Another reason why we don’t like the meme of viewing the ECB’s three-year liquidity as (or in any way analogous to) “quantitative easing” for sovereigns:
Mark-to-market risks remain key
As the LTRO is a repo transaction the ECB takes in collateral in order to back any loans. This highlights one of the key differences from QE, which entails the asset risk being removed from the bank’s balance sheet and replaced with cash. Under the LTRO framework the economic benefit remains with the bank. With the majority of the funding being utilised by banks in peripheral countries, their asset base is likely to be for the most part correlated with the performance of domestic sovereign bonds. Figure 1 below shows the increase in realised volatility across peripheral bonds over the past few years. The significant increase in volatility, although currently working positively for collateral correlated with sovereign bonds, could lead to greater variation margin call risk as collateral deviates from threshold levels. Asymmetry persists, with a major problem being a discrete risk event leading to a margin call, which could trigger a chain of margin calls.
That’s from Nomura rates strategist Guy Mandy, who’s had a thing or two to say about haircuts at the ECB before…
To go back to basics, what Nomura’s Mandy is talking about here are the ECB’s risk controls on the assets it accepts as collateral. In the first place, if the asset’s market price drops, the ECB normally requires the pledging bank to stump cash or more assets, using daily marking to the market. (The Banque of France, for example, outsourced its collateral management ops to a third-party agent to make the processes more efficient.)
In this case, we’re talking about the MTM or variation margin-call risk, which is a point that’s been in the air since the three-year LTROs were announced in December. Mandy’s also talking about broader ECB management of credit risk though, including ratings thresholds. He notes that if Moody’s and Fitch join S&P in cutting Italy to a BBB-level credit, it could lead the ECB to impose an additional five per cent margin on Italian bonds pledged into the LTRO. We wonder if that’ll start a debate about the central bank waiving its ratings requirement for Italian debt, like it’s done with other peripheral sovereigns…
But while that kind of detail should trigger a few doubts about the LTROs as an “indirect” QE for eurozone sovereigns, it’s actually Nomura’s points about ECB haircuts on other kinds of collateral that we find especially interesting. Notably credit claims – un-securitised bank loans that have only just been accepted as eligible collateral at the ECB itself, (as opposed to the national eurozone central banks) in tandem with the announcement of the three-year LTROs. February’s three-year LTRO will probably see quite a few credit claims being pledged, as it’s taken all this time for banks to prepare them.
Though Mandy has charted the likely haircuts applicable to these loans, and used them to make some interesting points about the ‘real’ funding costs for banks who tap ECB liquidity with low-quality collateral subject to high haircuts:
The cost of the haircuts and existing funding costs on leveraged banks balance sheets means that the cost of funding from the ECB is not 1%. For example, if a firm uses a 5-year maturity or greater loan book as collateral there is at minimum a 29% haircut, but there would also be a capital charge levied under Basel rulings given there is a requirement to hold varying amounts of capital against assets related to their riskiness. As an example, we assume for simplicity a blended cost (cost of capital and funding) of 5% and a capital charge of 5% against the loan book. In this case against a €100 asset there would be a €29 haircut and €5 capital charge giving €34 to fund through other sources. At a 5% blended rate the funding cost would be €1.70. Adding to this the €1 cost via the ECB, the actual all-in cost is approximately 2.7% on one year money. This effectively creates a floor to the front end of sovereign curves.
The really interesting thing is what happens if an ECB margin call hits a bank for cash, when (as with today’s conditions in money markets) that bank can probably only find the cash from the secured market. Unsecured funding is closed (to all but the bestest of the best), ergo the bank scrapes together assets to pledge for cash somewhere, running the gauntlet of the collateral crunch. As Mandy says, it’s actually likely that ‘somewhere’ would be the ECB itself. Pledge assets to the ECB, to meet ECB haircuts on pledged assets…
In short, it’s not unlike the margin spiral effect seen in repo markets in 2008. It’s partly why we do see the ECB’s actions as more like the open market operations of autumn that year rather than full QE to take assets completely off bank balance sheets. Although we wonder what the ECB’s margin calls might mean when the market as a whole is far more dependent on the secured funding market than in 2008…
Full note in the usual place.
Related links:
What the repo markets *want* the ECB to do – FT Alphaville
Risk mitigation and credit assessment framework (section 6.3) – ECB
But what does Mrs Watanabe think? – FT Alphaville
ECB reserves related to margin calls - FT Alphaville