The ECB's experience of monetary policy in a financially fragmented euro area.
Following more than 6 years of a mutating financial crisis, this paper recaps the various monetary policy decisions made by the European Central Bank (ECB) and attempts to describe their underlying rationale in a non-judgemental way.
What was so challenging for the ECB is that the crisis morphed from a liquidity crisis into a solvency crisis, which was then followed by the 'Great Recession'. Thereafter, in May 2010, the sovereign debt crisis of the euro area erupted, halting a long period of low and homogeneous financing costs. Euro area countries were no longer treated equally in terms of sovereign risks, which had been persistently underpriced for some. The euro area was subsequently faced with a sudden hybrid crisis, combining the features of a financial crisis in some countries with those of a balance of payments crisis or sluggish growth in others, which overshadowed a group of countries. All these factors have struck Europe before, but never all at the same time, and in so many countries with a single currency, limited adjustment mechanisms and no supranational crisis management framework for sovereigns or banks which might have stemmed some of the contagion. At times there were strong risks of a financial meltdown in some financial market segments, or a financial implosion in a euro area country, or even a break-up of the euro area. What contributed to making things worse is that flaws in the governance of the euro area became obvious--these are alluded to by references to the 'incompleteness of the Economic and Monetary Union (EMU)'. The dysfunctional political debate across countries, increasing mistrust in European institutions and time taken to agree upon institutional reforms also did not help.
The aim of this paper is to guide the reader through the ECB's monetary policy decisions in increasingly fragmented financial markets. What do we mean by financial fragmentation? In general, financial fragmentation captures a decline in competition in respect of cross-border banking activity. This might be the result of a growing national bias leading to a stronger household and investor preference for domestic banks, at the expense of foreign banks. Yet, this definition does not fully capture the drama in the euro area. Why? This is because financial fragmentation in the euro area was exacerbated by a sudden stop of cross-border transactions within the banking system. For example, cross-border claims of euro area banks on monetary and financial institutions (MFIs) located in other euro area countries decreased by 670 billion [euro] between September 2008 and September 2012, as compared with only 206 billion [euro] for domestic claims. Similarly, for the whole euro area banking sector, loans to the domestic non-financial private sector increased by 570 billion [euro], whereas cross-border loans decreased by 450 billion [euro] over the same period (see Laeven and Tressel, 2013).
There were 'doom loops' between weak sovereigns and weak banks in ever more distressed economies, with the:
(a) freezing of most unsecured transactions due to mistrust of counterparties' financial health;
(b) freezing of transactions in money market structured instruments (asset-backed securities, certificates of deposit, commercial paper and covered bonds in the first instance) due to mistrust of the quality of certain underlying assets;
(c) freezing of market activity in the sovereign bond markets of stressed countries due to the possible prospect of default and/or accentuated expectations of contagion.
Such acute fragmentation was combined with a bank run in some distressed banking systems--like in Greece and Ireland--and a flight of households' and investors' funds to 'safer' national banking systems, such as those of AAA and AA-rated countries. As discussed in the paper, such tensions and heterogeneity in financial conditions started to slowly recede only in the Summer of 2012 (see Constancio, 2013).
What mattered greatly to the ECB is that such fragmentation endangered the transmission of the (single) monetary policy stance along the yield curve. This has put the ECB in a more complex position than, for example, the US Fed, the Bank of Japan (BoJ) or the Bank of England (BoE). It is such fragmentation that explains in large part the distinct response of the ECB with respect to the other major central banks in recent years.
As already mentioned, the tension relented only in the Summer of 2012. This paper tells the story of what happened in three steps by:
(d) describing the evolution of financial integration since the launch of the euro and some of the factors that have enabled self-fulfilling speculative attacks;
(e) explaining how the ECB's monetary policy was conducted between August 2007 and mid-2013 in such a challenging and vulnerable environment. Some limited comparisons with the US Fed, BoJ and the BoE are also made;
(f) reporting some empirical findings concerning the functioning of the two main transmission channels before and during the crisis.
Although the intermediation role grew for all central banks, through changes in the composition and size of their balance sheets, the underlying motivation was sometimes quite different. Indeed, whereas the US Fed and the BoE implemented non-standard measures as a substitute for further cuts in the policy rate in a zero lower bound environment, the measures adopted by the ECB were aimed at reducing the impact of increasing financial market fragmentation.
The paper is organised as follows. The section 'How did financial integration evolve: the implications of the institutional set-up' reviews the evolution of financial integration in the run-up to the crisis, the main phases of the crisis and then the rise of financial fragmentation across the euro area; events that can only be understood against the backdrop of the EMU's very unique institutional setting. The section 'What was the ECB's response to financial fragmentation?' recaps the ECB's main responses to the financial crisis in general, and to financial fragmentation in particular. The purpose of this section is to describe the various measures implemented by the ECB and to present (not evaluate) the ECB's motivation as well as the constraints involved. The section 'How did the main transmission channels evolve?' examines the evolution of the two main transmission channels of monetary policy decisions under discussion, namely the interest rate and credit channels. Both were severely impaired by the financial tensions that emerged across the various phases of the crisis, and recovered only after the Summer of 2012. The section 'Concluding remarks' concludes. It should be noted that, as the crisis is not over yet, various caveats and qualifications would apply.
HOW DID FINANCIAL INTEGRATION EVOLVE: THE IMPLICATIONS OF THE INSTITUTIONAL SET-UP
The early years with the euro (from 1 January 1999 to 8 August 2007)
Following the launch of the euro on 1 January 1999, the national money markets of the 11 founding countries that jointly adopted the euro merged into a single integrated money market, which now encompasses 18 countries. This initiated a long period of low and homogeneous financing costs across the whole euro area. Reference prices in the various segments of the money markets rapidly converged across euro area countries, and some new reference rates were also created. These play an important role in the discussion and are listed below.
(g) For the unsecured cash segment, the two main reference interest rates became the euro overnight index average (EONIA) for overnight lending transactions and the euro interbank offered rate (EURIBOR) for interbank deposits with a maturity of 1 week and 1-12 months.
(h) For the derivatives segments, the main references became the EONIA swap interest rates (overnight index swap (OIS) rates) and EURIBOR futures prices. (1)
Financial integration materialised very quickly through a significant increase in cross-border interbank lending and an acceleration of mergers and acquisitions in the euro area, which also reflected a growing trend around the world. (2) As an illustration, claims of euro area banks on other European banks were of a magnitude of up to 70% of Europe's gross domestic product (GDP) in 2007, out of which 30% were related to cross-border claims. Sovereign spreads also narrowed down to within a few tenths of basis points while banks' financing costs were rather similar across countries. As demonstrated by Figure 1, financial market segmentation was non-existent. Figure 1 is based on median absolute deviations across several euro area financial market segments. It shows, for instance, the absence of cross-country dispersion within the euro area in the pre-crisis period, namely from January 2004 to July 2007 (see the solid black line). Indeed, bank lending rates, 10-year government bond yields, banks' cost of funding and volumes of both loans to the private sector and interbank deposits were homogenous within the euro area. Furthermore, it was a period during which no Target2 imbalances materialised.
