The CFO just talked at a conference where he knew how bad revenue would be, and was asked about it, but refused to say! And then he stopped by the IR department to tell them how important it was for them to call analysts and explain that revenue would be low.
Anyway it really was a priority; the IR team “held private, one-on-one phone calls with approximately 20 sell-side analyst firms covering AT&T,” in which they allegedly “disclosed, among other things, AT&T’s internal upgrade rate and wireless equipment revenue data for the first quarter on these calls and otherwise communicated to the analysts, in sum and substance, that the analysts’ revenue estimates were above what AT&T was expecting to report and therefore needed to be reduced,” and “each of the twenty analysts issued revised research reports reducing their revenue estimates shortly after the calls.” And it worked, barely:
On April 25, 2016—the day before AT&T reported its 1Q16 earnings—the last of the approximately 20 analyst revenue reductions brought the consensus estimate just below what AT&T knew it would ultimately report.
AT&T’s CFO emailed the IR Director: “We may just beat revenue consensus.”
The IR Director replied: “I think we will :)”
The CFO forwarded this email to AT&T’s Chief Executive Officer and wrote: “These two updates may do it for us—we may beat revenue consensus—not by much but a beat nonetheless.”
The CEO replied, “Good.”
AT&T’s reported $40.535 billion in revenue for 1Q16, beating the final consensus revenue estimate by less than $100 million.
From an efficient-markets perspective this is all bizarre. One day before AT&T reported its results, the revenue was what it was. Reporting a lot of revenue would be good; reporting less revenue would be less good. But one day before the results, the CFO and the IR director were not worried about the revenue; they were worried about the estimates. Getting analysts to lower their estimates—to say that AT&T would make less money—was good, because then, when AT&T reported how much money it made (again, a number that it already knew), that number would be higher than the consensus estimate, which would allow everyone to say that AT&T “beat revenue consensus.” 2 AT&T’s actual revenue, reported on April 26, would be higher than the average of analyst estimates on April 25, though not higher than the average of those estimates on April 24. It is a completely arbitrary bar, but it mattered to AT&T.
AT&T completely disagrees with all of this and plans to fight:
Tellingly, after spending four years investigating this matter, the SEC does not cite a single witness involved in any of these analyst calls who believes that material nonpublic information was conveyed to them.
The information discussed during these March and April 2016 conversations concerned the widely reported, industry-wide phase-out of subsidy programs for new smartphone purchases and the impact of this trend on smartphone upgrade rates and equipment revenue. Not surprisingly, without device subsidies, customers upgraded their smartphones less frequently, leading to a reduction in equipment revenue.
Not only did AT&T publicly disclose this trend on multiple occasions before the analyst calls in question, but AT&T also made clear that the declining phone sales had no material impact on its earnings. Analysts and the news media frequently wrote about this trend and investors understood that AT&T's core business was selling connectivity (i.e., wireless service plans), not devices, and that smartphone sales were immaterial to the company's earnings.
The evidence could not be clearer—and the lack of any market reaction to AT&T's first quarter 2016 results confirms—there was no disclosure of material nonpublic information and no violation of Regulation FD.
I mean, it’s possible that the lack of any market reaction was because AT&T guided expectations precisely to its actual revenue?