Euro short-term rate (€STR) questions and answers
The €STR is published on every TARGET2 business day at 08:00 CET. If errors are detected following standard publication that affect the published €STR by more than 2 basis points, the €STR is revised and republished on the same day at 09:00 CET.
The ECB rate is based exclusively on transaction-by-transaction data reported in accordance with the MMSR Regulation. The MMSR Regulation specifies that data shall be transmitted once per day to the ECB between 18:00 CET on the trade date and 07:00 CET on the first TARGET2 settlement day after the trade date. The complete dataset is therefore only available for the computation of the €STR after 07:00 CET on the following TARGET2 day.
In order to ensure timely publication, the publication process is highly automated, using algorithms to automatically filter out trades that deviate from usual patterns. Such trades, however, can be re-integrated upon confirmation by the reporting banks.
The identifiers for the euro short-term rate are:
German WKN: A2X2A2
FISN: ECB/EUR EURO SHORT-TERM RATE IR
The ECB decided to develop an unsecured rate, instead of a secured rate, for a number of reasons. First, the €STR is intended to complement and serve as a backstop to existing critical benchmark rates such as the euro overnight index average (EONIA), which reflect the unsecured money market. In this respect, the €STR should have features that would make it comparable to these rates. Second, the repo market has a number of unique characteristics related to
- the motivation for entering into a trade
- the difficulty of differentiating between general collateral and special collateral if the aim is primarily to measure the price of cash, and
- the type of collateral, which would affect the formation of the final rate of a repo transaction.
For example, the price of a repo can vary considerably depending on the availability and use of collateral and the credit rating of the issuers of the collateral. Furthermore, the share of general collateral versus special collateral and the degree of “specialness” vary significantly over time, which reflects the respective countries’ issuance cycle in the absence of a homogeneous European collateral market and the influence of certain reporting dates, such as year-end reporting. As a result, it would be very challenging to develop a rate that is expected to have broad euro area coverage and meaningful, consistent prices in the underlying transactions at the same time. Moreover, when comparing similar notions based on secured transactions to determine the price of overnight cash (only general collateral trades, with the same overnight tenor as the €STR), money market statistical reporting (MMSR) data show that the daily average overnight general collateral repo volume in 2017 (excluding “specials”), as traded by the 52 reporting agents, would have amounted to around €60 billion. This is higher than, but not fundamentally different from, the volumes captured in the unsecured market by the €STR, although pricing remains subject to significant fluctuations on reporting dates.
Finally, there are already a number of existing repo benchmarks, which the ECB welcomes as the availability of more benchmarks will allow users to choose the most suitable one for their needs.
The ECB decided to publish an overnight benchmark because the absence of a reliable private benchmark could result in a potential adverse impact on the transmission mechanism of monetary policy and may have repercussions for financial stability.
However, the ECB is not in a position to provide longer-term reference rates (i.e. beyond the overnight maturity). To judge by the results of the pre-live verification exercise of the European Money Markets Institute (EMMI) and also by the data provided under the ECB MMSR Regulation, it appears that there are currently not enough transactions to construct purely transaction-based longer-tenor reference rates. This means that some expert judgement may be required in order to sustain daily benchmark publications on such tenors. Such judgement cannot come from a central bank for two main reasons. First, the central bank may not have the same overview of the prevailing market conditions and the funding costs of banks as credit institutions have. Second, expert judgement, if provided by a central bank, might be interpreted as being related to the (desired) monetary policy stance; this might create, or be perceived as creating, a conflict of interest.
Tell me more: Why are benchmark rates so important?
Counterparties and instruments
They have one feature in common – they both rely on transactions from the euro-denominated overnight unsecured money market segment.
However, they differ in several ways. First, EONIA is administrated by the private sector via the European Money Markets Institute (EMMI), while the €STR will be administrated by the ECB. EONIA relies on voluntary data input by 28 panel banks (with one contribution per bank), while the ECB’s new rate will be built on the daily data submissions of the banks reporting in accordance with the MMSR Regulation. Moreover, EONIA is a weighted average rate of the submitted contributions; the €STR relies on individual transactions rather than on a single contribution per bank. Furthermore, the €STR is based on unsecured overnight borrowing deposit transactions, while EONIA is calculated using unsecured overnight lending transactions.
For more information about the methodology for the €STR, please refer to the “Euro short-term rate methodology and policies”.
The broader scope of the €STR is intended to respond to the developments of the wholesale market in recent years. More specifically, the share of the interbank market in the wholesale market became smaller owing to a reassessment of counterparty risks, changing regulations and liquidity conditions. However, banks developed significant money market activity with other entities, such as money market funds, insurance companies and other financial corporations. For that reason, all of these counterparties play an important role in the wholesale funding mix of banks and are therefore considered relevant for determining wholesale borrowing costs.
Nevertheless, and as mentioned in the first public consultation on developing a euro unsecured overnight interest rate, other counterparty sectors such as governments or non-financial corporations will not be taken into account in the €STR in order to reduce the influence of possible idiosyncratic factors on the final rate.
The broad scope of the €STR guarantees that the rate is a fair reflection of the overnight borrowing cost for banks in the wholesale market, in which not only banks but also a number of other different entities interact. Some of these entities may not have access to the Eurosystem monetary policy operations (because they are non-banks or are located outside the euro area), which means that the rates of the Eurosystem facilities will not strictly serve as a lower or upper bound for the rate of their transactions. As a result, such transactions may be conducted at a rate below the deposit facility rate or above the marginal lending facility rate. The €STR is intended to capture this market reality. For example, in conditions of abundant excess liquidity, the €STR would be expected to be below the deposit facility rate. The position of the rate in relation to the Eurosystem policy rates, however, does not mean that the rate will be unable to respond to changes in the policy rates. In fact, since the €STR reflects a liquid market with multiple participants and therefore competitive pricing, these prices are expected to follow the direction of the policy rates.
