Oil company consolidation and the price of gasoline: Studying the studies

Elizabeth M. Bailey, clinical assistant professor of economics at the W. P. Carey School of Business, has been studying the debate over oil industry consolidation and prices. She finds that critics of the oil giants, as well as those who defend them, may skew their conclusions by the assumptions they make. "In order to understand the results, people need to understand the assumptions," she says.

If anything has been more volatile than gasoline prices in recent years, it is public reaction to the instability at the pumps. Consumers, politicians, and influential voices in the media all have railed against the oil companies over the soaring prices, including the big spike after hurricanes Katrina and Rita in 2005.

A frequent accusation: That the big petroleum firms, which got even bigger in a wave of mergers between 1998 and 2002, exploited natural disasters and used their market power to gouge consumers. Elizabeth M. Bailey, clinical assistant professor of economics at the W. P. Carey School of Business, has been studying the debate over oil industry consolidation and prices.

She finds that critics of the oil giants, as well as those who defend them, may skew their conclusions by the assumptions they make. Even in otherwise rigorous analyses, the underlying assumptions about what variables to control and even what prices to study can go unexamined, according to Bailey. "In order to understand the results, people need to understand the assumptions," she says.

Dueling reports

Bailey's research of late has focused on federal government studies of consolidation and market power in the petroleum industry, including a pair of dueling 2004 reports — one by the General Accounting Office and the other by the Federal Trade Commission. The two studies, both econometric analyses and issued within weeks of each other, came to strikingly different conclusions.

The GAO report found that consolidation in the industry in the 1990s and early 2000s did result in higher U.S. gasoline prices. The FTC's Bureau of Economics found that while the industry had experienced significant consolidation, this was not the cause of higher prices. The wave of petroleum industry mergers and joint ventures that occurred starting in the late 1990s is well-documented.

Among the larger ones: Shell-Texaco and BP-Amoco in 1998; Exxon-Mobil in 1999; BP-ARCO in 2000; Chevron-Texaco, Phillips-Tosco, and Valero-UDS in 2001; and Conoco-Phillips in 2002. And the price rise during this time is all too familiar to anyone who puts gas in a vehicle: From about $1 a gallon in 1998 to $2 in late 2004 and over $3 after Katrina swept across the Gulf of Mexico in 2005.

Starting with these agreed upon facts, how did the technical economists at the GAO and FTC get such different results? One problem right off, according to Bailey, is that the two federal agencies studied different prices. The GAO examined wholesale gasoline prices, while the FTC looked at retail. This is a possible reason for the differing conclusions, Bailey says.

"The GAO's study assumes movements in wholesale gasoline prices are a reasonable proxy for movements in retail gasoline prices," Bailey writes in a recent article. "The FTC's study, on the other hand, focuses on retail gasoline prices directly because they argue it is possible for a merger to affect upstream (wholesale) prices but not downstream (retail) prices."

The FTC and GAO also took very different approaches to exploring possible links between higher prices and oil industry consolidation, according to Bailey. The GAO actually did two studies, both of which came to the same conclusion. The first focused on eight oil company mergers and used an econometric model to control for all other supply and demand factors that could affect prices.

The study concluded that the mergers were, indeed, a reason for the upward shift in prices. The second GAO study used an econometric model to measure industry concentration. The model found that the dominant market share of the top four petroleum firms could explain at least some of the price increases. In its report, the FTC undertook a case study of a 1998 joint venture between Marathon and Ashland Petroleum. The study examined the price effect of the merger on a test city and three control cities.

The test city, Louisville, experienced an increase in concentration after the joint venture. The three control cities — Chicago, Houston, and the northern Virginia suburbs of Washington, D.C. — were not affected by the Marathon-Ashland Petroleum deal but were similar in all other respects to the test city. The FTC found price effects to be substantially the same in the test and control cities.

