Post-crisis money markets, monetary policy and financial stability

Fecha: enero 2020

Miguel Carrión Álvarez, Funcas Europe

The money market, where cash is lent and borrowed wholesale, is a key component of the financial system. Broadly speaking money market transactions can be secured, when the borrower pledges some guarantee of repayment, typically using as collateral high-quality liquid assets such as government bonds; or unsecured, when the lending takes place without any guarantee. One of the consequences of the global financial crisis and its aftermath has been the replacement of unsecured by secured lending in the euro money markets. This is a very significant, but somewhat underappreciated, development because interbank lending is an important part of the unsecured money market. This note explores the causes and implications of the virtual disappearance of unsecured lending from money markets.

One of the factors at work is the European Central Bank’s persistent provision of extraordinary amounts of liquidity to the eurozone banking system. However, loose monetary policy cannot be the whole story. The outbreak of the global financial crisis was marked by a breakdown of unsecured interbank lending, and there is evidence that this has not been repaired. The eurozone’s interbank market in particular remains fragmented along national borders. When there is need for liquidity to move between national silos it does so almost exclusively by means of secured lending.

The decreasing importance of unsecured interbank lending means that an increasing share of money-market transactions involve non-banks, namely regulated financial institutions such as insurers, unregulated entities such as private investment funds or nonfinancial corporations, as well as foreign entities. Most money market participants are outside the direct reach of the central bank, be it through monetary policy or banking supervision. A lasting consequence of the financial crisis is therefore an expansion of the shadow banking system.

Additionally, the broader financial markets are underpinned by interest-rate benchmarks based on the money markets. Risk-free interest-rate benchmarks based on surveys of the unsecured interbank market, such as Euribor, are in the process of being replaced by benchmarks based on actual transactions and therefore to some extent on the shadow banking system.

One conclusion is that perhaps repo rates, representative of the secured money market, have become more relevant for the transmission of monetary policy than unsecured lending rates. Another is that the ECB might consider giving non-banks, at least regulated entities such as insurers, access to its monetary policy operations. This would be a significant structural change, however, and would require a focused debate.

1. The disappearance of unsecured lending

The main source of information on the eurozone money markets is the ECB’s biannual euro money market study. The latest issue (European Central Bank, 2019) covers up to the year 2018. Figure 1 captures the evolution of transaction volumes in the various segments of the euro money markets. These include unsecured lending/borrowing or deposit transactions, secured lending/borrowing such as repo and reverse repo transactions, currency swaps and other foreign exchange derivatives, interest-rate swaps such as overnight index swaps (OIS), and the issuance of short-term securities (STS). Derivatives such as swaps used to be mostly uncollateralised and therefore akin to unsecured lending, but since the crisis they have typically become at least partly collateralised and therefore more like secured lending.

Figure 1 quarterly volume by segment (2018 euro money market study)

Until 2011, that is before the height of the eurozone government debt and banking crisis, unsecured money market volumes exceeded €5tn per quarter, with a peak of over 10tn in 2007, corresponding to daily volumes under 100bn and €200bn respectively (on the basis of business days). Since 2011, however, and particularly after 2015 which corresponds to the start of QE, unsecured money market volumes have collapsed to the order of €1tn quarterly, or below €20bn daily. Meanwhile, most of money market volume growth has occurred in the secured segment, which has reached volumes of over 25tn quarterly or over €400bn daily. The foreign-exchange swaps market, also secured, has grown too, in particularly since the start of QE in 2015. On the other hand, the overnight index swap (OIS) market and short-term securities (STS) issuance have shrunk in recent years.

As a result, the unsecured money market has gone from 20%-25% of total transactions before the crisis, i.e. comparable to the volume of the secured money market, to negligible levels of about 5% of the volume in the secured money market in recent years. In addition, ECB data show that only about 15% of unsecured money market transactions are currently in the interbank market.

One of the first indications that a serious financial crisis was under way in August 2007 was a liquidity squeeze in the interbank lending market, which is unsecured. This was because banks suddenly mistrusted each other’s creditworthiness. Later, as a result of the eurozone government debt crisis, the eurozone’s interbank market fragmented along national borders. Essentially banks from creditor countries retained access to the interbank market while banks from debtor countries lost it. The interbank market should have recovered by now, given the efforts made to repair banks’ balance sheets and improve their capital buffers. However, because of the persistence of the so-called “doom loop” between banks and governments, the eurozone’s interbank market remains nationally fragmented.

