The ESEF die is cast…but let’s not forget that investors do read annual reports

Our response to the FCA consultation on ESEF (closed on 1 November)

I was chatting to my lovely AstraZeneca colleagues a few weeks ago to pick their brains for our book, when Martin Blaxall (usually so kind) cruelly burst my Brexit bubble. ‘The best thing about Brexit,’ I said, ‘is that if we do leave on 31 October – or by the end of the year as the FCA says – then we won’t have to do that pesky ESEF! Maybe I won’t bother putting in a response to the consultation…’

‘Oh yes we will,’ said Martin (or words to that effect), ‘you must have misread it. We do still have to do it, because [and I was so disappointed I can’t remember now what the reason was], and besides, all those like us who have a listing in Europe have to do it anyway.’

All this, of course, is academic now, because since then political affairs have continued along their rollercoaster ride, and we are indeed remaining until 31 January 2020 (probably).

So, I thought, I can hardly complain about how the regulations are implemented if I don’t offer my say when I have the chance. So last Friday I squeaked our response in just before the deadline.

My main message (to save you reading the whole thing – although if you want to, please see below) is twofold: 1) let’s make sure the laudable aim of having machine readable annual reports doesn’t have the unintended consequence of compromising their readability by human beings*; and 2) the way the ESEF is implemented could add a huge burden to companies’ already complex process of producing annual reports – but it needn’t.

* If anyone wants to challenge this and say ‘well no one reads them anyway so what does it matter’ – I refer you to our forthcoming book! I have been delighted to hear from our research that investors (and other people) do indeed read them.


Our response to FCA Consultation CP19/27** – Chapter 5, Amendments to the DTR to implement ESEF

I am writing to give you our input into the consultation on how ESEF is to be implemented. As reporting consultants and writers, we are heavily involved in the writing and production of annual reports, principally for FTSE companies, including five of the FTSE 100. I have been to a number of seminars and have had many discussions with our clients about the ESEF requirement, and there are two key points at issue here, that could be addressed in a relatively straightforward way – but it all depends on how the new requirement is implemented.

These points are:

1. It is essential that the ‘human’ readability of the annual report is not compromised in order to enable the ‘machine’ readability which is at the heart of this new requirement – or the purpose of reporting will be lost.

2. The production of the ESEF could add considerably to the burden of an already highly intense, complex and lengthy project – but it needn’t.

Here is my analysis of each of these points, with some thoughts on implementation.

1. Ensure the ‘human’ readability of the annual report

The new requirement to report in XHTML and tag the numbers is all about making it possible for machine readers to extract and compare information from annual reports for investors. This is a very useful thing when it comes to information that is taggable and easily comparable – precisely the data that is tagged. However, it does not work for narrative reporting, which is where the human reader comes in. The real strengths of reporting above and beyond providing data lie in what is not easily compared and codified – namely how a company chooses to write and present its story, which ultimately often reveals far more to an investor, particularly about the longer-term picture, than the data, which by its nature cannot speak of the future. The importance of this has been emphasised recently in the requirements of the new Code to avoid a ‘tick box approach’ and the advice from the FRC about ‘telling your story’.

When it comes to reporting, ‘telling your story’ isn’t just about words on a page. An annual report is not a novel that has its own narrative pull, with cliff-hangers at the end of chapters to make you want to read on. For annual reports to be effective for the human reader, they have to be well designed, clearly signposted, and work on a number of levels. People generally do not read annual reports from cover to cover – they dip in and out; some simply skim read, while others read in more depth. A good annual report that serves all its human readers needs graphics and images as well as text; design, writing and layout work together to tell a clear story.

Why might ESEF cause a problem in this respect? Because of the XHTML format. If companies were to prepare their reports directly in XHTML, in a more ’20-F’ style presentation, and move away from a properly designed and written annual report, then readership will suffer, and the whole purpose of reporting – to ensure companies are accountable to and transparent with their stakeholders, to engender trust – could be lost. This is because such reports will be far harder to read, and therefore those company stories that are so important a part of reporting will not reach their audiences.

