Reliability Examination in Horizontal-Merger Price Simulations: An Ex-Post Evaluation of the Gap between Predicted and Observed Prices in the 1998 Hyundai-Kia Merger
40 Pages Posted: 24 Feb 2014 Last revised: 14 Jul 2016
Date Written: July 13, 2016
Abstract
Horizontal-merger price simulations, which rely upon pre-merger data to predict post-merger prices, have been proposed and used in antitrust policymaking. However, a dearth of closely observed large mergers in differentiated-product industries makes empirical investigations of simulation performance extremely difficult, and raises many questions regarding the accuracy of simulation performance. Although a handful of previous studies exist, they focus on short-term simulation performances and ignore long-run effects of mergers. This research investigates the long-run simulation performance and long-run pricing effects of merger in the Korean automobile industry for the period 1991-2010. This period saw the merger of Hyundai and Kia Motors in 1998, a merger caused by the Asian economic crisis and which resulted in the conglomeration of 70 percent of the Korean automobile market. By taking Nevo's (2000, 2001) method as a base and measuring its performance against this real-world merger, I find that post-merger prices can be predicted reasonably well in the short term, but that large discrepancies appear in the long-run simulation. To account for this discrepancy, I confirm four further factors that appear essential to move toward a more accurate post-merger price simulation model: change in marginal costs, change in product lines, and change in consumer incomes and preferences. I counterfactually investigate each factor's contribution to price change, confirming their significance. In my investigation I estimate consumer preferences and substitution patterns leading up to the merger, then I calculate marginal costs, and simulate post-merger prices. In addition, I estimate automobile assembly plant-level production functions to evaluate merger synergy effects. By incorporating changes in the four factors I mention, I can account for 61 percent of the long-run price discrepancies.
Suggested Citation: Suggested Citation
Do you have negative results from your research you’d like to share?
Recommended Papers
-
By Jerry A. Hausman and J. Gregory Sidak
-
Models, Mathematics and Critical Loss
By Malcolm B. Coate and Joseph J. Simons