The impacts of avian flu on the poultry
Download
Report
Transcript The impacts of avian flu on the poultry
The impacts of avian flu on the poultry-related
stocks listed in the US public stock markets
Wei Huang
David Bessler
Texas A&M University, College Station
September 28, 2008
Outline of Presentations
•
•
•
•
•
Introduction
Theoretical analysis
Empirical methodologies
Empirical results
Conclusions
Introduction
Research Objectives
This study plans to explore the impacts of recent outbreaks of avian
influenza (AI) on security values of poultry-related firms listed in
the United States (US) stock market, especially the different impacts
between upstream and downstream poultry-related firms, and
between AI outbreaks in US and out of US.
Previous Studies
• Explore possible effects of AI outbreaks on trade,
and industrial, regional or national economy
(Djunaidi et al., 2007; Brown et al., 2007;
Paarlberg et al., 2007)
• Study influences of BSE outbreaks on stock prices
in the meat and other related industries in the
United Kingdom (UK) and the United States.
(Henson et al, 2002; Jin and Kim, 2008)
Theoretical Analysis:
demand - supply analysis for poultry
meat(egg)/cooked food of a producer
Demand and supply for a producer’s poultry
meat(eggs) before and after AI outbreaks
outside US
Pu
S1u
P2u
u
P1
D1u
D2u
Qu
Q1u
Q2u
Demand and supply for a producer’s poultry
meat(eggs) before and after AI outbreaks
inside US
Pu
S2u
S1u
P1u
D1u
P2u
D2u
Q2u
Q1u
Qu
Demand and supply for a producer’s poultry
cooked food before and after AI outbreaks
outside US
Pdus
S2dus
ΔSdus
P1dus
S1dus
D1dus
Q2dus
Q1dus
Qdus
Demand and supply for a producer’s poultry
cooked food before and after AI outbreaks
inside US
• Price decease effect > quantity
decrease effect
Pdus
• Price decease effect < quantity
decrease effect
S1dus
Pdus
ΔSdus
P1dus
S2dus
ΔSdus
D1dus
Q1dus
Q2dus
S2dus
P1dus
s
D1dus
Q2dus
Qdus
S1du
Q1dus
Qdus
Summary on theoretical impacts of AI outbreaks
on poultry-related stock prices
Location of AI outbreaks
Firm type
Stock price behavior
Outside the country
Poultry meat/eggs producers
Stock price
Inside the country
Poultry meat/eggs producers
Stock price
Outside the country
Poultry food producers
Stock price
Inside the country
Input price effect >
input quantity effect
Stock price
Input price effect <
input quantity effect
Stock price
Poultry food producers
Empirical Methodologies:
historical decomposition
Historical decomposition
• Here we write the time series stock price vector X in its
moving average form (MAR)
Xt
e
i 0
i
(3)
t i
• Where the vector X is written as an infinite series of
orthogonalized innovations, et-i. From Equation (3), we
can calculate a historical partition of the vector X at any
date T+k into information available at time t = T and
information which is revealed at period t = T+1, T+2, … ,
T+k. We can write the vector X at period T+k as:
k 1
X T k s eT k s [ s eT k s ]
s 0
s k
(4)
How to obtain MAR for Historical decomposition?
• Through (cointegrated) Vector Autoregression form (VAR)
• The VAR can be illustrated using a set of m variables each
measured at time t; t= 1, 2, 3,…,T:
• xt' = (x1t, x2t, x3t, . . . , xmt); t= 1,2,3,…,T .
• This vector, xt , can be written as equation (1):
•
K
• (1) xt = Σ α(k)xt-k + et
k=1
•
• Here α(k) is an autoregressive matrix of dimension (mxm) at lag
k which connects xt and xt-k. K is the maximum lag in the VAR.
et is a vector residual term of dimension (mx1). The integer K is
large enough such that et is white noise.
Empirical Results
Data
• Data source: Yahoo finance
• Daily adjusted closing stock prices for five
firms: AFC Enterprises (AFCE), Cal-Maine
Foods (CALM), Industrias Bachoco (IBA),
Sanderson Farms (SAFM), and Tyson
Foods (TSN)
• Study period: Jun.1, 2001 to Jan.1, 2005
Summary on Firms
AFCE
CALM
IBA
SAFM
TSN
Business Type
Restaurant of
chicken foods
Egg producer,
processor and
packer
Chicken and
egg
producer,
processor and
packer
Chicken
producer,
processor and
packer
Chicken, pork
and beef
producer,
processor and
packer
Stock market
NASDAQ
NASDAQ
NYSE
NASDAQ
Market Cap.