1.1 Banking union fragmentation

Recent evidence for this fragmentation, discussed in a recent speech by Benoît Coeuré (Coeuré, 2019), comes from the introduction of tiering in the ECB’s deposit rate. This is the remuneration of excess central-bank reserves kept by banks over and above the required minimum of 1% of eligible short-term liabilities, mostly demand deposits.

When the ECB lowered the deposit rate from -0.40% to -0.50% last September, it also exempted some of the excess reserves from the negative deposit rate. The threshold was set at 6 times the minimum requirement, which in the aggregate is about two thirds of the total reserves. Because of the uneven distribution of reserves some banks, mostly in creditor countries, had a large excess reserve position. The tiering of the deposit rate made it profitable for them to move part of these surplus reserves to debtor countries, mostly Italy. This transfer of reserves, however, largely took part by means of secured lending and not through the unsecured interbank market.

Figure 2: the lack of effect of deposit rate tiering on €STR (Coeuré 2019)

We know this because the introduction of tiering had no discernible impact on €STR, the interest-rate benchmark representative of the unsecured interbank marked. This is shown by figure 2.The €STR rate is compiled by ECB compiles from unsecured borrowing transactions settled through the TARGET2 system (European Central Bank, 2020), which is the real-time gross settlement system run by the Eurosystem and through which eurozone banks clear payments.

On the other hand repo rates, which are the prices paid for short-term cash borrowing guaranteed by various kinds of high-quality bonds, were visibly affected by the introduction of tiering. This was the case especially for banks located in Italy and Spain, as shown in Figure 3.

Figure 3: the effect of deposit-rate tiering on the secured money markets (Coeuré 2019)

The way the secured and unsecured market benchmark rates responded, or didn’t respond, to the introduction of deposit rate tiering shows that cross-border interbank lending within the eurozone takes place largely through secured channels such as repo, rather than through unsecured lending. Cross-border movement of funds did take place as illustrated by the right-hand panel of Figure 3. Coeuré also cites data that over 85% of unsecured money-market lending by banks is with counterparties in the same country.

1.2 Quantitative easing

One reason why the unsecured interbank lending market has shrunk is that banks have abundant liquidity as a result of loose ECB monetary policy, in particular quantitative easing. The main structural function of the interbank lending market is for banks to transfer reserves in order to settle payments among themselves and their customers. However, as a result of ECB asset purchases, eurozone banks have on average about nine times the amount of required reserves. This means they do not generally need to borrow reserves from each other to settle transactions.

Indeed, according to ECB data reported by Benoît Coeuré, only 3% of eurozone banks’ unsecured borrowing is with other banks in the eurozone. The rest of the borrowing is from counterparties without access to the ECB’s deposit facility, be it 59% with eurozone non-banks or the remaining 38% with non-eurozone counterparties. This is seen in Figure 4.

Figure 4: shares of euro unsecured money market (Coeuré 2019)

To some extent this is because by-and-large ECB asset purchases have not been from eurozone banks but from eurozone non-banks or non-eurozone financial institutions. When the ECB buys a bond in the open market it pays for it with reserves, but only eurozone banks which are counterparties to ECB monetary policy operations can hold reserve accounts. As a result, ECB bond purchases from entities other than eurozone banks result in some bank intermediating between the ECB and a non-bank or a foreign entity. The bank will be holding reserves at the central bank, and a deposit which is an unsecured liability to the bond seller. So, ECB bond purchases from non-banks or foreign entities will appear as banks borrowing from them unsecured money markets.

This is largely a mechanical consequence of the way the payment system uses banks’ reserve accounts with the Eurosystem for final settlement. Therefore, QE leaves banks with no choice but to borrow from non-banks in the form of deposits.

