Abstract
This paper builds a structural model for both advertising and readers’ demand in the Belgian newspaper industry, taking into account cross network effects that exist between advertisers and newspaper readers We combine our parameter estimates with the publishers’ first-order conditions to simulate the impact of a merger. We find a limited impact of the merger on reader and advertiser welfare. The effect on welfare is easily offset by possible efficiency gains. We compare the simulation outcome with the actual outcomes and find—in line with the predictions—no clear aggregate effect of the merger. The responses of individual newspapers, however, differ from the simulation.
This is a preview of subscription content, log in via an institution to check access.
Similar content being viewed by others
Notes
For the reasoning behind this assumption we refer to Sect. 3.
For the Dutch speaking part of Belgium alone, this fraction is even higher: almost 80% in both the reader and advertising markets.
In elite newspapers the emphasis is more on international and political news while popular newspapers spend more attention to human interest, criminal news, and entertainment.
Cover prices of French language newspapers follow a similar pattern.
One should be careful in interpreting the increase in advertising revenue from 2001 to 2002. Because in 2001 a new system to measure advertising revenue was introduced, it is likely that this revenue was under-reported in that year.
Revenue from subscriptions and daily distribution through newspaper shops.
Note that in fact we do not observe advertisements but advertising spending per newspaper/advertiser combination. To be correct, the statement should be: “< 1% of the advertisers that bought advertising space in Het Volk in a given month, did not advertise in another newspaper in the same month”.
The methodology, for example, does not take into account price changes of other firms, nor does it incorporate how the cost pass-through depends on the demand system.
Rysman (2004) applies this model to describe advertising demand in the Yellow Pages industry.
This is equivalent to the assumption that a consumer reads only one newspaper a day. We keep this assumption in our model of readers’ demand.
If there are marketing budget constraints for advertisers, the marketing department will allocate its budget over the newspapers in order to retrieve a maximum return on its investments. Consequently, advertisers will trade off different newspapers. However, this would also mean a departure from the assumption of profit-maximizing advertisers.
Throughout the paper, the quantity of advertising is expressed as the number of black and white pages.
Time subscripts are omitted for simplicity.
We assume G(.) is the same for elite and popular newspapers. The difference between the two types of newspapers lies in the profit per reader of the newspaper: \(\widetilde{\pi }_{j}\).
We make the assumption that an advertiser has an ad on only one page of the newspaper. If advertisers buy advertising space on multiple pages, the returns to scale have to be interpreted as referring to the total ad quantity that is bought in the newspaper.
The model is set up such that irrespective of the value for \(p_{Aj}\), the optimal amount of advertising chosen by the representative advertiser is always larger than zero if \(\pi _{j}\) and \(R_{j}\) are strictly positive. Consequently, the advertisers will always advertise in all newspapers. A solution would be to constrain the minimal amount of advertising to be (for example) \(\frac{1}{16}\) of a page.
The assumptions on the functional form of advertiser profits result in a constant elasticity demand function. This functional form puts certain restrictions on price changes in response to changes in demand/costs as well as to the division of consumer versus producer surplus. As a robustness check, we perform the demand estimation and merger simulation as well using a linear function for advertising demand and the results remained qualitatively the same.
Again, time subscripts are omitted.
Note that the amount of advertising in each newspaper remains limited. It could very well be that for large numbers of ad pages, advertising becomes a nuisance—for example in free newspapers—but for the observed range of advertising quantities we believe that ad intensity does not play a role in consumer choice. For example Sokullu (2015) finds nonlinearities in advertising aversion for the German magazine market.
Belgium is divided into ten provinces: five Dutch speaking and five French speaking. The population in each province varies between 200,000 in the smallest province and 1,400,000 in the largest province.
In October 2000, a free newspaper was launched. We do not include this newspaper in our analysis as we lack detailed distribution figures. Moreover, we do not believe that our results would be affected by the inclusion of the free newspaper. The free newspaper contains only short factual articles—which are taken literally from press agencies—and is consequently not a good substitute for the traditional newspapers from the perspective of the reader. This is confirmed in a regression of readers’ demand where we include a dummy variable that is equal to one for the period when free newspapers were available. The coefficient on the dummy variable is not significantly different from zero. For advertising demand, the introduction of the free newspaper will be picked up by the time dummy variables.
