The relationship was examined between government smoking control levels and eight health, social and economic indicators in 93 countries. Governmental smoking control levels were quantified by assigning a point to each control measure. The controls were as follows: health warning on cigarette package; tar, nicotine or carbon monoxide level indications on cigarette package; ban on advertising on TV or radio; and prohibiting sales to minors. The eight indicators were annual rate of population increase, infant mortality rate, population per hospital bed, number of TV sets per 1000 people, protein supply per capita per day, military expenditure as % of GNP, gross national product (GNP) and % of primary school enrollment. We analyzed predictors of the smoking control level using multiple linear regression analysis with these eight indicators. This regression model indicates that the higher the GNP is, the stronger the smoking control level is, and GNP alone accounts for 42% of the variance in smoking control levels. The set of eight indicators explained 49% of it. GNP had the largest partial regression coefficient in the standardized model. Therefore, a low GNP might be an important factor behind the difficulty in strengthening government controls on smoking.