* Stylized Data on Fish As Household Pets (99 Quarterly Observations *

EXHIBIT B

|_*------*_|

|_* Stylized data on fish as household pets (99 quarterly observations *_|

|_* 1968:1 to 1992:3 *_|

|_* petfish = number of households per thousand who keep fish as pets *_|

|_* home = percent of hhlds with at least one person at home during day *_|

|_* price = price of guppies, goldfish as a proxy for price of pet fish *_|

|_*------*_|

|_sample 1 99

|_read(petfish.dat) petfish home price

|_genr year=1968+(time(0)/4)

|_stat / pcor

NAME N MEAN ST. DEV VARIANCE MINIMUM MAXIMUM

PETFISH 99 86.057 9.0935 82.692 63.565 103.57

HOME 99 54.026 1.7712 3.1371 50.031 57.623

PRICE 99 58.827 7.0859 50.210 46.939 72.000

YEAR 99 1980.5 7.1807 51.563 1968.3 1992.8

CORRELATION MATRIX OF VARIABLES - 99 OBSERVATIONS

PETFISH 1.0000

HOME 0.28811 1.0000

PRICE -0.25743 0.26103 1.0000

YEAR 0.28097 -0.18905 -0.94618 1.0000

PETFISH HOME PRICE YEAR

REGRESSION B1

|_olspetfish home price / auxrsqrnoanovaexactdwresid=e predict=fitols

REQUIRED MEMORY IS PAR= 86 CURRENT PAR= 500

OLS ESTIMATION

99 OBSERVATIONS DEPENDENT VARIABLE = PETFISH

...NOTE..SAMPLE RANGE SET TO: 1, 99

DURBIN-WATSON STATISTIC = 0.17784

DURBIN-WATSON PROBABILITY = 0.000000

R-SQUARE OF HOME ON OTHER INDEPENDENT VARIABLES = 0.0681

R-SQUARE OF PRICE ON OTHER INDEPENDENT VARIABLES = 0.0681

R-SQUARE OF CONSTANT ON OTHER INDEPENDENT VARIABLES = 0.0000

R-SQUARE = 0.2017 R-SQUARE ADJUSTED = 0.1851

VARIANCE OF THE ESTIMATE-SIGMA**2 = 67.384

STANDARD ERROR OF THE ESTIMATE-SIGMA = 8.2088

SUM OF SQUARED ERRORS-SSE= 6468.9

MEAN OF DEPENDENT VARIABLE = 86.057

LOG OF THE LIKELIHOOD FUNCTION = -347.367

VARIABLE ESTIMATED STANDARD T-RATIO PARTIAL STANDARDIZED ELASTICTY

NAME COEFFICIENT ERROR 96 DF P-VALUE CORR. COEFFICIENT AT MEANS

HOME 1.9576 0.4850 4.036 1.000 0.381 0.3813 1.2290

PRICE -0.45810 0.1212 -3.779 0.000-0.360 -0.3570 -0.3131

CONSTANT 7.2439 25.31 0.2862 0.612 0.029 0.0000 0.0842

|_genr e2=e*e

REGRESSION B2

|_ols e2 home price / auxrsqrnoanova

REQUIRED MEMORY IS PAR= 10 CURRENT PAR= 500

OLS ESTIMATION

99 OBSERVATIONS DEPENDENT VARIABLE = E2

...NOTE..SAMPLE RANGE SET TO: 1, 99

R-SQUARE OF HOME ON OTHER INDEPENDENT VARIABLES = 0.0681

R-SQUARE OF PRICE ON OTHER INDEPENDENT VARIABLES = 0.0681

R-SQUARE OF CONSTANT ON OTHER INDEPENDENT VARIABLES = 0.0000

R-SQUARE = 0.0252 R-SQUARE ADJUSTED = 0.0049

VARIANCE OF THE ESTIMATE-SIGMA**2 = 6367.4

STANDARD ERROR OF THE ESTIMATE-SIGMA = 79.796

SUM OF SQUARED ERRORS-SSE= 0.61128E+06

MEAN OF DEPENDENT VARIABLE = 65.342

LOG OF THE LIKELIHOOD FUNCTION = -572.520

VARIABLE ESTIMATED STANDARD T-RATIO PARTIAL STANDARDIZED ELASTICTY

NAME COEFFICIENT ERROR 96 DF P-VALUE CORR. COEFFICIENT AT MEANS

HOME -5.1860 4.714 -1.100 0.137-0.112 -0.1148 -4.2879

PRICE 1.6218 1.178 1.376 0.914 0.139 0.1437 1.4601

CONSTANT 250.12 246.0 1.017 0.844 0.103 0.0000 3.8278

|_genr e1=lag(e,1)