Broadsheet measures 540 × 385 mm (8 columns); Belgian format is slightly smaller: 490 × 336 × mm (7 columns); tabloid is the smallest format: 385 × 250 × mm (5 columns).
We obtained the revenue data from Aegis Media, which compute revenues from observing ads in each newspaper and multiplying these with the appropriate list prices.
Note that advertising prices are not expressed per square millimetres but per fraction of a full page advertisement.
In principle, we can obtain a time-varying measure for these characteristics; but given the relatively short time period for the advertising data, these characteristics do not vary much, and year-to-year variation is most likely to be due to sampling variation. Therefore we opt to use time-invariant reader characteristics that are taken from the 2003 survey.
This figure probably overestimates the real share of advertising revenue in total revenue, since rebates for advertisers are not taken into account.
So, also the number of copies that are distributed for free are included.
However, note that this is not a formal test since we cannot control in this regression for the endogeneity of advertising price due to a lack of exogeneous instruments.
Given the evolution of cover prices in Fig. 1, one could as well argue that the own cover price is exogenous to newspaper-specific advertising demand shocks. Using the own cover price as instrument instead of cover prices of other newspapers renders similar results.
In all our regressions, we pool the Dutch-language and French-language newspapers. In unreported results, we have tested for the equality of the coefficients for Dutch language versus French language newspapers and failed to reject the hypothesis of equality of coefficients. The same is true for readers’ demand.
We do not have information about reader characteristics of Grenz-Echo, so the newspaper is dropped from the sample in this specification.
The dummy variables are not reported but are jointly significant at the 1% level. They pick up seasonal and year effects as well as seasonal effects that are specific to 1 year—such as for example major sports events. The aforementioned problem of misreporting in 2001, is picked up by the dummy variables (as long as there is no systematic difference in reporting across newspapers). To check whether the misreporting is driving our results, we executed the regression excluding the year 2001 and results remained the same.
This is a partial \(R^{2}\) measure that takes into account the intercorrelations among instruments; see Shea (1997). When there is only one endogeneous regressor, the measure is equal to “normal” partial \(R^{2}\) statistics.
As a rule of thumb, Staiger and Stock (1997) suggest that we do not have to worry about weak instruments if the \(F-\)statistic is above 10 in the case of a single endogenous regressor.
These characteristics are time-invariant and are taken from the 2003 survey. We do not expect these characteristics to vary substantially over the sample period.
The German language newspaper Grenz-Echo is excluded from the analysis since the CIM surveys do not report information on it. Moreover, this newspaper is only available in one province and is not expected to be a substitute for French-language newspapers.
This choice will not influence our parameter estimates, since we include province dummy variables in the equation. However, it can have an impact on the computed price elasticities since these depend on the market share of each newspaper.
As Nevo (2001) notes, prices are a function of marginal costs and a markup term and the markup term depends on the unobserved taste parameter.
Recall that prices are set at the national level, so a single newspaper has the same cover price in each province.
For example, the prices of inputs such as paper are not expected to differ across regions in Belgium, given the limited geographic size of the country. Moreover, wage bargaining in Belgium takes place mainly at the national level.
For example, Gazet van Antwerpen was founded in Antwerp and large parts of its content are devoted to news from the province of Antwerp.
Both the province and newspaper dummy variables are jointly significant in all specifications.
However, the Hansen statistic only makes sense when there are at least as many valid instruments as there are endogenous regressors. Given that the overidentification comes from the use of both average cover prices in the same group (elite or popular) but other region, and average cover prices of both groups in the other region, the Hansen p-value should be interpreted with care. It is hard to imagine a situation where average prices of other region newspapers in the same group are valid instruments and average prices over all other region newspapers are not or vice versa.
Price elasticities are computed for each newspaper in each province in the year 2004. Subsequently, a weighted average of these elasticities has been taken, using as weights the population in the province, to get an estimate of the total price elasticities for each newspaper across all provinces. Those are the figures that are reported in Table 6. Consequently, the cross-price elasticities between newspaper j and the other newspapers at the national level are not the same any more for each newspaper in the same group but will be (for example) higher between newspapers that have a strong presence in the same provinces.
