The Bivariate Regular Variation of randomly weighted sums revisited in the presence of interdependence
Dimitrios G. Konstantinides, Charalampos D. Passalidis · 2025
Read The Bivariate Regular Variation of randomly weighted sums revisited in the presence of interdependence by Dimitrios G. Konstantinides, Charalampos D. Passalidis (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2506-17895.txt.
Introduction
Dimitrios Konstantinides and Charalampos Passalidis revisit how two risky sums behave together when extremes strike, in a 2025 paper. Each sum multiplies random outcomes by random weights, as when investment returns meet portfolio shares or claims meet reinsurance fractions. The danger lives in the tails, where one huge jump dominates.
The authors assume the main variables share a heavy-tailed pattern that treats the two risks unevenly, and let the weights depend weakly on them. Under moment conditions they prove the weighted sums inherit that tail pattern, with an explicit formula capturing the interplay. A stronger version holds uniformly, even for endless sums. Insurers and risk managers use such results to price the bad days when several large losses land together.
Catalog record
| CREATOR | Dimitrios G. Konstantinides, Charalampos D. Passalidis |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 72 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2506-17895
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
The Bivariate Regular Variation of randomly weighted sums revisited in the presence of interdependence Dimitrios G. Konstantinides, Charalampos D. Passalidis Abstract . We study the closure properties of the class of Bivariate Regular Variation, symbolically B R V BRV , in standard and non-standard cases, with respect to the randomly weighted sums. However, we take into consideration a weak dependence structure among the random weights and the main random variables, providing more flexibility in applications. In the first main result, under the condition that the main random variables belong to B R V BRV , we provide a dependent bi-variate extension of Breiman’s theorem, under some sufficient conditions. Furthermore, we continue the extension into randomly weighted sums. The main target motivation of this work is stochastic modeling of regular variation under the condition of economic instability. For this reason we provide the ruin probability estimation over finite-time horizon in a two-dimensional discrete-time risk model, with interdependence among the insurance and financial risks, as also the asymptotic behavior of a risk measure, known as joint expected shortfall. Keywords: Breiman’s theorem; closure properties; interdependence; insurance and financial risks; bidimensional discrete-time risk model; multivariate linear single big jump principle. Mathematics Subject Classification : Primary 62P05; Secondary 60G70. 1. Introduction Let \{(X_{i},\,Y_{i})\,,\;i\in{\mathbb{N}}\} be a sequence of independent, identically distributed random vectors, which are coppies of the general random vector (X,\,Y) , with distributions (F,\,G) . We assume that the (X,\,Y) represents a pair of non-negative random variables with tails \overline{F}=1-F , \overline{G}=1-G such that the inequalities \overline{F}(x)>0 , \overline{G}(x)>0 are true for any x\in{\mathbb{R}} . Let us also consider the sequences \{\Theta_{i},\,\Delta_{i},\;i\in{\mathbb{N}}\} of arbitrarily dependent, non-negative and non-degenerated to zero random variables. We assume that the vector (X,\,Y) follows a bivariate regular varying ( BRV ) distribution, see in Section 2.1 below, and we are interested in the behavior of the pair (S_{n}^{\Theta},\,T_{n}^{\Delta}) , in the sense of BRV structure, where \displaystyle S_{n}^{\Theta}:=\sum_{i=1}^{n}\Theta_{i}X_{i}\,,\qquad T_{n}^{\Delta}:=\sum_{i=1}^{n}\Delta_{i}Y_{i}\,, (1.1) for n\in{\mathbb{N}} , under some dependence conditions on (X_{i},\,\Theta_{i}) and on (Y_{i},\,\Delta_{i}) . Although the asymptotic behavior of the tail of a randomly weighted sum S_{n}^{\Theta} was studied extensively, see for example [ 24 ] , [ 62 ] , [ 37 ] , [ 63 ] , [ 14 ] , among others, only few papers spend attention on the two dimensional case. Even more in these few publications the independence assumption among the main variables \{(X_{i},\,Y_{i})\,,\;i\in{\mathbb{N}}\} and among the discount factors (or random weights) \{\Theta_{i},\,\Delta_{i},\;i\in{\mathbb{N}}\} makes the condition of practical usage to be pretty narrow as in case of independence between insurance and financial risks, see for example in [ 39 ] , [ 12 ] , [ 52 ] , [ 33 ] , [ 58 ] . We have to recognize that the modern insurance companies can