annual vs. 10k, what’s the difference?
Both documents contain audited financials (btw. the first thing to do in reading either document is to make sure the auditors say that the statements give an “unqualified” opinion of their audit– i.e., that they give a “true and complete” picture of company operations. If it’s not a clean audit, the opinion will typically state “except for xyz, where…”
The annual, though, is a marketing document. It’s printed in color on glossy paper, with smiling employees and customers and products portrayed as objects of desire. And it’s ok for the annual to skip over uglier stuff, say, a product recall, or loss of market share or a mass exodus by top management (which might actually be good)…
The 10k, on the other hand, can run to hundreds of pages of small-print prose, with most often no pictures. The annual is sometimes attached–usually as an exhibit, rather than part of the official report. This is because the 10-k is SEC-mandated disclosure, and may contain things that the company doesn’t want put in neon lights for copmpetitors and you and me as investors to see.
An example:
…early in my career, when I was an oil analyst, a colleague left the firm and I took over his coverage temporarily. One of his stocks had two main assets: a very simple offshore oil refinery in the Caribbean and a contract for cheap oil from Iran. The oil ended up in the US, but the company could avoid tariffs by importing refined proeducts rather than crude.
Anyway, the stock was plunging, for no apparent reason. The just-released annual was all sunshine. My numerous calls to the company headquarters were unreturned. I finally got to speak to an employee, who said simply “Look at page 73 in the 10-k,” and hung up. I did. Buried in the middle of that page was a single, small-print statement that the Iran contract had been cancelled.
Nothing anywhere else. But the company had fulfilled its legal disclosure obligation.
This is where AI comes in. I’ve thought for some time that once AI can do the tedious task of reading 10s with at least some measure of understanding, AI users will have a significant information advantage over other market participants.
I’ve recently been using Gemini to find out about the operations of a couple of companies I’m interested in. It’s surprisingly good in getting information like the case of the offshore refinery above. Eventually, Wall Street will catch on. And, although this may be too old-fashioned, my sense is that brokers are so highly trading oriented, rather than research based, that it will take longer than one might think for them to figure this out.
Even then, the game may only shift to diagnosing hallucinations faster than the sell side does.