By using the Hirschman-Herfindahl index (HHC) the traditional approach
to the tax complexity hypothesis introduces a restriction into the fi
scal illusion model which has no theoretical foundation. We analyse th
e existing framework of the tax complexity hypothesis in detail and pr
opose to capture this complexity through a Hannah and Kay index. We ex
tend the theoretical framework by considering the expected return on i
nvestment in information. The empirical tests show that the HHC overes
timates the importance of size inequalities between different taxes, w
hile underestimating the impact of the number of taxes as a source of
informational costs. The expected revenue hypothesis is not supported.