I would like to thank the Bank of Albania for the kind invitation. Costs and benefits of ECB negative deposit policy have been extensively discussed in the euro area, and I understand that it also raises technical challenges for you in the Western Balkans.
1. Motivation and Overview of Accommodating ECB Monetary Policies
First, let me recall the rationale of the low interest rate policy and unconventional monetary policies implemented by the ECB, as well as by other central banks in developed financial markets.
Central bankers use their policy rate to pursue their inflation objective. Central banks consider that in principle, there is at any moment in time, a level of short-term risk-free interest rates, that is in line with reaching in the medium-term the central bank’s inflation target. As explained in Figure 1, the theoretical starting point is that the short-term interest rate needs to be equal to the expected return on capital (which, in the steady state, should correspond to expected GDP growth) and expected inflation. If the central bank wishes to exert a disinflationary impulse, it will increase short-term interest rates above the neutral level, that is, the sum of the expected return on capital and expected inflation, and vice versa. Measuring expected inflation is relatively straightforward, while measuring expected return on capital is more difficult. It could be derived from an assessment of GDP growth potential, but various other methods have been proposed. Lately, the neutral interest rate level in the euro area has been very low due to the subdued inflationary expectations and low GDP growth expectations.
So, in principle, the central bank can provide inflationary impulses to address deflationary pressures as it controls the interest rate. However, this assumption is confronted with three problems:
a) The first issue is how a central bank’s interest rate decisions are transmitted to the effective cost of funding the economy. Factors, that are not under the direct control of the central bank, such as credit and risk premia, influence the cost funding of the economy. Particularly high levels of intermediation spreads would imply an even lower neutral central bank interest rate, everything else equal.
b) The second issue is that the interest rate transmission channel may change once the interest rate reaches a low level, because of other effects of low interest rates that could interact negatively with the inflationary impulse of lower interest rates.
c) Finally, the central bank’s choice of the short-term risk-free interest rate is constrained by the so-called zero-lower bound problem. If interest rates fall below a certain negative level, banknotes demand would likely increase very significantly, destabilize the financial system, and undermine low interest rate policies.
To address the challenges described above, after reducing its policy rates to low levels, the ECB has launched policies to compress spreads: term spreads, liquidity spreads, and credit spreads to reduce actual funding cost. Also, the deposit facility rate was lowered somewhat below zero. This is illustrated by Figure 2.
Figure 2 represents the effective cost of funding for households, 10-year maturity, mainly mortgages, and non-financial corporates (NFCs), six-year maturity. In addition, the slide presents the interbank overnight unsecured borrowing cost in the euro area (EONIA), over which the ECB has better control. Finally, the Figure shows 6- and 10-year EONIA swaps, which indicate the risk-free rate for maturities equal to the household interest rate and the NFC interest rate.
In 2006, the spread between the risk-free rate and the actual customer rate was about 50 basis points, with a tendency to increase during the run up toward the first financial crisis. It increased with the first financial crisis (2008) and, even more, during the sovereign debt crisis in the euro area (2011–2012). Since 2012, it has remained in a range between 200 and 250 basis points, despite the unconventional measures taken to compress this spread.
Inflation forward rates (in particular the five to 10 years forward rate) dropped in 2014 below 2 percent. This was a matter of concern since, as suggested by the equation in Figure 1, if expectations on inflation drop, the implied ex ante real interest rate rises, thereby leading to a tightening of monetary policy when the policy rate is not lowered proportionally. However, the lower bound issue could compound the credibility and monetary policy effectiveness issues, as the central bank has limited room to counteract the lower inflation expectations.
Therefore, central banks should act quickly and forcefully in such a situation. That is what the ECB did.
Figure 3 shows the ECB policy rates and EONIA since 2009. One could notice that the deposit facility rate and EONIA moved in negative interest rate territory. Is it a conventional or unconventional monetary policy action? On one hand, it is conventional, as it follows standard monetary policy decision and implementation logic, based on the policy rate. On the other hand, it is unconventional, as negative interest rates had never been applied before in the euro area.
The other side of unconventional monetary policy is what you can see in the balance sheet of the euro system. Figure 4 shows key measures, announcements, and programs. The first dot in the chart is the fine-tuning operation of August 9, 2007, a long time ago. Then follow different announcements, like the ECB President’s speech in London in 2012, the OMT announcement, and the PSPP announcement. The programs and their duration are indicated at the bottom. The top part of the chart represents Eurosystem balance sheet items: on the up-side are the asset items and on the lower side are the liability items, including the so-called autonomous factors.
