IN THE UNITED STATES DISTRICT COURT
DAVID REITMAN. Ph.D. DATED MARCH 31, 2000
Dated: June 14, 2000
COUNSEL FOR PLAINTIFF
Judith M. Kinney (DSB #3643)
Mark J. Botti
Rebuttal Expert Report
My earlier expert report analyzed the efficiency claims that Dentsply had made in its White Paper and its presentations to the Antitrust Division. I concluded that the efficiencies from the exclusionary provision of the Dealer Criteria were negligible. In his expert report, Prof Marvel reiterates one of the efficiency arguments espoused by Dentsply regarding free riding on manufacturer pre-sale services (pages 13-22 of his report), and develops a new argument based on tooth returns (pages 22-28 of his report), though he does not quantify the magnitude of these efficiencies. Having reviewed Prof. Marvel's report, I continue to believe that the efficiency claims are negligible for the reasons addressed in my earlier report, and that any efficiency benefits from the exclusionary provision of the Dealer Criteria cannot justify the anticompetitive harm identified in my expert report.
This rebuttal report responds to several issues raised in Prof. Marvel's expert report, but is not intended to be a comprehensive response to his opinions.1 Instead, I focus on several arguments related to efficiencies introduced by Prof. Marvel that I believe to be incorrect. I may supplement my opinions based on defendant's rebuttal reports as well as any additional information that becomes available. In that regard, I will continue to review the record, including evidence developed since the date of my initial report, prior to my testimony.
B. Do Dealers Switch Customers?
Both the pre-sale services and the tooth return efficiency rationales in Prof. Marvel's expert report rely on dealers being able to switch customers from one brand of artificial teeth to another brand from which the dealer earns more money. However, the evidence that dealers could or would switch customers is weak, at best. Various dealers state that they do not attempt to switch customers from one brand to another.2 Dealers try to provide whatever products will best satisfy their customers' needs. When a lab customer calls in or transmits a request for a card of a particular brand, mold, and shade of tooth, the lab itself would know best if that tooth card would be most appropriate for the corresponding denture case. Consequently, labs have no incentive to adopt alternative tooth suggestions from dealers. Moreover, there is a risk that the customer would not be satisfied with the suggested alternative tooth, if the suggestion were accepted, and would possibly stop buying from the dealer as a result of what was perceived as poor service. As Michael Crane, formerly Senior vice president at Dentsply, testified, "What I'm saying is that there was no incentive to do so [switch customers from one brand of tooth to another] because if they're already buying their products from the dealer, what incentive would there be to change them from one brand to another brand? As a matter of fact, there's a huge risk in doing so because you may make your customer unhappy. And if they're unhappy with that specific product, they're unhappy with all the business that is processed through that sales rep and that dealer." [Crane at 30-31]
If dealers can not or will not switch customers from one brand to another, then the only direct effect that adding additional brands has on sales of Dentsply teeth is the procompetitive result that some customers who would now have access to other brands through their preferred distribution channel choose to buy more of the newly added brands and fewer Dentsply teeth. Thus the inability or unwillingness of dealers to switch customers from one brand of artificial teeth to another is enough to negate these efficiency claims. However, even if one were willing to assume that dealers have some control over the sales of different brands of teeth, there are a number of other problems with both of Prof. Marvel's efficiency rationales that are discussed, in turn, in the rest of this report.
