A connection between supermodular ordering and positive/negative association

In this paper, we show that a vector of positively/negatively associated random variables is larger/smaller than the vector of their independent duplicates with respect to the supermodular order. In that way, we solve an open problem posed by Hu (Chinese J. Appl. Probab. Statist. 16 (2000) 133) refering to whether negative association implies negative superadditive dependence, and at the same time to an open problem stated in Muller and Stoyan (Comparison Methods for Stochastic Modes and Risks, Wiley, Chichester, 2002) whether association implies positive supermodular dependence. Therefore, some well-known results concerning sums and maximum partial sums of positively/negatively associated random variables are obtained as an immediate consequence. The aforementioned result can be exploited to give useful probability inequalities. Consequently, as an application we provide an improvement of the Kolmogorov-type inequality of Matula (Statist. Probab. Lett. 15 (1992) 209) for negatively associated random variables. Moreover, a Rosenthal-type inequality for associated random variables is presented.

[1]  I. Olkin,et al.  Inequalities: Theory of Majorization and Its Applications , 1980 .

[2]  Michael V. Boutsikas,et al.  On the distance between convex-ordered random variables, with applications , 2002, Advances in Applied Probability.

[3]  Henry W. Block,et al.  L-superadditive structure functions , 1989, Advances in Applied Probability.

[4]  Pingyan Chen,et al.  The Hájeck-Rényi inequality for the NA random variables and its application , 1999 .

[5]  Alfred Müller Stochastic Ordering of Multivariate Normal Distributions , 2001 .

[6]  Moment inequalities and weak convergence for negatively associated sequences , 1997 .

[7]  D. Walkup,et al.  Association of Random Variables, with Applications , 1967 .

[8]  Alfred Müller,et al.  MODELING AND COMPARING DEPENDENCIES IN MULTIVARIATE RISK PORTFOLIOS , 1998 .

[9]  Michel Denuit,et al.  Smooth generators of integral stochastic orders , 2002 .

[10]  N. Bäuerle Inequalities for stochastic models via supermodular orderings , 1997 .

[11]  H. Rosenthal On the subspaces ofLp(p>2) spanned by sequences of independent random variables , 1970 .

[12]  Qi-Man Shao,et al.  A Comparison Theorem on Moment Inequalities Between Negatively Associated and Independent Random Variables , 2000 .

[13]  J. George Shanthikumar,et al.  Regularity of Stochastic Processes: A Theory Based on Directional Convexity , 1993 .

[14]  Donald M. Topkis,et al.  Minimizing a Submodular Function on a Lattice , 1978, Oper. Res..

[15]  A. Khursheed,et al.  Positive dependence in multivariate distributions , 1981 .

[16]  A. Müller,et al.  Comparison Methods for Stochastic Models and Risks , 2002 .

[17]  Moshe Shaked,et al.  Stochastic orders and their applications , 1994 .

[18]  R. Szekli Stochastic Ordering and Dependence in Applied Probability , 1995 .

[19]  Przemysław Matuła,et al.  A note on the almost sure convergence of sums of negatively dependent random variables , 1992 .

[20]  A. Müller,et al.  Some Remarks on the Supermodular Order , 2000 .

[21]  Michel Denuit,et al.  Does Positive Dependence between Individual Risks Increase Stop-Loss Premiums? , 2001 .

[22]  H. Taizhong Negatively Superadditive Dependence of Random Variables with Applications , 2000 .

[23]  A. Tchen Inequalities for distributions with given marginals , 1976 .

[24]  Moshe Shaked,et al.  Supermodular Stochastic Orders and Positive Dependence of Random Vectors , 1997 .

[25]  K. Fan Subadditive functions on a distributive lattice and an extension of Szász's inequality , 1967 .

[26]  A. Müller Stop-loss order for portfolios of dependent risks , 1997 .

[27]  K. Joag-dev,et al.  Negative Association of Random Variables with Applications , 1983 .