252regress078 12/4/07
I. (25+ points) Do all the following. Note that answers without reasons and/or citation of appropriate statistical tests receive no credit. Most answers require a statistical test, that is, stating or implying a hypothesis and showing why it is true or false by citing a table value or a p-value. If you haven’t done it lately, take a fast look at ECO 252 - Things That You Should Never Do on a Statistics Exam (or Anywhere Else)
In the Lees’ 2000 text they noted that before 1979 the Federal Reserve targeted interest rates, letting the money supply grow in such a way that the interest rates would remain stable. After 1979, the Fed switched to targeting the money supply. The Lees did a regression of Money supply against GNP (I had to replace this with GDP.), the prime rate(PrRt) and a dummy variable (Dummy)that is 1 before 1979 and zero from 1979 till 1990, when their analysis stops, They report a high R-squared, and extremely significant coefficients for the Prime Rate, GNP and the dummy variable, which seems to tell us that the Fed’s change of regime had a real effect on the money supply. Later in the text they suggest the addition of an interaction variable (GDPPR), which is the product of the Prime rate and the GDP, and a second interaction variable (GDPPR). I added the year and its square measured from 1958, population, and GDP squared. My attempt to update the Lees results was terrible discouraging. The dependent variable is M1 or its logarithm (logM1).
————— 12/3/2007 11:31:46 PM ————————————————————
Welcome to Minitab, press F1 for help.
MTB > WOpen "C:\Documents and Settings\RBOVE\My Documents\Minitab\M1PrRGDP.MTW".
Retrieving worksheet from file: 'C:\Documents and Settings\RBOVE\My
Documents\Minitab\M1PrRGDP.MTW'
Worksheet was saved on Mon Dec 03 2007
MTB > print c5 c2 c4 c6 c7 c8 c9 c10 c11 c12 c13 c14 c15
Data Display
Row C5 M1 PrRt GDP Dummy GDPPr GDPdum year yearsq
1 1959 140.0 4.50 $506.60 1 2280 506.6 1 1
2 1960 140.7 5.00 $526.40 1 2632 526.4 2 4
3 1961 145.2 4.50 $544.70 1 2451 544.7 3 9
4 1962 147.8 4.50 $585.60 1 2635 585.6 4 16
5 1963 153.3 4.50 $617.70 1 2780 617.7 5 25
6 1964 160.3 4.50 $663.60 1 2986 663.6 6 36
7 1965 167.8 4.50 $719.10 1 3236 719.1 7 49
8 1966 172.0 5.52 $787.80 1 4349 787.8 8 64
9 1967 183.3 5.50 $832.60 1 4579 832.6 9 81
10 1968 197.4 6.50 $910.00 1 5915 910.0 10 100
11 1969 203.9 8.23 $984.60 1 8103 984.6 11 121
12 1970 214.4 8.00 $1,038.50 1 8308 1038.5 12 144
13 1971 228.3 5.50 $1,127.10 1 6199 1127.1 13 169
14 1972 249.2 5.04 $1,238.30 1 6241 1238.3 14 196
15 1973 262.9 7.49 $1,382.70 1 10356 1382.7 15 225
16 1974 274.2 11.54 $1,500.00 1 17310 1500.0 16 256
17 1975 287.1 7.07 $1,638.30 1 11583 1638.3 17 289
18 1976 306.2 7.20 $1,825.30 1 13142 1825.3 18 324
19 1977 330.9 6.75 $2,030.90 1 13709 2030.9 19 361
20 1978 357.3 8.63 $2,294.70 1 19803 2294.7 20 400
21 1979 381.8 11.65 $2,563.30 0 29862 0.0 21 441
22 1980 408.5 12.63 $2,789.50 0 35231 0.0 22 484
23 1981 436.7 20.03 $3,128.40 0 62662 0.0 23 529
