Stay Connected!
ETF News Daily - Europe
ETF News Weekly - Europe
Sign up to's newsletters.
Print this article
Deconstructing Eurozone Bond ETFs
By Paul Amery | 21 July, 2011

In recent weeks the strains in the government debt markets of the Eurozone’s member countries have escalated to a new and more dangerous level. Concerns over possible sovereign defaults have spread from those peripheral countries—Greece, Ireland and Portugal—which have already required external financial support, to larger, core members of Europe’s single currency.

Between early April and early July, five-year credit default swap (CDS) spreads on Italian and Spanish government debt jumped by 2 percent a year, nearly trebling in Italy’s case. It now costs nearly 3 percent a year to insure against the default of Italy for five years, and around 3.5 percent a year to insure against Spanish default. Even France, previously relatively untouched by Eurozone contagion, saw a doubling of its CDS spread over the same period, with default insurance now costing over 1 percent per annum.

Investors in exchange-traded funds that track indices of European sovereign debt face one obvious concern, relating to the composition of their benchmarks. An index that allocates higher percentage weights to the debt of the Eurozone’s perceived “problem” countries is undoubtedly riskier—and will pay investors a higher interest yield to compensate for that risk.

However, there’s a secondary and arguably no less important type of risk facing investors in European ETFs, particularly as European banking systems come under pressure. That’s the question of ETFs’ structure and the counterparty and collateral risks which they incur.

To illustrate these two levels of risk, in this article we’ve taken a sample of seven ETFs that track very similar benchmarks. All the ETFs aim to reproduce the returns of short-maturity European government bond indices (i.e., containing bonds with over one, but less than three years to maturity). The survey includes all European ETFs of this type that have over €100 million in assets. After describing the funds we look, in turn, at the two types of risk that we’ve just described: index-related and structural.

Short-Maturity Eurozone Government Bond ETFs

ETF Name

Fund Ticker

AUM (€m)

Replication Type

(% p.a.)

iShares Barclays Capital Euro Government Bond 1-3





Lyxor ETF EuroMTS 1-3Y





db x-trackers II iBoxx Euro Liquid Sovereigns Eurozone 1-3





Amundi ETF Government Bond EuroMTS Broad 1-3Y





Comstage ETF iBoxx Euro Liquid Sovereigns Diversified 1-3





CS ETF (IE) on iBoxx Euro Government 1-3





ETFlab iBoxx Euro Liquid Sovereigns Diversified 1-3





Index-Related Risk Exposures

The investment exposures of these seven short-maturity Eurozone government bond ETFs are determined by their underlying indices’ country weightings. These are illustrated in the bar chart below.



For a larger view, please click on the image above.

The German, French and Italian bond markets constitute the most significant weightings in these indices, collectively adding up to two-thirds or more of the benchmarks. Spain, Belgium and the Netherlands also command sizeable positions, while weightings in other countries’ government bond markets—including those of Ireland, Portugal and Greece—are typically in the low single figures.

The single most noticeable difference between the indices used by the seven ETFs in the survey is that between the concentrated nature of the Barclays Capital index tracked by iShares’ ETF, IBGS (which allocates a 40 percent-plus country weight to Italy, for example) and the more diversified nature of the iBoxx and EuroMTS indices that are used by the other six ETFs.

There are other, more subtle differences too. Markit’s iBoxx indices set a minimum ratings requirement for membership, and so Portugal’s recent downgrade to below investment grade means that the country (and its bonds) will be expelled from the iBoxx indices at the end of this month. EuroMTS, by contrast, sets no such ratings requirement, explaining the continuing presence of Greece in its 1-3 year benchmarks (albeit with a small, 2 percent weighting).



Counterparty and Collateral Exposures

On Friday July 15, asked the managers of all seven ETFs in the survey to provide information on the counterparty and collateral exposures that had been incurred by their funds during the week of July 11-15.

We asked managers of those ETFs that use physical replication (i.e., funds that own the bonds in the indices they track, a sample of them or an optimised basket of similar bonds) to tell us the percentage of their funds’ bonds that were on loan on the individual days of this week, provide details of the bonds lent and the collateral received in return, and give the names of the counterparties used.

We asked managers of synthetic ETFs (which use derivatives issued by a third party to track their indices, holding a basket of potentially unrelated assets to back this promise) to send us the daily breakdown of assets held by their funds, the net (uncollateralised) swap exposure via derivatives contracts and the names of counterparties used.