By contrast, equity markets and corporate bond markets did not merge immediately following the introduction of the euro. Yet, over time, increased cross-border financial cooperation and investments were observed. The rapid integration of wholesale capital markets contrasts with more modest changes on the lending side: most banks basically kept their exposure, and thus risks, domestic. The euro area has also been a magnet for internal foreign direct investment activities. Companies and, to a lesser extent, banks and other financial institutions started to have a more continental reach. Integration among banks made slow progress at first, but gathered momentum later on. Integration was faster in the case of the sovereign bond market. Substantial cross-country holdings of public debt instruments issued by euro area governments built up (as would be expected in a monetary union). The proportion of domestic investors holding national government bonds in the various euro area countries gradually decreased (see the ECB report on 'Financial Integration in Europe' of April 2013). International ownership of the public debt of euro area countries increased from 32.5% in 1999 to 53.5% in 2009.
Deeper and more integrated financial markets contributed to a compression of spreads in money and sovereign bond markets (see Figures 2 and 3). In the money market, the spread between the unsecured interest rate (ie, the EURIBOR) and the secured (OIS) interest rate declined quite markedly to a few basis points for all maturities after the launch of the euro and remained at that level until the emergence of the financial crisis in August 2007 (see Figure 2). In the case of the sovereign bond market, the reduced price of risk as characterised by the spread between the yields of various 10-year government bonds and the corresponding German Bund--was substantial, even exceeding the expected reduction coming solely from the disappearance of the foreign exchange risk premium (see Figure 3). The narrowing of sovereign spreads also reflected strong investor confidence that the euro would be as safe and as strong a currency as the strongest legacy currencies--first of all, the Deutsche Mark. To a certain extent, this also implicitly reflected confidence in the institutional framework surrounding the euro.
Crisis times (from 9 August 2007 onwards)
On 9 August 2007, the announcement by BNP Paribas, a big global money market player, that it had frozen redemptions for three of its investment funds, marked the start of growing distortions in financial markets through tensions gradually affecting the fair value of various financial assets, which invariably systematically deteriorated the liquidity position of banks. The 'confidence crisis' among market participants came from: (a) uncertainty about the quality of counterparty balance sheets; and (b) uncertainty about the quality of structured assets on their own balance sheets. This created incentives among market participants to hoard liquidity, hence dramatically reducing market activity in various segments (Heider et al, 2009). As displayed by Figure 2, spreads between the unsecured EURIBOR interest rate (this being a proxy of liquidity and credit risks in the money market) and secured OIS interest rates (these being a proxy of monetary policy expectations) skyrocketed in the light of market disruptions. These spreads started rising above levels that would be consistent with the monetary policy stance desired by the central bank. The conditions for a normal liquid market were then no longer fulfilled, which led to weaker representativeness of financial asset prices.
These phenomena intensified over time, subsequently taking the form of a short-term money market crisis, a 'too big to fail' crisis and a sovereign bond crisis centred on the euro area. Since the cost of interbank credit represents the first step in the transmission process, its significant and persistent deviations from the monetary policy stance defined by the key ECB interest rates could render the ECB unable to shape the yield curve and to anchor the market's expectations over time.
What happened was that the financial turmoil (from 9 August 2007 to 14 September 2008) affected the prevalent bank funding model by introducing distortions between collateralised (or secured) and uncollateralised (or unsecured) interbank transactions. As shown in Figure 1, this first crisis phase marked the start of growing financial market fragmentation within the euro area. In practice, financial fragmentation materialised in the greater dispersion across euro area countries, both in terms of pricing of key financial instruments (namely lending rates, banks' costs and bond yields) and of volumes. (3) This initial liquidity crisis then evolved into a full-blown crisis of confidence related to growing solvency concerns in the second phase: the global financial crisis (from 15 September 2008 to 7 May 2010) during which all financial institutions felt vulnerable (Brunnermeier, 2009; Reinhart and Reinhart, 2010). This in turn dramatically intensified the liquidity crisis--freezing some money market segments and the covered bond market--which, in contrast to the first phase, ultimately affected prospects for economic growth (for a detailed analysis, please refer to Drudi et al., 2012).
Finally, the most recent phase, the euro area sovereign debt crisis (from 8 May 2010 onwards), (4) strengthened the concerns manifested in the two previous phases through a mutually reinforcing negative spiral between sovereign and banking risks all over the euro area--the so-called 'diabolic feedback loop' (see Shambaugh et at, 2012). In addition, the initial tensions in the Greek sovereign debt market precipitated doubts about the sustainability of public finances and economic prospects in other euro area countries via a contagion spiral. This was soon followed by a crisis in Ireland and then Portugal. Other countries followed suit and five euro area countries finally ended up with various types of adjustment programmes (albeit Ireland exited its programme in mid-December 2013 and Spain left its banking sector adjustment programme on 23 January 2014). As of early 2010, capital flows reversed and there were sudden stops in financing and even runs on some banks. This is when financial fragmentation started escalating and seriously impairing the transmission of the monetary policy impulses that euro area economies needed.
Liquidity and credit risk premia soared in stressed economies and spreads widened (see Cassola et al, 2011). Increasing interest rates for financial products perceived as riskier by the market did not reflect the monetary policy stance, implying a de facto tightening of monetary conditions. These developments not only blurred monetary policy signals, but also seriously impaired their transmission to the money market. Furthermore, the risk that the central bank would become unable to shape the yield curve was quite real for at least two reasons. First, the reduced activity in the cash segments of the euro area money market (both in the EONIA and the EURIBOR segments) could negatively affect the activity and thus price representativeness in the derivatives segments (both in terms of swap and futures contracts). Hence, the risk of a potential de-anchoring of the money market yield curve was serious. Second, pricing problems with structured financial products and public debt instruments, which represent a large part of banks' assets and thus affect their liquidity ratios, could have ultimately led to an uncertain and volatile determination of interest rates in the economy.
These risks are simply illustrated by Figures 2,3,4 and 5. Figure 4 shows that due to the liquidity hoarding phenomenon, whereby banks are encouraged to hold liquidity well above their liquidity needs, (5) the uncollateralised overnight interest rate (EONIA) dropped below the main monetary policy interest rate for collateralised ECB refinancing operations. This deviation of the overnight interest rate became particularly striking when the ECB decided to accommodate all bank liquidity demands. (6) At the same time, the apparent cheapness of overnight uncollateralised transactions did not support market activity since the EONIA volume was also decreasing. (7) In the light of the deterioration in banks' liquidity ratios caused by distorted money markets and illiquid sovereign bond markets, adverse effects on both the supply and pricing of bank loans to the real economy were substantial (see Figure 5). Figure 5 reports the composite nominal lending rate to non-financial corporations (NFCs) (left-hand chart) and the annual growth rate of volume (right-hand chart) for various euro area countries, together with the euro area average. As illustrated by the grey area in both charts, the cross-country dispersion increased significantly from 2008 onwards. Last, but not the least, the emergence of tensions in the sovereign bond market (see Figure 3) also renewed stress in the money market (see Figure 4), reflecting a 'feedback loop' between the sovereign bond market and the banking sector. (8) These pressures further accentuated financial market segmentation, as reported in Figure 1. (9)
Some of the root causes of the sovereign debt crisis
What were some of the root causes of the crisis, how are they being addressed and what might be their legacy?
