Looking for the integral:

With $b>0, L>0, \alpha_0>b,\sigma>0, x>L$ ,

$$\phi(x;\alpha_0,\sigma_)=\int_b^\infty\alpha L^{\alpha } x^{-\alpha -1} \frac{e^{-\frac{\left(\log (\alpha -b)-\log (\alpha_0-b)+\frac{\sigma ^2}{2}\right)^2}{2 \sigma ^2}}}{\sqrt{2 \pi } \sigma (\alpha -b)}d\alpha$$

$\textbf{Background}$: This is the density of a Pareto distribution $\alpha L^{\alpha } x^{-\alpha -1} $ with its tail exponent $\alpha$ distributed according to a shifted lognormal with mean $\alpha_0$ and lower bound $b$. In other words $\alpha + b$ follows a $\text{LogNormal}\left(\log (\alpha_0-b)-\frac{\sigma ^2}{2},\sigma \right)$.


The integral doesn't appear to admit a closed form solution. But, with b=1 (which is the lower bound for b),we get:

$$\phi(x;\alpha_0,\sigma)=\frac{1}{x^2 }\sum _{i=0}^\infty \frac{1}{i!}L (\alpha_0-1)^i e^{\frac{1}{2} i (i-1) \sigma ^2} \frac{(i+\log (\frac{L}{x})}{\log (\frac{L}{x})^{1-i} }$$

This result is obtained by expanding $\alpha$ around its lower bound $b$ (which we simplified to $b=1$) and integrating each summand.

We know that as $x \to \infty$ the density $\phi$ collapses to power laws with a tail $\alpha=1$, i.e., where K is a norming constant, $\phi(x)= K x^{-2}$.

$\textbf{Expectation: }$ Even in the absence of a fully explicit density, we can extract the explicit expectation as follows. By integrating first with respect to $x$: $$ \int_b^\infty \int_L^\infty x \phi(x;\alpha)\,\mathrm{d}x \mathrm{d}\alpha= \int_L^\infty \frac{\alpha L \exp \left(-\frac{\left(2 \log (\alpha -b)-2 \log (\text{$\alpha $0}-b)+\sigma ^2\right)^2}{8 \sigma ^2}\right)}{\sqrt{2 \pi } (\alpha -1) \sigma (\alpha -b)}\mathrm{d}\alpha$$ With, again, b=1: $$ \mathbb{E}(X)= \frac{L \left(\alpha_0+e^{\sigma ^2}-1\right)}{\alpha_0-1}$$


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.