ON THE UNDERINVOICING OF IMPORTS**This paper was written during my assignment as an Organization for Economic Development and Co-operation (OECD) Consultant in Turkey during April–July 1964. Its originality consists neither in discovering that underinvoicing of imports can occur nor in finding out that c.i.f. import statistics can be less than the corresponding f.o.b. export statistics, a fact well known to statisticians of international trade. It consists in (i) inferring the former from the latter; (ii) linking it to the phenomenon of tariffs and import control; and (iii) using the exercise to make informed inferences about the areas into which ‘black’ exchange is leaking and hence to assess better the ‘losses’ imposed by such a leakage. In my statistical work on the Turkish trade data, I have profited greatly from conversation with R. Gross, who had already noticed discrepancies between Turkish c.i.f. and partner-countries' f.o.b. values for a single year. My thanks are also due to Professors Bent Hansen and Besim Ustunel for useful discussions in the preparation of this paper.