Words: 3340
We also find that the more viewers dislike ads, the ore likely it is that welfare is increasing in the number of advertising financed TV channels. A publicly owned TV channel can partly correct market distortions, in some cases by having a larger amount of advertising than private TV channels. It may even have advertising in cases where advertising is wasteful per SE. JELL classification: ALL, MOM Keywords: Television industry; Advertising Correspondence should be sent to Tore Nielsen, Department of Economics, University of Oslo, P. O. Box 1 095 Blinder, NO-0317 Oslo, Norway. E-mail: tore. nilssen@econ. Ii. O. We are grateful to Steve William, three anonymous referees, and seminar participants in Antwerp, Helsinki, and Milan for helpful comments. We would like to thank the Research Council of Norway (the KIM program) for its financial support through SIN Institute for Research in Economics and Business Administration. Kind would like to thank Chiefs in Munich for excellent working conditions while revising this work. The TV industry is important both in terms of the time people spend watching TV and the amount of advertising it transmits. 1 However, advertising- financed channels are potentially a mixed blessing.
On the one hand, TV immemorial may be the most efficient way for firms to advertise their products and can generate a surplus both for individual firms and for society as a whole. On the other hand, viewers may dislike being interrupted by commercials. 2 We thus have an ambiguity that raises the questions of whether there is over- or underproduction of advertising on TV and of whether there is a need for some kind of public intervention in the sector. Would it for instance be advantageous to restrict entry of commercial TV channels if consumers dislike ads?
In this paper, we set out to provide answers to these questions with the help of a simple model in which TV stations sell advertising space to advertisers. The basis for the advertisers’ willingness to pay for such advertising space is the attention of the viewers that the stations attract. And in order to attract viewers, the stations offer TV programs. Thus, the TV industry is an example of a two-sided market: TV stations offer programs to viewers and advertising space to advertisers, with externalities in both directions. We find that there is too little advertising on TV when the channels’ programs are close substitutes, and that the scope for such interspersion of advertising becomes higher if the number of TV channels is enlarged. We further show that the more viewers dislike ads, the more likely is that welfare is increasing in the number of advertising financed TV channels. Well-known discussions of the welfare effects of advertising, such as Dixie and Norman (1978) and Becker and Murphy (1 993), do not take into account the role of media firms as transmitters of advertising. An early attempt to do so is in Spence and Owen (1977). However, in their discussion of advertising-financed TV versus pay TV, the presence of advertising is assumed to have no effect on viewers. William and Owen (1985) extend the Spence-Owen model to take into account that commercials are a nuisance to TV viewers. Our analysis differs from the work of Spence-Owen and William- Owen in that we model strategic interactions beet”en TV stations in an oligopoly, whereas they assume monopolistic competition without any strategic interaction.
Moreover, in the Spence-William-Owen models the price of commercials 2 is exogenously determined, whereas we let it be endogenously determined. This is important for understanding strategic interaction in media markets. For example, we show that tougher competition caused by greater substitutability between TV channels leads to higher prices per minute Of advertising, while tougher competition caused by an increase in the number of TV channels leads to lower such prices. In other respects, however, our approach is similar to that of Spence and Owen (1977) and William and Owen (1985).
In particular, we follow them in modeling the demand for TV programs by way of a quadratic utility function of a representative viewer. In recent analyses of the media market, such as Gal-or and Dukes (2003), Zwieback, Lasses, and Cognacs (2004), Petit and Valletta (2005), and Anderson and Zwieback (2006), viewers are located along a Whittling (1929) line, varying according to their preferences for Waveform content, and TV’ channels choose positions on that line. 5 Thereby product differentiation is endogenously determined.
But this comes at a cost: First, these analyses are limited to discussions of duopoly and have small chances of being generalize to an oligopoly. Secondly, a Whittling analysis fixes the total size of the market (I. E. , the total amount of TV viewing). Thirdly, each viewer is assumed to watch one channel only. We choose a different angle and fix the degree of product differentiation, in line with the earlier work of Spence and Owen (1977) and William and Owen (1985). A major benefit is that we can discuss oligopoly and do not have to limit ourselves to duopoly.
This is particularly interesting since most democratic countries have reduced or eliminated regulatory entry barriers for TV channels over the last decades. A major insight from our analysis is that a larger number of advertising- financed TV channels is more likely to have positive welfare effects, the higher s the consumers’ disability from advertising. The intuition for this somewhat paradoxical result is that higher competition through a larger number of TV channels forces each TV channel to sell less ad time when the viewers dislike commercials.
The positive welfare effects of less ad time are greater the more the consumers dislike advertising, other things equal. Our approach allows for a framework in which the total time people spend watching TV is endogenous, depending on the extent of competition. As in most other markets, our model features the appealing property that stronger intention leads to higher output, that is, to people spending more time watching TV. Finally, our formulation allows a viewer to allocate his time between different channels. This increases the competition in the market for advertising space, as 3 each n. Channel can offer advertisers a little of each viewers attention. 6 In many European countries there are mixed oligopolies in the TV industry with both publicly and privately owned TV channels. 7 Introducing a welfare- maximizing publicly owned TV channel into our model, we show that, for sufficiently differentiated TV channels, the public TV channel sells less advertising space than the private channels. Thereby the provisions Of advertising in a system with only privately owned TV channels is mitigated. Conversely, the public TV channel brings more advertising than the private ones if TV programs are sufficiently close substitutes.
