It was supposed to come with a "bang." Instead, the double trial balloon launched by the ECB on Thursday night, delivered at the same time by both Reuters and Bloomberg to make sure that everyone got it and according to which the ECB was considering slashing its QE in half from €60 billion to €30 billion and keeping the program active for at least 9 months, revealed one of the biggest market reaction "whimpers" to an ECB leak in years.
For those who missed it, here again are the bullets:
- ECB has consensus to extend asset purchases at lower volumes on Oct 26
- Agreed on reducing buys from €60 bln/month for nine months
- Debating buys between €25-40 bln - with a likely final bogey of €30 bln - and whether program should be open-ended
Perhaps the market reaction was unexpectedly muted because the announcement was i) largely in line with expectations and ii) still to be finalized. Naturally, the lack of an acute selloff in Bunds (and rates globally) was welcome news to the European central bank, for which another major bond tantrum would be the worst case scenario, one which could promptly unravel the entire carefully planned pre-tapering edifice.
The problem, however, is that the complete lack of a reaction failed to give Mario Draghi a sense of whether the €30 billion bogey was too high, too low or, maybe, just right.
Still, as Citi's Harvinder Sian writes in a Friday note, the market impact from the ECB meeting is likely to be determined by the signaling channel for policy rates which is imbedded in the QE extension. In other words, the final framework of the ECB's QE will certainly move markets, "although the purchase horizon is not the only variable that matters: the size of purchases will signal whether QE is most likely to come to a hard-stop or be open to continuation."
So what are the possible scenarios for the ECB to achieve it aim of tapering without tantrum, and via signaling. That will require a judgement on how the rates market will react to the new calibration of size and horizon, of which there are many options.
To answer that question, Citi take a look at the market impact from the various ECB policy permutations.
The starting point is that rates do not rise until ‘well past’ the net asset purchase period. The purchase horizon, however, is not the only variable to assess because a hard-stop at €20bn is easier to envisage than a hard-stop on a €40bn announcement. We look through these size/horizon combinations and model the EONIA curve for timing of the first rate hike, time to zero rates and the average hike per meeting thereafter. That is the framework for assessing euro swaps and Bund impacts. The bullish to bearish scenarios for 10yr Bunds range from -25bp to +24bp (one-week impacts) from current levels. The market neutral level of APP extension appears to be around +€250bn.
Or, just under €30 billion over 9 months, precisely what Reuters and BBG "leaked."
Below, Citi presents its EONIA-signaling driven model which aims to judge the near-term market impact (around one-week) across the euro swap and Bund curve for a range of QE scenarios.
Below are its key findings:
- For 10yr Bunds, yields fall around 25bp in the most dovish scenario (€40bn x 12mth) and rise around 24bp in the most hawkish scenario (€20bn x 6mth). Figure 1 above summarizes the full set of results for Bunds.
- The most market neutral scenarios, according to the model, are €20bn x 12mth, €30bn x 9mth and €40bn x 6mth.
- This is broadly consistent with the Reuters poll (taken 11-14 September) which suggested the consensus amongst economists was for €40bn (range €30- 50bn) over 6mths (range 3-12mths).
- Cross-checking the model output (based on policy signals) with the total size of APP extension shows a clear relationship (Figure 2). The model therefore assumes that there is less of a role for the ‘intensity’ of purchases.
- The market neutral size for APP upsizing appears to be around +€250bn
- The Citi house view is for an extension in the form of an ‘envelope’ (without specifying a monthly purchase rate) of €150bn (with upside risk of €210bn). That could lead to a near-term sell-off of around 15-20bp.
- The scenarios presented assume deliverability. But, the most dovish options undoubtedly would be more challenging to implement (see below) given scarcity constraints. In terms of likelihood, we would put less weight on these scenarios which skews the risk towards a bearish reaction on 26 October.
Is Every scenario as plausible?
The scenarios covered above include the four cited in the ECB sources story following the September ECB meeting (€20bn x 6mth, €20bn x 9mth, €40bn x 6mth, €40bn x 9mth).