Physical ETFs

Among the three issuers of physically replicated ETFs included in our survey, Credit Suisse said it does not lend out the assets of its European government bond fund, CSBGE3 (though it does lend securities from its Swiss government bond trackers). But both iShares and ETFlab said they do lend the bonds held by their short-maturity European government bond ETFs, IBGS and ETFES13.

Neither firm, however, was willing to disclose specific details of its fund’s lending operations for the last week, even though iShares recently said it aims to provide daily disclosure of lending collateral in its European ETFs. ETFlab says it plans a similar move.

iShares referred us to its latest quarterly report on securities lending, in which the firm said that the average percentage of the assets lent out by its short-maturity Eurozone government bond fund, IBGS, during the year to end-March 2011, was 23 percent. Lending activities produced a small net yield for the fund of 2.3 basis points for the year.

iShares’s lending agent, its parent company, BlackRock, took a 50 percent cut of its funds’ gross lending revenues until November 2010, then reduced its percentage share of earnings to 40 percent. So one can assume that gross lending revenues earned by IBGS in the year to March were around 4-4.5 basis points. Given the percentage out on loan for the fund during the year, the average lending operation must therefore have produced an annualised return of 17-20 basis points.

iShares’ lending report details the collateral held on behalf of all its Dublin-based ETFs as at March 31 this year, though the firm doesn’t give a breakdown of collateral held by individual funds. Nor does the report specify the amounts on loan to individual counterparties at that date, although iShares had twelve approved borrowers at the end of March, all investment banks.

The firm’s lending guidelines specify that, if government bonds are lent out, collateral may be taken in the form of cash, government bonds, equities and certificates of deposit, subject to varying levels of “haircut”. The firm says it determines the acceptability of collateral after an analysis of liquidity, volatility, and correlation with the loaned securities, which presumably means that it’s not usual to take equities as collateral for loaned bonds.

In general, iShares’ securities lending activities produce higher net returns to the firm’s equity ETFs than to its bond ETFs, although the actual levels of lending in bond ETFs may be significant. On average, the iShares Barclays Capital Euro Government Bond 3-5 and 7-10 year bond ETFs had 55 percent and 72 percent of their assets on loan, respectively, during the year to March, for example.

ETFlab says it undertakes two types of loans in its short-maturity European government bond fund: on a bilateral basis to a maximum extent of 10 percent of the value of the fund with a single counterparty, against collateral; and via Clearstream Banking, a lending system approved by the German regulator and which allows up to 100 percent of fund’s value to be lent on a secured basis. The firm gave no further detail of its lending policies.

Synthetic ETFs - Coverage

In general, providers of synthetic ETFs were more forthcoming than those of physically backed funds in response to our request for information.

All the synthetic ETF providers from whom we requested details of their funds’ underlying holdings for the five days of last week—db x-trackers, Lyxor, Amundi and Comstage—provided the necessary details. Of the four, however, only Deutsche Bank publishes its funds’ full composition on its website daily. The other issuers in the survey publish details of fund assets less frequently, though stated to that they can disclose this information to clients on request.

The four synthetic funds in the survey all used a single derivatives counterparty: db x-trackers held a swap with its firm’s parent bank, Deutsche Bank; Lyxor and Amundi both used Société Générale (which owns 100 percent and 25 percent, respectively, of these two ETF issuers); and Comstage also used its parent bank, Commerzbank.

Europe’s UCITS rules cap exposure to derivatives counterparties at 10 percent of a fund’s net asset value. During last week and in the funds we surveyed, db x-trackers adopted the most conservative position in terms of the absolute level of assets covering its ETF’s swap, maintaining an substitute basket that slightly exceeded, on average, the fund’s net asset value.

Lyxor’s fund MTA was next, averaging a net counterparty exposure under its swap contract of just under 1.5 percent to Société Générale. The ETFs operated by Comstage and Amundi maintained an average swap counterparty exposure of just under, and just over 7 percent of fund NAV, respectively, for the week.

Comstage adds an additional level of protection to its ETFs, slightly overcollateralising the outstanding mark-to-market value of its swap contracts with German government bonds.

Synthetic ETFs—Asset Make-Up

The recent concerns that have been expressed by regulators about potential structural risks in synthetic ETFs have focused less on the absolute levels of coverage that are mandated by the UCITS rules than on the nature of the assets being held to back the swaps held by such funds.

Specifically, regulators have pointed out that there’s an economic incentive for bank-backed issuers of synthetic ETFs to use these funds to finance the parent banks’ holdings of less liquid assets.

