Public companies release financial results every quarter, and earnings season is always met by hoopla and speculation in the press: Will Company X meet or beat analyst estimates? It’s the kind of bottom-line, win-or-lose drama the media loves. But are those earnings reports, in 10-Q filings and press releases, actually useful to investors?
That might sound like a silly question; after all, the whole point is to give people the information they need to participate in the stock market. So it may surprise you to learn that accounting scholars, who are more acutely aware than anyone of the limitations of financial reporting, have been debating that very question for decades.
“We have a whole industry in the accounting profession whose aim it is to produce financial statements. You want to know, to what end?” laughs Maureen McNichols, a professor of accounting at Stanford Graduate School of Business. “You know, does it make a difference?”
Nearly 50 years ago, Bill Beaver, now an emeritus professor of accounting at Stanford, published a famous study in which he answered that question in the affirmative. That paper won the Seminal Contributions to Accounting Literature Award, an honor that’s been conferred just six times.
But the world has changed, and some have argued that those 10-Q filings might not be as useful or necessary as they once were. So McNichols decided to team up with Beaver (who was her mentor and an inspiration at the start of her own career) and newly minted PhD Zach Wang — “three generations of Stanford accounting faculty,” as she puts it — to revisit the issue.
The researchers compiled a database covering every firm on the three major exchanges from 1971 through 2011 and analyzed stock swings following earnings announcements. What they found, as reported in a new paper, is that earnings reports still convey valued information to investors. But here’s the kicker: Their value has increased over time, and dramatically so since 2001.
Old News, Squishy Numbers
There are good reasons to have expected otherwise. For starters, financial statements look at past performance, and only future performance matters to an investor’s buy/sell decision. Even then, the numbers are iffy. Any earnings figure is contingent on a raft of arbitrary choices concerning cost allocation, depreciation, accruals, deferrals, and so on. As the author of a classic accounting text wrote in 1965, “The task of measuring corporate income and position objectively is an impossible assignment.”
What’s more, the quality of the data may have declined. One of the biggest flaws in our accounting model is its failure to capture the value of intangible assets. In particular, the inability to capitalize intellectual property, a key driver of success in today’s economy, is a serious blind spot.
“R&D is treated as an operating cost when it’s typically an investment,” McNichols says. We do it that way because it’s hard to know what its value will turn out to be. But the result is clearly a distortion of both profitability and assets. “On that score, you’d expect accounting numbers to do less well over time as an indicator of performance,” she says.
Finally, that formal earnings report is, well, just so last-century. Digital technology has lowered the cost of producing and disseminating information, and investors today have access to timelier insights from analysts, blogs, and statistical services, not to mention on-the-fly updates from the companies themselves. In this continuous-information environment, can the period-end financial statement tell investors anything they don’t already know?
Gauging the Information
To answer that question, the researchers used an analytical technique pioneered by Beaver in 1968. The basic idea is that you can quantify the amount of new information contained in a given disclosure by measuring how much the company’s stock price changes afterward. If it’s old news, it will have already been impounded in the market’s valuation and the movement will be minimal.
When they applied this approach to their sample of 700,000 quarterly earnings reports from 1971 through 2011, they found that price revisions were much larger in the days surrounding an announcement than at other times. And that held true in every year, for every randomly selected comparison interval. Over the sample period as a whole, a statistical measure of price volatility averaged 1.18 in non-announcement periods and 2.54 for announcements — confirming and updating the findings in Beaver’s original paper.
But the researchers also noticed something remarkable: That difference widened over time, gradually at first and then sharply after 2001. Between 2002 and 2011, the volatility statistic for announcements averaged 3.51, reaching 4.15 in the final year. In other words, the informational value of earnings announcements has actually increased at a time when the Internet and Internet-based business models were supposed to have made them less informative.
Of course, McNichols points out, there were other things going on, including major reforms in accounting standards and reporting requirements following the dot-com bust and high-profile scandals at Enron and WorldCom. One interpretation of the data is that new regulations under the Sarbanes-Oxley Act (SOX) of 2002 increased public confidence in financial statements.
“A lot of people have looked at the information content of earnings announcements,” McNichols says. “As far as we know, this is the first paper to carry the analysis into the 2000s, so discovering this sudden shift is exciting — it makes you want to know what’s behind it.”
The Role of Analysts
The researchers also took some cross-sectional cuts on their large database to find out whether the price response differs for firms facing different information environments. For instance, you might think that big corporations, dogged by analysts and journalists dissecting their every move, would see some of the news value of their earnings reports eroded.
Indeed, research in the 1980s observed that the price response was higher for small-cap firms. But McNichols and her colleagues find that the pattern flipped in the 1990s, and since then it has been just the opposite: The larger the company and the more analysts who follow it, the larger the reaction to its announcements. “It could be that analysts serve as intermediaries,” she says, “amplifying the information and maybe even influencing disclosure decisions.” But then, why did the pattern change over time?
New Avenues for Research
That’s just one of several fascinating questions raised by the findings in this study — none more intriguing than why the value of earnings announcements jumped after 2001. “We don’t yet know what caused that,” McNichols says, but her intergenerational A-team is launching a follow-up study to find out. “Has the process changed? Are companies bundling additional information in their quarterly reports? What role have SOX and post-Enron regulatory reforms played? These are things we’ll be looking at closely.”
But one thing is clear: The hoopla over earnings season appears to be justified — investors really do pay attention when companies turn in their 10-Qs. “You could say, well, firms put out so much information nowadays, why do we need financial statements? These results show that the information they provide isn’t preempted through other channels,” McNichols says.
“In my mind, the reporting process, the legal environment, the standards and SEC rules, the fact that statements are audited or reviewed — it all makes for a unique kind of information that investors can’t get any other way.”
Looks like accountants won’t need to worry about job security anytime soon.
Maureen McNichols is the Marriner S. Eccles Professor of Public and Private Management at Stanford GSB and a professor by courtesy at Stanford Law.