ECB data (Grandia, Hänling, Lo Russo, & Åberg, 2019) suggests that most ECB net purchases of bonds have not been from eurozone banks. As discussed in an earlier note (Carrión Álvarez, ECB Quantitative Easing as carry trade, 2019), and as shown in Figure 5, eurozone banks have not made significant net sales of eurozone government bonds to the ECB. If they had, the growth of the blue bars representing new reserves would have been at the expense of a reduction in the coloured bars representing liquid assets. But that is not what is observed.

Figure 5: effect of QE on banks' high quality liquid assets (ECB, 2019)

1.3. Securities lending

Accessing the deposit facility is also a driving factor of the secured money market. Due to the ample liquidity available to both banks and non-banks, the secured money market is no longer primarily about borrowing cash. Instead, it has become dominated by securities lending, as financial market participants seek to hold particular kinds of repo collateral, for example German government bonds which are scarce and in high demand. Borrowing securities against cash ultimately results in lending cash to eurozone banks, because they are the only ones with access to the ECB’s deposit facility and that means they don’t need to lend their excess cash in the repo market. The end result is that eurozone banks participate in the secured money markets mainly as securities lenders. This makes banks appear, despite their excess liquidity, as cash borrowers in the repo market.

1.4. Post-crisis regulation

The euro money market study also attributes banks’ shift from unsecured to secured transactions to the post-crisis regulatory environment. There are two reasons why participating in the unsecured money market has become onerous for banks. Both as lenders and as borrowers, banks favour the secured markets because repo transactions are essentially asset swaps and have lower regulatory impacts on both capital and liquidity.

By contrast, unsecured lending incurs higher capital charges than secured lending due to counterparty credit risk. In case the lending is to a non-eurozone entity, cross-border risk also contributes to a bank’s capital cost.

Unsecured borrowing increases a bank’s liquidity needs when the loan matures. Because of post-crisis liquidity regulations, this forces banks to hold high-quality liquid assets to compensate. This effect is particularly strong for banks borrowing for terms of less than one month. This is because of the liquidity coverage ratio, which requires banks to hold high quality liquid assets to cover 100% of their liquidity needs for the next 30 days at any point in time.

2. Policy implications

The changing character of the eurozone money markets presents peculiar challenges for monetary policy and financial stability. One way to see this is to consider interest-rate benchmarks such as Euribor and €STR. These are rates from the unsecured interbank market which underpin the financial markets in everything from consumer credit to derivative contracts. But, as illustrated above, the transaction volume in the unsecured money market has shrunk markedly and the interbank market is a negligible fraction of money market transactions, even restricting attention just to banks’ money-market activity.

This means, first of all, that benchmarks based strictly on unsecured interbank lending as before the crisis are less representative after the crisis. But the ECB’s interest-rate corridor is designed to constrain the range of variation of the overnight unsecured interbank rate only. One possible policy implication is that the ECB could consider adopting a policy to also constrain repo rates.

Secondly, benchmarks representative of the broader money markets, such as repo rates, are exposed to the shadow banking system, which presents its own set of challenges. The policy relevance of both issues is explained in the remainder of this section.

2.1. Interest-rate benchmarks

Before the global financial crisis, Euribor and related benchmarks used to be set through surveys of banks. But, as a result of scandals surrounding the manipulation of Libor and Euribor at the time of the financial crisis, the survey method fell into disrepute. In the EU the 2016 Benchmark Regulation (European Securities and Markets Authority, 2020) requires interest-rate benchmarks to be based on actual transactions to the greatest possible extent. This was deemed to be technically possible for Eonia, the overnight interest rate benchmark, because it was originally based partly on transactions. However, the collapse in the volume of banks’ mutual unsecured term deposits means that the Euribor rates could not be based on transaction data. In the end Eonia was itself replaced by a new euro short-term rate €STR (previously “ester”), and Euribor is likely to be replaced by some elaboration of €STR forward rates.

€STR is an overnight rate compiled by the ECB on the basis of eurozone banks’ borrowing costs in the unsecured money market. It was chosen over two repo-rate alternatives (European Central Bank, 2018). These would have reflected conditions in the secured money market, and not just the borrowing costs of banks but of other market participants. The repo rate alternatives to €STR were STOXX’ GC Pooling Deferred rate, and NEX’ RepoFunds rate. Both are one-day, secured, centrally-cleared, general-collateral repo rates.

