Predicting Currency Crises Using Interbank Rates

January 31, 2010

Share
We propose exploiting the term structure of relative interest rates to obtain estimates of changes in the timing of a currency crisis as perceived by market participants. Our indicator can be used to evaluate the relative probability of a crisis occurring in one week as compared to a crisis happening after one week but in less than a month. We give empirical evidence that the indicator performs well for two important currency crises in Eastern Europe: the crisis in the Czech Republic in 1997 and the Russian crisis in 1998.

The indicator looks for a downward sloping yield curve using a 7 day vs 30 day interbank interest rate.
  • Fixed and crawling pegs or fluctuation bands are most prone to speculative attack
  • Current early warning systems use Fundamental data
    • Current account Balances
    • Exchange rate overvaluation
    • Export growth
    • HistorySquared notes that Carmen and Reinhart and others have found many other Leading Indicators such as
      • Jumps in FDI
      • Credit Growth Surges
      • Money Supply Spikes
      • Debt Spikes
      • Global Banking Crises
      • Commodity price collapses
      • Housing Bubbles
  • Problem with current Early Warning Systems is limited availability and attempt to predict up to 2 years out
  • Early Warning System indicator by Kaminsky, Lizondo, and Reinhart
    • Failed to warn of thai crises until ex poste, more explanatory than predictive
    • Uses economic indicators
    • Weighted factors by relevance
    • Incorporates deviations and thresholds
  • Early Warning Systems by Collins approach
    • Based on the Uncovered Interest Rate Parity theorum (UIP)
    • Looks at the Term Structure of Interest rates
    • Forecasts 1 week out
  • Based on the Collins approach, this paper uses the daily interbank rates with maturity one week (it;7and it;7) and one month (it;30 and it;30) looking for a downward sloping yield curve
  • Example :
  • 1997 Czech Republic crisis
    • Tight peg against DM and USD with bands of +0.5% to 7.5% until crises in May 1997
    • Trade Balance turned negative in 1996
    • Economic growth slowed
    • Upward Trend in currency from 1992-1997
    • Heavy central bank intervention in week before crises kept it within the band
    • Indicator was positive since early January 1997, accelerating in the week before the crises
      • Compared 1 week to 1 month interest rates
      • Alternative is the forward rate
    • Simple rate of change performed well to determine the threshold
    • Leading indicator in Tranquil times
      • Performed ok 5 days out for the Czech Republic but not for DM
  • Conclusion:
    • “To sum up, our indicator is able to identify both speculative pressures that were successfully combated by the central bank in the pre-crisis period, and starts signaling the impending crisis six days before the official announcement of the devaluation.”
    Jesús Crespo Cuares

Predictable Changes in Yields and Forward Rates

Related Posts Plugin for WordPress, Blogger...

Popularity: 5% [?]

Share

Previous post:

Next post: