One thing is becoming clear about Wednesday's cheap money give away from the ECB and that's no one has any idea what the take up will be. Estimates are all over the place, ranging from €100bn to €550bn with lots in between.
Steve Englander says Citigroup's economists are looking for banks to take €400-500bn in the 3-year LTRO, but RBC thinks the demand at this week's tender could be as low as €100bn.
We reckon there is a serious risk that the 3y tender could come out as low as €100bn (the consensus of a Reuters poll last week, bearing in mind expectations have probably increased significantly since then). Estimating the actual amount is difficult given that this tender takes monetary policy into uncharted territory. But applying some rough estimation we forecast a take-up between €200bn-€250bn assuming that banks will not allow a net drain of their cash reserves to take place. If already high ‘excess reserves’ get pared back, there is a risk that the actual tender results even undershoot.
Expectations have in part increased because of speculation that some euro area banks are backing up the truck for a massive carry trade. They have been loading up on PIIG paper with a view to funding them in ECB operations.
Market participants are recalling the first ever 1yr LTRO back in June 2009 where banks demanded EUR442bn and similar stories circulated about banks purchasing short-dated Spanish and Italian bonds, yielding substantially more than the 1% it cost to fund the position back then.
Further complicating matters is that fact the fact that hundreds of billions of ECB debt comes due or has to be rolled this week, says Barclays Capital.
This week a total of €430bn will mature, €292bn at the MRO that will be rolled on Tuesday 20 December and €138bn at the 3m LTRO that will rolled on Wednesday 21 December. In addition, the €57bn that were allotted at the 12m LTRO in October will potentially be rolled into this three-year operation.
For what it's worth BarCap's Giuseppe Maraffino reckons total demand for the 3-year LTRO will be €250-300bn. The take up will come mainly from peripheral countries, for which the long-term funding (both unsecured and secured) market is currently closed.
The table in figure 2 shows an educated guess at the amount likely to be taken at the 3y LTRO based on banks’ behaviour on 24 June 2009 when the first 1y LTRO was allotted. Basically we assume for each operation the same % shift (estimated) as in June 2009. At the previous day’s MRO, in June 2009, borrowing declined by €142.7bn (to €167.9bn), while at the 3m LTRO allotted the same day as the 1y LTRO, the demand for liquidity declined by €22.3bn to €6.4bn. Note also that the borrowing at the STRO (settled on 10 June) declined by €59.3bn (to 56.8bn), which probably was shifted to the same day’s MRO (where the borrowing rose by €74.5bn). Also at the 3m and 6m LTROs held at the beginning of the month (this kind of operations has already been dismissed), borrowing declined by €15.7bn and €20bn, respectively.
Table 2

Netting everything out, RBS reckons there will be €150bn of new borrowing on Wednesday, so the headline figure could be in the range of €450bn-€550bn.
The trade-off for the European banks sits between the risk of negative stigma vs an attractive financial arbitrage opportunity. The average benefit from lower-term funding costs is over 300bp, and significantly more for the weaker. Feedback from discussions with all the banks in our coverage universe points to an intense level of focus on other banks’ willingness to participate. Our working assumption is that most of the large European banks ‘dip their toes’: €5bn each for the biggest 30 would mean €150bn. We suspect that the majority if not all of this would be used to reduce the €800bn of term wholesale maturities to come in 2012, by c20%. However, usage rates could rise dramatically with the second operation on 29 February 2012. Nonetheless, we expect the large European banks to continue to reduce their non-domestic holdings of stressed sovereign bonds.

So RBS is not in the carry trade camp.
Here's why.
Stressed euro zone sovereign bonds command a zero RWA weighting; so combined with their eligibility as collateral for ECB funding, in theory the European banks could benefit from an RoE enhancing, match-funded, positive carry trade. In reality, the impact of collateral margin will slightly increase the ‘all-in’ cost of the ECB’s three-year LTRO operation, but for simplicity the chart below compares the yield of the major euro zone sovereign bonds with a 1% benchmark. Leaving aside the three in-programme euro zone members, this would offer an interesting theoretical arbitrage opportunity by buying three-year sovereign bonds in Italy, Spain or Belgium.
Irrespective of the headline attractions of this carry trade, we expect that until the EU authorities tackle the insolvency concerns at the heart of the euro zone financial crisis the European banks will continue to reduce their holdings of non domestic stressed sovereign bonds. By contrast, it is more reasonable to assume that the domestic banks with stressed sovereigns may well take advantage of this arbitrage opportunity as their exposures are already high.