Questioning assumptions

Bailey dissects the assumptions made in each of these studies. She finds questionable the GAO's assumption that it is possible to control for all other supply and demand variables besides industry consolidation. She notes an FTC criticism that the GAO failed to account for several important variables, including mandated gasoline formulations and seasonal changes in demand.

She also finds problems with the industry concentration model. For one thing, the model assumes that the relationship between concentration and prices is a one-way street: more concentration equals higher prices. In fact, says Bailey, higher prices can encourage new firms to enter a market, and this can lead to less concentration, which reverses the direction of the causal arrow. Another possible problem she notes with the model is that the data were compiled and aggregated across several U.S. states.

The FTC contends this calculation uses too broad a geographic area to yield meaningful results. Similarly, Bailey questions the FTC's assumption in its case study that the only difference between test city and control city is the impact of the merger. She underscores a GAO point that it is unlikely that any control city will have the same supply and demand characteristics as the test city.

Raising questions about underlying assumptions is not to say these particular studies or others should be discarded, according to Bailey. She writes, "The key to overcoming questions and criticisms is testing whether the assumptions upon which the econometric model is constructed are valid and, if it is not possible to test certain assumptions, demonstrating that the results are not sensitive to those assumptions."

She offers some suggestions for how to do this. On the matter of the relationship between higher prices and industry concentration, a statistical test could be applied to determine the direction of causation, she says. Another test could be used to determine how data should be grouped for analysis. And a little exploring of business documents could determine whether firms make pricing decisions separately for cities or states or groups of states.

But what if certain assumptions cannot be tested, say, for example, the documentation doesn't exist or the needed information hasn't been compiled? Conclusions still can be defended, according to Bailey, if the results are overwhelming. For example, concerns about city-level vs. state-level analyses fade in importance if similar results come in from a collection of zip codes, a collection of counties, or an entire state.

Did big oil gouge consumers?

Using these principles, Bailey examines another important document on oil prices, the FTC's investigation into post-Katrina gas prices. Issued in May 2006, the FTC study concludes that there was no evidence of price manipulation by refiners and distributors and that increases in gas prices were consistent with principles of supply and demand. The FTC found no evidence that firms reduced their refinery output, diverted supplies, or withheld inventories to influence prices.

Bailey's view is that the FTC's report is inconclusive because the report does not look at the magnitude of the oil companies' responses to the hurricanes. A firm can boost its output, divert supplies, and draw down reserves and still behave in a non-competitive manner — if the actions are less than what they would have been in a competitive environment. "Did they respond to the degree they would have if consolidation had not taken place? It's a very tough question to answer and one that the FTC did not do," she says.

So were the oil shocks of the past three years a result of gouging by big oil companies? Bailey declares herself "an agnostic" on the question. "My purpose is to examine whether the FTC and GAO did their studies correctly," she says. And, she doubts anyone will ever find a smoking gun. "It's unlikely that the oil companies will say, 'Oh, yes, I did this because I have market power.' These are very difficult questions to untangle."

Bailey does have some advice for policymakers dealing with the problem of soaring prices. The standard response of slapping on regulatory price controls when prices rise is likely to backfire, she says. Higher prices prompt new firms to enter the market or existing firms to produce more, she says, which is a much better solution in the long run.

"My view is that higher prices are a signal," says Bailey. "They are an invitation for entry into a market. That's what we want to see happen."

Bottom Line

• Gasoline prices tripled between 1998 and 2005. The increases occurred immediately after a wave of mergers and joint ventures involving the biggest players in the petroleum industry.
• In May 2004, the federal government's General Accounting Office and the Federal Trade Commission both issued reports on the effect of consolidation in the petroleum industry on oil prices — and came to opposite conclusions.
• A May 2006 FTC investigation found no evidence that big oil companies manipulated prices after Katrina. While noting that firms increased production and took other steps to boost supplies, the study did not consider whether the response would have been greater if the industry had not become so concentrated.
• Higher prices aren't all bad. They can attract new producers and alert existing firms to produce more.