Unlock Editor’s Digest Lock for Free
FT editor Roula Khalaf will select your favorite stories in this weekly newsletter.
The writer is the managing director of Frontline Analysts and the author of “The Accaubleableability Machine.”
The company’s results season for 2024 has been going well this month, leading to that strange addition. This is when the annual report fell on your inbox or doormat. Revenue counts have arrived in the public domain, with conference calls being finished and price action is underway. Why do we care?
Because that is still a statutory requirement. However, due to the extent that the annual report has grown over the years, we cannot blame the regulators for the full extent. There is an old joke among stock analysts that if you really want to keep secret, somewhere between the management pension section and the statement on net zero emissions targets, they will be published in the company’s annual report. As actual actions moved to headline announcements and investor relations calls, the annual report has been transformed into a reservoir of all disclosures that everyone feels the company should do but no one wants to read.
But are the numbers themselves really useful? Unfortunately, this is not that much.
A major advantage that an annual report has on a press release or an investor presentation is that all audit numbers exist. A skilled investor relationship expert can spin gold from the most uncompromising straws. Costs can be “adjusted”, “normalized” revenues, and bad events that are treated as one-off events. Annual reports are where you need to make a confession. Where did the cash come from and where did you go?
If you are having trouble reading a series of annual accounts in detail, you can reveal insights, have the skills to undo the job in the investor relations department, and try to draw as much as possible. In fact, there are probably more opportunities than ever before. The permutation of “adjusted revenue” is worse every year. But do you really give advice to intelligent new graduates who enter the financial industry to develop this skill?
For one thing, being able to decipher the annual report is less useful than before. Over the past five years, many active fund managers have recommended it, but outperformance has been a matter of wandering around, rather than finding hidden treasures. From Tesla’s inventory to Nvidia’s accounts receivable, the ability to ignore the red flags is a more reliable source of alpha than the ability to detect them.
Also, in another case, this is a game that processes a lot of information and examines it to find patterns and inconsistencies. That would certainly be handed over to artificial intelligence, and perhaps earlier than that. Large language models not only allow investors to remove numbers faster than the relevant teams can spin them, but they may even be able to dig into disclosures and find nuggets from time to time.
Therefore, it may be necessary to reinvent the entire concept of the annual report and utilize new technologies to do so. And we should think big. What can we do if we use artificial intelligence to free us from the constraint that a set of accounts must be understandable to humans?
One place to start may be the gap between management accounting and financial reporting. The most misleading number in an annual report is often the date at the top of each column. These often mistakenly mean that 12 months is the relevant period in which performance is assessed.
Recommended
All other numbers in your account are subject to a compromise between the reality of your business and the need to provide investors with a set of comparable, consistent numbers that can be used to make your choices. We call these compromises “generally accepted accounting principles” and it feels like blasphemy for financial analysts to say bad things about them, but they are not laws of nature but trade-offs of convenience.
If all companies publish their accounts today based on what they feel their management system is best represented, it becomes a disaster. But it may be pointless to do so in the future. GAAP may evolve into a set of principles and may ensure that AI always has enough data and metadata to re-register numbers based on what investors may prefer.
This allows everyone to find their own way to understand numbers based on understanding their business. The future may be brighter and more enjoyable for investors and analysts than we were afraid.
