Transcript Non-fundamentals and Stock Market: New Evidence

Non-fundamentals and Stock
Market: New Evidence
Moshfique Uddin (LUBS)
Anup Chowdhury (TYMS)
Keith Anderson (TYMS)
Introduction and Motivation
• There is growing evidence that the standard
valuation model has failed to capture the
stock market movements.
• Therefore, it is important to identify and
evaluate drivers of volatility other than
conventional dividends and earnings, such as,
national elections, political uncertainty,
government policy, regulatory changes, along
with monetary and fiscal policies.
2
Introduction and Motivation
• The aim of this paper is to provide new evidences by
analysing the reaction of an emerging stock market to nonfundamental factors. Particularly, in a first step, we
investigate whether stock prices behaviour of this market
changes over time or switches over states with respect to
timing of monetary policy, fiscal policy, political events,
national election, changes in government policies and
changes in capital markets regulations.
• In second step, we extend our analysis to provide firm-level
evidence related to this interdependence. We examine
whether any of the macro and non-macroeconomic news
has specific effect on portfolios with different
characteristics, i.e. size, dividend yield and sectors.
3
Introduction and Motivation
• The stock market i.e. Dhaka Stock Exchange (DSE)
is one of the fastest growing equity markets of
this region and named as one of the best
performing markets in the world (see The
Economist, 2011; Rintoul, 2012).
• The economy and the stock market in Bangladesh
have some interesting characteristics, which are
distinctly different than most other developed
and emerging economies.
4
Introduction and Motivation
• From 1990-2011, it has had a system of interim nonpolitical governments (commonly known as Caretaker
Government (CG)) between politically elected
governments. The CG has presented three national budgets
and conducted four national elections.
• The political government of this country is very powerful
and intervene into the equity market, such as to increase
the liquidity government has allowed the black money
(undisclosed) in the stock market.
• Bangladesh is a commonplace of political uncertainty. Since
1990 there have been more than 1100 days (equal to about
4 working years) of nation-wide strike (commonly known as
hartal) called by political parties till 2012 (UNDP, 2005).
5
Introduction and Motivation
• As a Muslim country, Bangladesh has stocks of different
sets of industries or firms listed other than conventional
operation, e.g. banks with Islamic banking norms. Chau et
al., (2014) identify that conventional and Islamic financial
market indices react heterogeneously to the political
turmoil and the volatility of Islamic indices significantly
increases during the period of political unrests.
• More than 90 percent equity investors are individuals in
Dhaka Stock Exchange (DSE), who usually have less capacity
to diversify their portfolios may be due to small scale of
investment and lack of knowledge about the capital
market. Hence, they are very sensitive to any shocks or
surprise, particularly to any form of negative information.
6
Monetary Policy shocks and the capital
market
• Equity market performance not only responds to monetary
policy decisions and affect the economy but also provide
feedback effect to the central bank regarding the private
sector’s expectations about the future course of key
macroeconomic variables (Mishkin, 2001).
• Monetary policy can influence the asset prices (e.g. stock
market) via different channels, such as via changes in cost
of capital, via subsequent changes in the investment
opportunity set faced by firms, via other mechanisms, such
as exchange rate and transfer of funds between stock and
bond markets and adjusting investors’ expectation (see
Tobin, 1969; Mishkin, 2001, Bernanke and Kuttner, 2005;
Chatziantoniou et al., 2013).
7
Fiscal Policy shocks on equity markets
• According to Keynesian approach fiscal policy can support
aggregate demand, boosting the economy and thus positively
contributing to the financial market.
• They argue that contractionary and expansionary fiscal policy
shocks related to government spending, deficit financing, public
sector investment and tax policy could directly influence equity
markets.
• Darrat (1988, 1990) in his two papers empirically tested stock
market efficiency with respect to both monetary and fiscal policy
variables. He asserts that fiscal policy stance plays a significant role
in determining stock returns even when the path through interest
rates is excluded. In a separate study, Darrat and Brocato (1994)
claim that the Federal budget deficit exerts a significant lagged
impact on current US stock returns.
8
Political and other non-macroeconomic
shocks on stock markets
• Empirical evidence strongly suggests that the stock market
is not only influenced by monetary and fiscal policies but
also by other macro and non-macroeconomic factors, such
as, political uncertainty, national elections, changes in rules
and regulations related to capital markets etc.