The choice of €STR made sense at the time because, being based only on banks’ borrowing costs, it was qualitatively the closest to the Eonia rate it hoped to replace. And it seemed to satisfy the condition of being reactive to changes in the policy rate and other market factors. But, as pointed out by Benoît Coeuré, after the deposit rate was lowered from -0.40% to -0.50% last September, the subsequent introduction of deposit-rate tiering did not have a discernible effect on €STR (see Figure 3 above). Therefore it seems that interest rates based on unsecured transactions (such as €STR and Euribor) have become relatively insensitive to the most recent monetary policy decisions.

This is because, as already mentioned, in order to take advantage of tiering, banks exchanged liquidity though the secured money markets, not the unsecured money markets. This means the repo-rate alternatives such as the STOXX GC Pooling Deferred repo rate (STOXX, 2020) might have been better suited than €STR to replace Eonia.

However, repo-rate benchmarks don’t measure the funding costs of banks specifically. They include all secured borrowers including other regulated non-banks such as insurers, unregulated entities in the shadow-banking system, and foreign entities. Entities other than eurozone banks are not eligible counterparties for ECB monetary policy operations except for outright asset purchases in the open market. Thus, most participants in the secured money markets are outside the direct reach of the ECB. And therefore the ECB cannot influence secured money market benchmarks directly.

2.2. Monetary policy and financial stability

The ECB does not target any particular rate with its monetary policy, but it keeps interbank lending rates within a “corridor”. This is bound below by the overnight deposit rate and above by the so-called marginal facility rate, which is what the ECB charges banks to borrow reserves overnight in case they fall short of minimum requirements. This is so because banks can always borrow at the marginal rate if other banks won’t lend to them below it, and they can lend at the deposit rate if other banks won’t borrow from them above it. Under what used to be normal conditions, before the financial crisis, the ECB managed to anchor the interbank rate by means of its main refinancing operations, which are weekly repo transactions. However, since 2009 the ECB provides banks with unlimited liquidity at the main refinancing rate, and the resulting excess of bank reserves means that the interbank rate hovers just above the deposit rate instead of around the main refinancing rate.

Nevertheless, the ECB used Eonia, and currently uses €STR, as an indicator of conditions in the interbank market. If these get out of kilter the ECB can either intervene with injections or extractions of liquidity from the reserve system, or else it can shift the corridor to accommodate prevailing interbank market conditions.

It should be clear by now that the decreasing turnover of the unsecured money market, let alone the interbank lending market, is a problem for the use of €STR as an indicator. And this is not just a theoretical problem, as the introduction of deposit-rate tiering shows.

More fundamentally, the fact that deposit tiering moved the repo rates but not the €STR raises questions about the channels through which ECB monetary policy is transmitted. Evidently it is transmitted through banks, but then not through the interbank market or even through the broader unsecured money market. As pointed out above, cross-border interbank lending of reserves takes place through the secured money markets and likely involves repo transactions with non-banks. It follows that, to the extent that monetary policy is actually transmitted through the repo market, the ECB may need to change the way it conducts it.

This is important because, as shown by the recent scare in the US dollar repo market, (Carrión Álvarez, What the eurozone can learn from the US repo crisis, 2019) market stress can manifest itself in the repo market only, without affecting the interbank market which is unsecured. In an environment where most interbank lending actually takes place on the secured market, as is the case in the eurozone (given the market behaviour when the ECB introduced deposit-rate tiering), the central bank should be able to intervene in the repo market.

An important consideration in this regard is that, unlike the US Federal Reserve, which conducts monetary policy mostly though trading US Treasury bonds, the ECB already conducts its monetary policy by means of repo transactions. But these are only with banks, which presently don’t really need ECB liquidity as a result of Quantitative Easing. This means that stress could manifest itself in the repo market among non-bank entities that the ECB cannot reach directly. For this reason, Benoît Coeuré has suggested that the ECB could look into extending eligibility for refinancing operations to non-banks. At a minimum this could be done for regulated entities such as pension funds.