C. Prof. Marvel's efficiency arguments
1. Free Riding on Manufacturer Pre-sale Services
One premise of Prof. Marvel's pre-sale services argument is that Dentsply must charge a price premium to cover the cost of promotion and training. This is not true. Since Dentsply earns positive margins on its tooth sales, any additional sales generated by its promotional efforts or from having better trained dealers would lead to incremental profits for Dentsply. Promotional and training expenditures would be justified if they bring in enough additional sales to cover their costs. As Prof. Marvel states, the size of the sales force and training efforts as well as the budget for other promotional activities are fixed expenses with regard to Dentsply's pricing decisions for teeth. But fixed costs generally affect pricing only in a long run competitive equilibrium, in which firms must cover their fixed costs in order to stay in the market. Given the substantial profits earned by the Trubyte division, it is clear that exit considerations do not play a role in the pricing of Dentsply teeth. Apart from such entry and exit dynamics, firms set prices based on their variable costs of supplying output as well as demand conditions for their product, but not on their fixed costs.3 Thus if Dentsply raises prices as a result of its sales, marketing, and training efforts, it is only because these efforts have increased Dentsply's market power in the tooth market, enabling it to profitably set a higher price for teeth.4
Even if one accepts Prof. Marvel's assertion that Dentsply must charge a price premium to cover these promotional and training expenditures, it does not follow that Dentsply will have higher prices than its tooth competitors. The Trubyte division does spend considerable money on promotion and training, more than most, if not all, of its tooth competitors.5 But Dentsply's tooth sales are roughly ten times as large as those of its next largest competitor, so it is no great surprise that its expenditures are larger. The relevant measure for Prof. Marvel's argument is the expenditure on promotion and training per unit of tooth sales, since that average cost is what would be recovered through a price premium.6 However, Prof. Marvel has not attempted to quantify the expenditures on pre-sale services per volume of sales. On a per dollar basis, Dentsply's Trubyte expenditures on these services do not appear to be larger than those of its competitors. Judging from one measure, which is the number of sales representatives per dollar of sales, Trubyte expenses are less than some of its competitors, and perhaps on par with others. Thus, even if it made sense that firms must charge a price premium to cover their promotional and training expenses, it is not evident that Dentsply's premium would be any greater than that of its competitors.
Finally, Prof. Marvel's pre-sales services argument presumes that these services are efficient because they bring new customers or additional sales into the market. If there is no change in the market output of artificial teeth as a result of these services, then there is no clear efficiency associated with providing them. But much of Dentsply's promotional activity is directed toward switching customers from other brands of teeth to its own Trubyte brand, rather than increasing output.7 Thus even if the other conditions for Prof. Marvel's pre-sales services argument were satisfied, and even if Dentsply were to curtail some of these services in the absence of Dealer Criterion six, any resulting efficiency losses in the economy are likely to be minimal.
This discussion has focused on four places where the manufacturer pre-sale services argument falls short: Dentsply need not charge a price premium to recover its promotional and training expenditures; even if it does charge a premium, it is not clear that its premium would be any higher than that of other tooth competitors for their promotional and training expenditures; even if Dentsply did charge a higher premium, dealers would not steer customers to brands of teeth that offer the dealer greater returns; and even if Dentsply did reduce its pre-sale services in the absence of the exclusionary provision of the Dealer Criteria, there may not be much inefficiency in the resulting market allocation. In addition, my earlier report discussed two other weaknesses in the free riding argument: much of Dentsply's promotional and training expenditures are not free rideable by other manufacturers, and the evidence suggests that Dentsply does not vary its investment in these expenditures depending on the opportunity for competing brands to free ride on them. Therefore I conclude that the Prof. Marvel's pre-sale services efficiency argument can not justify the exclusionary provision of the Dealer Criteria.
2. Tooth Returns
Prof. Marvel's tooth returns efficiency argument also relies on dealers' alleged ability and willingness to steer customers from more expensive Dentsply teeth to a cheaper brand. However, for this argument, Prof. Marvel's explanation for why Dentsply teeth would be more expensive is that Dentsply teeth would be returned more, and that a higher proportion of returned teeth leads to greater costs of supplying teeth.
Teeth are returned for two reasons. One is that broken sets of teeth are returned when some of the teeth on a card are used to make a partial denture. The second reason is that dealers may adjust their inventory by returning slow moving cards of teeth to the manufacturer. Prof. Marvel asserts that Dentsply has a more favorable returns policy than its competitors, and that dealers would steer returns to Dentsply by stocking only more popular lines of competitive brands and by selling Dentsply teeth when the teeth will be used for partial dentures.
