Economic Modeling and Simulations

Cornerstone Research regularly formulates and implements empirical analyses to respond to economic and financial issues in litigation. We frequently use real-world data with sophisticated statistical and econometric methods, such as advanced structural modeling and regression analysis, numerical techniques, time series analysis, forecasting, simulations, and optimization

  • Performed 10,000-iteration simulations, randomly selecting peer funds across years for performance comparison analyses, reducing runtime from hours down to minutes.
  • Developed GARCH modeling tools that feature a C++ base for efficiency and an R interface for convenient integration into the case codebase. This framework fits large-scale models with external regressors more quickly than standard software packages, while allowing for complete control over the model specification.
  • Performed demand model simulation by replicating Matlab simulations in R, with plots to illustrate simulation results.
  • Developed a custom optimization method for merger simulations, decreasing runtime by 80 percent with improved stability over standard methods.
  • Developed state-of-the-art econometric models of the demand for cigarettes, using very large datasets of store-level data.
  • Developed and implemented a detailed simulation model to estimate the competitive impact of a merger in the electric power industry. The flexibility and computational efficiency of the model enabled us to estimate the merger’s impact under a large number of market scenarios in a timely manner.
  • Used payment data from millions of customers to determine the personal characteristics and transactional history of customers. We then designed a sophisticated Monte Carlo simulation to project the future cash flows from each customer under thousands of scenarios and estimated the present value of the expected cash flows associated with each scenario.
  • Developed an institutional damages model using Monte Carlo simulations to generate a confidence interval for damages in a securities class action.
  • Refactored code and used parallel processing to reduce the runtime for comparing thousands of distributions from three days to less than 30 seconds.
Statistical Analysis and Sampling

Using samples drawn from large databases can create efficiencies. We are experienced in developing and implementing reliable and defensible sampling techniques, as well as assessing the sampling methodologies of opposing experts.

  • Evaluated sampling analyses performed by opposing parties to prove liability and damages in False Claims Act litigation.
  • Provided statistical analyses of the incidence rates of alleged defects in automotive product liability cases, based on data from the National Highway Traffic Safety Administration or manufacturers, including warranty claims, customer recall repairs, or customer complaints.
  • Performed statistical assessments of “no poach” allegations in hiring agreements between tech firms.
  • Conducted statistical analyses of a university’s admissions process.
  • In a labor discrimination class action, provided statistical analysis of wages and promotion outcomes across individual employees and, on average, across different store locations.
  • Created a representative sample of homeowners and conducted an independent customer satisfaction survey in a putative class action against a national window manufacturer.