Asset purchases as an instrument of monetary policy

Thank you very much for inviting me to open this session on whether central banks should continue to target longer-term bond yields in normal times. My first instinct when presented with this question was to provide a rhetorical answer and say that to continue targeting long-term bond yields one has first to have started doing it. And indeed, the ECB has so far only intervened in long-term bond markets, such as the covered bond market, to restore monetary policy transmission in malfunctioning market segments and help enforcing a given monetary policy stance. We have not so far resorted to a policy of targeted asset purchases that aim to alter the monetary policy stance – what is universally, if in my view inaccurately, referred to as quantitative easing.

On second thoughts, however, this would be the wrong way to reply. It would not only be wrong because we have recently made clear that asset purchases are an instrument that we are ready to use if we deem necessary. It would also be the wrong answer because the question of whether central banks should target longer-term bond yields provides a good focal point to reflect on what monetary policy is trying to achieve in the first place. In other words, instruments are just the tail that wags the dog.

Let me try and explain what I mean as concisely as possible.

Focusing specifically – and at the risk of over-simplifying the issue – on the interest rate channel of monetary transmission, monetary policy operates by raising or lowering the interest rate in the economy. A lower (real) interest rate lowers the cost of capital for firms, encourages investment spending and stimulates consumption. A higher real interest rate has the opposite effect.

But the point of course is that there is no such thing as one interest rate to which all economic agents respond. There are at least three ways in which interest rates are differentiated in the euro area. There is vertical differentiation – different economic agents are sensitive to interest rates with different maturities. There is spatial differentiation – different interest rate curves provide the reference rates in different jurisdictions. And there is horizontal differentiation – within jurisdictions, different markets determine firms’ and households’ cost of borrowing.

What this implies is that the levels of medium- and long-term real interest rates across jurisdictions and markets will always be relevant to the formulation of monetary policy. The difference between normal and abnormal times is therefore not what we are trying to achieve – it is how we strive to achieve it.

In normal times, central banks affect this array of interest rates by steering the overnight interest rate and influencing expectations. We do the former through setting policy rates and the latter by communicating transparently our reaction function. This allows market participants to form reasonable assumptions about future short-term rates, which in turn allows our monetary impulses to be transmitted along the yield curve to longer-term interest rates.

Monetary policy-making becomes more complicated, however, when short-term nominal rates reach the effective lower bound. Here, our ability to alter the monetary policy stance relies comparatively less on the level of the overnight rate, and comparatively more on influencing expectations, i.e. on the shape of the yield curve. This is why central banks have introduced forward guidance, which is enhanced communication on the expectations and reaction function of the central bank. It is essentially a natural extension of ‘normal’ monetary policy.

Forward guidance influences the long-term interest rate by acting on both its determinants. It helps steer expectations about the level of future rates. And it lowers uncertainty about this level – that is, reduces the time-value of money or the term premium. Yet, as an instrument it also has natural limitations. A monetary policy committee cannot credibly commit to keep interest rates low beyond its horizon of visibility. This is reflected in the specific wording of our forward guidance, in particular the fact that it is outcome-contingent.

Hence, should further monetary accommodation be needed, it is reasonable to consider other operations aimed at lowering the term premium. This is where targeted asset purchases enter the toolset of monetary policy. Targeted asset purchases mean operations aimed primarily at impacting the level of the term premium across maturities and market segments.

Consistent with public statements made by members of the ECB Governing Council recently – in particular our President, Mario Draghi, and my colleague on the board of the ECB, Peter Praet – I would like to make a few qualifying remarks about the possible use of such purchases by the ECB.

First, at its meeting last week the Governing Council confirmed our baseline scenario – that is, a prolonged period of low inflation followed by a gradual rise in inflation rates towards 2%. Further monetary easing is therefore not excluded, but remains contingent on outcomes.

Second, if such easing is called for, the Governing Council is unanimous in its commitment to use also unconventional instruments within its mandate. The question we face is whether asset purchases would be the appropriate unconventional measure . And if we deem that asset purchases would be appropriate, the question then becomes how we could implement such a policy in way that is useful and that complies with our mandate.

This means, among other things, ensuring that any such purchases comply with the Treaty, keeping in mind that Article 18 of the ESCB Statute allows the ECB and national central banks to operate in the financial markets by buying and selling outright claims and marketable instruments. It also means guarding against operations that unduly distort market allocations or worse, that would have intended distributive effects. And it means designing a programme that is effective given the financial structure of the euro area – where, for instance, bank-based intermediation is pre-dominant and wealth effects through equity and real estate prices are less important transmission channels than, say, in the US.

These considerations in turn imply thinking in terms of the three forms of interest rate differentiation that I described above.

First, vertical differentiation – the relevant maturities at which asset purchases should take place. In practice, purchases would naturally be linked to the interest rate maturities that are most important for firms’ and households’ investment and consumption decisions. In the euro area, this tends to be the intermediate to longer part of the yield curve.

Second, spatial differentiation – the jurisdictions across which asset purchases should be spread. Here we would have to take into account the interest rates in different jurisdictions that provide the benchmarks for loan pricing. In the euro area, remember, there is no single yield curve that refers to a “commoditised” reference asset and that is equally relevant for loans to firms and households. Creating such an asset would ease the implementation of our monetary policy, but this cannot be a short term project.

Third, horizontal differentiation – the markets within jurisdictions that asset purchases should target. When financial markets are highly integrated with a high degree of substitutability between assets, purchases in one asset class, such as government bonds, are more likely to affect term premia across all asset classes. This is because the process of portfolio reallocation facilitates a relatively homogenous transmission. But given the segmentation of euro area financial markets, this effect cannot just be assumed. To achieve a homogenous reduction of term premia across relevant interest rates, segmentation would have to be taken into consideration in our strategy.

Overall, the yardstick for the success of any targeted asset purchases would not be the size of our balance sheet, but the observable effect of our operations on term premia across markets and jurisdictions. Or put differently, asset purchases in the euro area would not be about quantity, but about price. This is why I said that quantitative easing is a misnomer.

However, whether reducing term premia would in turn be effective in reducing real interest rates for firms and households depends on complementary policies, in particular the continued repair of the banking sector. For example, the sizeable reduction of spreads we have witnessed since July 2012 has not in itself translated into higher credit volumes and lower bank lending rates. This is largely because the restructuring of the euro area banking sector remains ongoing. Thus, an essential complement to any monetary policy strategy is a strict implementation of the ongoing Comprehensive Assessment of euro area banks and swift corrective action to bridge identified capital shortfalls.

Let me conclude.

Unconventional monetary policy tools are less unconventional than the word implies. They are unusual, because they respond to highly unusual circumstances. They imply risks that have to be carefully weighed and mitigated. But fundamentally, unconventional tools are only a means for central banks to continue doing what they have always done: managing aggregate demand, by influencing the level of real interest rates and other monetary transmission channels, to maintain price stability. To borrow from Giuseppe Tomasi di Lampedusa, in these unusual times “everything must change, so that everything stays the same”. It is this that will determine both the appropriateness of using targeted asset purchases in our monetary policy operations, and the design of any such purchases.