Without taking into account the acquisition of De Tijd and L’Echo.
Belgian newspaper publishers set a cover price that is the same in each single province. Consequently, we compute average demand elasticities, weighted by sales in each province, at the national level to infer markups. As such we retrieve an estimate for marginal costs only at the national level—not at the provincial level.
Note that the absolute advertising markup is lower for elite newspapers. However, elite newspapers typically have fewer readers such that the ad markup per reader is higher.
As mentioned before, the co-movement of cover prices displayed in Fig. 1 could be the result of collusion among newspaper publishers. If newspaper publishers are colluding instead of competing in prices, our markup estimates are too low, and thus marginal costs estimates are too high. We computed as well markups and marginal costs under the assumption of joint profit maximization. Now, the average markup on the readers’ side increases to €0.593 and €− 0.056 for popular and elite newspapers, respectively. Note as well that if publishers are colluding, there will be no impact of the merger (discussed below) on prices.
For example, Belang van Limburg and Gazet van Antwerpen have the lowest marginal costs.
Note that at that point, De Persgroep was already distributing the French-language counterpart of De Tijd.
We focus on the Dutch-language newspapers in Flanders since we lack precise data on the number of Dutch-speaking people in Brussels. This number is low however, so ignoring Brussels will not introduce a bias in our merger simulation.
Note that the impact of the merger on the advertising side is limited to the changes in readership that shift the inverse advertising demand curve. This because newspaper publishers act as monopolists on the advertising side.
These are weighted averages with circulation as weights.
Again, in a previous version of the paper, we executed the merger simulation exercise for the specification with linear advertising demand. The main results were not dependent on the chosen advertising demand system.
References
Affeldt, P., Filistrucchi, L., & Klein, T. J. (2013). Upward pricing pressure in two-sided markets. The Economic Journal, 123(572), F505–F523.
Argentesi, E., & Filistrucchi, L. (2007). Estimating market power in a two-sided market: The case of newspapers. Journal of Applied Econometrics, 22(7), 1247–1266.
Armstrong, M. (2006). Competition in two-sided markets. The RAND Journal of Economics, 37(3), 668–691.
Ashenfelter, O., & Hosken, D. (2010). The effect of mergers on consumer prices: Evidence from five mergers on the enforcement margin. The Journal of Law and Economics, 53(3), 417–466.
Berry, S. T. (1994). Estimating discrete-choice models of product differentiation. The RAND Journal of Economics, 25(2), 242–262.
Berry, S., Levinsohn, J., & Pakes, A. (1995). Automobile prices in market equilibrium. Econometrica, 63(4), 841–890.
Bjornerstedt, J., & Verboven, F. (2016). Does merger simulation work? Evidence from the Swedish Analgesics market. American Economic Journal: Applied Economics, 8(3), 125–64.
Brenkers, R., & Verboven, F. (2006). Liberalizing a distribution system: The European car market. Journal of the European Economic Association, 4(1), 216–251.
Budzinski, O., & Ruhmer, I. (2010). Merger simulation in competition policy: A survey. Journal of Competition Law & Economics, 6(2), 277–319.
Busse, M., & Rysman, M. (2005). Competition and price discrimination in yellow pages advertising. The RAND Journal of Economics, 36(2), 378–390.
Chandra, A., & Collard-Wexler, A. (2009). Mergers in two-sided markets: An application to the Canadian newspaper industry. Journal of Economics & Management Strategy, 18(4), 1045–1070.
De Bens, E., & Raeymaeckers, K. (2007). De Pers in België. Tielt: Lannoo Uitgeverij.
Evans, D. S., & Noel, M. D. (2008). The analysis of mergers that involve multisided platform businesses. Journal of Competition Law & Economics, 4(3), 32.
Fan, Y. (2013). Ownership consolidation and product characteristics: A study of the US daily newspaper market. American Economic Review, 103(5), 1598–1628.
