetary policy was in place in the euro area. A significant incompleteimmediate pass-through for the money market rates with a longer maturityand no statistical relationship between the official interest rate andgovernment bond yields have been found, in line with the full sample results.

Table 8 summarizes the retail bank interest rate pass-through process sincethe introduction of the euro and compares the results based on the EMUsample with the estimated coefficients based on the full sample. Comparing

Table 7 Official interest rate pass-through since January 1999 based on a

univariate error-correction model

Retail bank rate


Finalpass-through

Adjustmentspeed inmonths


complete?

Finalpass-through

complete?a2 b2 (1 a2)/b1 a251 b251

Money market rates1-month deposits 0.90** 0.94** 0.14** Yes** Yes**

3-month deposits 0.61** 0.73** 2.66** Yes** Yes**

6-month deposits 0.37** No** No**12-month deposits 0.31** No** No**

Government bond yields2 years 0.12 No** No**5 years 0.16 No** No**10 years 0.34 No** No**

Notes: Results based on equation (5a) using a sample 1999.12001.5. ** and * denote significanceof the F-statistic at the 1% and 5% levels, respectively.Sources: ECB, Reuters and authors estimations.

G. J. de Bondt



25/42

Table 8 Retail bank interest rate pass-through since January 1999 based on a univa

Retail bank rate

Pass-through EMU sampleChow test

Immediate Final Speed Complete Cointegration LR a2 b2 (1 a2/b1) b251 b150 Break 1999.

Deposit rateOvernight 0.04** 0.18** 1.6** No** Yes** 24.6**

Up to 3 months notice 0.01 0.26** 4.1** No** Yes** 15.4**Over 3 months notice 0.35** 0.72** 0.8** No** Yes** 53.0**Maturity up to 2 years 0.38** 0.76** 1.1** No** Yes** 53.2**

Maturity over 2 years 0.47** 0.65** 1.0** No** Yes** 28.1**

Lending rateUp to 1 year to firms 0.19* 0.88** 2.8** Yes* Yes** 38.7**Over 1 year to firms 0.55** 0.80** 2.5 Yes* No** 9.5*Consumer lending 0.08 0.61** 6.3** No** Yes** 3.8 House purchase 0.46** 1.04** 2.6** Yes** Yes** 18.7**

Notes: Results based on equation (5b) using a sample 1999.12001.5. ** and * denote significance of the

respectively, unless stated otherwise; LR denotes log likelihood ratio statistic.aF-statistic.Sources: ECB, Reuters and authors estimations.

r


Ltd.2005

61


26/42

the findings based on the EMU sample with the results from the full sample,three observations emerge.

The first observation is that Chow breakpoint tests, analysing a structural

break for the error-correction model at January 1999, indicate a significantchange in the pass-through process since 1 January 1999 in all cases with theexception of the interest rate on consumer lending. The impact of theintroduction of the euro is, for all retail bank interest rates, reflected in aquicker speed of adjustment, which in most cases is statistically significant.The mean adjustment lag since January 1999 is typically 1 month for depositinterest rates and 3 months for bank lending rates. Exceptions are the interestrate on deposits redeemable at notice of up to 3 months and on consumerlending with an adjustment speed of 4 and 6 months, respectively.

Second, the proportion of a given market interest rate change that is

immediately passed through to the interest rate on mortgages and ondeposits with a maturity of over two years has significantly increased sincethe start of Stage Three of EMU by around 20 and 10 percentage points,respectively. This suggests a rise in the prevailing competitive forces and/or afall in switching and asymmetric information costs in these segments of theretail bank market.

A third difference between both sets of findings is that for the period since1 January 1999 the final pass-through to interest rates on loans to enterprisesup to one year is no longer more than 100%. At the same time, for the interestrate on consumer lending an incomplete long-term pass-through is found forthe period since the introduction of the euro, suggesting a certain degree ofcredit rationing in this segment of the credit market during this period. Moregenerally, for all retail bank interest rates, except the mortgage interest rate,the final pass-through found for the EMU sample is lower than for the pre-EMU sample. In several cases this effect is statistically significant.

