650 likes | 742 Views
Product Variety and Quality. Introduction. Most firms sell more than one product Products are differentiated in different ways horizontally goods of similar quality targeted at consumers of different types how is variety determined? is there too much variety vertically
E N D
Introduction • Most firms sell more than one product • Products are differentiated in different ways • horizontally • goods of similar quality targeted at consumers of different types • how is variety determined? • is there too much variety • vertically • consumers agree on quality • differ on willingness to pay for quality • how is quality of goods being offered determined?
Modeling horizontal differentiation • Address models • Consumers have preferences over the characteristics of products • Monopolistic competition model • Consumers have preferences over goods and a taste for variety
Horizontal product differentiation • Suppose that consumers differ in their tastes • firm has to decide how best to serve different types of consumer • offer products with different characteristics but similar qualities • This is horizontal product differentiation • firm designs products that appeal to different types of consumer • products are of (roughly) similar quality • Questions: • how many products? • of what type? • how do we model this problem?
A spatial approach to product variety • The spatial model (Hotelling) is useful to consider • pricing • design • variety • Has a much richer application as a model of product differentiation • “location” can be thought of in • space (geography) • time (departure times of planes, buses, trains) • product characteristics (design and variety) • consumers prefer products that are “close” to their preferred types in space, or time or characteristics
A Spatial approach to product variety 2 • Assume N consumers living equally spaced along Main Street – 1 mile long. • Monopolist must decide how best to supply these consumers • Consumers buy exactly one unit provided that price plus transport costs is less than V. • Consumers incur there-and-back transport costs of t per mile • The monopolist operates one shop • reasonable to expect that this is located at the center of Main Street
Suppose that the monopolist sets a price of p1 The spatial model Price Price p1 + tx p1 + t.x V V All consumers within distance x1 to the left and right of the shop will by the product t t What determines x1? p1 z = 0 x1 x1 z = 1 1/2 Shop 1 p1 + tx1 = V, so x1 = (V – p1)/t
Suppose the firm reduces the price to p2? The spatial model 2 Price Price p1 + t.x p1 + t.x V V Then all consumers within distance x2 of the shop will buy from the firm p1 p2 z = 0 x2 x1 x1 x2 z = 1 1/2 Shop 1
The spatial model 3 • Suppose that all consumers are to be served at price p. • The highest price is that charged to the consumers at the ends of the market • Their transport costs are t/2 : since they travel ½ mile to the shop • So they pay p + t/2 which must be no greater than V. • So p = V – t/2. • Suppose that marginal costs are c per unit. • Suppose also that a shop has set-up costs of F. • Then profit is p(N, 1) = N(V – t/2 – c) – F.
Monopoly pricing in the spatial model • What if there are two shops? • The monopolist will coordinate prices at the two shops • With identical costs and symmetric locations, these prices will be equal: p1 = p2 = p • Where should they be located? • What is the optimal price p*?
