[Mike Scherer, the Aetna Professor Emeritus at the John F. Kennedy School of Government, Harvard University, is a member of the Advisory Board of the American Antitrust Institute. The AAI transmitted this memorandum to the D.O.J. prior to issuance of a second request to the merging parties.]
MEMORANDUM BY F. M. SCHERER ON THE PROPOSED YELLOW-ROADWAY MERGER
July 22, 2003
1. This memorandum offers some thoughts on possible anti- competitive aspects of the proposed merger between Yellow Corporation and Roadway Corporation, announced July 8, 2003.
2. The author has written extensively on the economics of mergers and the relationship between market structure and economic performance. I have no background of research on the transport- ation industries, except to the extent that their services are inputs in determining economies of scale relationships in other industries. However, during the 1940s and 1950s I was a part-time and (during summers) full-time employee of Scherer Freight Lines Inc., which served a territory roughly bounded by Chicago, Milwaukee, Rock Island, Peoria, Decatur, and St. Louis. My assignments included working as rate clerk, billing clerk, dispatcher, shipment tracer, truck loader, peddle route helper, and (in the summer of 1957) principal sales representative for north central Illinois. From all this I retain a basic understanding of how the motor freight industry is organized and functions.
3. Since that time the economics of the trucking industry have changed significantly with deregulation under the Motor Carrier Act of 1980 and subsequent legislation. Both entry and rate-setting have been freed, and rates have on average fallen. See e.g. John S. Ying and Theodore Keeler, "Pricing in a Deregulated Environment: The Motor Carrier Experience," RAND Journal of Economics, Summer 1991, pp. 264-273. Before deregulation, the task of a rate clerk was to locate the commodity to be transported in a book the size of a metropolitan telephone directory and then combine the rating characteristics with other formulaic aspects of the shipment -- e.g., weight, distance, territory, etc. The published rate parameters were worked out cooperatively in regional "rate bureaus" and, at least in principle, were identical for all common inter-city truck carriers with authorization to serve a given territory.
4. What initially triggered my interest upon announcement of the proposed Roadway and Yellow merger was curiosity as to how truck transportation rates are now set under deregulation. What gives the question salience for one who has written extensively on the economics of pricing is the complexity of the rate-setting problem. Following the pioneering monograph by Charles Ellet Jr., An Essay on the Laws of Trade in Reference to Works of Internal Improvement in the United States (1839, yes, 18), transportation services are priced in the first instance according to "what the market will bear" -- i.e., the more valuable the commodity per hundredweight, the higher the rate on a given shipment, all else equal. In the old rate books, there must have been thousands of different named commodity ratings. Further variables that must be factored into the pricing equation are the distances between the origin and destination points, local traffic congestion, the weight and cubage of the shipment (with a basic distinction between "truckload" (TL) and "less-than-truckload" (LTL) volumes), whether the shipment must be interlined (i.e., transferred from one carrier to another), whether the origin and destination points have loading and unloading from truck-height docks, fragility and perishability, and claims experience, among other things. What strikes me is that the pricing problem is in principle extraordinarily complex -- about as complex as any pricing problem I have studied in my many years as an academic industrial organization economist. Under regulation, it was solved by applying crude rules of thumb. I cannot claim to know, but wonder, how it is solved under deregulation. The recent scholarly literature is silent on this question.
4. Inter-city truck transportation might seem to be a highly unconcentrated industry with easy entry. Therefore mergers in the industry, and even a merger that by the terms of the announcement creates "the largest U.S.-based transportation service provider with a focus on big shipments for business-to-business customers," might be expected to pass merger antitrust screening without any hesitation at all. But appearances can deceive.
