# ou are to work on the problem sets individually. For this problem set…

You are to work on the problem sets individually. For this problem set, please find an excel data set on the course website

(pset3DATA.xls), which contains monthly data for an unnamed variable from January 1971

though August 2012.

1. Begin by restricting the sample to the time period from January 1971 through December

2011. Observations in 2012 will be used to compare the forecasting abilitily of your selected

model to actual observations of the data.

Document Preview:

ECON 4325-001

Assignment 3

October 23, 2012

The due date for this assignment is Tuesday, November 1, 2012.

Reading assignment: Chapters 7-8

Instructions: The assignment is due at the beginning of class on November 1, 2012. Please

note that late assignments will not be accepted. You are to work on the problem sets

individually. For this problem set, please find an excel data set on the course website

(pset3DATA.xls), which contains monthly data for an unnamed variable from January 1971

though August 2012.

1. Begin by restricting the sample to the time period from January 1971 through December

2011. Observations in 2012 will be used to compare the forecasting abilitily of your selected

model to actual observations of the data.

2. a. We will assume that the data series is generated from a model that is stationary.

Additonally, you can assume that there is neither a deterministic trend nor deterministic

seasonality in the data. Begin, by making an initial guess regarding the number of ARMA

terms (e.g. p and q). For this step, provide a printout of the correlogram. In addition,

provide a detailed write-up that indicates the models that could appropriately capture the

observed patterns in the sample autocorrelations and partial autocorrelations.

b. Estimate several models based on your guesses above. For each model, record the SIC

and AIC (please provide these values).

c. Select an appropriate model. Provide output related to the correlograms of the residuals

from the estimated model you selected. Provide a write-up that explains why you believe

your model is appropriate. Provide statistical details.

3. a. Provide a 8 step ahead forecast (a forecast for January 2012 through August 2012)

directly from EViews for the data series. Provide a graph that contains each of the following:

i. The forecasted series.

ii. The 95% confidence intervals about your estimated forecast.

iii. Actual observations for the data for several periods before the start of…

Attachments: