The imputation of the labour income of the self-employed typically relies upon the assumption that individuals of this group earn the same average hourly compensation as employees, either at the total economy or industry level. While this assumption is convenient in that it relies upon readily available information on the composition of the labour force and on the compensation of employees, it nevertheless remains somewhat simplistic and thus questionable in its validity. This shortcoming is addressed here by investigating a more refined method to impute the labour income of the self-employed in the United States. Imputations are based on the assumption that the labour income of the self-employed equals the average earnings of employees of the same sex and within the same age group, working in the same industry and having the same level of education. The proposed estimation of the labour income of the self-employed is followed by an analysis of how adjusted total labour income might impact the value of the labour share of output. Results for the United States show that applying this alternative methodology leads to a 2.5 percentage point rise in labour shares of output at the total economy level, led by larger increases of this indicator in sectors such as agriculture and hunting as well as professional, business and other service industries. The time profile in recent years, i.e. 2003-2009, of the labour share of output remains nevertheless unchanged when applying the proposed adjustment methodology.