That is the topic of the paper By Şebnem Kalemli-Özcan Liliana Varela that I had the opportunity to discuss at the NBER International Seminar on Macroeconomics (Stockholm, June 24-25).
This paper, which includes both industrialized and emerging market currencies (against the dollar) investigated this factor — the UIP premium from a local market perspective:
Where i local interest rate, iIn the US is the US interest rate, and us is the log exchange rate (in foreign currency units per USD). The superscript denotes the expected value of the response. If equal interest is held, the expected changes in the exchange rate will be equal to the difference in interest rates
Using research data – therefore the distribution of rational expectations in the equilibrium hypothesis that assumes that the errors of expectations are new strategies – the authors find 22 EM’s, 18 AE’s, 1996-2018 (from Consensus Economics):
- EM UIP premium > AE UIP Premium, σEM > σAM
- EM UIP premium driven by local factors
- EM Int is highly variable associated with local factors
- Local and global factors affect exchange rate expectations, hence the interest rate differential
- A predictable local factor is the country’s time-varying policy uncertainty
This finding is quite interesting. To summarize some of the main findings (there are many interesting ones in the paper):
The authors found this to be the case when they developed Baker-Bloom-Davis models of policy uncertainty indicators for countries outside the United States (and other risk factors).
In other words, deviations from the UIP can arise from interest rate differences, or from changes in the expected exchange rate. Another, more common, decomposition is used by Chinn and Frankel (2000) and Chinn and Ito (2024):
Once one uses this decomposition, it becomes clear that different markets will have different decisions to deviate from the UIP.
In general, before 2008, it covered the equivalent interest held in developed economies’ currencies, so deviations from UIP had to come from exchange risk. Chinn and Frankel (2000), using FX4casts survey data, wrote that in developed economies, the UIP was held too high, so that the risk premium could not explain the forward rate bias (ie, as in Fama regression):
Figure 1: The depreciation coefficient of the forward discount. Source: Chinn and Frankel (2020).
In other words, the findings of Kalemli-Ozcan & Varela were consistent with the findings of Chinn-Frankel that in developed economies, the results of Fama’s regression were mainly due to the bias of research data; while in emerging market funds, it was more likely to be due to greater risk.
I say it’s compatible, but I can’t guarantee it. That’s because Chinn and Frankel didn’t use interest rates, so it’s probably a combination of risk premia. again it includes differences in returns (due to political risk, different default risk, or simple yield) that explain the greater role of domestic factors in emerging markets.
That being said, while Klaemli-Ozcan and Varela found a key role for domestic factors in the volatility of emerging market returns, Chinn and Frankel found a variable role for global factors (VIX).
Figure 2: The regression coefficient of risk premium on log Vix. Source: Chinn and Frankel (2020).
The global factor has a small effect on developed currencies, but the effect is not the same. Integrating across currencies (as in Kalemli-Ozcan and Varela) may be appropriate, but these results suggest otherwise.
As of June 26, FX4casts shows a 2.5% depreciation of the dollar against the euro over the next year (compared to a prior discount of 1.7%) but a 2.6% appreciation against the Korean won (-1.3% prior discount). On average, the dollar will appreciate against the euro, and depreciate against the Korean won under these conditions.
