Topic 6: Empirical Evidence on CIP

Question 1

Before the 2007-09 Financial Crisis

Answer 1

Frenkel and Levich (1975, 1977)
Akram, Rime and Sarno (2008)

Question 2

Frenkel and Levich (1975, 1977)

Answer 2

Many non-zero CIDs with 3-month treasury bills and eurocurrency rates (the UK-USA and for Canada-USA)

Question 3

Frenkel and Levich (1975, 1977) CASE 1: T-BILLS

Answer 3

Around 80% of profit opportunities using T-bills are not exploitable because of transaction costs

20% could reflect political risk/sovereign default risk (defaultable T-bills?)

Question 4

Frenkel and Levich (1975, 1977) CASE 2: EUROCURRENCY RATES

Answer 4

No to political risk (issued in the same, external financial centre)
No to default/credit risk (provided by a group of banks with similar credit risks)

After taking into account for the bid-ask spreads, 100% of profit opportunities using eurocurrency interest rates are gone

Question 5

Akram, Rime, and Sarno (2008)

Answer 5

Use intra-daily transaction data for EURUSD; GBPUSD; USDJPY

Very short-lived (e.g., a second) violations of CIP (CID) arise but arbitrage opportunities are exploited very rapidly

Question 6

The 2007-09 crisis and afterward

Answer 6

(1) Baba and Packer (2009)
(2) Du, Tepper, and Verdelhan (2018)
(3) Rime, Schrimpf, and Syrsatd (2019)
(4) Borio, McC8auly McGuire and Shushko (2016)

Question 7

(1) Baba and Packer (2009)

Answer 7

many non-zero CIDs during the financial crisis with eurocurrency rates (EURUSD, GBPUSD, CHFUSD)

i. No to “transaction costs”, “default risks in the bonds”, and “political risk”
ii. the counterparty risk in the forward contract

NOT AN ARBITRAGE STRATEGY BUT A RISK PREMIUM

Question 8

(2) Du, Tepper, and Verdelhan (2018)

Answer 8

many non-zero CIDs for forward contracts that appear on banks’ (i.e., arbitrageurs) balance sheets at the end of the quarter.

To avoid raising capital for quarter-end positions, one week (month) before the end of the quarter, European banks will not enter one-week (month) forward. (i.e., they don’t exploit the CIP violation.)

Question 9

(3) Rime, Schrimpf, and Syrsatd (2019)

Answer 9

Money markets are segmented.
Therefore carefully need to select rates that capture the mariginal $ funding (borrowing) costs for the critical arbitrageurs.

CIP arbitrage opportunities do indeed exist but only for high-rated banks. Even for these high-rated banks, there are limits to the scalability of the arbitrage trades because much of the CIP arbitrage profit is wiped out by the increase in $ funding costs, when they scale up their arbitrage strategy

Question 10

(4) Borio, McCauley, McGuire, and Sushko (2016

Answer 10

Typically, exploiting CIP deviations requires the arbitrageur to borrow and lend (i.e., higher leverage).

Since the crisis, pressure from shareholders, creditors and prudential authorities has reinforced and hard-wired participants’ awareness on the risk caused by high leverage.

As a result, leverage has declined since the crisis and there has been less willingness to deploy the balance sheet for activities that make heavy demands on the leverage. These make it harder to narrow the persistent deviations from CIP

Question 11

Regression based tests on UIP and profitability or carry trade strategies

Question 12

Implementing carry trade strategies in portfolios PROCESS

Answer 12

Rank all countries (except for U.S.) according to their one-month interest rates and form 6 portfolios (Ca1, …, Ca6). Thus, Ca1 is the low-interest rate portfolio and Ca6 is the high-interest rate portfolio. For each portfolio Ca, we compute excess returns of borrowing $ and investing into foreign T-Bill investments for one month (i.e., $ carry trade) by averaging across the different countries in each portfolio

(Zero-cost $ carry trade portfolio): buy portfolio Ca6 and (short) sell Ca1. The $ excess return is called as 𝐇𝐌𝐋𝐅𝐗

Question 13

Implementing carry trade strategies in portfolios THEORY (APT)

Answer 13

(1) Lustig, Roussanov and Verdelhan (2011)
(2) Menkhoetal (2011)
(3) Mancini et al.(2013)

Question 14

(1) Lustig, Roussanov and Verdelhan (2011)

Answer 14

propose the following a two-factor model (‘alpha becomes beta’)

E(rj,t) − rf,t = δjE(RXt) +γjE(HMLFX,t)

where the dollar risk factor (RX): average of six $ carry trade returns portfolios and the carry risk factor (𝐇𝐌𝐋𝐅𝐗)

Question 15

Menkho et al.(2011)

Answer 15

E(r )−r = δE(RXt)+λE(un_GVfx,t )

Idea: ‘unexpectedly high volatility’ is bad news to investors. Thus, investors ask risk premium for an asset with negative covariance with unexpected increases in global FX volatility (e.g., an asset with lower returns during unexpectedly volatile FX markets). They propose a new return factor un_GVFX

Carry trades perform poorly during times of (FX) market turmoil and, thus, their high returns can be a compensation for taking volatility risk

Question 16

Mancini et al.(2013)

Answer 16

E(rj,t) − rf,t = δjE(RXt) + λjE(un_GiLFX,t)

Idea: ‘unexpectedly high illiquidity’ is bad news to investors. Thus, investors ask risk premium for an asset with negative covariance with unexpected increases in global FX illiquidity (e.g., an asset with lower returns during unexpectedly illiquid FX markets).

They propose a new return factor un_GiLFX

Carry trades perform poorly during times of illiquid FX markets (i.e., when difficult to trade) and, thus, their high returns can be a compensation for taking
illiquidity risk

Question 17

Market friction and profitability of carry trade strategies (Brunnermeier, Nagel, and Pedersen (2009)):

Answer 17

Assume risk-neutral investors, but with short- run market frictions, UIP violations (i.e., profitable carry trades) exist in the short run but disappear in the long run.