During the first 10 years of the euro, euro area countries became more interconnected, but not all went according to plan. Nonetheless, it is worth stressing that the whole institutional architecture to secure a well-functioning EMU was dependent on the respect of strict ex ante rules for effective coordination between monetary and fiscal policies in the absence of a federal budget. Three 'fault lines' were allowed to build up over time. The first is the result of weak public finances: the favourable first decade with the euro was not used to reduce public debt ratios faster in the countries that later became 'stressed'. The second fault line involves persistent current account imbalances --and an erosion in competitiveness--in another overlapping group of countries. The third fault line is the slow pace of productivity growth and overall GDP growth--which require structural reforms and innovation--in some other euro area countries. The global financial crisis and the ensuing Great Recession (note that this originated elsewhere) exacerbated these fault lines.
Moreover, several preventive mechanisms put in place to hinder 'breaks' at the above fault lines did not work as expected. In particular, when action was needed, the preventive arm of the Stability and Growth Pact (SGP) did not work as intended. At the time, the European Commission did not have enough support to fully enforce the corrective arm of the SGP and to push for more timely reductions in deficits and debt at the national level. Similarly, financial market participants and credit rating agencies did not discriminate sufficiently between national issuers with different legacy debt and credit rankings during the first decade of the euro. Hence, both European Union institutions and financial markets did not encourage faster deficit and debt reduction where necessary.
In Spring 2010, the belief that euro area membership would act as a shield against exchange rate volatility and credit risks vanished with the start of the sovereign debt crisis in Greece. That is the point when flaws in the EMU's design started emerging. Two design flaws stand out, in particular. The first is that the governance of the EMU was designed without a framework to deal with a sovereign debt crisis. The second design flaw is that a common resolution framework for large pan-European banks, the so-called 'systemically important financial institutions', was also missing. Such frameworks were still national; however, several banking groups had already branched out across countries and a few had even acquired balance sheets worth several multiples of their home-country GDP. The EMU had initially been built without some necessary financial firewalls and backstops. Hence, it is now widely agreed that the incomplete institutional setting of the euro area contributed to spreading and accelerating the sovereign debt crisis. It certainly enabled speculative attacks against stressed peripheral euro area countries. Naturally, the latter phenomenon has strengthened financial market segmentation further by widening the median absolute deviations in sovereign bond yields and banks' funding costs, while making Target2 imbalances more persistent (Figure 1).
Here, it is worth recalling that the EMU was designed over 20 years ago without a risk-sharing mechanism to absorb asymmetric macroeconomic shocks. There is no common facility to help in buffering, and then resolving, sovereign debt and banking crises. With hindsight, the EMU's institutional framework was inadequate at the onset of the 2007-2013 financial crisis (see Drudi et al., 2012; Mongelli, 2013). What was the logic behind this? The implicit logic was to ensure the conditions for an active monetary policy and a passive fiscal policy regime (in the sense of Leeper, 1991) within the EMU, whereby, ex ante, governments assure debt sustainability and the central bank focuses on maintaining price stability over the medium term, as discussed in detail by Durre and Smets (2014).
However, after it became obvious that fiscal rules were not respected and that idiosyncratic macroeconomic shocks reduced economic convergence, the absence of ex post correction mechanisms in the EMU institutional framework raised questions about the irreversibility of the euro and led to the sovereign debt crisis spreading from Greece to other countries (the contagion effect) through the phenomenon of 'self-fulfilling speculative attacks' (Giavazzi and Pagano, 1990; Alesina et al, 1992).
WHAT WAS THE ECB'S RESPONSE TO FINANCIAL FRAGMENTATION?
For all central banks, it is fundamental that monetary policy decisions are transmitted effectively and smoothly to the economy in general, and to private sector price-setting decisions in particular. This in turn requires well-functioning financial markets. In fact, as witnessed after August 2007, there was a rapid deterioration in the liquidity and solvency positions of many market players, followed by a fragmentation of various financial market segments. Euro area countries were not all affected equally by these tensions. Fragmentation went hand-in-hand with national legislation and/or regulations shielding the domestic banking sector, and was thus determined by local considerations. Given such exceptional circumstances, it became evident that it was necessary for central banks to adopt non-standard measures. In particular, it appeared crucial to react rapidly to preserve the effectiveness of the transmission of monetary policy through the yield curve. Indeed, the mitigation of financial market dysfunctionalities via non-standard measures aims at maintaining the distribution of liquidity among financial institutions and ultimately to the real economy.
By nature, the actions of the central bank in charge of the single currency area are providing an implicit risk-sharing mechanism to preserve the integrity of the common monetary policy among euro area countries. In the euro area, the observable part of this risk-sharing mechanism--as reflected by the intra-Eurosystem claims and liabilities referred to in Target 2 balances--has recently led to a lively debate among economists (see, for instance, Sinn, 2012 and Bindseil et al., 2012), although these balances would not appear problematic as long as no member country exits the currency area. To support the latter view, it is recalled that Target2 balances have been declining in recent months on account of improvements in both the euro area money market and national banking systems. At the same time, one should refrain from the tendency to believe that the central bank has to resolve structural imbalances in the economy and within the single currency area.
In the ECB's view, the main goal of the 'non-standard' measures was to preserve the effectiveness of the interest rate, bank lending and financial asset channels. (10) 11 This naturally increased the ECB's intermediation role via a temporary substitution of distorted market segments (when necessary) to ensure a continuation in the provision of bank loans to the real economy at conditions in line with the monetary policy stance decided by the Governing Council. In fact, the prospect of a decline in banks' deposits together with reduced access to market liquidity, if not compensated for by other sources of funding, would have strengthened the constraints on the asset side of banks' balance sheets. Combined with the decreasing value of structured products and other financial assets, the likelihood of a disorderly deleveraging by banks (eg, via fire sales) was a real risk--and would eventually have led to a negative impact on the amount of bank lending to the economy.
The increased number of refinancing operations (ie, via more frequent fine-tuning operations in the first phase and then the full accommodation of banks' liquidity needs (11) by the ECB in the second crisis phase) aimed at reassuring market participants, and ultimately banks, about the availability of liquidity independently of prevailing market conditions and uncertainties. By increasing the weight of longer-term refinancing operations in the total amounts provided, the ECB has aimed at reassuring market participants about their access to liquidity in the longer term. Interestingly, it has recently been observed that when there are tensions in the money market, or market participants' uncertainties about the availability of liquidity receded, banks' appetite for excess liquidity also decreased.
In the same vein, the implementation of securities purchase programmes, such as the Covered Bond Purchase Programme and the Securities Markets Programme (SMP), was required given the sudden stop of market activity in the respective segments and the above-mentioned possible negative impact on bank activity. In the specific case of the SMP, the main goal was to prevent the sovereign debt crisis from causing a systemic financial crisis, while making it clear that such non-standard measures were temporary in nature and conditional on the affected governments reforming their public finances. This conditionality was made explicit in the last asset purchase programme (12) announced on 6 September 2012, geared at reducing the tail risks associated with fears of the reversibility of the euro (also, for an explanation of the recently announced 'forward guidance' policy, see Praet, 2013). The last 3 years have indeed been characterised by increasing heterogeneity and various market distortions across stressed and programme countries, as well as various types of borrowers and lenders.
Meanwhile, in the light of deteriorating economic developments, it was also necessary to support economic growth and to avoid deflationary pressures through 'standard' policy measures, namely the determination of the key ECB interest rates. Having a more accommodative monetary policy stance was thus necessary to avoid a sudden and drastic (demand-driven) reduction in bank lending in the economy. (13)
In short, the main goal of both standard and non-standard measures was to ensure the continuation of lending flows in the economy between the initial lender and the final borrower, while insulating the real economy from tensions in financial markets. (14) Beyond the standard portfolio balance approach to understanding central bank behaviour in times of crisis, the ECB's actions must be seen as essentially aiming to overcome market malfunctioning. This explains why, in addition to an increase in the frequency and maturity of direct refinancing operations, the ECB intervened in specific market segments through asset purchase programmes. Of course, the natural consequence of market substitution by the ECB is an expansion in the size of its balance sheet. Certainly, the size of the ECB's balance sheet increased from a level of 10% of euro area GDP in the pre-crisis period to a peak of more than 50% in 2013, before falling back to its current level of around 40%. It should also be recalled that the crisis has multiple causes: persistent imbalances, governance failings, incomplete institutions and so forth.
The ECB has been among the advocates for change, but there is now acceptance of reforms and the need for strengthening institutions (Coeure, 2013). As a result, elements of a new 'constitutional framework' are emerging and a new political economy is now within reach. There are financial backstops for banks and sovereigns, reducing the likelihood of future systemic risks. These responses to the crisis emerged from a series of EU and euro summits and joint decisions: in fact, there is now a financial backstop as the temporary European Financial Stability Facility has been transformed into the permanent European Stability Mechanism, which has a number of instruments at its disposal.
The prospect of a new political economy is rather encouraging. The so-called 'six-pack' of legislative changes has established, among other things, a new Macroeconomic Imbalance Procedure that will complement the revised SGP. There is also a new macro-prudential supervisory framework centred on the European Systemic Risk Board: it can issue warnings and macroprudential recommendations whenever necessary. And a Single Supervisory Mechanism will start operating later this year under the responsibility of the ECB (see Mersch, 2013).
HOW DID THE MAIN TRANSMISSION CHANNELS EVOLVE?
Acute financial strains generated persistent deviations of the cost of retail loans from their fundamentals, countervailing the accommodating monetary policy stance. Expectations became unstable and risk premia became extremely volatile. Since the banking system in the various euro area countries faced different stress conditions, the interest rates applied by banks in their retail banking activity started to differ from one country to another (see Figure 5). In these circumstances, the risk existed that the pricing of real economy lending in the euro area would become de-anchored from the monetary policy stance decided by the ECB. As explained in the section 'What was the ECB's response to financial fragmentation?', such risks became the main concern of the ECB. They also explain the course that it has taken since August 2007. Naturally, the question arises whether or not the main transmission channels of monetary policy decisions in the euro area did in fact break down. Two channels appear of particular relevance in this respect, namely the interest rate channel and the bank credit channel. The purpose of this section is thus to discuss the main concerns regarding both channels and to illustrate the crisis dynamics.
Interest rate channel
Empirical evidence for the pre-crisis period suggests that the interest rate channel is a powerful transmission channel in Europe in general, and in the euro area in particular. (15) In the light of the various tensions that emerged in various segments of the financial market, the magnitude of the severity of the interest rate channel's impairment appears relevant. Hence, the paper now shows the degree of impairment of this important transmission channel against the backdrop of the ECB's actions. In other words, what was its performance during the crisis in comparison with the pre-crisis period? To answer, the relationship between a retail interest rate (the dependent variable) and both short- and longer-term interest rates (the former being used to quantify the immediate pass-through and the latter to quantify the final pass-through) as explanatory variables is estimated by applying a multivariate equation framework. This type of relationship involves the pass-through of market interest rates to the interest rates applied for retail banking activity. More specifically, we estimate the interest rate pass-through via a vector error correction mechanism model using the following variables: the interest rate applied by banks to mortgage credits; the interest rate on consumption loans (with a horizon of more than 1 year); the EURIBOR and the 10-year yield on government bonds. In the pre-crisis context, it can be argued that the short-term (EURIBOR (16)) and the long-term (10-year public bond yields) market interest rates are used to evaluate the degree of distortion in the pricing of retail banking products.
If, in the crisis period, loan interest rates would no longer be explained by the expected arbitrage against the market, this would reflect severely impaired transmission; something that would perhaps put into question the ability of the ECB to control the shape of the yield curve.
For this exercise, we estimate the relationship between these interest rates for just two countries with a very different experience of risk pressures, namely Germany and Spain. (17) The following equation is thus estimated using a two-step approach (Engle and Granger, 1987):
[MATHEMATICAL EXPRESSION NOT REPRODUCIBLE IN ASCII] (1)
where [[epsilon].sub.t] represents the long-term relationship and is obtained from the following specification (18):
[[epsilon].sub.t] = [r.sub.t] - [[alpha].sub.0] - [[alpha].sub.1][mrs.sub.t] - [[beta].sub.1][mrl.sub.t] + [u.sub.t] (2)
with [r.sub.t] denoting the retail bank interest rate, [mrs.sub.t] being the short-term market interest rate (ie, the EUR1BOR), [mrl.sub.t], being the long-term interest rate (ie, 10-year government bond yields) and [u.sub.t] the residuals of the long-term relationship. By design, the set of coefficients in equations 1 and 2 denote, respectively, the immediate ([[delta].sub.0] and [[phi].sub.0]) and the final [[[alpha].sub.1] and [[beta].sub.1]) pass-through of the corresponding variable. All these coefficients are expected to be significantly different from 0 and positive. The coefficient [rho] refers to the (error correction) adjustment coefficient and is expected to be statistically significantly negative. Data are collected from the ECB and Thomson Reuters. Equations 1 and 2 are estimated over the sample period ranging from January 2003 to December 2012 on a monthly basis, that is, there are 120 observations in total. The final results are reported in Table 1.
The following observations can be made. The related empirical evidence for the pre-crisis period, namely the time before August 2007, generally points to a strong (both immediate and final) pass-through of changes in policy rates to retail bank interest rates, albeit with some differences across countries. These differences between short- and long-term elasticity are usually explained by banks' strategy, the state of their balance sheets and by fundamentals (eg, the business cycle and inflation), against the background of some cross-country heterogeneity. Interestingly, empirical studies suggest that the introduction of the euro has strengthened the pass-through recorded in various countries before the euro era (see, eg, Sander and Kleimeier, 2004).
In general, the results reported in Table 1 for the pre-crisis period from January 2003 to July 2007--based on an updated version of the sample period with harmonised data for the euro area as a whole--are very close to those reported in previous studies based on national data. In normal times, both the short-term money market interest rate (3-month EURIBOR) and the long-term bond interest rate (10-year bond yield) play a significant explanatory role in the determination of retail interest rates.
