A new analysis by Goldman looks at the post-ECB QE world and finds finds something troubling: having purchased virtually no Italian bonds in the past 2 years, the private sector will have to aggressively step up in 2019, and absorb 80% of gross Italian sovereign bond issuance.
First, a quick reminder of where we currently stand after last week's announcement by the world's biggest hedge fund that it will be further tapering and, soon, ending it purchases of newly-issued European bonds.
The ECB communicated last Thursday that, subject to incoming data, the Asset Purchase Programme (APP) will be extended from October to December at a pace of €15bn per month, and that net new purchases will then be wound down to zero.
Initiated in March 2015, the APP will terminate after three years and nine months, having accumulated a total portfolio of approximately €2.6trn in assets, mostly government securities (~€2.1trn – 80% of total QE holdings) and corporate/covered bonds (~€500bn).
At the end of this year, the ECB’s balance sheet will reach a total of €4.75trn, or more than 40% of the area’s GDP – a higher level than the Fed’s at the end of its QE. We calculate that the central bank will own around 27% of the stock of PSPP eligible and marketable government bonds issued by EMU member states, and just under half of the float of supranational bonds (e.g., ESM, EFSF) used to fund conditional rescue packages for the troubled peripheral countries during the crisis.
That said, it's not like after January 1, 2019 the ECB will simply go cold turkey, or as Goldman says "this is not
adieu for the APP but rather au revoir" for two reasons:
- First, the QE purchases will continue through reinvestments for an extended period of time. According to Goldman calculations, "the central bank will roll-over €15-25bn a month from 2019 onwards, as shown in Exhibit 2. These purchases will constitute the pace of monthly flows once the net purchases are terminated in December."
- Second, the APP will remain a policy tool of the ECB, which could be reactivated should the need arise. To Goldman, "this suggests that, from being an ‘unconventional’ instrument, asset purchases could be used in the future as a more ‘standard’ policy tool to pursue the ECB’s inflation objective."
There are some other complications, including that the ECB is largely tapped out when it comes to eligible German Bund purchases, which are now at the 33% limit of German bond holdings.
The ECB could also need to exploit some additional flexibility in conducting net purchases from now until December. The 33% issue/issuer limit on the ECB’s holdings, and the consequential scarcity of QE eligible securities in Germany and other smaller ‘core’ markets, could still require small adjustments to purchases of government bonds via substitution towards supranational agencies, or a further expansion of the pool of suitable bonds to buy under the public sector programme.
Still while for some eurozone states the ECB's reinvestment of bonds could be a sufficient "buffer" source of duration demand, for others it could be the source of the next crisis.
In Germany, for example, things couldn't be better in 2019, as not only is gross issuance of long-term bonds expected to decline, but - since it already owns so many Bunds - the central bank will reinvest around €50-60bn Bunds in its portfolio, continuing to absorb around 47% of the gross supply via purchases in the secondary market also next year.
The same can not be said for Italy, however, which is about to suffer the "perfect storm." Among major European markets, even though gross issuance of medium-/long-term bonds could diminish slightly next year - although it probably won't with the new government's plans to boost fiscal stimulus, and as the country with the third-largest ECB capital key, Italy will still see the lowest share of issuance absorbed by the ECB in 2019, with the private (or foreign official) sector required to absorb more than 80% of newly issued long-term government securities.
One look at the table above and it becomes clear that for the private sector to purchase €182BN in Italian bonds next year, one thing will have to happen: yields will have to go higher: far, far higher to make the paper attractive to investors who in addition to the loss of ECB as a backstop buyer, will have to soak up more Italian bonds than they have in years.
Recall this Citi chart which shows that in the past 2 years, the only net buyer of Italian debt was the ECB. Well, not anymore.
To be sure, the phase out of QE will impact other nations too: according to Goldman, the cumulative purchases of government bonds by the ECB have reduced German 10-year Bunds by as much as 50-80bp relative to where they would otherwise be. And naturally, despite the recent increase in EMU rates volatility, the effect on their Italian and Spanish counterparts is somewhat larger also on account of a compression in systemic risk embedded in these bonds. Goldman's conclusion:
All else equal, such effects should remain in place for some time after the end of QE purchases, until the market starts to discount a forthcoming reduction in the size of the ECB portfolio. The ‘flow’ effect of QE purchases, at the current QE pace close to -10bp for ‘core’ rates and around -15bp for ‘peripherals’ according to our estimates, will instead diminish from October, as a consequence of the forthcoming tapering. On a medium-term horizon, such a reduction in QE purchases constitutes, in our view, an additional source of uncertainty for peripheral rates.
In other words, the Italian "perfect storm" is about to get very windy, starting some time in October, and only get worse from there.