Adding to ProMarket’s discussion of the Robinson-Patman Act, Herbert Hovenkamp argues that – among other issues– the law was captured by special interests when it was written. The only way the Act can be salvaged without inducing harm is with a serious market power requirement.

The Robinson-Patman Act (RPA) became law during the Great Depression in 1936. Representative Wright Patman spearheaded a Congressional campaign to limit the growth of “chain stores,” who were hurting small independent retailers. The chains, defined by the census as retailers owning four or more stores, were targeted by populists in search of scapegoats. Huey Long railed at them: “I would rather have thieves and gangsters than chain stores in Louisiana.”  The RPA was not Patman’s last battle.  Two years later he unsuccessfully proposed  “death sentence” legislation intended to tax many chains out of existence.

The RPA’s enactment reveals a captured Congress.  The FTC’s chain store investigation, summarized in its 1935 Annual Report, acknowledged dangers from monopoly price discrimination and secret rebates but praised the chains’ many economic advantages, including:

“…the integration of production and of wholesale and retail distribution, from the savings involved in avoiding credit and delivery service, and from the ability of chains to realize the benefits of large-scale advertising….  Nor did the Commission recommend any change in the law in order to eliminate such advantages. Such a program would involve radical interference with the rights of private ownership and initiative, virtual abandonment of the competitive principle, and destruction of the public advantage represented by lower prices and lower cost of living.”

Unfortunately, Congress never asked the FTC to testify at the RPA hearings, nor were the Department of Justice, consumer groups, or economists invited.  It listened almost exclusively to trade associations of independent retailers.  Representative Patman candidly acknowledged that Henry B. Teegarden, General Counsel of the Wholesale Grocer Association, actually drafted the bill.

As protection of small business from powerful distant competitors, RPA was a drafting catastrophe. Unlike other Clayton Act provisions, the RPA required neither market power nor the often evoked “buyer power.” It demanded only “injury to the competitor,” rejecting any requirement of “general injury to competitive conditions.” The RPA case law followed, never requiring either market power nor even remote ownership or a minimum size.  Further, it gave vertical integration a free pass. The statute made it unlawful to sell the same good to two competing resellers at different prices. However, a transfer from a firm’s own production facility or warehouse was not a “sale.” If a chain store’s growth resulted from vertical integration – the true source of harm to small independents – the Act simply did not apply. No more than 15% of the cost advantage enjoyed by the chains came from lower buying prices, and even these were cost justified. Further, an unintegrated firm could avoid the statute by integrating vertically; so the RPA encouraged firms to do the very thing that most injured independent grocers.

The RPA’s drafters woefully misunderstood competitive markets. Cartels adhere to uniform prices. By contrast, a competitor’s price varies with market circumstances as well as the nature and terms of the deal, including processing, delivery, stocking, differing quantities, joint costs, customized production, or the combining of multiple products into a single order.  Nevertheless, a mere price difference triggered the statute. To be sure, a monopsony buyer might force a lower price, but the RPA would never discover it because it lacked a market power requirement.

While the RPA allowed ‘cost justified’ discounts, the 1956 Attorney General’s National Committee Antitrust Report complained that they were almost impossible to prove. It required defendants to compute customer delivery and processing costs individually, through heroics such as timing labor steps with a stopwatch so that one buyer would not get an unjustified deal. Further, the defense applied only to the “differing methods or quantities” by which goods were sold, thus favoring small firms with minimal fixed costs. Price discrimination with higher fixed cost assets, such as warehousing and transportation, results from efforts to keep output high by pricing competitively to marginal buyers.  A seller who bid a lower price (but above marginal cost) out of available capacity violated the Act unless it retroactively gave that price to everyone. Similarly, a seller who cut local prices in order to compete violated the Act unless it reduced the price in less competitive areas as well.

