Japan J. Indust. Appl. Math. (2011) 28:301–313
DOI 10.1007/s13160-011-0040-2
ORIGINAL PAPER
Area 1
On the probability of completeness for large markets
John A. Wright · Phillip S. C. Yam ·
Hailiang Yang
Received: 19 July 2010 / Revised: 3 February 2011 / Published online: 15 May 2011
© The JJIAM Publishing Committee and Springer 2011
Abstract We consider a family of discrete multiperiod multinomial market models
F
n
, each of which contains n - 1 stocks and one bond. All the securities are allowed
to be risky and we assume that the number of states in each period is finite. We
let the securities’ prices follow probability distributions that reflect the traders’ view
of the market. Under mild restrictions on the probability structure of F
n
, we show that
the probability that a market, chosen at random from F
n
, is complete tends to one as
n approaches infinity.
Keywords Market completeness · Large markets · Single period model · Multiperiod
model · General linear groups over a finite field
Mathematics Subject Classification (2000) 91G99 · 60B20
1 Introduction
The completeness of real-world security markets has been the subject of much the-
oretical and empirical study over the years. Several papers (such as [1, 3, 5]) reject
the hypothesis of real-world market completeness statistically. Indeed, the assump-
tion of market completeness has been blamed for the poor predictive power of many
J. A. Wright
Department of Mathematics, The University of Hong Kong, Hong Kong, China
P. S. C. Yam
Department of Statistics, The Chinese University of Hong Kong, Hong Kong, China
H. Yang (B )
Department of Statistics and Actuarial Science, The University of Hong Kong, Hong Kong, China
e-mail: hlyang@hkusua.hku.hk
123