Keio
Symposium on Risk Assessment

Date: September 21st, 2016

Place: BLDG 14 Room 201, Yagami campus, Keio
University

Access (English) http://www.st.keio.ac.jp/english/access/index.html

(Japanese) http:/www.st.keio.ac.jp/access/index.html

10:25 - 10:30 Opening

10:30 - 11:00 Hiroshi Shiraishi, Keio University, Faculty of Science and
Technology

"Nonparametric estimation for optimal dividend barrier based on Laplace
transformation"

11:00 - 11:30 Junichi Imai, Keio
University, Faculty of Science and Technology

"Enhancement
of QMC: severity and tail dimension乭

11:30 - 11:50 Taka-aki Koike, Keio University, Graduate school of Science and
Technology

"Efficient Computation of Risk Contributions by using
MCMC乭

11:50 – 13:00 Lunch Break

13:00 - 14:00 Paul Embrechts, ETH Zürich, Department of Mathematics

"Bernoulli and tail-dependence compatibility"

14:00 - 14:20 Coffee Break

14:20 – 14:50 Takuji Arai, Keio University, Department of Economics

"Convex risk measures for cadlag processes on Orlicz
hearts"

14:50 - 15:20 Takaki Hayashi, Keio University, Graduate School of
Business Administration

"On lead-lag analysis with
high-frequency data"

15:20 - 15:30 Break

15:30 – 16:00 Mihoko Minami, Keio University, Faculty of Science and
Technology

"Doubly Cyclic Smoothing Splines and Analysis of Seasonal Daily
Pattern of CO2 Concentration in Antarctica"

16:00 - 16:20 Tomoshige Nakamura, Keio University, Graduate school of
Science and Technology

"The Problem of Treating Imputed Data as Observed Data
When We Estimate the Effect of Exposure to Particulate Matter乭

16:20 - 16:40 Ayato Kashiyama, Keio University, Graduate school of
Science and Technology

"Annual Maximum Rainfall Analysis Using Extreme Value
Theory乭

16:40 - 16:45 Closing

Abstracts

Professor Paul Embrechts, ETH

Title: "Bernoulli and
tail-dependence compatibility"

Abstract: Based on a practical
example from stress testing within a solvency study of an insurance company, in
this talk I will present a so-called inverse-dependence problem. By this I mean
a general class of problems where a specific dependence structure between the
various components of a vector of risks is given; based on this information,
one has to characterize the risk vectors exhibiting this dependence structure,
if at all such vectors exist. In particular for the current example, given a
symmetric dxd matrix with [0,1]
entries, can this matrix result as the lower (or upper) asymptotic
tail-dependence coefficient matrix of a d-dimensional risk vector? A full
solution of this problem is given. It is shown that this problem is closely
related to the determination of matrices of second order cross-moments of
general Bernoulli vectors.

Title: Convex risk measures for cadlag processes on Orlicz hearts

Abstract: Our purpose is to
study properties and representations of convex risk measures for possibly
unbounded cadlag processes. As the underlying space
on which we define convex risk measures we consider spaces of cadlag processes whose supremum belongs to an Orlicz heart. In order to obtain concrete representations
for such convex risk measures, we shall investigate representations of
continuous linear functionals on the underlying
space. Moreover, many examples of such risk measures are introduced. In
particular, we deal with risk measures associated with hedging and pricing
problems for American claims. Among others, we look into shortfall risk measure
in detail.

Title: On lead-lag analysis with
high-frequency data

Abstract:Lead-lag analysis in the high-frequency
domain has not been drawn attention more than it deserves in the literature,
albeit its potential importance in practice. The talk will first introduce
approaches proposed recently by some researchers, then will present our own (H.
and Koike (2016)). As an empirical application of these methodologies, we
explore lead-lag relationships between market prices of identical stocks traded
concurrently on multiple trading venues. Results on such "cross-market,
single-asset" analyses with Japanese stock market datasets will be
presented. (Joint work with Yuta Koike, Tokyo Metropolitan
University.)

Key words: high-frequency data;
lead-lag analysis; limit order book; nonsynchronicity;
quadratic covariation; timestamp.

Junichi Imai

Title: Enhancement of QMC:
severity and tail dimension

Abstract: In this talk we
discuss how to further enhance numerical efficiency of a quasi-Monte Carlo method.It is well-known that the quasi-Monte Carlo method
often outperforms a Monte Carlo method, especially in applying it to lower
dimensional problems.In the paper, we introduce new
notions; severity control, tail dimension control,and
provide new principles, based on the notions, that enable us to improve the
numerical efficiency of the quasi-Monte Carlo methods.These
principles are useful when a function of interest has a complex form and/or it
contains multiple discontinuity/non-differentiability.

Title: Annual Maximum Rainfall
Analysis Using Extreme Value Theory

Abstract: In recent years,
natural disasters occur frequently caused by extreme weather events. Extreme
value theory aims at modeling maximum or minimum data, and in meteorological
data, such data corresponds when natural disaster occurs. In this presentation,
I will talk about extreme value theory and show an analytical result of annual
maximum daily rainfall data from a region of Japan.

Taka-aki Koike

Title: Efficient Computation of
Risk Contributions by using MCMC

Abstract: In most of financial
institutions, the risk of their portfolios is measured by the economic capital.
For the purpose of more detailed risk analysis, it is necessary to decompose
the portfolio-wide economic capital into the sum of risk contributions by unit
exposures. Despite high practical demands, computing the risk contributions is
a challenging task in general. No explicit solutions are available for most
risk models. In this talk, we will introduce a Markov chain Monte Carlo
(MCMC)-based estimator of risk contributions when economic capital is computed
by Value-at-Risk. We will demonstrate that the estimator is available and
high-performing in a wide variety of risk models.

Title: Doubly Cyclic Smoothing
Splines and Analysis of Seasonal Daily Pattern of CO2 Concentration in
Antarctica

In order to flexibly estimate a continuously varying daily pattern throughout a year, we propose a doubly cyclic smoothing spline method and present its properties. We apply the method along with other tensor product smoothing methods to estimate the daily pattern of CO2 concentration at Syowa station in Antarctica and the effects of wind direction and velocity to CO2 concentration.

Tomoshige Nakamura

Title: The Problem of Treating
Imputed Data as Observed Data When We Estimate the Effect of Exposure to
Particulate Matter

Abstract: When we estimate the
effects of exposure to particulate matters, community health survey data are
often used. In such a case, information of the amount of the exposure of each
subjects need to be provided. However, in many cases, the amount of exposure of
some subjects cannot be observed, so estimated values computed by some methods
are imputed to them, and the effects of particulate matter is analyzed as if
they were observed. In this talk, I will discuss problems of estimating the
effect of exposure to Particulate Matter using imputed values as observed
values, which have not been paid much attention by environmental
epidemiologists.

Hiroshi
Shiraishi

Title: Nonparametric Estimation
for Optimal Dividend Barrier based on Laplace Transformation

Abstract: Dividends are defined
as premium income whenever the insurance surplus attains a barrier level. Under
the aggregate claims process taken as a compound Poisson model, optimal
dividend barrier is defined as a barrier level that maximizes the expectation
of the discounted dividends until ruin. We derive the optimal dividend barrier
as a solution of a function following Gerber and Shiu
(1997, 1998). Then, we consider the non-parametric estimation based on the empirical
version of the Laplace transformation.