1

ECON 6313-001

Fall Semester, 2014

Exam 2

Professor C. Brown
October 15, 2014

The data (in Excel format)you will need to complete this examination can be found at

Answer all questions. The Exam is due on Wednesday, October 22.

  1. This question is worth 35 points—3 points each, except for subparts (a) and (j), which are worth 5 ½ points each. You will need to download data for question 1 from the course web page (see URL above) to complete this question.This file contains 200monthly observations (January 1998 to August 2014) on the following variables:
    CCM:Construction machine manufacturing, seasonally adjusted, in millions of dollars.
    BPI: Number of building permits issued, all building types, U.S.
    DJIA: Dow Jones Industrial Average, monthly opening
    TWDI: Trade-weighted dollar index (broad).

a)Use regression to estimate the following model specification. Report the results of the regression—that is, report your estimates of β0, β1 , β2, and β3.

b)Are the signs of the (estimated) coefficients consistent with your (prior) expectations? Please explain.

c)Building permits issues (BPI) fell by 24,547 between June and July of 2008. What does the equation say about the impact of the decline in BPI on CCM in that month, other things being equal?

d)What is the predicted impact of a 200 point increase in the Dow Jones Industrial Average (DJIA) on the monthly value of new small trucks and SUVs produced, other things being equal?

e)Report the t-statistic for the estimate of the TWDI() coefficient. Set up a null and alternative hypothesis for this coefficient. Can you reject the null hypothesis at the .01 (1 percent) level? Briefly Explain.

f)Use the equation you estimated above to obtain a fitted value of CCM for November 2001. Is the actual value ofCCMwithin one standard error of its fitted value for this month? Explain.

g)Prepare a chart illustrating actual and fitted values of CCMfor the period January 2004 to August 2014.

h)Report the value of R2 and give an interpretation.

i)Set up an F-test. Can you reject null hypothesis at the 1 percent (.01) confidence level?

j)Use the data contained in “sheet 2” of your spreadsheet to forecast the value of construction machine manufacturing (CMM) inNovember and December, 2014. Report your results.

  1. This question is worth 35 points—3 points for each part except for subparts (e) and (j), which are worth 5 ½ points each. You will need to download data for question 2 from the website (see URL above) to complete this question. You have 151 monthly values of retail jewelry sales in the U.S. beginning in January 2002 and running through July 2014 (in millions of dollars, not seasonally adjusted).

a)Forecast jewelry sales for October 2014 using a 4-month prior moving average technique.

b)Compute root mean square error () for the in-sample forecast using the same technique as in part (a) above.

c)Find the 2-decimal point smoothing constant () which gives the best fit for (based on the criterion) for the in-sample forecast using exponential smoothing technique. (Note: the “damping factor” in Excel is equal to).

d)How does for the exponential smoothing technique compare to its value for the 5-period moving average technique?

e)Forecast jewelry sales in September 2014 using the exponential smoothing technique (using the smoothing constant you found in part (c)).

f)Estimate and report a linear trend component for the jewelry time series using the ordinary least squares (OLS) technique.

g)Compare the trend value of your series for February 2012 with its actual value in that month. What factors might account for the difference between the trend value and the actual value of jewelry sales for February 2012?

h)Compute a seasonal index using a 12 month centered moving average of the jewelry series. What are the best months in the jewelry business? What are the worst months?

i)Do an in-sample forecast on jewelry sales using the multiplicative time series technique (assume the cyclical components is equal to 1).

j)Use the information contained in following table to perform a forecast of Jewelry sales for November and December, 2014 using the multiplicative time series technique (Note: you will need to compute a trend component for these months using the equation you obtained in part (f)).

Month / Trend
Component / Seasonal
Component / Cyclical
Component / Forecast
(in millions of US$)
Nov-14 / ? / Use Nov-13 value / 0.997
Dec-14 / ? / Use Dec-13 value / 0.995
  1. This question is worth 15 points. Over a span of 170 years, the men’s world record for the mile run was reduced from 4 minutes and 56 seconds in 1804 to Roger Bannister’s 3:59.4 in 1954 to John Walker’s 3:49.4 in 1975. Regression analysis was used to fit a linear trend line to the data on world record performances since 1804. The trend function estimated is given by:

    where Time is the time in minutes and Y is the year.

a)According to the equation, the passing of one decade should reduce the record by how many seconds?

b)Sebastian Coe’s 1981 world record was 3:47.3. Hichman El Guerouj broke the world record in 1999 with a time of 3:43.1.How accurate is the equation’s forecast for each of these times?

c)What is the predicted world record time for 2010? Would it be reasonable to use the equation for 2050? Explain.

  1. This question is worth 15 points. The demand for Wanderlust Travel Services (X) is estimated to be Qx = 22,000 - 2.5Px + 4Py - 1M + 1.5Ax, where Py is the price of good Y, M is income, and Axrepresents the amount of advertising spent on X and the other variables have their usual interpretations. Suppose the price of good X is $450, good Y sells for $40, the company utilizes 3,000 units of advertising, and consumer income is $20,000.
    a. Calculate the own price (point) elasticity of demand at these values of prices, income, and advertising.
    b. Is demand elastic, inelastic, or unitary elastic?
    c. How will your answers to parts a and b change if the price of Y increases to $50?