I have developed a regression model to help forecast bank’s income and explain the year-on-year income variance. I have 5 years of monthly data. R-squared is 60%. Only three factors are important: interest rates, stock market and seasonality.
My questions, which I hope you can help me with, are:
My questions, which I hope you can help me with, are:
- My regression line will give me monthly forecast of income but If I want to do yearly forecast, do I simply sum up the 12 months?
- What about the standard error? If my monthly income forecast is $100 and a standard error of $7, when I project yearly income forecast, do I simply multiply the standard error by 12?
- If I add unemployment rate data to the model, I can increase R-squared to 67% which I like but it will cause the model to have negative serial correlation (lower bound 2.411 and upper bound 4 with model DW value at 2.414). I am tempted to ignore that given its much higher R-squared but your advice will be much appreciated.
- In explaining the variance year-on-year, do I simply take the differences in interest rate and stock market levels x their respective coefficients?