Advertising for Attention Marco A. Haan∗
Jos´e Luis Moraga Gonz´alez† March 21, 2007
Abstract We model the idea that when consumers search for products, they first visit the firms which advertise more, or the firms whose advertising is more salient. Equilibrium prices and advertising efforts are increasing in consumer search costs. By contrast, equilibrium profits are nonmonotonic in search costs so firms are not necessarily better off if search costs rise. In our model, firms are engaged in a battle for consumer attention so advertising is purely wasteful. If advertising were banned, equilibrium prices would not be affected and firms would be better off.
---PRELIMINARY--JEL Classification Numbers: D83, L13, M37. Keywords: Advertising, search.
1
Introduction
Advertising is an increasingly important aspect of everyday life. According to PriceWaterhouseCoopers worldwide advertising in 2005 amounted to a staggering $385 billion (PriceWaterhouseCoopers, 2005). This amount is set ∗
Department of Economics, University of Groningen, PO Box 800, 9700 AV Groningen, The Netherlands, e-mail:
[email protected]. † Department of Economics, University of Groningen, PO Box 800, 9700 AV Groningen, The Netherlands, e-mail:
[email protected].
1
to grow to over half-a-trillion dollars in 2010. Economists have dedicated a significant amount of effort to understand the role of advertising in markets. Traditionally, advertising has been thought of as a sunk cost firms incur with the purpose of enhancing consumers’ willingness-to-pay for their products. This type of advertising has been termed persuasive advertising. The idea behind its utilization is that persuasive advertising creates fictitious product differentiation, which softens product competition (Kaldor, 1950, Galbraith 1967, Solow 1967). Since Telser (1964) and Nelson (1970, 1974), the view of advertising as a device to transmit information has been gradually gaining support by economists. Generally speaking, informative advertising helps to reduce the lack of information existing in product markets by, for instance, communicating firms’ existence, their prices, qualities or locations. Informative advertising increases the information economic agents possess, which increases competition and raises social welfare. It is hard to believe that all of this advertising is purely of the informative or the persuasive kind that is often studied in Industrial Organization models. Rather, it seems that firms are competing to increase consumer attention. As a matter of fact, as the amount of advertising grows, it is increasingly difficult for a firm to stand out from the crowd, to get the attention of consumers, and to convince them to come to your firm, or to try your product. As Comanor and Wilson (1974, pg. 47) argued: To the extent that the advertising of others creates ‘noise’ in the market, one must ‘shout’ louder to be heard, so that the effectiveness of each advertising message declines as the aggregate volume of industry advertising increases.1 In his extensive literature survey on advertising, Bagwell (2007) argues that this approach of advertising warrants formalization: ”Future work might revisit this noise effect, in a model that endogenizes the manner in which consumers with finite information-storage capabilities manage (as possible) their 1
For an empirical study on how the “salience” of one brand inhibit the recall of alternative brands see Alba and Chattopadhyay (1986).
2
exposure to advertising” (pg. 111/112). In this paper, we try to contribute to this line of enquiry by proposing a novel model where advertising serves to increase the likelihood consumers recall a particular firm and visit this firm earlier than other firms. Suppose that a consumer wants to purchase a certain product, but is not sure where she can purchase that product. After some thinking, she remembers an ad that she has seen from a shop that sells the product that she is interested in and goes there to see what exactly the shop has on offer. Such a visit is costly. If the product is not to her liking, she has the option to walk away from the shop, recall an alternative vendor and visit such firm, again at some cost. This process continues until the consumer finds a deal that is so attractive that searching further is not worth her while, or after she has visited all firms. In the latter case, she will return to the firm that offers the best deal. We assume that the probability that a consumer will remember a shop at every stage in the thought process is proportional to that shop’s share in total industry advertising expenditures. In this model, firms are thus engaged in a battle for attention, in the sense that a consumer is more likely to visit a firm if it advertises more. This also implies that we need to couch our analysis in terms of a search model. Otherwise, consumers would simply visit all suppliers and there would be no advantage of advertising over and above rivals’ advertising efforts. Technically, we have a search model in which we endogenize the order in which firms are visited. Our model of search builds on the work of Anderson and Renault (1999), who add search costs to the analysis in Perloff and Salop (1985). In a standard search model, firms are visited at random.2 That is not true in our model: a firm is more likely to be visited if it advertises more. The fact that firms sell horizontally differentiated products implies that, even in equilibrium when all firms charge the same price, a consumer does not necessarily buy from the first firm that she visits. If she does not like this firm’s product(i.e. if she happens to get a low matching value with this firm), 2
Note however that this is not the case in Arbatskaya (2007). However, in her model the order of visit is still exogenously given, whereas in our model it is determined endogenously.
