Event Study Methodology in Cartel Litigation

1. Introduction

Event study methodology is an econometric technique used in cartel litigation to determine whether a particular event—such as the formation, operation, discovery, or termination of a cartel—caused a measurable change in market prices, output, stock values, or other economic variables.

In competition-law damages cases, event studies are particularly useful where the claimant must establish a before-and-after economic effect attributable to cartel conduct. The central question is generally:

What would the relevant price or economic outcome have been if the cartel had not existed?

This hypothetical outcome is commonly described as the counterfactual or “but-for” price. The difference between the actual cartel-affected price and the estimated counterfactual price may be used to calculate the overcharge and, ultimately, damages.

2. Meaning of an Event Study

An event study examines the economic effect associated with a particular event by comparing observed outcomes around the event with an estimated outcome that would otherwise have occurred.

In cartel litigation, relevant events can include:

• commencement of cartel discussions;
• agreement to fix prices;
• coordinated price increase;
• exchange of competitively sensitive information;
• allocation of customers or territories;
• termination of cartel arrangements;
• cartel investigation or dawn raid;
• public disclosure of the cartel;
• leniency application;
• infringement decision;
• cessation of collusive conduct.

The methodology attempts to isolate the economic effect associated with these events from ordinary market fluctuations.

3. Why Event Studies Matter in Cartel Litigation

Cartels generally operate in markets where prices are affected by many variables simultaneously.

For example:

Observed price = competitive price + cartel effect + demand effects + input-cost effects + seasonal effects + other market factors

Consequently, simply showing that prices increased during the cartel period is usually insufficient.

An econometric analysis seeks to determine whether the increase was statistically and economically associated with cartel activity after controlling for other factors.

4. Basic Structure of the Methodology

A typical cartel event study involves five stages:

Stage 1 — Identify the cartel event

The economist identifies a legally and economically meaningful event.

Examples:

• cartel formation;
• first coordinated price increase;
• termination;
• investigation announcement.

Stage 2 — Establish the relevant period

The analysis normally separates the market into:

1. pre-cartel period;
2. cartel period;
3. post-cartel period.

Stage 3 — Collect economic data

Potential data include:

• transaction prices;
• list prices;
• quantities sold;
• production costs;
• input prices;
• demand indicators;
• exchange rates;
• freight costs;
• capacity utilisation;
• market shares;
• competitor prices.

Stage 4 — Construct the counterfactual

The economist estimates what prices would probably have been absent the cartel.

Stage 5 — Estimate cartel effect

The estimated difference between actual and counterfactual prices becomes the basis for assessing the cartel's economic effect.

5. Price Event Studies

The most important application in cartel damages litigation is the analysis of price movements.

Suppose:

• actual cartel-period price = ₹120;
• estimated competitive price = ₹100.

The estimated overcharge is:

120−100=₹20120-100=₹20

Percentage overcharge:

120−100100×100=20%\frac{120-100}{100}\times100=20\%

Thus, the estimated cartel overcharge would be 20%.

The calculation can subsequently be applied to affected purchases, subject to questions concerning pass-on, mitigation and other damages principles.

6. Regression-Based Event Study

A simplified model can be expressed as:

Pt=α+βCartelt+γXt+ϵtP_t=\alpha+\beta Cartel_t+\gamma X_t+\epsilon_t

Where:

• PtP_t = observed price;
• CarteltCartel_t = variable indicating the cartel period;
• XtX_t = other explanatory variables;
• β\beta = estimated cartel effect;
• ϵt\epsilon_t = error term.

If β\beta is statistically significant and positive, the analysis may support the proposition that cartel activity increased prices.

A more sophisticated model could be:

Pit=α+βCartelt+γCostit+δDemandt+μi+λt+ϵitP_{it}=\alpha+\beta Cartel_t+\gamma Cost_{it}+\delta Demand_t+\mu_i+\lambda_t+\epsilon_{it}

This can incorporate:

• firm-specific effects;
• time effects;
• input costs;
• demand conditions;
• product characteristics.

7. The Importance of the Counterfactual

The principal difficulty in cartel damages litigation is that the competitive price is unobservable.

Once a cartel has operated, courts cannot directly observe what prices would have existed without the cartel.