24 1982 474.8 16.50 $3,255.00 0 53708 0.0 24 576
25 1983 521.4 10.50 $3,536.70 0 37135 0.0 25 625
26 1984 551.6 12.60 $3,933.20 0 49558 0.0 26 676
27 1985 619.8 9.78 $4,220.30 0 41275 0.0 27 729
28 1986 724.7 8.50 $4,462.80 0 37934 0.0 28 784
29 1987 750.2 8.25 $4,739.50 0 39101 0.0 29 841
30 1988 786.7 9.00 $5,103.80 0 45934 0.0 30 900
31 1989 792.9 11.07 $5,484.40 0 60712 0.0 31 961
32 1990 824.7 10.00 $5,803.10 0 58031 0.0 32 1024
33 1991 896.9 8.50 $5,995.90 0 50965 0.0 33 1089
34 1992 1024.8 6.50 $6,337.70 0 41195 0.0 34 1156
35 1993 1129.7 6.00 $6,657.40 0 39944 0.0 35 1225
36 1994 1150.7 7.25 $7,072.20 0 51273 0.0 36 1296
37 1995 1127.4 9.00 $7,397.70 0 66579 0.0 37 1369
38 1996 1081.4 8.25 $7,816.90 0 64489 0.0 38 1444
39 1997 1072.8 8.50 $8,304.30 0 70587 0.0 39 1521
40 1998 1095.9 8.50 $8,747.00 0 74350 0.0 40 1600
41 1999 1123.0 7.75 $9,268.40 0 71830 0.0 41 1681
42 2000 1087.7 9.50 $9,817.00 0 93262 0.0 42 1764
43 2001 1182.0 6.98 $10,128.00 0 70693 0.0 43 1849
44 2002 1219.5 4.75 $10,469.60 0 49731 0.0 44 1936
45 2003 1305.5 4.22 $10,960.80 0 46255 0.0 45 2025
46 2004 1375.2 4.01 $11,685.90 0 46860 0.0 46 2116
47 2005 1373.2 6.01 $12,433.90 0 74728 0.0 47 2209
48 2006 1365.9 8.02 $13,194.70 0 105821 0.0 48 2304
Row Pop GDPsq log M1 logM1l
1 176289 256644 4.94164 4.89222
2 179979 277097 4.94663 4.94164
3 182992 296698 4.97811 4.94663
4 185771 342927 4.99586 4.97811
5 188483 381553 5.03240 4.99586
6 191141 440365 5.07705 5.03240
7 193526 517105 5.12277 5.07705
8 195576 620629 5.14749 5.12277
9 197457 693223 5.21112 5.14749
10 199399 828100 5.28523 5.21112
11 201385 969437 5.31763 5.28523
12 203984 1078482 5.36784 5.31763
13 206827 1270354 5.43066 5.36784
14 209284 1533387 5.51826 5.43066
15 211357 1911859 5.57177 5.51826
16 213342 2250000 5.61386 5.57177
17 215465 2684027 5.65983 5.61386
18 217583 3331720 5.72424 5.65983
19 219760 4124555 5.80182 5.72424
20 222095 5265648 5.87858 5.80182
21 224567 6570507 5.94490 5.87858
22 227225 7781310 6.01249 5.94490
23 229466 9786887 6.07925 6.01249
24 231664 10595025 6.16289 6.07925
25 233792 12508247 6.25652 6.16289
26 235825 15470062 6.31282 6.25652
27 237924 17810932 6.42940 6.31282
28 240133 19916584 6.58576 6.42940
29 242289 22462860 6.62034 6.58576
30 244499 26048774 6.66785 6.62034
31 246819 30078643 6.67570 6.66785
32 249623 33675970 6.71502 6.67570
33 252981 35950817 6.79894 6.71502
34 256514 40166441 6.93225 6.79894
35 259919 44320975 7.02971 6.93225
36 263126 50016013 7.04813 7.02971
37 266278 54725965 7.02767 7.04813
38 269394 61103926 6.98601 7.02767
39 272647 68961398 6.97803 6.98601
40 275854 76510009 6.99933 6.97803
41 279040 85903239 7.02376 6.99933
42 282217 96373489 6.99182 7.02376
43 285226 102576384 7.07496 6.99182
44 288126 109612524 7.10620 7.07496
45 290796 120139137 7.17434 7.10620
46 293638 136560259 7.22635 7.17434
47 296507 154601869 7.22490 7.22635
48 299398 174100108 7.21957 7.22490
I followed the course suggested by the textbook to find what variables were actually important in predicting the money supply.