3
then she will simply search further. This also implies that we do not assume that advertising is an all-or-nothing venture. A firm that manages to capture the attention of our consumer will be visited first, but that does not imply that the consumer will also buy from this firm. Hence, there still is a role for competition in prices as well. Of course, firms that are visited first, do have an advantage, as the consumer has to incur additional search costs when visiting another firm. Strictly speaking, our model primarily applies to a set-up where shops try to attract consumers. Yet, we do believe that our model can also serve as a metaphor for other environments in which advertising is a battle for attention. Also note that, in the strict interpretation of our model, we need that shops cannot advertise the price of their products. This may be the case if shops sell many products and it is impossible to list the prices of all products. An alternative interpretation is that consumers simply cannot remember all the prices they have seen in advertisements. The best an advertiser can hope for is that consumers remember its identity. In our model, advertising is socially wasteful. In equilibrium, all firms use the same level of advertising, which implies that a consumer visits each firm with equal probability. This would also be the case absent advertising. The firms thus have a classic prisoners’ dilemma. We find that both prices and advertising are increasing in search costs. If searching is more costly, each firm will have more market power over each consumer that pays it a visit, which will raise prices. At the same time, it also becomes more profitable to attract a consumer in the first place, which implies that firms will advertise more. The effect of an increase in search costs on equilibrium profits, is ambiguous. Such an increase has two effects. On the one hand, firms can charge higher prices, which is good news for them. On the other hand, they advertise more, which is bad news. If search costs are small, the good news dominates, and equilibrium profits increase. With high search costs, however, the bad news dominates and equilibrium profits decrease in search costs. The remainder of this paper is structured as follows. In section 2 we give the set-up of the model. The equilibrium is derived in section 3.1. We give 4
the comparative statics results of our model in section 4. Section 5 concludes.
2
The Model
On the supply side of the market there are n firms selling horizontally differentiated products. They employ a constant returns to scale technology of production and we normalize unit production costs to zero. On the demand side of the market, there is a unit mass of consumers. A consumer m has tastes described by a indirect utility function umi (pi ) = v − pi + εmi , if she buys product i at price pi . The parameter εmi can be thought of as a match value between consumer m and product i. Match values are independently distributed across consumers and products. We assume that the value εmi is the realization of a random variable with distribution F and a continuously differentiable density f with support that is normalized to [0, 1]. The consumer must incur a search cost s in order to learn the price charged by any particular firm as well as her match value for the product sold by that firm. No firm can observe εmi so firms cannot practise price discrimination. The constant v is assumed to be large enough such that, in equilibrium, all consumers buy.3 Consumers search sequentially with costless recall. To model advertising for attention, we assume that the consumer is more likely to go to a firm if she has been relatively more exposed to ads from that firm, or if the ads from that firm have been relatively more salient. This assumption captures the idea in the marketing and business literatures of “top-of-the-mind awareness”.4 More precisely, we assume that the probability that a consumer will visit a firm in first place, is proportional to this firm’s share in total industry advertising. This technology is similar to that in the rent-seeking contest described by Tullock (1980). Schmalensee (1976) uses a similar idea in the context of advertising, but in his model prices are 3 4
This assumption can be relaxed without affecting the qualitative nature of our results. See
5
exogenously given. Intuitively, one can think of each advertising dollar that a firm spends, as a ball that this firm puts in an urn. Each firm can put as many balls in the urn as it likes. When the consumer happens to need a product and has to start searching for it, intuitively it is as if she takes one ball from the urn and visits the firm that has put this particular ball in the urn. Note that this advertising technology is somewhat similar to that in Butters (1977). In his model, the consumer is already informed about a firm once she receives one ad from that firm. In our model, it is the relative number of ads that matters when deciding which firm to visit when. The timing in our model is as follows. First, firms simultaneously set both advertising levels and prices. Second, the consumer sequentially searches for the best deal. Let ai denote the amount of advertising by firm i, and pi be the price firm i charges, i = 1, 2, ..., n. We will focus on symmetric pure-strategy equilibria, i.e., where ai = a∗ and pi = p∗ for all i.