Let me go very quickly through these non-conventional measures. One is our TLTRO-target longer-term refinancing operations. The latest one was launched in June 2016. It is a four-year operation, which is very attractive for the banks. The purpose is to give an extra incentive to expand lending and also to compete via loan rates, to support the transmission from our low rates to the actual funding cost of NFCs. In order to benefit from lower interest rates in the TLTRO, banks should expand lending to corporates, but the loan increase ratio has been set at a relatively moderate level. If banks expand their loans to non-financial corporate by 2 percent over a two-year period, they would pay the deposit facility rate on the Eurosystem refinancing, so minus 40 basis points, as a fixed rate over four years. That explains why the demand has been relatively large. In total now, the Eurosystem provided 761 billion of this cheap credit. While the Eurosystem was not the first central bank to introduce negative rates on its liabilities, it was the first central bank to provide banks’ funding at negative rates.
Then comes the asset purchase program of the Eurosystem. Its main purpose is to compress term spreads. Indeed, by buying long-term assets, the central bank compresses term spreads, which should affect funding costs of the economy as the term spread element is a component of the funding cost. The Eurosystem is currently buying government bonds, corporate bonds, covered bonds, and ABS.
One could ask why does the Eurosystem buy all those different assets. The starting point was the universe of debt instruments issued in the euro area, applying certain minimum credit quality criteria. The Eurosystem does not want to favor governments as issuers, and, therefore, also included various private sector bonds in its programs, even if purchasing these often less liquid bonds is technically less convenient.
One can note in particular that the ABSPP in Figure 6 is not very important from a mere quantitative perspective. Still, the ECB considered ABS an important asset class that should be included in the Asset Purchase Programme (APP) in view of its potentially important role in financing the euro area real economy. Figure 7 shows a 10-year German bund yield and the spreads of other countries toward the German 10-year bund.
German bund 10-year yields and spread of other 10-year government bond yields vis-à-vis German bund
It is difficult to assess the impact of the different purchase programs. There have been a lot of discussions on the need of such programs before they were announced. Therefore, the announcement and implementation only partially came as a surprise and their impact was already partially reflected in prices well before the program announcement. The start of the actual daily purchase flow did not seem to have had an impact on yields.
As the Eurosystem initially decided not to buy securities below the deposit facility rate, the duration of the purchases went higher. The more the yields went down, the higher rose the duration of term purchases, which meant even lower yields. The Eurosystem was subsequently authorized to purchase below the deposit facility rate.
When analyzing the effects on the spreads between the various euro area government bonds, it is important to note that higher rated countries tend to be over-weighted in the purchases relative to market capitalization since the Eurosystem buys according to capital key and not according to the size of the market of the securities to be purchased. Therefore, for example, Italy is underweighted and Germany overweighted in the purchases from a securities market perspective. This could have led to a spread widening. However, there may have been some opposite effects since the perceived overall positive effects on financial stability and economic developments may have benefitted the somewhat lower rated countries more than the highest rated countries. The overall net effect of the APP on euro area spreads is therefore difficult to assess.
CSPP, the corporate sector purchase program, shows a nice announcement effect, because it was not discussed in the public beforehand. The announcement was relevant. The right-hand chart of Figure 8 shows the iBoxx insurance sector index. Here, the modalities of the CSPP, when they were released in April 2016, had an effect, because it was not clear whether or not the insurance sector would be part of it.
2. Experience with Negative Interest Rates
The Eurosystem did not invent negative policy rates. However, it was the first one to introduce the negative rates in a large monetary area, which did not need it in terms of preventing exchange rate appreciation pressures, but needed it from a domestic monetary policy perspective.
Overview of negative interest rate policies
Overview of negative interest rate policies
Overview of negative interest rate policies
What are the benefits of negative rates in short? Those are conventional benefits of lowering interest rates in the Wicksellian logic. If you need an inflationary impulse and you are close to zero, then you may want to use the possible remaining leeway going negative.
This is the standard logic of conventional monetary policy. In fact, the policy rate cut below zero not only lowered the short-term interest rates, but the entire yield curve went down. Markets beforehand believed that interest rates could not go below zero. Once they actually moved below zero, the whole expectations on future interest rates also changed. The probability distribution of future interest rates now had a tail in negative territory. As a result, the long-end of the curve moved down very effectively with the policy rate cut below zero. This is also what the Bank of Japan experienced when introducing negative rates.
As the benefits of negative rates are standard, much of the literature focuses on the possible negative side effects. Did negative interest rates have side effects that undermined the obvious standard positive effects? The following list of questions comes to mind:
Would negative interest rates lead to lower money market activity?
