Question

Suppose the standard deviation of X is 6 and the standard deviation of Y is 8. Answer the following two questions, rounding to the nearest whole number. What is Var[3X - 7Y] if the covariance of X and Y is 2?

Answer 1

Recall that

Var[aX + bY] = a ² Var[X] + 2ab Cov[X, Y] + b ² Var[Y]

Then

Var[3X - 7Y] = 9 Var[X] - 42 Cov[X, Y] + 49 Var[Y]

Now, standard deviation = square root of variance, so

Var[3X - 7Y] = 9×6² - 42×2 + 49×8² =  3376

The general result is easy to prove: by definition,

Var[X] = E[(X - E[X])²] = E[X ²] - E[X]²

Cov[X, Y] = E[(X - E[X]) (Y - E[Y])] = E[XY] - E[X] E[Y]

Then

Var[aX + bY] = E[((aX + bY) - E[aX + bY])²]

… = E[(aX + bY)²] - E[aX + bY]²

… = E[a ² X ² + 2abXY + b ² Y ²] - (a E[X] + b E[Y])²

… = E[a ² X ² + 2abXY + b ² Y ²] - (a ² E[X]² + 2 ab E[X] E[Y] + b ² E[Y]²)

… = a ² E[X ²] + 2ab E[XY] + b ² E[Y ²] - a ² E[X]² - 2 ab E[X] E[Y] - b ² E[Y]²

… = a ² (E[X ²] - E[X]²) + 2ab (E[XY] - E[X] E[Y]) + b ² (E[Y ²] - E[Y]²)

… = a ² Var[X] + 2ab Cov[X, Y] + b ² Var[Y]