not operate with only one risky portfolio, and this brings up the need for study of multivariate risk models. But also we feel necessary to take into account the essential dependence betweeen the insurance and financial risks, from where we get motivated to examine the closure properties of BRV for the pair (S_{n}^{\Theta},\,T_{n}^{\Delta}) , with n\in{\mathbb{N}} , under a weak dependence structure among the random variables X_{i} , \Theta_{i} , Y_{i} , \Delta_{i} , that contains independence as a special case, and also allows the vectors (X_{i},Y_{i}) and (\Theta_{i}\,\,\Delta_{i}) to have arbitrarily dependent components in such a way that we achieve generalization of the results by [ 12 ] . We should recall that the simultaneous dependence between the main random variables (X,\,Y) , but also the dependence of them with the random weights (\Theta,\,\Delta) is called interdependence, as it appears in [ 28 ] , where the uni-variate asymptotic behavior of S_{n}^{\Theta} was studied, see also in [ 7 ] for similar studies on interdependence. Nevertheless, we meet very often the interdependence effect in multivariate risk models and in risk measures with dependent portfolios, see for example in [ 23 ] , [ 16 ] , [ 31 ] . However, in all these papers was considered only the standard case of multivariate regular variation. As main domain of application of the closure properties about the pair (S_{n}^{\Theta},\,T_{n}^{\Delta}) with condition of BRV and under some interdependence structures can be the risk theory and risk management and our applications of main results, also focus on these two topics. For some results that focus on the closure properties for heavy-tailed distributions see in [ 36 ] for the uni-variate case and in [ 23 ] , [ 51 ] , [ 20 ] , [ 32 ] , [ 35 ] for the multivariate case. The rest of this paper is organized as follows. In section 2, we provide the necessary preliminary results either for the distribution classes or for the dependence structures. In section 3, we present the first main result, as an extension of Breiman’s theorem in BRV set up under interdependence. In section 4, we establish the closure property of distribution class BRV with respect to pair (S_{n}^{\Theta},\,T_{n}^{\Delta}) . We also discuss about some extensions on the randomly weighted and stopped sums. In section 5 are presented two applications of our main results. The first one relates to the Risk Theory. Namely, restricting ourselves in standard BRV case, was established the asymptotic behavior of finite-time ruin probability, in a discrete time risk model, under the presence of interdependent insurance and financial risks. In opposition to similar papers, except the generalization through interdependence, we avoid any special condition for the random measure of BRV structure of insurance risks, since usually some assumptions of asymptotic dependence is needed. The second application relates to risk management. Under the BRV structure, either standard or non-standard, when the products of these risks satisfy asymptotic dependence conditions, we show the asymptotic behavior of a risk measure, called joint expected shortfall. 2. Preliminaries For any two positive functions f,\,g , the asymptotic relation \displaystyle f(x)=o[g(x)]\,, as x\to\infty , means that it holds \displaystyle\lim_{x\to\infty}\dfrac{f(x)}{g(x)}=0\,. The asymptotic relation f(x)=O[g(x)] , as x\to\infty , means that it holds \displaystyle\limsup_{x\to\infty}\dfrac{f(x)}{g(x)}<\infty\,, while the asymptotic relation f(x)\asymp g(x) , as x\to\infty , means that both f(x)=O[g(x)] , and g(x)=O[f(x)] , as x\to\infty hold. For any event E , we denote by {\bf 1}_{E} the indicator function of the event E . By E^{c} we denote the complementary event to E . For any x,\,y real numbers, we denote x\vee y:=\max\{x,\,y\} and x\wedge y:=\min\{x,\,y\} , x^{+}:=x\vee 0 . The random variables follow distributions with support is on {\mathbb{R}}_{+}:=[0,\,\infty) , unless otherwise stated. The support of a distribution V is denoted by S_{V} and the random variable Z has distribution V . 2.1. Bivariate regular variation For a distribution V , such that \overline{V}(x)>0 , for any x\in{\mathbb{R}} , we say that it is regularly varying, with some finite index \alpha>0 if it holds \displaystyle\lim_{x\to\infty}\dfrac{\overline{V}(t\,x)}{\overline{V}(x)}=t^{-\alpha}\,, (2.1) for any t>0 , and we write symbolically V\in\mathcal{R}_{-\alpha} . The class of regular variation plays crucial role in the frame of heavy-tailed distribution, with several application in actuarial science, finance, credit risk and risk management, see for example [ 22 ] , [ 30 ] , [ 64 ] , [ 40 ] , [ 42 ] , [ 44