• Wang and Theobald (2008) examine the regime-switching
behaviour of equity return volatility of six East Asian
markets following the introduction of liberalization policies
in the mid to late 1980s and early 1990s. Their model
detects two or three volatility states and results suggest
that the switching between regimes is associated with
international and country-specific events, such as – Asian
financial crisis, political instability, failed military coup
attempts, Gulf-war and oil price shocks.
9
Political and other non-macroeconomic
shocks on stock markets
• Bialkowshi et al. (2008) investigate stock market volatility
around national election from 27 OECD countries. They
document that the index return variance become double
during the week around an election and stock markets can
become very unsettled during the period of important
political changes. Mei and Guo (2004) also observe
increased market volatility during political election and
transition periods.
• Bengtsson et al. (2014) investigate a series of enforcement
actions taken by the Securities and Exchange Commission
(SEC) of the US and their results indicate enforcement
actions influence the public equity investors and thus stock
prices.
10
Firm level evidences of economic
shocks
• It is documented in earlier literature that firms are not
homogeneously affected by news. For example,
Wasserfallen (1989) explains that the effect of
macroeconomic events may also depend on characteristics
specific to a firm or an industry, such as, the amount of
international trade, inflation, changes in money supply etc.
• In one of the early study, Gertler and Gilchrist (1994)
suggest that monetary policy should have disproportionate
impact on borrowers with limited access to capital markets,
everything else equal. They argue that small firms are
strongly affected by monetary policy shocks since they are
likely to face more constrained in financial markets for
borrowing.
11
Firm level evidences of economic
shocks
• Ehrmann and Fratzscher (2004) explain that the effect
of monetary policy on stock market returns is likely to
differ across industries for various reasons; firms in
cyclical industries, capital-intensive industries, and
industries that are relatively open to trade are affected
more strongly.
• Dedola and Lippi’s (2005) use 21 manufacturing sectors
from five OECD countries (i.e. France, Germany, Italy,
the UK and the US). Their analysis further reveals that
the impact of monetary policy is stronger in industries
that produce durable goods, have greater financing
requirements, lower borrowing capacity and small size.
12
Data and Analysis
• 𝑅𝑡 = 𝛼0 + 𝑎𝑖=1 𝜃𝑖 𝑅𝑡−𝑖 + 𝑏0 𝑅𝐿,𝑡−1 + 𝑏1 𝑅𝑅,𝑡−1 + 𝑏2 𝑅𝑊,𝑡−1 +
𝑏3 𝑅𝑆𝑡 + 𝑢𝑡
• Where, 𝑅𝑡 is the daily return of the DSE all-share price index, 𝑅𝐿 ,
𝑅𝑅 , and 𝑅𝑊 are the daily return of local, regional and world equity
indices over the sample period. The autoregressive terms 𝑅𝑡−𝑖 , are
included in the return equation to account for the problem of
autocorrelation potentially induced by nonsynchronous trading,
which is particularly severe in emerging markets given their low
level of liquidity (see Lee and Rui, 2001). 𝑅𝑆𝑡 is day-of-the-week
dummy for Sunday, as Chowdhury, Uddin and Anderson (2013) have
identified a Sunday effect for the DSE all-share price index. We use
the residual {𝑢𝑡 } as our new filtered return series for this analysis
and renamed it as 𝑅𝑡 .
13
Data and Analysis
• We use Markov Regime Switching GARCH
(MS-GARCH) model of Haas et al., (2004) to
check the robustness of structural breaks and
interdependence between economic
information and stock markets.
14
Data and Analysis
• In order to discuss the influence of macro and non-macro
information on daily returns and variance of firms based on size,
dividend and sectors, the empirical model assumes returns (𝑅𝑡 )
follow the following process, similar to equation (ii), (iii) and (iv)
• 𝑅𝑡 = 𝛼0 + 𝑎𝑖=1 𝜃𝑖 𝑅𝑡−𝑖 + 𝑎𝑖=1 𝜗𝑖 𝑑𝑖 + 𝜀𝑡 ,
(xiii)
• 𝜀 2 𝐼𝑡−1 ~ 𝐺𝐸𝐷(0, ℎ𝑡 )
(xiv)
• ℎ𝑡 = 𝜔 +
𝑏
𝑖=1 𝜉𝑖 𝜍𝑖
+
𝑝
𝑖=1 𝛽𝑖
ℎ𝑡−𝑖 +
𝑞
2
𝛼
𝜀
𝑗
𝑡−𝑗
𝑗=1
+
𝑞
𝑗=1 𝛾𝑗
−
2
𝑍𝑡−𝑗
𝜀𝑡−𝑗
(xv)
• Where, 𝑅𝑡 is the filtered return of each of the value-weighted
indices for firm characteristics; 𝑑𝑖 and 𝜍𝑖 are the dummy variables
for macro and non-macro variables in mean and volatility equations
respectively. The significance of 𝜗𝑖 and 𝜉𝑖 implies the reaction of
each portfolio to each of the macroeconomic and non-macro
information around the structural breaks or regimes switching.