V V Location with two shops Delivered price to consumers at the market center equals their reservation price Suppose that the entire market is to be served Price Price If there are two shops they will be located symmetrically a distance d from the end-points of the market p(d) The maximum price the firm can charge is determined by the consumers at the center of the market p(d) What determines p(d)? Now raise the price at each shop Start with a low price at each shop d 1/2 1 - d z = 0 z = 1 Shop 1 Shop 2 Suppose that d < 1/4 The shops should be moved inwards
V V Location with two shops 2 Delivered price to consumers at the end-points equals their reservation price The maximum price the firm can charge is now determined by the consumers at the end-points of the market Price Price p(d) p(d) Now what determines p(d)? Now raise the price at each shop Start with a low price at each shop d 1/2 1 - d z = 0 z = 1 Shop 1 Shop 2 Now suppose that d > 1/4 The shops should be moved outwards
V V Location with two shops 3 It follows that shop 1 should be located at 1/4 and shop 2 at 3/4 Price at each shop is then p* = V - t/4 Price Price V - t/4 V - t/4 Profit at each shop is given by the shaded area c c z = 0 1/4 1/2 3/4 z = 1 Shop2 Shop 1 Profit is now p(N, 2) = N(V - t/4 - c) – 2F
V V By the same argument they should be located at 1/6, 1/2 and 5/6 What if there are three shops? Three shops Price Price Price at each shop is now V - t/6 V - t/6 V - t/6 z = 0 1/6 1/2 5/6 z = 1 Shop 1 Shop 2 Shop 3 Profit is now p(N, 3) = N(V - t/6 - c) – 3F
Optimal number of shops • A consistent pattern is emerging. • Assume that there are n shops. • They will be symmetrically located distance 1/n apart. How many shops should there be? • We have already considered n = 2 and n = 3. • When n = 2 we have p(N, 2) = V - t/4 • When n = 3 we have p(N, 3) = V - t/6 • It follows that p(N, n) = V - t/2n • Aggregate profit is then p(N, n) = N(V - t/2n - c) – nF
Optimal number of shops 2 Profit from n shops is p(N, n) = (V - t/2n - c)N - nF and the profit from having n + 1 shops is: p*(N, n+1) = (V - t/2(n + 1)-c)N - (n + 1)F Adding the (n +1)th shop is profitable if p(N,n+1) - p(N,n) > 0 This requires tN/2n - tN/2(n + 1) > F which requires that n(n + 1) < tN/2F.
An example Suppose that F = $50,000 , N = 5 million and t = $1 Then tN/2F = 50 For an additional shop to be profitable we need n(n + 1) < 50. This is true for n< 6 There should be no more than seven shops in this case: if n = 6 then adding one more shop is profitable. But if n = 7 then adding another shop is unprofitable.
Some intuition • What does the condition on n tell us? • Simply, we should expect to find greater product variety when: • there are many consumers. • set-up costs of increasing product variety are low. • consumers have strong preferences over product characteristics and differ in these • consumers are unwilling to buy a product if it is not “very close” to their most preferred product
Empirical Application: Price Discrimination and Imperfect Competition Although we have presented price discrimination and product design (versioning) issues in the context of a monopoly, these same tactics also play a role in more competitive settings of imperfect competition Imagine a two-store setting again Assume N customers distributed evenly between the two stores, each with maximum willingness to pay of V . No transport cost—Half of the consumers always buys at nearest store. Other half always buys at cheapest store.
Price Discrimination and Imperfect Competition 2 If both stores operated by a monopolist, set price = V. Cannot set it higher of there will be no customers. Setting it lower though gains nothing. What if stores operated by separate firms? Imagine P1 = P2 = V. Store 1 serves N/4price-sensitive customers and N/4 price-insensitive ones. The same is true for Store 2. If Store 1 cuts its price below V. It loses N/2 from all current customers It gains N(V - )/4 by stealing all price-sensitive customers from Store 2
Price Discrimination and Imperfect Competition 3 MORAL 1: Both firms have a real incentive to cut price. This ultimately proves self-defeating In equilibrium, both still serve N/2 customers but now do so at a price closer to cost. This is especially frustrating in light of the “brand-loyal” or price-insensitive customers Cutting their price does not increase their likelihood of shopping at a particular place. It just loses revenue. MORAL 2: Unlike the monopolist who sets the same price to everyone, these firms have an incentive to discriminate and so continue to charge a high price to loyal consumers while pricing low to others.
Price Discrimination and Imperfect Competition 4 The intuition then is that price discrimination may be associated with imperfect competition and become more prominent as markets get more competitive (but still less than perfectly competitive). This idea is tested by Stavins (2001) with airline prices. Restrictions such as a required Saturday night stay-over or an advanced purchase serve as screening mechanism for price-sensitive customers. Hence, restrictions lead to lower ticket price. Stavins (2001) idea is that price reduction associated with flight restrictions will be small in markets that are not very competitive.
Price Discrimination and Imperfect Competition 5 Stavins (2001) looks at nearly 6,000 tickets covering 12 different city-pair routes in September, 1995. She finds strong support for the dual hypothesis that: a) passengers flying on a ticket with restrictions pay less; b) price reduction shrinks as concentration rises In highly competitive (low HHI) markets, a Saturday night restriction leads to a $253 price reduction but only a $165 reduction in less competitive ones. In highly competitive (low HHI) markets, an Advance Purchase restriction leads to a $111 price reduction but only a $41 reduction in less competitive ones.
