Since Adam Smith, most economists have held the belief that trade fosters economic growth, although it has not been possible to establish a strong causal relationship. The results of growth regressions are, at best, mixed, and several historical studies have found a positive relationship between tariffs and economic growth in the nineteenth century. This paper adopts a different strategy. We look for cointegration between GDP per capita and openness for about thirty countries since 1830. About half return no cointegration – i.e. no relationship. The rest show mixed results, which change through time. An ordered probit model suggests that significantly positive relationships are more likely at low-to-middle income levels.