The Salomon Brothers Electric Utility Model: Another Challenge to Market Efficiency

In each year from 1977 to 1981, portfolios of stocks identified as undervalued by Salomon Brothers Inc's electric utility model outperformed portfolios identified as overvalued. Furthermore, monthly returns on the undervalued portfolios ranked in the top three deciles in 34 of 58 months, whereas the most overvalued portfolios ranked in the lowest three deciles in 31 of the 58 months. The performance of the portfolios cannot be explained in terms of traditional risk measures. The two most undervalued portfolios fall in the middle of the sample in terms of beta and offer excess returns after adjustment for systematic risk. The most overvalued portfolio has the highest beta and negative risk-adjusted returns. Tests of the portfolios using Arbitrage Pricing Theory risk factors yield similar results. Could risk factors other than the market-wide variables captured by the Capital Asset Pricing Model and Arbitrage Pricing Theory affect the model's results? Undervaluation by the Salomon model does appear to be associated with small firm size and large investment in uncompleted nuclear plant. Even when these variables are taken into account, however, significant association between undervaluation and high return remains. Finally, inclusion of transaction costs diminishes the returns on the most undervalued portfolios but does not destroy their return advantage. The evidence indicates that the model distinguishes between over and undervalued stocks and suggests that investors may make profitable use of this information.