Monopoly and product quality • Firms can, and do, produce goods of different qualities • Quality then is an important strategic variable • The choice of product quality determined by its ability to generate profit; attitude of consumers to q uality • Consider a monopolist producing a single good • what quality should it have? • determined by consumer attitudes to quality • prefer high to low quality • willing to pay more for high quality • but this requires that the consumer recognizes quality • also some are willing to pay more than others for quality
Choosing quality • Quality – vertical attributes of a product (ALL consumers agree that a product X is of higher quality than another product Y) • Firms choose both quantity and quality • Profit maximization • MR of an increment of quality is equal to its MC • Key problem: asymmetric information
Asymmetric information • Lemons problem • How would you describe an equilibrium? • Why is this a problem? • Adverse selection • Moral hazard • One side of a transaction has an incentive to change the terms of the exchange, unobserved by the other side
Asymmetric information: types of goods • Search goods: consumers have sufficient everyday knowledge or can accurately predict quality of a product BEFORE purchase • Experience goods: quality can be determined by consumers only AFTER purchase
Search goods • Why manufacturers and service companies do not provide a moderate quality at a moderate price (why price/quality ratio rare works in real life)? • Aim: increase profit by capturing consumer surplus (we assume imperfect competition here) • Mechanism: quality discrimination (Highest quality level is chosen solely on grounds of independent profit maximization, but a firm needs preventing switching high-end consumers to low-end products) • versioning / damaged goods • widening the range of quality offered
Experience goods: strategies • Reputation • Aim: repeat sales (customer loyalty) • Reputation is transferrable across consumers (“word of mouth” as a promotion tool) and across markets (exploiting established reputation in new markets) • Commitment • warranty (how to use for quality discrimination?) • reputation (why restaurants in tourist areas are so bad?) • Investment as a form of commitment: what is a signal of quality in this case?
Modeling commitment • “Pure reputation model” (Shapiro, 1983) • Only new products are of unknown quality in the first period • Consumers learn true quality after purchase and inform other potential buyers • Decision: investment in reputation in the first period vs. earning a rent from selling high-quality products in all subsequent periods • Assume that companies don’t “milk” their reputation • “Advertising models” (Nelson, 1970, 1978) • Decision: price / advertising expenses combination to prevent entry of low-quality producers • Why “burning money” / noninformative advertising exists? (Warning: explanation for new experience goods only)
Price and advertising as signals of quality • Higher price – Higher Quality • Theory: Yes • More advertising – High Quality • Theory: it depends (on the cost of information)
Empirical testing • Caves and Greene (1996) • Source of information: Consumer Reports (ranking products by objective characteristics – use as measures of quality) • Method: correlation analysis • Findings: • rank correlation coefficient for price-quality 0.38 for list prices, 0.27 for transaction prices • Is it a strong or weak correlation?
Why weak? • Price-quality correlation is higher for product categories that include more brands (greater scope for vertical differentiation) • Lower for “convenience goods” (heavy advertising and frequent repeat purchase) • Weakest for product that can use image advertising to build customer loyalty (horizontal differentiation)
Testing Nelson model • Conclusion: quality signaling is not a particularly important determinant of advertising in consumer goods (median values of the rank correlations are close to zero) • In some product categories correlation in very strong positive, in some – strong negative) • Advertising outlays tend to increase with quality for innovative goods (providing information) • Advertising is less correlated with quality of convenience goods (horizontal differentiation)
Introduction • Firms often bundle the goods that they offer • Microsoft bundles Windows and Explorer • Office bundles Word, Excel, PowerPoint, Access • Bundled package is usually offered at a discount • Bundling may increase market power • GE merger with Honeywell • Tie-in sales ties the sale of one product to the purchase of another • Tying may be contractual or technological • IBM computer card machines and computer cards • Kodak tie service to sales of large-scale photocopiers • Tie computer printers and printer cartridges • Why? To make money!
