The ECBC released 2012 covered bond statistics this week, the tenth time it has collected year-end data forming what is said to be the most complete single source of data about the market, with certain category changes having been made to boost the value of the data.

ECBC offices, Brussels

The statistics are compiled by the European Covered Bond Council (ECBC) statistics and data working group, which is chaired by Florian Eichert, senior covered bond analyst at Crédit Agricole, and capture outstanding volumes and annual gross supply.

Information about covered bonds issued out of Belgium, Iceland and Panama was added to the database for the first time as part of the 2012 data. The working group also took the 10th anniversary of the data exercise as an opportunity to make some changes to how the statistics are categorised so as to reflect the market as accurately as possible.

The breakdown by collateral type was changed to replace the “mixed” category with “other”, which sits alongside public sector, mortgage and ship collateral categories. This was done to reflect the increased heterogeneity of the covered bond market in recent years, with the range of collateral having widened to include aircraft mortgages, for example. Commerzbank has sold the first public benchmark covered bond backed by SME loans, but this took place this year (in February) and is therefore not included in the latest data set.

The working group decided that although non-traditional collateral represents only a small share of the market it cannot be ignored.

“The new breakdown will enable investors to track the evolution of covered bonds backed by these newer collateral types versus those still backed by traditional collateral,” said Eichert.

The working group also changed the deal size and placement type categories that had hitherto been used, by merging non-jumbos and jumbos, and privately and publicly placed deals into one new category. This is then broken down into public benchmark covered bonds above Eu1bn, public benchmarks between Eu500m and Eu1bn, and public issues for less than Eu500m, and private placements.

The working group also changed the definition of what counts as public and private, with the former [corrected] for the 2012 data and onwards being defined as a bond that has been listed and syndicated.

It did this for two reasons, one being that the definition of private versus public placements was not always consistently applied across countries, which meant that the figures were misleading. In addition, benchmark covered bonds between Eu500m and Eu1bn have become more common and accepted in recent years, thereby eroding much of the relevance of the jumbo concept, with bonds in the past also having been inaccurately allocated to this category despite not fulfilling the non-size related jumbo criteria, such as public syndication via at least three lead managers.

“We are aware that especially the size and placement type category changes are very substantial changes to how data is displayed,” said Eichert. “Backdating data to fit the new categories and maintain a consistent data history for previous years is a major challenge.”

Because not all countries were able to successfully carry out the recategorisation, on an aggregate level the ECBC is only able to present data for the new categorisation for 2012, with a full dataset going back to 2003 only for some countries.

“We still decided to make the change and accept this flaw because the advantage of painting a more realistic picture of market realities weighs more in our view,” said Eichert.

The ECBC statistics are based on publicly available data gleaned from sources such as issuer reports, rating agencies and supervisory authorities, among others. Because in some cases the information provided by these databases differs, the working group intensified its dialogue with issuers for the 2012 data collection exercise in a bid to provide the most accurate possible statistics and to resolve discrepancies across sources. The working group then updated the ECBC database, which means that there are some slight differences in the numbers for 2011 compared with what was published last year, according to Eichert.

“In our view, these adjustments are perfectly normal and we would rather adjust historic data to reflect a more realistic picture than mechanically hold on to data that was once published but proven incorrect,” he said.