Looking specifically at the crisis period, several observations can be made. First, it appears that the error correction mechanism model still provides an explanation of retail bank interest rates--[rho] coefficients are still significantly different from zero and negative for both countries. (19) However, it is worth noting that, in Germany, the adjustment process captured by this coefficient seems to be broadly unchanged for the interest rate applied to long-term mortgage loans, whereas it decreases somewhat for the interest rate applied to loans to NFCs. In Spain, it decreases more substantially for both types of loans. This decrease thus suggests that banks may have taken account of other considerations in the short term for the pricing of loans to the real economy without fundamentally affecting the cointegration relationships. Second, the evidence in each country differs significantly when it comes to short- and long-term elasticity. In Germany, elasticities tend to remain unchanged (short term) or to decrease (long term) for both types of loans. By contrast, pass-through in Spain tends to increase for both types of loans.
These findings might be related to the nature of loans and the related inherent competition in the market and/or to idiosyncratic developments in each country. On the one hand, competition for loans to NFCs is usually more intense than for mortgage loans, which are nationally based. On the other hand, German government bonds benefited from a 'flight to safety' phenomenon pushing the yield for some maturities to zero or below, which may explain why banks have attached more importance to money market developments when setting the price of their mortgage loans. In Spain, the influence of money market and government bond
rates tends to increase during the crisis period in both the short term and the long term, independent of the type of loan involved. This may, for instance, reflect the deterioration of Spanish banks' funding and a need to rebuild profitability.
To sum up, these results confirm an increased heterogeneity in the determination of retail interest rates for loans to the real economy across both countries during the crisis period. For example, taking just the case of the bond market, decreasing yields in Germany have allowed cheaper loans to the German real economy while, in Spain, upward pressures on yields have naturally made loans in this country more expensive. In general, fluctuations in the short-term interest rate remain of significant importance, perhaps reflecting the influence of non-standard monetary policy measures in terms of limiting the extent to which the transmission of the monetary policy stance to the real economy is impaired.
Impairment of the bank credit channel through increased heterogeneity across both groups of countries and borrowers and lenders
Let us now turn to the analysis of the bank credit channel, which is another important transmission channel in the euro area. According to theory, informational and contractual frictions between lenders and borrowers--a common feature of bank credit markets--tend to worsen significantly in periods of stress in financial markets and when monetary policy is particularly restrictive (Bernanke and Gertler, 1995). In such situations, the external finance premium, that is, the premium that borrowers have to pay to obtain external financing, increases and amplifies the effects of changes in monetary policy rates. This implies that monetary policy can have real effects through both credit supply and demand. It has been shown, for example, that a tightening of monetary policy reduces the supply of loans by increasing banks' external financing costs, which in turn impacts their capacity to extend credit (see, eg, Allen et al, 2004). At the same time, the demand for loans declines due to the higher external finance premia faced by borrowers (the firm and household balance sheet channel) and the higher direct cost of loans (the classical interest rate channel).
Ciccarelli et al. (2011) focus on the transmission of monetary policy impulses in the euro area during the financial crisis by investigating responses to the Bank Lending Survey (BLS). (20) Since lending standards and loan demand may react to business cycle fluctuations, the information from the surveys is embedded into a vector autoregressive model to account for the linkages between the credit and the business cycle. (21) And the credit channel is identified by using survey data on bank lending standards, namely the conditions that banks apply to borrowers depending on their balance sheet position, competitive pressure and the risk profile and creditworthiness of borrowers.
The evidence reported by Ciccarelli et al. (2011) suggests that frictions in credit markets increase in periods of stress, and changes in market interest rates have an amplified effect on the ability and willingness of banks to extend credit to borrowers. Indeed, these frictions impact both credit supply and demand. The increased recourse to secured financing and the refinancing operations of the ECB allowed these pressures to ease somewhat, but the emergence of the sovereign debt crisis again worsened the situation, as banks in countries under stress found it increasingly difficult to secure financing using national government bonds. Furthermore, this analysis suggests that the credit channel is operational and amplifies the impact of a monetary policy shock on GDP and inflation via the balance sheets of households, firms and banks. This is true for all types of loans, with differences in the size and timing of the impact across borrowers. In the euro area, for business loans, the amplification of monetary policy shocks is higher via the bank lending channel than via the demand and balance sheet channels. Credit demand is the most important channel for mortgage loans. (22)
As regards the situation during the crisis, the study implies that tighter lending standards for firms, due to weaker bank capital and liquidity positions, contributed significantly to the decline of GDP growth, thus suggesting that there was a credit crunch for firms with very real implications. The immediate policy actions undertaken by the ECB, notably a significant reduction in policy interest rates and the implementation of a full allotment policy for repo refinancing operations, have supported GDP growth by relaxing the balance sheet constraints of banks. Maddaloni and Peydro (2013) show that lending standards were relaxed more for banks that borrowed more long-term liquidity from the ECB. Therefore, the results suggest that monetary policy rates and public provision of long-term liquidity have complemented each other in reducing a credit crunch for firms.
What about the issue of heterogeneity? To address this issue, Ciccarelli et al. (2013) look at changes in monetary policy transmission that have occurred over time during the crisis periods, against the backdrop of possible heterogeneity across countries and agents. In particular, to take into account the effect of the sovereign debt crisis, the countries of the euro area are divided into two groups: countries under stress and other countries. These two groups are then divided according to the value of the CDS spreads on government bonds. Interestingly, Figure 6 shows that, already in 2008, the dispersion of the CDS spreads was increasing across the euro area and countries like Italy and Spain were showing signs of distress by this measure, even though worries about the sustainability of their public debt were not an issue at the time.
In the exercise mentioned above, the transmission of monetary policy is analysed through different sub-channels, namely the bank lending channel and the borrower balance sheet channel. The amplification effect of monetary policy throughout the credit channel is significant (both in statistical and economic terms) only for the countries under stress. Moreover, transmission via the bank lending channel is important until the end of 2009, suggesting that the non-standard measures implemented by the ECB, in particular the provision of long-term liquidity, were somewhat able to neutralise the liquidity frictions affecting bank balance sheets by reducing deleveraging pressures and therefore supporting balance sheet capacity. From mid-2010 onwards, the recursive correlation between (private) interbank volumes and (public) long-term ECB liquidity provision became negative for the countries under stress. This suggests a substitution of private interbank liquidity for ECB liquidity for these countries whereas, in the other group of countries, the two sources of funding remained complementary (see Ciccarelli et at, 2013). However, frictions due to the risk and creditworthiness of borrowers (the borrower's balance sheet channel) remained acute, particularly in countries under stress, and so did the related amplification effect of monetary policy shocks. In this respect, the deleveraging process reinforced tensions in some cases. On the one hand, deleveraging was necessary to put banks' balance sheets into good order, but it also had the potential to exacerbate the ongoing recession, possibly through a credit crunch.
Last, but not the least, Ciccarelli et al. (2013) find uneven impacts on smaller banks and small- and medium-sized enterprises (SMEs), that is, size matters in the transmission of monetary policy through the credit channel (see also Gertler and Gilchrist, 1994 and Kashyap and Stein, 2000). Smaller banks--which lend primarily to smaller firms--are typically more financially constrained, as they may have more limited access to the wholesale capital market and may be more affected by a freezing of the interbank market. This is why changes in monetary policy rates are likely to affect more the credit granted to smaller firms.