6.4. Explanations for the observed differences in the retail bank interestrate pass-through

6.4.1. Quicker interest rate pass-through since the euroPrima facie, three likely explanations emerge for the observed quicker retailbank interest rate pass-through process since January 1999.

First, one could argue that the observed change around January 1999relates to the reversal in the direction of the change in market interest ratesand therefore is in favour of an asymmetric interest rate pass-through, e.g.bank lending rates are quicker adjusted upward in an environment of risinginterest rates than adjusted downward when interest rates fall. In the courseof 1999 the declining pattern in interest rates as observed since January 1996reversed into a rise in interest rates, which continued more or less up to theend of 2000, after which market interest rates started to decline again.Although the period of rising market interest rates is a substantial part of theEMU sample, an asymmetric pass-through argument cannot explain why the

G. J. de Bondt



27/42

speeding up is not only found for lending rates, but also for deposit rates forwhich the opposite effect is expected.

The second and more likely economic explanation is that the prevailing

competitive forces in most segments of the retail bank market have increasedsince 1999, both from a supply and from a demand perspective. The develop-ments towards a rise in the prevailing competitive forces in the differentsegments of the retail bank market, in turn, may have been triggered by theintroduction of a common monetary policy (see third explanation below).

As regards the supply side, the ongoing restructuring of the banking sectorseems to have promoted higher efficiency in the banking markets and morecompetitive pricing. A declining role of traditional banking, e.g. grantinglong-term loans and funding them by short-dated deposits, as reflected in anincreasing use by banks of non-deposit or market-based funding sources is

observed in the EMU sample (see Figure 6). As regards the euro area interbankmarket, which plays an important role in the overall banking activity in theeuro area, an increasing use of cross-border transactions vis-a-vis the domesticones exists since the introduction of the euro. In addition, notwithstandingthe steady decline in the number of credit institutions, new (foreign) playershave become active in the retail bank credit market. Euro area banks also facemore and more competition from other financial intermediaries such asinsurance companies and pension funds (ECB, 2002).

Turning to the demand side, there are also indications that competitionhas increased as bank customers have become more sensitive to the interestrates on bank loans and deposits compared with those charged by other banksand by financial markets. This may be related to the higher stability ofnominal interest rates in an environment of price stability, which facilitatesthe comparison of prices across different suppliers, and to an increase in the

31

32

33

34

35

1997 1998 1999 2000 2001 200246

47

48

49

50

51

52

53

Deposit funding (left-hand scale) Alternative funding (right-hand scale)

Figure 6 Deposit and alternative funding of MFI loans in the euro area (in per centof balance sheet total)

Note: Alternative funding defined as deposits from resident MFIs and from non-residents(interbank and foreign deposits) and securities other than shares issued by MFIs.Source: ECB.




28/42

availability of alternative non-bank sources of finance, at least for someenterprises. For example, the commercial paper and corporate bond marketsin the euro area have become deeper and more liquid in recent years (de

Bondt, 2002). In addition, money market mutual funds, which offer market-based returns, have grown considerably during the EMU sample.

The third and final explanation, closely related to the competitionargument, is that the price-setting behaviour by euro area banks may havechanged due to the shift from national central banks to the ECB. Baglianoet al. (2000) theoretically show that in particular for German bank lendingrates a strong impact of the change towards a common monetary policy atthe euro area level is expected, since German banks, from January 1999, face amonetary policy which has responded to the euro area instead of Germanconditions. Empirical results for Germany using the univariate ECM are

indeed in favour of this hypothesis (see the table in Appendix C). Germanbank lending rates adjust significantly quicker to changes in market interestrate developments since January 1999. For the other euro area countries theimpact of the introduction of a common monetary policy is ambiguous. Onthe one hand, the other euro area countries may experience less moneymarket interest rate variability from external conditions than before,suggesting a quicker interest rate pass-through. On the other hand, a largerECB responsiveness to their country or local shocks than their nationalcentral banks did before 1999 under the EMS may result in a slower pass-through. Preliminary results at the euro area country level for four lendingrates and one time deposit rate, point in the direction of a quicker adjustmentspeed since the introduction of a common monetary policy. In 34 out of 36cases the adjustment speed of retail bank interest rates to changes in theshort- and long-term market interest rates has increased since January 1999(de Bondt et al., 2002).

