David Solomon, Assistant Professor of Finance and Business Economics at the University of Southern California Marshall School of Business, has a humdinger of a research paper out on the damaging influence IR firms can have on their clients’ stock prices.

It’s an absolute must-read for anyone in the investor relations profession, and for investors. Funny enough, I’ve not seen a single IR firm representative mention this study. After you read it, you’ll know why.

“I focus on whether IR firms spin their clients’ news by creating more positive media coverage and perceptions of the company. The specific method of spinning the news that I examine is increasing coverage of a company’s good news relative to its bad news. Bushee and Miller (2007) find that IR firms increase overall media coverage. The spin hypothesis is more linked to the role of the press in generating information about a story’s importance or accuracy. Media coverage of a story is likely to increase its credibility or perceived importance with investors, which companies may wish to take advantage of by selectively promoting their good news. The most likely benefit for companies is to temporarily increase share prices by affecting investor expectations. If investors use media reports when forming expectations of the company’s prospects, then more positive media coverage may cause investors to bid up the price. However, higher expectations cannot be sustained indefinitely without real effects at the level of company operations. As a consequence, media spin seems likely to result in eventual disappointment. The evidence in this paper indicates that IR firms are significantly involved in spinning their clients’ news stories, and moreover that this impacts stock prices. My main findings are twofold. First, I find significant evidence that IR firms generate greater media coverage of their clients’ good news relative to their bad news, consistent with the spin hypothesis. Second, I find that this positive media coverage increases returns around news announcements. However, it leads to subsequent lower returns around earnings announcements, where IR firms show no ability to generate disproportionately positive media coverage. I argue that the lower earnings announcement returns represent investor disappointment due to the effects of past spin.

I find that during 2002 to 2006, employing an IR firm is associated with an increase in media coverage on announcement days by 25.4% overall. Consistent with the spin hypothesis, the increase is 32.1% for press releases with positive headlines, but only 18.8% for press releases with neutral or negative headlines. This good news/ bad news disparity in the IR firm effect exists only in non-earnings announcements, and is not evident in earnings news. This is consistent with non-earnings announcements being easier to spin. Relative to earnings, non-earnings news is less anticipated (so fewer journalists will be aware of it without IR influence) and likely to contain more soft information (allowing IR firms to push a particular interpretation of ambiguous news events). I also find evidence that IR firms affect investor expectations and stock returns. On non-earnings press release days (when the IR firm can spin the news), IR firm clients have significantly higher characteristic-adjusted returns by 11.1 basis points, after controlling for a large number of other factors. On earnings announcement days (when the IR firm cannot spin the news), IR firm clients lower returns by 33.5 basis points. When the earnings news is negative, the effect is even stronger - the reaction to a given level of negative earnings news is around 67% larger for companies using an IR firm. Moreover, the lower earnings announcement returns appear to be a direct consequence of investor disappointment due to past spin. Earnings returns are significantly more negative if there were higher returns since the previous earnings announcement, and after greater media coverage of positive press releases.

One of the challenges of this paper is to show that the results above are causal, and not driven by unobserved characteristics of companies who hire an IR firm. I address this in several ways. For returns, I examine connections between IR firms and reporters at newspapers, based on reporters who wrote about multiple clients of the IR firm. Turnover among connected reporters should reduce the effectiveness of IR firms in spinning the news. However, it is exogenous to company characteristics, because reporters leaving newspapers seems largely unrelated to which companies they wrote about. I show that connected reported turnover predicts lower returns around non-earnings announcements and higher returns around earnings announcements – that is, it weakens the IR firm effects on announcement returns. This is strong evidence that the patterns in returns are driven by the ability of IR firms to generate media coverage. Moreover, IR firms that lack any connections to reporters show very little ability to influence returns in the first place. The IR firm effect on earnings announcement returns is roughly four times as large for IR firms with some past journalist connections (18 basis points for clients of unconnected IR firms vs. 70 basis points for connected IR firms). For non-earnings announcements, the effect of unconnected IR firms is less than 1 basis point, compared with 22 basis points for connected IR firms.

The effects on media coverage also appear to be causal. IR firms can spin the news more in newspapers where we would expect them to have more influence – papers that are geographically close, and papers that are historically susceptible to IR firm influence. IR firms increase positive news coverage by about 3 times more in newspapers in the same state as the IR firm (after controlling for whether the company is in the same state as the newspaper). Further, when a company adds an IR firm, its positive coverage goes up more in papers that were more susceptible to IR firm influence in the previous year. I also find that that IR firm use is more common for companies with greater incentives to increase share prices in the short term. A higher proportion of CEO pay in stock and option compensation predicts a higher probability of using an IR firm. IR firm clients also engage in other types of manipulative behavior - they are around 3% more likely to restate their earnings in a given year. IR firm clients are also more likely to selectively time their news announcements by releasing positive press releases from Monday to Thursday, when investor attention is higher (DellaVigna and Pollet (2008). These suggest companies that are more interested in pushing up prices in the short term.

The predictable differences in announcement returns also create the potential for profitable trading strategies. I consider a strategy that around earnings announcements buys companies that don‟t use IR firms and shorts companies that use IR firms. Holding stocks for two days produces an alpha of 19.8 b.p. per day or around 50% per year, while holding stocks for 21 days lowers the alpha to 2.8 b.p. per day, or around 7% per year. The fact that very high turnover is required to capture the large differences in returns suggests that transaction costs may partly explain why the mispricing is not eliminated. During a two day window, IR firm clients have a negative earnings announcement premium of around -13.3 b.p. This provides evidence that investors are actually disappointed after the spin.

Posted via email from IR Web Report Bits | Comment »