The ECB May Run Out Of German Bonds To Buy In Just A Few Months: Here Is How It Can Fix This

With the monthly ECB governing council meeting and press conference due on Thursday morning, many market participants have turned their attention to a major problem we flagged nearly two years ago which has re-emerged as a result of the global collapse in yields: Mario Draghi may run out of eligible German debt to buy as soon as a few months from now, which has prompted speculation of what the ECB can do to address this issue. 

As UBS' Nishay Patel writes, market participants have questioned whether the ECB will be able to source enough German bonds to help meet the monthly target of €80bn of asset purchases due to the eligibility criteria applied on bond purchases. "Under the current rules, we estimate the Eurosystem may hit holding limits of German debt within the next 6 months. In all likelihood this could prompt the ECB to change the eligibility criteria on asset purchases."

This contrasts sharply with what Mario Draghi said last month when asked about whether the purchasing program was facing roadblocks, stressing that he saw “ample liquidity.” He said the program “continues to proceed smoothly” and has enough “flexibility” to deal with any limits that might come up.

That appears to no longer be the case.

As a reminder, the problem facing the ECB is that the securities that yield less than the ECB’s minus 0.4% deposit rate have grown to more than 60% based on a $1.13 trillion Bloomberg German bond index. That means they’re ineligible for the purchases (see figure 3). And as Bloomberg adds, "analysts from UBS Group AG and SEB AB are estimating the central bank may run out of German targets within six months, and as soon as August, unless the rules are broadened."

The full eligibility criteria is as follows:

"for a bond to be eligible for purchase in the PSPP, it must have a yield above the deposit rate (currently -0.40%) and the Eurosystem can own up to 33% of an individual bond issue (in some cases it is 25% and for supranational bonds the issue limit is 50%). On top of these criteria, bonds must have a minimum remaining maturity of 2yrs and a maximum remaining maturity of 30yrs and 364 days to qualify for purchase and an investment grade credit rating3. Purchases are distributed across jurisdictions in accordance with a country's capital key share. Germany has a (rebased) capital key share of 25.6%."


As German government bonds with a maturity of up to 7yrs have a yield trade below the ECB's deposit rate, a vast pool of German bonds are ineligible for purchase (Figure 3). The deposit rate constraint has resulted in the Eurosystem having to purchase longer-dated German bonds, yielding above the deposit rate (Figure 4). This is one of several reasons why core euro area government bonds yield below where fundamentals imply.


Should the ECB not change the eligibility criteria of its purchases it faces a purchase "precipice." Based on current yield levels UBS estimates that the ECB could hit the issuer limit for all German debt within the next 6 months (with the earliest time being 1-2 months due to the large amount of German debt that is currently ineligible to be purchased). Out of the €80bn/month of asset purchases, the Eurosystem tends to buy ~€19bn/month of German bonds (regional, agency and sovereign).

“They can only keep the current show on the road for a couple of months, and the reason is all the self-imposed limits,” adds Peter Schaffrik, head of European rates strategy at Royal Bank of Canada.

Here is UBS's calculation:

As the ECB does not provide a breakdown of bond holdings, estimates of the precise time when holding limits could be reached can vary widely, as one cannot determine how much of the current ineligible pool of German bonds were purchased when they were previously eligible and whether the total amount of PSPP purchases will account for a similar portion of all monthly asset purchases (APP) as they have done so far. Furthermore, if yields were to rise from current levels, resulting in ineligible shorter-dated bonds becoming eligible again, the time taken to reach the holding limit would increase.


Figure 5 shows that if the ECB were to abandon the deposit rate floor (while keeping other parameters unchanged, such as the 2-31yr maturity restriction and holding limits), purchases of German bonds could run for another 11months or so.



If all of the German bonds that the Eurosystem currently owns are within the current eligible pool of assets, the holding limit would reached very soon. This is highly unlikely of course as a number of bonds that now trade below the deposit rate are likely to have been purchased when they had a yield above the deposit rate in the past.

UBS' assessment of this problem:

The ECB may not immediately announce that the holding limit of a country has been reached as national central banks (NCB) can carry out substitute purchases of "marketable debt securities issued by international organisations and multilateral development banks", as well as non-financial corporate debt located in their jurisdiction under exceptional circumstances. However, these are subjected to holding limits as purchases of international or supranational institutions located in the euro area are subsumed under the 10% allocation for these securities in the PSPP. As a result, the ECB faces some additional pressure to amend the rules of the asset purchase programme

Making the problem worse, Bloomberg notes that reaching the precipice would affect a broad range of investors, because a decision by the ECB to open up new groups of bonds for its quantitative-easing program may support their prices even more, helping extend this year’s rally.

In other words, the ECB may be damned if it does, and damned if it doesn't. For now, however, it will most likely have to pick one option from those presented below in how to expand the universe of QE-eligible bonds.

The full Draghi "menu" courtesy of UBS:

(1) Scrap the deposit rate floor


If the ECB were to abandon this constraint, the eligible pool of German government bonds would increase dramatically (Figure 7) and asset purchases would be able to continue until the ECB held 33% of all 2-31yr German debt. Should this option be brought to the table there could be some push-back from some members within the governing council as purchases of bonds with a yield less than the deposit rate would mean that the Eurosystem would be crystallising a loss at the time of purchases. However, it could be argued that the ECB could still make a loss on purchases of bonds with a yield above the deposit rate at the time of purchase. Overall, we think this option is possible and could potentially face less resistance than some of the other options such as deviating away from the capital key, as it could be argued that because a large amount of sovereign debt trades above the deposit rate a loss may not be crystalised on an aggregate level.