3
Analysis
Given the strategies of the rival firms (a−i , p−i ) = (a∗ , p∗ ), to derive the (expected) payoff to a firm charging a price pi and advertising with intensity ai , we need to consider the order in which the firm may be visited and the probability to make a sale conditional on being visited at a given point in time. We start by computing the joint probability that consumers visit firm i first and decide to accept the offer of firm i without searching further. Suppose that a buyer has approached firm i in her first search and her current purchase option yields utility v + εi − pi ≥ 0. In the Nash equilibrium, a visit to a new firm j will yield utility v + εj − p∗ . The consumer will prefer to buy from firm j if εj > ∆ + εi ≡ x, with ∆ = p∗ − pi . Therefore, the expected benefit from searching once more R1 is x (ε − x)f (ε)dε. Searching is worthwhile if and only if these incremental 6
benefits exceed the cost of searching one more time, s. We thus have that the buyer is exactly indifferent between searching once more and stopping and accepting the offer at hand if x ≥ xˆ, with xˆ implicitly defined by Z
1
(ε − xˆ)f (ε)dε = s.
(1)
x ˆ
Given this, the probability that a buyer stops at firm i given that firm i is sampled, is equal to P r(x > xˆ), or5 Pr(∆ + εi > xˆ) = 1 − F (ˆ x + ∆). If we denote the probability that a consumer visits firm i in her first search and buys there as λi1 (ai , pi ; a∗ , p∗ ) we have: λi1 (ai , pi ; a∗ , p∗ ) =
ai (1 − F (ˆ x + ∆)) ai + (n − 1)a∗
(2)
We now compute the joint probability that a consumer patronizes firm i in her second search and the consumer decides to acquire the offering of firm i. For this we need to take into account that the consumer has visited some other firm first but walked away from that firm because the offering was not satisfactory enough. If we denote by λi2 (ai , pi ; a∗ , p∗ ) the chance that firm i is visited second and makes a sale we have:
λi2 (ai , pi ; a∗ , p∗ )
µ = 1−
ai ai + (n − 1)a∗
¶
ai F (ˆ x)(1 − F (ˆ x + ∆)) ai + (n − 2)a∗ (3)
More generally, the joint probability that a consumer visits firm i in her kth search and buys there, k = 3, . . . , n − 1, is k−1 Y (n − `) a∗ ai F (ˆ x)k−1 [1 − F (ˆ x + ∆)] ai + (n − k) a∗ `=1 ai + (n − `) a∗ (4) and for k = n we get
λik (ai , pi ; a∗ , p∗ ) =
5
For details, see Anderson and Renault (1999).
7
λin (ai , pi ; a∗ , p∗ ) =
n−1 Y `=1
(n − `) a∗ F (ˆ x)n−1 [1 − F (ˆ x + ∆)] ai + (n − `) a∗
(5)
To complete firm i’s payoff calculation, we need to compute the joint probability that a consumer walks away from every single firm in the market and happens to return to firm i to conduct a transaction, which occurs when maxk6=i {uk } < ui − ∆. Since this probability is independent of the order in which firms are visited we will denote it as R(pi ; p∗ ). We then have: Z
x ˆ+∆
∗
R(pi ; p ) =
F (ε − ∆)n−1 f (ε)dε.
