Overall, McBrady and Schill find compelling evidence
that the bond issuers in their sample opportunistically
vary their issuance currencies in an attempt to
lower borrowing costs both before and after the costs
of hedging foreign currency risk are taken into account.
With regard to the former, McBrady and Schill identify
potential ‘uncovered currency bargains’ directly in
terms of the UIP relationship established in Eq. (15.2).
Beginning with this equation, they take logs on both
sides and rearrange terms to identify potential deviations
from the UIP in a manner similar to the deviations
from the CIP defined in Eq. (15.5)
As evident in Eq. (15.6), FC debt issuers’ potential uncovered
currency bargains can be decomposed into two
distinct parts: a nominal interest rate premium and an
expected rate of foreign currency appreciation.
In a series
of regressions, McBrady and Shill find that issuers
respond opportunistically to both.
Collectively, they issue
a greater share of bonds in those currencies, whose
nominal interest rates are relatively low in comparison
to other currencies in their sample.
Across all currencies
in the sample, currency shares increase by 2.4% on an average
for every 10 bps increase in nominal interest cost
savings.
McBrady and Schill also find that borrowers collectively
issue a greater share of bonds in currencies,
whose exchange rates have appreciated over the previous
year.
While estimated sensitivities are lower, with
currency shares increasing by an average of 20 bps for
each 10 bps increase in relative rates of appreciation,
the economic sensitivity is arguably comparable given
the much greater variability observed in exchange rate
movements.
Overall, this evidence suggests that bond issuers
do not believe that the UIP holds.
Instead, they appear
to believe that lower nominal interest ratesultimately translate into lower realized borrowing costs over time.
Given that borrowers issue more bonds in currencies
that have recently appreciated, they also appear
to believe that exchange rates do not tend to offset differences
in local and foreign interest rates, but instead that
they are mean-reverting over time. Interestingly, this belief
subsequently appears to be validated in the data.
The
average bond offering in McBrady and Schill’s sample
precedes a beneficial 150 bps depreciation in the issue
currency in the year following the bond’s issuance.