its coffee. These rebates were designed to achieve a “price discriminating” reduction in the
export tax paid by some importers, thereby increasing Brazil’s exports and net export revenues.
However, since the rebates conveyed substantial income to recipients, coffee importers avidly
sought them. Brazilian policy makers/bureaucrats were apparently persuaded to issue growing
amounts of rebates even after the export quota was filled, presumably by rent seeking activity.
There is also evidence of significant irregularities in the rebates’ use that seems to have benefited
domestic exporters as well as some policy makers and/or bureaucrats (Jarvis 2001).
I show econometrically that the export tax rebates stimulated foreign demand for
Brazilian coffee, increasing its price relative to those of its competitors. However, Brazil’s
export price rose less than the amount of the export tax rebate so that the net price fell. This
caused a large real transfer of domestic coffee quota rents to foreign importers and/or consumers.
Indeed, as a result of the export tax rebates, foreign roasters may have gained more from the ICA
quota than did Brazil.
The ICA export quota created a market context within which the export tax rebates
appeared attractive to Brazil, i.e., the export quota encouraged further government intervention in
the market and, though the rents created, provided the tax revenues that facilitated payment of
the rebates. Nonetheless, it is surprising that Brazil used coffee export tax rebates to effectively
transfer a large share of its domestic ICA quota rent to foreign roasters. The ICA was designed
to increase coffee export revenues for the benefit of coffee exporting countries, not foreign
roasters, and foreign roasters had no political muscle in Brazil. Brazilians with whom I talked at
the beginning of this study, both in the private and the public sector, consistently expressed a
belief that foreign roasters had not received any transfer of rents. Certainly none of the
Brazilians that I have talked with believed that such a transfer was warranted. I conclude that
25