UNCERTAINTY, LONG‐RUN, AND MONETARY POLICY RISKS IN A TWO‐COUNTRY MACRO MODEL

Published date01 August 2024
AuthorKimberly A. Berg,Nelson C. Mark
Date01 August 2024
DOIhttp://doi.org/10.1111/iere.12697
INTERNATIONALECONOMIC REVIEW
Vol. 65, No. 3, August 2024 DOI: 10.1111/iere.12697
UNCERTAINTY, LONG-RUN, AND MONETARY POLICY RISKS IN A
TWO-COUNTRY MACRO MODEL∗
By Kimberly A. Berg and Nelson C. Mark
Miami University, U.S.A.; University of Notre Dame and N.B.E.R, U.S.A.
We study the currency risk premium and the forward premium bias in a two-country New Keynesian model
with production, no physical capital, and recursive utility. Monetary policy follows an interest rate feedback
rule and exogenous total factor productivity (TFP) growth follows a long-run risk process with stochastic
volatility, which we estimate from data. Withcross-country heterogeneity in TFP and monetary policy, reason-
able currency risk premia emerge under complete and incomplete markets but the forward premium bias is
trivial. We diagnose the challenge faced by this fairly standard production model to explain the forward pre-
mium bias.
1. introduction
We study the currency risk premium (deviation from uncovered interest rate parity) and
the forward premium bias in a two-country dynamic stochastic general equilibrium New
Keynesian model with production, no physical capital, and recursive utility. Monetary pol-
icy follows a Taylor-type interest rate feedback rule and exogenous total factor productivity
(TFP) growth follows a long-run risk process with stochastic volatility, which we estimate
from Australian, Canadian, Japanese, and U.S. data. We examine complete market and incom-
plete market environments and export pricing conventions of local currency pricing (LCP),
producer currency pricing (PCP), and dominant currency pricing (DCP). The model is driven
by four exogenous shocks: an outright shock to productivity growth, a shock to long-run risk,
a shock to stochastic volatility, and a monetary policy shock. In the article, country 1 is home
and country 2 is foreign. The currency risk premium always refers to the expected excess re-
turn from borrowing currency 2 and lending in currency 1. The forward premium bias always
refers to the deviation from 1 of the slope coeff‌icient from regressing the future depreciation
of currency 1 on the country 1 minus country 2 interest rate differential.
Under log-normality of the underlying shocks, the interest rate is linear in the log stochas-
tic discount factor’s (SDF) conditional mean and conditional variance. This conditional vari-
ance represents the precautionary saving component driving interest rate dynamics and we re-
fer to it as an uncertainty-risk factor.1An increase in country 2’s uncertainty-risk factor will
raise their precautionary saving and lower country 2’s interest rate. A positive currency risk
premium emerges to incentivize country 1 to borrow from country 2 to satisfy their excess
precautionary saving. Cross-country heterogeneity in monetary policy rules and productivity
growth are shown to produce systematic uncertainty-risk factor differentials.
∗Manuscript received September 2020; revised October 2023.
The useful comments and suggestions from four anonymous referees and useful correspondence with Charles En-
gel, Eric Swanson, and Chris Telmer helped to improve the article. Please address correspondence to: Nelson C.
Mark, University of Notre Dame 3060 Jenkins Nanovic Hall Notre Dame, IN 46556.E-mail: nmark@nd.edu.
1We refer to “uncertainty” not in the Knightian sense, but in the modern macro sense of high dispersion in the
underlying probability distributions. Our analysis of the uncertainty-risk factor pays particular attention to the condi-
tional variance of the log nominal SDF. It serves as a useful, preference-based measure of economic uncertainty, that
accounts for attitudes toward risk and the psychological ease of intertemporal consumption substitution.
1387
© 2024 the Economics Department of the University of Pennsylvania and the Osaka University Institute of Social
and Economic Research Association.
1388 berg and mark
Under complete markets, the currency risk premium is determined entirely by this
uncertainty-risk factor differential. One might think that the country with low uncertainty risk
would be safe and its currency, serving as a hedge asset, would earn a negative currency risk
premium. Instead, it is the high uncertainty-risk country that “pays” the currency risk pre-
mium, as in Ready et al. (2017). Under complete markets, factors that drive differences in the
uncertainty-risk factor alter the relative amount of precautionary saving, and hence, the inter-
est rate differential, but have no effect on the expected exchange rate depreciation. That is,
the risk premium is the interest rate differential induced by differences in precautionary sav-
ing. This suggests why the carry trade (the strategy of shorting the low interest rate currency
and investing in the high interest rate currency) is prof‌itable. The source of the prof‌it comes
from the interest rate differential, not from currency movements. Under incomplete markets,
the risk premium also depends on the conditional mean of the exchange rate “wedge” of
Lustig and Verdelhan (2019), but the uncertainty-risk factor differential is the dominant fac-
tor.
The conditional mean of the log SDF component of the interest rate is what drives inter-
est rates and exchange rates in a deterministic world. Fluctuations ref‌lect how agents adjust
their valuation of future consumption due to nonuncertainty-risk–related economic f‌luctua-
tions. Any effects that variations in these components have on the interest rate differential are
exactly offset by the expected exchange rate depreciation. Hence, they have no effect on the
risk premium, just as there is no risk premium in a deterministic world.
The forward premium bias/anomaly, on the other hand, is a phenomenon that is distinctly
different from the risk premium.2There is a forward premium bias when the slope coeff‌i-
cient in the regression of the one-step-ahead depreciation of country 1’s currency on the in-
terest rate differential between countries 1 and 2 deviates from 1. There is a forward premium
anomaly is when the slope coeff‌icient is negative. This regression will be referred to as the
Fama regression (Fama, 1984), and the coeff‌icient of interest will be denoted as βFand re-
ferred to as the Fama coeff‌icient. Under complete markets and log-normality, one of the terms
in the numerator of the Fama coeff‌icient is the covariance between the country 2 minus coun-
try 1 log SDF differential in conditional means and the country 2 minus country 1 log SDF dif-
ferential in conditional variances. The forward premium anomaly will be present if the covari-
ance between these differentials is a suff‌iciently large negative quantity. The forward premium
anomaly/bias is a feature of covariances and not the risk premium per se. There can be a risk
premium with no forward premium anomaly.
With heterogeneity in both monetary policy and productivity growth, the model can gen-
erate sizable risk premia with signs of mean values that correspond correctly to those found
in the data. The model generates a modest forward premium bias, but no forward premium
anomaly. This is the case under complete markets and incomplete markets and LCP, PCP, and
DCP. Although endowment models under complete markets (Bansal and Shaliastovich, 2012;
Backus et al., 2013) can generate the forward premium anomaly, we f‌ind this to be much more
challenging in a conventionally specif‌ied two-country dynamic stochastic general equilibrium
model with production. The challenge faced by our production model is, although the covari-
ances between the “own” country conditional means and conditional variances are negative
and sizable, the cross-country covariances are also negative and nearly the same size. When
the cross-country covariances are subtracted from the “own” country covariances, the two ef-
fects largely offset each other and render the Fama coeff‌icient near 1.
The remainder of the article is as follows. The next section discusses related literature.
Section 3 provides a brief presentation of the model and Section 4 discusses how the cur-
rency risk premium and the forward premium bias emerges in our setup. Section 5 reports
the model parameterization. The main results from the model are presented in Section 6 and
Section 7. Section 8 concludes.
2The distinction between the forward premium bias and the currency risk premium was explored in a different
context in Hassan and Mano (2019).

Get this document and AI-powered insights with a free trial of vLex and Vincent AI

Get Started for Free

Unlock full access with a free 7-day trial

Transform your legal research with vLex

  • Complete access to the largest collection of common law case law on one platform

  • Generate AI case summaries that instantly highlight key legal issues

  • Advanced search capabilities with precise filtering and sorting options

  • Comprehensive legal content with documents across 100+ jurisdictions

  • Trusted by 2 million professionals including top global firms

  • Access AI-Powered Research with Vincent AI: Natural language queries with verified citations

vLex

Unlock full access with a free 7-day trial

Transform your legal research with vLex

  • Complete access to the largest collection of common law case law on one platform

  • Generate AI case summaries that instantly highlight key legal issues

  • Advanced search capabilities with precise filtering and sorting options

  • Comprehensive legal content with documents across 100+ jurisdictions

  • Trusted by 2 million professionals including top global firms

  • Access AI-Powered Research with Vincent AI: Natural language queries with verified citations

vLex

Unlock full access with a free 7-day trial

Transform your legal research with vLex

  • Complete access to the largest collection of common law case law on one platform

  • Generate AI case summaries that instantly highlight key legal issues

  • Advanced search capabilities with precise filtering and sorting options

  • Comprehensive legal content with documents across 100+ jurisdictions

  • Trusted by 2 million professionals including top global firms

  • Access AI-Powered Research with Vincent AI: Natural language queries with verified citations

vLex

Unlock full access with a free 7-day trial

Transform your legal research with vLex

  • Complete access to the largest collection of common law case law on one platform

  • Generate AI case summaries that instantly highlight key legal issues

  • Advanced search capabilities with precise filtering and sorting options

  • Comprehensive legal content with documents across 100+ jurisdictions

  • Trusted by 2 million professionals including top global firms

  • Access AI-Powered Research with Vincent AI: Natural language queries with verified citations

vLex

Unlock full access with a free 7-day trial

Transform your legal research with vLex

  • Complete access to the largest collection of common law case law on one platform

  • Generate AI case summaries that instantly highlight key legal issues

  • Advanced search capabilities with precise filtering and sorting options

  • Comprehensive legal content with documents across 100+ jurisdictions

  • Trusted by 2 million professionals including top global firms

  • Access AI-Powered Research with Vincent AI: Natural language queries with verified citations

vLex

Unlock full access with a free 7-day trial

Transform your legal research with vLex

  • Complete access to the largest collection of common law case law on one platform

  • Generate AI case summaries that instantly highlight key legal issues

  • Advanced search capabilities with precise filtering and sorting options

  • Comprehensive legal content with documents across 100+ jurisdictions

  • Trusted by 2 million professionals including top global firms

  • Access AI-Powered Research with Vincent AI: Natural language queries with verified citations

vLex