How to do it: companies should keep the designed pdf format and simply convert it into XHTML for tagging and filing purposes, so that both human and machine readers are well served – and that the spirit of recent regulation around reporting is preserved. It is therefore essential that any requirements for ESEF implementation enable companies to do this.

2. Make the production of the ESEF as least burdensome as possible

In the nearly 20 years that I have been working in corporate reporting, the requirements have ballooned, and the processes have got more complicated, while the time available to produce the annual report has shortened, as the pressure to report faster has increased. For example, one FTSE company with whom I have been working since 2001 has seen their report go from 78 pages to 192, without the business really changing all that much – and I’m sure we’ve all heard the HSBC story from a few years ago when their annual report got so large that postmen couldn’t deliver it. There are good reasons why companies have to report more things in more depth – but the point I am making here is that reporting is already a very lengthy, intensive process, so when adding further requirements, it will help everyone – including Board directors – if they are as least burdensome as possible.

So how might we do that? There has been much debate about whether the ESEF version itself has to be ‘the thing’ that is signed off by the Board, replacing the pdf version, and if it isn’t and the ESEF is produced afterwards (which it would have to be if the pdf remains and the ESEF is created from it), that there would have to be another Board meeting to sign off this second version. This is nonsense, for two reasons.

First, there would be no difference in the two versions, except that the ESEF version would have tags applied – and the Board is concerned with the real content of the report, not the details of the tags, which, in the case of other tagging, for HMRC for example, they delegate to others within the company. How would the Board actually read the XHTML version with the tags applied, given that the tags can’t be seen on a print-out – and, in my experience, it’s a print-out of the designed pdf that Board members will read and approve? Are the Board directors really going to go into the XHTML, click on each tag, read it and check it too? I doubt it.

Second, there are other forms of the annual report that are produced after the pdf is signed off by the Board – the printed copies coming off the printing press, for example, or the Companies House version, which is different from the signed-off pdf. There is no separate Board meeting to sign off either of these. The process for producing the Companies House version is simply delegated, as is the sign-off at the printers to ensure that what comes off the press is what has been approved by the Board. And both of these examples come out after the pdf is published online.

The reason that it would be far easier for companies to finalise and file the ESEF version after the Board has approved the pdf is because these things take time to produce and check. It is already an enormous undertaking for companies to publish their annual reports on the day of the annual results, as some have to do, given that the Board may make changes up to the night before; adding the tagging requirement into that as well could make it nearly impossible. And, as noted above, Board directors aren’t likely to read the tags anyway.

If, therefore, the ESEF does become ‘the thing’ signed off by the Board, then the result could be that companies abandon the designed pdf form in favour of a 20-F style report which could be created directly in XHTML. This would defeat the object of the human readability side of the reporting exercise – of communicating a company’s story to its stakeholders – as noted in my first point above. And, it would have the added effect of making the Board directors’ reading of the annual report far more onerous, because a 20-F style report is far harder work to read than a designed pdf.

How to do it: make the filing of the XHTML ESEF report simply another, and later, output of the existing reporting process, akin to the Companies House filing, and keep the existing pdf format for the official sign-off and publication. Do not require the ESEF to be filed on the same day that the pdf is published. Boards can delegate this duty in the same way that they delegate other reporting sign-off duties.

In summary – the idea of the ESEF is great, and I hope that a machine reader and a single repository for European company reports will soon be created to enable all stakeholders, not just professional investors, to benefit from the analysis such a machine reader would enable.

But like anything, the success or otherwise comes down to implementation – and, for reporting to remain useful and to achieve all its objectives, the ESEF must be an addition to rather than a replacement of the pdf report that, at its best, is a great piece of communication for the human reader.

If I can be of any further help to you in your deliberations around the implementation of the ESEF, please do get in touch. I was recently commissioned by the Governance Institute to write a book on how to produce the annual report (due to be published in June 2020), and have had many discussions with people from the corporate, investor and regulatory worlds as part of my research that I’d be happy to share with you*.

Many thanks,


Claire Bodanis


* Anonymously, of course, dear readers.

Click here to leave a comment

Leave a Reply

Your email address will not be published. Required fields are marked *