Farrel, J., & Shapiro, C. (2010). Antitrust evaluation of horizontal mergers: An economic alternative to market definition. B.E. Journal of Theoretical Economics, 10(1), 1–39.
Filistrucchi, L., Klein, T. J., & Michielsen, T. O. (2012). Assessing unilateral merger effects in a two-sided market: An application to the Dutch daily newspaper market. Journal of Competition Law & Economics, 8(2), 297–329.
George, L. (2007). What’s fit to print: The effect of ownership concentration on product variety in daily newspaper markets. Information Economics and Policy, 19(3–4), 285–303.
Hausman, J., Leonard, G., & Zona, J. (1994). Competitive analysis with differentiated products. Annales d’Economie et de Statistique, 34, 423–446.
Ivaldi, M., & Verboven, F. (2005). Quantifying the effects from horizontal mergers in European competition policy. International Journal of Industrial Organization, 23(9–10), 669–691.
Jeziorski, P. (2014a). Effects of mergers in two-sided markets: The US radio industry. American Economic Journal: Microeconomics, 6(4), 35–73.
Jeziorski, P. (2014b). Estimation of cost efficiencies from mergers: Application to US radio. The RAND Journal of Economics, 45(4), 816–846.
Kaiser, U., & Wright, J. (2006). Price structure in two-sided markets: Evidence from the magazine industry. International Journal of Industrial Organization, 24(1), 1–28.
MediaMarketing. (2004). Hoe Tabloid tot Modules Leidt. Media Marketing, 93. http://www.mm.be/nl/archief.php?id=231. Accessed 10 June 2005.
Miller, N. H., Remer, M., Ryan, C., & Sheu, G. (2017). Upward pricing pressure as a predictor of merger price effects. International Journal of Industrial Organization, 52, 216–247.
Nevo, A. (2000). Mergers with differentiated products: The case of the ready-to-eat cereal industry. RAND Journal of Economics, 31(3), 395–421.
Nevo, A. (2001). Measuring market power in the ready-to-eat cereal industry. Econometrica, 69(2), 307–342.
Pinkse, J., & Slade, M. E. (2004). Mergers, brand competition, and the price of a pint. European Economic Review, 48(3), 617–643.
Rochet, J.-C., & Tirole, J. (2003). Platform competition in two-sided markets. Journal of the European Economic Association, 1(4), 990–1029.
Rysman, M. (2004). Competition between networks: A study of the market for yellow pages. Review of Economic Studies, 71(247), 483–512.
Shea, J. (1997). Instrument relevance in multivariate linear models: A simple measure. Review of Economics and Statistics, 79(2), 348–352.
Sokullu, S. (2015). Network effects in the German magazine industry. Economics Letters, 143, 77–79.
Staiger, D., & Stock, J. (1997). Instrumental variables regression with weak instruments. Econometrica, 65(3), 557–586.
Verboven, F. (1996). International price discrimination in the European car market. RAND Journal of Economics, 27(2), 240–268.
Wilbur, K. C. (2008). A two-sided, empirical model of television advertising and viewing markets. Marketing Science, 27(3), 356–378.
Acknowledgements
We thank the editor, two anonymous referees, Filip Abraham, Anneleen Forrier, Lisa George, Joep Konings, Jo Van Biesebroeck, Ambarish Chandra, Gregory Crawford, Jan De Loecker, Lapo Filistrucchi, Damiaan Persyn, Charles Romeo, Ricardo Ribeiro, John Sutton, Pasquale Schiraldi, Frank Verboven, and participants at the IIOC Conference in Savannah, EARIE Conference in Valencia, MIE Spring Camp at KU Leuven, LSE Work in Progress Seminar Series, and IESE SP-SP Luch Seminar Series for useful comments and suggestions.
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Conflict of interest
Patrick Van Cayseele has been an expert member of the Belgian antitrust authority and was involved in consulting practice in many merger cases in the Media industry.
Rights and permissions
About this article
Cite this article
Van Cayseele, P., Vanormelingen, S. Merger Analysis in Two-Sided Markets: The Belgian Newspaper Industry. Rev Ind Organ 54, 509–541 (2019). https://doi.org/10.1007/s11151-018-9650-z
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11151-018-9650-z