Study links emerging-market currency premia to domestic uncertainty

A paper by Şebnem Kalemli-Özcan and Liliana Varela finds that emerging-market currencies carried larger and more volatile uncovered interest parity premia than advanced-economy currencies against the dollar from 1996 to 2018. The study, discussed at the NBER’s International Seminar on Macroeconomics in Stockholm on June 24-25, uses Consensus Economics survey data covering 22 emerging markets and 18 advanced economies.

Uncovered interest parity, or UIP, is a benchmark condition in international finance. It says that the gap between a local interest rate and the US interest rate should be matched by the expected move in the exchange rate, measured here as foreign currency units per dollar. If UIP holds, investors should not earn a systematic expected excess return by borrowing in one currency and lending in another without hedging the exchange-rate risk.

Kalemli-Özcan and Varela examine the premium from the perspective of local markets and use survey expectations rather than assuming that forecast errors are purely random under rational expectations. That distinction matters because measured deviations from UIP can reflect what market participants expected at the time, not only what later happened to exchange rates.

The authors report five main findings: emerging-market UIP premia exceeded those in advanced economies; emerging-market premia were more volatile; local factors played a larger role in emerging-market premia; emerging-market interest differentials were more closely tied to domestic variables; and both local and global factors affected exchange-rate expectations and therefore interest differentials. They also find that a local factor is predicted by time-varying country policy uncertainty, using policy-uncertainty measures in the spirit of the Baker-Bloom-Davis indices.

The mechanism can operate through either side of the UIP condition. A deviation may arise because the interest-rate spread is large relative to expected currency movements, or because expected exchange-rate changes do not line up with the yield differential. A separate decomposition used by Chinn and Frankel in 2000 and by Chinn and Ito in 2024 separates the exchange risk premium from covered interest differentials, highlighting that different currency markets may have different sources of UIP failure.

Before the 2008 financial crisis, covered interest parity generally held for advanced-economy currencies, according to the discussion of earlier research. Under that condition, UIP deviations for those currencies had to come mainly through exchange risk premia. Chinn and Frankel’s work using FX4casts survey data found that UIP mostly held for advanced-economy currencies, which meant risk premia could not account for the forward-rate bias seen in Fama-style regressions.

The Kalemli-Özcan and Varela results are described as consistent with that earlier evidence rather than as a confirmation of it. Chinn and Frankel did not use interest rates in that analysis, so the stronger role of domestic factors in emerging markets could reflect a mix of exchange risk premia and covered interest differentials. Those covered differentials may be associated with political risk, different default risk, or convenience yield, according to the discussion.

Global factors also appear in the earlier evidence. Chinn and Frankel found that the VIX had a variable relationship with the exchange risk premium. The impact was smaller for non-advanced currencies on average, but differed across currencies, suggesting that pooling countries may obscure important variation.

FX4casts data cited for June 26 showed an expected 2.5% depreciation of the dollar against the euro over the following year, compared with a 1.7% forward discount. The same data showed an expected 2.6% dollar appreciation against the Korean won, compared with a minus 1.3% forward discount.

This story draws on original reporting from Econbrowser.