Therefore, the economist must construct a counterfactual using techniques such as:

A. Before-and-after analysis

Compare prices before and during the cartel.

B. After-and-before analysis

Use post-cartel prices to estimate competitive prices.

C. Yardstick analysis

Compare the affected market with a comparable unaffected market.

D. Difference-in-differences

Compare changes in the affected market with changes in an unaffected control market.

E. Regression analysis

Control for observable determinants of price.

F. Synthetic-control methodology

Construct a statistically weighted combination of unaffected markets or products to approximate the counterfactual.

8. Difference-in-Differences

Difference-in-differences is especially useful where an unaffected control market exists.

Suppose:

Affected market increase:

25%25\%

Control market increase:

10%10\%

Estimated cartel effect:

25%−10%=15%25\%-10\%=15\%

The methodology therefore attempts to remove general market-wide price inflation.

9. Event Studies and Cartel Termination

Termination events can be particularly informative.

Suppose prices:

• rise substantially when cartel coordination begins;
• remain elevated during cartel operation;
• decline sharply after cartel termination.

Such a pattern may support the inference that cartel activity affected prices.

However, correlation does not automatically establish causation.

A price decline after termination could also result from:

• falling input costs;
• declining demand;
• new competitors;
• technological change;
• recession;
• exchange-rate movements.

Consequently, appropriate controls are essential.

10. Statistical Significance

Courts must distinguish between:

Statistical significance

Whether the estimated cartel effect is unlikely to have occurred merely through random variation.

Economic significance

Whether the estimated effect is sufficiently large to matter economically.

A coefficient can be statistically significant but economically trivial.

Conversely, a large estimated overcharge may have a wide confidence interval and therefore substantial statistical uncertainty.

Courts may therefore examine:

• standard errors;
• confidence intervals;
• p-values;
• robustness tests;
• model specifications;
• sample size.

11. Structural Break Analysis

A cartel may create a structural break in a price series.

For example:

Price  | 130|                  _________ 120|                / 110|              / 100|_____________/  |  +-------------------------- Time       Cartel begins

A structural-break model asks whether the statistical relationship governing prices changed at a particular date.

This can be useful where:

• cartel formation has a known date;
• coordination began suddenly;
• prices changed abruptly;
• cartel termination produced another identifiable break.

12. Event Windows

The economist must carefully define the event window.

For example:

• 30 days before the event;
• event date;
• 30 days after the event.

A very narrow window may fail to capture the effect.

A very broad window may introduce unrelated economic developments.

The appropriate window depends upon the market and event.

13. Multiple Events

Cartels rarely involve only one event.

A cartel may involve:

1. initial meeting;
2. agreement;
3. first price increase;
4. subsequent coordination;
5. investigation;
6. public announcement;
7. termination.

An event-study model can therefore include multiple event indicators.

For example:

Pt=α+β1Formationt+β2PriceIncreaset+β3Investigationt+β4Terminationt+ϵtP_t=\alpha+\beta_1 Formation_t+\beta_2 PriceIncrease_t+\beta_3 Investigation_t+\beta_4 Termination_t+\epsilon_t

This allows the economist to study how prices reacted to different cartel-related events.

14. Stock-Market Event Studies

Event studies can also examine share prices.

Suppose a cartel investigation becomes public and the defendant's stock price falls sharply.

An economist may estimate:

ARit=Rit−E(Rit)AR_{it}=R_{it}-E(R_{it})

Where:

• ARAR = abnormal return;
• RR = actual stock return;
• E(R)E(R) = expected return.

Cumulative abnormal returns can then be calculated over the event window.

However, stock-market evidence generally addresses a different question from cartel overcharge analysis.

A stock-price reaction may demonstrate that investors considered the cartel revelation economically important, but it does not by itself establish the amount of overcharge suffered by customers.

15. Event Study vs. Traditional Cartel Damages Analysis

In practice, courts may consider several methodologies together.

16. Evidentiary Problems

Event studies present several difficulties in litigation.

16.1 Confounding variables

Other events may coincide with cartel formation.

For example:

• oil-price increases;
• supply shortages;
• inflation;
• currency fluctuations;
• war;
• regulatory changes.

16.2 Endogeneity

Cartel members may increase prices precisely when market conditions would independently have caused prices to rise.

16.3 Selection of control markets

A supposedly unaffected market may itself have experienced indirect cartel effects.

16.4 Data problems

Data may be:

• incomplete;
• aggregated;
• inconsistent;
• affected by product changes;
• affected by discounts or rebates.

16.5 Cartel duration

The legally identified infringement period may not perfectly correspond to the economically effective period.

17. Judicial Treatment: Six Important Case Laws

1. J&E Eversheds LLP v. RWE Npower plc

This litigation is important in the English competition-damages context because it demonstrates the courts' willingness to scrutinise the methodological and evidential foundations of economic loss calculations.

The broader lesson is that damages cannot simply be inferred from the existence of an infringement. The claimant must establish a sufficiently reliable methodology connecting the infringement to the alleged loss.

Principle: Economic modelling must be sufficiently grounded in evidence to support a damages assessment.

2. Devenish Nutrition Ltd v Sanofi-Aventis SA

This English litigation arose from the vitamins cartel and is important for understanding the relationship between a competition infringement and the claimant's damages.

The case illustrates that cartel damages involve more than proving that a cartel existed. Issues such as loss, causation, restitutionary principles and pass-on can materially affect recovery.

Principle: The existence of cartel conduct does not automatically determine the amount of compensable loss.

3. Sainsbury's Supermarkets Ltd v Mastercard Inc

The litigation concerned interchange fees and involved extensive economic evidence concerning competitive counterfactuals.

The Supreme Court considered the difficulty of determining what would have happened in a hypothetical competitive environment.

The case is particularly valuable for understanding how courts approach counterfactual economic analysis and causation.

Principle: Competition damages require a realistic assessment of the counterfactual, rather than a purely theoretical alternative market.

4. BritNed Development Ltd v ABB AB

This is one of the most significant English competition-damages decisions for econometric analysis.

The case involved the power-cables cartel and extensive expert evidence concerning the estimation of the overcharge.

The court considered competing economic models and expert approaches to estimating the price effect of cartel conduct.

Principle: Courts can engage closely with sophisticated econometric evidence when determining cartel overcharge, but the model must be supported by reliable data and economically reasonable assumptions.

5. United States v. U.S. Gypsum Co.

The U.S. Supreme Court's competition jurisprudence concerning price effects and economic evidence illustrates the importance of distinguishing legitimate parallel pricing from conduct producing anticompetitive effects.

The case is relevant to the broader evidentiary problem faced by event studies: price movement alone cannot establish unlawful coordination without appropriate economic and factual context.

Principle: Economic evidence must be interpreted in conjunction with evidence concerning competitive conduct and market circumstances.

6. In re High Fructose Corn Syrup Antitrust Litigation

This U.S. cartel litigation is important because it illustrates the role of economic evidence, pricing patterns and circumstantial evidence in establishing coordinated conduct.

The case demonstrates that economists may identify pricing behaviour that is consistent with coordination, while courts must determine whether the totality of evidence supports the inference of cartelisation.

Principle: Statistical and economic evidence can reinforce other evidence of collusion but should not be treated as an automatic substitute for proof of the underlying unlawful agreement.

18. Important Lessons From the Case Law

The cases collectively demonstrate several propositions.

First

Cartel existence and cartel effect are different questions.

A competition authority's infringement finding may establish the unlawful conduct, but damages litigation may still require an economic assessment of its consequences.

Second

Counterfactual analysis is central.

The court must determine what price or economic outcome would probably have existed without the cartel.

Third

Econometric models are evidence, not mathematical truth.

An economic model depends upon:

• assumptions;
• data;
• specification;
• control variables;
• comparator selection.

Fourth

Courts can reject unreliable economic models.

A sophisticated regression is not persuasive merely because it uses advanced mathematics.

Fifth

Multiple methodologies may be preferable.

A court may find a damages estimate more reliable where:

• regression analysis;
• before-and-after evidence;
• comparator markets; and
• documentary evidence

point in the same direction.

19. Event Study in a Hypothetical Cartel Case

Assume manufacturers A, B and C are found to have operated a cartel from January 2020 to December 2022.

Observed prices:

A simple analysis suggests that prices increased during the cartel period and declined after termination.

But the economist cannot simply claim:

₹120−₹102=₹18₹120-₹102=₹18

as the cartel overcharge.

