Significant increases in energy and food prices are likely to keep inflation above the ECB's target of 2% over the next two years. Eurozone inflation expectations over the next five years are also above 2%. With high inflation starting to dampen household consumption and growth, the ECB has clearly indicated that the time has come to normalize monetary policy. As the ECB's president, Christine Lagarde, stated in a recent blog post, not adjusting policy rates in this environment would amount to unnecessary policy easing. GDP growth in the eurozone of 0.6% in the first quarter of 2022 versus the prior quarter, a strong upward revision of the original estimate, supports the ECB's view.

The ECB reiterated its intention to normalize policy in today's meeting, announcing the end of both preferential rates on TLTROs and net asset purchases, alongside the hikes in all three policy rates in July and September 2022. The ECB has added several tools to its arsenal since the last normalization cycle prior to the financial crisis of 2007-2008. Normalization therefore not only means increasing interest rates, but also phasing out unconventional tools such as QE and TLTROs.

The Effective Policy Rate Will Reach 1.5% In Late 2023

The ECB intends to raise the three policy rates by 25 bps in July, and again in September, possibly at a larger increment depending on the new inflation forecasts at the time. Beyond September, we expect the ECB to continue raising rates in 25 bp increments on a quarterly basis until its effective policy rate reaches 1.5%, which is what we understand the ECB's current estimate of the neutral rate to be. In a speech at the beginning of May 2022, the governor of the French central bank, Villeroy de Galhau, put real neutral rates between -1% and 0% in the euro area, echoing previous estimates by ECB board member Isabel Schnabel at the end of 2021 (see "Related Research").

We expect the neutral rate to be the terminal rate for this cycle, since the output gap is unlikely to become strongly positive in the coming two years and inflation should recede from its current level. Under our baseline assumptions, we think that the rate-hiking cycle could end as soon as the final quarter of 2023. That said, the actual pace at which the policy rate rises to neutral will ultimately depend on the data. What's more, although the neutral rate is subject to slow-moving dynamics--that is, demographics, productivity cycles, the return on capital, and the demand for safe assets--the ECB may change its estimates of the neutral rate as its tightening cycle progresses if it sees any of these factors changing in the post-pandemic world.

For now, the ECB's de facto policy rate will continue to be the deposit rate as long as excess liquidity remains in the system (see chart 1). Banks have accumulated abundant liquidity over the past decade because of the ECB's credit easing (TLTROs) and asset purchases. This, and the full allotment procedure--banks' ability to satisfy all their bids for liquidity at the central banks--mean that they have made little use of the interbank market in past years. Furthermore, as long as banks keep large amounts of excess reserves, the euro short-term market rate is likely to continue tracking the deposit rate.

It is unclear what level of excess liquidity would bring the interbank market, and thus the refinancing rate, back into play. According to its Financial Stability Review from May 2022, the ECB expects that banks will repay €700 billion of TLTROs by year-end 2022. However, a quick back-of-the-envelope calculation suggests that it would take excess liquidity dropping at least below €1 trillion for the euro short-term market rate to move closer to the refinancing rate (see chart 2). Whether or not excess liquidity will fall as low as this threshold will also depend on market and bank lending conditions in an environment of rising short-term interest rates. The ECB might be tempted to inject long-term liquidity into the system in the event of an unwarranted tightening of financing conditions. Therefore, it could take time for the rate instruments to normalize fully, and the ECB could maintain its floor policy rate framework throughout this rate-tightening cycle.

Chart 1

Chart 2

The ECB's Balance Sheet Is Likely To Remain Large, Even As Interest Rates Rise

The ECB's balance sheet has expanded by up by €6.6 trillion since 2015. One-quarter of this, €1.7 trillion, came from the ECB increasing the TLTROs it granted to banks, and three-quarters, €4.7 trillion, came from outright purchases of securities (see chart 3). The ECB could normalize the size of its balance sheet with the same two instruments, but the pace of the endeavor, and the extent to which it would reduce each instrument, remain unclear.

Chart 3

The end of preferential rates on TLTROs

The ECB's announcement of the end of the preferential interest rate for TLTROs is coherent with the start of the interest rate normalization cycle. In a higher rate environment, it wouldn't make sense for the ECB to further incentivize lending at a lower interest rate than its policy rate as it would cause the economy to overheat, something the ECB seeks to prevent. The ECB expects banks to repay some €700 billion until the end of 2022. Nevertheless, the ECB could choose to engage in another round of TLTROs in the event of an unwarranted tightening of financing conditions that risks upsetting the smooth transmission of its interest rate policy across the euro area. Flexibility, optionality, and gradualism are the ECB's stated guiding principles of its monetary policy normalization efforts.

The end of net asset purchases

While the ECB has confirmed an end to its net asset purchases as of July 1, 2022, there is still little clarity on the start and extent of the downsizing yet. So far, the ECB has committed to reinvesting maturing securities under the pandemic emergency purchase program until the end of 2024. Unlike the U.S. Federal Reserve and the Bank of England, the ECB has not yet communicated a willingness to reduce the size of its balance sheet through outright sales of bonds held under QE. Since the average maturity of the securities on its balance sheet is around seven years, it would take the ECB until at least 2030 to unwind its asset purchases by not reinvesting.