250.15 M
668.44 M
1.44 B
759.37 M
6.03 B
1,820
1,536
22,983(07)
9,705
104,000
EBITDA
49.20 M
222.49 M
194.65 M
175.52 M
1.07 B
Net Income
23.10 M
133.59 M
120. 90 M
87.90 M
245.00 M
Operation
location
USA,
Columbia,
Puerto Rico,
Guam, and
internationally
USA
Mexico
USA (SE, WE,
W)
USA, export
Headquarter
location
Georgia,
USA
Mississippi,
USA
Celaya,
Mexico
Mississippi,
USA
Arkansas,
U.S.A.
Employees
NYSE
Timeline of AI outbreaks early 2004
Timeline
Country
Virus Type
January 8, 2004
Vietnam
H5N1
January 11, 2004
Japan
H5N1
January 20, 2004
Taipei China
H5N1
January 23, 2004
Thailand
H5N1
January 26, 2004
Cambodia, and Hong Kong
H5N1
January 27, 2004
Laos
H5N1
February 2, 2004
Indonesia
H5N1
February 4, 2004
China
H5N1
February 11, 2004
Delaware, USA
H5N2
February 19, 2004
Canada
H7N2
February 23, 2004
Texas, USA
H7N2
March 15, 2004
Canada
Estimated results of cointegrated VAR
model
PCALM 0.008
PCALM
0.25 0.002
P 0.006
P
0.15 0.004
IBA
IBA
PSAFM 0.004 0.004 0.34 1.01 2.01 1.00 PSAFM 0.58 0.001PSP500 1t 1 t
P
P
0
.
012
0
.
63
0
.
006
TSN
TSN
PAFCE 0.009
PAFCE
0.68 0.004
t
t 1
Historical decompositions for Cal-Maine Foods from 8
January 2004 to 10 February 2004
Here we see information arising in CALM after
January 15 is “pushing” the CALM price up;
whereas information arising in AFCE after
January 12 is pulling CALM price down.
Historical decompositions of Industrias Bachoco from 8
January 2004 to 10 February 2004
Here we see information in IBA after January 15
is pulling the price down; while information
arising in AFCE after January 15 is pushing the
price up.
Historical decompositions of Sanderson Farms from 8
January 2004 to 10 February 2004
Information arising in Sanderson farms after
early February is pulling Sanderson Farms
price down; whereas information arising in
TSN after January 28 is puling Sanderson
farms stock price up.
Historical decompositions of Tyson Foods from 8 January
2004 to 10 February 2004
Here the up-ward movement in TSN after
January 23 is due to information arising in its
own price (own errors are positive) and
information arising in AFCE.
Historical decompositions of AFC Enterprises from 8
January 2004 to 10 February 2004
Here the upward movement in AFCE after
January 13 appears to all be due to
information arising in its own information
shocks – no contribution from information
arising first in other companies’ stock prices.
Historical decompositions of Cal-Maine Foods from 11
February 2004 to 15 March 2004
CALM appears to be exogenous over post
February 11 data.
Historical decompositions of Industrias Bachoco from 11
February 2004 to 15 March 2004
IBA appears to be exogenous over post
February 11 data.
Historical decompositions of Sanderson Farms from 11
February 2004 to 15 March 2004
The downward movement in SAFM equity price
after February 11 is generated by information first
noted in its own equity prices (its own errors) and
those of TSN. SAFM shocks later on appear to
move SAFM prices upward (say after February 13).
TSN information continues to have a negative affect
on SAFM well into late February.
Historical decompositions of Tyson Foods from 11
February 2004 to 15 March 2004
The negative movement in TSN after
February 11 appears to be self generated.
Historical decompositions of AFC Enterprises from 11
February 2004 to 15 March 2004
AFCE appears to be exogenous over post
February 11 data.
Conclusions
• the AI outbreaks in Asia raised the stock prices of CalMaine Foods, Sanderson Farms, Tyson Foods and AFC
Enterprises;
•
the AI infections in the United States dropped the stock
prices of Industrias Bachoco, Sanderson Farms and Tyson
Foods and raised stock prices of Cal-Maine Foods and
AFC Enterprises.
• Most of empirical findings are consistent with the
theoretical results. AI outbreaks appear cause volatility of
firm values through its effects on international trade.
Thank You