Mountain or Molehill? Downward Biases in the Conglomerate Discount Measure

Abstract

The Berger and Ofek (1995) excess value measure, comparing a conglomerate's actual market value to an imputed value based on standalones, has become the standard method to determine value effects of diversification. In this paper, we address a significant bias in this procedure stemming from the difference in cash holdings between diversified and standalone firms. Excess values are based on firm values, including corporate cash positions. As standalones hold significantly more cash, the imputed cash value is higher than the conglomerate's actual cash value, resulting in a downward biased excess value. We thus propose to calculate excess values based on enterprise values, replacing total debt by net debt. Based on an extensive U.S. sample, we show that there is significantly less evidence of a diversification discount when adjusting for the cash bias. In terms of average dollar losses, the firm value-based models overestimate the conglomerate discount by at least 25%. Apart from removing the cash bias, we propose a second modification to the excess value measure, arguing that standalone industry multipliers should be calculated using geometric mean aggregation instead of median aggregation.

Keywords: conglomerate discount, cash bias, multiplier aggregation bias

JEL Classification: G31, G32, G34