The discrimination provision applied to goods “of like grade and quality,” forcing courts to confront product differentiation. This included variations in the physical dimensions of containers, the difference between house brands and nationally advertised brands, or goods that were specially configured for specific purchasers. It was also deeply suspicious of other price differences, such as discounts for brokerage, which larger buyers often did not need; promotional allowances because large retailers often did their own advertising; or the provision of processing facilities. One contentious example was “backhaul allowances,” which sellers might reimburse to customers who had two-way business between two destinations and could return with a full truck. The FTC saw an unfair advantage to smaller firms that had no business requiring a backhaul. Trucks returning empty resulted “in the waste of thousands of gallons of fuel annually.”

In any event, monopoly was rare. The chains competed not only with independents but also with each other, a trend that only increased as the century progressed. Even giant A&P’s national market share in the 1930s was roughly 10% and did not exceed 20% in any region.  By 1929 there were 50000 competing grocery chains alone, but they constituted only 18% of total grocers. Between 1929 and 1949 the number of all chains declined from 159,000 to 105,000 but their revenue tripled, suggesting consolidation. Overall numbers indicate robust competition today, as much as in the 1930s. While individual chains have grown much larger, so have the markets they serve. Today no single retailer claims even 10% of consumer retail spending.  Worldwide, the ten largest firms together share less than 7% of the retail market. This high degree of competition among large retailers has injured small independents severely. They are simply not in this game.

Prior to the neoliberal right’s critique of the RPA in the late 1970s, pushback came from the Civil Rights Movement’s discovery of urban poverty. The poor and people of color paid too much for food. The FTC’s chain store investigation had already observed in 1934 that those of “lower means” spent twice as much of their income at chains (35% vs. 17%) than the wealthier, because the chains had lower prices. In a stinging report on the RPA in 1978 the Justice Department linked harm to low income people to RPA enforcement in the areas of medical products and general merchandising.

Confronted with a statute that imposed enormous litigation and operational costs, and that harmed consumers, particularly those of low income, without showing much benefit for small retailers, the Justice Department largely repudiated enforcement. While a technical study of the cost of a particular statute is impossible, the DOJ’s estimate that the RPA cost the economy $3 to $6 billion annually was almost certainly too low. It included higher prices, but not compliance costs or job losses. Labor and consumers are both vertically related to production. They rise or fall togetherHigher prices harm consumers and also lead to fewer jobs.

After the 1970s the Supreme Court also concluded that because the RPA is an “antitrust law” it should be interpreted consistently with overall antitrust enforcement goals. It limited RPA decisions that required competitors to verify one another’s prices in cartel fashion in order to satisfy the statute’s “meeting competition” defense. It also imposed stricter requirements on proof of competitive harm.

Like many populist campaigns, the RPA was the worst possible solution to the problem at hand: struggling Americans in 1936 did not really need “protection” from low prices. Nor do they today. Nevertheless, even a badly designed statute might still be of use. Is anything in the RPA worth salvaging? For example, “buyer power” can be used anticompetitively. If enforcement agencies actually started to require it, perhaps the Act would find a purpose.  Most-favored-nation agreements, loyalty or bundled discounts, or other compelled price discrimination can be anticompetitive, but only in the presence of market power. However, differential pricing by firms without market power is competitively harmless. Indeed, it almost always increases rather than decreases output. So the Act could be improved if enforcers required market power and reasonable proof that a price difference led to reduced output and higher prices. Few decided RPA cases would meet that test, and the Sherman Act would address them in any event.

Another problem, which prosecutorial discretion cannot control, is that the RPA is enforceable in private treble damages actions. While we might trust the antitrust agencies not to abuse a badly designed statute, we cannot be so sanguine about private plaintiffs. That then creates the hazard of a statute that does not insist on any showing of market power, including buyer power, and private plaintiffs that are not as public regarding as the Agencies should be.

One possible way out would be a statutory amendment that explicitly requires market power, either as buyer or seller, and proof of an anticompetitive output reduction. Otherwise there is little reason to return to a distribution economy held hostage by a bad statute that a misguided Congress passed nearly a century ago.

Read about our disclosure policy here.