Long-term correlations in European socioeconomic conditions create a bias to conclusion that an invasion debt occurs.

Essl et al. (1) tested whether historical socioeconomic conditions have created an invasion debt in Europe (i.e., a suite of nonnative species that have not yet been discovered but whose establishment has been determined by previous socioeconomic conditions). Essl et al. (1) tested this theory by asking whether the cumulative numbers of nonnative species presently established in 28 European countries were more closely correlated to socioeconomic conditions in the year 1900 or 2000. In this letter, we describe a flaw in their logic that biases the analysis to concluding that an invasion debt does occur. Essl et al. (1) modeled the response variable (cumulative establishments in 2009) as a function of socioeconomic variables from a single year. Because of data limitations, they did not estimate a more complete model in which cumulative establishments are a function of cumulative human activity (1). Because socioeconomic variables are correlated over time, using an explanatory variable from a single year served as a proxy for preceding and subsequent years; socioeconomic data for those surrounding years were excluded. A socioeconomic data point from 1900 in this model served as a good proxy for years immediately surrounding 1900 and a rough proxy for an extended time frame that extended for decades. For example, examining data for the 16 countries in the dataset by Essl et al. (1) for which gross domestic product (GDP) levels were available from 1960 to 2006 (2) shows that interannual correlation in GDP with respect to the middle year of 1983 declines only to 0.962 in 1960 and to 0.992 in 2006. This shows that relative levels of GDP among these countries are highly correlated for at least two decades before and after any given year. This stability drives a bias against the explanatory power of socioeconomic explanatory variables from 2000, because the degree to which they can serve as proxies for human activity in years after 2000 is truncated at 2009, the cutoff for the response variable. This bias would be greater if relative changes in socioeconomic conditions among these countries were occurring more slowly around 1900 than around 2000. Essl et al. (1) concluded that there is an invasion debt in Europe, because socioeconomic conditions in 1900 are better correlated to the current number of established species than socioeconomic conditions in 2000. The logical flaw described above biases the analysis to this conclusion. Although the invasion debt hypothesis is a plausible conjecture and an important dynamic to understand, the results presented in the work by Essl et al. (1) are not sufficient to support this claim for Europe.