Medical Progress Report No. 1  December 2004


Are Drug Price Controls Good for Your Health?

Appendix: Multiple Regression Findings

Dependent Variable: D (LRPP)
Method: Least Squares
Date: 05/25/04 Time: 10:57
Sample: 1960–2001
Included observations: 42

Variable

Coefficient

Standard Error

t-Statistic

P-Value

D(LGDP(-1))

-0.389742

0.104793

-3.719164

0.0007

D(LPUB(-1))

-0.086158

0.036614

-2.353126

0.0242

D(LPUB(-1))*D92

-0.579867

0.223008

-2.600213

0.0134

D(LRPP(-1))

0.854943

0.131145

6.519069

0.0000

D(LRPP(-2))

-0.375947

0.127824

-2.941118

0.0057

WH

0.035633

0.007753

4.596012

0.0001

R-squared

0.865946

Mean dependent variable

-0.002188

Adjusted R-squared

0.847327

S.D. dependent variable

0.039870

S.E. of regression

0.015578

Akaike info criterion

-5.354293

Sum squared residuals

0.008737

Schwarz criterion

-5.106055

Log likelihood

118.4402

Durbin-Watson statistic

1.950186

 

 

 

 

 

Definitions of the Variables:

D(LRPP) = change from one year to the next of the logarithm of the real drug price, defined as the ratio of the pharmaceutical price index to general consumer price index.

D(LGDP(-1)) = lagged value of the change from one year to the next of the logarithm of real gross domestic product per capita (measure of the growth of economic activity).

D(LPUB(-1)) = lagged value of the change from one year to the next of the logarithm of the government’s share of pharmaceutical spending (growth of government influence).

D92 = 0/1 dummy variable taking on the value of 1 for years after 1991 to capture the effects of OBRA of 1990 and the Veterans Act of 1992 (change in the growth of government influence).

D(LRPP(-1)), D(LRPP(-2)) = lagged values of the change from one year to the next of the real drug price (i.e., the prior one and two years of real drug price growth).

WH = 0/1 dummy variable taking on the value of 1 after 1983 to capture the presence of the Waxman/Hatch Act.

[figure 1] [figure 2] [figure 3] [figure 4] [appendix]