There are a number of problems with this argument. First of all, Dentsply imposes restrictions on returns (such as "like for like" tooth credits and a two for one limitation) that raise the cost of returns to dealers. It is not clear that it is, in fact, cheaper for dealers to handle Dentsply tooth returns versus returns to other tooth suppliers.8
With regard to broken sets of teeth, dealers generally have no way of knowing which of the tooth cards that they sell will be used for full dentures and which will be used for partials. As a result, the main driver for the amount of returns of broken sets for a given brand is the overall fraction of usage of that tooth line for full versus partial dentures. However, Dentsply teeth tend to be used relatively more frequently for full dentures, while other brands typically are used relatively more for partials.9 Thus other brands would tend to have a greater proportion of broken sets of teeth to handle, and would have higher tooth returns costs than Dentsply.
As for returns of full sets, it is somewhat unclear from Prof. Marvel's report whether or not he assumes that dealers are able to anticipate which are the slow moving cards of teeth. But either way, the argument does not work. If dealers know which cards will be slow moving, then they would choose their inventory accordingly and not need to return full sets. If there is some uncertainty, then a dealer who carries the full line of Dentsply teeth may need to adjust inventory somewhat as it learns what cards or lines are more popular. But returns would be more of an issue for competing brands since, under Prof. Marvel's assumption, the dealer wants to only carry the most popular cards of those brands. As the dealer learns which cards are more popular, it would need to send back the entire inventory of unpopular cards in exchange for unstocked cards that turn out to be popular, rather than making a more moderate adjustment of the relative inventory levels, as it would for Dentsply teeth. Thus once again, returns would be relatively more frequent with the added competitive lines then with Dentsply teeth. Prof. Marvel seems to want to assume that dealers are uncertain about which Dentsply cards are more popular, but completely certain about which cards of competitive brands are more popular. If anything, the reverse would be true, since the dealers would have been selling Dentsply teeth to their customers for a number of years, but would have only just added competing lines of teeth. Thus, for both full sets and broken sets of teeth, newly added competitive lines of teeth would have a greater share of returned teeth, and therefore higher costs, than Dentsply would.
Even if it were true that Dentsply teeth were returned more than competitive brands, and even if it were true that Dentsply offered a more favorable returns policy than its competitors, and even if it were true that dealers could switch customers from Dentsply teeth to competitive brands, there is a much simpler solution to the problem than imposing exclusionary restrictions on dealers; this solution would also be procompetitive rather than anticompetitive. Dentsply could simply charge dealers for returns directly, rather than bundling the returns cost into the price of teeth. Dealers would then have the option of whether to pass along a tooth returns charge to their customers. Prof. Marvel correctly states that this policy could raise the price of teeth used for partial dentures relative to the price of teeth used for full dentures. However, this price change would reflect more accurately the cost of providing teeth for each type of denture. Teeth used for partial dentures are more costly to provide to customers due to the handling costs associated with broken sets of teeth. Under the current bundled pricing policy, full denture patients are subsidizing partial denture patients because the price paid for full denture teeth includes an average return cost which overstates the cost of handling returns for full dentures. By charging for returns separately, the price of full dentures would fall relative to the price of partial dentures. To the extent that the subsidy under the bundled pricing scheme distorts the allocation of resources in the economy between full and partial dentures, removing this subsidy by charging prices for teeth more in line with their actual costs would increase, not decrease, the efficiency of the economy.
My earlier report discussed Dentsply's policies towards dealers currently authorized to carry its teeth in the context of the free riding efficiency argument. Since some dealers currently carry various grandfathered competitive brands while other dealers sell Dentsply teeth exclusively, any free riding that was possible when dealers add competitive brands would already be occurring at the dealers carrying grandfathered brands. Dentsply currently has a Trubyte Exclusivity Program for dealers that carry only Dentsply teeth. However, Dentsply makes no attempt to differentiate the allegedly free rideable services offered to exclusive versus non-exclusive dealers, either in the context of this program or through any other means. This suggests that free riding cannot be a very serious problem in the distribution of artificial teeth.