The €STR is intended to be a borrowing rate, which means that it is more representative if it captures trades with all significant counterparties in the wholesale market, including international counterparties. Furthermore, excluding transactions with non-euro area counterparties would not be sufficient to ensure that the only eligible transactions are those conducted with counterparties that have access to the Eurosystem facilities. If that were the intention, the scope of the ECB rate would have to be reduced to only the interbank market, where counterparties are banks with access to the ECB facilities. This, however, would result in a lack of data and therefore the final rate may not be considered robust.
The selection of eligible instruments for the €STR was presented in the first ECB public consultation on developing an unsecured overnight interest rate; it was argued that only money market deposits should be used for the computation of the €STR, because deposits are standardised products with easily understandable pricing rules that ensure the consistency of the rate. As shown in the second ECB public consultation, there are sufficient data on deposit transactions to produce a reliable daily reference interest rate.
Call accounts as captured by the MMSR have been analysed from three perspectives:
- contribution to data sufficiency
- level of standardisation (homogeneous product type with pricing and understanding of the rate)
- rate behaviour (level and volatility in line with market conditions)
With regard to data sufficiency, including call accounts would have increased the volume underlying the computation of the rate by around €10 billion on average, which may have supported their inclusion. However, call accounts would have improved neither the country representativeness of the rate nor the concentration, given that call accounts are used in very few jurisdictions, Germany being one example.
With regard to the level of standardisation, including call accounts would reduce the clarity of the envisaged scope (deposits) and make the rate more vulnerable to idiosyncrasies as discussed in the first public consultation. Indeed, the definition of call accounts is quite vague owing to the various non-harmonised legal frameworks in the euro area for this financial product. The definition includes savings accounts, which are also defined in a relatively broad manner in the MMSR Reporting Instructions.
With regard to rate behaviour, the rates of call accounts as captured by the MMSR appear quite “sticky”. Data suggest that including call accounts would have been likely to reduce the responsiveness of the €STR to ECB policy rate changes. This observation was even clearer at individual reporting agent level. Rates often remained at exactly the same levels for extended periods of time suggesting the rates were not renegotiated in the market, as otherwise there would have been daily fluctuations.
Finally, short-term papers as reported under the MMSR were also analysed. However, the very limited volumes captured by the MMSR and the quite volatile rate behaviour were seen as reasons not to include short-term papers in the computation of the €STR at this stage.
The money market statistical reporting (MMSR) sample currently covers the 50 largest banks in the euro area in terms of balance sheet size at the time of selection. The 50 reporting banks are spread across ten euro area countries (Belgium, Germany, Ireland, Greece, Spain, France, Italy, Netherlands, Austria and Finland).
With regard to the possible impact of the expansion of the reporting population, the analysis on data sufficiency conducted in the context of the second ECB public consultation and earlier evidence from the ECB’s Euro Money Market Surveys suggest that the unsecured money market tends to be a concentrated market, as also shown in the first ECB public consultation.
The legal status of the reporting banks as MMSR reporting agents will not change following the release of the €STR. The €STR will be based exclusively on the statistical information on transactions reported to the ECB or the NCBs under the MMSR.
The reporting banks will continue to have obligations pursuant to the MMSR Regulation and the overall ECB statistical framework. Amendments to the MMSR Regulation will follow the established rules and procedures, and where required will be announced publicly well in advance and will involve consultation with the European Commission.
The banks reporting MMSR data on which the €STR is based are listed here: List of reporting banks.
The thresholds ensure that the €STR is published on the basis of data provided by a sufficient number of banks, although none of those banks would have too large an influence on the final rate.
The pre-€STR showed that there would have been very few cases of data insufficiency in recent years.
Any change in market dynamics that leads to deterioration in market liquidity would need to be considered in a regular or ad-hoc reassessment of the methodology of the rate.
A volume-based trigger was considered but ultimately not deemed desirable. As explained in the second ECB public consultation, day-to-day fluctuations in volume can be considered part of how markets function. Such changes could relate to calendar effects or local holidays in the various euro area countries. MMSR data show that, even on days with reduced volumes, those volumes are generated by a fairly large number of reporting banks with no additional concentration of activity, therefore a rate calculation based on lower volumes could be seen as robust and unbiased.
In the second ECB public consultation, a number of respondents expressed concern that the proposed trimming value of 25% would be too high and could undermine the rate’s representativeness. However, the trimming value does not affect the rate representativeness, and in fact improves the stability and resilience of the €STR.
Regarding the representativeness, the trimmed mean – like the arithmetic mean and the median – is a measure of the central tendency of the distribution of rates, and existing MMSR data confirm that the characteristics of the distribution of rates are such that the trimmed mean is a proper measure of this central tendency.
Moreover, trimming is used to reduce the impact of significant outliers on the computation of the €STR; the threshold of 25% was found in the second public consultation to be close to optimal in reducing the variability of the rate in a day to day basis while ensuring a broad calculation basis. The difference between trimming at 25% compared to 10% on the trimmed mean is very limited, with only around 0.1 basis points on average (see Second public consultation), while the 25% trimming shows less day to day volatility, and is thus the selected choice for the €STR.