Yves Mersch: Ructions in the repo market - monetary easing or regulatory squeezing?
Speech by Mr Yves Mersch, Member of the Executive Board of the European Central Bank, at the GFF (Deutsche Börse Group Global Funding and Financing) summit, Luxembourg, 26 January 2017.
There is no doubt that European repo markets today are operating in unprecedented territory. In the past year we have faced the extraordinary situation of collateralised transactions taking place well below the ECB's deposit rate, and not just for the highest-rated issuers. There are increasing signs that this is indicative of market stress.
Though market depth remains relatively stable and the bid offer spreads are not a major liquidity restriction, the decreased average ticket size seems reflective of collateral scarcity in some market segments.
If one looks at the distribution of trades taking place by rates towards the end of last year, while German collateral usually traded at around -70 bps, it traded at -4.88% at year-end, reflecting increased demand for "special" German bonds (while overall turnover declined about 50%). In a "special" the repo contract specifies a particular bond to be exchanged for cash, rather than the specification of a broad collateral class such as 'German government bond'. As a result, specials tend to trade at lower interest rates.
The volume of trades conducted at special rates also increased in other markets, pushing the distribution of rates lower. Average French collateral rates have moved from -45 bps before summer 2016 to below -60 bps before Christmas. Spain is trading between -45 bps and -50 bps. Italy is now trading below -40 bps1. And for Austria, Belgium and The Netherlands, rates have been moving downwards as well.
One could say that whereas in the past the repo market was largely characterised by "collateral chasing cash", today the main dynamic in some markets appears to be "cash chasing collateral". The evolution of repo rates for different collateral classes over reporting dates suggest that the old pattern still exists in parallel with the new. As I will argue below, this observation provides a key to understanding the roots of these developments.
New dynamics in the short term funding markets unsurprisingly lead to debate in the financial industry. There are concerns about short-term borrowing costs swinging sharply at the end of reporting periods, creating volatility and potentially spilling over into government bond markets - an effect we have indeed seen for core general collateral (GC) collateral. And some observers fear collateral scarcity might herald more general disruptions to the "plumbing" of the financial system, which would be especially problematic given ongoing structural changes in financial intermediation. Activity is after all shifting increasingly from banks to non-banks, and market-based finance is, by and large, organised around collateralised lending.
Perhaps inevitably, some point the finger at monetary policy, and in particular at the side-effects of the ECB's asset purchase programme. The substantial creation of liquidity by the APP is argued to have pushed unsecured lending rates to the deposit floor, causing secured lending rates to fall even lower. At the same time, the APP is said to have withdrawn increasing volumes of high-quality liquid collateral from the market, exacerbating pricing tensions, especially for special bonds.
While monetary policy may, in some circumstances, exacerbate repo market tensions, I believe it to be far from the only factor at play today, or indeed the most important one. The situation in the repo market today in fact reflects a confluence of longer-term trends that have affected both the supply and demand for good collateral. I will review some of these trends in a moment.
Before doing so, I would like to recall an important point. Our mandate is clear: to deliver price stability. This is our primary objective and the sole orientation for our policy. What we aim to do is to act in a way that is necessary, proportionate and that limits side-effects. This is why we have designed our programmes in such a way as to deliver our objective without creating excessive distortions in market functioning and price discovery. And it is also why we have taken measures to mitigate some effects of our measures on repo markets. These measure, too, are governed by the principle that they must be necessary and proportionate, as well as not being in conflict with our price stability mandate.
Contributing factors to the current situation in repo market
Let me now review some medium to long-term trends contributing to the current situation in the repo market, affecting both the supply and demand of collateral.
It should first be noted that the repo market has seen a reduction in its role for funding, also due to the excess liquidity injected and the securities absorption implied by APP.
On the supply side, the availability of high-quality collateral is inevitably pro-cyclical. From a pure economic view point, the financial crisis, its length, rising debt levels and rating downgrades to sovereigns reduced the amount of available assets that are considered 'safe', particularly for non-banks (although regulators treat all sovereigns as risk-free). This is most easily visible in the current clustering of GC collateral rates by country of issuer, which was largely absent prior to the crisis.
