Yesterday, in the "ECB Matrix" we showed [19]what the probable outcomes out of today's "most anticipated ever" ECB meeting could be, alongside some suggestions from DB on what the best "delivery" and "disappointment" trades could be. For those who missed it, here it is again.
[20]
Disappointment trades
- On this simplistic risk-reward perspective iTraxx Main spread wideners, long euro/usd and Bund ASW wideners offer the best risk-reward for ECB disappointment trades.
Delivery trades
- The conclusion regarding ECB delivery is a bit more complicated. If
we ignore 10Y Bunds which could rally if the ECB announces a broad based
QE including government bonds we are left with 2Y Euribor-OIS
spread tighteners, long 2Y1Y inflation swaps (which might remain under
pressure due to the decline in oil prices) and paid positions on the
long-end of the EUR curve.
[21]
But since nobody knows better what the former Goldman Sachs head of the ECB will do than Goldman, here is Goldman's preview of today's ECB announcement:
Reflections on ECB sovereign QE and Asset Prices
- The ECB’s press conference is the main event today. We expect no changes in policy.
- Our baseline case calls for ongoing ‘credit easing’ measures through private assets.
- Over time these should support the recovery and a rebuild of inflation expectations.
- Consistently, we forecast a modest but sequential pick-up in GDP growth after the Q3 lull.
- We see a 1-in-3 chance that the ECB will launch sovereign QE ...
- ... reflecting the probability of a downside scenario for economic activity.
- We would expect sovereign QE to operate mainly through expectations and portfolio rebalancing channels.
- Tighter long-dated intra-EMU spreads, and a weaker EUR, would be the most direct ways to capitalize on this.
- While the latter is already part of our market strategy, low growth, financial fragmentation and political uncertainties have been headwinds to spread compression.
- These factors have kept us neutral on spread direction since the Summer.
- Looking into 2015, the case for a flattening of EMU peripheral yield curves deserves attention given the high premium.
1. Overview
The key event today is the ECB’s press conference. It comes at a time when media reports have highlighted tensions within and around the Governing Council's policy stance. One of the points of friction appears to be related to the public suggestions by part of the central bank’s leadership that EMU countries that have room to expand fiscal policy should use it.
Although the ECB’s comments on fiscal matters may have interfered with delicate political balances between EMU Member States, we think that a relaxation of the structural fiscal stance is appropriate with monetary policy at the zero bound. And it is what our European Economics team has built into our baseline growth forecasts, based on fiscal plans announced by the main EMU members.
Back to today’s policy meeting, we do not expect any additional easing to be announced in addition to the various measures adopted between June and September. We expect Mr Draghi’s remarks to be focused on the Comprehensive Assessment of Euro area banks, and on the fact that the decline in oil prices is lowering headline inflation in most advanced economies.
Our base case has long been that ECB purchases would be concentrated in Asset Backed Securities and, possibly, be extended to other eligible assets, such as supranational bonds, and private bonds. Consistently, we still assign only a 1-in-3 chance to large-scale purchases of EMU area government bonds. Such subjective probability reflects the odds of the Euro area economy not responding positively to the measures in support of private credit already administered, and hence requiring a heavier and quicker boost. In the event, we think the ECB would be willing to bear the high political ‘fixed cost’ of engaging in sovereign QE in order to fulfil its inflation mandate.
In today’s Daily, building on questions we frequently receive from our clients, we sketch out some of the market issues surrounding sovereign bond purchases in the Euro area, should the ECB decide to take a turn in this direction against our central case. We comment on four aspects in particular: 1) is sovereign QE priced, 2) if implemented how it would look, 3) what would the transmission channels be, and 4) what could be the broad asset price response. We would expect sovereign QE to operate mainly through expectations and portfolio rebalancing channels. Tighter long-dated intra-EMU spreads, and a weaker EUR, would be the most direct ways to capitalize on this, alongside a broader easing of financial conditions.
A lower EUR/US$ is already part of our market strategy, as we believe it will be also helped by higher Dollar real rates. Looking into 2015, the case for a bull-flattening of peripheral EMU spread curves would be supported by our baseline economic case, and in the event of sovereign QE. However, current low growth, residual financial fragmentation across the Euro area and political uncertainties still weigh on peripheral EMU bond markets. We recognize that we are at an important inflection point both in the economic cycle and the fiscal/monetary policy stance, and we reiterate our neutral recommendation on spread direction.
2. Are ECB sovereign purchases already priced in?
The term structure of ECB policy rates, as captured by the EONIA curve, is very low and flat. The current 5-year EONIA yield trades at 11bp, while the 5-year rate in 5-years time stands at 130 - well below the expected rate of Euro area inflation at that horizon, currently at 1.8% (this implies that 5-year 5-year forward EUR real rates are negative to the tune of 50bp). On our analysis, such depressed level of future yields largely reflects the ECB’s policy of negative overnight deposit rates, and ‘forward guidance’. The latter is reinforced by the possibility afforded to Euro area banks growing their loan books to lock in funding rates at 15bp out to 4-years, which, in turn, has contributed to the flattening of the yield curve. Therefore, low rates and a flat curve provide per se no conclusive evidence that large-scale purchases are ‘in the price’.