(6)
0
Using the notation introduced above, we can now write the expected profits of firm i: Πi (ai , pi ; a∗ , p∗ ) = pi
" n X
# λik (ai , pi ; a∗ , p∗ ) + R(pi ; p∗ ) − ai .
(7)
j=1
We look for a Nash equilibrium in prices and advertising levels. Thus, we need a price p∗ and an advertising level a∗ that are implicitly defined by the following first-order conditions: n X λik (a∗ , p∗ ) ∂Πi (a∗ , p∗ ) ∗ =p − 1 = 0, ∂ai ∂ai k=1
(8)
n
∂Πi (a∗ , p∗ ) X λk (a∗ , p∗ ) + R(p∗ ) + = ∂pi k=1 " n # i X ∂λ (a∗ , p∗ ) ∂R(p∗ ) k +p∗ + = 0. ∂p ∂p i i k=1
8
(9)
3.1
Equilibrium prices and advertising intensities
Using the expressions (2)-(5), we can compute ∂λi1 (n − 1) a∗ = (1 − F (ˆ x + ∆)) ∂ai (ai + (n − 1) a∗ )2 ··· "
# " k−1 `−1 X Y ai ∂λik − (n − `) a∗ (n − m) a∗ = ∂ai ai + (n − k) a∗ `=1 (ai + (n − `) a∗ )2 m6=` ai + (n − m) a∗ # k−1 Y (n − `) a∗ (n − k) a∗ + F (ˆ x)k−1 (1 − F (ˆ x + ∆)) (ai + (n − k) a∗ )2 `=1 ai + (n − `) a∗ ··· ∂λin ∂ai
=
n−1 X
"
`=1
∗
− (n − `) A (A + (n − `) A∗ )2
n−1 Y
∗
#
(n − m) A F (ˆ x)n−1 (1 − F (ˆ x + ∆)) ∗ A + (n − m) A m6=`
In symmetric equilibrium we have ∂λi1 n−1 = 2 ∗ (1 − F (ˆ x)) ∂ai na ··· "
" # k−1 k−1 X ∂λik − (n − `) Y n − m 1 = ∂ai n − k + 1 `=1 (n − ` + 1)2 a∗ m6=` n − m + 1 # k−1 Y n−` n−i + F (ˆ x)k−1 (1 − F (ˆ x)) (n − i + 1)2 a∗ `=1 n − ` + 1 ···
" # n−1 n−1 ∂λin X − (n − `) Y n − m = F (ˆ x)n−1 (1 − F (ˆ x)) . 2 ∗ ∂ai n − m + 1 (n − ` + 1) a `=1 m6=` Note that k−1 Y `=1
n−1 n−2 n+1−k n+1−k n−` = · · ··· · = , n+1−` n n−1 n+2−k n
9
which allows us to write " µ ¶ ¸ k−1 · X ∂λik −1 n+1−k 1 (n − ` + 1) = ∂ai n − k + 1 `=1 (n − ` + 1)2 a∗ n ¸ n−k F (ˆ x)k−1 (1 − F (ˆ x)) + (n − k + 1) na∗ " # k−1 X 1 n−k 1 = − F (ˆ x)k−1 (1 − F (ˆ x)) . ∗ na n − k + 1 `=1 (n − ` + 1) and
n−1 1 X ∂λn 1 =− ∗ F (ˆ x)n−1 (1 − F (ˆ x)) ∂ai na `=1 n − ` + 1
Using these derivations and the expression for (6) above, the first order conditions in (8) and (9) can be rewritten as: ! k−1 X 1 n−k − F (ˆ x)k−1 (1 − F (ˆ x)) − 1 = 0, n − k + 1 `=1 (n − ` + 1) Z xˆ 1 − F (ˆ x)n + F (ε)n−1 f (ε)dε n 0 ¶ µ Z xˆ x)n 1 1 − F (ˆ n−2 2 n−1 ∗ F (ε) f (ε)dε − F (ˆ x) f (ˆ x) = 0. f (ˆ x) + (n − 1) +p − n 1 − F (ˆ x) 0
n X 1 ∗ p na∗ k=1
Ã
Let us denote
k−1
Ck ≡
X 1 n−k − . n − k + 1 `=1 n − ` + 1
Then these equations can be simplified to n X 1 p (1 − F (ˆ x)) Ck F (ˆ x)k−1 − 1 = 0, na∗ k=1 µ ¶ Z xˆ n 1 1 1 − F (ˆ x) ∗ n−1 0 +p − f (ˆ x) + (n − 1) F (ε) f (ε)dε = 0. n n 1 − F (ˆ x) 0 ∗
10
Solving for p∗ we obtain p∗ =
1−F (ˆ x)n f (ˆ x) 1−F (ˆ x)
−n
1 R xˆ −∞
f 0 (ε)F (ε)n−1 dε
and solving for a∗ we get: n X p∗ (1 − F (ˆ x)) Ck · F (ˆ x)k−1 , a = n k=1 ∗
Note that à Ck − Ck−1 = =
k−1
X n−k 1 − n − k + 1 `=1 n − ` + 1
!