Indeed, the analysis conducted by Ciccarelli et al. (2013) shows that, during the crisis, the amplification effect of a monetary policy shock has proceeded mainly through the non-financial borrower balance sheet channel of small banks. This means that, in distressed countries in particular, the lower net worth of SMEs, due to the higher discount rate applied to the valuation of their assets, makes lending to these firms highly unattractive for banks, notwithstanding the actions of the central bank. The non-standard monetary policy measures adopted by the ECB so far, including allowing the use of loans to SMEs as eligible collateral in central bank repo operations, may have had only a limited impact here. The very high financial fragmentation across countries and borrowers has not been overcome. Consequently, further policy actions, such as some form of direct lending to the corporate sector (eg, an adaptation of the funding for lending schemes implemented by the BoJ and the BoE), are still openly discussed within and outside the ECB--and may prove beneficial in the euro area.
In crisis times it is expected that central banks become more active by stepping up their intermediation role. This holds in particular when the roots of the crisis are as complex as in the case of the sovereign debt crisis of the euro area. Yet, the ECB has faced a wholly new and dramatic situation, namely liquidity problems in large parts of the banking system of stressed countries and impairments in the transmission of monetary policy. Moreover, at some points, growing fears of a break-up of the euro area have enhanced financial fragmentation and vice versa.
Three particular comments can be made in response to the above. First, the mutating crisis has motivated the ECB to implement various non-standard measures since August 2007 in order to foster a smoother transmission of the monetary policy stance along the yield curve. Without such transmission, changes in the monetary policy stance would have no impact on the real economy, both in terms of prices and economic growth. Thus, in crisis times, non-standard measures help support the effectiveness of standard measures, that is, the setting of short-term ECB interest rates when financial markets are fragmented.
Second, even under such difficult circumstances, a pass-through of market interest rates to retail bank interest rates, as well as the bank lending channel, has been preserved throughout the crisis. While, on average, it has decreased somewhat in comparison with the pre-crisis period, the pricing of retail bank interest rates has remained stable in both Germany and Spain, despite the differences in financial market developments. In addition, although a counterfactual exercise is not unproblematic, it could be argued that the continuation of a standard level of pass-through was probably made possible through the significantly increased intermediation role of the ECB, which substituted the market to a large extent. This finding is in line with various studies suggesting that the ECB's measures during the financial crisis assisted in avoiding a more severe credit crunch (and the possible related deflationary effects) through the stabilisation of inflation expectations and the provision of support to economic activity. (23)
Third, there should be limits in the scope and duration of a central bank's interventions in order to safeguard its mandate. This holds in particular for the ECB, which has a narrow mandate limited to maintaining price stability over the medium term and operates within an institutional framework with a clear demarcation of responsibilities between monetary and fiscal authorities--one that has served as an anchor for expectations throughout the crisis. Moreover, the provision of liquidity to illiquid but solvent financial institutions must be seen as a natural, if not necessary, central bank task in times of crisis. By contrast, it should not be expected that the central bank can or should solve structural imbalances in the economy.
Thus, a credible single monetary policy can provide a shield against outside shocks, support the functioning of the financial system and help promote internal stability. Yet, in the long run, the sustainability of the EMU will depend on the completion of a new governance framework and structural reforms (where necessary). Addressing institutional weaknesses is also required to prevent the likelihood of a future systemic financial crisis centred on the euro area.
The authors would like to thank Paul Wachtel, Lucio Vihnas de Souza, Tom Kokkola, Jyoti Patel and two anonymous referees for helpful comments and suggestions. A special thanks to Marco Corsi, Julie Sipura, Vozinos Anastasios and Eric Rodriguez Ramos for research assistance. The views expressed are those of the authors exclusively and do not necessarily reflect those of the ECB or of the Eurosystem. The paper only reflects data and information up to the end of 2013.
Alesina, A, De Broek, M, Prati, A and Tabellini, G. 1992: Default risk on government debt in OECD countries. Economic Policy 15(7): 427-463.
Allen, F, Chui, M and Maddaloni, A. 2004: Financial systems in Europe, the U.S.A. and Asia. Oxford Review of Economic Policy 20(4): 490-508.
Andrews, D. 1991: Heteroskedasticity and autocorrelation consistent covariance matrix estimation. Econometrica 59(3): 817-858.
Bagehot, W. 1873: Lombard Street: A description of the money market, 3rd Edition. Henry S. King & Co: London Electronic Version.
Bernanke, BS and Gertler, M. 1995: Inside the black box: The credit channel of monetary policy transmission. Journal of Economic Perspectives 9(4): 27-48.
Bindseil, U, Cour-Thimann, P and Konig, P. 2012: Target2 and cross-border interbank payments during the financial crisis. CESifo Forum 13 (Special Issue 2): 83-92.
Brousseau, V, Chailloux, A and Durre, A. 2013: Fixing the fixings: What road towards a more representative money market benchmarking system? IMF Working Paper, No WP/13/131.
Brunnermeier, MK. 2009: Deciphering the liquidity and credit crunch 2007-2008. Journal of Economic Perspectives 23(1): 77-100.
Cassola, N, Durre, A and Holthausen, C. 2011: Implementing monetary policy in crisis times: The case of the ECB. In: Jarocinski, M, Smets, F and Thimann, C (eds). Approaches to Monetary Policy Revisited--Lessons from the Crisis. Session 5. Frankfurt am Main: European Central Bank: 280-321.
Ciccarelli, M, Maddaloni, A and Peydro, JL. 2011: Trusting the bankers: A new look at the credit channel of monetary policy (mimeo). Earlier draft published in the ECB Working Paper Series, No 1228.
Ciccarelli, M, Maddaloni, A and Peydro, JL. 2013: Heterogeneous transmission mechanism: Monetary policy and financial fragility in the Euro area. Economic Policy 28(75): 459-512.
Constancio, V. 2013: Fragmentation and rebalancing in the euro area, Joint EC-ECB Conference on Financial Integration, Brussels, 25 April, available at http://www.ecb.europa.eu/press/key/date/ 2013/html/spl 30425_l.en.pdf.
Corsetti, G, Kuester, K, Meier, A and Muller, G. 2011: Sovereign risk and the effects of fiscal retrenchment in deep recessions. Federal Reserve Bank of Philadelphia Working Paper, No 11-43, September.
Cottarelli, C and Kourelis, A. 1994: Financial structure, bank lending rates and the transmission mechanism of monetary policy. IMF Staff Papers 41(4): 587-623.
Cceure, B. 2013: The political dimension of European economic integration, Speech at the 19th Universite d'automne de la Ligue des droits de l'Homme, Paris, 23 November, available at http: // www.ecb.europa.eu/press/key/date/2013/html/spl31123.en.html.
de Bondt, G. 2005: Interest rate pass-through: Empirical results for the euro area. German Economic Review 6(1): 37-78.
De Santis, RA. 2012: The euro area sovereign debt crisis: Safe haven, credit rating agencies and the spread of the fever from Greece, Ireland and Portugal. ECB Working Paper Series, No 1419.
