## Facts

- Plaintiffs (purchasers) brought an antitrust action against defendant suppliers, alleging an unlawful conspiracy to fix prices for ferrosilicon, an input used in making carbon steel.
- The case proceeded in the U.S. District Court for the Western District of Pennsylvania as consolidated litigation.
- Plaintiffs designated economist Dr. Laurits Christensen to offer opinions on antitrust impact and damages.
- Defendants did not dispute Christensen’s qualifications (Ph.D. in economics, long-time college instructor, editorial-board service at a leading economics journal); they challenged the reliability of his methods under Federal Rule of Evidence 702 and Daubert.
- Christensen used a multiple-regression analysis to explain defendants’ ferrosilicon prices (the dependent variable) using several independent variables, including measures such as iron and steel production, gross domestic product, an import price index for ferrosilicon, electricity prices, and capacity utilization in the iron and steel industries.
- The regression results showed a strong, statistically significant relationship between import prices and defendants’ domestic prices (reported at greater than a 95% confidence level), while other proposed variables contributed little to explaining price movements.
- Using the import-price relationship, Christensen estimated a “but-for” price path absent the alleged agreement and calculated overcharges from the difference between actual and predicted prices.
- Defendants argued the model was unreliable because it omitted additional explanatory variables and because it did not yield a single, simple predicted price that a jury could apply mechanically.
- In response to the omitted-variable attack, Christensen tested additional variables suggested by defendants’ economist; those additions increased his estimated damages.
- The court held a multi-day evidentiary hearing under Rule 104(a) on the admissibility of Christensen’s testimony. The court had already denied the motion by prior order and issued the written decision to explain its reasoning.

## Issues

1. Whether Christensen’s multiple-regression methodology for proving antitrust impact and estimating damages was sufficiently reliable under Federal Rule of Evidence 702 and Daubert to be presented to the jury.
2. Whether alleged flaws in model specification (including suggested missing variables) and objections about how the results would be presented required exclusion rather than resolution through cross-examination and competing proof.

## Decision

- The court denied defendants’ motion to exclude Christensen’s testimony.
- The court ruled that Christensen would be permitted to testify at trial in accordance with his report.
- The court found that multiple regression is a generally accepted statistical tool for evaluating how several factors relate to price and for estimating “but-for” prices in antitrust cases.
- The court concluded that defendants’ objections—such as disputes over variable selection, alternative specifications, and how easily a jury could use the results—were matters affecting weight and credibility, not a basis for barring the testimony altogether.

## Legal Principles

- Under Rule 702 and Daubert, the trial court acts as a gatekeeper to screen proposed opinion testimony for reliability and relevance to the issues the jury must decide.
- The Daubert inquiry is flexible and focuses on the soundness of the methods and their application, not whether the court agrees with the result.
- Multiple-regression analysis is an accepted method in economics and can be used to estimate antitrust impact and damages when the model is tied to record data and standard statistical practice.
- Disagreements about omitted variables, alternative model choices, and sensitivity to assumptions ordinarily go to weight and are tested through cross-examination and competing economists, rather than by exclusion.
- Opinion evidence need not be reduced to a single simplistic number to be helpful; if the analysis can be explained and tested in the adversary process, it may assist the trier of fact.

## Conclusion

In In re Industrial Silicon Antitrust Litigation, the court rejected a Daubert challenge to plaintiffs’ economist and admitted his multiple-regression testimony on impact and damages, holding that disputes over model inputs and presentation were issues for cross-examination and competing evidence, not grounds to exclude the methodology under Rule 702.


In re Industrial Silicon Antitrust Litigation, 1998 U.S. Dist. LEXIS 20464 (1998)