à −
k−2
n−k+1 X 1 − n − k + 2 `=1 n − ` + 1
n−k 1 n−k+1 −1 − − = . n−k+1 n−k+2 n−k+2 n−k+1
We thus have Ck = Ck−1 −
1 , n−k+1
which implies that k−1
Ck =
n−1 X 1 − . n n − ` `=1
11
!
For the equilibrium advertising level we then have " # n k−1 X p∗ n−1 X 1 a = (1 − F (ˆ x)) − · F (ˆ x)k−1 n n n−` k=1 `=1 " # n n X k−1 X p∗ n−1X 1 = (1 − F (ˆ x)) F (ˆ x)k−1 − · F (ˆ x)k−1 n n k=1 n − ` k=1 `=1 " !# Ã n n n−1 X X n−1X p∗ 1 = (1 − F (ˆ x)) F (ˆ x)k−1 − F (ˆ x)k−1 n n k=1 n − ` `=1 k=`+1 " Ã !# n−1 n−1 n−1 ∗ X X p n−1X 1 = (1 − F (ˆ x)) F (ˆ x)k − F (ˆ x)k n n k=0 n − ` `=1 k=` " # µ ¶ n−1 ∗ ` n n X p 1 F (ˆ x) − F (ˆ x) n − 1 1 − F (ˆ x) = (1 − F (ˆ x)) · − n n 1 − F (ˆ x) n−` 1 − F (ˆ x) `=1 " # µ ¶ n−1 X ¡ ¢ 1 p∗ n − 1 n ` n−` · (1 − F (ˆ x) ) − F (ˆ x) 1 − F (ˆ x) = n n n−` `=1 ∗
Therefore: Proposition 1 Equilibrium prices and advertising are given by
p∗ a∗ = n p∗ =
4
Ã
¡ ¢! n−1 k n−k X F (ˆ x ) 1 − F (ˆ x ) 1 − F (ˆ x)n − n−k k=0
1−F (ˆ x)n f (ˆ x) 1−F (ˆ x)
−n
1 R xˆ −∞
f 0 (ε)F (ε)n−1 dε
Comparative Statics
In the previous section, we derived the equilibrium of our model, in terms of prices and advertising levels. In this section, we will look at the comparative statics effects of search costs on prices, advertising intensity, and profits. We first have the following: Proposition 2 Equilibrium prices are increasing in search costs. 12
Proof. Since in symmetric equilibrium the probability a firm is visited in kth place is equal for all firms, the proof of this result is identical to the proof in Anderson and Renault (1999). This result is intuitive: as search costs increase, a firm has more market power over the consumers who happens to venture the firm. As a result, the firm will charge a higher price if search costs increase. We also have: Proposition 3 Equilibrium advertising levels are increasing in search costs. Proof. For simplicity, we rewrite a∗ as a∗ = with
p∗ A , n
¡ ¢ n−1 X F (ˆ x)k 1 − F (ˆ x)n−k A ≡ (1 − F (ˆ x) ) − . n − k k=0 n