5. The 10-K reports of both Yellow and Roadway companies indicate that the companies specialize in less-than-truckload (LTL) inter-city, and usually relatively long-distance, shipments. There are important differences between the TL and LTL businesses. To carry truckload-quantity freight efficiently, all one needs is a rig, a map, a telephone, and perhaps an internet connection. LTL carriage, defined conventionally as shipments of less than 10,000 pounds, is very different. The shipper communicates its need for a pickup to the transportation company. A peddle truck is dispatched to the point of shipment to load the shipment. Usually the shipment is too small to be carried directly to the consignee (i.e., destination point), so it is transported to a terminal, off- loaded, and then reloaded onto a truck that will carry truckload quantities of mixed freight to the general destination area. When the line-haul truck reaches a terminal in the vicinity of the destination, the shipment is off-loaded again and loaded onto a peddle truck that carries it to its final destination. The amount of handling required to provide this kind of service is much greater than the handling needed on truckload shipments, and as a result, LTL rates for a given distance are several times higher per hundredweight than TL rates. I should note that UPS imposes weight maxima (from a web search, 150 pounds) that severely limit its competition with truckers on LTL shipments.
6. One key requisite of an LTL specialist is a network of terminals. According to their most recent 10-K reports, Yellow has 345 terminals, with an average shipment size of 1000 pounds and average shipping distance of 1200 miles; Roadway has 371 terminals and reports an average shipment length of 1301 miles. This seems roughly consistent with nationwide coverage plus my experience on the "reach" of terminals during the 1950s. The home base terminal of Scherer Freight Lines was Ottawa, Illinois, where I worked. >From there our peddle trucks fanned out to surrounding cities within a radius of roughly 50 miles. This was before the construction of the interstate highway system. Peddle route radii may be somewhat longer now, but I doubt that the basic economics of LTL freight haulage have changed fundamentally since then. The total area encompassed by the "lower 48" United States is 3.096 million square miles. Assuming that circular territories fill that space (hexagonal markets would be better), the U.S. area would be filled by 377 circular markets with maximum shipping radii of 50 miles. This approximates the number of terminals actually operated by Yellow and Roadway.
7. According to Thomas Gale Moore in Leonard Weiss and Michael Klass, eds., Regulatory Reform: What Actually Happened (1986), p. 28, "Apparently the cost of setting up terminals and building a business from scratch inhibited entry into the LTL business [after deregulation], but entry into truckload carriage was easy." This too is consistent with my experience. Within limits, there are appreciable economies of scale in setting up and operating a terminal. (A very large terminal may suffer from diseconomies due to more complex and costly cross-platform transfers.) Markets are intrinsically localized by the time required for peddle trucks to perform their required pickup and delivery jobs, with consolidation or deconsolidation occurring at terminals. Following Adam Smith's precept that "the division of labour is limited by the extent of the market," a larger number of efficient-scale terminals can be accommodated in densely populated markets than in sparsely populated markets. My recollection is that there were many dozens and perhaps even a hundred terminals, one of them my company's, within the metropolitan Chicago area, from which truckload shipments were dispatched to all parts of the United States. But within the 50 mile radius around my less populated home town in north central Illinois, with population totalling about 100,000, there were only half a dozen terminals handling long-distance LTL freight. Typically at that time, we competed directly with one to three carriers providing the same kind of service. Localization of LTL markets led to oligopoly structures.
8. Again, I doubt that, despite deregulation, much has changed in this fundamental respect since the 1950s. There must be many parts of the United States, typically more than 50 miles from large cities, in which the number of competing LTL carriers can be counted on the fingers of one hand, with some fingers to spare.
9. According to the 1997 Economic Census of Transportation and Warehousing, Subject Series, Summary volume, EC97T48S [www.census.gov/epcd/www/econ97.html, click on transportation and warehousing], total revenues in 1997 were $51.1 billion in long- distance truckload shipping, with 23,111 establishments reporting, and $25.0 billion in long-distance LTL trucking (NAICS 484122), with 6,210 establishments (mostly terminals) reporting, 697 of which were inoperative. According to that same report, p. 143, the four largest long-distance LTL companies, two of which were almost surely Yellow and Roadway, generated 39.3 percent of total category revenues. They operated 1271 establishments, or an average of 318 each -- a bit less than half the number of terminals reported separately by Yellow and Roadway. Companies 5-8, reporting 897 establishments together, generated 18.6 percent of industry revenues, and companies 9-20 an additional 14.6 percent of industry revenues. Assuming that Yellow and Roadway were in fact members of the leading four and assuming very conservatively that market shares were evenly divided within size distribution groups, the increase in the LTL segment HHI index is roughly 117 points from a pre-merger base somewhat in excess 490. Again, if the market is viewed as national in scope, the proposed merger would not violate conventional guidelines.
