1/ Carry Trade and Systemic Risk: Why are FX Options so Cheap? (Caballero, Doyle)
"VIX rolldown performs at least as well as carry.
"However, hedging carry with exchange rate options produces large returns that are not a compensation for systemic risk."
papers.ssrn.com/sol3/papers.cf…
"VIX rolldown performs at least as well as carry.
"However, hedging carry with exchange rate options produces large returns that are not a compensation for systemic risk."
papers.ssrn.com/sol3/papers.cf…
2/ Realized returns for currency carry vs. returns required based on its β (in this case, β exposure to VIX futures rolldown)
Hedged with 1-month, ATM exchange rate options back-calculated from IV data
Note: The authors don't explain why this works or whether it will continue.
Hedged with 1-month, ATM exchange rate options back-calculated from IV data
Note: The authors don't explain why this works or whether it will continue.
3/ Colm O'Shea implemented a similar strategy in real time in 2006-7.
In his interview with Jack Schwager, he notes that implied volatility was too low during that period (and suggests that carry hedged with options may not always work).
In his interview with Jack Schwager, he notes that implied volatility was too low during that period (and suggests that carry hedged with options may not always work).
4/ Jamie Mai did a trade like this in 2010:
Bhansali looks at a similar strategy going further back in time and gets similar results (but does not control for exposure to VIX futures):
Bhansali looks at a similar strategy going further back in time and gets similar results (but does not control for exposure to VIX futures):
5/ Full sample: March 26, 2004 to August 24, 2012
Pre-crisis: May 19, 2004 to August 20, 2008
EQL = equal weights
SQL = weights ∝ interest rate spreads
β refers to VIX futures rolldown, not market beta
Most of carry's alpha is gone after controlling for β to VIX futures.
Pre-crisis: May 19, 2004 to August 20, 2008
EQL = equal weights
SQL = weights ∝ interest rate spreads
β refers to VIX futures rolldown, not market beta
Most of carry's alpha is gone after controlling for β to VIX futures.
6/ "Alphas are almost uniformly below average carry returns across individual currencies.
"Hedging the carry trade with a long position in VIX rolldowns leaves investors with no significant excess returns."
"Hedging the carry trade with a long position in VIX rolldowns leaves investors with no significant excess returns."
7/ "Hedging with FX options [statically; no dynamic delta hedge] leaves a significant carry return.
"Such hedging removes much of the trades' systemic [VIX futures] exposure.
"The results hold for individual currencies as well."
"Such hedging removes much of the trades' systemic [VIX futures] exposure.
"The results hold for individual currencies as well."
8/ "The gains from hedging with options comes primarily from the low interest rate quintiles.
"For the two lowest quintiles [short foreign currencies], carry trade returns were actually negative on average, but positive payoffs from options were enough to outweight the losses."
"For the two lowest quintiles [short foreign currencies], carry trade returns were actually negative on average, but positive payoffs from options were enough to outweight the losses."
9/ "The new puzzle is that there is a premium to selling systemic risk insurance when it is hedged with currency options.
After controlling for β to VIX futures, "the source of the excess return is in the low cost of the hedge, not in the high return of carry itself."
After controlling for β to VIX futures, "the source of the excess return is in the low cost of the hedge, not in the high return of carry itself."
10/ FULL-SAMPLE RESULTS
"Our full sample overstates the empirical frequency of crises. We do not see strikingly positive carry trade returns as in the pre-crisis period."
Controlling for VIX futures exposure makes alphas insignificant as it did in the pre-crisis sample.
"Our full sample overstates the empirical frequency of crises. We do not see strikingly positive carry trade returns as in the pre-crisis period."
Controlling for VIX futures exposure makes alphas insignificant as it did in the pre-crisis sample.
11/ "As before, the hedged carry trade is significant and remains so after controlling for its systemic exposure. This happens even though the unhedged carry results are not broadly significant.
"All but six currencies' t-statistics were higher for their hedged strategy."
"All but six currencies' t-statistics were higher for their hedged strategy."
12/ "All interest rate quintiles of the hedged carry strategy outperformed.
"While the gains for hedging came largely from increased expected returns for lower quintiles, the gains for high quintiles were entirely from reduced volatility."
"While the gains for hedging came largely from increased expected returns for lower quintiles, the gains for high quintiles were entirely from reduced volatility."
13/ "Finally, we implement our direct test of whether the appropriately constructed bundle of FX options provides a cheap form of systemic insurance. The results presented in Table 6 confirm our findings from the pre-crisis sample."
