Motivation

  1. Explain the UIP puzzle - why the high-interest rate currencies appreciate in comparison to low-interest rate currencies in the future. Also explains,
    1. why the carry-trade works so wonderfully.
    2. asymmetric exchange rate behavior
    3. time-series momentum
  2. Why does it matter - as its property of open economy models.

Mechanism

  1. Assumes the investors are ambiguity averse meaning they place greater emphasis on negative news than the positive news.
  2. UIP Explanation
    1. Agent acts as if “observed increase” in the interest rate different are caused by temporary volatile shocks whereas the “observed decrease” in the interest rate differential is the result of persistent shocks.
    2. Due to this asymmetric beliefs, the UIP agent perceives positive innovations when updating her initial estimate. $\Rightarrow$ Leads to more PV of payoff as interest rate differential increases $\Rightarrow$ higher demand for the investment currency $\Rightarrow$ The investment currency appreciates.
  3. Asymmetric Exchange Rate Behavior Explanation
    1. Meaning: High interest rate currencies appreciate slowly but depreciate suddenly.
    2. Why: Agents underreacts to increases in the interest rate in high-interest rate currencies (as they assume it is caused by temporary shock) but also overreacts to any interest rate decrease as they assume its caused by persistent shocks
  4. Time-Series Momentum or “delayed overshooting puzzle”
    1. Meaning: Currency past payoff predicts its future payoff.
    2. Currency appreciate gradually rather than suddenly as the postive innovations are gradually factored into the agent estimate. $\Rightarrow$ Building the time-series momentum

My Critique

  1. The agent believe system is endogenous to her portfolio. If she is long on the foreign bond, then any increase in the foreign interest rate is likely to be perseived as temporary. However, if she is short on the foreign bond, then any increase in the foreign interest rate is likely to be perceived as persistent. Thus, the belief system is dependent on the portfolio choice, rather than portfolio choice being dependent on the belief system.
  2. The model assumes a very high degree of persistence of state variable $\rho = 0.98$, almost a random walk. If the persistence falls to $\rho = 0.7$, then the model fails to explain the UIP puzzle. Is it realistic to assume such a high degree of persistence in real world data? The authors says this help explain the postive UIP coefficient in the Emerging Market currencies (Bansal and Dahlquist, 2000). However, for the Develeoped Markets economies is it realistic to assume such a high degree of persistence?

Slides on Discussion on UIP Puzzle and Ambiguity Aversion