In fact, we find that a welfare-maximizing public TV channel brings advertising even in some cases where advertising is per SE wasteful (I. E. , where the disability of viewers exceeds the surplus that the advertising generates for the advertisers). This article is organized as follows. The formal model is presented in the next section. In the three subsequent sections, we discuss and compare equilibrium outcomes and social optimum. Thereafter, we analyze the implications of introducing a welfare-maximizing TV channel owned by the government. In the final section we offer some concluding remarks.
The model We consider a model with m 2 TV stations and a continuum Of identical viewers with measure one. The time that each viewer spends watching n. ‘ programs on channel I = 1 , … ,m is denoted by Vi. We follow Motto (2004) and assume that consumers’ preferences are given by the Subintervals utility unction, originally introduced by Suburb and Levitate (1 980): m m 02 0 1 C We may interpret Vi both as the time that each viewer spends watching channel I and as the number of viewers of channel I, since we have normalized the population size to 1.
The parameter s e (0,1) is a measure of product differentiation: The higher is s, the closer substitutes are the TV channels from the viewers’ point of view. The Suburb-Levitate formulation ensures that the parameter s only captures product differentiation and has no effect on market size. 8 The TV channels are financed by advertising, and an be watched free of charge. However, the viewers have a disability of being interrupted by commercials.
To capture this, we assume that the viewers’ subjective cost of watching channel I is Chi = waive, where Ai is the ad 4 time on TV channel I and O is a parameter that measures the viewers’ disability from advertising. A viewer’s consumer surplus is thus given by CSS -?? U-heavily. Ii=mm In a sense, advertising is an indirect price: It has the same function in the market for TV programs as prices have in other markets. By setting docs DVD = 0 , we find the audience’s demand for viewing TV channel I: (2) Vi where A -?? CIA-SOCIO mi=l Ai is the average level of advertising on the m channels.
The time viewers spend watching channel I is thus strictly decreasing in the level of ad inventory on that channel, and more so the higher their disability of being interrupted by commercials (captured by y), and increasing in the levels of ad inventory’ on the competing channels: The more commercials there are on the rival channels, the more attractive is channel for the viewers. Channel I charges the price Ri per advertising slot, and we set operating profit of channel I equal to in Ri m . Throughout the article we abstract from both variable and fixed costs for the TV channels. This is not to say that such costs are unimportant (see, e. G. Motto & Polo, 1997, for a discussion). However, this allows us to highlight some basic competitive forces and strategic effects without having to worry about free entry conditions. Let Kaki denote advertiser KS advertising level on channel I. The advertiser’s gross gain from advertising at channel is naturally increasing in its advertising level and in the number of viewers exposed to its advertising. We make it simple by assuming that the gross gain equals Kiev. This implies that the net gain for advertiser k from advertising on equals C m Accommodation Ri C, e n, 0 Chi=l advertiser k. The advertisers’ aggregate profit equals n A k TTT k . Where n is the number of advertisers. With a slight abuse of terminology, we label ask the profit of 5 Most of the analyses in the literature consider the advertisers to be price takers and derive demand for advertising by way of a zero-profit condition on the marginal advertiser’s profit. 9 We find it useful to do this differently, and more in line with models of successive oligopoly, such as Slinger (1988), here producers and retailers set quantities sequentially. We consider the following two-stage game: Stage 1 : TV channels set levels of advertising space. Stage 2: The advertisers choose how much advertising space to buy.
One noteworthy feature of our set-up is that the TV channels are quantity setters in advertising. If program choice is inflexible in the short run – with a given amount of time been each program – such an assumption is plausible. However, there might be arguments indicating that TV channels are more flexible concerning the amount of advertising. 10 If so, price setting on advertising is a more natural choice. It can be shown that our main results still hold if we assume price setting rather than quantity setting among TV channels-II Unless stated otherwise, we assume that the TV channels act non-cooperatively.
Equilibrium outcomes We solve the game by backward induction. At stage 2, the advertisers simultaneously determine how much to advertise on each of the m channels, taking prices of advertising space as given. Solving k Dais = 0 simultaneously for the n advertisers and then using Ai = k Kaki we find that n demand for advertising at channel I equals Ai=l n [1 -??m (1 -??s) Ri -?? ms y n +1 I = 1 where R = m IM 1 Ri is the average advertising price on the m channels.
