First Four Moments of the Standard Normal

Random Variables · Easy · Free problem

Let $X \sim N(0, 1)$ be a standard normal random variable. Compute

$$E[X], \qquad E[X^2], \qquad E[X^3], \qquad E[X^4].$$

Derive them (for example from the moment generating function or by integration by parts) rather than quoting them.

Hints

  1. The density is symmetric about 0. What does that immediately say about every odd moment?
  2. The MGF is $M(t) = E[e^{tX}] = e^{t^2/2}$. The $n$-th moment is $M^{(n)}(0)$; expand $e^{t^2/2}$ as a power series and read off coefficients.
  3. $e^{t^2/2} = 1 + \tfrac{t^2}{2} + \tfrac{t^4}{8} + \cdots$, and $E[X^n] = n! \times [\text{coefficient of } t^n]$, giving $E[X^4] = 4!/8 = 3$.

Worked Solution

How to Think About It: Symmetry kills the odd moments instantly. For the even moments, the MGF of a standard normal is $e^{t^2/2}$; differentiating it four times is tedious, but expanding it as a power series and matching against $M(t) = \sum E[X^n]\, t^n / n!$ takes one line. Integration by parts gives the same recursion $E[X^n] = (n-1)E[X^{n-2}]$.

Quick Estimate: $E[X^2] = 1$ by definition of a unit variance, and $E[X^4]$ must exceed $(E[X^2])^2 = 1$ by Jensen. The normal's fourth moment is the benchmark "kurtosis 3."

Formal Solution:

*Step 1 -- Odd moments by symmetry.* The density $\phi(x) = e^{-x^2/2}/\sqrt{2\pi}$ is even, so $x^n \phi(x)$ is odd for odd $n$ and integrates to zero over $\mathbb{R}$ (the integrals converge absolutely). Hence $E[X] = E[X^3] = 0$.

*Step 2 -- MGF.* Completing the square,

$$M(t) = E[e^{tX}] = \int \frac{e^{tx - x^2/2}}{\sqrt{2\pi}}\,dx = e^{t^2/2} \int \frac{e^{-(x-t)^2/2}}{\sqrt{2\pi}}\,dx = e^{t^2/2}.$$

*Step 3 -- Match coefficients.* Expanding both sides,

$$\sum_{n \ge 0} \frac{E[X^n]}{n!}\, t^n = e^{t^2/2} = \sum_{k \ge 0} \frac{t^{2k}}{2^k\, k!}.$$

So $E[X^{2k}] = \dfrac{(2k)!}{2^k k!} = (2k - 1)!!$ and the odd moments vanish. In particular $E[X^2] = 2!/2 = 1$ and $E[X^4] = 4!/(4 \cdot 2) = 24/8 = 3$.

*Step 4 -- Integration-by-parts check.* Since $\phi'(x) = -x\phi(x)$,

$$E[X^n] = \int x^{n-1} \cdot x\phi(x)\,dx = \left[-x^{n-1}\phi(x)\right]_{-\infty}^{\infty} + (n-1)\int x^{n-2}\phi(x)\,dx = (n-1)E[X^{n-2}],$$

giving $E[X^2] = 1 \cdot E[X^0] = 1$ and $E[X^4] = 3 \cdot E[X^2] = 3$.

Answer: $E[X] = 0$, $E[X^2] = 1$, $E[X^3] = 0$, $E[X^4] = 3$ (in general $E[X^{2k}] = (2k-1)!!$ and odd moments are zero).

Intuition

Odd moments vanish by symmetry, $E[X^2] = 1$ is the variance, and $E[X^4] = 3$ is the number behind "the normal has kurtosis 3." The fastest derivation is to expand the MGF $e^{t^2/2}$ and match coefficients, which also gives the general pattern $E[X^{2k}] = (2k-1)!!$. Knowing $E[X^4] = 3$ cold matters in practice: it is the baseline against which excess kurtosis of returns is measured, and it drives the variance of a sample variance and the vega/volga terms in option pricing.

Open the full interactive solver →