In order to do this the ECB would have to allow non-banks to keep reserve accounts. This would mean banks would lose their privileged position in the euro payment system. The ECB would potentially not want to do this for entities not supervised by itself through the Single Supervisory Mechanism. Last year the ECB withdrew a proposal to give central counterparties access to its refinancing operations because it was not given supervisory authority over them (European Central Bank, 2019). So, if the ECB were to give serious thought to opening its repo facility to non-banks, there would need to be a debate about closer coordination between the SSM and the other European Supervisory Authorities; namely, the Banking regulator EBA, the securities markets supervisor ESMA, and the insurance supervisor EIOPA. It is odd that the SSM is not part of the European System of Financial Supervision alongside these other authorities, and it has already had consequences in the case of central counterparties.

3. Summary and conclusions

A structural shift that has occurred as a consequence of the global financial crisis and its aftermath is that the unsecured money market, and in particular unsecured interbank lending, have shrunk in volume. These unsecured lending markets are now essentially negligible compared with the secured money market made up of repo and derivative transactions.

More worryingly, the unsecured money market is proving unresponsive to monetary policy actions such as the introduction of deposit-rate tiering by the ECB, which did have a visible impact on the euro overnight repo market.

Two possible policy consequences can be drawn from this. The first is that the ECB’s euro short term rate or €STR, which has been designated as the replacement for the Eonia overnight benchmark rate, may not be the best replacement after all. Both are unsecured bank borrowing rates, but a repo rate may currently be a preferable benchmark for the conduct of monetary policy.

The second, related, consequence is that it may be reasonable for the ECB to give regulated non-banks such as insurers access to its monetary policy operations, which it conducts by means of weekly repo transactions. But the ECB has already made it clear it will not extend its monetary policy operations to central counterparties that it does not supervise. Therefore, a debate on closer coordination between the SSM and other European supervisory authorities such as ESMA and EIOPA seems necessary and even urgent.

Acknowledgement

The author thanks Raymond Torres of Funcas for helpful comments and suggestions throughout the writing of this note.

References

Carrión Álvarez, M. (2019, August 2). ECB Quantitative Easing as carry trade. Retrieved from Funcas Europe: https://www.funcas.es/funcaseurope/ECB-Quantitative-Easing-as-carry-trade

Carrión Álvarez, M. (2019, October 25). What the eurozone can learn from the US repo crisis. Retrieved from Funcas Europe:https://www.funcas.es/funcaseurope/What-the-eurozone-can-learn-from-the-US-repo-crisis

Coeuré, B. (2019, November 12). A tale of two money markets: fragmentation or concentration. Retrieved from Media:https://www.ecb.europa.eu/press/key/date/2019/html/ecb.sp191112~5808616051.en.html

European Central Bank. (2018, September 13). Private sector working group on euro risk-free rates recommends ESTER as euro risk-free rate. Retrieved from Media: https://www.ecb.europa.eu/press/pr/date/2018/html/ecb.pr180913.en.html

European Central Bank. (2019, March 20). ECB withdraws initiative on conferral of powers in relation to central counterparties (CCPs). Retrieved from Media:https://www.ecb.europa.eu/press/pr/date/2019/html/ecb.pr190320~df3e12da5a.en.html

European Central Bank. (2019, September 27). Euro Money Market Study 2018. Retrieved from Research and Publications: http://ecb.europa.eu/pub/euromoneymarket/html/ecb.euromoneymarket201909_study.en.html

European Central Bank. (2020, January 10). Overview of the euro short-term rate. Retrieved from Statistics: https://www.ecb.europa.eu/stats/financial_markets_and_interest_rates/euro_short-term_rate/html/eurostr_overview.en.html

European Securities and Markets Authority. (2020, January 10). Benchmarks. Retrieved from Policy Activities: https://www.esma.europa.eu/policy-rules/benchmarks

Grandia, R., Hänling, P., Lo Russo, M., & Åberg, P. (Eds.). (2019, February). Availability of high-quality liquid assets and monetary policy operations: an analysis for the euro area. Retrieved from ECB Occasional Papers:https://www.ecb.europa.eu/pub/pdf/scpops/ecb.op218~801632b377.en.pdf?044f2bab998318e649ca0bc899b3095c

STOXX. (2020, January 13). STOXX® GC Pooling EUR Deferred Funding Rate. Retrieved from Indices: https://www.stoxx.com/index-details?symbol=SGCPDFR