Results for: M1PrRGDP.MTW
MTB > Regress c2 5 c4 c6 c7 c10 c12; Regression 1
SUBC> Constant;
SUBC> VIF;
SUBC> Brief 2.
Regression Analysis: M1 versus PrRt, GDP, Dummy, year, Pop
The regression equation is
M1 = 2874 - 19.1 PrRt + 0.0714 GDP - 115 Dummy + 46.2 year - 0.0149 Pop
Predictor Coef SE Coef T P VIF
Constant 2874 1232 2.33 0.025
PrRt -19.116 3.941 -4.85 0.000 2.241
GDP 0.07138 0.01762 4.05 0.000 62.461
Dummy -114.81 48.62 -2.36 0.023 8.260
year 46.23 15.57 2.97 0.005 668.523
Pop -0.014888 0.007176 -2.07 0.044 917.418
S = 57.7863 R-Sq = 98.4% R-Sq(adj) = 98.2%
Analysis of Variance
Source DF SS MS F P
Regression 5 8498077 1699615 508.98 0.000
Residual Error 42 140249 3339
Total 47 8638326
Source DF Seq SS
PrRt 1 3746
GDP 1 8260319
Dummy 1 139454
year 1 80187
Pop 1 14371
Unusual Observations
Obs PrRt M1 Fit SE Fit Residual St Resid
23 20.0 436.70 361.08 37.33 75.62 1.71 X
35 6.0 1129.70 982.60 18.35 147.10 2.68R
36 7.3 1150.70 986.80 14.01 163.90 2.92R
37 9.0 1127.40 975.89 11.81 151.51 2.68R
R denotes an observation with a large standardized residual.
X denotes an observation whose X value gives it large leverage.
So the regression above was my first attempt. There are several questions that can be asked at this point.
1) Why does this regression look awfully good as far as significance and the amount of the variation in the Y variable that is explained by the equation? (3)
2) There are only two coefficients here whose sign you can predict in advance. What are they, what did you predict and why and were you right? (2)
3) What does the Analysis of Variance tell us? What hypothesis did it cause you to reject?(1)
MTB > Regress c2 4 c4 c6 c7 c10 ; Regression 2
SUBC> Constant;
SUBC> VIF;
SUBC> Brief 2.
Regression Analysis: M1 versus PrRt, GDP, Dummy, year
The regression equation is
M1 = 321 - 20.7 PrRt + 0.0415 GDP - 174 Dummy + 14.5 year
Predictor Coef SE Coef T P VIF
Constant 321.24 66.06 4.86 0.000
PrRt -20.668 4.016 -5.15 0.000 2.160
GDP 0.04152 0.01055 3.94 0.000 20.791
Dummy -173.71 40.96 -4.24 0.000 5.444
year 14.530 3.077 4.72 0.000 24.254
S = 59.9651 R-Sq = 98.2% R-Sq(adj) = 98.0%
Analysis of Variance
Source DF SS MS F P
Regression 4 8483706 2120927 589.83 0.000
Residual Error 43 154620 3596
Total 47 8638326
Source DF Seq SS
PrRt 1 3746
GDP 1 8260319
Dummy 1 139454
year 1 80187
Unusual Observations
Obs PrRt M1 Fit SE Fit Residual St Resid
23 20.0 436.70 371.34 38.39 65.36 1.42 X
35 6.0 1129.70 982.21 19.04 147.49 2.59R
36 7.3 1150.70 988.13 14.53 162.57 2.79R
37 9.0 1127.40 980.00 12.08 147.40 2.51R
R denotes an observation with a large standardized residual.
X denotes an observation whose X value gives it large leverage.
MTB > Regress c2 3 c4 c6 c7 ; Regression 3
SUBC> Constant;
SUBC> VIF;
SUBC> Brief 2.