On the demand side, the evolution of prudential regulation has affected the repo market. A number of features, which have in principle been known since the reform of Basel III in 2010, are still being phased in and directly affect banks' participation in the repo markets, particularly the liquidity coverage ratio (LCR) and the net stable funding ratio (NSFR). A third measure - the leverage ratio - is now subject to reporting requirements, but there is not as yet an agreed international regulatory standard.
- The LCR requires banks to hold sufficient high-quality liquid assets (HQLA) to cover projected net cash outflows. It started being phased in at the start of 2015 and reaches full application by 2018. The LCR consequently makes short-term funding less attractive to banks and holding HQLA more attractive. The requirement to meet this ratio is likely to be a contributory factor in market volatility around reporting dates.
- The NSFR is designed to reduce bank reliance on short-term wholesale funding relative to more stable sources of funding. It penalises short-term funding, including repos of shorter maturity than one year, particularly those under six months. The NSFR also penalises matched book transactions, since the negative impact of the reverse repo leg more than offsets the positive impact of the repo leg. Overall, the impact from the NSFR on short-term repo is reckoned to be lower supply, reducing volumes and increasing the price.
- The leverage ratio, which has been implemented in the United States for some time, is a deliberately simplified way to assess banks' riskiness. Traditional bank capital ratios are calculated as capital over risk-weighted assets. The leverage ratio is calculated approximately as capital over total assets, with no attempt to employ risk weights.
This is important for the repo market because sovereign bonds typically attract a zero risk weight. As such, carrying out repo as a large volume, low margin activity is threatened by the leverage ratio. Although the exact parameters of the leverage ratio regulations have not been finalised, euro area banks are required to report leverage ratios. Given the higher relative penalty given to holding sovereign bonds under the leverage ratio as opposed to the capital ratio, it is expected to increase the cost of banks' repo, and reduce incentives for money market arbitrage and market making.2
Reduced incentives for arbitrage and market making in turn affect market liquidity and can lead to greater periods of uncertainty. Furthermore, since European banks are required to report average month-end positions, there are increasing signs of market stress at month-end dates.
The leverage ratio contains provisions for netting of offsetting cash positions when they are with the same counterparty and settled through the same system. This has resulted in a much greater use of central counterparties (CCPs). The use of CCPs has been strongly supported by authorities and brings benefits in terms of smooth market functioning, as well as for individual institutions. Eurosystem contacts report a wide difference in the pricing of trades cleared bilaterally and those cleared through CCPs - sometimes in the tens of basis points.
CCPs, however, are also important for repo markets for a different reason. The increased volume of derivative trades centrally cleared at CCPs due to new regulations such as EMIR, as well as sporadic bouts of volatility, means that those institutions tend to hold significant quantities of cash. Placing that cash as unsecured deposits would represent wrong-way risk for CCPs as the collapse of a significant financial institution would create significant market volatility at the same time as potentially the loss of some of the unsecured deposits. CCPs are therefore increasingly large participants in the repo markets.
The requirements for trades to be centrally cleared are still being introduced, so the demand from market infrastructure to exchange cash for collateral will rise.
The rise in importance of CCPs is indicative of the generally increasing role played by non-bank financial institutions in recent years. Since these institutions do not hold accounts at the central bank, they are unable to access the deposit facility. The demand from these non-banks to find a secure place to deposit cash is a contributory factor to repo rates trading below the Eurosystem's Deposit Facility Rate (DFR).
Last but not least, changes in the interbank lending market have led to increased collateral demand. In the period since the onset of the crisis, loan losses and weak profitability have constrained the availability of capital for banks. Unsurprisingly, there has been a marked trend in interbank markets away from unsecured towards secured lending in order to achieve more favourable risk weights for interbank lending.
I would also like to acknowledge at this point the role of the Eurosystem's monetary policy in shifting the balance of the repo market. Since the onset of the crisis, the ECB has undertaken a number of necessary and proportionate measures to ensure the efficacy of monetary policy. The first measure relevant here was introduced to repair the transmission mechanism, which had become impaired and threatened the singleness of monetary policy within the euro area.
Prior to the crisis, internal trade imbalances in the euro area had been primarily financed via transactions between country banking sectors. Following the onset of the crisis, the interbank market became fragmented, with a marked reduction in cross-border transactions. It became increasingly clear to the ECB that this impairment in the interbank market interrupted the even transmission of monetary policy across the euro area. The ECB responded by introducing full allotment - we provided as much liquidity as the market required at our main refinancing rate. By statute we only lend against good collateral and so more lending by us means more demand for collateral.