Instead, the economist must determine whether the price increase was caused by:

• input costs;
• inflation;
• demand;
• exchange rates;
• capacity constraints;
• product quality;
• transportation costs;
• technological changes.

After controlling for these factors, suppose the estimated competitive price is ₹104.

Then:

Overcharge=₹120−₹104=₹16Overcharge=₹120-₹104=₹16

Percentage overcharge:

16104×100≈15.38%\frac{16}{104}\times100\approx15.38\%

The estimated cartel effect would therefore be approximately 15.38%.

20. Event Study and Passing-On

Another important issue is pass-on.

Suppose a cartel causes an intermediate purchaser to pay an additional ₹16 per unit.

That purchaser may increase its own selling price and transfer part of the additional cost to downstream consumers.

Therefore:

Cartel Overcharge≠Necessarily Final LossCartel\ Overcharge \neq Necessarily\ Final\ Loss

The court may need to consider:

• whether the claimant passed the overcharge downstream;
• how much was passed on;
• whether demand changed;
• whether margins were compressed;
• whether lost sales resulted.

Event-study evidence may therefore form only one part of the overall damages analysis.

21. Strengths of Event Study Methodology

1. Empirical

It uses actual market data rather than purely theoretical assumptions.

2. Flexible

It can be applied to:

• prices;
• quantities;
• profits;
• share prices;
• market shares.

3. Useful for identifying timing

It can help identify whether a market response coincided with cartel-related events.

4. Quantitative

It can generate an estimated magnitude of the cartel effect.

5. Adaptable

It can be combined with:

• regression analysis;
• difference-in-differences;
• structural models;
• comparator analysis.

22. Limitations

Event studies should not be treated as conclusive merely because they produce a numerical result.

Important limitations include:

1. uncertainty concerning the event date;
2. simultaneous economic shocks;
3. inadequate control variables;
4. non-comparable control markets;
5. changes in product composition;
6. anticipation effects;
7. data limitations;
8. serial correlation;
9. model specification problems;
10. potential manipulation or selection of the event window.

23. Best-Practice Approach for Courts

A robust cartel event study should ideally contain:

A. Clear event definition

The economist should explain precisely what event is being studied.

B. Reliable data

The underlying dataset should be sufficiently complete and representative.

C. Appropriate controls

Relevant cost and demand variables should be incorporated.

D. Counterfactual justification

The economist should explain why the selected counterfactual is economically plausible.

E. Sensitivity analysis

Results should be tested under alternative specifications.

F. Robustness checks

The conclusion should not depend upon one arbitrary model.

G. Confidence intervals

The court should understand the degree of uncertainty surrounding the estimate.

H. Independent corroboration

Econometric evidence should, where possible, be tested against documentary and factual evidence.

24. Role in Modern Digital and Algorithmic Cartels

Event studies are becoming increasingly relevant to algorithmic pricing and digital markets.

For example, if competing firms use pricing algorithms and suddenly adopt coordinated pricing behaviour, an event study could examine:

• algorithm deployment;
• software updates;
• API integration;
• common pricing rules;
• data-sharing arrangements;
• sudden price convergence;
• termination of algorithmic coordination.

The methodology could compare pricing before and after the algorithm's introduction while controlling for demand and cost variables.

However, algorithmic environments create an additional challenge: the event itself may be gradual rather than a single identifiable date.

Consequently, conventional event studies may need to be supplemented with:

• rolling-window analysis;
• structural-break analysis;
• machine-learning techniques;
• panel-data models;
• algorithmic interaction analysis.

25. Conclusion

Event study methodology is a powerful evidentiary and econometric tool in cartel litigation because it attempts to connect a legally relevant cartel event with measurable economic consequences.

Its central contribution is the estimation of the counterfactual market outcome and the resulting cartel effect.

However, a price increase surrounding a cartel event does not automatically establish an overcharge. Courts must consider alternative explanations, data quality, statistical significance, economic significance, model assumptions and the reliability of the counterfactual.

The most persuasive cartel damages analysis therefore combines event-study evidence with regression techniques, comparator markets, documentary evidence and factual evidence concerning cartel operation.

The fundamental analytical sequence is:

Cartel Event → Market Response → Econometric Identification → Counterfactual → Estimated Overcharge → Causation → Damages