Question

1. Describe why the MSE or MAPE is not suitable for financial data such as stock and foreign exchange!

Answer 1

Both these statistical measures are the most widely accepted for the Calculation of Std Error in the data , But in case of financial data of stocks and forex , where there is more possibility of unusually data points ( like zero(0) return achieved in a period …) this model fails to provide appropriate results and the problems associated are as follows

Error= 1/n E (At-Ft )/At …

No	Problems
1	Both “MSE” and “MAPE” are relative measurements to the (At) or the actual value in the denominator. As such if this value drops to Zero 0, the model fails.
2	The model suffers from inherent weakness in case of smaller data set .
3	Results obtained on the basis of error percentage are biased towards the smaller values and as such any single large variation are ignored
4	In case of time series based historical data predication of error becomes difficult as more than one influential factors are involved in the model
5	Model lacks relative adjustment factors required for VaR (Value art risk ) adjustments based on calculated errors

Thanks