Next, we examine whether our results are robust to the use of an alternative
exchange-rate index to estimate a firm’s exposure in the first stage of the estimation.
Instead of using J.P. Morgan’s narrow, nominal dollar index (against 18 currencies),
we use the broad, real, dollar index published by the Dallas Fed (against 101
currencies). Although firm exposure might be captured better using a broader currency
index, the impact of derivative use might be captured better using a narrow,
nominal currency index. However, the choice of index does not affect our results
(Regression 2, Table 3), since the use of derivatives (foreign sales) is also significantly
negatively (positively) related to a firm’s exchange-rate exposure estimated
based on the Dallas Fed exchange-rate index.