In peripheral EMU bond markets, by contrast, the term structure of interest rates is very steep, as we documented in a recent Global Markets Daily, which compares sovereign spreads in Italy and Germany. Using updated figures, the differential between Germany on the one side and the average of Italy and Spain on the other goes from 65bp at the 2-year maturity to 235bp at the 10-year in 10-years' time horizon. Cross-country spreads at these longer maturities have all but widened since April, when the ECB introduced its policy statement that ‘unconventional instruments’ (i.e., QE) could be deployed. This would also suggest that government bond purchases are not being fully discounted.
Persistently high long-dated intra-EMU spreads reflect a combination of factors, including (i) credit risk, as reflected in the large and growing (according to IMF projections) differential in public debt ratios between Germany on the one side and Italy Spain on the other, and (ii) a fragmentation in Euro area financial markets, where more than two-thirds of all the public debt of Italy and Spain is now held by domestics, with the ‘home bias’ on longer-maturity bonds anecdotally higher.
3. How would sovereign QE be carried out?
The stock of Euro area sovereign bonds is around EUR7trn. The US, UK and BoJ have bought the equivalent of 20% of their respective stocks of public debt over an average of 3-4 years. Crudely applying this ratio to the Euro area would lead to purchases of EUR400-500bn per annum. Should the ECB decide to undertake sovereign QE, our European Economists would expect it to act boldy, exploiting announcement effects.
Given how low front-end rates already are (the 1-year rate in 2-years' time in Italy trades at 1.2%, and zero in Germany), the maximum impact of sovereign purchases along this dimension would be achieved through the removal of duration risk from private hands - a key aspect of the Fed’s QE3 and the BoJ’s current policy. The international experience suggests that central banks tend not to restrict purchases up to a certain maturity in order to avoid distorting the sovereign yield curve (the Bank of England did so at its first attempt at QE in 2009, when it announced it would purchase bonds with residual maturity between 5- and 25-years; but soon switched to all bonds with maturity beyond 3-years).
As for the country allocation of the purchases, the criterion would likely be the ECB’s capital key. This is not without issues, however. How to deal with countries that are still under a stabilization program would be a point of contention. A good chunk of the debt of countries that have recently emerged from a program (Ireland, Portugal) is in the form of non-tradable official sector loans. Other EMU countries have relatively small bond markets, given the size of their economy. Finland provides a fitting example: with EUR500bn-worth of total purchases, the ECB would completely absorb the estimated gross issuance of Finnish bonds for an entire year. Anomalies in the relative pricing of EMU sovereigns along cross-country lines would be likely.
4. What would the transmission channels be?
Empirical evidence from other experiences of sovereign QE suggests that the impact of bond purchases tends to be front-loaded, when the size of the program, or the objective to which this is tied (e.g., ‘until inflation returns to target’, as in Japan) is announced, rather than when the purchases are actually carried out. A rapid expansion of the ECB’s balance sheet would lead to a faster decline in the EUR, particularly against the Dollar, helping reflate the economy.
Countries that adopted sovereign QE in recent years ran primary (i.e., net of interest payments) budgetary deficits. The average of the US, UK and Japan for the period 2009-13 was around 7.5% of GDP. The monetary expansion therefore interplayed with the looser fiscal policy stance, lowering funding costs and leaving more room for banks and other domestic financial institutions to accumulate other assets (including cash balances) rather than government bonds (see, for example, the small increase in government bonds relative to total assets held by US and UK banks in the wake of the Great Recession relative to the historical norms post downturns).
By contrast, in the Euro area, two of the main beneficiaries of sovereign QE – Germany and Italy – are both running primary surpluses. Unless the fiscal stance is relaxed much more than we anticipate (which could be conceivable in a scenario of deep supply-side reforms, but unlikely), the transmission to asset prices and the real economy would come through a combination of expectation shifts and portfolio rebalancing effects.
As Huw Pill and Dirk Schumacher have argued, sovereign QE would entail risk transfers among taxpayers of the different EMU constituents through the ECB’s balance sheet. Relative to the experience of other countries, this ‘risk sharing’ characteristic of open-market operations would magnify the effect of the policy. Relatedly, government bond QE would likely lead to upgrades in sovereign ratings.
ECB purchases of government bonds would also increase the fungibility of longer-dated securities across the major EMU issuers, and encourage a broadening of demand. Using Italy and Spain as an example, part of the stock of public debt currently held by domestic financial institutions would probably be sold to the ECB. Concomitantly, foreign ownership – estimated at around 30% of the stock – would expand, and its composition would likely see an increased participation of Asian investors.
5. What would the main implications for macro asset prices be?
Should the ECB engage in sovereign QE, the main macro asset price response would likely include:
- A tightening in EMU spreads, with a large part of the adjustment coming from a decline in peripheral rates. For reference, 10-year yields 10-years forward in Italy and Spain, currently trading at 515bp and 415bp respectively, would likely experience the largest drop relative to Germany (230bp). Estimating by how much rests on a number of key assumptions (relative size and distribution of the purchases, the response of debt management offices, intervening changes in the domestic policy stance).
- A slight steepening in the nominal term structure of EONIA, also reflecting an increase the market price of inflation. We would envisage a gradual rebuild of term premium in the German bond market.
- A decline in the Euro, particularly against the USD, which continues to trade slightly expensive relative to long-run ‘fair value’ estimates.
* * *
Perhaps focus on those "disappointment" trades, then?
