Marginal Conditional Stochastic Dominance (MCSD) is an extension of the second order stochastic
dominance that considers the joint nature of return distributions. It is a useful tool for examining marginal
dominance of one asset to another conditionally to a given market return distribution for all risk-averse
investors. MCSD is superior to conventional market models in that it requires no modeling specification
and is distributional free. Although the size and value effect of equity portfolio performance has been
well documented, most of analysis relies on statistical regression description and/or linear factor models.
This manuscript applies MCSD to re-exanimate the size/value effects for international equity markets.
The empirical MCSD test reveals that U.S. value stocks outperformed the market and dominated growth
stocks for the post 1975 period. However, the phenomenon of value over growth is generally
insignificant in markets around the world, and it varies with different valuation criteria.