Why should integrating the product of PDF and CDF have the same result given the same mean to standard deviation ratio?

83 Views Asked by At

Using the following expression where $f(x)$ is the PDF of a normal distribution with a mean $\mu$ and standard deviation $\sigma$, and $F$ is the CDF of another normal distribution with mean of $0$, but the same standard deviation:

$$ \int_{-\infty}^{\infty}f(x|\mu,\sigma)(F(x|\mu=0, \sigma)^V)dx $$

$V$ is always an integer $\ge 1$. I've noticed from simulations that the results are always the same whenever the ratio $\mu/\sigma$ is the same. In other words:

$$ \int_{-\infty}^{\infty}f(x|\mu=\mu_a,\sigma=\sigma_a)(F(x|\mu=0, \sigma=\sigma_a)^V)dx = \int_{-\infty}^{\infty}f(x|\mu=\mu_b,\sigma=\sigma_b)(F(x|\mu=0, \sigma=\sigma_b)^V)dx $$

whenever $\frac{\mu_a}{\sigma_a} = \frac{\mu_b}{\sigma_b}$, for any $V\geq 1$.

It is not clear to me why this should be true, or how to show it. Is there a way to prove this relationship?

1

There are 1 best solutions below

1
On BEST ANSWER

Let $\phi$ and $\Phi$ be the standard normal pdf and cdf. When the mean is $\mu$ and the standard deviation is $\sigma$ the pdf $f(x)$ and cdf $F(x)$ are $$ f(x)=\frac{1}{\sigma}\phi\left(\frac{x-\mu}{\sigma}\right),\quad\quad F(x)=\Phi\left(\frac{x-\mu}{\sigma}\right). $$ By a change of variable it is obvious that the integral \begin{eqnarray} \frac{1}{\sigma}\int_{-\infty}^{+\infty}\phi\left(\frac{x-\mu}{\sigma}\right)\Phi^V\left(\frac{x}{\sigma}\right)\,dx =\int_{-\infty}^{+\infty}\phi\left(x-\frac{\mu}{\sigma}\right)\Phi^V(x)\,dx,\quad\quad V\ge 1, \end{eqnarray} only depends on $\mu/\sigma$. The same holds for the integral \begin{eqnarray} \frac{1}{\sigma}\int_{-\infty}^{+\infty}\phi\left(\frac{x-\mu}{\sigma}\right)\Phi^V\left(\frac{x-\mu}{\sigma}\right)\,dx =\int_{-\infty}^{+\infty}\phi\left(x-\frac{\mu}{\sigma}\right)\Phi^V\left(x-\frac{\mu}{\sigma}\right)\,dx. \end{eqnarray}