5.1 U.S. Equity Trailing P/E Ratios (January 1926-June 2002) However, looking only at the U.S. data probably biases our inferences, because our interest and access to this long data series on the United States is conditional on the U.S. market growing from a small, emerging market two centuries ago into by far the largest market in the world today. This survivor bias can only be corrected by painstakingly creating equivalent data sets for every market that existed over the entire time period. Fortunately, Philippe Jorion and Will Goetzmann (2002) have done this for us. Starting in 1926, they collect equity prices on 39 different equity markets and construct real price return (without dividend) approximations over periods of market disruption, mostly wars and nationalizations. Figure 5.2 displays their real capital gain estimates as a function of length of market survival. Notably, using this measure the United States was the best-performing market in the world. Whereas the real price return in the United States was 4.32 percent per annum, the median across all markets was only 0.75 percent. This difference does not appear to be explained by higher dividend returns in countries outside the United States, either: The dividend return in the United States was over 4 percent per year during this period and is about the same as a subset of other countries in the sample where Jorion and Goetzmann obtained dividend returns. On a brighter side, a gross domestic product (GDP) weighted estimate across all countries yields a real price return of approximately 4 percent, only 0.3 percent below that of the United States.2 2Jorion and Goetzmann report that the United States was 46 percent of worldwide GDP in 1921 versus only about 30 percent today.