There is empirical evidence of a substitution effect. For example, Barton (2001) finds
thatf inanciald erivativesa re used as a substitutef or accountinge arningsm anagementB. lack
et al. (1998) compare the use of asset sales in countries that differ with respect to the option
of revaluing noncurrent assets. While upward asset revaluations are not allowed in the
United States, they are an available option in the United Kingdom, Australia, and New
Zealand. They find that the use of asset sales to smooth income is less pronounced in
countries that allow revaluations. It is not clear from these papers what the source of the
substitution effect is.