Multinationals companies and productivity spillovers: a meta-analysis with a test for publication bias

This paper presents the results of a meta-analysis of the literature on multinational companies and productivity spillovers. Studies in this literature examine spillovers usually within the framework of an econometric analysis in which labour productivity in domestic firms is regressed on a number of covariates assumed to have an effect on productivity, one of which is the presence of foreign firms. A positive and statistically significant coefficient on the foreign presence variable is then taken as evidence that spillovers exist. For a sample of published and unpublished studies, we collect the different coefficients on the foreign presence variable reported in the different studies, and their associated values of the t-statistic. We then regress the value of the t-statistics on a number of study characteristics, such as sample size, variable definitions used, etc. Some of these characteristics, namely, variable definitions, and whether it is a cross-section or panel analysis, have an effect on the size of the coefficient found in the productivity studies. Using a similar regression approach, we also find evidence that there may be publication bias in the literature on productivity spillovers.