Would negative interest rates lead to a surge in banknotes or in the demand for alternative secure deposits? Does that create an effective lower bound (lower bound type 1)?
Would the negative interest rate reduce banks’ profit? And, then, would banks try to recoup the loss via higher lending rates, which is the opposite effect that the one sought? This is a different kind of effective lower bound (effective lower bound type 2).
Would negative interest rates encourage investors to take excessive risks, which could endanger financial stability?
Would negative interest rates discourage structural reforms by easing too much government funding conditions?
The pass-through to money markets was smooth and identical to the one in positive interest rate territory. The Eurosystem announced that it would introduce negative rates a few months in advance so that technical (IT) issues could be solved. And, they were solved.
Pass-through of negative policy rates to money market rates
Pass-through of negative policy rates to money market rates
Pass-through of negative policy rates to money market rates
In term of money market volumes, did negative rates affect the turnover in the money markets? More than by negative rates, money market turnover was reduced by large excess liquidity. Every bank has excess reserves, so there is less opportunity for trading liquidity. That is what Figure 11 is showing. The yellow line indicates when the Eurosystem introduced the negative deposit facility rate. On volumes, nothing at all happened. The other line, the green line, shows the excess liquidity. As excess reserves increased, EONIA volumes went down. But, the decline in volumes is not due to negative rates.
Negative interest rates, excess liquidity, and money market activity
Negative interest rates, excess liquidity, and money market activity
Negative interest rates, excess liquidity, and money market activity
Figure 12 shows the interest rate pass-through to household and corporate deposits up to one-year maturity in the euro area. It happens that negative interest rates were not passed on to customer or corporate deposits. While remaining positive, corporate deposit rates declined to lower levels than household deposit rates because some corporate deposits were charged negative rates. The zero lower bound seems to have been almost universally applied only on retail deposits.
Negative interest rates, banks’ profitability, and pass-through to customers’ deposits
Negative interest rates, banks’ profitability, and pass-through to customers’ deposits
Negative interest rates, banks’ profitability, and pass-through to customers’ deposits
Banknotes’ demand might have increased slightly above trend, as shown in Figure 13. But one cannot say that negative rates have made a fundamental difference.
Negative interest rates and banknotes’ demand
Negative interest rates and banknotes’ demand
Negative interest rates and banknotes’ demand
In a certain way, negative rates are merely a continuation of very low rates already characterizing the euro area since the financial crisis, and low rates should already have an upward impact on banknotes demand. The Eurosystem does seem to have reached the effective lower bound type 1. A general view seems to have emerged, however, that the lower bound of type 1 is only reached when negative rates are decreased below the minus 1 percent threshold.
Reserve management accounts, which include accounts of non-euro area public sector investors, went up slightly. However, it is a deliberate policy of the Eurosystem to subject these accounts, as well as government accounts, to −40 basis points, at least to avoid encouraging an increase in the amount deposited on these accounts, while banks are charged 40 basis points on their accounts in the Eurosystem.
With regard to the bank profitability issue, it is important to distinguish the general issue of lower net interest rate margins from the specific one that arises when banks have to hold excess liquidity with the central bank at a negative remuneration rate.
Negative interest rates and other central bank liabilities towards non-banks
Negative interest rates and other central bank liabilities towards non-banks
Negative interest rates and other central bank liabilities towards non-banks
Figure 15 shows the results of the analysis conducted by the Eurosystem on the effects on profitability of the different policy instruments deployed. The analysis indicates that the net effect on bank profitability of monetary policy of the last three years is positive.
Monetary policy and effects on banks’ profitability
Monetary policy and effects on banks’ profitability
Monetary policy and effects on banks’ profitability
The black dot is a net effect. The effects captured include the positive impact on credit quality and NPLs of accommodative monetary policy. The net interest income impact is indeed negative. This is shown by the blue bar at the bottom.
Cost of external financing of NFCs and household and loan growth
Cost of external financing of NFCs and household and loan growth
Cost of external financing of NFCs and household and loan growth
Finally, critics of non-standard monetary policy also often give prominence to the associated financial stability risks. Indeed, some investors may be more exposed than others to low interest rates, and might end up taking excessive risks. For example, parts of the insurance industry had been running a maturity mismatch (with a longer duration of liabilities compared to assets), not having considered a scenario of risk-free interest rates becoming negative. However, it can also be argued that low for long and stable interest rates may also support a more precise and more prudent pricing of assets than in a period of high interest rate volatility. So far, there is no convincing evidence that low rates may have encouraged excessive risk-taking and asset misallocation.