6.4.2. Quicker bank deposit than lending rate pass-throughThe generally slower speed of adjustment of bank lending rates comparedwith bank deposit rates may relate to credit risk considerations due to

asymmetric information costs (Wong, 1997; Winker, 1999). The spreadbetween yields on long-term BBB-rated corporate bonds and governmentbonds with a comparable maturity can provide an indication of the degree ofprevailing corporate credit risk as viewed by corporate bond marketparticipants. This proxy for long-term corporate credit risk is added as anadditional explanatory factor in the univariate error-correction modelspecification for the long-term corporate lending rate:

Dbrt a1 a2Dmrt a3Dcrt b1brt1 b2mrt1 b3crt1 et 6

Estimates of equation (6) show that the immediate adjustment of the long-term corporate bank lending rate to a change in the market interest rateremains around 55%, but that the adjustment speed substantially increases

G. J. de Bondt



29/42

due to a larger adjustment coefficient (see Table 9). The adjustment speed onthe basis of the model with a credit risk factor is found to be 1 month insteadof 212 months. The coefficient estimates for the corporate bond spread suggestthat, following a 100 basis point change in the BBB-rated corporate bondspread, banks immediately adjust the long-term corporate lending rate witharound 15 basis points and in the long term by around 35 basis points.

Another observation is that the final pass-through for deposit rates istypically lower than for lending rates. This may be explained by a lowerinterest rate elasticity of the demand for deposits than of loans as shown bymoney and loan demand studies for the euro area. For money demand in theeuro area Calza et al. (2001) find a semi-elasticity with respect to theopportunity cost of M3 between 0.6 and 0.9. As regards the demand of firmsand households for loans in the euro area, Calza et al. (2003) find a semi-elasticity with respect to the short-term interest rate of 0.4 to 1.0 and to thelong-term interest rate of 1.8 to 3.1. However, it should be kept in mind thatthese interest rate elasticities are based on total deposits and loans, whereasthis study examines specific components of total deposits and loans.

7. CONCLUDING REMARKS

This study empirically analyses for the first time the interest rate pass-throughprocess at the euro area level. The bank and market interest rates consideredhave, in contrast to previous studies, a comparable maturity to avoid distortionsfrom yield curve effects and to accurately capture the marginal pricing costs ofbanks. Besides the pass-through of changes in market interest rates to retailbank interest rates, the pass-through of official interest rates, approximated by

Table 9 Long-term corporate bank lending rate pass-through taking

account of credit risk based on a univariate error-correction model

a1


Adjustmentcoefficient

b1

Finalpass-through

Adjustmentspeed

(1 a2)/b1 R2

Marketrate

Creditriska

Marketrate

Creditriska

a2 a3 b2 b3

0.46 0.55** 0.18 0.80** 2.5 0.70(0.31) (0.08) (0.11) (0.10)0.96 0.58** 0.15 0.37* 0.66** 0.37* 1.1* 0.72

(0.50) (0.09) (0.13) (0.18) (0.07) (0.18)

Notes: Estimates of equation (6) using a sample period 1999.12001.5. ** and * denotesignificance at the 1% and 5% levels, respectively; standard errors are reported in parentheses.aLong-term corporate credit risk is approximated by the long-term BBB-rated corporate bond spread.

Sources: ECB, Merrill Lynch, Reuters and authors estimations.




30/42

the overnight interest rate, to market interest with longer maturities isexamined. Notwithstanding the fact that the empirical findings presented inthis paper have to be interpreted with more than usual caution because the

sample period is short and the interest rate cycles covered are limited, threemain conclusions emerge from the empirical results presented. These results arefairly robust across the three different empirical methods applied.