The same argument applies to the tooth returns efficiency claim. If tooth returns cause a distortion in the behavior of those dealers that carry multiple brands of teeth, then Dentsply could impose additional charges or restrictions on returns for nonexclusive dealers (while still requiring all of its dealers to operate the same liberal returns policy with respect to their customers). The changes need not raise the overall cost to dealers, since, by unbundling some of the returns cost from the price of teeth, the price of teeth could be lowered accordingly. These additional charges or restrictions could simply be incorporated into the existing Trubyte Exclusivity Program, or else handled separately. The fact that Dentsply makes no attempt to differentiate among its dealers with regard to tooth returns suggests that there are no significant efficiency concerns about the way that dealers who carry multiple brands of teeth handle tooth returns.
Finally, as a variation on the tooth returns rationale, Prof. Marvel argues that "dealers that can promise more inventory, 'turns' are in a position to negotiate more favorable treatment from their manufacturer-suppliers." This, too, is wrong. Assuming that dealers own their inventory of teeth, suppliers only care about the volume of sales through that dealer. The average amount of inventory on hand at the dealer used to sustain that level of sales is irrelevant to the supplier. Thus if there is any variation in a dealer's bargaining power, it is due to sales volume, not inventory turns. But even if dealers add competitive brands, Dentsply would likely continue to have greater sales through those dealers than its competitors for the time being. Thus, this argument actually suggests that dealers would if anything pay less, not more, for Dentsply teeth. To the extent that dealers switch customers among brands at all, their incentive would be to switch customers from competing brands to Dentsply. Therefore, once again, this argument could not be a justification for Dentsply's exclusionary policies.
4 Dentsply could recover some or all of its promotional and training expenditures through a fee levied periodically on dealers (this could be collected on a monthly, quarterly, or yearly basis). Prof. Marvel asserts that such a fee would not work due to "ex post opportunism," but it is not at all clear what opportunism he has in mind, given the ongoing relationship between Dentsply and the dealers that Dentsply has authorized to carry its teeth, and also given the possibility of making the fee collection period relatively short.
5 However, note that the relevant comparison for this efficiency analysis is what Dentsply and its competitors would spend in the absence of the exclusionary provision of the dealer criteria, not what they are currently spending.
6 Depending on whether the price premium is thought of as cents per tooth or a percentage increase in tooth prices, the appropriate units for tooth sales in this calculation would be, respectively, number of teeth sold or dollar sales of teeth.
8 Prof. Marvel argues in a footnote that a similar result holds if returns policies are the same across brands because there are fixed costs in handling returns. However, this argument assumes that a dealer could completely eliminate returns for the newly added competitive lines of teeth. Since some returns are inevitable for all but the cheapest lines of denture teeth, this argument does not apply.
9 See Thumim at 190 and Turner at 25-26. Vita and Ivoclar teeth are favored for partials because their shades match existing crown or bridge work using porcelain in Vita or Ivoclar shades. The same would be true for other tooth brands that use Vita shades. Austenal is one of the leading manufacturers of partial framework systems, and may derive additional sales for partial denture teeth through that connection.
AFFIDAVIT OF SERVICE
I, Roberta J. Nava, an employee in the Office of United States Attorney for the District of Delaware, hereby attest under penalty of perjury that on June 14,2000, two copies of the foregoing REBUTTAL EXPERT REPORT OF DAVID REITMAN, Ph.D. DATED MARCH 31,2000 were served, as indicated, upon:
BY Hand Delivery
BY Hand Delivery (served by Nana D. Osafo, U.S. Department of Justice, Washington, D.C.)
AFFIDAVIT OF SERVICE
I, Nana D. Osafo, an employee of the U.S. Department of Justice, Antitrust Division, 325 Seventh Street, N.W., Suite 400, Washington, D.C. 20530, hereby attest under penalty of perjury that on June 14, 2000, two copies of the foregoing REBUTTAL EXPERT REPORT OF DAVID REITMAN, Ph.D. DATED MARCH 31, 2000 were served, as indicated, upon:
BY Hand Delivery