The quality of the assets put into the substitute basket by the ETF issuer is therefore important, both in absolute terms and relative to the quality of the assets in the index being tracked. Other things being equal, and assuming that ratings are a true reflection of risk, an ETF tracking an index of A-rated bonds, for example, and with a substitute basket of AAA-rated bonds, would incur a cost to the provider, since it's more expensive to source a portfolio of AAA-rated than A-rated securities. By contrast, the same ETF with a substitute basket of bonds rated lower than the index securities—say BB or BBB—would provide the issuer with an implicit subsidy. Such operations are broadly referred to as "collateral upgrade" trades—though it seems it's usually the ETF issuer that's receiving the upgrade, rather than the fund's investors.

So what did the four Eurozone bond ETFs in our survey actually own during the last week?

There were some significant differences between the funds surveyed. db x-trackers’ X13E was the only ETF in the sample to hold only government bonds in its substitute basket. The Lyxor and Amundi ETFs, MTA and C13, held a mixture of government, agency, municipal and covered bonds. Comstage’s CBOXES1, meanwhile, owned European equities in its substitute basket (in fact, Comstage uses the same substitute basket make-up for all its ETFs).

For the three ETFs owning bonds as substitute assets—those of db x-trackers, Lyxor and Amundi—there were also significant variations from the index being tracked, both in terms of country exposure and in the maturities of the bonds held. In some cases the maturities of basket bonds were much longer than the short, 1-3 year maturity of the index bonds.

For its fund X13E, db x-trackers said that it aims to hold a similar composition of bonds in the substitute basket to that represented by the index holdings. The substitute basket’s country weightings during the last week varied slightly from those in the index, though only to the extent of a few percentage points in each case.

Lyxor said that it accepts as assets for its fund’s substitute basket both government-backed bonds (those issued by a national or regional government, municipality, or government agency) and covered bonds (bonds issued by firms, often banks, that are collateralised by mortgages and/or public sector assets), subject to a minimum ratings requirement of A- in both cases.

The firm says it sets a minimum weighting target of 90 percent for the first category, although during the week under review it appears that the covered bonds owned by MTA, which included issues from German, Spanish, Dutch and Austrian banks, represented slightly more than 10 percent of MTA’s assets in aggregate. The bonds owned by MTA also varied in maturity, with one extending to seventeen years, well beyond the maturity profile of the index. Overall, MTA’s bond holdings were heavily weighted towards agency and municipal debt issues, which are generally regarded as less liquid than mainstream sovereign bonds.

During the five-day period surveyed, Amundi’s short-maturity bond tracker, C13, owned more sovereign (i.e. central government) debt issues than Lyxor’s MTA. However, the Amundi fund also owned some agency bonds and bank-issued covered bonds. The most notable mismatch between C13’s holdings and those of the index, however, came from a heavy aggregate fund weighting in zero coupon bonds, often of long maturity. The fund also had a significant overweight position in Italian government debt when compared with the index.

The issuer sets a minimum ratings requirement of BBB- for asset eligibility in its eurozone bond ETF, said Valérie Baudson, Amundi’s managing director, while also monitoring the fund holdings’ correlation with the index, their diversification and liquidity, she added.

Comstage’s decision to hold European equities as its bond trackers’ assets is motivated by considerations of tax efficiency for the ETF issuer’s largely German client base, explained Arne Scheel, who is responsible for sales to institutional clients at the firm. Holding equities rather than bonds as fund assets allows investors to delay recognition of an income-related tax liability until the fund is sold, Scheel said. Scheel acknowledged that holding a basket of equities to track a bond index presents correlation risks, but argued that collateralising the Comstage fund’s swap exposure via additional holdings of German government bonds is a significant mitigating factor.

Does All This Matter?

What an ETF actually holds, rather than what it tracks, may seem of secondary importance to most owners of these funds. However, in extremis, it’s the actual assets of a swap-based fund which make sure that investors can be repaid their money. Likewise, if fund holdings are lent out in a physically backed ETF, to whom assets have been lent and what collateral has been taken in return could be of critical importance.

This survey of a small subset of Eurozone government bond ETFs and our analysis of the key mismatches between individual funds and the indices they track is necessarily incomplete. A full investigation of the differences between index holdings and fund assets (and of their implications for fund risk exposures) would require a detailed examination by specialists in securities finance. Such analyses seem increasingly necessary, and not just for this category of fund.

As far as trackers of Europe’s government debt markets are concerned, investors should undoubtedly pay attention both to the structural, as well as the investment exposures of their funds. With sovereign and banking sector risks now closely intertwined, it would seem foolish not to do so.

Comment Using:

blog comments powered by Disqus

Europe Blog

Larry Swedroe

The market “correction” of August provided investors with an opportunity to learn some important lessons