(2) Cut the deposit rate solely to increase the pool of eligible bonds


We would be extremely surprised if the ECB were to cut the deposit rate solely to increase the amount of eligible bonds for purchases. Additional cuts to the deposit rate would compress the net interest margins for banks further and increase the  chance of negative rates being passed onto retail customers; ultimately resulting in a net tightening of credit conditions. Instead a more likely option would be to "scrap the deposit rate" floor (an option mentioned earlier).


(3) Increase the 33% issue limit on bonds that do not contain a collective action clause (CAC)


Increasing (or abandoning) the 33% limit on non-CAC bonds would result in an increase in the pool of eligible bonds and ensure that purchases can take place for slightly longer, particularly if the 33% limit is no longer applied to regional and agency debt. We estimate the amount of sovereign bonds with and without a model CAC in Figure 8 below. It must be noted that the ECB has repeatedly stated that the asset purchase programme is intended “to be market-neutral”, and, “the Eurosystem wants to create as little distortion as possible”. If the 33% issue limit is increased for non-CAC bonds, certain segments in government bond markets could see distortions between bonds containing a CAC and ones that do not. For example, a number of countries have non-CAC bonds with a similar maturity to bonds containing a model CAC.


(4) Abandon the capital key allocation for PSPP purchases


In order to reduce the time taken for the ECB to potentially hit the issuer limit for German sovereign debt (where a significant portion trades below the deposit rate) a commonly cited option by market participants would be for the ECB to deviate from the capital key share. Moving away from the capital key would open up ways for less German debt to be purchased relative to the capital key share. However, a deviation away from capital key shares is likely to be politically sensitive, particularly in Germany, if the Eurosystem were to increase the amount of purchases in a country that has not been complying with deficit targeting measures and is at risk of being fined. We see this option of changing the allocation of purchases away from the capital key share as highly unlikely. Instead, the ECB could make the rules on substitute purchases more flexible (which we explore below).


(5) Expanding the 2-31yr maturity restriction for sovereign, regional and agency debt


The ECB could easily widen (or remove) the maturity restrictions on sovereign, agency and regional bond debt (from the current range of 2yrs to 30yr 364days). Should this option materialise, the eligible pool of assets would increase slightly. For Germany, the impact on eligible sovereign debt would be limited as bonds with a maturity of zero to 2yrs already have a yield below the deposit rate (making them ineligible for purchases) and the longest maturity bond (Aug-46) is already eligible for purchases. However, for countries such as France and Belgium, there are several bonds which have a maturity greater than 31 years and have a yield above the deposit rate


(6) Expanding the range of assets to include bank bonds


Since the asset purchase programme began in March 2015, the ECB has expanded the range of eligible assets for purchases on several occasions. Purchases of regional bonds, local government and investment grade corporate bonds (nonbank) began this year, and the list of eligible agency bonds has been increased regularly. Expanding the range of assets that can be purchased would ensure that the ECB will still be able to expand its balance sheet without resulting in severe market distortion. For the time being we do not expect the ECB to include new types of assets, such as bank bonds, as part of the asset purchase programme, especially as purchases of non-bank corporate bond purchases only began in June 2016. As regards bank bonds specifically, we would expect resistance from the hawkish camp of the Governing Council who will likely be concerned about the risk of a Eurozone-wide mutualisation of bank risks via the ECB balance sheet.


(7) Making the rules on substitute purchases more flexible


Under the current rules, if a national central bank (NCB) cannot make purchases of central government and recognised agency debt to "implement the relevant NCB's share of purchases through the end of the APP5", it can carry out substitute purchases of "marketable debt securities issued by international organisations and multilateral development banks", as well as non-financial corporate debt located in their jurisdiction under exceptional circumstances. Purchases of international or supranational institutions located in the euro area are subsumed under the 10% allocation for these securities in the PSPP. Therefore, if the NCB has exhausted all of these measures (including non-financial corporate debt) and the relevant share of purchases for a NCB still cannot be met the ECB may face additional pressure to amend the  rules on the asset purchase programme. Rather than formally changing the allocation of purchases away from the implied capital key share, the ECB could permit NCBs that are unable to make purchases in their home country to purchase eligible assets in other jurisdictions. This could be particularly beneficial for countries with a limited amount of sovereign, agency and regional debt, such as Estonia, Slovenia, Slovakia and Lithuania. However, it cannot be taken for granted that NCBs would choose to make substitute purchases of debt in other jurisdictions.

* * *

In less than 24 hours Draghi may pick one or more of the options presented above. Or maybe not: SEB’s Marius Daheim says it’s unlikely the institution will make major tweaks to its program this month and will likely wait until September when it will also present growth and inflation projections. The problem is that should the ECB choose to wait that long, the market could simply frontrun the ECB's decision, and push yields even further into negative territory, in the process undoing what little pick up in long-dated yields we have seen in recent months, and further forcing Draghi's hands to engage in even more of the above menu selections.

Of course, there is a simple solution to the problem: all the ECB needs to do is force European government to open the debt spigots and issue much more debt. Should there be a surge in supply, the ECB will have no problem with matching this new issuance under its existing €80BN/month mandate. Of course, that is also somewhat synonymous with helicopter money, and at least as of this moment, we doubt it will be glowingly endorsed in Germany, one of the last bastions of fiscal responsibility left in the entire world.