We take the derivative of a∗ with respect to xˆ. From (1), we immediately have that xˆ is decreasing in s. Hence, for a∗ to be increasing in s, we need it to be decreasing in xˆ. First note that Theorem 2 now implies that ∂p∗ /∂ xˆ < 0. Dropping subscripts, we have ¡ ¢ n−1 X kF k−1 1 − F n−k − (n − k) F k F n−k−1 ∂A n−1 n−1 = −nF f −F f− f ∂ xˆ n−k k=1 ¡ ¢ n−1 n−1 X X kF k−1 1 − F n−k n−1 n−1 = −nF f −F f− f+ F n−1 f n−k k=1 k=0 ¢ ¡ n−1 X kF k−1 1 − F n−k = −F n−1 f − f < 0. n−k k=1 This implies
∂a∗ A ∂p∗ p∗ ∂A = + < 0, ∂ xˆ n ∂ xˆ n ∂ xˆ which in turn implies the desired result. 13
The intuition for this result is as follows. Given that an increase in search costs confers a firm more market power over the consumers that it attracts, it becomes more and more profitable for a firm to lure consumers into its shop as search costs increase. As a result, firms will advertise with greater intensity as search costs rise. Our next result concerns the effect of an increase in search costs on equilibrium profits: Proposition 4 The effects of search costs on equilibrium profits are ambiguous. For sufficiently small search costs, profits are increasing in search costs. For sufficiently large search costs, profits are decreasing. Proof. Plugging our expression for a∗ into the profit function yields: Z xˆ n p∗ X 1 j−1 ∗ · F (ˆ x) · (1 − F (ˆ x)) + p F (ε)n−1 f (ε) dε Πi (a , p ) = n j=1 n −∞ Ã ¡ ¢! n−1 k n−k X F (ˆ x ) 1 − F (ˆ x ) p∗ − (1 − F (ˆ x)n ) − n n−k k=0 "Z ¡ ¢# n−1 k n−k x ˆ X F 1 − F 1 = p∗ F (ε)n−1 f (ε) dε + . n k=0 n−k −∞ ∗
∗
For ease of exposition, we will write this as Πi = p∗ T, with
Z
x ˆ
T ≡
n−1
F (ε) −∞
¡ ¢ n−1 1 X F k 1 − F n−k , f (ε) dε + n k=0 n−k
and p∗ is given by (??). Again, we take derivatives with respect to reservation utility xˆ. Dropping subscripts, we have · ¶¸ µ ∂p∗ 1 f 0 (1 − F ) + f 2 1 − F n n−1 =− 2 − nF ∂ xˆ p 1−F 1−F
14
and
¡ ¢ n−1 X F k−1 k − nF n−k ∂T = f. ∂ xˆ n − k k=1
To get the last derivative, we have first taken the case k = 0 outside of the summation. The derivative of this term then exactly drops out against the derivative of the first term in T. The total effect on profits is now given by ∂Π ∂p∗ ∂T = T + p∗ . ∂ xˆ ∂ xˆ ∂ xˆ Consider a case where search costs are very small: s → 0. Then xˆ → 1, so F (ˆ x) → 1. From Anderson and Renault (1999, pg. 724), we have lim p∗ =
s→0
n (n − 1)
Since
R1 0
1 f (ε)2 F (ε)n−2
> 0.