First, monetary policy fully controls the short-end of the yield curve, i.e.money market interest rates with a maturity of up to three months. The pass-through of official interest rates to money market interest rates up to threemonths has become immediately complete since January 1999, in line withthe introduction of a common monetary policy in the euro area. The impactof changes in official interest rates on market interest rates is found to declinewith the maturity of the market instrument. No significant statistical

relationship has been found between official interest rates and governmentbond yields, suggesting a credible monetary policy.

Second, the immediate pass-through of market interest rates to retail bankinterest rates is incomplete, in line with previous cross-country studies. Theproportion of a given market interest rate change that is passed throughwithin one month is found, at its highest, to be around 50%. The pass-through is higher in the longer term and notably for bank lending rates closeto 100%. The most sticky retail bank interest rates in the euro area are theinterest rates on overnight deposits and deposits redeemable at notice of upto three months with a long-term pass-through of at most 40%.

The third and final conclusion is that the empirical results suggest aquicker retail bank interest rate pass-through process in the euro area sincethe introduction of the euro. For all retail bank interest rates the meanadjustment speed has become quicker since January 1999. The mean speed atwhich bank rates finally adjust to market interest rate developments since theintroduction of the euro is found to be typically around one month fordeposit rates and around three months for lending rates. A likely factorunderlying this speeding up is the tendency towards a rise in the prevailingcompetitive forces in the different segments of the euro area retail bankmarket during the first years of Stage Three of EMU. The latter, in turn, mayhave been triggered by the introduction of the euro and/or the commonmonetary policy. This is a topic for further research. Another field of futureresearch is the role of credit risk in the determination of bank lending rates.Tentative results as presented in this paper in this respect may form a startingpoint for this. The possibility of an asymmetric interest rate pass-throughprocess in the euro area also warrants future research, as soon as a moreextended EMU interest rate cycle becomes available.

G. J. de Bondt



31/42

APPENDIX A. IMPULSE RESPONSE ANALYSIS OF INTERESTRATE PASS-THROUGH

Official interest rate pass-through (response of the market interest rate to a onestandard deviation innovation of the official interest rate)




32/42

G. J. de Bondt



33/42

Retail bank interest rate pass-through (response of the retail bank rate to a onestandard deviation innovation of the market interest rate)




34/42

Notes: Dotted lines denote 95% confidence interval based on analyticalstandard errors of estimated VAR models of interest rate pairs using a sample1996.12001.5.Sources: ECB and authors estimations.

G. J. de Bondt



35/42

APPENDIX B. IMPULSE RESPONSE ANALYSIS OF INTERESTRATE PASS-THROUGH SINCE JANUARY 1999

Official interest rate pass-through since January 1999 (response of the marketinterest rate to a one standard deviation innovation of the official interest rate)




36/42

G. J. de Bondt



37/42

Retail bank interest rate pass-through since January 1999 (response of the retailbank rate to a one standard deviation innovation of the market interest rate)




38/42

Notes: Dotted lines denote 95% confidence interval based on analyticalstandard errors of estimated VAR models of interest rate pairs using a sample1999.12001.5.Sources: ECB and authors estimations.

G. J. de Bondt



39/42

APPENDIX C. RETAIL BANK INTEREST RATEPASS-THROUGH IN GERMANY

ACKNOWLEDGEMENTS

Comments by John Fell, Hans-Joachim Klockers, Manfred Kremer and ananonymous referee are appreciated. All views expressed are those of theauthor alone and do not necessarily reflect those of the ECB or theEurosystem.