That said, the ECB could choose to reinvest maturing securities in bonds with shorter or longer maturities if it seeks more tightening or loosening in the future. Moreover, should financing conditions remain favorable, the ECB could also decide to actively sell bond purchases to unwind its balance sheet. On the other hand, should financing conditions worsen to the point of excessive fragmentation among members states, the ECB could be tempted to launch a new asset purchase facility, as several ECB members have suggested (see Christine Lagarde's blogpost "Monetary policy normalisation in the euro area"). Such options could make the normalization of the ECB's balance sheet take even longer.

Long-Term Yields Should Not Rise Much Further As The ECB Starts To Normalize Policy

Bond market conditions have become less favorable since the ECB started signaling monetary policy tightening in March 2022. For example, German ten-year yields have risen by around 140 bps since December 2021. While there is a relatively good understanding of how increases in the main policy rate affect financing, there is still little clarity on what will happen to yields when the ECB ends QE and starts to normalize its balance sheet.

Our modelling results suggest that what matters most in the depression of long-term yields is the size of the ECB's balance sheet (the stock effect), while the monthly increase in the ECB's balance sheet (the flow effect) has an insignificant impact on rates. The ECB's balance sheet expansion since 2015 has meant that the stock effect had a growing impact on yields over time, depressing 10-year bund yields by around 110 bps after one year and as much as 200 bps after four years (see chart 4). By contrast, the flow effect has had a small to insignificant effect on yields (see chart 5), suggesting that the end of net asset purchases shouldn't trigger a rise in yields.

In other words, as long as the ECB holds a large proportion of sovereign bonds on its balance sheet (currently around 40% of outstanding eurozone government bonds) and banks hold excess reserves at the ECB, this will continue to depress yields by constraining the supply available to price-sensitive investors.

Chart 4

Chart 5

The ECB's balance sheet expansion has affected both the risk-free rate and the term premium

The stock effect arises partly from investors' expectations, and not just the ECB's balance sheet holdings. As such, a large part of the effects on yields is priced in ahead or around the timing of the QE announcement. This is because the QE effect has two channels--a signaling channel and a supply scarcity channel. The signaling channel--which is largely tied to the ECB's forward guidance on interest rates--is likely priced in immediately, whereas the effects on supply (the scarcity channel) may take more time to unfold, explaining why the stock effect grows stronger over time. In the context of balance sheet unwinding, it implies that the signaling effect dissipates when rate tightening is on the cards, while the scarcity effect is likely to last longer.

To illustrate these dynamics, we disentangled the two main channels by splitting the German ten-year yield curve into a term premium--the risk premium investors pay for holding a bond with a longer maturity date, and a risk-free rate component--investors' policy rate expectations across the yield curve. The term premium shows the size of the scarcity effect, while the risk-free rate shows the effect on the signaling channel.

The results confirm that the ECB's balance sheet expansion has affected both the risk-free rate and the term premium. Specifically, it has lowered the term premium by around 90 basis points (see table 1), and has had a significant 120 bp impact on risk-free-rate expectations (see table 2). The recent rise in the risk-free rate confirms that the signaling effect has already dissipated (see chart 6), since the ECB has already communicated the end of its purchases. At the same time, the scarcity channel is still at work, with our term premium estimates still close to zero (the recent rise is mostly linked to higher inflation). Our estimates hold firm using different specifications, with three different models showing similar results. The total impact is similar to that under the SVAR model we used in charts 4 and 5. Our results are also very close to the estimates of ECB staff.

Chart 6

Chart 7

Chart 8

Table 1

Table 2

Yields Should Remain Depressed Until The ECB Starts To Reduce Its Balance Sheet

The results of our models suggest that yields are likely to stay depressed for some time, even as the ECB normalizes monetary policy. As risk-free-rate expectations are already close to the neutral rate, lower yields will mostly materialize through a lower term premium of around 90 bps lower until the ECB starts to reduce the size of its balance sheet. Meanwhile, although the ECB may not be buying additional assets, eurozone governments are starting to reduce their net issuances, which means that safe-asset supply constraints are likely to remain broadly the same for investors. This suggests that the ECB's actions will keep a lid on the tightening of long-term financing conditions in the eurozone, at least until it starts to unwind its balance sheet.

Related External Research

• Financial Stability Review, ECB, May 2022
• Monetary policy normalisation in the euro area, blog post by Christine Lagarde, May 23, 2022
• The Eurosystem and its monetary policy - from an "impossible dilemma" to a possible roadmap for normalization, speech by Villeroy de Galhau, Paris, May 6, 2022
• A new strategy for a changing world, speech by Isabel Schnabel, July 14, 2021
• Eser F., Lemke W., Nyholm K., Radde S., Vladu A. (2019), Tracing the impact of ECB's asset purchase programme on the yield curve, ECB Working Paper
• Adrian, Crump and Moench (2008), Pricing the Term Structure with Linear Regressions, Federal Reserve Bank of New York Staff Reports, No. 340