1 − Fn = n, F →1 1 − F lim
we have that
∂p∗ = 0. F →1 ∂ x ˆ lim
Also ∂T = − (n − 1) f < 0, ∂ xˆ lim T = 1,
lim
F →1
F →1
Taken together, this implies ∂Π = − (n − 1) p∗ f < 0. s→0 ∂ x ˆ lim
Now consider the case in which search costs are very high. In the most extreme case, search costs can become so high that the consumer is almost certain not to search beyond the first firm that she visits, regardless of the matching value that she encounters there. Note that this implies that the 15
consumer is still willing to visit the first firm. After her first visit, her utility is at least v − p∗ which, by assumption, is higher than the utility of 0 that she holds before her first visit. We thus consider a case in which s is so high that xˆ → 0, which implies that F (ˆ x) → 0, and lim p∗ =
F →0
1 , f (0)
lim T = 0,
F →0
¤ ∂p∗ 1 £ = − 2 f0 + f2 , F →0 ∂ x ˆ p ∂T 1 lim = f > 0. F →0 ∂ x ˆ n−1 lim
Taken together, this implies lim
F →0
∂Π 1 = > 0. ∂ xˆ n−1
Hence for small enough s, equilibrium profits are increasing in search costs. For large enough s, they are decreasing. This establishes the result. We now elaborate on the intuition behind this result. An increase in search costs has two opposite effects on firm profits. First, with an increase in s, firms gain market power, which allows them to charge a higher price. Yet, this also implies that it becomes more attractive for a single firm to gain the battle for attention so firms advertise more as search costs increase. When search costs are small, the price effect has a dominating influence and firms gain from an increase in search costs. Advertising is a rent-seeking activity which leads to a dissipation of the rents generated by market power. When search costs are large, this negative advertising effect dominates the price effect and profits decrease with higher search costs. In our model, lowering search always increases welfare. As we assume that the market is always covered, total welfare is maximized if the costs of advertising is minimized. From Proposition 3, we know this to be the case if search costs are zero. If we were to consider a case in which industry demand is not completely inelastic, this result will only be reinforced, as lower search
16
costs imply lower prices and hence a lower deadweight loss. Interestingly, when search costs are sufficiently high, it would even be a Pareto improvement to have lower search costs. Consumers are better off as equilibrium prices decrease, while firms are better off as equilibrium profits increase.
5
Conclusion
In this paper, we have modelled the idea that, in an attempt to being visited as early as possible in the course of search of a consumer, firms engage themselves in a battle for attention. Through investments in more and more appealing advertising, a firm can achieve a salient place in consumer awareness so that consumers will visit this firm sooner when searching for a product they need. Advertising is not a winner-takes-all contest in our setting: when a consumer does come to a firm first, she can still decide to go to a different firm if she does not like the product of this particular firm, or if she thinks it is too expensive. We found that prices and advertising levels are increasing in consumers’ search costs. Yet, the effect on profits is ambiguous. If search costs are small to start with, then firms are better off if search costs increase. Instead, when search costs are already high a further increase in search costs lowers firm profits. In the latter case, getting the attention of a consumer becomes so important that firms over-dissipate the rent that is generated by a visiting consumer. Our model highlights the importance of looking at the interaction of advertising and search costs, rather than only looking at search costs in isolation. Increased search costs not only allow firms to increase prices, but also induce them to increase their efforts in trying to attract consumers in the first place. As a result, firms may very well be hurt if search costs increase. We believe this to be a general phenomenon, that applies beyond the scope of this particular model.
17
References Anderson, S., and R. Renault (1999): “Pricing, Product Diversity, and Search Costs: a Bertrand-Chamberlin-Diamond model,” RAND Journal of Economics, 30, 719–735. Arbatskaya, M. (2007): “Ordered Search,” RAND Journal of Economics, p. forthcoming. Bagwell, K. (2007): “The Economic Analysis of Advertising,” in Handbook of Industrial Organization, ed. by M. Armstrong, and R. Porter, Handbooks in Economics, chapter 2. North-Holland, Amsterdam. Butters, G. (1977): “Equilibrium Distribution of Prices and Advertising,” Review of Economic Studies, 44, 465–492. Comanor, W., and T. Wilson (1974): Advertising and Market Power. Harvard University Press, Cambridge, MA. Joseph W. Alba, A. C. (1986): “Salience Effects in Brand Recall,” Journal of Marketing Research, 23(4), 363–369. Perloff, J., and S. Salop (1985): “Equilibrium with Product Differentiation,” Review of Economic Studies, 52, 107–120. PriceWaterhouseCoopers (2005): “Global Entertainment and Media Outlook: 2006-2010,” Discussion paper. Schmalensee, R. (1976): “A Model of Promotional Competition in Oligopoly,” The Review of Economic Studies, 43(3), 493–507. Tullock, G. (1980): “Efficient Rent Seeking,” in Toward a Theory of the Rent Seeking Society, ed. by G. T. J.M. Buchanan, R. Tollison, pp. 224– 232. Texas A&M Press.
18
