In his latest note this morning, DB's Jim Reid admits that "I've no idea why Bunds are rallying so hard at the moment." That said, he does attempt to provide some reasons noting that 10y yields (-4.7bps) hit 0.228% yesterday, down from their YTD peak of 0.495% intraday on the 26th of January. 2y yields also closed another -3.0bps lower yesterday at -0.932%. They traded as ‘high’ as -0.648% back on the same day.
The most obvious explanation is of course Euro systemic risk - especially from France and perhaps Italy. However other markets (equities, equity vol, the Euro, broader credit spreads etc) aren't moving much to price in redenomination risk in Europe. A lack of high quality collateral has been cited as an explanation but it's not clear there is much new info on this over recent days to explain the move. Perhaps it's as simple as government bond investors are generally by nature ultra conservative and Bunds seemingly offer complete safety from redenomination risk.
Whatever the reason, the demand for German paper is nowhere more obvious than the "schatz", the German 2 Year whose yield fell earlier in the week to what was then a record 0.92%, and has since continued to fall, earlier in the session sliding to a new all time low of -0.95%, down 15 bps on the week, before rebounding modestly as ravenous credit traders snapped up every German asset they can find amid rising political fears, and the avocementioned concerns about a collateral shortage.
While in the earlier part of the week the appetite for German paper was as a result of a sudden spike for Marine Le Pen in the presidential polls, this has since moderate and yet the flight to German safety has not faded.
Schatz prices have also been boosted by reports the ECB has been buying the paper at recent prices, after it revised its bond-buying parameters. Previously, the central bank limited its sovereign debt purchases to yields above -0.4 per cent.
But perhaps the big catalyst for today's move is a note out of Citi's harvinder Sian and Jamie Searle, who expect to see the Schatz to drop to -1%, for one main reason which has nothing to do with French politics: the ECB needs to buy around EU80b 1-6y Germany by year-end, and as a result traders are merely frontrunning the ECB. They also expect the Bund yield to plunge again, dropping as low as -0.10%.
They write that in QE simulations, based on assumptions and judgments due to the lack of QE holdings data, there is no credible alternative to front-end buying in Germany. The simulation suggests that by Dec. 2017 the available pool of Bunds will be exhausted across all sectors, and this dwindling pool of Bunds raises questions over the ability to extend QE.
Here are the highlights from their note:
Schatz made record yield lows this week at -0.91% (and Bobls at -0.56% were only Gbp off their record). That is despite strong data and ECB hawks wanting to remove rate cut possibilities from forward guidance. We think Schatz yields can go even lower, initial targeting -1%.
This week's rally has been driven by French politics and associated tail risk redenomination concerns. But, the rally owes a lot to the ECB beginning to purchase bonds below the depo rate (which started on the 16 January according to the ECB accounts). This is likely to be the lasting driver.
That is because prioritising QE purchases above the depo rate doesn't work for Germany. Just to complete the QE already announced (to Dec-17) requires a heavy dependency on the 1.6yr sector of the German curve, according to our estimates. Our simulation suggests that the monthly average maturity of purchases will fall to just above 6 years very quickly. and stay there. To put a number on it, we think the ECB will buy around €80bn 1-6yr Bunds just to complete QE to year-end.
QE should keep up the downward pressure on Schatz and Bobl yields, regardless of developments in France. It should also drive yields lower further out given the curve is already steep (we forecast a low for lOs of 0.10%). It also means euro swap spreads can continue to richen, even from here.
Another really important point falls out from our simulation of German QE to year-end. It is surprisingly challenging to extend QE beyond Dec-17 without a hard break of the capital key. That greatly reinforces our concern that this heralds the end of the ECB's lender-of-last-resort function. Huge risks lie ahead for the EMU periphery.
This week: 1) French politics and the chart that scared the market; 2) Simulating GE - why Schatz will stay rich; 3) QE extension beyond Dec-17 is harder than it seems; 4) German yields to fall, regardless of French risks.
And here is Citi's simulation of QE, according to which the central bank will keep buying this most desired of German financial assets:
Simulating QE - why Schatz will stay rich
To understand the reliance of the German portion of QE (specifically the PSPP) on the 1-6yr sector, we have run a simulation to Dec-17 (announced QE).
We stress from the outset that this requires many assumptions and judgments on our part due to the lack of holdings data for ECB QE.
The starting point for our simulation is how much we think there is left to buy while respecting the 33% issue/issuer limits. As a reminder, the German portion of PSPP can be made up of sovereign bonds, agencies and regional debt. The estimates in Figure 4 are derived from our model' for QE purchases to date.
An important assumption in the model is that purchases so far have been split 85% in Bunds, 10% in agencies and 5% in regional debt. This is tilted towards Bunds -which are only 60% of the eligible pool - to reflect relative liquidity the and relatively small outstanding sizes of agencies and regional debt. As we will see, how the asset split evolves is crucial to the longevity of QE.
How much left to buy?
Figure 4 makes it clear that the bulk of Bunds available to buy are in the front-end. And this is only thanks to the ECB now allowing purchases below the depo rate. Before this parameter change, Bunds were running out on our model.
There is quite a bit of agency and regional debt available to buy (respecting 33%), but we expect that we are overestimating. On regional debt in particular we may even be doubling the real potential (as we include all listed bonds which fit the ECB criteria of which quite a sizeable portion may be registered bonds). In reality, only benchmark format laender bonds are likely to be bought and may be hard to find.
Running a QE simulation to Dec-17
The model gives us a starting point for our simulation for monthly PSPP purchases of German debt to Dec-17. We calibrate February purchases to have an average maturity of around 7.5 years given the January figure. That involves 85% allocation to Bunds, 9% to agencies and 6% to regional debt. Within Bunds, the bulk is allocated to 1-6yr. A higher allocation to 15yr+ means it would run out very quickly. The allocation to 6-15yr is largely determined by new supply (which we allow for together with bonds dropping below lyr). Further, by running this simulation it quickly becomes apparent that the 85% allocation to Bunds is hard to sustain without prematurely running out of Bunds before Dec-17. Our simulation has this dropping to 80% then 75%.
* * *
If Citi is correct, expect a continuation of the bifurcated paradox of stocks rising even as the German "flight to safety" asset continues to show a level of near panic by market investors, leading to even more confusion.