The other important monetary policy measure with regards the functioning of the repo markets is the asset purchase programme (APP), and in particular the public sector purchase programme (PSPP). It is worth recognising firstly that although the APP has led to a shrinking pool of liquid securities, it has also resulted in a net increase in liquid assets through bank reserves in exchange of inter alia non-HQLA marketable securities.
But the APP has clearly reduced the available pool of L1 securities for other investors. To the extent that non-banks seek to maintain their holdings of government securities, this is likely to cause additional pressure on the repo market.
These monetary policy measures carried out by the Eurosystem have been necessary to ensure the continued orderly transmission of monetary policy throughout the euro area and to provide sufficient stimulus to bring inflation back to our objective. They have also been proportionate solutions to the scale of the problems they seek to solve.
The ECB is aware of - and closely monitors - the potential impact of its monetary policy measures on market functioning. Indeed, as I mentioned earlier, full allotment was put in place precisely because the market wasn't functioning. But in assessing the impact of its policies, the ECB must have first regard to its primary objective of price stability. Under our mandate, considerations of the robustness of the financial system and market functioning are only relevant once we have fulfilled that primary objective.
Nonetheless, in designing the APP programme, the ECB has implemented a number of features designed to minimise the impact on market functioning. The hard legal constraints on issue and issuer limits reduce the influence the APP has on particular bonds. Our recent widening of eligible maturities and removal of the DFR floor permit purchases to be spread more widely and reduce pressure on specials. We have also put in place a number of securities lending facilities. Recent moves to improve these facilities, including the use of cash collateral, have been received positively by market participants.3
Future direction for the repo market
Having described the longer-term trends driving current conditions in the repo market, I would like to spend the rest of my speech today discussing prospects for the future.
I would argue that the supply of safe assets will increase again over time. As I mentioned earlier, the volume of safe assets is pro-cyclical. The economic recovery underway in the euro area will strengthen government finances. Furthermore, once inflation is sustainably back to our objective, monetary policy will normalise, reducing any additional strain our unconventional measures may be placing on repo markets today.
The LCR and NSFR are relatively new as regulations, and the exact future form of leverage ratios is still not finalised. European banks are currently somewhat less advanced than their US counterparts in re-optimising their business models in the face of these regulations. Yet already Eurosystem contacts tell us that repo desks are being reorganised and integrated more tightly into banks' overall decisions on balance sheet and liquidity usage. Certainly a range of potential tools exist, from tri-party repo, to trade compression, intra-day margining and so forth.
Regulation may also encourage a greater use of unsecured lending. With a potentially binding leverage ratio, banks may shift their balance sheets towards more risk-weighted assets, trading off capital encumbrance against balance sheet encumbrance. Such a shift would reverse the current trend towards more secured lending in the interbank market. Furthermore, as the economic recovery continues, bank profitability will increase, bolstering equity. With less constrained equity, conditions will become more favourable for a return to more unsecured lending.
Let me conclude.
Banks and other market participants report a decrease in market making activities and collateral scarcity in repo-markets. The perception of a challenging environment can therefore not be denied.
Pro-cyclical market reactions during and after the financial crisis and - related - changes in banking regulation clearly had an impact on the functioning of the repo markets. But with the continued economic recovery market stress should fade and conditions for unsecured lending might re-gain some attractiveness. Likewise, market players will adjust their business models over time amid a new regulatory environment.
Regarding the ECB's asset purchase program, two important points have to be made:
Although the APP might have some negative side effects on the functioning of the repo market as compared to the pre-crisis levels, other more potent factors were at work. To reduce spill-over effects the ECB has also introduced several mitigation measures.
Even more important: the ECB has one mandate, which is to ensure price stability. Elevated risks to its goal have made unconventional measures necessary on an unpreceded scale and on a temporary basis. And they prove successful. Although we try to minimise negative side effects, any of these considerations are without prejudice to the obligation to honour our mandate.
1 Rates based on RepoFunds Germany, France, Italy and Spain.
3 ICMA (2017), Regulatory Policy Newsletter, January.