«13TH INTERNATIONAL PUBLIC RELATIONS RESEARCH CONFERENCE “Ethical Issues for Public Relations Practice in a Multicultural World” Holiday Inn ...»
positive and negative returns are significantly correlated with coverage, and the same is true for the two interaction terms in column . On the other hand, when adding the ownership dummies, only the interaction of ads with positive returns is mildly significant at the 10% confidence level. If anything, the estimated reactivity of coverage is larger for positive returns than for negative ones (as shown by the p-value on the corresponding t-test), but one cannot reject at ordinary confidence levels the null hypothesis that the differential reactivity of coverage to stock returns as a function of ads purchases is not significantly different for positive and negative returns.
Up to now, in order to estimate our effects of interest, we have exploited the fact that -for each listed company in our sample and each trading day- we have media coverage data for six different dailies, which differ on the basis of the monetary amount of ads being purchased by that company. In other terms, we have checked whether coverage of a given company, conditionally on the issuance of a press release and the last available absolute return on the stock market, systematically depends on the amount of ads being purchased on the different newspapers. From this point of view, our identification strategy relies on both the time series and the cross-sectional variation across newspapers.
One could also try and explore the time variation in the amount of ads being purchased by company c on newspaper n, and correlate it again with media coverage. More formally, we modify equation (1) by including fixed effects for each (company-newspaper) pair. This is of course a more demanding specification, since we are controlling for all time-invariant factors that are specific of a given (company-newspaper) pair, and we are solely exploiting the time series variation in newspaper coverage as a function of the time variation in monthly expenditure. Results of this exercise are shown in Table 5, which replicates the format of Table 4, with the sole exception of not including the specifications with the ownership dummies, as (company × newspaper) fixed effects would absorb them.
First, it is interesting to notice that in this setup the partial correlation with monthly ad expenditure is no longer statistically significant, with the sole exception of column , i.e. when controlling for lagged absolute return. On the other hand, it is still the case that newspaper coverage of a given company is significantly larger the day after a press release, and that this increase in coverage is systematically larger the more ads are purchased by that company on a given newspaper.
Columns - show that past absolute return is positively and significantly associated with newspaper coverage. Similarly to what found in Table 4, the interaction between the absolute return and ads expenditure is positive and mildly significant when not controlling for the ownership dummies (column ), while it is not significant when doing so. In column  we again distinguish between positive and negative returns. Both positive and negative returns are significantly correlated with coverage, but only the interaction of ads expenditure with positive returns is estimated to be positive and mildly significant at the 10% confidence level, while the one with negative returns is positive but not statistically significant. However, one cannot reject the null hypothesis that the two interaction terms are equal.
In Table 6 we present some robustness checks of our results. The table is organized as follows: in columns - we add past day trading volume and its interaction with ads expenditure as explanatory variables. As argued by Barber and Odean (2008), larger than usual trading volume for a given stock is likely to be associated with the arrival of relevant news pertaining to that company. In fact, investors might disagree on how to interpret those pieces of news, so that there is a larger amount of transactions on the stock (see Frazzini and Lamont 2007 for additional references). In columns - we control for company and newspaper fixed effects, while in columns - we control for (company × newspaper) fixed effects: for both specifications, we first control for absolute returns and then distinguish between positive and negative returns.
342 We find that trading volume is a positive and significant predictor of newspaper coverage;
moreover, when controlling for time-invariant features of each company-newspaper pair, the interaction between trading volume and ads expenditure is positive and mildly significant. On the other hand, the interaction between returns and ads expenditure is no longer significant at ordinary confidence level. The other results are reasonably robust to this specification, with the only exception of the positive correlation between newspaper coverage and ads expenditure, which is no longer significant with this more demanding specification. Of course, one should take into account that here we are interacting ads expenditure with the press release dummy, stock returns and trading volume within the same specification, so that approximate multicollinearity might lower the estimated precision of individual coefficients.
In columns - we replace last month’s ads expenditure with the contemporaneous value thereof, again interacted with the press release dummy and previous day absolute return.
