Among other properties of earnings forecasts, Das and Saudagaran (1998) document a difference in analyst forecast accuracy between non‐U.S. incorporated firms listed in U.S. capital markets and U.S. incorporated domestic firms. In this paper we replicate their analysis using an extended sample period (1984–1993) and a control sample of U.S. incorporated multinational firms and conclude that the results reported in Das and Saudagaran (1998) are not driven by their use of a control sample comprised of domestic firms nor are their results specific to the sample period used. We also extend their analysis to demonstrate that the differences in forecast accuracy documented in Das and Saudagaran (1998) do not extend to firms listed on the NYSE. This result that differences in analyst forecast accuracy between non‐U.S. firms listed in the U.S. and U.S. firms vary by U.S. exchange has not been examined in prior research. More importantly, we extend Das and Saudagaran (1998) by documenting the existence of cross‐country differences in analyst forecast accuracy among our sample of non‐U.S. incorporated firms listed in U.S. markets. This difference extends to firms listed on all three U.S. stock exchanges represented in our sample (the NYSE, NASDAQ, and AMEX).