..NOTE.LAG VALUE IN UNDEFINED OBSERVATIONS SET TO ZERO

|_genr e2=lag(e,2)

..NOTE.LAG VALUE IN UNDEFINED OBSERVATIONS SET TO ZERO

|_genr e3=lag(e,3)

..NOTE.LAG VALUE IN UNDEFINED OBSERVATIONS SET TO ZERO

|_genr e4=lag(e,4)

..NOTE.LAG VALUE IN UNDEFINED OBSERVATIONS SET TO ZERO

|_sample 5 99

REGRESSION B3

|_ols e e1 e2 e3 e4 / auxrsqrnoanova

REQUIRED MEMORY IS PAR= 14 CURRENT PAR= 500

OLS ESTIMATION

95 OBSERVATIONS DEPENDENT VARIABLE = E

...NOTE..SAMPLE RANGE SET TO: 5, 99

R-SQUARE OF E1 ON OTHER INDEPENDENT VARIABLES = 0.8305

R-SQUARE OF E2 ON OTHER INDEPENDENT VARIABLES = 0.9116

R-SQUARE OF E3 ON OTHER INDEPENDENT VARIABLES = 0.9128

R-SQUARE OF E4 ON OTHER INDEPENDENT VARIABLES = 0.8318

R-SQUARE OF CONSTANT ON OTHER INDEPENDENT VARIABLES = 0.0000

R-SQUARE = 0.8376 R-SQUARE ADJUSTED = 0.8304

VARIANCE OF THE ESTIMATE-SIGMA**2 = 11.162

STANDARD ERROR OF THE ESTIMATE-SIGMA = 3.3410

SUM OF SQUARED ERRORS-SSE= 1004.6

MEAN OF DEPENDENT VARIABLE = -0.31556

LOG OF THE LIKELIHOOD FUNCTION = -246.825

VARIABLE ESTIMATED STANDARD T-RATIO PARTIAL STANDARDIZED ELASTICTY

NAME COEFFICIENT ERROR 90 DF P-VALUE CORR. COEFFICIENT AT MEANS

E1 0.94800 0.1031 9.194 1.000 0.696 0.9488 0.7881

E2 -0.19885E-01 0.1413 -0.1407 0.444-0.015 -0.0201 -0.0089

E3 -0.16693E-01 0.1410 -0.1184 0.453-0.012 -0.0170 -0.0031

E4 -0.17305E-02 0.1014 -0.1707E-01 0.493-0.002 -0.0018 -0.0001

CONSTANT -0.70688E-01 0.3432 -0.2060 0.419-0.022 0.0000 0.2240

|_sample 2 99

REGRESSION B4

|_auto petfish home price / predict=fitautonoanova

REQUIRED MEMORY IS PAR= 15 CURRENT PAR= 500

DEPENDENT VARIABLE = PETFISH

..NOTE..R-SQUARE,ANOVA,RESIDUALS DONE ON ORIGINAL VARS

LEAST SQUARES ESTIMATION 98 OBSERVATIONS

BY COCHRANE-ORCUTT TYPE PROCEDURE WITH CONVERGENCE = 0.00100

ITERATION RHO LOG L.F. SSE

1 0.00000 -344.165 6443.8

2 0.90319 -248.008 889.99

3 0.94588 -247.344 873.02

4 0.94974 -247.371 872.86

LOG L.F. = -247.374 AT RHO = 0.95003

ASYMPTOTIC ASYMPTOTIC ASYMPTOTIC

ESTIMATE VARIANCE ST.ERROR T-RATIO

RHO 0.95003 0.00099 0.03153 30.12935

R-SQUARE = 0.8921 R-SQUARE ADJUSTED = 0.8899

VARIANCE OF THE ESTIMATE-SIGMA**2 = 9.1880

STANDARD ERROR OF THE ESTIMATE-SIGMA = 3.0312

SUM OF SQUARED ERRORS-SSE= 872.86

MEAN OF DEPENDENT VARIABLE = 86.093

