Using Self-employment as Proxy for Entrepreneurship: Some Empirical Caveats

Research on entrepreneurship has received an increased amount of interest inrecent years, with self-employment being used as the most common proxy for“entrepreneurship” in empirical studies. However, there are various ways of defining selfemployment,making it a somewhat dubious proxy. This may flaw the analysis, especially incross-country studies, since the documentation of data often is insufficient and difficult toaccess due to language barriers. We present an analysis of Swedish self-employment data. Weshow that the measurement of self-employment has changed over time to noticeably affect thereported number of self-employed in the two major statistical sources on self-employment.The reported development of self-employment sometimes differs diametrically depending onsource. Sweden is occasionally erroneously reported to show the largest increase in self-employment in cross-country studies. Our study mimics the results of other country-specificanalyses and we conclude that well-grounded conclusions require that the advantages anddisadvantages of different statistical sources are recognized.