Objective:  Using the drugs5.xls data file described in problem 14 on pp. 156-158, answer the following modified questions.  The data file contains information for 32 countries on P (pharmaceutical price levels), GDPN (per capital GDP), CVN (per capita drug consumption), PP (a 1,0 dummy indicating a country recognizes drug patents), DPC (a 1,0 dummy indicating a country applies strict price controls) and IPC (a 1,0 dummy indicating a country encourages price competition).  The P, GDPN and CVN variables are measured as a % of US values, e.g., 100 indicates equal to the US level, 50 indicates half of the US level and 150 indicates 50% greater than the US value).

Using the drugs5.xls data file. 

1.  Generate descriptive statistics (i.e., means, std. devs. & correlations) for the variables.  Briefly discuss.
2.   Run a regression of P on GDPN, CVN, PP, DPC and IPC.  Print out the regression results. 
3.  What regression "model" are you estimating?  Describe verbally and algebraically.
4.  Test whether any of the coefficients on GDPN, CVN, PP, DPC are zeros.
5.  Specification bias questions:

     a.  Rerun the regression but leave the GDPN variable out.  What should happen to the estimated coefficient of CVN if you rerun the regression but leave the GDPN variable out?  Why?  Validate your conclusions by rerunning the regression leaving the GDPN variable out.
     b.  Rerun the regression but leave the PP variable out.  What should happen to the estimated coefficients of GDPN and CVN?  Why?  Validate your conclusions by rerunning the regression leaving the GDPN variable out.