During the 1990s, academic research and reports in the popular press presented the view that corporate diversification destroyed value. These reports suggested that conglomerate companies were discounted by as much as 15 percent from the value that could be attained by divesting and operating the divisions as stand-alone companies. These results followed from a methodology that compared the divisions with stand-alone proxy companies used as benchmarks. The authors contend that if the divisions of conglomerates systematically differ from the benchmarks, then failure to account for these differences can lead to incorrect inferences regarding valuation effects of conglomerate diversification. To assess the extent to which these selection-bias issues are important in measuring the effect of diversification on company value, the authors examine two samples of companies that expanded via acquisition and/or increased their reported number of business segments.

Typically, a business unit is priced at a significant discount to the median benchmark company in its industry prior to its merging with a larger company. Thus, when this discounted business unit is merged into an existing company, the excess value of the combined business is negatively affected when measured using the two-step methodology of Berger and Ofek (Journal of Financial Economics, 1995). For the sample of acquisitions, the authors find that accounting for the characteristics of the acquired units explains nearly all the reduction in the acquirer's excess value as a result of the acquisition. Consistent with the view that characteristics are important in determining valuation discounts, the authors also find that excess value is not reduced when a company increases its number of business segments through internal growth rather than acquisition.