Estimating the time-varying reserve elasticity of money market rates in the euro area

Prepared by Flavia Ungarelli and Thomas Kostka

Published as part of the ECB Economic Bulletin, Issue 1/2026.

Understanding the demand for central bank reserves by commercial banks is important for the implementation of monetary policy. If the supply of central bank reserves increases relative to demand, money market rates will decline, up to the point where the most attractive option for banks is to deposit reserves with the central bank. If the demand for central bank reserves increases relative to supply, money market rates will rise, up to the point where the most attractive option for banks to satisfy their demand is to borrow from the central bank. Hence, the relation between central bank reserves and money market rates is non-linear: it is broadly flat around the respective central bank lending and deposit rates when reserves in the system are low or high and it slopes downwards when there are intermediate quantities of reserves. A stylised representation of this pattern is shown in Chart A. For central banks to gauge the elasticity of money market rates to changing central bank liquidity conditions, it is therefore relevant to have reliable estimates of the slope of this relation.

Chart A

Illustrative representation of the demand curve for central bank reserves

(percentages)

This box presents a new method for estimating in real time the time-varying elasticity of euro money market rates to excess liquidity. The method was originally developed by staff at the Federal Reserve Bank of New York (FRBNY) for the unsecured federal funds market.^[1] Given the potential divergence between secured and unsecured segments of the money market, we apply this approach to both the euro short-term rate (€STR) and euro general collateral repo rates. This dual application aims to provide a more comprehensive view of liquidity dynamics and their sensitivity across different market segments, which may respond differently to shifts in central bank liquidity conditions.

Estimating the reserve demand elasticity comes with methodological challenges owing to the endogeneity between the price and volume of liquidity and also to shifts in the demand curve over time. First, the relation between money market rates and liquidity volumes is endogenous because money market rates not only reflect exogenous shifts in liquidity demand and supply but these also influence the liquidity uptake by banks in their own right. Moreover, confounding factors – such as changes in government deposits – can simultaneously affect liquidity supply and broader money market conditions. Second, the demand curve itself is subject to both horizontal and vertical shifts over time, which may occur, for instance, on account of changes in banks’ structural liquidity needs or evolving market structures. For example, during the euro area sovereign debt crisis, the demand curve appears to have shifted horizontally (to the right) (Chart B, panel a) as banks increased their precautionary reserve demand amid heightened uncertainty and regulatory changes following the global financial crisis. More recently, a vertical (downward) shift in the demand curve seems to have taken place (Chart B, panel b). The shift likely reflects that the reduction in excess liquidity in 2022 owed largely to the repayment of targeted longer-term refinancing operations and had no bearing on the volume of non-bank deposits placed with banks, while the latter has been the main driver of recent movements in the spread between the €STR and the deposit facility rate, according to ECB staff analysis.

Chart B

Shifts in the reserve demand curve

a) Horizontal shift around August 2011

(percentages)

b) Vertical shift around August 2022

(percentages)

The methodology developed by Afonso et al. (2025) offers a robust solution to these challenges and provides a reliable tool for tracking the reserve demand elasticity in real time. Rather than seeking to estimate the full shape of the historical demand curve, the approach simply estimates its local slope on any given day. This makes the estimation invariant to whether changes in the elasticity arise from movements along the curve or from (horizontal or vertical) shifts in its position. Additionally, the approach addresses the endogeneity challenges by employing lagged forecast errors as instruments for excess liquidity.

Currently, there is no statistically significant evidence of heightened rate sensitivity to liquidity conditions in the euro area. Chart C, panel a shows the estimated reserve demand elasticity of the spread between the €STR and the deposit facility rate. The chart tracks the basis-point impact of a one percentage point exogenous change in reserves on money market rates. Three observations stand out. First, after liquidity operations were first conducted on a fixed rate full allotment basis in autumn 2008, within approximately one year unsecured market rates had stabilised and stopped reacting strongly to fluctuations in liquidity supply. Second, the elasticity temporarily became significantly negative during two distinct episodes: in 2013/14, when longer-term refinancing operations matured and liquidity levels declined sharply, and, to a lesser extent, at the onset of the COVID-19 pandemic, when heightened risk aversion exacerbated pre-existing trends toward gradually waning excess liquidity. Third, in all other periods – particularly during the ECB’s asset purchase programmes – liquidity conditions remained abundant, rendering money market rates largely insensitive. This constellation also seems to apply in the current situation.

Chart C

Time-varying estimates of the liquidity demand elasticity of euro money market rates

a) Spread between the €STR and the deposit facility rate

(left-hand scale: basis points/percentage points; right-hand scale: percentages)

b) Spread between the general collateral repo rate and the deposit facility rate

(left-hand scale: basis points/percentage points; right-hand scale: percentages)

The elasticity of repo rates to liquidity appears to be reacting more strongly than in unsecured rates recently, mirroring global trends. In contrast to unsecured rates, repo rates remained mildly sensitive to liquidity supply fluctuations throughout the period of balance sheet expansion (Chart C, panel b). In fact, repo rate sensitivity strengthened during periods of liquidity growth (e.g. in 2021 and early 2022) and weakened during times of balance sheet reductions (e.g. in 2018/19 and late 2022). In contrast to the traditional relation depicted in Chart A, this pattern suggests the presence of a collateral scarcity channel, where Eurosystem asset purchases constrain collateral availability in repo markets. As collateral becomes scarce, repo rates are subject to more – rather than less – downward pressure relative to the deposit facility rate when excess liquidity expands further.^[2] Conversely, when collateral supply increases, the scarcity premium diminishes, reducing repo rate sensitivity, all else being equal.^[3] Since early 2023 repo rate sensitivity has been rising again in line with the traditional patterns, and secured rates have become mildly more sensitive than unsecured rates.

These findings contrast with recent money market developments in other regions. While euro secured and unsecured rates have exhibited limited reactivity to changes in the supply of excess liquidity up to now, more notable upward moves in secured money market rates were recently recorded in the United States and the United Kingdom. Central bank officials from the two jurisdictions explicitly linked these market moves to a diminishing supply of reserves.^[4] To calm market conditions, the Federal Open Market Committee (FOMC) decided to conclude the reduction of sovereign bond holdings. The Bank of England anticipates a greater reliance by banks on its repo facilities. These developments underscore the usefulness of tools to detect changes in the liquidity environment at an early stage.

References

Afonso, G., Giannone, D., La Spada G. and Williams, J.C. (2022, revised 2025), “Scarce, abundant, or ample? A time-varying model of the reserve demand curve”, Federal Reserve Bank of New York Staff Reports, No 1019.

Hartung, B., Linzert, T., Rahmouni-Rousseau, I., Schneider, Y. and Skrzypińska, M. (2025), “The first year of the Eurosystem’s new operational framework”, The ECB Blog, 25 April.

Saporta, V. (2025), “The evolving liquidity landscape”, speech at the market panel of the ECB Conference on Money Markets.