15
Data and Analysis
• For detecting the structural breaks and regime shifts
we use daily index of DSE (Dhaka Stock Exchange) allshare price from Datastream over a period from 1
January, 1990 till 31 December, 2012.
• The daily market capitalization, dividend yield and
market price for each of the 265 firms are also
collected from Datastream. However, firm level data is
only available in Datastream from 1 January 2000 till 31
December 2012; therefore we take this dataset which
includes more than one million observations.
16
Data and Analysis
• We divide all the 265 firms into four sectors, namely,
manufacturing, service, financial and miscellaneous and
the value weighted index for each of the sector is
calculated based on the algorithm given in the Dhaka Stock
Exchange.
• Finally, we also consider international benchmark indices to
proxy for the world, regional and local influences, those
indices are – MSCI (i.e. Morgan Stanley Capital
International) World, MSCI Emerging Market, MSCI
Emerging Market Asia and Industry specific (i.e. Financial
Sector, Manufacturing and Service) indices from MSCI
World and MSCI Emerging Market. Macroeconomic and
non-macroeconomic events, which are considered over the
sample period for this study, are hand collected.
17
Main Findings
• Results indicate that DSE is highly sensitive to
political uncertainty, electoral system, money
growth policy and government debt policy.
• Interestingly, market has shown strong
confidence and less volatility during the period of
caretaker government system. This finding is
relatively novel because prior research in
emerging markets, such as Diamonte et al.
(1996), Bilson et al. (2002) and Chau et al. (2014)
do not empirically examine the stock market
response to electoral system.
18
Main Findings
• This paper provides new evidence on the interdependence
between firm characteristics (i.e. size, dividend yield and
industry) and non-macroeconomic factors (i.e. political risk,
national election, electoral system and regulatory changes),
where the empirical substantiations are limited, particularly
from emerging market it is rare.
• Using modified GJR-GARCH model we find top and bottom
20% firms are sensitive to any information; smaller firms
are significantly influenced by the changes in monetary
policy variables; political uncertainty of 2007 has greater
impact on the market; financial and manufacturing sectors
are more sensitive to both macro and non-macro news.
19
Main Findings
• the impact of lagged regional and world
market index (i.e. 𝑏0 to 𝑏2 ) is not found
significant for most of the portfolios.
• However, local and regional factors are
reported significant only for DSE itself, top
10% firms, and financial industry. This result,
therefore, support the argument that market
exhibit strong home bias.
20
Main Findings
• Estimated dates for structural changes in DSE
conditional variance are 22 November 1996, 25 July
2000, 21 September 2004, 9 January 2007 and 15
September 2010. The structural breaks dates for equity
returns, which are on 21 November 1996 and on 12
April, 2005 respectively.
• We link the breakpoints with the daily price index of
DSE over the sample period 1990-2012. It shows that
the break dates coincide with the major price
fluctuations of this stock market, such as two big
slumps those happened in DSE in 1996 and 2010 and
other two minor crashes those happened in 2005 and
2008.
21
Main Findings
• Financial sector is mostly affected by the information
under this study; which is probably due to their strong
involvement with the capital market and methods of
their business
• Changes in exchange rate regime (i.e. from pegged to
floating system in 2003) do not change the price
behaviour significantly.
• The impact of CRR is stronger to control the excess
money supply in Bangladesh. Particularly, the effect of
raising the CRR has a relatively higher impact on
investment, such as in capital market.
22
Main Findings
• The government of Bangladesh allows the black
money into stock market to satisfy the investors,
which initially increases the demand of stocks
and the price but, unfortunately, that growth
does not sustain in the long run.
• Government pressure on the central bank and
SEC to stabilize the market after plunge of 2010
did not work
• Unfortunately, in Bangladesh, remittance only
raises the size of the market but not the
efficiency.
23
Many thanks for your time.
For any further information
Please contact
Moshfique Uddin
Leeds University Business School, UK
[email protected]
24