Decomposing exchange rate risk of Chinese foreign exchange reserves with extreme loss data

To meet the requirement of Forex reserves management, Forex reserves decomposition is required because it can offer the way each asset and its marginal change influences the whole exchange rate risk of Forex reserves. But the existing methods can not applied to highly non-linear portfolios. So this paper presents the linear local approximation method and adjusted conditional mean model to decompose the exchange rate risk of Forex reserves with the extreme loss data. Empirical study shows that under extreme market circumstance the marginal changes of assets add more exchange risk to Chinese Forex Reserves than normal market circumstance, all assets increase the exchange rate risk to the Chinese Forex Reserves, and the new added asset doesn't obviously diversify the exchange rate risk of Chinese Forex Reserves.