Endogeneity concerns clearly induce us to prefer the baseline specification -with previous month’s advertising expenditure as explanatory variable-, but one could argue that newspaper coverage more immediately reacts to contemporaneous expenditure. We replicate the set of specifications being used in columns -, and we similarly find that advertising expenditure is no longer a significant predictor of newspaper coverage, even when not controlling for (company × newspaper) fixed effects. Compared to the baseline results shown in Table 4 and 5, the interaction between ads expenditure and the press release dummy is still positive and significant, but at a lower confidence level; on the other hand, the interactions between (absolute and positive) returns and ads expenditure do reach higher levels of statistical significance here.
Finally, in columns - we control for the sum of ads expenditure during the last three months, again properly interacted with our variables of interest. The purpose of this exercise is to check whether the cumulative amount of ads being purchased by a given company on a given newspaper is a less noisy signal than previous month’s expenditure. The sign and statistical significance of ads expenditure, the press release dummy and their interaction is comparable to what found with the baseline specification. Interestingly, while the interaction of ads expenditure with neither positive nor negative returns is significantly different from zero at ordinary confidence levels, here one can reject the null hypothesis that those interactions are equal, i.e. the interaction with positive returns appears to be significantly larger than the one with negative returns.
In this paper we have investigated how –in a sample of Italian newspapers- coverage of listed companies is correlated with advertising. More specifically, we find that the amount of advertising a given company purchases on a given newspaper is positively and significantly associated with the number of articles mentioning that company. This result is robust to controlling for time-invariant features of newspapers and companies.
We have also matched coverage and advertising data with data on the exact days when companies issue their press releases. Unsurprisingly, newspaper coverage of a given company is significantly larger the day after a press release. But it is also the case that this increase in coverage is significantly larger on newspapers where that company has purchased more ads. This result is statistically stronger when controlling for ownership links between companies and newspapers, and when generally controlling for (company × newspaper) fixed effects.
We use the previous day absolute return of the company’s stock as a proxy for the presence of company-specific newsworthy events, which are not fully captured by the issuance of press releases, and find some evidence that positive returns obtain systematically more attention on those newspapers that receive more advertising from the company in question.
From this point of view, strategic actions by firms -in the shape of ads purchases- appear to influence the amount of information regarding them that is provided by actors like newspapers, which in principle should not behave as agents for a principal other than their readers.
In the future we plan to check whether this pattern of results is robust to extending our sample of companies and newspapers. From a broader perspective, one can replicate the analysis performed here in other country settings. Ex ante, in the case of daily newspapers, it would be interesting to check whether the correlations we have found are typical of countries where national level advertising is widespread, while they are less strong in countries like the U.S., where local advertising is comparatively more relevant.
Barber, Brad M., and Terrance Odean (2008). “All That Glitters: The Effect of Attention and News on the Buying Behavior of Individual and Institutional Investors.” Review of Financial Studies, 21(2): 785-818.
Bertrand, Marianne, Esther Duflo and Sendhil Mullainathan (2004). “How Much Should We Trust Difference in Differences Estimates?” Quarterly Journal of Economics, 119(1): 249-275.
Besley, Timothy and Andrea Prat (2006). “Handcuffs for the Grabbing Hand? Media Capture and Government Accountability.” American Economic Review, 96(3): 720-736.
Bushee, Brian J. and Gregory S. Miller (2007). “Investor Relation, Firm Visibility and Investor Following” Mimeo, Wharton School.
Clark, Fiona and Deborah L. Illman (2006). “A Longitudinal Study of the New York Times Science Times Section.” Science Communication, 27(4): 496-513.
Nguyen, Bang Dang (2005). “Is More News Good News? Media Coverage of CEOs, Firm Value and Rent Extraction.” Paper presented at the AFE/ASSA 2006 Boston Meetings.
Depken, Craig A., II and Dennis P. Wilson, “Is Advertising a Good or a Bad? Evidence from U.S.
Magazine Subscriptions.” Journal of Business, 77(2): S61-S80.
Di Tella, Rafael and Ignacio Franceschelli (2009). “Government Advertising and Media Coverage of Corruption Scandals.” NBER working paper no. 15402.
Durante, Ruben and Brian Knight (2009). “Partisan Control, Media Bias, and Viewer Responses:
Evidence from Berlusconi's Italy” NBER working paper no. 14762.
Ellman, Matthew, and Fabrizio Germano (2009). “What do the Papers Sell? A Model of Advertising and Media Bias.” Economic Journal, 119(537): 680-704.
Erjavec, Karmen (2004). “Beyond advertising and journalism: hybrid promotional news discourse.” Discourse & Society, 15(5): 553-578.
Fang, Lily, and Joel Peress (2008). “Media Coverage and the Cross-Section of Stock Returns.”
Journal of Finance, forthcoming. Available at:
http://www.insead.edu/facultyresearch/faculty/personal/jperess/research/documents/Fang_PeressMed ia_Coverage_Stock_Returns.pdf Fengler, Susanne and Sthephan Ruß-Mohl (2008). “Journalists and the information-attention market.” Journalism, 9(6): 667-690.
Gentzkow, Matthew A., Edward L. Glaeser and Claudia Goldin (2006). “The Rise of the Fourth Estate: How Newspapers Became Informative and Why it Mattered.” In Edward L. Glaeser and Claudia Goldin (eds.), Corruption and Reform: Lessons from America's History, National Bureau of Economic Research.
Gentzkow, Matthew A., and Jesse M. Shapiro (2009). “What Drives Media Slant? Evidence from U.S. Newspapers.” Econometrica, forthcoming.
Groseclose, Tim, and Jeffrey Milyo (2005). “A Measure of Media Bias.” Quarterly Journal of Economics, 120: 1191-1237.
Grullon, Gustavo, George Kanatas and James P. Weston (2004). “Advertising, Breadth of Ownership and Liquidity”, Review of Financial Studies, 17(2): 439-461.
Gurun, Umit G. and Alexander Butler (2009). “Don’t believe the hype: Local media slant, local advertising and firm value”, AFA 2010 Atlanta Meeting Paper.
Hamilton, James T. (2004). All the News that’s Fit to Sell. How the Market Transforms Information into News. Princeton, Princeton University Press.
Ho, Daniel E., and Kevin M. Quinn (2008). Assessing Political Positions of the Media.'' Quarterly Journal of Political Science, 3(4): 353-377.
Knight, Brian G., and Chun-Fang Chiang (2008). “Media Bias and Influence: Evidence from Newspaper Endorsements.” NBER working paper no. 14445.
Lamont, Owen and Andrea Frazzini (2007). “The Earnings Announcement Premium and Trading Volume.” NBER Working paper no. 13090.
Larcinese, Valentino, Riccardo Puglisi, and James M. Snyder, Jr. (2007). “Partisan Bias in Economic News: Evidence on the Agenda-Setting Behavior of U.S. Newspapers.” NBER working paper no.
Lott, John R., Jr., and Kevin A. Hassett (2004). “Is Newspaper Coverage of Economic Events Politically Biased?” Unpublished manuscript, University of Maryland and American Enterprise Institute.
Peress, Joel (2008). “Media Coverage and Investors’ Attention to Earnings Announcements.”
Mimeo, INSEAD. Available at:
http://www.insead.edu/facultyresearch/faculty/personal/jperess/research/documents/Peress_Media_C overage-Attention.pdf Norris, Claire E. and Andrew M. Colman (1992). “Context Effects on Recall and Recognition of Magazine Advertisements.” Journal of Advertising, 21(3): 37-48.
Puglisi, Riccardo, and James M. Snyder, Jr. (2008). “Media Coverage of Political Scandals.” NBER working paper no. 14958.
Notes: the relative frequency of articles on newspaper n about company c is calculated by dividing the daily count of articles mentioning company c on newspaper n by the daily number of articles where the word "il" (Italian definite article for masculine nouns) appears. This relative frequency is expressed in percentage points.