Bundling: an example How much can be charged for Godzilla? How much can be charged for Casablanca? If the films are sold separately total revenue is $19,000 • Two television stations offered two old Hollywood films • Casablanca and Son of Godzilla • Arbitrage is possible between the stations • Willingness to pay is: $7,000 Willingness to pay for Casablanca Willingness to pay for Godzilla $2,500 Station A $8,000 $2,500 Station B $7,000 $3,000
How much can be charged for the package? Bundling is profitable because it exploits aggregate willingness pay Bundling: an example 2 Now suppose that the two films are bundled and sold as a package If the films are sold as a package total revenue is $20,000 Willingness to pay for Casablanca Willingness to pay for Godzilla Total Willingness to pay Station A $8,000 $2,500 $10,500 Station B $7,000 $3,000 $10,000 $10,000
Bundling • Extend this example to allow for mixed bundling: offering products in a bundle and separately
Mixed bundling • What should a firm actually do? • There is no simple answer • mixed bundling is generally better than pure bundling • but bundling is not always the best strategy • Each case needs to be worked out on its merits
An Example Four consumers; two products; MC1 = $100, MC2 = $150 Reservation Price for Good 1 Reservation Price for Good 2 Sum of Reservation Prices Consumer A $50 $450 $500 B $250 $275 $525 $300 $220 $520 C D $450 $50 $500
The example 2 Good 1: Marginal Cost $100 Consider simple monopoly pricing Price Quantity Total revenue Profit $450 1 $450 $350 $300 2 $400 $600 Good 1 should be sold at $250 and good 2 at $450. Total profit is $450 + $300 = $750 $250 $250 3 $750 $450 $50 4 $200 -$200 Good 2: Marginal Cost $150 Price Quantity Total revenue Profit $450 $450 1 $450 $300 $275 2 $200 $550 $220 3 $660 $210 $50 4 $200 -$400
The example 3 Now consider pure bundling Reservation Price for Good 1 Reservation Price for Good 2 Sum of Reservation Prices Consumer The highest bundle price that can be considered is $500 All four consumers will buy the bundle and profit is 4x$500 - 4x($150 + $100) = $1,000 A $50 $450 $500 B $250 $275 $525 $300 $220 $520 C D $450 $50 $500
The example 4 Now consider mixed bundling Take the monopoly prices p1 = $250; p2 = $450 and a bundle price pB = $500 All four consumers buy something and profit is $250x2 + $150x2 = $800 Reservation Price for Good 1 Reservation Price for Good 2 Sum of Reservation Prices Consumer Can the seller improve on this? A $50 $450 $500 $500 B $250 $275 $500 $525 $300 $250 $220 $520 C D $250 $450 $50 $500
The example 5 This is actually the best that the firm can do Try instead the prices p1 = $450; p2 = $450 and a bundle price pB = $520 All four consumers buy and profit is $300 + $270x2 + $350 = $1,190 Reservation Price for Good 1 Reservation Price for Good 2 Sum of Reservation Prices Consumer A $50 $450 $450 $500 B $250 $275 $520 $525 $300 $220 $520 $520 C D $450 $450 $50 $500
Bundling again • Bundling does not always work • Mixed bundling is always more profitable than pure bundling • Mixed bundling is always better than no bundling • But pure bundling is not necessarily better than no bundling • Requires that there are reasonably large differences in consumer valuations of the goods • Bundling is a form of price discrimination • May limit competition
Tie-in sales • What about tie-in sales? • “like” bundling but proportions vary • allows the monopolist to make supernormal profits on the tied good • different users charged different effective prices depending upon usage • facilitates price discrimination by making buyers reveal their demands
Tie-in sales 2 • Suppose that a firm offers a specialized product – a camera – that uses highly specialized film cartridges • Then it has effectively tied the sales of film cartridges to the purchase of the camera • this is actually what has happened with computer printers and ink cartridges • How should it price the camera and film? • suppose also that there are two types of consumer, high-demand and low-demand, with one-thousand of each type • high demand P = 16 – Qh; low demand P = 12 - Ql • the company does not know which type is which