nations as in others. And the levels do not vary simply by region. Sweden and Portugal are similarly low, whereas Germany and Spain are similarly high. English-speaking nations outside Europe have relatively low smoking rates (22 and 20 percent in the United States and Canada, respectively).

Gender complicates the picture. According to the Tobacco Atlas, the gap in male prevalence between Sweden (17 percent) and Greece (47 percent) reaches 30 percent; that between Sweden and Germany (37 percent) reaches 20 percent. For women, the gaps are smaller but still substantial. Lower levels of 10 percent in Portugal and 18 percent in Sweden contrast with higher levels of 28 percent in Germany, 28 percent in the Netherlands, and 29 percent in Greece. The United States shows similar prevalence among men (24 percent) and women (19 percent), whereas Japan has a huge gap between men (47 percent) and women (14 percent).

Differences in smoking between the United States and European nations generate particular interest. Throughout the 1950s, the United States had higher levels of cigarette consumption than other countries (Forey et al., 2002), perhaps because it was a major source of tobacco leaf, the location of many large tobacco companies, and the source of innovative and misleading advertising about the safety of smoking (Brandt, 2007). In more recent years, however, smoking among Americans has dropped faster than in Europe, particularly among men. Cutler and Glaeser (2006) highlight this change in their paper “Why Do Europeans Smoke More Than Americans?” As discussed below, their answer to the question—differences in beliefs about the harm of smoking—offers one of several explanations for country differences.

The addictive attractions of nicotine and widespread access to cigarettes certainly play a role in the persistence across countries. But for insight into the national differences, other factors relating to government policies, social patterns of smoking, beliefs, and the timing of adoption need to be considered. The next sections review explanations of the cross-national patterns of smoking and then examine variation in smoking prevalence among high-income nations.

Epidemiologists note that population changes in smoking take a form analogous to an epidemic that spreads from relatively small parts of a population to other parts and then eventually recedes (Lopez, 1995; Lopez, Collishaw, and Piha, 1994; Mackenbach, 2006). More than changes in level, the epidemic involves a diffusion process that changes the socioeconomic composition of the smoking population. In the early stages, smoking emerges first among high-SES groups, who are most open to innovations