0
$\begingroup$

can anyone help me on this?

A claim size distribution of an insurance portfolio follows a Pareto distribution given by:

$f(x)=(\frac{\alpha}{\beta})*(\frac{\beta}{(\beta+x)})^{(\alpha+1)}$

I am trying to derive a formula for the $r^{th}$ moment $\alpha_{r}$ of the Pareto distribution in terms of its $(r-1)^{th}$ moment $\alpha_{r-1}$ and from this expression find, $\alpha_2$, $\alpha_3$ and $\alpha_4$ using the known result for the mean $\mu$ of the above Pareto distribution.

I found that the mean of the Pareto distribution is given by:

$$\mu=\int_{\beta}^{\infty} (u-\beta)\frac{\alpha}{\beta}\frac{\beta^{^{(\alpha+1)}}}{u^{(\alpha+1)}} du$$which is:

$$\alpha\beta^{\alpha}\int_{\beta}^{\infty}\frac{1}{u^{\alpha}}-\frac{1}{u^{(\alpha+1)}} du=\frac{\beta}{\alpha-1}$$

I also find that the second moment is given by:

$$\alpha_2=\alpha\beta^{\alpha}\int_{0}^{\infty}\frac{x^{2}}{(\beta+x)^{\alpha+1}} dx$$, where $\beta+x=u$ then,

$$\alpha\beta^{\alpha}\int_{\beta}^{\infty}\frac{u^{2}-2u\beta+\beta^{2}}{(u)^{\alpha+1}} du=\frac{2\beta^{2}}{(\alpha-2)(\alpha-1)}$$

I am not sure if this is correct and I could not find anything regarding $\alpha_3$ and $\alpha_4$ which I believe they are the skewness and kurtosis of this Pareto distribution.

In terms of the $r^{th}$ moment formula I really cannot understand what I should do to derive the formula. I read that there are the $r^{th}$ moment about the origin that for a continues is given by

$$\alpha_r=E(x^{r})=\int_{}^{} x^r f(x) dx$$

And also the $r^{th}$ moment about the mean is given by:

$\mu_r=E[(x-\mu)^{r}]$

Can anyone give me any tip in how to continue and let me know if what I did is correct and can be somehow of any use for this question?

Thanks

$\endgroup$
1
$\begingroup$

In general for positive integers $k$, $$\begin{align*} \operatorname{E}[X^k] &= \alpha\beta^\alpha \int_{x=0}^\infty x^k (x+\beta)^{-(\alpha+1)} \, dx \\ &= \alpha\beta^\alpha \int_{x=\beta}^\infty (x-\beta)^k x^{-(\alpha+1)} \, dx \\ &= \alpha\beta^\alpha \int_{x=\beta}^\infty x^{-(\alpha+1)} \sum_{i=0}^k \binom{k}{i} x^{k-i} (-\beta)^i \\ &= \alpha\beta^\alpha \sum_{i=0}^k \binom{k}{i}(-\beta)^i \int_{x=\beta}^\infty x^{k-(i+\alpha+1)} \, dx \\ &= \alpha\beta^\alpha \sum_{i=0}^k \binom{k}{i}(-\beta)^i \left[ \frac{x^{k-(\alpha+i)}}{k-(\alpha+i)} \right]_{x=\beta}^\infty \\ &= \alpha\beta^\alpha \sum_{i=0}^k \binom{k}{i}\frac{-(-\beta)^i \beta^{k-(\alpha+i)}}{k-(\alpha+i)}, \quad k < \alpha \\ &= \alpha\beta^\alpha \beta^{k-\alpha} \sum_{i=0}^k \binom{k}{i} \frac{(-1)^{i+1}}{k-(\alpha+i)} \\ &= \alpha\beta^k \sum_{i=0}^k \binom{k}{i} \frac{(-1)^i}{\alpha+i - k}. \end{align*}$$

So for $k = 1$, we get $$\operatorname{E}[X] = \alpha\beta \left( \frac{1}{\alpha-1} - \frac{1}{\alpha} \right) = \frac{\beta}{\alpha-1},$$ and for $k = 2$, $$\operatorname{E}[X^2] = \alpha\beta^2 \left( \frac{1}{\alpha-2} - \frac{2}{\alpha-1} + \frac{1}{\alpha} \right) = \frac{2\beta^2}{(\alpha-2)(\alpha-1)}.$$ Next, $$\operatorname{E}[X^3] = \alpha\beta^3 \left(\frac{1}{\alpha-3} - \frac{3}{\alpha-2} + \frac{3}{\alpha-1} - \frac{1}{\alpha} \right) = \frac{6\beta^3}{(\alpha-3)(\alpha-2)(\alpha-1)}.$$ Now you should suspect a pattern; namely, $$\operatorname{E}[X^k] = \frac{k! \beta^k}{\prod_{i=1}^k (\alpha-i)} = \beta^k \binom{\alpha-1}{k}^{-1}, \quad k < \alpha.$$ The proof of this last step is not too difficult, so I have left it as an exercise.

$\endgroup$
  • $\begingroup$ Hi, how the first formula fot he general k is given in terms of its $(r-1)^{th}$ moment $\alpha_{r-1}$? I will continue the exercise for $\alpha_4$. Thank you very much for your help. $\endgroup$ – user290335 Mar 10 '18 at 14:03

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.