Econ 421: Economic Forecasting and Big Data Assignment 2 In this assignment you will produce a number of simple (naive) forecasts and use the methods described in class to evaluate each forecasting model. The EXCEL spreadsheet SP500_assignment2.xlex contains monthly data on the S&P 500 stock price index from January 2000 until December 2020. In this assignment you will create some simple forecasts for a number of different horizons and evaluate them. FRED - SAP 3.40 2.00 1.An 1 207 MiA.Forecasting Models In this assignment you will evaluate the performance of the following models for forecasting SEtP index level data, at forecast horizons of h = 1 month ahead and h = 12 months ahead; The models you will evaluate are the following: 1. Naive Mean Model: Here the h-step ahead forecast is E(yl+.t I II}=j?r+hp :3?\" where It is the mean of the data from the start of the sample until period t. 2. Naive Drift Model: Here the h-step ahead forecast is in!\" = y: + hi5" where if is a measure of the drift at time period t. I want you to use two different drift models. 23' a: 2 [yr -yr]) 1 213.15! =ELFI _yl]2) The first drift model estimates the drift using the most recent two months of data and the second model estimates the drift using only the last year {local estimate of drift}. 3. Seasonal Naive Model: Here the h-step ahead forecast is yr+|r : FullI I Forecasting Exercise Use the data from cocoa until 2U1D:12 as your initial training sample. The \"holdout sample" starts January 2D11 {i.e.. 211:1). Thus+ you are to produce forecasts for January 2D11 until December EDED. Your out-of-sample forecast sample is thus 12D monthly forecasts for the period 2011:1-2D2I]:12. Plot your forecasts and forecast errors forecasts for Models 1, 2a. 2b. and 3 for your 1-step ahead predictions. Comment on the plots. What do you see? What can you say? Now, compare the forecasts from the different models and choose one model as your preferred one. Which one did you choose and why? Show all details. Now. do the same for your 12-step ahead forecasts. Which models \"wins\" this time. Comment on your findings. Output and Hints In order to carry out the forecasting exercise, you should use historical data that are in your "holdout sample", and calculate "real-time" (i.e., out-of-sample or ex ante) forecast errors for each model across the forecasting period from 2011:1-2020:12. This should be done separately for all models for h=1, and also for h=12. As stated earlier, you should plot the forecasts and forecast errors from each model. When comparing the forecast performance of the models for h-1 or for h=12, compare the MSFE (mean square forecast error) from each of the models. MSFE = Here, P denotes the number of forecasts.Hint: 1When calculating MSFEs. do it separately for each model and each forecast horizon, h, as discussed above. Separately choose a "best\" model for the h=1 horizon and for the h=f2 horizon. {i.e., the best model could he di'erent for each forecast horizon). Expectations 1 do not want to see the raw calculations (Le. the spreadsheet} but I do want to see summary statistics. You answer will be evaluated in'part on how you describe what you did. You should write up your answer under the assumption that the reader is college educated but may not have taken any statistics or econometric classes