Dradi, F, Durre, A and Mongelli, FP. 2012: The interplay of economic reforms and monetary policy: The case of the euro area. Journal of Common Market Studies 50(6): 881-898.
Durre, A and Nardelli, S. 2008: Volatility in the euro area money market: Effects from the monetary policy operational framework. International Journal of Finance and Economics 13(4): 307-322.
Durre, A and Smets, F. 2014: Sovereign debt and monetary policy in theory and practice: The case of the euro area. In: Chadha, JS, Durre, A, Joyce, M and Sarno, L (eds). Developments in MacroFinance Yield Curve Modelling. Chapter 2. Cambridge: Cambridge University Press: 56-89.
ECB. 2009: The implementation of monetary policy since August 2007. Monthly Bulletin July: 75-89.
ECB. 2013: Early repayment of funds raised through three-year longer-term refinancing operations: Economic rationale and impact on the money market. Monthly Bulletin February (Box 3): 31-35.
Engle, RF and Granger, CWJ. 1987: Cointegration and error correction: Representation, estimation and testing. Econometrica 55(2): 251-276.
Fahr, S, Motto, R, Rostagno, M, Smets, F and Tristani, O. 2011: Lessons for monetary policy strategies from the recent past. In: Jarocinski, M, Smets, F and Thimann, C (eds). Approaches to Monetary Policy Revisited--Lessons from the Crisis. Session 1. Frankfurt am Main: European Central Bank: 26-66.
Gambacorta, L. 2004: How do banks set interest rates? NBER Working Paper, No 10295.
Giannone, D, Lenza, M, Pill, H and Reichlin, L. 2012: The ECB and the interbank market. CEPR Working Paper Series, No 8844.
Giavazzi, F and Pagano, M. 1990: Confidence crisis and public debt management. In: Draghi, M and Dornbusch, R (eds). Capital Markets and Debt Management. Cambridge University Press: Cambridge, UK.
Gertler, M and Gilchrist, S. 1994: Monetary policy, business cycles, and the behavior of small manufacturing firms. Quarterly Journal of Economics 109(2): 309-340.
Heider, F, Hoerova, M and Holthausen, C. 2009: Liquidity hoarding and interbank market spreads: The role of counterparty risk. ECB Working Paper, No 1126.
IMF. 2013: Financial integration and fragmentation in the European Union, Financial Sector Assessment Program, Technical note, March.
Kashyap, AK and Stein, JC. 2000: What do a million observations on banks say about the transmission of monetary policy? American Economic Review 90(3): 407-428.
Laeven, L and Tressel, T. 2013: Financial Sector Assessment Program, European Union, Financial Integration and Fragmentation in the European Union. Technical Note, IMF Country Report, No 13/71.
Leeper, EM. 1991: Equilibria under 'active' and 'passive' monetary and fiscal policies. Journal of Monetary Economics 27(1): 129-147.
Lenza, M, Pill, H and Reichlin, L. 2010: Monetary policy in exceptional times. Economic Policy (62): 295-339.
Maddaloni, A and Peydro, JL. 2013: Monetary policy, macroprudential policy and banking stability: Evidence from the euro area. International Journal of Central Banking 9(1): 121-169.
Mersch, Y. 2013: Towards a European Banking Union, Speech at The Bridge Forum Dialogue, Luxembourg, 30 September, available at http://www.ecb.europa.eu/press/key/date/2013/html/ spl30930_1.en.html.
Mishkin, FS and White, EN. 2002: U.S. stock market crashes and their aftermath: Implications for Monetary Policy. NBER Working Paper, No 8992.
Mongelli, FP. 2013: The mutating euro area crisis - Is the balance between 'sceptics' and 'advocates' shifting? ECB Occasional Paper Series, No 144.
Praet, P. 2013: Forward guidance and the ECB, column in VoxEU.org published on 6 August, available at http://www.voxeu.org/article/forward-guidance-and-ecb.
Reinhart, C and Reinhart, V. 2010: After the fall. NBER Working Paper, No 16334.
Sander, H and Kleimeier, S. 2004: Convergence in euro-zone retail banking? What interest rate pass-through tells us about monetary policy transmission, competition and integration. Journal of International Money and Finance 23(3): 461-492.
Shambaugh, JC, Reis, R and Rey, H. 2012: The euro's three crises. Brookings Papers on Economic Activity 2 (Spring): 157-231.
Sinn, HW. 2012: Europe's Path to Disunity. Project Syndicate, available at http.y/www.projectsyndicate.org/commentary/why-a-united-states-of-europe-has-become-impossible-by-hans-werner-sinn.
Thornton, H. 1802: An Enquiry into the Nature and Effects of the Paper Credit of Great Britain, printed for J. Hatchard, Bookseller to the Queen, Piccadilly, and Messrs. F. and C. Rivington. St Paul's Church Yard: London.
van Leuvensteijn, M, Kok-Sorensen, C, Bikker, JA and van Rixtel, A. 2008: Impact of bank competition on the interest rate pass-through in the euro area. CPB Discussion Paper, No 103.
Weth, MA. 2002: The pass-through from market interest rates to bank lending rates in Germany. Deutsche Bundesbank Discussion Paper, No 11/02.
Zoli, E. 2013: Italian sovereign spreads: Their determinants and pass-through to bank funding costs and lending conditions. IMF Working Paper, WP/13/84.
ALAIN DURRE , ANGELA MADDALONI  & FRANCESCO PAOLO MONGELLI 
 European Central Bank and Institut d'Economie Scientifique et de Gestion (IESEG) School of Management.
 European Central Bank.
 European Central Bank and Goethe University Frankfurt.
(1) An explanation of the functioning of the euro area money market and its intermediation and dynamics is to be found in Durre and Nardelli (2008).
(2) According to McKinsey Global Institute, cross-border bank flows increased by about 10 times between 1990 and 2007, reaching a peak of around US$5 trillion in 2007. See also IMF (2013) and Laeven and Tressel (2013).
(3) In Figure 1, this greater dispersion is reflected by the different dotted lines for the three main crisis phases which move away from point 0 at the centre of the graph.
(4) This euro-specific crisis was ignited by the announcement in early November 2009 by the then newly elected Greek government of a huge revision of the public deficit left by the former coalition government. See Drudi et al. (2012).
(5) Note that these 'liquidity needs' refer to the liquidity demand necessary to cover liquidity shocks that banks face in the short run, mostly due to reserve requirements (ie, the minimum reserve (2% of deposits with a maturity of up to 2 years) to be held at the central bank), government deposits and demand for banknotes).
(6) This full accommodation of banks' liquidity demands refers to the implementation of the so-called 'fixed-rate full allotment (FRFA) procedure', which made, de facto, the supply of liquidity by the ECB endogenous to banks' requests.
(7) For further details about the malfunctioning of the unsecured term money market segments (ie, both the London interbank offered rate (LIBOR) and the EUR1BOR), please refer to the discussion in Brousseau et al. (2013).
(8) The downward pressures on government bond prices weakened banks' balance sheets and thus necessitated extra funding, which in turn created tensions on money market interest rates (Figure 3). See, for instance, the discussion in De Santis (2012) and Corsetti et al. (2011).
(9) See, in particular, the dotted lines for the period between July 2011 and October 2013.
(10) Since the seminal analyses of Thornton (1802) and Bagehot (1873), it has been widely recognised in the literature that a fractional banking system increases the sensitivity of market participants to liquidity shocks, which can eventually generate a panic. This in turn may imply a significant decrease in the money stock, possibly leading to a collapse in economic activity. Indeed, history also shows that financial instability as a result of liquidity gaps requires increased intermediation by central banks (Mishkin and White, 2002).
(11) As explained earlier in Footnote 6, this ECB tender procedure to inject liquidity into the money market was called the FRFA procedure and was formally started on 13 October 2008.
(12) This is the 'Outright Monetary Transactions' (OMT) programme. In contrast to the SMP, OMT offers the possibility to conduct asset purchases in unlimited amounts, though subject to conditionalities in the case of public debt instruments.
(13) For a description of the ECB measures, see Cassola et al. (2011), Drudi et al. (2012) and ECB (2013).
(14) Recent empirical results suggest that the ECB measures during the financial crisis helped avoid a credit crunch (with its possible
deflationary effects) through the anchoring of inflation expectations and support to economic activity. See, for example, Lenza et al. (2010), Fahr et al. (2011), Giannone et al. (2012) and Ciccarelli et al. (2011). For an analysis of ECB actions on the liquidity risk and credit risk components of spreads, please refer to Cassola et al. (2011).
(15) For empirical evidence on the euro area (as a whole or at a national level) in the pre-crisis period, see, among others, Cottarelli and Kourelis (1994), Weth (2002), Gambacorta (2004), Sander and Kleimeier (2004), de Bondt (2005) and van Leuvensteijn et al. (2008).
(16) The short-term interest rate is usually seen as a proxy of the monetary policy stance. Therefore, the EONIA could have been used instead. However, the 3-month EURIBOR interest rate is used here, in line with standard specifications applied in the literature. In general, the correlation between these two short-term interest rates remains high during the sample period, including the crisis sub-period.
(17) Given the illustration purpose of this exercise and for space reasons, it is assumed that the results for Germany are quite representative of those for the euro area countries facing less financial distress, whereas those for Spain represent more the results for the peripheral euro area countries under stress. See Zoli (2013] for empirical evidence for Italy.
(18) Equation 2 is estimated with autocorrelation and heteroskedasticity-consistent estimators to take account of the integration of variables in level and thus to obtain non-biased standard errors (Andrews, 1991).
(19) Therefore, one rank of cointegration between the variables in equation 1 still holds true during the crisis sample. Note that the results of the tests are available upon request.
(20) The BLS collects quarterly information on the lending standards that banks apply and on the loan demand that banks receive from firms and households (see ECB website). The information refers to the actual lending standards that banks apply to the whole pool of borrowers (not only to accepted loans) and is thus less subject to sample bias due to the restriction of borrowers that have effectively been granted loans.
(21) The EONIA is used as the monetary policy instrument, since it is a good proxy of the monetary policy stance--also for the crisis period, when a full allotment policy and credit enhancement actions were introduced (see ECB, 2009 and Lenza et at, 2010).
(22) This analysis also shows that a sudden stop of the lending channel would have significant economic effects. In this case, the positive median effect on GDP growth from a monetary policy shock would be reduced at the peak by about 35% for both GDP and inflation. See also Allen et al. (2004).
(23) See, for example, Lenza et al. (2010), Fahr et al. (2011) and Giannone et al. (2012).
Table 1: Interest rate pass-through in retail banking activity in Germany (DE) and Spain (ES) Retail interest rate on long-term credit to NFCs (above 1-year maturity) Short-term elasticity (a) Pre-crisis Euribor 10-year ([[delta].sub.0]) ([[phi].sub.0]) DE -0.010 0.312 ** ES 0.385 * 0.155 Retail interest rate on long-term credit to NFCs (above 1-year maturity) Short-term elasticity (a) Crisis Euribor 10-year ([[delta].sub.0]) ([[phi].sub.0]) DE 0.204 ** 0.216 ** ES 0.623 ** 0.155 Retail interest rate on long-term credit to NFCs (above 1-year maturity) Long-term elasticity (b) Pre-crisis Euribor 10-year ([[alpha].sub.1]) ([[beta].sub.0]) DE 0.304 ** 0.526 ** ES 0.910 ** 0.038 Retail interest rate on long-term credit to NFCs (above 1-year maturity) Long-term elasticity (b) Crisis Euribor 10-year ([[alpha].sub.1]) ([[beta].sub.0]) DE 0.317 ** 0.261** ES 0.594 ** 0.625** Retail interest rate on long-term credit to NFCs (above 1-year maturity) ECM (c) Pre-crisis Crisis DE -0.663 ** -0.235 ** ES -0.834 ** -0.436 ** Retail interest rate on long-term mortgage credit Short-term elasticity (a) Pre-crisis Euribor 10-year ([[alpha].sub.1]) ([[beta].sub.0]) DE 0.071 0.149 ** ES 0.586 ** -0.076 Retail interest rate on long-term mortgage credit Short-term elasticity (a) Crisis Euribor 10-year ([[alpha].sub.1]) ([[beta].sub.0]) DE 0.053 ** 0.117** ES 0.247 ** -0.049 Retail interest rate on long-term mortgage credit Long-term elasticity (b) Pre-crisis Euribor 10-year ([[alpha].sub.1]) ([[beta].sub.0]) DE 0.093 ** 0.650 ** ES 0.860 ** 0.154 ** Retail interest rate on long-term mortgage credit Long-term elasticity (b) Crisis Euribor 10-year ([[alpha].sub.1]) ([[beta].sub.0]) DE 0.194 ** 0.481 ** ES 0.723 ** 0.331 ** Retail interest rate on long-term mortgage credit ECM (c) Pre-crisis Crisis DE -0.430 ** -0.404 ** ES -0.348 ** -0.289 ** Note: NFCs stands for non-financial corporations. For each equation, a one-rank cointegration relationship exists. The whole sample period ranges from January 2003 to December 2012 on a monthly basis, that is, there are 120 observations in total. It is divided into a 'pre-crisis' sub-period (from January 2003 to July 2007) and a 'Crisis' sub-period (from August 2007 to December 2012). The symbols * and ** denote the 5% and 1% significance levels, respectively. (a) Short-term elasticity refers to parameters [[delta].sub.0] and [[phi].sub.0] of equation 1. (b) Long-term elasticity refers to parameters [[alpha].sub.1] and [[beta].sub.1] of equation 2. (c) ECM denotes the error correction mechanism and corresponds to the parameter [rho] in equation 1.
|Printer friendly Cite/link Email Feedback|
|Title Annotation:||Symposium Article; European Central Bank|
|Author:||Durre, Alain; Maddaloni, Angela; Mongelli, Francesco Paolo|
|Publication:||Comparative Economic Studies|
|Date:||Sep 1, 2014|
|Previous Article:||Integration of Central and Eastern European and the euro-area financial markets: repercussions from the global financial crisis.|
|Next Article:||Conflict and social and political preferences: evidence from World War II and civil conflict in 35 European countries.|