Table A.1 German retail bank interest rate pass-through based on aunivariate error-correction model

Retail bankinterest rate

Market interestrates with a

comparable maturity

Immediatepass-

througha

Finalpass-

throughb

Adjustmentspeed

(in months)a2 b2 (1 a2)/b1

Bank deposit rate1996.12001.5Up to 3 months notice 3 months 0.16** 0.31* 11.4*

Over 3 months notice 12 months 0.27** 0.75** 2.3**Maturity of 1 month 1 month 0.43** 0.72** 1.1**

Maturity of 3 months 3 months 0.45** 0.83** 1.0**

1999.12001.5Up to 3 months notice 3 months 0.04 0.22** 2.7**Over 3 months notice 12 months 0.32** 0.78** 2.1**Maturity of 1 month 1 month 0.42** 0.71** 1.0**Maturity of 3 months 3 months 0.42** 0.82** 1.0**

Bank lending rate1996.12001.5

Up to 1 year to firms 12 months 0.12* 1.05** 6.8**Over 1 year to firms 58 years 0.57** 1.26** 4.9*Consumer lending 35 years 0.04 0.82** 10.5**House purchase 35 years 0.69** 1.06** 1.0**1999.12001.5Up to 1 year to firms 12 months 0.02 0.89** 4.3**Over 1 year to firms 58 years 0.56** 1.14** 1.9**

Consumer lending 35 years 0.08 0.83** 10.2House purchase 35 years 0.75** 0.98** 0.3**

Notes: Results based on equation (5b). ** and * denote significance at the 1% and 5% levels,

respectively.a In all cases the immediate pass-through is statistically different from 1.b In all cases the final pass-through is statistically different from 1 for deposit rates, but not forlending rates.Sources: Bundesbank and authors estimations.




40/42

Address for correspondence: European Central Bank, Kaiserstrasse 29, D-60311, Frankfurt am Main, Germany. Tel.: 49 69 1344 6477; fax: 49 691344 6514; email: [email protected]

REFERENCES

Angbazo, L. (1997), Commercial Bank Net Interest Margins, Default Risk, Interest-RateRisk, and Off-Balance Sheet Banking, Journal of Banking and Finance 21, 5587.

Bagliano, F. C., A. Dalmazzo and G. Marini (2000), Bank Competition and ECBsMonetary Policy, Journal of Banking and Finance 24, 967983.

Bank for International Settlements (BIS) (1994), National Differences in Interest RateTransmission, Basle, March.

Bindseil, U. and F. Seitz (2001), The Supply and Demand for Eurosystem Deposits: TheFirst 18 Months, ECB Working Paper No. 44, European Central Bank.

Bondt, G. J. de (2000), Financial Structure and Monetary Transmission in Europe: A Cross-Country Study, Edward Elgar, Cheltenham.

Bondt, G. de (2002), Euro Area Corporate Debt Securities Market: First EmpiricalEvidence, ECB Working Paper No. 164, European Central Bank.

Bondt, G. de, B. Mojon and N. Valla (2002), Interest Rate Setting by Universal Banksand the Monetary Policy Transmission Mechanism in the Euro Area, CEPRConference Paper, Conference entitled Will Universal Banking Dominate orDisappear? Consolidation, Restructuring and (Re)regulation in the BankingIndustry, Madrid, 15 and 16 November.

Borio, C. E. V. and W. Fritz (1995), The Response of Short-Term Bank Lending Rates toPolicy Rates: A Cross-Country Perspective, in BIS, Financial Structure and theMonetary Transmission Policy Transmission Mechanism, Basle, March, pp. 106153.

Boswijk, H. P. (1994), Testing for an Unstable Root in Conditional and Structural ErrorCorrection Models, Journal of Econometrics 63, 3760.

Calza, A., C. Gartner and J. Sousa (2003), Modelling the Demand for Loans to thePrivate Sector in the Euro Area, Applied Economics 35, 107117.

Calza, A., D. Gerdesmeier and J. Levy (2001), Euro Area M3: Measuring theOpportunity Costs Appropriately, IMF Working paper No. 01/179.

Corvoisier, S. and R. Gropp (2002), Bank Concentration and Retail Interest Rates,Journal of Banking and Finance 26, 21552189.

Cottarelli, C. and A. Kourelis (1994), Financial Structure, Bank Lending Rates, and theTransmission Mechanism of Monetary Policy, IMF Staff Papers 41, 587623.

interest rate pass-through - empirical results for the euro area

Documents