Even though informality in labor markets have been studied widely, envelope wages,
a hybrid form of informal employment has only recently attracted more attention. In
this form, some employees receive two wages from their employers, an ocial wage
that is reported to the tax authorities and an envelope wage that is undeclared. This
underreporting of actual wages by the rms reduce the tax burden of the rms, but
result in unfair competition, and limits the access of employees to social protection sys-
tems. Research on envelope wages is gaining momentum recently as more data becomes
available. The institution set-up in Turkey provides a unique setting to study envelope
wages as the wages are taxed at the source. Therefore the rms have an incentive to
underreport and the households have no incentive to lie about their wages. In this
paper, we use data from two dierent data sources, one collected at the rm level and
the other at the household level, to identify individuals who may be receiving envelope
wages. Following a Heckman sample selection estimation strategy, we estimate the
coecients in a Mincerian wage equation, which are later used to construct estimated
wages for individuals whose wages may have been underreported. This paper is the
rst in the relevant literature to provide estimates of underreporting and the associated
tax losses when data on underreporting is not directly available.