The maximum implied extension of the four is +€360bn, which presumably is seen as achievable if it is on the table as an option. But, that would seem a stretch given existing scarcity constraints on Bunds and the extra challenge presented by reinvestments, which will grow to around €3.2bn/month for Bunds in Q4 2018.
The Citi scenarios include an even more dovish scenario which totals +€480bn. This really does seem to be pushing it. But, the ECB has proven to be resourceful in ‘using the full flexibility of programme’.
That said, there are various technical adjustments that could be deployed:
- The QE rules could change (again), such as allowing purchases below 1yr maturity.
- The composition of overall APP could also be altered. By that, we mean that the bulk of the reduction in purchases falls in PSPP. Note this would not mean an increase in the nominal purchase amount of CBPP3 and CSPP. Rather, that could be left unchanged allowing them to take up a greater share of a smaller overall APP (Figure 8 and Figure 9). There are challenges in these markets too, in terms of sourcing paper, especially with re-investments growing in CBPP, but new issuance may allow purchases to continue at around the same pace. The legal limits are also further away with the ECB estimated to hold around 40% of the covered bond universe vs a legal limit of 70%.
- Another option within PSPP is for some of the Bund share to be substituted for supras (ECB probably hold around 35-40% vs the limit of 50%).
- A greater portion could also be directed to agencies and regional debt.
- Deviations from the capital key could also be stretched further, though a complete disregard of the capital key is unlikely.
A long extension at €40bn/month, if deemed necessary from a macro perspective, is not easy to implement, but might be possible at a stretch. Perhaps this is why recent ECB speak (Praet, Smets) has been pointing to a more plausible lower-for-longer.
* * *
Citi's conclusion is that the market risk, as of this moment, appears tilted to the bearish side. And even as the ECB looks to their models for QE calibration to meet the inflation target, ultimately the decision will be a compromise.
- The hawks will argue that deflation risk has disappeared, the benefits of QE are diminishing, the growth backdrop has improved significantly, and that financial stability risks are on the rise.
- The doves will point out that core inflation is not yet sustainable, wages are still low against the backdrop of a persistent inflation undershoot, and the euro has strengthened 12% since April.
The compromise reached is likely to respect the ECB’s demonstrated mantra of slow and steady policy withdrawal (a.k.a confident, persistent, patient, and prudent). The aim is a taper without tantrum. Anything else would risk a premature tightening in financial conditions.
Citi's analysis shows that a super-dovish extension could drive 10yr Bund yields back down by 25bp to around 0.20% (close to the year-to-date lows). A super-hawkish extension could push 10yr Bund yields back up by close to 25bp to 0.70% (not seen since 2015).
Neither seems most probable. Emboldened hawks and implementation problems act against the dovish end of the spectrum while the desire to avoid a tantrum – especially if the euro rises again into the meeting – argues against the most hawkish scenario.
- If the middle ground seems most likely then the model indicates that a total extension worth around +€250bn or so is needed to avoid higher Bund yields. That can be achieved in various versions of a taper - €20bn x 12mth, €30bn x 9mth and €40bn x 6mth, with the shorter horizons and larger sizes leaving the scope for further extension on the table. This is where we think the consensus expectation sits.
- The risks are however skewed towards higher yields and bear-steepening, in Citi's view. That’s because issuer limit constraints, which are probably a pivotal factor in recent ECB communication of lower sizes for longer, suggest more forcefully that the next QE extension should be more clearly interpreted as the last in this cycle.
- It is not hard to see the ECB take up the longer 9mth extension at a net €20bn, as discussed already at the September meeting. This could also be presented as 9mth x gross €30-35bn (including re-investments). That would lift 10yr Bunds towards 0.60%
Citi's trade advice: "We continue to like steepeners, long USD rates versus EUR and are waiting for levels in 10yr Bunds near 0.65% and will then buy on expectation of a year-end squeeze as the ECB purchase operations have to chase higher duration Bund holdings."