Outline of paper Suggestions

Carry trades in Asia and the Pacific:

Evidence on ultra-low and negative interest rates
in advanced economies

Pornpinun Chantapacdepong, Hiroyuki Ito and Kieran Hull

Discussion by Michael T. Chng

Xian-Jiaotong Liverpool University, China
Deakin University, Australia

December 2, 2016
The views expressed in this presentation are the views of the author and do not necessarily reflect the views or policies of the Asian
Development Bank Institute (ADBI), the Asian Development Bank (ADB), its Board of Directors, or the governments they represent. ADBI
does not guarantee the accuracy of the data included in this paper and accepts no responsibility for any consequences of their use. Terminology
used may not necessarily be consistent with ADB official terms.
Outline of paper Suggestions

Main Research Question

Is the empirical rejection of UIP caused by an omitted variable

bias?

Time-varying exchange rate risk premium on UIP.

Risk factors and/or limits to arbitrage on CIP.

Both (zt ) are unaccounted for in the regression test for UIP.

If Cov(zt ; it,k it,k ) 6= 0, this can cause the for (it,k it,k ),
or IDt,k , to exhibit the wrong sign.
Outline of paper Suggestions

Contribution

This paper has potential in terms of the issues that it aims to

address.

To test UIP:

Fama regression: st,t+k = + (IDt,k ) + t,t+k (1)

t,t+k = t,t+k t,t+k + t,t+k (2)

where,
t,t+k : time-varying exchange risk premium from UIP

t,t+k : political or financial systems risk from CIP

t,t+k : white noise

Outline of paper Suggestions

Contribution

Unconventional monetary policy (UMP) e.g. QE, and negative

interest rate policy (NIRP) could affect both IDt,k , and
{t,t+k , t,t+k } i.e.

Cov(IDt,k , t,t+k ) 6= 0

Cov(IDt,k , t,t+k ) 6= 0

If so, the omitted variable bias induced by UMP and NIRP has two
empirical implications:
biased coefficient i.e. reject UIP

time-varying, or conditional t from rolling estimation of

Eq(1).
Outline of paper Suggestions

Suggestion 1: Low vs. Negative interest rates

The paper could further distinguish between low and negative

interest rate regimes.

Consider 3 sub-samples on 2001-2015:

Pre-GFC: Normal interest rate regimes.

Post-GFC: Low to ultra-low interest rate regime.

In recent years: Negative interest rate regime.

Outline of paper Suggestions

Suggestion 2: Time-varying betas

Consider the shift from normal to low to ultra-low to negative

interest rates on IDt,k .
Lending rates continue to fall.

Deposit rates floored at zero.

This causes downward trends in IDt,k , and reduced

cross-country dispersion in IDt,k .

Implication: Country s lose significance over time; Less

cross-country dispersion in s.
Outline of paper Suggestions

Suggestion 3: CIP risk or mispricing

Define CIP mispricing as t,k .

[ft,t+k st ] IDt,k = t,k (3)

Analyze time-series and cross-sectional properties of it,k ,

i = {1, 2, ..., N } for N country sample.
See if moving from ultra-low to negative rates affects limits to
arbitrage on CIP.
Outline of paper Suggestions

Suggestion 4: Augmented Fama regression

The paper tests an Augmented Fama regression using proxy

variables for t,t+k and t,t+k .
Stock market volatility, VIX etc.

These variables could indicate the level of financial risk. But

are they risk premiums?

See Sargo, Schneider and Wagner (2012 JFE) for ideas to

proxy t,t+k .
See Lustig, Roussanov and Verdelhan (2011 RFS) for ideas to
proxy t,t+k .
Outline of paper Suggestions

Suggestion 4: Example to proxy t,t+k

If t,t+k proxy for political or financial system risk, then:

Sort countries based on Sovereign credit rating, or SCDS
spreads.

Form a currency portfolio that goes long in high risk (H)

countries, and short in low risk (L) countries.

The HML currency portfolio return proxy for t,t+k in the

augmented Fama regression.
Outline of paper Suggestions

Conclusion

This paper is addressing an interesting and contemporaneous issue.

It could benefit from more clarifications on

whether there is a structural change from ultra low interest
rate to negative interest rate regime;
Outline of paper Suggestions

Conclusion

This paper is addressing an interesting and contemporaneous issue.

It could benefit from more clarifications on

whether there is a structural change from ultra low interest
rate to negative interest rate regime;
the rationale for using the variables to proxy t,t+k and t,t+k ;
Outline of paper Suggestions

Conclusion

This paper is addressing an interesting and contemporaneous issue.

It could benefit from more clarifications on

whether there is a structural change from ultra low interest
rate to negative interest rate regime;
the rationale for using the variables to proxy t,t+k and t,t+k ;
how the CIP mispricing of various countries behave going
normal to low to ultra-low to negative interest rates.
Outline of paper Suggestions

Conclusion

This paper is addressing an interesting and contemporaneous issue.

It could benefit from more clarifications on

whether there is a structural change from ultra low interest
rate to negative interest rate regime;
the rationale for using the variables to proxy t,t+k and t,t+k ;
how the CIP mispricing of various countries behave going
normal to low to ultra-low to negative interest rates.
I suspect mispricing will increase due to increased bank fees +
transaction costs to substitute for loss in interest margin.