LOG OF THE LIKELIHOOD FUNCTION = -247.374

VARIABLE ESTIMATED STANDARD T-RATIO PARTIAL STANDARDIZED ELASTICTY

NAME COEFFICIENT ERROR 95 DF P-VALUE CORR. COEFFICIENT AT MEANS

HOME 0.82179 0.3572 2.301 0.988 0.230 0.1595 0.5159

PRICE 0.26696 0.2212 1.207 0.885 0.123 0.2053 0.1821

CONSTANT 26.280 24.54 1.071 0.857 0.109 0.0000 0.3053

|_genr home1=lag(home)

|_genr price1=lag(price)

|_stat home home1 price price1 / pcor

NAME N MEAN ST. DEV VARIANCE MINIMUM MAXIMUM

HOME 98 54.042 1.7728 3.1427 50.031 57.623

HOME1 98 54.055 1.7571 3.0874 50.031 57.623

PRICE 98 58.710 7.0252 49.354 46.939 72.000

PRICE1 98 58.945 7.0249 49.349 46.939 72.000

CORRELATION MATRIX OF VARIABLES - 98 OBSERVATIONS

HOME 1.0000

HOME1 0.87540 1.0000

PRICE 0.28115 0.25601 1.0000

PRICE1 0.27664 0.24089 0.98239 1.0000

HOME HOME1 PRICE PRICE1

REGRESSION B5

|_auto petfish home home1 price price1 / noanova

DEPENDENT VARIABLE = PETFISH

..NOTE..R-SQUARE,ANOVA,RESIDUALS DONE ON ORIGINAL VARS

LEAST SQUARES ESTIMATION 98 OBSERVATIONS

BY COCHRANE-ORCUTT TYPE PROCEDURE WITH CONVERGENCE = 0.00100

ITERATION RHO LOG L.F. SSE

1 0.00000 -341.192 6064.4

2 0.89840 -245.464 845.36

3 0.93639 -245.033 834.14

4 0.93774 -245.041 834.10

5 0.93778 -245.042 834.10

LOG L.F. = -245.042 AT RHO = 0.93778

ASYMPTOTIC ASYMPTOTIC ASYMPTOTIC

ESTIMATE VARIANCE ST.ERROR T-RATIO

RHO 0.93778 0.00123 0.03508 26.73638

R-SQUARE = 0.8969 R-SQUARE ADJUSTED = 0.8925

VARIANCE OF THE ESTIMATE-SIGMA**2 = 8.9688

STANDARD ERROR OF THE ESTIMATE-SIGMA = 2.9948

SUM OF SQUARED ERRORS-SSE= 834.10

MEAN OF DEPENDENT VARIABLE = 86.093

LOG OF THE LIKELIHOOD FUNCTION = -245.042

VARIABLE ESTIMATED STANDARD T-RATIO PARTIAL STANDARDIZED ELASTICTY

NAME COEFFICIENT ERROR 93 DF P-VALUE CORR. COEFFICIENT AT MEANS

HOME 0.85134 0.3543 2.403 0.991 0.242 0.1652 0.5344

HOME1 -0.36913E-01 0.3599 -0.1026 0.459-0.011 -0.0071 -0.0232

PRICE 0.29479 0.2174 1.356 0.911 0.139 0.2268 0.2010

PRICE1 -0.45778 0.2171 -2.109 0.019-0.214 -0.3521 -0.3134

CONSTANT 52.014 31.79 1.636 0.947 0.167 0.0000 0.6042

|_test

|_test home1=0

|_test price1=0

|_end

F STATISTIC = 2.2316429 WITH 2 AND 93 D.F. P-VALUE= 0.11307

WALD CHI-SQUARE STATISTIC = 4.4632859 WITH 2 D.F. P-VALUE= 0.10735

UPPER BOUND ON P-VALUE BY CHEBYCHEV INEQUALITY = 0.44810