10. However, within parts of the United States lacking very large cities, it is possible or even likely that the merger leads to a consolidation of viable LTL terminal-operating rivals with much larger market share impacts. I have no information on the geographic distribution of terminals operated by Yellow and Roadway and their principal local rivals. Such information could be obtained through a "second request" subpoena. Alternatively, for a first look, one might consult the Yellow Pages of cities ranked, say, 100 through 140 in the Census Bureau's annual SMSA population estimates -- covering cities with populations between 250,000 and 384,000 -- and a similar sample of smaller, more isolated, SMSAs, checking to see how many long-distance general LTL freight carriers are listed. I lack the resources to perform such an analysis, but, given the probability of guideline-exceeding concentration changes, it could be done at the Library of Congress with a modest investment of antitrust agency resources.
11. This leads back to my ignorance on how the LTL carriers determine their prices. My expectation is that in the less densely populated areas of the United States, LTL market structures are oligopolistic or even tightly oligopolistic. What would I do to solve the very complex pricing problem under these structural conditions if I were still in the trucking business? Most likely I would try to establish some sort of price leadership, using formulas announced by the largest service provider and followed more or less closely by smaller rivals. A merger with substantial market share impacts could strengthen adherence to such price leadership. And in an industry with a long tradition of overt collusion in rate-setting, it is not inconceivable that there could be actual collusion. Further concentration of the market structure could reduce remaining proclivities toward independent pricing. Again, on this I have no information at all. But a second request could illuminate the pricing mechanisms used. Because the pricing problem is so complex and must be solved on a decentralized basis, there is probably substantial documentation on it. I would ask, for example, for rate-setting guidelines and formulae available to rate-quoting personnel along with statements of how deviations from standard rates are to be made by field sales personnel and approved by higher authority. For a sample of standardized commodity shipments between some named smaller city pairs in which Yellow and Roadway face one another, one might request actual rate quotations. Or similar price information might be sought from shippers known to be small businesses. More rate-conforming behavior is to be expected with such shippers as compared to shipments for the largest customers.
12. In tightly oligopolistic localized markets, a merger from say three to two rivals could also have an adverse impact on service competition. I know from first-hand experience that even under price regulation, our firm competed vigorously on a service basis with our one or two rivals in particular small-town markets.
13. Anticipating that the merger could have appreciable price-raising or tacit collusion-entrenching effects, I add that I would expect a vigorous efficiencies defense. In their announcement statement of July 8, the parties state that "Annual synergies of $45 million should be achieved by the end of the second year. By year five annual synergies could be in excess of $125 million." I find this statement plausible on its face. There are undoubtedly economies of scale in terminal operation, and there is probably substantial overlap between some territories served by Yellow and Roadway terminals. Given this, it is puzzling to see that the announcement goes on to say that "We expect minimal employee displacement among Field Sales & Operations at either company." This seeming inconsistency should be explored. It is also possible, as in the 1996 merger between the Union Pacific and Southern Pacific railroads, that computer system incompatibilities and divergent corporate cultures could lead to chaos rather than expected rationalization. If the evidence shows pricing to be formulaic, realized efficiencies might not be passed on in the form of lower rates. Such a case would be pose difficult enforcement problems under the April 7, 1997, Horizontal Merger Guidelines implication that merger-specific efficiencies should lead to lower prices and/or other consumer benefits.