Regression Analysis: M1 versus PrRt, GDP, Dummy
The regression equation is
M1 = 451 - 14.3 PrRt + 0.0865 GDP - 240 Dummy
Predictor Coef SE Coef T P VIF
Constant 450.99 73.19 6.16 0.000
PrRt -14.269 4.605 -3.10 0.003 1.914
GDP 0.086456 0.005548 15.58 0.000 3.875
Dummy -239.76 46.90 -5.11 0.000 4.809
S = 73.0515 R-Sq = 97.3% R-Sq(adj) = 97.1%
Analysis of Variance
Source DF SS MS F P
Regression 3 8403519 2801173 524.91 0.000
Residual Error 44 234807 5337
Total 47 8638326
Source DF Seq SS
PrRt 1 3746
GDP 1 8260319
Dummy 1 139454
Unusual Observations
Obs PrRt M1 Fit SE Fit Residual St Resid
23 20.0 436.7 435.7 43.7 1.0 0.02 X
35 6.0 1129.7 941.0 20.6 188.7 2.69R
36 7.3 1150.7 959.0 16.0 191.7 2.69R
37 9.0 1127.4 962.1 14.0 165.3 2.30R
R denotes an observation with a large standardized residual.
X denotes an observation whose X value gives it large leverage.
4) What did I do to get from Regression 1 to regression 3 and why? (2)
5) Why was I now ready to quit dropping variables and do a ‘best subsets’ regression? (1) [9]
6) What would the money supply be that would be predicted for 1970 assuming that the numbers given for 1970 are correct? By what percent is it off the actual value? (2)
7) Can you make this into a rough prediction interval? Does this include the actual value for 1970? (2) [13]
MTB > BReg c2 c4 c6 c7 ; Regression 4
SUBC> NVars 1 3;
SUBC> Best 2;
SUBC> Constant.
Best Subsets Regression: M1 versus PrRt, GDP, Dummy
Response is M1
D
P u
r G m
Mallows R D m
Vars R-Sq R-Sq(adj) Cp S t P y
1 95.6 95.6 26.5 90.432 X
1 67.8 67.1 477.7 246.02 X
2 96.7 96.5 11.6 79.727 X X
2 95.7 95.5 28.1 91.197 X X
3 97.3 97.1 4.0 73.051 X X X
8) What is Regression 4 telling me to do? Why can you say that? (2)
MTB > Regress c2 3 c4 c6 c7 ; Regression 5
SUBC> GFourpack;
SUBC> RType 1;
SUBC> Constant;
SUBC> VIF;
SUBC> DW;
SUBC> Brief 2.
Regression Analysis: M1 versus PrRt, GDP, Dummy
The regression equation is
M1 = 451 - 14.3 PrRt + 0.0865 GDP - 240 Dummy
Predictor Coef SE Coef T P VIF
Constant 450.99 73.19 6.16 0.000
PrRt -14.269 4.605 -3.10 0.003 1.914
GDP 0.086456 0.005548 15.58 0.000 3.875
Dummy -239.76 46.90 -5.11 0.000 4.809
S = 73.0515 R-Sq = 97.3% R-Sq(adj) = 97.1%
Analysis of Variance
Source DF SS MS F P
Regression 3 8403519 2801173 524.91 0.000
Residual Error 44 234807 5337
Total 47 8638326
Source DF Seq SS
PrRt 1 3746
GDP 1 8260319
Dummy 1 139454
Unusual Observations
Obs PrRt M1 Fit SE Fit Residual St Resid
23 20.0 436.7 435.7 43.7 1.0 0.02 X
35 6.0 1129.7 941.0 20.6 188.7 2.69R
36 7.3 1150.7 959.0 16.0 191.7 2.69R
37 9.0 1127.4 962.1 14.0 165.3 2.30R
R denotes an observation with a large standardized residual.
X denotes an observation whose X value gives it large leverage.
Durbin-Watson statistic = 0.445619
Residual Plots for M1
9) Regression 5 is just a repeat of regression 3, but now I am doing residual analysis. What are the Durbin-Watson statistic and the plot of residuals vs. order telling me is present? What 2 conditions for regression seem to be being violated? (3) [18]
MTB > Regress c2 4 c4 c6 c7 c13; Regression 6
SUBC> GFourpack;
SUBC> RType 1;
SUBC> Constant;
SUBC> VIF;