When macroeconomic conditions change rational inattention and cognitive frictions plausibly prevent markets from adjusting expectations for futures interest rates immediately and fully. This is an instance of information inefficiency. The resulting forecast errors give rise to joint predictability of currency and bond market returns. In particular, an upside shock to the rates outlook in a country heralds positive (rationally) expected returns on its currency and negative expected returns on its long-term bond. This proposition has been backed by empirical evidence for developed markets over the past 30 years.
Granziera, Eleonora and Markus Sihvonen (2020), “Bonds, currencies and expectational errors”, Bank of Finland Research Discussion Papers 7 – 2020.
The below are quotes from the paper. Cursive text and text in brackets have been added for clarity.
The post ties up with this site’s summary on macro trends.
The relation between predictability and sticky expectations
“The [academic] literature’s key findings concerning currency and bond return predictability are related: while a high short-term interest rate predicts high returns for a currency, it predicts low returns for long-term bonds denominated in this currency. Similarly, a steep slope of the yield curve predicts low returns for a currency but high returns for corresponding long-term bonds.”
“Forecast errors concerning short-term interest rates give rise to joint predictability patterns in bond and currency markets… In a sticky expectations model, short rate forecast errors tend to be particularly high after recent short rate changes. This is because it takes time for forecasters to update their predictions. This implies that high short-term interest rates relative to past short rates should predict high returns for a currency but low returns for the corresponding long-term bond.”
“Our model is based on the well-documented finding that forecasters update their short rate predictions sluggishly…[This] can be caused indirectly due to slow updating concerning factors driving interest rates…Forecast errors are [presumably] related to cognitive frictions and consistent with [rational] inattention…Moreover, [sticky expectations] follow from ambiguity averse preferences.”
“Rational models only include the risk premium channel. However, here misperceptions about economic states affect the actual currency premium in three ways.
- First, they create time-variation in (actual) expected returns due to misperceptions about future interest rates.
- Second, they create additional time-variation in (actual) expected returns due to expectational errors about future risk adjustments.
- Third, they can create time-variation in expected returns due to misperceptions concerning the long-run FX rate.
The first two channels also affect the bond risk premium. The long-run component misperception effect does not affect standard bonds due to finite maturity.”
Intuition from a model
“The economic intuition behind our key results is simple. The current home and foreign short-term interest rates are known but agents must forecast their future values. The value of a foreign currency is increasing in expected foreign short-term interest rates and the value of foreign long-term bond decreasing in expected (foreign) short-term interest rates.
When agents underpredict the path of future foreign interest rates, the value of the foreign currency is lower than under rational expectations but the value of the foreign bond higher than under rational expectations. This implies high actual [subsequent] returns for the currency but low returns for the corresponding bond.”
“This underprediction is associated with sticky expectations. When short-term interest rates increase, for example, due to a contractionary monetary policy shock, it takes time for forecasters to revise their future short-rate expectations up. This leads forecasters to underpredict the future path of short rates. As the forecasters slowly increase their expectations over future foreign short-term interest rates, the foreign currency appreciates but the value of the foreign bond falls. Before the forecasters have updated their expectations closer to rational values, the returns for a currency will be high but the returns for the bond low.”
“The key assumption…is that currencies and bonds are priced consistently with…biased expectations concerning short rates. Then, the return on a bond or currency can be decomposed into a rational risk premium, a short rate misperception effect and a risk premium misperception effect. This decomposition…holds in all models in which subjective expectations are given by a probability measure.”
“Sticky expectations gives rise to a relation between the level of short-term interest rates and the degree of underprediction concerning future interest rates. When short-term interest rates are high, they have on average increased recently. Therefore, high short-term interest rates are associated with larger underprediction concerning future interest rates. This implies that a high short-term interest rate predicts high returns for a currency but low returns for the corresponding long-term bond…Effectively the high-interest rate currency is undervalued because the investors do not expect the high-interest rate environment to persist.”
“We use our calibrated model to quantify the effect of the interest rate misperception channel. We find that it can account for most of the variation in bond and currency premia driven by changes in short rates and yield spreads.”
“We provide strong empirical evidence that supports the importance of short rate forecast errors for bond and currency returns…We focus on the G10 currencies (1980/1990s to 2019) of Australia, Canada, Germany, Japan, Norway, New Zealand, Sweden, Switzerland, U.K. and U.S. We utilize FRED to obtain data on end of month FX rates and interest rates on 3 month and 10 year government securities…In particular, we show that the same variables that predict bond and currency returns also predict survey-based expectational errors concerning FX rates and long-term interest rates.”
“For example, when (domestic or foreign) short-term interest rates are high, forecasters underestimate the future level of long-term interest rates and overestimate the future value of long-term bonds. Similarly, when foreign short-term interest rates are high relative to domestic interest rates, forecasters underestimate the future value of the foreign currency relative to the home currency. Moreover, we show that foreign currency returns tend to be particularly high, and bond returns low, when foreign short rates have recently increased.”