As expected, we thus have dad a downward-sloping demand curve for advertising ( drib As is standard in markets with this property, the TV channels’ advertising levels are strategic complements (see Ivies, 1999). Using (5), We can write the inverse aggregate demand curve for advertising on channel I as Ri = Note that 1 D n + 1 Ai-asњ01 y n +1 Ai-Adair = . Ads m n (1 -s This means that the marginal willingness to pay for ad time on a channel is increasing in s if and only if the ad time on that channel is lower than the average (Ai The TV channels set their ad inventory non-cooperatively at stage 1 . (For the case where TV channels collude on advertising levels, see below. ) Solving we find that the equilibrium ad time at each TV channel equals AiM -?? Inn(l (7) -??s),I=l,… , m, , m, subject to (6), -s)yen+l d in dad where the superscript M denotes market equilibrium. 12 From this equation we see that the ad time is decreasing in the viewer disability of advertising (y), which is quite natural. Moreover, Adam DNA > 0 : An increase in the number of advertisers increases the demand for advertising, and it becomes optimal for the TV channels to offer more advertising space. Inserting (7) into (2), we find equilibrium TV viewing on each channel: ViM = Multiplying this expression by m, we see that total viewing time ( movie M ) depends on the number and substitutability of TV channels. In other words, total output varies with the competitive pressure between the firms. This realistic feature of our model is in contrast to the widely used 7 Whittling framework, where the size of the market by definition is constant.
We note from (8) that TV viewers’ equilibrium consumption is unaffected by y: The TV channels thus completely internalize the consumers’ disability of advertising through the levels of advertising. From (7) we find that Adam Ads 0 , which means that the equilibrium ad time is smaller the less differentiated the TV’ channels. To understand this result, note that a TV channel attracts viewers by limiting the quantity of advertising space. The better substitutes the viewers perceive the TV channels to be, the more sensitive they are to differences in ad time.
A high s thus gives each TV Handel an incentive to set a relatively low ad inventory in order to capture viewers from the other channel. Corresponding to smaller ad inventories when s increases, TV viewers’ consumption increases: DVD M Ads Finally, we get the typical effects of an increase in the number of competitors: Each firm’s output is reduced ( d ( movie M ) DMS Adam DMS O , and DMS > O Again, these are results that cannot be obtained in the Whittling framework, where movie M is fixed.
Summarizing our main results so far, we have: Proposition 1: (a) The larger he number of TV channels (I) the less time do viewers spend on each individual TV channel, but (ii) the more time do they spend on TV viewing in total. (b) The equilibrium ad time on each channel is smaller (I) the less differentiated the TV chance Nell’ programs; (ii) the higher the viewers’ disability of advertising; and (iii) the higher the numbers of TV channels.
At the same time as competition forces the TV channels to reduce their ad inventories, it R allows them to charge a higher slotting price RiM and a higher contact price per viewer, r IM V IM : I M RiM = -s) ] Ads 8 rim trim>O. N+1)(m-s)+m(1 (n +1 )(m-??s) -s ) Ads (10) By insertions into the expressions for profit in (3) and (4), we can now find the equilibrium profit levels of TV stations and advertisers: M in = (11) (12) M RL Differentiation of (1 1) and (12) shows that profits are decreasing in both y and s.
This is natural, since the level of ad inventories is smaller the larger the consumers’ disability of advertising and the closer substitutes are the TV channels’ programs. However, equation (11) might leave the impression that the TV channels will make positive profit for any finite value of y and for any s If we had subtracted such costs from the operating profit in equation (1 1), zero profit constraints would have implied that there is room for fewer diversification’s TV-channels the less horizontally differentiated they are and the higher the consumers’ disability of ads, other things equal. Recall that the slotting price is higher the closer substitutes the consumers perceive the TV’ channels to be. This is because reduced differentiation increases the competitive pressure. A higher s thus implies that the TV channels will have less ad time ( M Adam Ads O) .
The same kind of reasoning might lead one to expect that the slotting price also Ads increases in the number of TV channels. This is not true. From equation (9) we find, on the contrary, that trim DMS O .
It thus becomes more expensive Ads for the advertisers to reach each viewer the larger the number of channels. We therefore get the somewhat surprising result that the profit level of the advertisers is decreasing in the number of 9 TV channels: M d art DMS We summarize our results concerning profits: Proposition 2: Equilibrium profits both for TV channels and for advertisers are higher (I) the more differentiated the TV channels’ programs, (ii) the lower the viewers’ disability f advertising, and (iii) the lower the number of TV channels.
Less advertising time on each channel is an advantage for consumers. Additionally, consumers gain subsequent to an increase in m because the diversity of TV channels increases. We thus unambiguously have docs DMS We end this section with an extension to the case where there is collusion among the TV channels about levels of ad inventories. When s = O, the TV channels’ programs are independent, and collusion has no effect at all. At the other extreme, we know that the TV channels compete away (almost) all advertising and have close to zero profits when s approaches 1 .
This is a prisoners’ dilemma situation, where the firms would have been jointly better off with more ad time on all channels. This suggests that collusion between the TV channels leads to more advertising than in the non-cooperative equilibrium for all s E (O, 1), and more so the less differentiated the TV programs. 14 We derive the first-order conditions for a collusive outcome from the TV channels’ joint profit-minimization problem, and find that the equilibrium ad time on channel I now equals IAC = 1 n, I = (13)