I see. I can always subtract two equations, even if they are not the same. E.g
\(y = 3\)
\(x= 2\)
\(y-x = 3-2\)
If you are not convinced, you can use your preferred method. For instance, move \(\beta_GV(r_m)\) and \(\frac{5}{2}\beta_HV(r_m)\) to the right-hand side so you get 0 in both left-hand sides and you can equate them. That will also lead to the same solution. You can also solve for \(\dfrac{5}{7}cov(r_G,r_H)\) in the first equation and substitute that value in the second equation, and it will lead to the same solution.
You can check